Original Research

Acute respiratory tract (ART) infections, such as common cold, bronchitis, pharyngitis, sinusitis, and otitis media, are among the most common problems seen in primary care practice.¹ Unnecessary use of antibiotics for these infections is a major worldwide problem both in terms of cost² and as a contributor to the development of antibiotic-resistant bacteria.³

Although there is some evidence that physicians misdiagnose many viral infections as bacterial,^4,5 recent studies suggest that the reasons for unnecessary antibiotic prescribing are more complex, having as much or more to do with patient and physician expectations as with physicians’ diagnostic skills.^6-8 These studies are limited to describing perceptions of behavior rather than actual behavior, because of their use of interview and focus group data. Consequently, we do not know what actually happens during outpatient visits for ART infections that leads to antibiotic prescribing.

Two studies by Stivers^9,10 underscore the importance of directly observing what transpires during encounters with pediatric ART infection patients. Stivers’ examination of videotaped visits found that, in some cases, parental pressure for antibiotics influenced the physician’s decision to prescribe. This finding has not been replicated, however, in family practice settings, where both adults and children are seen. We used direct observation of outpatient visits to family physicians for ART infections to analyze the effects of physician-patient communication on unnecessary antibiotic prescribing. By understanding the ways these communication patterns influence prescribing behavior, practicing family physicians can develop strategies to deliver more appropriate care for ART infections.

Methods

These data were collected as part of the Prevention & Competing Demands in Primary Care Study, which was an in-depth observational study begun in October 1996 and completed in August 1999 that examined the organizational and clinical structures and process of community-based family practices. Each of 18 purposefully selected practices was studied using a multimethod comparative case study design that involved extensive direct observation of clinical encounters and office systems by field researchers who spent 4 weeks or more in each practice. Field researchers directly observed and dictated descriptions of approximately 30 patient encounters with each of the more than 50 clinicians. Details of the sampling and data collection are available elsewhere in this issue.¹¹

Data Analysis and Interpretation

Encounters related to ART infection were identified in the database using search terms for symptoms and diagnoses including: sore throat, runny nose, congestion, cough, drainage, postnasal drainage, earache, cold, upper respiratory infection, pharyngitis, sinusitis, bronchitis, and otitis. ART infection was identified as the principal or associated diagnosis in 316 outpatient visits of a total of 1637 observed encounters; 298 had sufficiently rich data for analysis. The encounters were first coded for antibiotic use or nonuse.

Before any qualitative analysis began, visits during which antibiotics were prescribed were further characterized as appropriate or unnecessary according to guidelines by the Centers for Disease Control and Prevention (CDC) for judicious use of antibiotics for children¹² and adults.Table W1¹³* Two family physicians assigned appropriate/unnecessary codes independently. Inter-rater reliability was good (k=0.71). All disagreements were resolved by discussion.

Subsequently, the text for each outpatient visit was read independently by 2 family physicians, a medical anthropologist, a nurse, and a communication specialist. This research team discussed individual encounters as a group to identify emerging patterns of physician-patient interaction.

Results

Women made up 59% of the study sample; 64% were 16 years or older (and classified as adults). Antibiotics were prescribed in 204 of the 298 ART infection encounters (68%). Antibiotic use was unnecessary according to the CDC guidelines in 164 of these (80%). Adults were more likely than children to receive unnecessary antibiotics Table 1.

Our analysis identified 6 different types of patient behaviors that advocated for medication, particularly antibioticsTable 2. These behaviors fell into 3 broad categories: explicit requests, presentation of chief complaint, and appeals to lifeworld circumstances. Multiple pressures were noted in many encounters.

While patients occasionally made direct requests for antibiotics, they much more frequently positioned themselves indirectly for receiving antibiotic treatment by the way they presented the chief complaint. Four distinct approaches were identified: symptoms only,⁹ candidate diagnosis,⁹ implied candidate diagnosis,⁹ and portraying the severity and inability to shake the illness.

A second category of indirect approach used life-world circumstances¹⁰ (eg, an upcoming family vacation) or a past history with successful antibiotic treatment to formulate appeals for antibiotics in the current encounter. In those cases in which antibiotics were clearly unnecessary, physicians often rationalized their prescribing practices by finding symptoms or assigning diagnoses to justify antibiotic use. Each of these patient pressures, as well as the physician-rationalizing behavior, is illustrated with sample visits. The samples are taken directly from transcribed field notes, but the names have been altered to protect the identity of patients and clinicians.

Explicit Request

Explicit requests for antibiotics were observed in only 6% of cases (n=15). For example:

Claire asked the patient, “How are you doing?” and she said, “Well, I’m coughing up phlegm, I ache and I have chills and a sore throat.” Claire said, “You have bronchoconstriction, and 3 times a day, if you need to, you should use proventil.” The patient asked if she could have an antibiotic for her cold; cephalexin has worked in the past. Claire said that she would get her cephalexin and also some samples of an inhaler.

Presentation of the Chief Complaint

Patients frequently put pressure on the physician for treatment during the presentation of the chief complaint, the exception being the symptoms-only presentation. This is different from the other indirect pressures, which usually occurred during different parts of the medical encounter.

Symptoms-only presentation (eg, “I have a cough and a sore throat.”) In the symptoms-only approach (n=15), the patient reports his or her symptoms with little embellishment. This approach does not pressure physicians for antibiotic treatment.

Candidate diagnosis (eg, “I think I’ve got strep throat.”) In contrast, patients also presented their chief complaint to the physician by offering a candidate diagnosis (n=18). As shown in the following example, the patient responds by offering a diagnosis. This is a way of indirectly advocating for antibiotic treatment.

A 21-year-old white woman went to see Dr. Maxwell with an acute problem of congestion. Dr Maxwell said, “Well, how are you doing?” The patient said, “It sounds like bronchitis. It started about 4 days ago.”

Implied candidate diagnosis (eg, “My throat hurts; it’s red; and it has white spots.”) The implied candidate diagnosis is a hybrid of the symptoms-only and the candidate diagnosis approaches (n=48). When presenting their chief complaint, patients reported very specific symptoms that indexed a particular diagnosis. For example:

A 29-year-old woman went to see Dr Redmond with swollen glands, congestion, and white spots on her throat. When Dr Redmond and I went into the examination room, the patient had a pink paper top on, and Dr Redmond told her that her throat culture was negative.

The patient reports that she has swollen glands, congestion, and white spots on her throat. The symptoms specifically index a particular condition (strep throat). The patient’s presentation of symptoms clearly implies a diagnosis of strep throat, and the physician ordered a strep culture before seeing the patient.

Candidate diagnoses and implied candidate diagnoses delicately assert that the nature of the patient’s problem is already known. The reason for the medical visit is to seek treatment for the patient’s already known condition. When candidate and implied candidate diagnoses point to a condition the patient believes to be treatable (eg, bronchitis, strep throat, ear infection), this way of presenting the chief complaint looks directly ahead to a treatment involving a prescription for an antibiotic and thus indirectly pressures the physician to prescribe one.

Portraying the severity of one’s illness (eg, “I can’t shake this, Doc.”) The most common strategy was for patients to subtly pressure physicians for medication by portraying the severity of their condition and their inability to shake the illness on their own (n=99). For example:

The patient was sitting up on the table, and right away he told Dr Lamont, “I just can’t shake it. I feel like the back of my throat has raw hamburger hanging in it.” Dr Lamont checked the patient’s throat well, and the patient said, “This has lasted 4 days and it has been getting worse today.” Dr Lamont checked the patient’s ears, glands, and lungs. “I’m going to give you a shot of penicillin, slow release. It’s some kind of an infection. It may be a virus.”

Portraying the severity of one’s illness may not in and of itself advocate for medication; however, portrayals of the severity of one’s condition were usually accompanied by other actions implicating the need for medication. By opening the encounter with the announcement “I just can’t shake it,” the patient implies that he needs help in getting well. This subtly suggests the need for a prescription medication to alleviate his sore throat. At the end of this visit, the patient receives an antibiotic shot.

Appeals to Nonmedical Circumstances

Patients also used nonmedical circumstances to advocate for medication. These behaviors tended to occur after the problem presentation in the encounter and either centered on some important event, such as a big examination or a trip out of town (n=16), or focused on a previous positive experience with antibiotics for themselves or a family member (n=39).

Appealing to life-world circumstance (eg, “But I’m going to Disney World.”) This patient uses an upcoming trip to make an appeal to the clinician to prescribe medication:

The patient is a 33-year-old man coming in with an acute problem of a sore throat. The patient stated that he had been trying to manage this on his own, but he was taking his wife and 2 children to Disney World at the end of the week and was becoming worried that he was still going be sick and not able to enjoy a trip that they had saved so long for. He also told Dr Liam: “I know we’ll just get to Florida, and the kids will get sick, and then we’ll all be sick again. Dr. Liam said, “Well, we can have you bring them in, but then we’d be treating them for something that they haven’t gotten. Let me think about this a bit.” He does the rapid strep test, and it’s negative. Dr Liam reported the news of a negative strep test and said, “Many times we get a 50% false-negative, so I’m gonna go ahead and put you on an antibiotic and see if we can’t get you feeling better.” With this the patient said, “Well, what do you think I should do about my kids?” Dr. Liam asked if the kids were seen in this clinic, and the patient responded that they had never been seen there before. Dr Liam said, “Well, I’ll go ahead and give you a script for erythromycin in case these kids get sick down in Florida. If they do, go ahead and give them the medicine; if they don’t, throw away the prescription.”

This case is interesting because once he is treated with an antibiotic, the patient uses the same argument to make an appeal for antibiotics for his children (both of whom have never been seen by this physician).

Previous positive experience with antibiotics (eg, “I got an antibiotic for this before.”). Patients also appealed to other nonmedical contingencies to advocate for antibiotic treatment. For example:

Our next patient was a 51-year-old woman complaining of a cold and laryngitis. The doctor asked the patient about her symptoms. The patient responded, saying that she had been taking medication during the end of December for the same symptoms; they had cleared after taking antibiotics, and now they were back again.

The patient indirectly makes an appeal for antibiotic treatment by stating that she received antibiotics in the past for the same symptoms that she has now.

Patients used several variations of this approach. These included stating that another physician prescribed an antibiotic for this illness in the past; that others in the family are sick with an illness for which they received antibiotics; that they have a history of illness for which antibiotics are regularly prescribed; and that they were recently taking an antibiotic for an illness that has not improved (with the idea that an antibiotic is needed again).

Effectiveness of Patient Pressures

Physicians prescribed an antibiotic unnecessarily in 80% of the encounters in which some patient pressure was observed. They seemed able to resist certain types of pressures better than others. Unnecessary antibiotics were prescribed for a smaller percentage of implied candidate diagnoses and candidate diagnoses and for a larger percentage of direct patient requests and previous positive experiences with antibiotics Table 3.

Physicians’ Response to Prescribing an Unnecessary Antibiotic

When physicians prescribed an antibiotic unnecessarily, they often rationalized this practice by finding symptoms or assigning diagnoses that, to them, justified prescribing antibiotics. Physicians used various rationales, such as red throat or enlarged tonsils; severe, prolonged, or productive cough; yellow or green mucus; sinus tenderness on palpation; associated chronic disease; history of previous infection; and the desire to “cover” the patient “just in case.” None of these rationales are supported by evidence as correlating with bacterial infection. An example of this kind of rationalization follows:

This is a 20-year-old woman coming in with a complaint of a worsening cough. She said that her chest had a prickly, burning sensation, and it hurt to breathe. Dr Hart asked if she was able to bring anything up. She said that she really couldn’t. It was just a really terrible barky cough. Following the physician examination, the physician told the patient that her lungs basically sounded clear, but she could certainly hear some rough bronchial sounds. With this, she said. “What I think is happening here with your cold is that it is probably ending up in a bronchitis-type situation, and probably what we should do is put you on an antibiotic and order a decongestant.”

Discussion

This investigation, in agreement with the pediatric studies of Stivers,^9,10 suggests that the connection between patient diagnosis and physician prescribing is highly complex, involving patient presentation and physician-patient communication as much as, if not more than, physician diagnostic skills. Also, these data suggest that physicians are better able to resist patient pressures that are framed in medical terms such as candidate diagnoses or implied candidate diagnoses but are much less able to resist pressures that are not medicalized, such as portraying severity of illness and use of life-world circumstances. Thus, it is not surprising that past interventions designed to increase physician knowledge regarding when to prescribe antibiotics have had limited success.^14,15 Physicians appear to be trying to maximize patient satisfaction by giving antibiotic-seeking patients what they want. Our findings show the need to modify current thinking about the diagnostic and treatment process to reduce the use of antibiotics. Rather than thinking of these processes as physician controlled, the powerful role patients play in this interaction must be considered.

Our study has important implications for future research. From a methodologic standpoint, our findings illustrate the importance of qualitative evaluation of directly observed medical encounters. The patterns of patient behavior observed could not have been discerned using survey, interview, or focus group data.

Limitations

Because these data were collected by field researchers who were unaware that ART infection would be a focus of our study, it is possible that there were other patient symptoms and behavior related to ART infection, as well as physician behaviors related to antibiotic prescribing, that were not recorded. The data were sufficiently rich, however, to easily and reliably apply the CDC guidelines for appropriate use of antibiotics. Any unrecorded behaviors might add to, but not substantially change, our conclusions that patients indirectly pressure their physicians for treatment, and physicians respond by giving antibiotics. Studies using videotaped encounters might uncover such additional important patient and physician behaviors. Since the patient population studied was limited to a single midwestern state, it is possible that other populations with a different ethnic or racial mix might behave differently. Future research in this area should attempt to include such populations. Finally, too few encounters per physician were observed in this study to evaluate whether particular physicians were high or low prescribers (such a pattern has been reported by De Sutter and colleagues¹⁶).

Conclusions

Physicians should be educated about the subtle approaches patients use to pressure them for antibiotic treatment and should be shown techniques for responding to these pressures without prescribing antibiotics unnecessarily. Our findings also suggest the need to increase patients’ awareness both of the dangers and lack of effectiveness of using antibiotics for ART infections and of the amount of influence that patients have on antibiotic prescribing. Macfarlane and coworkers17 have shown that use of patient education materials reduces visits for ART infection. Additional approaches to decreasing patient pressure for antibiotic prescriptions are needed to diminish antibiotic overuse and its public health consequences.

Acknowledgments

Our study was funded by the Agency for Healthcare Research and Quality Grant R01 HS08776. Dr Scott is a postdoctoral fellow supported by the Health Resources and Services Administration (HRSA) PE1011 and the Agency for Healthcare Research and Quality (AHRQ) HS09788. Analysis of these data was supported by a Research Center grant from the American Academy of Family Physicians (Center for Research in Family Practice and Primary Care). Drs Jaen and Crabtree are associated with the Center for Research in Family Practicer and Primary Care, Cleveland, New Brunswick, Allentown. and San Antonio. The authors wish to thank the family physicians of Nebraska who were willing to open their practices to us. We also thank Kurt C. Stange, MD, PhD, for his thoughtful comments on drafts of this manuscript.

Related Resources

U.S. Centers for Disease Control and Prevention—Promoting Appropriate Antibiotic Use in the Community http://www.cdc.gov/antibioticresistance/tools.htm

A vast resource of of patient education resources.

Acute respiratory tract (ART) infections, such as common cold, bronchitis, pharyngitis, sinusitis, and otitis media, are among the most common problems seen in primary care practice.¹ Unnecessary use of antibiotics for these infections is a major worldwide problem both in terms of cost² and as a contributor to the development of antibiotic-resistant bacteria.³

Although there is some evidence that physicians misdiagnose many viral infections as bacterial,^4,5 recent studies suggest that the reasons for unnecessary antibiotic prescribing are more complex, having as much or more to do with patient and physician expectations as with physicians’ diagnostic skills.^6-8 These studies are limited to describing perceptions of behavior rather than actual behavior, because of their use of interview and focus group data. Consequently, we do not know what actually happens during outpatient visits for ART infections that leads to antibiotic prescribing.

Two studies by Stivers^9,10 underscore the importance of directly observing what transpires during encounters with pediatric ART infection patients. Stivers’ examination of videotaped visits found that, in some cases, parental pressure for antibiotics influenced the physician’s decision to prescribe. This finding has not been replicated, however, in family practice settings, where both adults and children are seen. We used direct observation of outpatient visits to family physicians for ART infections to analyze the effects of physician-patient communication on unnecessary antibiotic prescribing. By understanding the ways these communication patterns influence prescribing behavior, practicing family physicians can develop strategies to deliver more appropriate care for ART infections.

Methods

These data were collected as part of the Prevention & Competing Demands in Primary Care Study, which was an in-depth observational study begun in October 1996 and completed in August 1999 that examined the organizational and clinical structures and process of community-based family practices. Each of 18 purposefully selected practices was studied using a multimethod comparative case study design that involved extensive direct observation of clinical encounters and office systems by field researchers who spent 4 weeks or more in each practice. Field researchers directly observed and dictated descriptions of approximately 30 patient encounters with each of the more than 50 clinicians. Details of the sampling and data collection are available elsewhere in this issue.¹¹

Data Analysis and Interpretation

Encounters related to ART infection were identified in the database using search terms for symptoms and diagnoses including: sore throat, runny nose, congestion, cough, drainage, postnasal drainage, earache, cold, upper respiratory infection, pharyngitis, sinusitis, bronchitis, and otitis. ART infection was identified as the principal or associated diagnosis in 316 outpatient visits of a total of 1637 observed encounters; 298 had sufficiently rich data for analysis. The encounters were first coded for antibiotic use or nonuse.

Before any qualitative analysis began, visits during which antibiotics were prescribed were further characterized as appropriate or unnecessary according to guidelines by the Centers for Disease Control and Prevention (CDC) for judicious use of antibiotics for children¹² and adults.Table W1¹³* Two family physicians assigned appropriate/unnecessary codes independently. Inter-rater reliability was good (k=0.71). All disagreements were resolved by discussion.

Subsequently, the text for each outpatient visit was read independently by 2 family physicians, a medical anthropologist, a nurse, and a communication specialist. This research team discussed individual encounters as a group to identify emerging patterns of physician-patient interaction.

Results

Women made up 59% of the study sample; 64% were 16 years or older (and classified as adults). Antibiotics were prescribed in 204 of the 298 ART infection encounters (68%). Antibiotic use was unnecessary according to the CDC guidelines in 164 of these (80%). Adults were more likely than children to receive unnecessary antibiotics Table 1.

Our analysis identified 6 different types of patient behaviors that advocated for medication, particularly antibioticsTable 2. These behaviors fell into 3 broad categories: explicit requests, presentation of chief complaint, and appeals to lifeworld circumstances. Multiple pressures were noted in many encounters.

While patients occasionally made direct requests for antibiotics, they much more frequently positioned themselves indirectly for receiving antibiotic treatment by the way they presented the chief complaint. Four distinct approaches were identified: symptoms only,⁹ candidate diagnosis,⁹ implied candidate diagnosis,⁹ and portraying the severity and inability to shake the illness.

A second category of indirect approach used life-world circumstances¹⁰ (eg, an upcoming family vacation) or a past history with successful antibiotic treatment to formulate appeals for antibiotics in the current encounter. In those cases in which antibiotics were clearly unnecessary, physicians often rationalized their prescribing practices by finding symptoms or assigning diagnoses to justify antibiotic use. Each of these patient pressures, as well as the physician-rationalizing behavior, is illustrated with sample visits. The samples are taken directly from transcribed field notes, but the names have been altered to protect the identity of patients and clinicians.

Explicit Request

Explicit requests for antibiotics were observed in only 6% of cases (n=15). For example:

Claire asked the patient, “How are you doing?” and she said, “Well, I’m coughing up phlegm, I ache and I have chills and a sore throat.” Claire said, “You have bronchoconstriction, and 3 times a day, if you need to, you should use proventil.” The patient asked if she could have an antibiotic for her cold; cephalexin has worked in the past. Claire said that she would get her cephalexin and also some samples of an inhaler.

Presentation of the Chief Complaint

Patients frequently put pressure on the physician for treatment during the presentation of the chief complaint, the exception being the symptoms-only presentation. This is different from the other indirect pressures, which usually occurred during different parts of the medical encounter.

Symptoms-only presentation (eg, “I have a cough and a sore throat.”) In the symptoms-only approach (n=15), the patient reports his or her symptoms with little embellishment. This approach does not pressure physicians for antibiotic treatment.

Candidate diagnosis (eg, “I think I’ve got strep throat.”) In contrast, patients also presented their chief complaint to the physician by offering a candidate diagnosis (n=18). As shown in the following example, the patient responds by offering a diagnosis. This is a way of indirectly advocating for antibiotic treatment.

A 21-year-old white woman went to see Dr. Maxwell with an acute problem of congestion. Dr Maxwell said, “Well, how are you doing?” The patient said, “It sounds like bronchitis. It started about 4 days ago.”

Implied candidate diagnosis (eg, “My throat hurts; it’s red; and it has white spots.”) The implied candidate diagnosis is a hybrid of the symptoms-only and the candidate diagnosis approaches (n=48). When presenting their chief complaint, patients reported very specific symptoms that indexed a particular diagnosis. For example:

A 29-year-old woman went to see Dr Redmond with swollen glands, congestion, and white spots on her throat. When Dr Redmond and I went into the examination room, the patient had a pink paper top on, and Dr Redmond told her that her throat culture was negative.

The patient reports that she has swollen glands, congestion, and white spots on her throat. The symptoms specifically index a particular condition (strep throat). The patient’s presentation of symptoms clearly implies a diagnosis of strep throat, and the physician ordered a strep culture before seeing the patient.

Candidate diagnoses and implied candidate diagnoses delicately assert that the nature of the patient’s problem is already known. The reason for the medical visit is to seek treatment for the patient’s already known condition. When candidate and implied candidate diagnoses point to a condition the patient believes to be treatable (eg, bronchitis, strep throat, ear infection), this way of presenting the chief complaint looks directly ahead to a treatment involving a prescription for an antibiotic and thus indirectly pressures the physician to prescribe one.

Portraying the severity of one’s illness (eg, “I can’t shake this, Doc.”) The most common strategy was for patients to subtly pressure physicians for medication by portraying the severity of their condition and their inability to shake the illness on their own (n=99). For example:

The patient was sitting up on the table, and right away he told Dr Lamont, “I just can’t shake it. I feel like the back of my throat has raw hamburger hanging in it.” Dr Lamont checked the patient’s throat well, and the patient said, “This has lasted 4 days and it has been getting worse today.” Dr Lamont checked the patient’s ears, glands, and lungs. “I’m going to give you a shot of penicillin, slow release. It’s some kind of an infection. It may be a virus.”

Portraying the severity of one’s illness may not in and of itself advocate for medication; however, portrayals of the severity of one’s condition were usually accompanied by other actions implicating the need for medication. By opening the encounter with the announcement “I just can’t shake it,” the patient implies that he needs help in getting well. This subtly suggests the need for a prescription medication to alleviate his sore throat. At the end of this visit, the patient receives an antibiotic shot.

Appeals to Nonmedical Circumstances

Patients also used nonmedical circumstances to advocate for medication. These behaviors tended to occur after the problem presentation in the encounter and either centered on some important event, such as a big examination or a trip out of town (n=16), or focused on a previous positive experience with antibiotics for themselves or a family member (n=39).

Appealing to life-world circumstance (eg, “But I’m going to Disney World.”) This patient uses an upcoming trip to make an appeal to the clinician to prescribe medication:

The patient is a 33-year-old man coming in with an acute problem of a sore throat. The patient stated that he had been trying to manage this on his own, but he was taking his wife and 2 children to Disney World at the end of the week and was becoming worried that he was still going be sick and not able to enjoy a trip that they had saved so long for. He also told Dr Liam: “I know we’ll just get to Florida, and the kids will get sick, and then we’ll all be sick again. Dr. Liam said, “Well, we can have you bring them in, but then we’d be treating them for something that they haven’t gotten. Let me think about this a bit.” He does the rapid strep test, and it’s negative. Dr Liam reported the news of a negative strep test and said, “Many times we get a 50% false-negative, so I’m gonna go ahead and put you on an antibiotic and see if we can’t get you feeling better.” With this the patient said, “Well, what do you think I should do about my kids?” Dr. Liam asked if the kids were seen in this clinic, and the patient responded that they had never been seen there before. Dr Liam said, “Well, I’ll go ahead and give you a script for erythromycin in case these kids get sick down in Florida. If they do, go ahead and give them the medicine; if they don’t, throw away the prescription.”

This case is interesting because once he is treated with an antibiotic, the patient uses the same argument to make an appeal for antibiotics for his children (both of whom have never been seen by this physician).

Previous positive experience with antibiotics (eg, “I got an antibiotic for this before.”). Patients also appealed to other nonmedical contingencies to advocate for antibiotic treatment. For example:

Our next patient was a 51-year-old woman complaining of a cold and laryngitis. The doctor asked the patient about her symptoms. The patient responded, saying that she had been taking medication during the end of December for the same symptoms; they had cleared after taking antibiotics, and now they were back again.

The patient indirectly makes an appeal for antibiotic treatment by stating that she received antibiotics in the past for the same symptoms that she has now.

Patients used several variations of this approach. These included stating that another physician prescribed an antibiotic for this illness in the past; that others in the family are sick with an illness for which they received antibiotics; that they have a history of illness for which antibiotics are regularly prescribed; and that they were recently taking an antibiotic for an illness that has not improved (with the idea that an antibiotic is needed again).

Effectiveness of Patient Pressures

Physicians prescribed an antibiotic unnecessarily in 80% of the encounters in which some patient pressure was observed. They seemed able to resist certain types of pressures better than others. Unnecessary antibiotics were prescribed for a smaller percentage of implied candidate diagnoses and candidate diagnoses and for a larger percentage of direct patient requests and previous positive experiences with antibiotics Table 3.

Physicians’ Response to Prescribing an Unnecessary Antibiotic

When physicians prescribed an antibiotic unnecessarily, they often rationalized this practice by finding symptoms or assigning diagnoses that, to them, justified prescribing antibiotics. Physicians used various rationales, such as red throat or enlarged tonsils; severe, prolonged, or productive cough; yellow or green mucus; sinus tenderness on palpation; associated chronic disease; history of previous infection; and the desire to “cover” the patient “just in case.” None of these rationales are supported by evidence as correlating with bacterial infection. An example of this kind of rationalization follows:

This is a 20-year-old woman coming in with a complaint of a worsening cough. She said that her chest had a prickly, burning sensation, and it hurt to breathe. Dr Hart asked if she was able to bring anything up. She said that she really couldn’t. It was just a really terrible barky cough. Following the physician examination, the physician told the patient that her lungs basically sounded clear, but she could certainly hear some rough bronchial sounds. With this, she said. “What I think is happening here with your cold is that it is probably ending up in a bronchitis-type situation, and probably what we should do is put you on an antibiotic and order a decongestant.”

Discussion

This investigation, in agreement with the pediatric studies of Stivers,^9,10 suggests that the connection between patient diagnosis and physician prescribing is highly complex, involving patient presentation and physician-patient communication as much as, if not more than, physician diagnostic skills. Also, these data suggest that physicians are better able to resist patient pressures that are framed in medical terms such as candidate diagnoses or implied candidate diagnoses but are much less able to resist pressures that are not medicalized, such as portraying severity of illness and use of life-world circumstances. Thus, it is not surprising that past interventions designed to increase physician knowledge regarding when to prescribe antibiotics have had limited success.^14,15 Physicians appear to be trying to maximize patient satisfaction by giving antibiotic-seeking patients what they want. Our findings show the need to modify current thinking about the diagnostic and treatment process to reduce the use of antibiotics. Rather than thinking of these processes as physician controlled, the powerful role patients play in this interaction must be considered.

Our study has important implications for future research. From a methodologic standpoint, our findings illustrate the importance of qualitative evaluation of directly observed medical encounters. The patterns of patient behavior observed could not have been discerned using survey, interview, or focus group data.

Limitations

Because these data were collected by field researchers who were unaware that ART infection would be a focus of our study, it is possible that there were other patient symptoms and behavior related to ART infection, as well as physician behaviors related to antibiotic prescribing, that were not recorded. The data were sufficiently rich, however, to easily and reliably apply the CDC guidelines for appropriate use of antibiotics. Any unrecorded behaviors might add to, but not substantially change, our conclusions that patients indirectly pressure their physicians for treatment, and physicians respond by giving antibiotics. Studies using videotaped encounters might uncover such additional important patient and physician behaviors. Since the patient population studied was limited to a single midwestern state, it is possible that other populations with a different ethnic or racial mix might behave differently. Future research in this area should attempt to include such populations. Finally, too few encounters per physician were observed in this study to evaluate whether particular physicians were high or low prescribers (such a pattern has been reported by De Sutter and colleagues¹⁶).

Conclusions

Physicians should be educated about the subtle approaches patients use to pressure them for antibiotic treatment and should be shown techniques for responding to these pressures without prescribing antibiotics unnecessarily. Our findings also suggest the need to increase patients’ awareness both of the dangers and lack of effectiveness of using antibiotics for ART infections and of the amount of influence that patients have on antibiotic prescribing. Macfarlane and coworkers17 have shown that use of patient education materials reduces visits for ART infection. Additional approaches to decreasing patient pressure for antibiotic prescriptions are needed to diminish antibiotic overuse and its public health consequences.

Acknowledgments

Our study was funded by the Agency for Healthcare Research and Quality Grant R01 HS08776. Dr Scott is a postdoctoral fellow supported by the Health Resources and Services Administration (HRSA) PE1011 and the Agency for Healthcare Research and Quality (AHRQ) HS09788. Analysis of these data was supported by a Research Center grant from the American Academy of Family Physicians (Center for Research in Family Practice and Primary Care). Drs Jaen and Crabtree are associated with the Center for Research in Family Practicer and Primary Care, Cleveland, New Brunswick, Allentown. and San Antonio. The authors wish to thank the family physicians of Nebraska who were willing to open their practices to us. We also thank Kurt C. Stange, MD, PhD, for his thoughtful comments on drafts of this manuscript.

Related Resources

U.S. Centers for Disease Control and Prevention—Promoting Appropriate Antibiotic Use in the Community http://www.cdc.gov/antibioticresistance/tools.htm

A vast resource of of patient education resources.

Acute respiratory tract (ART) infections, such as common cold, bronchitis, pharyngitis, sinusitis, and otitis media, are among the most common problems seen in primary care practice.¹ Unnecessary use of antibiotics for these infections is a major worldwide problem both in terms of cost² and as a contributor to the development of antibiotic-resistant bacteria.³

Although there is some evidence that physicians misdiagnose many viral infections as bacterial,^4,5 recent studies suggest that the reasons for unnecessary antibiotic prescribing are more complex, having as much or more to do with patient and physician expectations as with physicians’ diagnostic skills.^6-8 These studies are limited to describing perceptions of behavior rather than actual behavior, because of their use of interview and focus group data. Consequently, we do not know what actually happens during outpatient visits for ART infections that leads to antibiotic prescribing.

Two studies by Stivers^9,10 underscore the importance of directly observing what transpires during encounters with pediatric ART infection patients. Stivers’ examination of videotaped visits found that, in some cases, parental pressure for antibiotics influenced the physician’s decision to prescribe. This finding has not been replicated, however, in family practice settings, where both adults and children are seen. We used direct observation of outpatient visits to family physicians for ART infections to analyze the effects of physician-patient communication on unnecessary antibiotic prescribing. By understanding the ways these communication patterns influence prescribing behavior, practicing family physicians can develop strategies to deliver more appropriate care for ART infections.

Methods

These data were collected as part of the Prevention & Competing Demands in Primary Care Study, which was an in-depth observational study begun in October 1996 and completed in August 1999 that examined the organizational and clinical structures and process of community-based family practices. Each of 18 purposefully selected practices was studied using a multimethod comparative case study design that involved extensive direct observation of clinical encounters and office systems by field researchers who spent 4 weeks or more in each practice. Field researchers directly observed and dictated descriptions of approximately 30 patient encounters with each of the more than 50 clinicians. Details of the sampling and data collection are available elsewhere in this issue.¹¹

Data Analysis and Interpretation

Encounters related to ART infection were identified in the database using search terms for symptoms and diagnoses including: sore throat, runny nose, congestion, cough, drainage, postnasal drainage, earache, cold, upper respiratory infection, pharyngitis, sinusitis, bronchitis, and otitis. ART infection was identified as the principal or associated diagnosis in 316 outpatient visits of a total of 1637 observed encounters; 298 had sufficiently rich data for analysis. The encounters were first coded for antibiotic use or nonuse.

Before any qualitative analysis began, visits during which antibiotics were prescribed were further characterized as appropriate or unnecessary according to guidelines by the Centers for Disease Control and Prevention (CDC) for judicious use of antibiotics for children¹² and adults.Table W1¹³* Two family physicians assigned appropriate/unnecessary codes independently. Inter-rater reliability was good (k=0.71). All disagreements were resolved by discussion.

Subsequently, the text for each outpatient visit was read independently by 2 family physicians, a medical anthropologist, a nurse, and a communication specialist. This research team discussed individual encounters as a group to identify emerging patterns of physician-patient interaction.

Results

Women made up 59% of the study sample; 64% were 16 years or older (and classified as adults). Antibiotics were prescribed in 204 of the 298 ART infection encounters (68%). Antibiotic use was unnecessary according to the CDC guidelines in 164 of these (80%). Adults were more likely than children to receive unnecessary antibiotics Table 1.

Our analysis identified 6 different types of patient behaviors that advocated for medication, particularly antibioticsTable 2. These behaviors fell into 3 broad categories: explicit requests, presentation of chief complaint, and appeals to lifeworld circumstances. Multiple pressures were noted in many encounters.

While patients occasionally made direct requests for antibiotics, they much more frequently positioned themselves indirectly for receiving antibiotic treatment by the way they presented the chief complaint. Four distinct approaches were identified: symptoms only,⁹ candidate diagnosis,⁹ implied candidate diagnosis,⁹ and portraying the severity and inability to shake the illness.

A second category of indirect approach used life-world circumstances¹⁰ (eg, an upcoming family vacation) or a past history with successful antibiotic treatment to formulate appeals for antibiotics in the current encounter. In those cases in which antibiotics were clearly unnecessary, physicians often rationalized their prescribing practices by finding symptoms or assigning diagnoses to justify antibiotic use. Each of these patient pressures, as well as the physician-rationalizing behavior, is illustrated with sample visits. The samples are taken directly from transcribed field notes, but the names have been altered to protect the identity of patients and clinicians.

Explicit Request

Explicit requests for antibiotics were observed in only 6% of cases (n=15). For example:

Claire asked the patient, “How are you doing?” and she said, “Well, I’m coughing up phlegm, I ache and I have chills and a sore throat.” Claire said, “You have bronchoconstriction, and 3 times a day, if you need to, you should use proventil.” The patient asked if she could have an antibiotic for her cold; cephalexin has worked in the past. Claire said that she would get her cephalexin and also some samples of an inhaler.

Presentation of the Chief Complaint

Patients frequently put pressure on the physician for treatment during the presentation of the chief complaint, the exception being the symptoms-only presentation. This is different from the other indirect pressures, which usually occurred during different parts of the medical encounter.

Symptoms-only presentation (eg, “I have a cough and a sore throat.”) In the symptoms-only approach (n=15), the patient reports his or her symptoms with little embellishment. This approach does not pressure physicians for antibiotic treatment.

Candidate diagnosis (eg, “I think I’ve got strep throat.”) In contrast, patients also presented their chief complaint to the physician by offering a candidate diagnosis (n=18). As shown in the following example, the patient responds by offering a diagnosis. This is a way of indirectly advocating for antibiotic treatment.

A 21-year-old white woman went to see Dr. Maxwell with an acute problem of congestion. Dr Maxwell said, “Well, how are you doing?” The patient said, “It sounds like bronchitis. It started about 4 days ago.”

Implied candidate diagnosis (eg, “My throat hurts; it’s red; and it has white spots.”) The implied candidate diagnosis is a hybrid of the symptoms-only and the candidate diagnosis approaches (n=48). When presenting their chief complaint, patients reported very specific symptoms that indexed a particular diagnosis. For example:

A 29-year-old woman went to see Dr Redmond with swollen glands, congestion, and white spots on her throat. When Dr Redmond and I went into the examination room, the patient had a pink paper top on, and Dr Redmond told her that her throat culture was negative.

The patient reports that she has swollen glands, congestion, and white spots on her throat. The symptoms specifically index a particular condition (strep throat). The patient’s presentation of symptoms clearly implies a diagnosis of strep throat, and the physician ordered a strep culture before seeing the patient.

Candidate diagnoses and implied candidate diagnoses delicately assert that the nature of the patient’s problem is already known. The reason for the medical visit is to seek treatment for the patient’s already known condition. When candidate and implied candidate diagnoses point to a condition the patient believes to be treatable (eg, bronchitis, strep throat, ear infection), this way of presenting the chief complaint looks directly ahead to a treatment involving a prescription for an antibiotic and thus indirectly pressures the physician to prescribe one.

Portraying the severity of one’s illness (eg, “I can’t shake this, Doc.”) The most common strategy was for patients to subtly pressure physicians for medication by portraying the severity of their condition and their inability to shake the illness on their own (n=99). For example:

The patient was sitting up on the table, and right away he told Dr Lamont, “I just can’t shake it. I feel like the back of my throat has raw hamburger hanging in it.” Dr Lamont checked the patient’s throat well, and the patient said, “This has lasted 4 days and it has been getting worse today.” Dr Lamont checked the patient’s ears, glands, and lungs. “I’m going to give you a shot of penicillin, slow release. It’s some kind of an infection. It may be a virus.”

Portraying the severity of one’s illness may not in and of itself advocate for medication; however, portrayals of the severity of one’s condition were usually accompanied by other actions implicating the need for medication. By opening the encounter with the announcement “I just can’t shake it,” the patient implies that he needs help in getting well. This subtly suggests the need for a prescription medication to alleviate his sore throat. At the end of this visit, the patient receives an antibiotic shot.

Appeals to Nonmedical Circumstances

Patients also used nonmedical circumstances to advocate for medication. These behaviors tended to occur after the problem presentation in the encounter and either centered on some important event, such as a big examination or a trip out of town (n=16), or focused on a previous positive experience with antibiotics for themselves or a family member (n=39).

Appealing to life-world circumstance (eg, “But I’m going to Disney World.”) This patient uses an upcoming trip to make an appeal to the clinician to prescribe medication:

The patient is a 33-year-old man coming in with an acute problem of a sore throat. The patient stated that he had been trying to manage this on his own, but he was taking his wife and 2 children to Disney World at the end of the week and was becoming worried that he was still going be sick and not able to enjoy a trip that they had saved so long for. He also told Dr Liam: “I know we’ll just get to Florida, and the kids will get sick, and then we’ll all be sick again. Dr. Liam said, “Well, we can have you bring them in, but then we’d be treating them for something that they haven’t gotten. Let me think about this a bit.” He does the rapid strep test, and it’s negative. Dr Liam reported the news of a negative strep test and said, “Many times we get a 50% false-negative, so I’m gonna go ahead and put you on an antibiotic and see if we can’t get you feeling better.” With this the patient said, “Well, what do you think I should do about my kids?” Dr. Liam asked if the kids were seen in this clinic, and the patient responded that they had never been seen there before. Dr Liam said, “Well, I’ll go ahead and give you a script for erythromycin in case these kids get sick down in Florida. If they do, go ahead and give them the medicine; if they don’t, throw away the prescription.”

This case is interesting because once he is treated with an antibiotic, the patient uses the same argument to make an appeal for antibiotics for his children (both of whom have never been seen by this physician).

Previous positive experience with antibiotics (eg, “I got an antibiotic for this before.”). Patients also appealed to other nonmedical contingencies to advocate for antibiotic treatment. For example:

Our next patient was a 51-year-old woman complaining of a cold and laryngitis. The doctor asked the patient about her symptoms. The patient responded, saying that she had been taking medication during the end of December for the same symptoms; they had cleared after taking antibiotics, and now they were back again.

The patient indirectly makes an appeal for antibiotic treatment by stating that she received antibiotics in the past for the same symptoms that she has now.

Patients used several variations of this approach. These included stating that another physician prescribed an antibiotic for this illness in the past; that others in the family are sick with an illness for which they received antibiotics; that they have a history of illness for which antibiotics are regularly prescribed; and that they were recently taking an antibiotic for an illness that has not improved (with the idea that an antibiotic is needed again).

Effectiveness of Patient Pressures

Physicians prescribed an antibiotic unnecessarily in 80% of the encounters in which some patient pressure was observed. They seemed able to resist certain types of pressures better than others. Unnecessary antibiotics were prescribed for a smaller percentage of implied candidate diagnoses and candidate diagnoses and for a larger percentage of direct patient requests and previous positive experiences with antibiotics Table 3.

Physicians’ Response to Prescribing an Unnecessary Antibiotic

When physicians prescribed an antibiotic unnecessarily, they often rationalized this practice by finding symptoms or assigning diagnoses that, to them, justified prescribing antibiotics. Physicians used various rationales, such as red throat or enlarged tonsils; severe, prolonged, or productive cough; yellow or green mucus; sinus tenderness on palpation; associated chronic disease; history of previous infection; and the desire to “cover” the patient “just in case.” None of these rationales are supported by evidence as correlating with bacterial infection. An example of this kind of rationalization follows:

This is a 20-year-old woman coming in with a complaint of a worsening cough. She said that her chest had a prickly, burning sensation, and it hurt to breathe. Dr Hart asked if she was able to bring anything up. She said that she really couldn’t. It was just a really terrible barky cough. Following the physician examination, the physician told the patient that her lungs basically sounded clear, but she could certainly hear some rough bronchial sounds. With this, she said. “What I think is happening here with your cold is that it is probably ending up in a bronchitis-type situation, and probably what we should do is put you on an antibiotic and order a decongestant.”

Discussion

This investigation, in agreement with the pediatric studies of Stivers,^9,10 suggests that the connection between patient diagnosis and physician prescribing is highly complex, involving patient presentation and physician-patient communication as much as, if not more than, physician diagnostic skills. Also, these data suggest that physicians are better able to resist patient pressures that are framed in medical terms such as candidate diagnoses or implied candidate diagnoses but are much less able to resist pressures that are not medicalized, such as portraying severity of illness and use of life-world circumstances. Thus, it is not surprising that past interventions designed to increase physician knowledge regarding when to prescribe antibiotics have had limited success.^14,15 Physicians appear to be trying to maximize patient satisfaction by giving antibiotic-seeking patients what they want. Our findings show the need to modify current thinking about the diagnostic and treatment process to reduce the use of antibiotics. Rather than thinking of these processes as physician controlled, the powerful role patients play in this interaction must be considered.

Our study has important implications for future research. From a methodologic standpoint, our findings illustrate the importance of qualitative evaluation of directly observed medical encounters. The patterns of patient behavior observed could not have been discerned using survey, interview, or focus group data.

Limitations

Because these data were collected by field researchers who were unaware that ART infection would be a focus of our study, it is possible that there were other patient symptoms and behavior related to ART infection, as well as physician behaviors related to antibiotic prescribing, that were not recorded. The data were sufficiently rich, however, to easily and reliably apply the CDC guidelines for appropriate use of antibiotics. Any unrecorded behaviors might add to, but not substantially change, our conclusions that patients indirectly pressure their physicians for treatment, and physicians respond by giving antibiotics. Studies using videotaped encounters might uncover such additional important patient and physician behaviors. Since the patient population studied was limited to a single midwestern state, it is possible that other populations with a different ethnic or racial mix might behave differently. Future research in this area should attempt to include such populations. Finally, too few encounters per physician were observed in this study to evaluate whether particular physicians were high or low prescribers (such a pattern has been reported by De Sutter and colleagues¹⁶).

Conclusions

Physicians should be educated about the subtle approaches patients use to pressure them for antibiotic treatment and should be shown techniques for responding to these pressures without prescribing antibiotics unnecessarily. Our findings also suggest the need to increase patients’ awareness both of the dangers and lack of effectiveness of using antibiotics for ART infections and of the amount of influence that patients have on antibiotic prescribing. Macfarlane and coworkers17 have shown that use of patient education materials reduces visits for ART infection. Additional approaches to decreasing patient pressure for antibiotic prescriptions are needed to diminish antibiotic overuse and its public health consequences.

Acknowledgments

Our study was funded by the Agency for Healthcare Research and Quality Grant R01 HS08776. Dr Scott is a postdoctoral fellow supported by the Health Resources and Services Administration (HRSA) PE1011 and the Agency for Healthcare Research and Quality (AHRQ) HS09788. Analysis of these data was supported by a Research Center grant from the American Academy of Family Physicians (Center for Research in Family Practice and Primary Care). Drs Jaen and Crabtree are associated with the Center for Research in Family Practicer and Primary Care, Cleveland, New Brunswick, Allentown. and San Antonio. The authors wish to thank the family physicians of Nebraska who were willing to open their practices to us. We also thank Kurt C. Stange, MD, PhD, for his thoughtful comments on drafts of this manuscript.

Related Resources

U.S. Centers for Disease Control and Prevention—Promoting Appropriate Antibiotic Use in the Community http://www.cdc.gov/antibioticresistance/tools.htm

A vast resource of of patient education resources.

In primary care practice, patients in the top 3% for attendance generate 15% of total office visits.¹ “Frequent attenders” are more likely to be older, divorced or widowed, in lower socioeconomic groups, and to have multiple physical and psychosocial ills, and vague physical symptoms with no obvious etiology.^2-7 At one extreme, frequent attenders are “heartsink patients” as described by O’Dowd⁸—individuals whose demands, behaviors, and dissatisfaction with care give the “doctor and staff a feeling of ‘heartsink’ every time they consult.” The memorable and sometimes overwhelming nature of encounters with such patients may lead to the conclusion that difficult encounters are the rule among frequent attenders.

Although the demographics and disease patterns among frequent attenders have been described, the detailed characteristics of their encounters with physicians have not been elucidated.

Multiple factors may influence the content of physician-patient encounters, including the character and severity of a patient’s symptoms, the level of patient concern, the patient’s interpersonal style, the interviewing style of the physician, the complexity of the patient’s medical and psychosocial difficulties, and the level of comfort and trust between patient and physician.^9,10 A better understanding of frequent attender visits could help guide more efficient management of patients’ problems and improve satisfaction for both the patient and the physician. The purpose of our paper is to describe the characteristics of encounters between family physicians and adult frequent-attender patients.

Methods

We drew the data used for this analysis from The Prevention and Competing Demands in Primary Care Study. This multimethod comparative case study of 18 community-based family practices involved extensive direct observation of clinical encounters and office systems by field researchers who spent 4 weeks or more in each practice. The primary data for this analysis were collected by field researchers who were trained to take chronological notes while observing outpatient encounters and later dictate them to create detailed descriptions of each encounter. Details of the sampling and data collection can be found elsewhere in this issue.¹¹

Field researchers directly observed and dictated descriptions of approximately 30 patient encounters with each of the more than 50 clinicians, and they audited the medical records of each of these patients. Visits with 1194 adults 21 years and older were observed. Non-pregnant patients were sorted by their number of visits in the previous 2 years as determined by medical record review. Those in the top 5% for visit frequency were selected for analysis.

Three family physician researchers with experience in qualitative methodology used an immersion/crystallization style to explore and characterize the physician-patient encounters.^12,13 In this interpretive style each researcher read the field notes independently and recorded his or her summary comments and possible themes. Over several sessions, the 3 reviewers met to compare observations, review each patient encounter in detail together, resolve disagreements regarding the characteristics of each physician-patient encounter, and identify major themes from the data. These themes gradually developed into a 3-dimensional framework for examining all encounters. This framework was refined through additional review of field notes and discussions with study collaborators familiar with the data. The field notes were further reviewed to identify individual encounters that illustrated the range of encounter types represented by the descriptive framework.

Although the initial purpose of the study did not include comparison of frequent attenders with other patients, it was unclear whether the descriptive framework that had emerged from the analysis was unique to frequent attenders or if it could be applied to patients who visit much less often. Thus, we reviewed additional encounters by non-frequent attenders to explore the applicability and relevance of the framework to other patients. Visits by patients with only 2 to 6 visits in the previous 2 years (non-frequent attenders) were matched by age and sex to the frequent attender sample (n=62). Each of the 3 reviewers independently read and characterized each non-frequent attender visit using the descriptive framework, then met to compare findings and determine major themes when comparing visits by frequent and non-frequent attenders.

Results

Frequent Attender Visits

Adult patients in the top 5% for visit frequency (n=62) had 25 or more visits over the previous 2 years Figure 1. The average number of encounters for frequent attenders over the previous 2 years was 32 (range= 25-52). The average age of these patients (65 years) was greater than that for all other adult patients (51 years), while the proportion of women was similar for frequent attenders and all other adult patients (68% and 66%, respectively).

Encounters with frequent attenders reflected a wide range of biomedical and psychosocial complexity and a variety of physician and patient communication patterns. We identified 3 major dimensions that distinguished differences in the content and characteristics of the encounters:

1. Biomedical complexity varied considerably in terms of the number and complexity of acute or chronic illnesses. During some visits only a single straightforward biomedical issue was addressed (biomedical=low), while other visits included review of more complex or more numerous acute or chronic conditions (biomedical=high).
2. Psychosocial complexity was evidenced by a patient’s need for reassurance or counseling from the clinician regarding family or relationship issues, discussions regarding lifestyle issues, a perceived need for general social support, an expression of nonspecific somatic symptoms, or presentation of symptoms of depression or anxiety. Some visits reflected minimal psychosocial issues (psychosocial=low), while others encompassed a wide range of expressed emotions, concerns, or symptoms related to psychosocial issues (psychosocial=high).
3. Dissonance between the patient and physician included a lack of mutual understanding and agreement between the patient and clinician, or frustration and/or confusion regarding diagnostic conclusions, goals and direction of treatment, and follow-up plans. Many visits included friendly chatting between the physician and the patient with evidence for a warm familiar understanding between patients, physicians, nurses, and office staff, and usually ended with an organized, mutually agreeable plan of action (dissonance=low). However, in some encounters patient concerns or symptoms were either only partially addressed by the physician or there seemed to be disagreement, confusion, frustration, or conflict about the diagnosis or therapeutic plan (dissonance=high).

A three-dimensional framework emerged from these themes to characterize the range of encounter types with frequent attenders Figure 2. Each encounter could be rated along a continuum from low to high on each axis, resulting in a wide range of visit types. The following description of frequent attender visits illustrates the 8 combinations of the lowest and highest dimension ratings within this framework:

1. Simple medical (biomedical=low, psychosocial=low, dissonance=low; n=12): An 83-year-old man came to the office with an acute foot injury. The physician completed a focused history and physical examination. A radiograph was negative for fracture. The patient expressed no other symptoms and had no questions about the prescribed treatment.
2. Ritual visit (biomedical=low, psychosocial=high, dissonance=low; n=9): A 54-year-old woman with chronic low back pain was seen first by a nurse practitioner who chatted with the patient about a number of relationship and lifestyle issues while giving her multiple lidocaine trigger point injections in the low back and buttocks, a procedure that had occurred during previous visits. When asked by the nurse practitioner whether she needed any other trigger points injected, the patient answered “no” and got dressed. The physician entered the room 15 minutes later and asked, “Should I do another spot for you?” During a number of additional trigger point injections by the physician, the patient discussed additional family and relationship issues with the physician.
3. Complicated medical (biomedical=high, psychosocial=low, dissonance=low; n=15): A 68-year-old man presented for follow-up of diabetes mellitus type 2, recent myocardial infarction, smoking cessation efforts, and an injured hand from a recent fall. After history taking and physical examination were completed, a plan for each problem was stated and readily agreed to by the patient. The encounter was efficient and friendly, including a few inside jokes between the physician and the patient. There was no mention of psychosocial issues or related symptoms.
4. The tango (biomedical=high, psychosocial=high, dissonance=low; n=11): A 49-year-old woman with multiple minor biomedical problems completed “the most unusual visit” that the nurse researcher had ever observed. The encounter began when the patient “just barged in” to the physician’s administrative office and asked to watch a TV game show. After watching the show together for a few minutes they moved to the examination room where the patient proceeded to “tell the doctor what to do” for her multiple complaints. The physician appeared to take all this in stride and facilitated a series of friendly negotiations for a list of medical issues including sore throat, arthritis pain, and chronic respiratory symptoms. There seemed to be a mutually acceptable resolution and a plan for each issue that emerged from this complicated “dance”.
5. Simple frustration (biomedical=low, psychosocial=low, dissonance=high; n=1): Only one visit in the sample had elements of this visit type but was not a clear exemplar, since the dissonance expressed by the patient was not directly related to the physician encounter itself. A 71-year old woman with severe back pain was unhappy about “waiting for a week to get something done” for her pain, and both she and her husband expressed significant frustration about difficulties with scheduling an epidural injection at the hospital. The physician assisted with rescheduling the procedure for the following day, but some confusion remained at the end of the visit as to where the patient was to meet with the anesthesiologist. After the visit the patient’s husband commented to the field researcher that “we live next door to (this physician), so we have to give him a hard time.”
6. Psychosocial disconnect (biomedical=low, psychosocial=high, dissonance=high; n=7): A 33-year-old man came to the office for a chronic leg ulcer caused by an underlying metabolic disorder. After he briefly checked the ulcer and applied a new dressing, the physician expressed dismay and frustration about the patient’s ongoing reluctance to quit smoking. He lectured the patient about the harm caused by smoking, particularly given his underlying condition. They discussed options for smoking cessation and other lifestyle issues. The physician then encouraged the patient to follow through with his previous suggestion to screen his infant son for metabolic disorders. The patient was reluctant, since the child’s pediatrician “didn’t seem too concerned.” The physician replied that “the pediatrician should be concerned” and to “consider taking him to another physician.” During their discussion of lifestyle and family concerns, there was no indication that the patient agreed with the physician’s advice or planned to take action.
7. Medical disharmony (biomedical=high, psychosocial=low, dissonance=high; n=4): A 52-year-old woman presented with left-sided chest and shoulder pain. There was confusion regarding which medications she was currently taking for hypertension, sleep disorder, headache, and gastrointestinal symptoms. During the physical examination, active range of motion of both shoulders elicited no increase in pain. The only stated plan by the physician was that a “short burst of steroid” was best because “it’s an inflammatory thing, I think.” After the physician left the room the patient repeated her concern to the observer that she had “a lot of pain in my shoulder.” There were no expressed psychosocial needs and little evidence for significant interpersonal connection between the physician and the patient during the visit.
8. The heartsink visit (biomedical=high, psychosocial=high, dissonance=high; n=3): A 57-year-old woman came to the office with multiple minor medical problems and chronic depression. With English as her second language, a communication barrier complicated the encounter. During the visit diagnoses of urinary tract infection, arthritis, and gastritis were discussed, an abdominal radiograph obtained, and new medications prescribed. The patient cried as she described a number of sick family members who lived in Mexico and implied that she was unable to visit them because the physician would not adequately treat her pain. She threatened to “go on the street and get any drug I want for pain.” Although there was evidence that the physician had been compassionate and persistent in his attempts to assist her, there seemed to be little agreement with the patient on the direction for further diagnosis or therapy. Seemingly frustrated by the encounter, the physician concluded the visit by saying “Well, we will see you back here in a month and you’ll be feeling better, right?” The patient “kind of looked at him” and said, “Okay, bye, Doctor.”

Non-Frequent Attender Visits

The researchers were able to easily categorize non-frequent attender visits using the descriptive framework Figure 2. A majority of non-frequent attender visits (87%) were classified along the “biomedical” continuum from type 1 (simple medical; n=34) to type 3 (complex medical; n=20) visits. Psychosocial complexity and dissonance were less prominent and were addressed less often than with frequent attenders, and only one visit approached the heartsink corner of the framework. Non-frequent attender visits encompassed fewer exchanges of humor and small talk than observed in the frequent attender encounters and generally showed less evidence of familiarity between patient and physician. The visits were less dramatic in the range of characteristics defined by the descriptive framework and were less memorable than those of the frequent attenders.

Discussion

Our study is the first to provide a detailed description of the characteristics of encounters between family physicians and adult frequent attenders. We found wide variation in the content of these encounters. A framework emerged that describes the degree of biomedical complexity, psychosocial complexity, and dissonance between the patient and the physician for each encounter. Although previous epidemiologic studies and case series suggest that frequent attenders may generate many difficult heartsink visits, the encounters we studied were scattered across the entire 3-dimensional framework from very simple single issue visits to highly complex emotionally taxing visits.^2-7,14 The dimensions of the descriptive framework were applicable to non-frequent attender visits, but the range of psychosocial complexity and dissonance was greater among visits by frequent attenders.

Many frequent attenders seemed to have developed an intricate and harmonious relationship with the physician and the office staff and nurses in the practice. Visits by frequent attenders often included friendly chatting and humor among patients, staff, and physicians. These findings are consistent with the Direct Observation of Primary Care (DOPC) study, where chatting was a part of 69% of all visits to family physicians and accounted for almost 8% of overall visit time.¹⁵ Older patients who had longer visits and spent more time chatting with their physicians in the DOPC study reported greater satisfaction with care.¹⁶ We hypothesize that patients who find a “medical home” where they can talk comfortably with physicians and staff and gain a level of general social support are likely to return often. If such familiarity is interrupted by seeing a physician other than the patient’s continuity provider, as happened in the “medical disharmony” visit we described, confusion about treatment or other visit-specific dissonance may result between the treating physician and the patient.

The content of an encounter is influenced by physicians’ interviewing skills and techniques and whether patients voice all of their concerns, symptoms, and health questions during the visit.^10,17,18 Many factors determine whether psychosocial concerns are elicited and addressed during a visit as described in another article in this issue.¹⁹ Over multiple visits, patients and physicians are likely to reach an equilibrium of expectations regarding patterns of communication and to develop mutually acceptable parameters for the relationship. Frequent attenders have many opportunities to learn a physician’s style and approach to medical and psychosocial problems. Some encounters in our study suggested that a ritual pattern of discussion or visit procedures had developed over time within a trust-filled continuity relationship.

Patients who develop a strong continuity relationship may be less likely to present a question or a request that they know will not be agreed to by the physician. This may explain why the “simple frustration” visit type was not well represented in the frequent attender sample. When patients disagree with physicians on straightforward treatment issues, such as antibiotic prescription for an uncomplicated upper respiratory infection, they usually will either come to some understanding and acceptance of the physician’s views or eventually seek care from another physician.

Limitations

Our descriptive study has a number of limitations. Field notes from the nurse observers described the interaction between the physician and patient and included subjective interpretations of each encounter that may not have accurately reflected the tone of the physician-patient interaction. The observers did not actively seek information about patients’ unvoiced concerns, thus our conclusions regarding the degree of dissonance in the encounter were implied only from written observations. Other than occasional field notes recorded from physicians’ comments after the patient left the office, the observers also did not measure the physician’s emotional response to encounters. These limitations may have led us to misclassify a given encounter on one or more of the 3 dimensions. Also, we cannot conclude from these data whether the sampled visit for each patient represented a typical visit for that patient, and we cannot judge the appropriateness of the care provided.

We reported the number of visits in each category to provide a general impression of the variation of visit types among frequent attenders and how it differed from the non-frequent attender group. Given the qualitative study design and data collection methods, the distribution of visit types may not accurately represent all frequent attender visits to the physician practices in this study and may not be generalizable to other clinical settings.

For some patients frequent attendance is appropriate, and a one-size-fits-all approach to reducing frequent attendance is unlikely to be effective. In his qualitative study of the management styles of 2 experienced family physicians, Miller20 described a visit typology that included routine, drama, and ceremony visits. He observed that visits with frequent attenders were often “maintenance ceremonies” with a prescribed repetitive format guided by a skilled physician. The “ritual” or “tango” visit types described in our study were the most obvious instances of “maintenance ceremonies” that were based on a high level of familiarity between patient and physician. The ceremonies in such visits may not be based on published medical guidelines or outcomes studies but are likely, in some instances, to represent an effective and appropriate therapeutic response by a physician who has developed an intimate understanding of the complex medical and psychosocial needs of a patient. In other cases such a visit pattern may be an inappropriate pattern of care that allows the familiar routines to take the place of consideration of different diagnostic or therapeutic directions.

Conclusions

Our study provides a foundation for further understanding the variation in family practice patient encounters and holds implications for a number of other issues that require further study. Any evaluation of the appropriateness of outpatient utilization patterns should recognize the wide range of encounter types that may not be apparent from studies of administrative databases. Interventions designed to reduce “inappropriate” utilization also should consider such factors. Our study also provides a basis for teaching students and physicians how to understand and manage the complexities and variation of outpatient primary care encounters. By categorizing and reflecting on the pattern of visit types over time for particular patients, physicians may gain insight into how best to care for “difficult” frequent attenders.

Acknowledgments

Our study was supported by a grant from the Agency for Healthcare Research and Quality (R01 HS08776), a grant from the Health Resources and Services Administration (5D32HP10231), and a Family Practice Research Center grant from the American Academy of Family Physicians. We are grateful to the physicians, staff, and patients from the 18 practices, without whose participation this study would not have been possible. The authors also wish to thank Kurt C. Stange, MD, PhD, who provided helpful comments on earlier drafts of this paper.

In primary care practice, patients in the top 3% for attendance generate 15% of total office visits.¹ “Frequent attenders” are more likely to be older, divorced or widowed, in lower socioeconomic groups, and to have multiple physical and psychosocial ills, and vague physical symptoms with no obvious etiology.^2-7 At one extreme, frequent attenders are “heartsink patients” as described by O’Dowd⁸—individuals whose demands, behaviors, and dissatisfaction with care give the “doctor and staff a feeling of ‘heartsink’ every time they consult.” The memorable and sometimes overwhelming nature of encounters with such patients may lead to the conclusion that difficult encounters are the rule among frequent attenders.

Although the demographics and disease patterns among frequent attenders have been described, the detailed characteristics of their encounters with physicians have not been elucidated.

Multiple factors may influence the content of physician-patient encounters, including the character and severity of a patient’s symptoms, the level of patient concern, the patient’s interpersonal style, the interviewing style of the physician, the complexity of the patient’s medical and psychosocial difficulties, and the level of comfort and trust between patient and physician.^9,10 A better understanding of frequent attender visits could help guide more efficient management of patients’ problems and improve satisfaction for both the patient and the physician. The purpose of our paper is to describe the characteristics of encounters between family physicians and adult frequent-attender patients.

Methods

We drew the data used for this analysis from The Prevention and Competing Demands in Primary Care Study. This multimethod comparative case study of 18 community-based family practices involved extensive direct observation of clinical encounters and office systems by field researchers who spent 4 weeks or more in each practice. The primary data for this analysis were collected by field researchers who were trained to take chronological notes while observing outpatient encounters and later dictate them to create detailed descriptions of each encounter. Details of the sampling and data collection can be found elsewhere in this issue.¹¹

Field researchers directly observed and dictated descriptions of approximately 30 patient encounters with each of the more than 50 clinicians, and they audited the medical records of each of these patients. Visits with 1194 adults 21 years and older were observed. Non-pregnant patients were sorted by their number of visits in the previous 2 years as determined by medical record review. Those in the top 5% for visit frequency were selected for analysis.

Three family physician researchers with experience in qualitative methodology used an immersion/crystallization style to explore and characterize the physician-patient encounters.^12,13 In this interpretive style each researcher read the field notes independently and recorded his or her summary comments and possible themes. Over several sessions, the 3 reviewers met to compare observations, review each patient encounter in detail together, resolve disagreements regarding the characteristics of each physician-patient encounter, and identify major themes from the data. These themes gradually developed into a 3-dimensional framework for examining all encounters. This framework was refined through additional review of field notes and discussions with study collaborators familiar with the data. The field notes were further reviewed to identify individual encounters that illustrated the range of encounter types represented by the descriptive framework.

Although the initial purpose of the study did not include comparison of frequent attenders with other patients, it was unclear whether the descriptive framework that had emerged from the analysis was unique to frequent attenders or if it could be applied to patients who visit much less often. Thus, we reviewed additional encounters by non-frequent attenders to explore the applicability and relevance of the framework to other patients. Visits by patients with only 2 to 6 visits in the previous 2 years (non-frequent attenders) were matched by age and sex to the frequent attender sample (n=62). Each of the 3 reviewers independently read and characterized each non-frequent attender visit using the descriptive framework, then met to compare findings and determine major themes when comparing visits by frequent and non-frequent attenders.

Results

Frequent Attender Visits

Adult patients in the top 5% for visit frequency (n=62) had 25 or more visits over the previous 2 years Figure 1. The average number of encounters for frequent attenders over the previous 2 years was 32 (range= 25-52). The average age of these patients (65 years) was greater than that for all other adult patients (51 years), while the proportion of women was similar for frequent attenders and all other adult patients (68% and 66%, respectively).

Encounters with frequent attenders reflected a wide range of biomedical and psychosocial complexity and a variety of physician and patient communication patterns. We identified 3 major dimensions that distinguished differences in the content and characteristics of the encounters:

1. Biomedical complexity varied considerably in terms of the number and complexity of acute or chronic illnesses. During some visits only a single straightforward biomedical issue was addressed (biomedical=low), while other visits included review of more complex or more numerous acute or chronic conditions (biomedical=high).
2. Psychosocial complexity was evidenced by a patient’s need for reassurance or counseling from the clinician regarding family or relationship issues, discussions regarding lifestyle issues, a perceived need for general social support, an expression of nonspecific somatic symptoms, or presentation of symptoms of depression or anxiety. Some visits reflected minimal psychosocial issues (psychosocial=low), while others encompassed a wide range of expressed emotions, concerns, or symptoms related to psychosocial issues (psychosocial=high).
3. Dissonance between the patient and physician included a lack of mutual understanding and agreement between the patient and clinician, or frustration and/or confusion regarding diagnostic conclusions, goals and direction of treatment, and follow-up plans. Many visits included friendly chatting between the physician and the patient with evidence for a warm familiar understanding between patients, physicians, nurses, and office staff, and usually ended with an organized, mutually agreeable plan of action (dissonance=low). However, in some encounters patient concerns or symptoms were either only partially addressed by the physician or there seemed to be disagreement, confusion, frustration, or conflict about the diagnosis or therapeutic plan (dissonance=high).

A three-dimensional framework emerged from these themes to characterize the range of encounter types with frequent attenders Figure 2. Each encounter could be rated along a continuum from low to high on each axis, resulting in a wide range of visit types. The following description of frequent attender visits illustrates the 8 combinations of the lowest and highest dimension ratings within this framework:

1. Simple medical (biomedical=low, psychosocial=low, dissonance=low; n=12): An 83-year-old man came to the office with an acute foot injury. The physician completed a focused history and physical examination. A radiograph was negative for fracture. The patient expressed no other symptoms and had no questions about the prescribed treatment.
2. Ritual visit (biomedical=low, psychosocial=high, dissonance=low; n=9): A 54-year-old woman with chronic low back pain was seen first by a nurse practitioner who chatted with the patient about a number of relationship and lifestyle issues while giving her multiple lidocaine trigger point injections in the low back and buttocks, a procedure that had occurred during previous visits. When asked by the nurse practitioner whether she needed any other trigger points injected, the patient answered “no” and got dressed. The physician entered the room 15 minutes later and asked, “Should I do another spot for you?” During a number of additional trigger point injections by the physician, the patient discussed additional family and relationship issues with the physician.
3. Complicated medical (biomedical=high, psychosocial=low, dissonance=low; n=15): A 68-year-old man presented for follow-up of diabetes mellitus type 2, recent myocardial infarction, smoking cessation efforts, and an injured hand from a recent fall. After history taking and physical examination were completed, a plan for each problem was stated and readily agreed to by the patient. The encounter was efficient and friendly, including a few inside jokes between the physician and the patient. There was no mention of psychosocial issues or related symptoms.
4. The tango (biomedical=high, psychosocial=high, dissonance=low; n=11): A 49-year-old woman with multiple minor biomedical problems completed “the most unusual visit” that the nurse researcher had ever observed. The encounter began when the patient “just barged in” to the physician’s administrative office and asked to watch a TV game show. After watching the show together for a few minutes they moved to the examination room where the patient proceeded to “tell the doctor what to do” for her multiple complaints. The physician appeared to take all this in stride and facilitated a series of friendly negotiations for a list of medical issues including sore throat, arthritis pain, and chronic respiratory symptoms. There seemed to be a mutually acceptable resolution and a plan for each issue that emerged from this complicated “dance”.
5. Simple frustration (biomedical=low, psychosocial=low, dissonance=high; n=1): Only one visit in the sample had elements of this visit type but was not a clear exemplar, since the dissonance expressed by the patient was not directly related to the physician encounter itself. A 71-year old woman with severe back pain was unhappy about “waiting for a week to get something done” for her pain, and both she and her husband expressed significant frustration about difficulties with scheduling an epidural injection at the hospital. The physician assisted with rescheduling the procedure for the following day, but some confusion remained at the end of the visit as to where the patient was to meet with the anesthesiologist. After the visit the patient’s husband commented to the field researcher that “we live next door to (this physician), so we have to give him a hard time.”
6. Psychosocial disconnect (biomedical=low, psychosocial=high, dissonance=high; n=7): A 33-year-old man came to the office for a chronic leg ulcer caused by an underlying metabolic disorder. After he briefly checked the ulcer and applied a new dressing, the physician expressed dismay and frustration about the patient’s ongoing reluctance to quit smoking. He lectured the patient about the harm caused by smoking, particularly given his underlying condition. They discussed options for smoking cessation and other lifestyle issues. The physician then encouraged the patient to follow through with his previous suggestion to screen his infant son for metabolic disorders. The patient was reluctant, since the child’s pediatrician “didn’t seem too concerned.” The physician replied that “the pediatrician should be concerned” and to “consider taking him to another physician.” During their discussion of lifestyle and family concerns, there was no indication that the patient agreed with the physician’s advice or planned to take action.
7. Medical disharmony (biomedical=high, psychosocial=low, dissonance=high; n=4): A 52-year-old woman presented with left-sided chest and shoulder pain. There was confusion regarding which medications she was currently taking for hypertension, sleep disorder, headache, and gastrointestinal symptoms. During the physical examination, active range of motion of both shoulders elicited no increase in pain. The only stated plan by the physician was that a “short burst of steroid” was best because “it’s an inflammatory thing, I think.” After the physician left the room the patient repeated her concern to the observer that she had “a lot of pain in my shoulder.” There were no expressed psychosocial needs and little evidence for significant interpersonal connection between the physician and the patient during the visit.
8. The heartsink visit (biomedical=high, psychosocial=high, dissonance=high; n=3): A 57-year-old woman came to the office with multiple minor medical problems and chronic depression. With English as her second language, a communication barrier complicated the encounter. During the visit diagnoses of urinary tract infection, arthritis, and gastritis were discussed, an abdominal radiograph obtained, and new medications prescribed. The patient cried as she described a number of sick family members who lived in Mexico and implied that she was unable to visit them because the physician would not adequately treat her pain. She threatened to “go on the street and get any drug I want for pain.” Although there was evidence that the physician had been compassionate and persistent in his attempts to assist her, there seemed to be little agreement with the patient on the direction for further diagnosis or therapy. Seemingly frustrated by the encounter, the physician concluded the visit by saying “Well, we will see you back here in a month and you’ll be feeling better, right?” The patient “kind of looked at him” and said, “Okay, bye, Doctor.”

Non-Frequent Attender Visits

The researchers were able to easily categorize non-frequent attender visits using the descriptive framework Figure 2. A majority of non-frequent attender visits (87%) were classified along the “biomedical” continuum from type 1 (simple medical; n=34) to type 3 (complex medical; n=20) visits. Psychosocial complexity and dissonance were less prominent and were addressed less often than with frequent attenders, and only one visit approached the heartsink corner of the framework. Non-frequent attender visits encompassed fewer exchanges of humor and small talk than observed in the frequent attender encounters and generally showed less evidence of familiarity between patient and physician. The visits were less dramatic in the range of characteristics defined by the descriptive framework and were less memorable than those of the frequent attenders.

Discussion

Our study is the first to provide a detailed description of the characteristics of encounters between family physicians and adult frequent attenders. We found wide variation in the content of these encounters. A framework emerged that describes the degree of biomedical complexity, psychosocial complexity, and dissonance between the patient and the physician for each encounter. Although previous epidemiologic studies and case series suggest that frequent attenders may generate many difficult heartsink visits, the encounters we studied were scattered across the entire 3-dimensional framework from very simple single issue visits to highly complex emotionally taxing visits.^2-7,14 The dimensions of the descriptive framework were applicable to non-frequent attender visits, but the range of psychosocial complexity and dissonance was greater among visits by frequent attenders.

Many frequent attenders seemed to have developed an intricate and harmonious relationship with the physician and the office staff and nurses in the practice. Visits by frequent attenders often included friendly chatting and humor among patients, staff, and physicians. These findings are consistent with the Direct Observation of Primary Care (DOPC) study, where chatting was a part of 69% of all visits to family physicians and accounted for almost 8% of overall visit time.¹⁵ Older patients who had longer visits and spent more time chatting with their physicians in the DOPC study reported greater satisfaction with care.¹⁶ We hypothesize that patients who find a “medical home” where they can talk comfortably with physicians and staff and gain a level of general social support are likely to return often. If such familiarity is interrupted by seeing a physician other than the patient’s continuity provider, as happened in the “medical disharmony” visit we described, confusion about treatment or other visit-specific dissonance may result between the treating physician and the patient.

The content of an encounter is influenced by physicians’ interviewing skills and techniques and whether patients voice all of their concerns, symptoms, and health questions during the visit.^10,17,18 Many factors determine whether psychosocial concerns are elicited and addressed during a visit as described in another article in this issue.¹⁹ Over multiple visits, patients and physicians are likely to reach an equilibrium of expectations regarding patterns of communication and to develop mutually acceptable parameters for the relationship. Frequent attenders have many opportunities to learn a physician’s style and approach to medical and psychosocial problems. Some encounters in our study suggested that a ritual pattern of discussion or visit procedures had developed over time within a trust-filled continuity relationship.

Patients who develop a strong continuity relationship may be less likely to present a question or a request that they know will not be agreed to by the physician. This may explain why the “simple frustration” visit type was not well represented in the frequent attender sample. When patients disagree with physicians on straightforward treatment issues, such as antibiotic prescription for an uncomplicated upper respiratory infection, they usually will either come to some understanding and acceptance of the physician’s views or eventually seek care from another physician.

Limitations

Our descriptive study has a number of limitations. Field notes from the nurse observers described the interaction between the physician and patient and included subjective interpretations of each encounter that may not have accurately reflected the tone of the physician-patient interaction. The observers did not actively seek information about patients’ unvoiced concerns, thus our conclusions regarding the degree of dissonance in the encounter were implied only from written observations. Other than occasional field notes recorded from physicians’ comments after the patient left the office, the observers also did not measure the physician’s emotional response to encounters. These limitations may have led us to misclassify a given encounter on one or more of the 3 dimensions. Also, we cannot conclude from these data whether the sampled visit for each patient represented a typical visit for that patient, and we cannot judge the appropriateness of the care provided.

We reported the number of visits in each category to provide a general impression of the variation of visit types among frequent attenders and how it differed from the non-frequent attender group. Given the qualitative study design and data collection methods, the distribution of visit types may not accurately represent all frequent attender visits to the physician practices in this study and may not be generalizable to other clinical settings.

For some patients frequent attendance is appropriate, and a one-size-fits-all approach to reducing frequent attendance is unlikely to be effective. In his qualitative study of the management styles of 2 experienced family physicians, Miller20 described a visit typology that included routine, drama, and ceremony visits. He observed that visits with frequent attenders were often “maintenance ceremonies” with a prescribed repetitive format guided by a skilled physician. The “ritual” or “tango” visit types described in our study were the most obvious instances of “maintenance ceremonies” that were based on a high level of familiarity between patient and physician. The ceremonies in such visits may not be based on published medical guidelines or outcomes studies but are likely, in some instances, to represent an effective and appropriate therapeutic response by a physician who has developed an intimate understanding of the complex medical and psychosocial needs of a patient. In other cases such a visit pattern may be an inappropriate pattern of care that allows the familiar routines to take the place of consideration of different diagnostic or therapeutic directions.

Conclusions

Our study provides a foundation for further understanding the variation in family practice patient encounters and holds implications for a number of other issues that require further study. Any evaluation of the appropriateness of outpatient utilization patterns should recognize the wide range of encounter types that may not be apparent from studies of administrative databases. Interventions designed to reduce “inappropriate” utilization also should consider such factors. Our study also provides a basis for teaching students and physicians how to understand and manage the complexities and variation of outpatient primary care encounters. By categorizing and reflecting on the pattern of visit types over time for particular patients, physicians may gain insight into how best to care for “difficult” frequent attenders.

Acknowledgments

Our study was supported by a grant from the Agency for Healthcare Research and Quality (R01 HS08776), a grant from the Health Resources and Services Administration (5D32HP10231), and a Family Practice Research Center grant from the American Academy of Family Physicians. We are grateful to the physicians, staff, and patients from the 18 practices, without whose participation this study would not have been possible. The authors also wish to thank Kurt C. Stange, MD, PhD, who provided helpful comments on earlier drafts of this paper.

In primary care practice, patients in the top 3% for attendance generate 15% of total office visits.¹ “Frequent attenders” are more likely to be older, divorced or widowed, in lower socioeconomic groups, and to have multiple physical and psychosocial ills, and vague physical symptoms with no obvious etiology.^2-7 At one extreme, frequent attenders are “heartsink patients” as described by O’Dowd⁸—individuals whose demands, behaviors, and dissatisfaction with care give the “doctor and staff a feeling of ‘heartsink’ every time they consult.” The memorable and sometimes overwhelming nature of encounters with such patients may lead to the conclusion that difficult encounters are the rule among frequent attenders.

Although the demographics and disease patterns among frequent attenders have been described, the detailed characteristics of their encounters with physicians have not been elucidated.

Multiple factors may influence the content of physician-patient encounters, including the character and severity of a patient’s symptoms, the level of patient concern, the patient’s interpersonal style, the interviewing style of the physician, the complexity of the patient’s medical and psychosocial difficulties, and the level of comfort and trust between patient and physician.^9,10 A better understanding of frequent attender visits could help guide more efficient management of patients’ problems and improve satisfaction for both the patient and the physician. The purpose of our paper is to describe the characteristics of encounters between family physicians and adult frequent-attender patients.

Methods

We drew the data used for this analysis from The Prevention and Competing Demands in Primary Care Study. This multimethod comparative case study of 18 community-based family practices involved extensive direct observation of clinical encounters and office systems by field researchers who spent 4 weeks or more in each practice. The primary data for this analysis were collected by field researchers who were trained to take chronological notes while observing outpatient encounters and later dictate them to create detailed descriptions of each encounter. Details of the sampling and data collection can be found elsewhere in this issue.¹¹

Field researchers directly observed and dictated descriptions of approximately 30 patient encounters with each of the more than 50 clinicians, and they audited the medical records of each of these patients. Visits with 1194 adults 21 years and older were observed. Non-pregnant patients were sorted by their number of visits in the previous 2 years as determined by medical record review. Those in the top 5% for visit frequency were selected for analysis.

Three family physician researchers with experience in qualitative methodology used an immersion/crystallization style to explore and characterize the physician-patient encounters.^12,13 In this interpretive style each researcher read the field notes independently and recorded his or her summary comments and possible themes. Over several sessions, the 3 reviewers met to compare observations, review each patient encounter in detail together, resolve disagreements regarding the characteristics of each physician-patient encounter, and identify major themes from the data. These themes gradually developed into a 3-dimensional framework for examining all encounters. This framework was refined through additional review of field notes and discussions with study collaborators familiar with the data. The field notes were further reviewed to identify individual encounters that illustrated the range of encounter types represented by the descriptive framework.

Although the initial purpose of the study did not include comparison of frequent attenders with other patients, it was unclear whether the descriptive framework that had emerged from the analysis was unique to frequent attenders or if it could be applied to patients who visit much less often. Thus, we reviewed additional encounters by non-frequent attenders to explore the applicability and relevance of the framework to other patients. Visits by patients with only 2 to 6 visits in the previous 2 years (non-frequent attenders) were matched by age and sex to the frequent attender sample (n=62). Each of the 3 reviewers independently read and characterized each non-frequent attender visit using the descriptive framework, then met to compare findings and determine major themes when comparing visits by frequent and non-frequent attenders.

Results

Frequent Attender Visits

Adult patients in the top 5% for visit frequency (n=62) had 25 or more visits over the previous 2 years Figure 1. The average number of encounters for frequent attenders over the previous 2 years was 32 (range= 25-52). The average age of these patients (65 years) was greater than that for all other adult patients (51 years), while the proportion of women was similar for frequent attenders and all other adult patients (68% and 66%, respectively).

Encounters with frequent attenders reflected a wide range of biomedical and psychosocial complexity and a variety of physician and patient communication patterns. We identified 3 major dimensions that distinguished differences in the content and characteristics of the encounters:

1. Biomedical complexity varied considerably in terms of the number and complexity of acute or chronic illnesses. During some visits only a single straightforward biomedical issue was addressed (biomedical=low), while other visits included review of more complex or more numerous acute or chronic conditions (biomedical=high).
2. Psychosocial complexity was evidenced by a patient’s need for reassurance or counseling from the clinician regarding family or relationship issues, discussions regarding lifestyle issues, a perceived need for general social support, an expression of nonspecific somatic symptoms, or presentation of symptoms of depression or anxiety. Some visits reflected minimal psychosocial issues (psychosocial=low), while others encompassed a wide range of expressed emotions, concerns, or symptoms related to psychosocial issues (psychosocial=high).
3. Dissonance between the patient and physician included a lack of mutual understanding and agreement between the patient and clinician, or frustration and/or confusion regarding diagnostic conclusions, goals and direction of treatment, and follow-up plans. Many visits included friendly chatting between the physician and the patient with evidence for a warm familiar understanding between patients, physicians, nurses, and office staff, and usually ended with an organized, mutually agreeable plan of action (dissonance=low). However, in some encounters patient concerns or symptoms were either only partially addressed by the physician or there seemed to be disagreement, confusion, frustration, or conflict about the diagnosis or therapeutic plan (dissonance=high).

A three-dimensional framework emerged from these themes to characterize the range of encounter types with frequent attenders Figure 2. Each encounter could be rated along a continuum from low to high on each axis, resulting in a wide range of visit types. The following description of frequent attender visits illustrates the 8 combinations of the lowest and highest dimension ratings within this framework:

1. Simple medical (biomedical=low, psychosocial=low, dissonance=low; n=12): An 83-year-old man came to the office with an acute foot injury. The physician completed a focused history and physical examination. A radiograph was negative for fracture. The patient expressed no other symptoms and had no questions about the prescribed treatment.
2. Ritual visit (biomedical=low, psychosocial=high, dissonance=low; n=9): A 54-year-old woman with chronic low back pain was seen first by a nurse practitioner who chatted with the patient about a number of relationship and lifestyle issues while giving her multiple lidocaine trigger point injections in the low back and buttocks, a procedure that had occurred during previous visits. When asked by the nurse practitioner whether she needed any other trigger points injected, the patient answered “no” and got dressed. The physician entered the room 15 minutes later and asked, “Should I do another spot for you?” During a number of additional trigger point injections by the physician, the patient discussed additional family and relationship issues with the physician.
3. Complicated medical (biomedical=high, psychosocial=low, dissonance=low; n=15): A 68-year-old man presented for follow-up of diabetes mellitus type 2, recent myocardial infarction, smoking cessation efforts, and an injured hand from a recent fall. After history taking and physical examination were completed, a plan for each problem was stated and readily agreed to by the patient. The encounter was efficient and friendly, including a few inside jokes between the physician and the patient. There was no mention of psychosocial issues or related symptoms.
4. The tango (biomedical=high, psychosocial=high, dissonance=low; n=11): A 49-year-old woman with multiple minor biomedical problems completed “the most unusual visit” that the nurse researcher had ever observed. The encounter began when the patient “just barged in” to the physician’s administrative office and asked to watch a TV game show. After watching the show together for a few minutes they moved to the examination room where the patient proceeded to “tell the doctor what to do” for her multiple complaints. The physician appeared to take all this in stride and facilitated a series of friendly negotiations for a list of medical issues including sore throat, arthritis pain, and chronic respiratory symptoms. There seemed to be a mutually acceptable resolution and a plan for each issue that emerged from this complicated “dance”.
5. Simple frustration (biomedical=low, psychosocial=low, dissonance=high; n=1): Only one visit in the sample had elements of this visit type but was not a clear exemplar, since the dissonance expressed by the patient was not directly related to the physician encounter itself. A 71-year old woman with severe back pain was unhappy about “waiting for a week to get something done” for her pain, and both she and her husband expressed significant frustration about difficulties with scheduling an epidural injection at the hospital. The physician assisted with rescheduling the procedure for the following day, but some confusion remained at the end of the visit as to where the patient was to meet with the anesthesiologist. After the visit the patient’s husband commented to the field researcher that “we live next door to (this physician), so we have to give him a hard time.”
6. Psychosocial disconnect (biomedical=low, psychosocial=high, dissonance=high; n=7): A 33-year-old man came to the office for a chronic leg ulcer caused by an underlying metabolic disorder. After he briefly checked the ulcer and applied a new dressing, the physician expressed dismay and frustration about the patient’s ongoing reluctance to quit smoking. He lectured the patient about the harm caused by smoking, particularly given his underlying condition. They discussed options for smoking cessation and other lifestyle issues. The physician then encouraged the patient to follow through with his previous suggestion to screen his infant son for metabolic disorders. The patient was reluctant, since the child’s pediatrician “didn’t seem too concerned.” The physician replied that “the pediatrician should be concerned” and to “consider taking him to another physician.” During their discussion of lifestyle and family concerns, there was no indication that the patient agreed with the physician’s advice or planned to take action.
7. Medical disharmony (biomedical=high, psychosocial=low, dissonance=high; n=4): A 52-year-old woman presented with left-sided chest and shoulder pain. There was confusion regarding which medications she was currently taking for hypertension, sleep disorder, headache, and gastrointestinal symptoms. During the physical examination, active range of motion of both shoulders elicited no increase in pain. The only stated plan by the physician was that a “short burst of steroid” was best because “it’s an inflammatory thing, I think.” After the physician left the room the patient repeated her concern to the observer that she had “a lot of pain in my shoulder.” There were no expressed psychosocial needs and little evidence for significant interpersonal connection between the physician and the patient during the visit.
8. The heartsink visit (biomedical=high, psychosocial=high, dissonance=high; n=3): A 57-year-old woman came to the office with multiple minor medical problems and chronic depression. With English as her second language, a communication barrier complicated the encounter. During the visit diagnoses of urinary tract infection, arthritis, and gastritis were discussed, an abdominal radiograph obtained, and new medications prescribed. The patient cried as she described a number of sick family members who lived in Mexico and implied that she was unable to visit them because the physician would not adequately treat her pain. She threatened to “go on the street and get any drug I want for pain.” Although there was evidence that the physician had been compassionate and persistent in his attempts to assist her, there seemed to be little agreement with the patient on the direction for further diagnosis or therapy. Seemingly frustrated by the encounter, the physician concluded the visit by saying “Well, we will see you back here in a month and you’ll be feeling better, right?” The patient “kind of looked at him” and said, “Okay, bye, Doctor.”

Non-Frequent Attender Visits

The researchers were able to easily categorize non-frequent attender visits using the descriptive framework Figure 2. A majority of non-frequent attender visits (87%) were classified along the “biomedical” continuum from type 1 (simple medical; n=34) to type 3 (complex medical; n=20) visits. Psychosocial complexity and dissonance were less prominent and were addressed less often than with frequent attenders, and only one visit approached the heartsink corner of the framework. Non-frequent attender visits encompassed fewer exchanges of humor and small talk than observed in the frequent attender encounters and generally showed less evidence of familiarity between patient and physician. The visits were less dramatic in the range of characteristics defined by the descriptive framework and were less memorable than those of the frequent attenders.

Discussion

Our study is the first to provide a detailed description of the characteristics of encounters between family physicians and adult frequent attenders. We found wide variation in the content of these encounters. A framework emerged that describes the degree of biomedical complexity, psychosocial complexity, and dissonance between the patient and the physician for each encounter. Although previous epidemiologic studies and case series suggest that frequent attenders may generate many difficult heartsink visits, the encounters we studied were scattered across the entire 3-dimensional framework from very simple single issue visits to highly complex emotionally taxing visits.^2-7,14 The dimensions of the descriptive framework were applicable to non-frequent attender visits, but the range of psychosocial complexity and dissonance was greater among visits by frequent attenders.

Many frequent attenders seemed to have developed an intricate and harmonious relationship with the physician and the office staff and nurses in the practice. Visits by frequent attenders often included friendly chatting and humor among patients, staff, and physicians. These findings are consistent with the Direct Observation of Primary Care (DOPC) study, where chatting was a part of 69% of all visits to family physicians and accounted for almost 8% of overall visit time.¹⁵ Older patients who had longer visits and spent more time chatting with their physicians in the DOPC study reported greater satisfaction with care.¹⁶ We hypothesize that patients who find a “medical home” where they can talk comfortably with physicians and staff and gain a level of general social support are likely to return often. If such familiarity is interrupted by seeing a physician other than the patient’s continuity provider, as happened in the “medical disharmony” visit we described, confusion about treatment or other visit-specific dissonance may result between the treating physician and the patient.

The content of an encounter is influenced by physicians’ interviewing skills and techniques and whether patients voice all of their concerns, symptoms, and health questions during the visit.^10,17,18 Many factors determine whether psychosocial concerns are elicited and addressed during a visit as described in another article in this issue.¹⁹ Over multiple visits, patients and physicians are likely to reach an equilibrium of expectations regarding patterns of communication and to develop mutually acceptable parameters for the relationship. Frequent attenders have many opportunities to learn a physician’s style and approach to medical and psychosocial problems. Some encounters in our study suggested that a ritual pattern of discussion or visit procedures had developed over time within a trust-filled continuity relationship.

Patients who develop a strong continuity relationship may be less likely to present a question or a request that they know will not be agreed to by the physician. This may explain why the “simple frustration” visit type was not well represented in the frequent attender sample. When patients disagree with physicians on straightforward treatment issues, such as antibiotic prescription for an uncomplicated upper respiratory infection, they usually will either come to some understanding and acceptance of the physician’s views or eventually seek care from another physician.

Limitations

Our descriptive study has a number of limitations. Field notes from the nurse observers described the interaction between the physician and patient and included subjective interpretations of each encounter that may not have accurately reflected the tone of the physician-patient interaction. The observers did not actively seek information about patients’ unvoiced concerns, thus our conclusions regarding the degree of dissonance in the encounter were implied only from written observations. Other than occasional field notes recorded from physicians’ comments after the patient left the office, the observers also did not measure the physician’s emotional response to encounters. These limitations may have led us to misclassify a given encounter on one or more of the 3 dimensions. Also, we cannot conclude from these data whether the sampled visit for each patient represented a typical visit for that patient, and we cannot judge the appropriateness of the care provided.

We reported the number of visits in each category to provide a general impression of the variation of visit types among frequent attenders and how it differed from the non-frequent attender group. Given the qualitative study design and data collection methods, the distribution of visit types may not accurately represent all frequent attender visits to the physician practices in this study and may not be generalizable to other clinical settings.

For some patients frequent attendance is appropriate, and a one-size-fits-all approach to reducing frequent attendance is unlikely to be effective. In his qualitative study of the management styles of 2 experienced family physicians, Miller20 described a visit typology that included routine, drama, and ceremony visits. He observed that visits with frequent attenders were often “maintenance ceremonies” with a prescribed repetitive format guided by a skilled physician. The “ritual” or “tango” visit types described in our study were the most obvious instances of “maintenance ceremonies” that were based on a high level of familiarity between patient and physician. The ceremonies in such visits may not be based on published medical guidelines or outcomes studies but are likely, in some instances, to represent an effective and appropriate therapeutic response by a physician who has developed an intimate understanding of the complex medical and psychosocial needs of a patient. In other cases such a visit pattern may be an inappropriate pattern of care that allows the familiar routines to take the place of consideration of different diagnostic or therapeutic directions.

Conclusions

Our study provides a foundation for further understanding the variation in family practice patient encounters and holds implications for a number of other issues that require further study. Any evaluation of the appropriateness of outpatient utilization patterns should recognize the wide range of encounter types that may not be apparent from studies of administrative databases. Interventions designed to reduce “inappropriate” utilization also should consider such factors. Our study also provides a basis for teaching students and physicians how to understand and manage the complexities and variation of outpatient primary care encounters. By categorizing and reflecting on the pattern of visit types over time for particular patients, physicians may gain insight into how best to care for “difficult” frequent attenders.

Acknowledgments

Our study was supported by a grant from the Agency for Healthcare Research and Quality (R01 HS08776), a grant from the Health Resources and Services Administration (5D32HP10231), and a Family Practice Research Center grant from the American Academy of Family Physicians. We are grateful to the physicians, staff, and patients from the 18 practices, without whose participation this study would not have been possible. The authors also wish to thank Kurt C. Stange, MD, PhD, who provided helpful comments on earlier drafts of this paper.

Primary care clinicians are being asked to deliver better-quality services with fewer resources. The literature has many examples of shortfalls in key physician services in primary care settings, including the delivery of preventive,^1-7 chronic disease,^8-15 and mental health services.^16-19 The Institute of Medicine of the National Academy of Sciences, recognizing the importance of systems in the delivery of high-quality health care, has called for new emphasis on health care teams as a way to reduce medical error and improve quality of care.²⁰ Collaborative team models have been proposed as a means to achieve a higher-quality level of clinical services.^20-24 Are physicians maximizing the human resources they have in their offices by fully involving clinical staff in the delivery of preventive care? Who are the clinical staff in physician offices, what is their training, and what roles are they being asked to play?

Integrated systems of care, where physicians, nurses, and other professional and nonprofessional care-givers deliver services, have promoted the theoretical notion of greater interdisciplinary collaboration in the practice setting.^20-25 At the same time, organizations such as the Medical Group Management Association have suggested optimal patient care staff-to-physician ratios for outpatient primary care practices. Their recommendations are based on surveys conducted in large group practices. However, the extent to which actual staffing patterns accommodate the diversity of practices is not well understood. Even less understood is the link between the idea of better systems of collaboration in patient care and the practical decisions that are made in determining the composition of clinical support staff. Also, assumptions about roles played by office staff underlie all staffing recommendations. Yet, competitive health care market forces may have forced many practices to seek less expensive help to provide patient care.^26-31 This could result in many traditional nursing roles being performed by non-nursing patient care staff whose task training is too limited in scope to enhance and contribute flexibly to recommendations for collaborative care.^30-35

The medical and nursing literature on collaborative staffing patterns has generally focused on the integration of nurse practitioners into the delivery of primary care services.^25,36-38 Because of their advanced practice status, we classify nurse practitioners as primary care clinicians and do not include them in our discussion of patient care staff. When we use the term “nurse” we are referring to registered nurses (RNs) and licensed practical nurses (LPNs), and do not differentiate between 2-, 3-, or 4-year nursing graduates with RN licensure. A review of preparatory programs for nursing roles reveals that American associate degree, diploma, and baccalaureate nursing programs have not emphasized outpatient office-based care roles for nurses Table 1.^39,40 Yet primary care practices employ large numbers of nonphysician patient care staff including professional and practical nurses to manage the day-to-day services required to provide care. These staff members are often generically called “nurses” by patients and those in office settings, but this is not always the case. Many auxiliary clinical staff are certified medical assistants (CMA), medical assistants (MA), and even radiology technicians (RTs) who have been cross-trained to perform patient care roles.

We explored the professional and practical nursing and auxiliary patient care staffing patterns of 18 community-based Midwestern family practices and describe the different roles patient care staff members assume in practice. We examine the education, training, and licensure of nursing and auxiliary staff and compare these with the roles these individuals play in patient care activities. The results have important implications for the design of efficient office staffing patterns that match human resources with service delivery goals and for the future education of nursing, mid-level professional, and auxiliary personnel.

Methods

The data used for this analysis were collected as part of the Prevention and Competing Demands in Primary Care study, a multimethod comparative case study that examined the organizational and clinical structures and process of 18 community-based family practices. Each practice was studied using extensive direct observation of office systems and of clinical encounters by field researchers who spent 4 weeks or more in each practice. Individual depth interviews with each clinician, many of the practice staff, and members of the community were used to obtain their perspectives of the practice. Details of the sampling and data collection are available elsewhere in this issue.⁴¹

This analysis was performed by a multidisciplinary team that had been involved in the analysis of the larger project. Team members began this analysis by independently reading and re-reading details of the contributions made by clinical support staff, both in the practice and in the patient encounters. Each team member independently made notes detailing important tasks and roles that staff performed. Afterward, team members met to compare and contrast their findings. This discussion led to the identification of staffing patterns that became our codebook. After establishing the codebook, team members met on several occasions to again methodically review staffing data on each practice. During these sessions, they constructed a large table with a column for each staffing pattern in the codebook and a row for each practice. In each of the table cells, they recorded whether the practice exhibited the staffing pattern characteristic. After the completion of this step and the attendant discussion, overall themes emerged.

Results

Details of the patient care staff in each of the 18 practices are presented in Table 2. Four practices had a solo physician, each with at least one physician assistant (PA) or nurse practitioner (NP); 6 practices had 2 physicians; 7 practices included 3 physicians; and another had 8 physicians. Ten of the 18 practices employed RNs, with 4 of these having both RNs and LPNs. Another 5 practices had LPNs but no RNs, leaving only 3 practices without either. These 3 employed CMAs as their most highly trained patient care staff. Rounding out the patient care staff in the 18 practices were combinations of CMAs, MAs, RTs, and on-the-job-trained assistants. The overall majority of practices used non-nursing personnel as the predominate patient care staff. The staff-to-clinician ratio ranged from a low of 0.5 to a high of 3.3. These figures have been corrected to indicate staff and physician full-time equivalents, since a number of staff members filled only part-time positions.

A key observation that emerged from the data was that the term “nurse” referred to any individual who performed clinical duties related to caring for patients in the practice. All practices employed varied combinations of patient care staff including RNs, LPNs, CMAs, both trained and untrained MAs, and even RTs who had been cross-trained to perform patient care duties. Typical duties ranged from simple tasks of moving patients in and out of examination rooms and taking vital signs to assisting with procedures and treatments and patient teaching Table 3. The generalized tasks of the “nurse”, however, belied the diversity of staffing roles and functions that characterized each individual practice. Approximately half of the practices (55%, 10 of 18) employed RNs, and 45% (8 of 18) did not. To analyze these data, we separated practices that hired RNs from those that did not, then compared and contrasted them.

Hiring Practices Related to Patient Care Staff

Nursing and other patient care staff roles appeared to be influenced by a number of complex factors embedded in the context of the practice situation that transcended staff educational background and training. Key factors influencing the hiring of patient care staff were the expectations and vision of physicians and administrators, and these were further influenced by economics and labor pool availability. Who was hired for patient care staff positions did not necessarily depend on the prospective employee’s clinical training but was determined by expectations of the practice leadership who tended to hire a “person” rather than a “position.” Explicit practice goals about staffing did not appear to be considered in these decisions, and the larger system ramifications were not recognized. Two examples illustrate differences in hiring practices, one where physicians set expectations about who to hire, and another where administrators (health systems) set those expectations.

Example 1. Suburban Family Practice is a 3-physician practice that was established by a hospital health system in an affluent suburban area of a large city. Health system management determined which types of staff to hire and decided not to hire RN staff, because they thought nurses were overqualified for the work they envisioned in the office setting. Management hired bright nonprofessional office managers whom they trained to teach unskilled staff members how to do tasks (using protocols and checklists) that are required in the delivery of patient services during encounters. Patient care staff were also trained using written scripts to learn how to communicate with patients both in the office and on the telephone. Training included attention to being pleasant and friendly, and to calling the patient by name. Overall, patients seemed pleased with the quality of care given by staff that fieldworkers reported to be pleasant but basic.

Example 2. Rural Community Family Practice is a rural 2-physician practice that employs 2 physician assistants. Unlike most practices in the sample, the patient care staff consisted of 2 full-time RNs, a part-time RN, and another on-call RN. Additional patient care staff included a full-time and a part-time MA and a part-time LPN. One of the full-time RNs worked as the head nurse for the practice. The office had a separate business manager. The 2 physicians were each paired with RNs, while one of the PAs was paired with a MA and the other with a LPN. The office personnel were cross-trained to help each other and did so effectively and cheerfully. Everyone’s attitude supported a universally held desire to see the practice run efficiently, and a team spirit was noticeable, making for a pleasant working environment. The practice philosophy, values, and goals appeared to flow from one physician’s selfless mission-driven patient care focus, and he hired staff willing to go the extra mile with him, putting patients first. Despite having RNs on the staff, they did fairly basic tasks such as counseling and patient education and not collaborative patient care.

In general we found that the leadership philosophy of physicians and administrators as in the examples was very important in the configuring of staff patterns, but it was not the only factor that influenced hiring practices. Other factors such as geography and economics were also influential but not always as might be expected, leading us to consider the importance of access to trained personnel. The Midwestern area of the study is predominately rural with one medium-sized and several smaller cities and towns. Colleges that prepare various types of nursing and technical personnel are clustered predominantly in the eastern section of the study area with a few educational facilities scattered through the middle east-west corridor of the region.

Although one might anticipate that rural practices have less access to hiring trained staff than more urban practices where colleges and technical schools graduate prepared personnel, this was not always the case. In some cases, staff members told of having left their small communities to obtain training elsewhere and later returned. Other factors such as practice economics played a role. One community with a 3-physician practice was able to recruit 2 RNs and a LPN, as well as other assistants, but the practice was in competition with the community hospital and nursing home in hiring and retaining its trained staff. Another 3-physician rural community practice found that trained staff were simply unavailable. The lead physician in this practice stated that he simply had to train staff himself. Yet, economics was a significant factor in hiring practices, as illustrated by the first example of the health system practice. We found that some health systems constrained the hiring of well-trained and more costly staff, while other health systems did hire RNs but used them more as managers than for patient care. If practices did not always hire on the basis of training but rather were influenced by geography and economics, what were the resulting roles that nurses and other patient care staff played?

Patient Care Staff Roles

Roles and responsibilities of patient care staff found in the practices are listed in Table 3. The roles patient care staff assumed in these primary care practices were not determined by education, training, or even by licensure as outlined in Table 1.Cross-training for patient care staff was encouraged by many practices. As an example, in practices where RN and LPN staff were simultaneously employed treatments, procedures, immunizations, and injections were often part of both of their responsibilities. In practices where RN and LPN staff were not employed, however, other less-skilled staff members were trained to assist with these tasks.

We found distinctions in roles between professional nursing staff and lesser-trained staff in some but not all practices, with most practices cross-training staff with different backgrounds to perform basic tasks. Although not common, distinctions were identified in the realm of patient management, particularly regarding the need to use independent judgment and the potential for leadership allowed by physicians and administrators. These differences are highlighted by the following 2 examples.

Example 3. Rural Group Practice is a high-volume rural multispecialty practice where registered nurses triaged patients using their judgment as to which patients should be seen and which handled using the telephone. RNs also did considerable patient education in addition to all the other duties they performed. One RN in particular did all cardiac rehabilitation and dietary patient education. Physicians and nurses worked in pairs, with physicians giving nurses considerable autonomy in managing patients. A sense of camaraderie between the physicians and nurses was evident in the working environment. One physician in this practice recognized the leadership potential of his nurse colleague and encouraged her continued education.

Example 4. Downtown Family Practice was an inner-city solo physician practice and was part of a health system where the physician brought the staff, including a CMA and an MA, with him from a previous practice. The physician saw approximately 30 to 35 patients a day in a practice with 2 examination rooms. Patient care staff members felt under stress trying to keep up with the physician’s pace and often became short-tempered managing telephone inquiries and moving patients in and out of examining rooms. Although the physician was extraordinarily patient centered, the staff did not express the same commitment to serving patients and at times exhibited discomfort or uncertainty going beyond limited patient care duties. The clinician assumed many “nursing” roles himself that were performed by other patient care staff in other practices; in fact, it was not uncommon for him to clean the examination room between patients.

Differences in attitude about the patient care staffing role and level of judgment are apparent in these 2 examples, as is the difference in staff capacity to assume a higher level of care. Nurses in the third example were willing and able to assume much more of a patient management role than staff in the fourth example, who were simply task oriented. As quality of care takes on more importance in team-oriented systems of care, these differences in training and capacity would seem to assume greater importance.

Leadership capacity differences among staff were marked and tied to professional training in the latter 2 examples. However, patient care staffing leadership was not tied to professional training in every practice situation. In 6 of the 10 practices with RNs, the nurse exerted little or no clinical leadership. Instead, leadership seemed to be related to the degree to which there were personal or professional connections to the population served and the degree to which an individual’s initiative was supported by practice authorities. We generally saw more leadership among professional and nonprofessional staff in rural areas where individuals knew the patients and were a part of the community. In one case, a CMA in a practice without any professional nurses exerted considerable leadership and had one of the more extensive roles of all staff members studied. Also, leadership seemed dependent on the blessing of the physician or administrative leaders within the practices. Since staff members were employees who are lower in the hierarchy of a practice, the encouragement of authorities was important for sustained leadership to emerge.

Discussion

We found that practices employ a wide range and different mixtures of professional nursing and non-nursing staff. Although patient care staff roles vary widely, they are not necessarily tied to professional training or particular skill sets. This appears to be due in part to physician and/or practice administration values and goals directly affecting the types of staff hired and the roles they ultimately assume. These findings have important implications and are of interest because of the recent articles in the medical, nursing, and management literature on the need to develop collaborative care models in primary care.^20,42

The results of this study indicate that physician and administrative values and goals shape the expectations of staff roles, but these values and expectations are more focused on economics than on larger patient care issues. In reviewing practice documents of vision, values, and goals (where they existed), only 2 affiliated health system practices had strategic matching of staff to the goals of the practice. Instead of looking critically at clinical goals and then matching staff available with those goals, most of practices tried to get by with the minimum educational preparation and number of staff—the values seemed tied to economic returns. Although we do not discount the importance of economics in organizational planning for effective primary care practice, other considerations such as expanding the practice’s ability to provide additional services and staff development opportunities for promotion of staff leadership may have even wider implications in the delivery of primary care services. Without challenges and appropriate development opportunities, staff may become disinterested and bored in their work. With encouragement and in-service training opportunities, nonprofessional staff can develop into excellent service providers as some of the staff we studied proved.

Until recently there has been little exploration of how physicians and others providing primary services to patients could collaborate more effectively with nursing and clinical support patient care staff in the day to day delivery of services. Our findings imply that either staffing patterns need to change to improve and enhance the skill mix of staff or administrative expectations of staff need to better correspond to training backgrounds if such collaboration is to succeed between primary care staff and nursing and auxiliary staff. Practices may be making sound economic decisions by hiring minimally trained staff; however, these hiring and staffing patterns fly in the face of recommendations emerging from other sources,^43,44 including a 4-part series of articles looking at the dimensions of ambulatory nursing role and staffing patterns in Nursing Economics.^45-48 Recent work takes an intensive look at the role of nurses in ambulatory settings that delineates key components and describes the staffing pilot projects and outcomes at the Group Health Cooperative of Puget Sound.^49,50 These studies argue that until there is an alignment of reimbursement with practice goals and corresponding staffing patterns, practices are unlikely to deliver the quality of care that patients deserve.

At the same time, practices do not seem to be taking into account the legal scope of nursing activities or encouraging their nurses to practice up to their levels of education. Also, nursing education, which is often focused on inpatient roles and responsibilities, needs to better address the task of preparing nurses for roles in ambulatory settings, particularly in practices. With physicians pressed with so many acute and chronic care needs, opportunities for teamwork abound, and registered nurses can fill in many of the gaps in primary care.

Limitations

Our study has a number of limitations. One of the most serious is that only 18 practices in a single state were studied, which limits the generalizability of the findings. We also did not link staffing characteristics to outcome measures or explicitly include patient perceptions of staff. Although these limitations are significant, we think the study has important implications for thinking about the configuration of staffing patterns.

Conclusion

Training and qualifications of staff alone do not tell the whole story about staffing patterns in family practices. These patterns are as varied as the practices themselves. Many opportunities exist for practices to engage their nursing and clinical support staff to enhance the quality of clinical services delivered and to provide opportunities for continual staff growth and development. It is clear that staff are malleable and can adapt to varied roles. Also, nursing and clinical support staff can potentially take greater leadership responsibility for patient care, which appears to be important for the creation of high-functioning primary care teams, regardless of staff titles and level of formal training.

Acknowledgments

The data used in our paper came from a study supported by a grant from the Agency for Healthcare Research and Quality (R01 HS08776). A Family Practice Research Center Grant from the American Academy of Family Physicians supported the analyses. We are grateful to the physicians, staff, and patients from the 18 practices, without whose participation our study would not have been possible. We also wish to thank dedicated work of Connie Gibbs and Jen Rouse, who spent countless hours collecting data and Mary McAndrews, who transcribed hundreds of taped interviews and dictated field notes. We would also like to thank Kurt C. Stange, MD, PhD, for reviewing earlier drafts of this manuscript. Dr Crabtree is associated with the Center for Research in Family Practice and Primary Care, Cleveland, New Brunswick, Allentown, and San Antonio.

Primary care clinicians are being asked to deliver better-quality services with fewer resources. The literature has many examples of shortfalls in key physician services in primary care settings, including the delivery of preventive,^1-7 chronic disease,^8-15 and mental health services.^16-19 The Institute of Medicine of the National Academy of Sciences, recognizing the importance of systems in the delivery of high-quality health care, has called for new emphasis on health care teams as a way to reduce medical error and improve quality of care.²⁰ Collaborative team models have been proposed as a means to achieve a higher-quality level of clinical services.^20-24 Are physicians maximizing the human resources they have in their offices by fully involving clinical staff in the delivery of preventive care? Who are the clinical staff in physician offices, what is their training, and what roles are they being asked to play?

Integrated systems of care, where physicians, nurses, and other professional and nonprofessional care-givers deliver services, have promoted the theoretical notion of greater interdisciplinary collaboration in the practice setting.^20-25 At the same time, organizations such as the Medical Group Management Association have suggested optimal patient care staff-to-physician ratios for outpatient primary care practices. Their recommendations are based on surveys conducted in large group practices. However, the extent to which actual staffing patterns accommodate the diversity of practices is not well understood. Even less understood is the link between the idea of better systems of collaboration in patient care and the practical decisions that are made in determining the composition of clinical support staff. Also, assumptions about roles played by office staff underlie all staffing recommendations. Yet, competitive health care market forces may have forced many practices to seek less expensive help to provide patient care.^26-31 This could result in many traditional nursing roles being performed by non-nursing patient care staff whose task training is too limited in scope to enhance and contribute flexibly to recommendations for collaborative care.^30-35

The medical and nursing literature on collaborative staffing patterns has generally focused on the integration of nurse practitioners into the delivery of primary care services.^25,36-38 Because of their advanced practice status, we classify nurse practitioners as primary care clinicians and do not include them in our discussion of patient care staff. When we use the term “nurse” we are referring to registered nurses (RNs) and licensed practical nurses (LPNs), and do not differentiate between 2-, 3-, or 4-year nursing graduates with RN licensure. A review of preparatory programs for nursing roles reveals that American associate degree, diploma, and baccalaureate nursing programs have not emphasized outpatient office-based care roles for nurses Table 1.^39,40 Yet primary care practices employ large numbers of nonphysician patient care staff including professional and practical nurses to manage the day-to-day services required to provide care. These staff members are often generically called “nurses” by patients and those in office settings, but this is not always the case. Many auxiliary clinical staff are certified medical assistants (CMA), medical assistants (MA), and even radiology technicians (RTs) who have been cross-trained to perform patient care roles.

We explored the professional and practical nursing and auxiliary patient care staffing patterns of 18 community-based Midwestern family practices and describe the different roles patient care staff members assume in practice. We examine the education, training, and licensure of nursing and auxiliary staff and compare these with the roles these individuals play in patient care activities. The results have important implications for the design of efficient office staffing patterns that match human resources with service delivery goals and for the future education of nursing, mid-level professional, and auxiliary personnel.

Methods

The data used for this analysis were collected as part of the Prevention and Competing Demands in Primary Care study, a multimethod comparative case study that examined the organizational and clinical structures and process of 18 community-based family practices. Each practice was studied using extensive direct observation of office systems and of clinical encounters by field researchers who spent 4 weeks or more in each practice. Individual depth interviews with each clinician, many of the practice staff, and members of the community were used to obtain their perspectives of the practice. Details of the sampling and data collection are available elsewhere in this issue.⁴¹

This analysis was performed by a multidisciplinary team that had been involved in the analysis of the larger project. Team members began this analysis by independently reading and re-reading details of the contributions made by clinical support staff, both in the practice and in the patient encounters. Each team member independently made notes detailing important tasks and roles that staff performed. Afterward, team members met to compare and contrast their findings. This discussion led to the identification of staffing patterns that became our codebook. After establishing the codebook, team members met on several occasions to again methodically review staffing data on each practice. During these sessions, they constructed a large table with a column for each staffing pattern in the codebook and a row for each practice. In each of the table cells, they recorded whether the practice exhibited the staffing pattern characteristic. After the completion of this step and the attendant discussion, overall themes emerged.

Results

Details of the patient care staff in each of the 18 practices are presented in Table 2. Four practices had a solo physician, each with at least one physician assistant (PA) or nurse practitioner (NP); 6 practices had 2 physicians; 7 practices included 3 physicians; and another had 8 physicians. Ten of the 18 practices employed RNs, with 4 of these having both RNs and LPNs. Another 5 practices had LPNs but no RNs, leaving only 3 practices without either. These 3 employed CMAs as their most highly trained patient care staff. Rounding out the patient care staff in the 18 practices were combinations of CMAs, MAs, RTs, and on-the-job-trained assistants. The overall majority of practices used non-nursing personnel as the predominate patient care staff. The staff-to-clinician ratio ranged from a low of 0.5 to a high of 3.3. These figures have been corrected to indicate staff and physician full-time equivalents, since a number of staff members filled only part-time positions.

A key observation that emerged from the data was that the term “nurse” referred to any individual who performed clinical duties related to caring for patients in the practice. All practices employed varied combinations of patient care staff including RNs, LPNs, CMAs, both trained and untrained MAs, and even RTs who had been cross-trained to perform patient care duties. Typical duties ranged from simple tasks of moving patients in and out of examination rooms and taking vital signs to assisting with procedures and treatments and patient teaching Table 3. The generalized tasks of the “nurse”, however, belied the diversity of staffing roles and functions that characterized each individual practice. Approximately half of the practices (55%, 10 of 18) employed RNs, and 45% (8 of 18) did not. To analyze these data, we separated practices that hired RNs from those that did not, then compared and contrasted them.

Hiring Practices Related to Patient Care Staff

Nursing and other patient care staff roles appeared to be influenced by a number of complex factors embedded in the context of the practice situation that transcended staff educational background and training. Key factors influencing the hiring of patient care staff were the expectations and vision of physicians and administrators, and these were further influenced by economics and labor pool availability. Who was hired for patient care staff positions did not necessarily depend on the prospective employee’s clinical training but was determined by expectations of the practice leadership who tended to hire a “person” rather than a “position.” Explicit practice goals about staffing did not appear to be considered in these decisions, and the larger system ramifications were not recognized. Two examples illustrate differences in hiring practices, one where physicians set expectations about who to hire, and another where administrators (health systems) set those expectations.

Example 1. Suburban Family Practice is a 3-physician practice that was established by a hospital health system in an affluent suburban area of a large city. Health system management determined which types of staff to hire and decided not to hire RN staff, because they thought nurses were overqualified for the work they envisioned in the office setting. Management hired bright nonprofessional office managers whom they trained to teach unskilled staff members how to do tasks (using protocols and checklists) that are required in the delivery of patient services during encounters. Patient care staff were also trained using written scripts to learn how to communicate with patients both in the office and on the telephone. Training included attention to being pleasant and friendly, and to calling the patient by name. Overall, patients seemed pleased with the quality of care given by staff that fieldworkers reported to be pleasant but basic.

Example 2. Rural Community Family Practice is a rural 2-physician practice that employs 2 physician assistants. Unlike most practices in the sample, the patient care staff consisted of 2 full-time RNs, a part-time RN, and another on-call RN. Additional patient care staff included a full-time and a part-time MA and a part-time LPN. One of the full-time RNs worked as the head nurse for the practice. The office had a separate business manager. The 2 physicians were each paired with RNs, while one of the PAs was paired with a MA and the other with a LPN. The office personnel were cross-trained to help each other and did so effectively and cheerfully. Everyone’s attitude supported a universally held desire to see the practice run efficiently, and a team spirit was noticeable, making for a pleasant working environment. The practice philosophy, values, and goals appeared to flow from one physician’s selfless mission-driven patient care focus, and he hired staff willing to go the extra mile with him, putting patients first. Despite having RNs on the staff, they did fairly basic tasks such as counseling and patient education and not collaborative patient care.

In general we found that the leadership philosophy of physicians and administrators as in the examples was very important in the configuring of staff patterns, but it was not the only factor that influenced hiring practices. Other factors such as geography and economics were also influential but not always as might be expected, leading us to consider the importance of access to trained personnel. The Midwestern area of the study is predominately rural with one medium-sized and several smaller cities and towns. Colleges that prepare various types of nursing and technical personnel are clustered predominantly in the eastern section of the study area with a few educational facilities scattered through the middle east-west corridor of the region.

Although one might anticipate that rural practices have less access to hiring trained staff than more urban practices where colleges and technical schools graduate prepared personnel, this was not always the case. In some cases, staff members told of having left their small communities to obtain training elsewhere and later returned. Other factors such as practice economics played a role. One community with a 3-physician practice was able to recruit 2 RNs and a LPN, as well as other assistants, but the practice was in competition with the community hospital and nursing home in hiring and retaining its trained staff. Another 3-physician rural community practice found that trained staff were simply unavailable. The lead physician in this practice stated that he simply had to train staff himself. Yet, economics was a significant factor in hiring practices, as illustrated by the first example of the health system practice. We found that some health systems constrained the hiring of well-trained and more costly staff, while other health systems did hire RNs but used them more as managers than for patient care. If practices did not always hire on the basis of training but rather were influenced by geography and economics, what were the resulting roles that nurses and other patient care staff played?

Patient Care Staff Roles

Roles and responsibilities of patient care staff found in the practices are listed in Table 3. The roles patient care staff assumed in these primary care practices were not determined by education, training, or even by licensure as outlined in Table 1.Cross-training for patient care staff was encouraged by many practices. As an example, in practices where RN and LPN staff were simultaneously employed treatments, procedures, immunizations, and injections were often part of both of their responsibilities. In practices where RN and LPN staff were not employed, however, other less-skilled staff members were trained to assist with these tasks.

We found distinctions in roles between professional nursing staff and lesser-trained staff in some but not all practices, with most practices cross-training staff with different backgrounds to perform basic tasks. Although not common, distinctions were identified in the realm of patient management, particularly regarding the need to use independent judgment and the potential for leadership allowed by physicians and administrators. These differences are highlighted by the following 2 examples.

Example 3. Rural Group Practice is a high-volume rural multispecialty practice where registered nurses triaged patients using their judgment as to which patients should be seen and which handled using the telephone. RNs also did considerable patient education in addition to all the other duties they performed. One RN in particular did all cardiac rehabilitation and dietary patient education. Physicians and nurses worked in pairs, with physicians giving nurses considerable autonomy in managing patients. A sense of camaraderie between the physicians and nurses was evident in the working environment. One physician in this practice recognized the leadership potential of his nurse colleague and encouraged her continued education.

Example 4. Downtown Family Practice was an inner-city solo physician practice and was part of a health system where the physician brought the staff, including a CMA and an MA, with him from a previous practice. The physician saw approximately 30 to 35 patients a day in a practice with 2 examination rooms. Patient care staff members felt under stress trying to keep up with the physician’s pace and often became short-tempered managing telephone inquiries and moving patients in and out of examining rooms. Although the physician was extraordinarily patient centered, the staff did not express the same commitment to serving patients and at times exhibited discomfort or uncertainty going beyond limited patient care duties. The clinician assumed many “nursing” roles himself that were performed by other patient care staff in other practices; in fact, it was not uncommon for him to clean the examination room between patients.

Differences in attitude about the patient care staffing role and level of judgment are apparent in these 2 examples, as is the difference in staff capacity to assume a higher level of care. Nurses in the third example were willing and able to assume much more of a patient management role than staff in the fourth example, who were simply task oriented. As quality of care takes on more importance in team-oriented systems of care, these differences in training and capacity would seem to assume greater importance.

Leadership capacity differences among staff were marked and tied to professional training in the latter 2 examples. However, patient care staffing leadership was not tied to professional training in every practice situation. In 6 of the 10 practices with RNs, the nurse exerted little or no clinical leadership. Instead, leadership seemed to be related to the degree to which there were personal or professional connections to the population served and the degree to which an individual’s initiative was supported by practice authorities. We generally saw more leadership among professional and nonprofessional staff in rural areas where individuals knew the patients and were a part of the community. In one case, a CMA in a practice without any professional nurses exerted considerable leadership and had one of the more extensive roles of all staff members studied. Also, leadership seemed dependent on the blessing of the physician or administrative leaders within the practices. Since staff members were employees who are lower in the hierarchy of a practice, the encouragement of authorities was important for sustained leadership to emerge.

Discussion

We found that practices employ a wide range and different mixtures of professional nursing and non-nursing staff. Although patient care staff roles vary widely, they are not necessarily tied to professional training or particular skill sets. This appears to be due in part to physician and/or practice administration values and goals directly affecting the types of staff hired and the roles they ultimately assume. These findings have important implications and are of interest because of the recent articles in the medical, nursing, and management literature on the need to develop collaborative care models in primary care.^20,42

The results of this study indicate that physician and administrative values and goals shape the expectations of staff roles, but these values and expectations are more focused on economics than on larger patient care issues. In reviewing practice documents of vision, values, and goals (where they existed), only 2 affiliated health system practices had strategic matching of staff to the goals of the practice. Instead of looking critically at clinical goals and then matching staff available with those goals, most of practices tried to get by with the minimum educational preparation and number of staff—the values seemed tied to economic returns. Although we do not discount the importance of economics in organizational planning for effective primary care practice, other considerations such as expanding the practice’s ability to provide additional services and staff development opportunities for promotion of staff leadership may have even wider implications in the delivery of primary care services. Without challenges and appropriate development opportunities, staff may become disinterested and bored in their work. With encouragement and in-service training opportunities, nonprofessional staff can develop into excellent service providers as some of the staff we studied proved.

Until recently there has been little exploration of how physicians and others providing primary services to patients could collaborate more effectively with nursing and clinical support patient care staff in the day to day delivery of services. Our findings imply that either staffing patterns need to change to improve and enhance the skill mix of staff or administrative expectations of staff need to better correspond to training backgrounds if such collaboration is to succeed between primary care staff and nursing and auxiliary staff. Practices may be making sound economic decisions by hiring minimally trained staff; however, these hiring and staffing patterns fly in the face of recommendations emerging from other sources,^43,44 including a 4-part series of articles looking at the dimensions of ambulatory nursing role and staffing patterns in Nursing Economics.^45-48 Recent work takes an intensive look at the role of nurses in ambulatory settings that delineates key components and describes the staffing pilot projects and outcomes at the Group Health Cooperative of Puget Sound.^49,50 These studies argue that until there is an alignment of reimbursement with practice goals and corresponding staffing patterns, practices are unlikely to deliver the quality of care that patients deserve.

At the same time, practices do not seem to be taking into account the legal scope of nursing activities or encouraging their nurses to practice up to their levels of education. Also, nursing education, which is often focused on inpatient roles and responsibilities, needs to better address the task of preparing nurses for roles in ambulatory settings, particularly in practices. With physicians pressed with so many acute and chronic care needs, opportunities for teamwork abound, and registered nurses can fill in many of the gaps in primary care.

Limitations

Our study has a number of limitations. One of the most serious is that only 18 practices in a single state were studied, which limits the generalizability of the findings. We also did not link staffing characteristics to outcome measures or explicitly include patient perceptions of staff. Although these limitations are significant, we think the study has important implications for thinking about the configuration of staffing patterns.

Conclusion

Training and qualifications of staff alone do not tell the whole story about staffing patterns in family practices. These patterns are as varied as the practices themselves. Many opportunities exist for practices to engage their nursing and clinical support staff to enhance the quality of clinical services delivered and to provide opportunities for continual staff growth and development. It is clear that staff are malleable and can adapt to varied roles. Also, nursing and clinical support staff can potentially take greater leadership responsibility for patient care, which appears to be important for the creation of high-functioning primary care teams, regardless of staff titles and level of formal training.

Acknowledgments

The data used in our paper came from a study supported by a grant from the Agency for Healthcare Research and Quality (R01 HS08776). A Family Practice Research Center Grant from the American Academy of Family Physicians supported the analyses. We are grateful to the physicians, staff, and patients from the 18 practices, without whose participation our study would not have been possible. We also wish to thank dedicated work of Connie Gibbs and Jen Rouse, who spent countless hours collecting data and Mary McAndrews, who transcribed hundreds of taped interviews and dictated field notes. We would also like to thank Kurt C. Stange, MD, PhD, for reviewing earlier drafts of this manuscript. Dr Crabtree is associated with the Center for Research in Family Practice and Primary Care, Cleveland, New Brunswick, Allentown, and San Antonio.

Primary care clinicians are being asked to deliver better-quality services with fewer resources. The literature has many examples of shortfalls in key physician services in primary care settings, including the delivery of preventive,^1-7 chronic disease,^8-15 and mental health services.^16-19 The Institute of Medicine of the National Academy of Sciences, recognizing the importance of systems in the delivery of high-quality health care, has called for new emphasis on health care teams as a way to reduce medical error and improve quality of care.²⁰ Collaborative team models have been proposed as a means to achieve a higher-quality level of clinical services.^20-24 Are physicians maximizing the human resources they have in their offices by fully involving clinical staff in the delivery of preventive care? Who are the clinical staff in physician offices, what is their training, and what roles are they being asked to play?

Integrated systems of care, where physicians, nurses, and other professional and nonprofessional care-givers deliver services, have promoted the theoretical notion of greater interdisciplinary collaboration in the practice setting.^20-25 At the same time, organizations such as the Medical Group Management Association have suggested optimal patient care staff-to-physician ratios for outpatient primary care practices. Their recommendations are based on surveys conducted in large group practices. However, the extent to which actual staffing patterns accommodate the diversity of practices is not well understood. Even less understood is the link between the idea of better systems of collaboration in patient care and the practical decisions that are made in determining the composition of clinical support staff. Also, assumptions about roles played by office staff underlie all staffing recommendations. Yet, competitive health care market forces may have forced many practices to seek less expensive help to provide patient care.^26-31 This could result in many traditional nursing roles being performed by non-nursing patient care staff whose task training is too limited in scope to enhance and contribute flexibly to recommendations for collaborative care.^30-35

The medical and nursing literature on collaborative staffing patterns has generally focused on the integration of nurse practitioners into the delivery of primary care services.^25,36-38 Because of their advanced practice status, we classify nurse practitioners as primary care clinicians and do not include them in our discussion of patient care staff. When we use the term “nurse” we are referring to registered nurses (RNs) and licensed practical nurses (LPNs), and do not differentiate between 2-, 3-, or 4-year nursing graduates with RN licensure. A review of preparatory programs for nursing roles reveals that American associate degree, diploma, and baccalaureate nursing programs have not emphasized outpatient office-based care roles for nurses Table 1.^39,40 Yet primary care practices employ large numbers of nonphysician patient care staff including professional and practical nurses to manage the day-to-day services required to provide care. These staff members are often generically called “nurses” by patients and those in office settings, but this is not always the case. Many auxiliary clinical staff are certified medical assistants (CMA), medical assistants (MA), and even radiology technicians (RTs) who have been cross-trained to perform patient care roles.

We explored the professional and practical nursing and auxiliary patient care staffing patterns of 18 community-based Midwestern family practices and describe the different roles patient care staff members assume in practice. We examine the education, training, and licensure of nursing and auxiliary staff and compare these with the roles these individuals play in patient care activities. The results have important implications for the design of efficient office staffing patterns that match human resources with service delivery goals and for the future education of nursing, mid-level professional, and auxiliary personnel.

Methods

The data used for this analysis were collected as part of the Prevention and Competing Demands in Primary Care study, a multimethod comparative case study that examined the organizational and clinical structures and process of 18 community-based family practices. Each practice was studied using extensive direct observation of office systems and of clinical encounters by field researchers who spent 4 weeks or more in each practice. Individual depth interviews with each clinician, many of the practice staff, and members of the community were used to obtain their perspectives of the practice. Details of the sampling and data collection are available elsewhere in this issue.⁴¹

This analysis was performed by a multidisciplinary team that had been involved in the analysis of the larger project. Team members began this analysis by independently reading and re-reading details of the contributions made by clinical support staff, both in the practice and in the patient encounters. Each team member independently made notes detailing important tasks and roles that staff performed. Afterward, team members met to compare and contrast their findings. This discussion led to the identification of staffing patterns that became our codebook. After establishing the codebook, team members met on several occasions to again methodically review staffing data on each practice. During these sessions, they constructed a large table with a column for each staffing pattern in the codebook and a row for each practice. In each of the table cells, they recorded whether the practice exhibited the staffing pattern characteristic. After the completion of this step and the attendant discussion, overall themes emerged.

Results

Details of the patient care staff in each of the 18 practices are presented in Table 2. Four practices had a solo physician, each with at least one physician assistant (PA) or nurse practitioner (NP); 6 practices had 2 physicians; 7 practices included 3 physicians; and another had 8 physicians. Ten of the 18 practices employed RNs, with 4 of these having both RNs and LPNs. Another 5 practices had LPNs but no RNs, leaving only 3 practices without either. These 3 employed CMAs as their most highly trained patient care staff. Rounding out the patient care staff in the 18 practices were combinations of CMAs, MAs, RTs, and on-the-job-trained assistants. The overall majority of practices used non-nursing personnel as the predominate patient care staff. The staff-to-clinician ratio ranged from a low of 0.5 to a high of 3.3. These figures have been corrected to indicate staff and physician full-time equivalents, since a number of staff members filled only part-time positions.

A key observation that emerged from the data was that the term “nurse” referred to any individual who performed clinical duties related to caring for patients in the practice. All practices employed varied combinations of patient care staff including RNs, LPNs, CMAs, both trained and untrained MAs, and even RTs who had been cross-trained to perform patient care duties. Typical duties ranged from simple tasks of moving patients in and out of examination rooms and taking vital signs to assisting with procedures and treatments and patient teaching Table 3. The generalized tasks of the “nurse”, however, belied the diversity of staffing roles and functions that characterized each individual practice. Approximately half of the practices (55%, 10 of 18) employed RNs, and 45% (8 of 18) did not. To analyze these data, we separated practices that hired RNs from those that did not, then compared and contrasted them.

Hiring Practices Related to Patient Care Staff

Nursing and other patient care staff roles appeared to be influenced by a number of complex factors embedded in the context of the practice situation that transcended staff educational background and training. Key factors influencing the hiring of patient care staff were the expectations and vision of physicians and administrators, and these were further influenced by economics and labor pool availability. Who was hired for patient care staff positions did not necessarily depend on the prospective employee’s clinical training but was determined by expectations of the practice leadership who tended to hire a “person” rather than a “position.” Explicit practice goals about staffing did not appear to be considered in these decisions, and the larger system ramifications were not recognized. Two examples illustrate differences in hiring practices, one where physicians set expectations about who to hire, and another where administrators (health systems) set those expectations.

Example 1. Suburban Family Practice is a 3-physician practice that was established by a hospital health system in an affluent suburban area of a large city. Health system management determined which types of staff to hire and decided not to hire RN staff, because they thought nurses were overqualified for the work they envisioned in the office setting. Management hired bright nonprofessional office managers whom they trained to teach unskilled staff members how to do tasks (using protocols and checklists) that are required in the delivery of patient services during encounters. Patient care staff were also trained using written scripts to learn how to communicate with patients both in the office and on the telephone. Training included attention to being pleasant and friendly, and to calling the patient by name. Overall, patients seemed pleased with the quality of care given by staff that fieldworkers reported to be pleasant but basic.

Example 2. Rural Community Family Practice is a rural 2-physician practice that employs 2 physician assistants. Unlike most practices in the sample, the patient care staff consisted of 2 full-time RNs, a part-time RN, and another on-call RN. Additional patient care staff included a full-time and a part-time MA and a part-time LPN. One of the full-time RNs worked as the head nurse for the practice. The office had a separate business manager. The 2 physicians were each paired with RNs, while one of the PAs was paired with a MA and the other with a LPN. The office personnel were cross-trained to help each other and did so effectively and cheerfully. Everyone’s attitude supported a universally held desire to see the practice run efficiently, and a team spirit was noticeable, making for a pleasant working environment. The practice philosophy, values, and goals appeared to flow from one physician’s selfless mission-driven patient care focus, and he hired staff willing to go the extra mile with him, putting patients first. Despite having RNs on the staff, they did fairly basic tasks such as counseling and patient education and not collaborative patient care.

In general we found that the leadership philosophy of physicians and administrators as in the examples was very important in the configuring of staff patterns, but it was not the only factor that influenced hiring practices. Other factors such as geography and economics were also influential but not always as might be expected, leading us to consider the importance of access to trained personnel. The Midwestern area of the study is predominately rural with one medium-sized and several smaller cities and towns. Colleges that prepare various types of nursing and technical personnel are clustered predominantly in the eastern section of the study area with a few educational facilities scattered through the middle east-west corridor of the region.

Although one might anticipate that rural practices have less access to hiring trained staff than more urban practices where colleges and technical schools graduate prepared personnel, this was not always the case. In some cases, staff members told of having left their small communities to obtain training elsewhere and later returned. Other factors such as practice economics played a role. One community with a 3-physician practice was able to recruit 2 RNs and a LPN, as well as other assistants, but the practice was in competition with the community hospital and nursing home in hiring and retaining its trained staff. Another 3-physician rural community practice found that trained staff were simply unavailable. The lead physician in this practice stated that he simply had to train staff himself. Yet, economics was a significant factor in hiring practices, as illustrated by the first example of the health system practice. We found that some health systems constrained the hiring of well-trained and more costly staff, while other health systems did hire RNs but used them more as managers than for patient care. If practices did not always hire on the basis of training but rather were influenced by geography and economics, what were the resulting roles that nurses and other patient care staff played?

Patient Care Staff Roles

Roles and responsibilities of patient care staff found in the practices are listed in Table 3. The roles patient care staff assumed in these primary care practices were not determined by education, training, or even by licensure as outlined in Table 1.Cross-training for patient care staff was encouraged by many practices. As an example, in practices where RN and LPN staff were simultaneously employed treatments, procedures, immunizations, and injections were often part of both of their responsibilities. In practices where RN and LPN staff were not employed, however, other less-skilled staff members were trained to assist with these tasks.

We found distinctions in roles between professional nursing staff and lesser-trained staff in some but not all practices, with most practices cross-training staff with different backgrounds to perform basic tasks. Although not common, distinctions were identified in the realm of patient management, particularly regarding the need to use independent judgment and the potential for leadership allowed by physicians and administrators. These differences are highlighted by the following 2 examples.

Example 3. Rural Group Practice is a high-volume rural multispecialty practice where registered nurses triaged patients using their judgment as to which patients should be seen and which handled using the telephone. RNs also did considerable patient education in addition to all the other duties they performed. One RN in particular did all cardiac rehabilitation and dietary patient education. Physicians and nurses worked in pairs, with physicians giving nurses considerable autonomy in managing patients. A sense of camaraderie between the physicians and nurses was evident in the working environment. One physician in this practice recognized the leadership potential of his nurse colleague and encouraged her continued education.

Example 4. Downtown Family Practice was an inner-city solo physician practice and was part of a health system where the physician brought the staff, including a CMA and an MA, with him from a previous practice. The physician saw approximately 30 to 35 patients a day in a practice with 2 examination rooms. Patient care staff members felt under stress trying to keep up with the physician’s pace and often became short-tempered managing telephone inquiries and moving patients in and out of examining rooms. Although the physician was extraordinarily patient centered, the staff did not express the same commitment to serving patients and at times exhibited discomfort or uncertainty going beyond limited patient care duties. The clinician assumed many “nursing” roles himself that were performed by other patient care staff in other practices; in fact, it was not uncommon for him to clean the examination room between patients.

Differences in attitude about the patient care staffing role and level of judgment are apparent in these 2 examples, as is the difference in staff capacity to assume a higher level of care. Nurses in the third example were willing and able to assume much more of a patient management role than staff in the fourth example, who were simply task oriented. As quality of care takes on more importance in team-oriented systems of care, these differences in training and capacity would seem to assume greater importance.

Leadership capacity differences among staff were marked and tied to professional training in the latter 2 examples. However, patient care staffing leadership was not tied to professional training in every practice situation. In 6 of the 10 practices with RNs, the nurse exerted little or no clinical leadership. Instead, leadership seemed to be related to the degree to which there were personal or professional connections to the population served and the degree to which an individual’s initiative was supported by practice authorities. We generally saw more leadership among professional and nonprofessional staff in rural areas where individuals knew the patients and were a part of the community. In one case, a CMA in a practice without any professional nurses exerted considerable leadership and had one of the more extensive roles of all staff members studied. Also, leadership seemed dependent on the blessing of the physician or administrative leaders within the practices. Since staff members were employees who are lower in the hierarchy of a practice, the encouragement of authorities was important for sustained leadership to emerge.

Discussion

We found that practices employ a wide range and different mixtures of professional nursing and non-nursing staff. Although patient care staff roles vary widely, they are not necessarily tied to professional training or particular skill sets. This appears to be due in part to physician and/or practice administration values and goals directly affecting the types of staff hired and the roles they ultimately assume. These findings have important implications and are of interest because of the recent articles in the medical, nursing, and management literature on the need to develop collaborative care models in primary care.^20,42

The results of this study indicate that physician and administrative values and goals shape the expectations of staff roles, but these values and expectations are more focused on economics than on larger patient care issues. In reviewing practice documents of vision, values, and goals (where they existed), only 2 affiliated health system practices had strategic matching of staff to the goals of the practice. Instead of looking critically at clinical goals and then matching staff available with those goals, most of practices tried to get by with the minimum educational preparation and number of staff—the values seemed tied to economic returns. Although we do not discount the importance of economics in organizational planning for effective primary care practice, other considerations such as expanding the practice’s ability to provide additional services and staff development opportunities for promotion of staff leadership may have even wider implications in the delivery of primary care services. Without challenges and appropriate development opportunities, staff may become disinterested and bored in their work. With encouragement and in-service training opportunities, nonprofessional staff can develop into excellent service providers as some of the staff we studied proved.

Until recently there has been little exploration of how physicians and others providing primary services to patients could collaborate more effectively with nursing and clinical support patient care staff in the day to day delivery of services. Our findings imply that either staffing patterns need to change to improve and enhance the skill mix of staff or administrative expectations of staff need to better correspond to training backgrounds if such collaboration is to succeed between primary care staff and nursing and auxiliary staff. Practices may be making sound economic decisions by hiring minimally trained staff; however, these hiring and staffing patterns fly in the face of recommendations emerging from other sources,^43,44 including a 4-part series of articles looking at the dimensions of ambulatory nursing role and staffing patterns in Nursing Economics.^45-48 Recent work takes an intensive look at the role of nurses in ambulatory settings that delineates key components and describes the staffing pilot projects and outcomes at the Group Health Cooperative of Puget Sound.^49,50 These studies argue that until there is an alignment of reimbursement with practice goals and corresponding staffing patterns, practices are unlikely to deliver the quality of care that patients deserve.

At the same time, practices do not seem to be taking into account the legal scope of nursing activities or encouraging their nurses to practice up to their levels of education. Also, nursing education, which is often focused on inpatient roles and responsibilities, needs to better address the task of preparing nurses for roles in ambulatory settings, particularly in practices. With physicians pressed with so many acute and chronic care needs, opportunities for teamwork abound, and registered nurses can fill in many of the gaps in primary care.

Limitations

Our study has a number of limitations. One of the most serious is that only 18 practices in a single state were studied, which limits the generalizability of the findings. We also did not link staffing characteristics to outcome measures or explicitly include patient perceptions of staff. Although these limitations are significant, we think the study has important implications for thinking about the configuration of staffing patterns.

Conclusion

Training and qualifications of staff alone do not tell the whole story about staffing patterns in family practices. These patterns are as varied as the practices themselves. Many opportunities exist for practices to engage their nursing and clinical support staff to enhance the quality of clinical services delivered and to provide opportunities for continual staff growth and development. It is clear that staff are malleable and can adapt to varied roles. Also, nursing and clinical support staff can potentially take greater leadership responsibility for patient care, which appears to be important for the creation of high-functioning primary care teams, regardless of staff titles and level of formal training.

Acknowledgments

The data used in our paper came from a study supported by a grant from the Agency for Healthcare Research and Quality (R01 HS08776). A Family Practice Research Center Grant from the American Academy of Family Physicians supported the analyses. We are grateful to the physicians, staff, and patients from the 18 practices, without whose participation our study would not have been possible. We also wish to thank dedicated work of Connie Gibbs and Jen Rouse, who spent countless hours collecting data and Mary McAndrews, who transcribed hundreds of taped interviews and dictated field notes. We would also like to thank Kurt C. Stange, MD, PhD, for reviewing earlier drafts of this manuscript. Dr Crabtree is associated with the Center for Research in Family Practice and Primary Care, Cleveland, New Brunswick, Allentown, and San Antonio.

Over the past decade health care has experienced rapid and sometimes volatile change that has affected the quality of patient care.¹ Changes in the structure and financing of health care have compromised family practice’s ability to maintain its core values of comprehensive, coordinated, and continuous care in the context of the family.² For example, the Direct Observation of Primary Care (DOPC) study found that one fourth of patients in participating northeastern Ohio practices were forced to change health care providers during a 2-year period.³ These patients reported less coordination of care, decreased continuity with their new provider, and less satisfactory interpersonal communication.

Reports from the DOPC study also found, however, that the family remains an important focus in patient care despite these disruptions in continuity and coordination. Family issues were discussed in more than 70% of patient encounters, accounting for 10% of visit time.⁴ Patients were accompanied by family members in one third of outpatient visits, providing opportunities for family involvement and even care for 18% of these "non-patients."⁵ These findings were consistent with those of a recent Canadian study that looked at the role of those people who accompany patients into the medical examination⁶ and a qualitative study of the ways the family was integrated into routine patient care.⁷

Although these and other studies confirm that the family is a salient feature in family practice, it is not known how knowledge of the family context affects the process of patient care. Data from the Prevention and Competing Demands in Primary Care Study (PCDPC) were used to answer this question through observation of visits in 18 family practices in Nebraska.

Methods

We used patient encounter data from the PCDPC practice study, an in-depth observational examination of the organizational and clinical structures and process of community-based family practices. Each of the 18 purposefully selected practices was studied using a multimethod comparative case study design that involved extensive direct observation of clinical encounters and office systems by field researchers who spent 4 weeks or more in each practice. Field researchers directly observed and dictated descriptions of approximately 30 patient encounters with each of the more than 50 clinicians and audited medical records of each of these patients. Detailed descriptive field notes documented day-to-day practice operations. Individual depth interviews with each clinician, many of the office staff, and members of the community were used to obtain different perspectives on each practice. Details of the sampling and data collection are available elsewhere in this issue.⁸

We analyzed 1600 of the 1637 outpatient encounters to determine how the family had an influence on patient care (37 visits had insufficient data for effective analysis). Each encounter was coded by one of the co-authors (S.H.) as a family-oriented visit if: (1) the patient’s family member was present in the examination room or (2) medical or health information about the patient’s family was discussed and/or addressed during the visit. Another author (D.S.M.) read each encounter using an editing organizing style⁹ by writing brief jottings about each encounter that described how the family had an impact on these medical encounters. Finally, 2 authors (S.H., D.S.M.) jointly examined these jottings to identify an organizational framework and to categorize all family-oriented care visits. The frequency of family-oriented visits was determined using SPSS for Windows (SPSS, Inc; Chicago, Ill).¹⁰ Encounter field notes were used to develop case examples to illustrate the ways the family context affected the process of patient care.

Results

Of the 1600 outpatient encounters that were analyzed, 923 (58%) were family-oriented in some way. In a total of 560 visits (35%), patients were accompanied into the examination room by at least one other person (usually a family member), while 363 (23% of total visits) included mention or discussion of the patient’s family in some way when the patients were by themselves. A large percentage of people who accompanied patients were family members (96%), indicating that such encounters were indeed “family visits.” Individuals accompanying patients into these family visits were much more likely to be women (73%), with this high percentage largely due to the high representation of mothers present during their children’s health care visits. For adult family visits, wives (29%), adult daughters (19%), and husbands (19%) most frequently accompanied patients into the examination room.

The family frequently came up in the medical visit when clinicians took patient health histories. Clinicians asked their adult patients about family history of heart disease, cancer, and diabetes, and sometimes use of alcohol, tobacco, and other drugs. Parents accompanying their children were frequently asked about their own history of disease, or they offered such information to help clinicians make a diagnosis or determine treatment for their children. For some patients, knowledge of their own family history led to their visit in the first place. For example, one patient came in to get a breast lump examined in part because she had a strong family history of breast cancer (her mother died of breast cancer). The clinician and patient explored her risk for breast cancer, and the patient was scheduled for a mammogram.

Although taking a patient’s health history was a routine part of patient care, some of these discussions led to visible changes in health care delivery. For example, during the history-taking for a 31-year-old man presenting with a constant dry cough that had lasted for 7 weeks, the physician discovered that the patient’s father and brother had died of lung cancer and his mother of a heart attack. The physician told the patient he wanted to treat him a little more aggressively because of this family history. Although this change in treatment may not have been necessary, it reflects how family-related concerns can affect patient management.

Among adults, women were more likely to be accompanied, often by their young children. When other family members accompanied adult men into the medical encounter, they were usually wives and sometimes mothers. Visits by adults older than 75 years were characterized by several factors: The patients were more likely to be men than any other visit type, and family involvement was greater than for any other adult category. Elderly patients were most likely to be accompanied by wives and daughters, many of whom served as primary caretakers. These visits were marked by frequent questioning and an exchange of information, and often care collaboration between providers and family members.

One of the most common patterns was for mothers and fathers to give and receive information about their children’s health, illness, and social context. Clinicians relied on parents to provide information about their child’s health condition, contextual information about family health history, or other familial, work, school or other environmental factors that may influence health and disease. These family encounters provided many opportunities for educating parents and providers about the important health issues of their children. Parental involvement progressively decreased through adolescence Figure 1

Qualitative analyses of patient visits identified categories of the different ways a family perspective made a difference in these patient encounters. Although many of the visits fell into a general category of family history of illness that did not appear to affect subsequent decision, 6 nonexclusive categories were identified in which a family-oriented perspective affected patient care: (1) using family social context to illuminate patient disease, illness, and health; (2) using family to discover the source of an illness; (3) discussing and managing the health and illness of family members; (4) family concern for patient’s health; (5) using the family as a care resource and care collaborator; and (6) family member receives unscheduled care. The following sections provide case examples of different ways the presence and/or incorporation of the family in patient visits can make a difference in the processes of care.

Using Family Context to Illuminate Patient Disease, Illness, and Health

In addition to discussions about the family and health history, the family social context provided important information for understanding and improving patient care. These discussions ranged from inquiring about or discussing the home and familial relationships as sources of support or stress, talking about the effects of a recent death or divorce on patient health and well-being, or determining family dynamics as they relate to a patients diagnosis or reason for visit. Patients and accompanying family members commonly initiated these discussions, but clinicians also asked about the patient’s social context, particularly when he or she knew the family. As illustrated in the following example, this contextual information can help to illuminate the patient’s “real” reason for the visit, guiding the clinician toward more productive discussions and effective patient care:

A 34-year-old man with a history of heart problems and very high cholesterol recently had stopped taking his medications. The clinician wanted to explore this further. He knew that the patient’s daughter had been killed in a car accident, so he initiated a conversation about the family. The patient admitted that he had quit taking his medications when this tragedy hit. “I just sort of gave up,” he said. “I know that I’m depressed over this.” The clinician took this opportunity to acknowledge the patient’s depression, and they talked about beginning antidepressants. He decided that medications might hamper the grieving process but encouraged the patient to talk. The clinician let the patient know that he was available any time of day.

Using Family to Discover the Source of an Illness

Clinicians frequently asked questions about the family to determine the source of a patient’s illness, for example, “Who else is sick in the family?” Patients were asked about their exposure to other family members (eg, passive smoke) and were also reminded of how they could spread disease to their families. These conversations were particularly productive when clinicians knew the entire family and both patients and clinicians could use the visit to problem-solve ways of improving the health of both patients and their family members:

A mother brought her 5-month-old infant in with complaints of a cough and runny nose. The clinician diagnosed the patient with asthmatic bronchitis and asked the mom if she smoked. The mom indicated that both she and the baby sitter smoke but not around the baby. The clinician took this opportunity to educate mom on how smoke permeates both clothing and the air. She stressed to mom that her smoking will aggravate the baby’s condition and that the baby will likely have more and longer episodes if she continued to smoke.

In this encounter, the clinician took advantage of a teachable moment to educate the parent on how her smoking was affecting her child’s health but also engaged the mother as a patient by spending time counseling her on smoking cessation.

Discussing and Managing the Health and Illness of Other Family Members

Patients often talked about the health of family members. Sometimes this came up because clinicians were also caring for other members of the patient’s family, and they wanted to know how they were doing (eg, “Is everyone else in the family well?”). In other patient encounters this arose because the patient was the primary caretaker of another family member and wanted information or support. In the following case illustration, the clinician makes recommendations that involve the husband and treats the family rather than just the patient:

A 53-year-old woman is visiting for a health maintenance visit. The patient is a breast cancer survivor who recently had a mastectomy and is currently on chemotherapy. The patient asks the clinician if she has had any experiences with women having breast cancer whom have had husbands lose interest in sex. The patient then confides that her husband has never said anything, but she senses a difference in him; he has absolutely no desire to have any sexual contact with her. The clinician listens and is very sympathetic. She encourages the patient to seek counseling for both of them and talks with her about talking with her husband about this sensitive issue.

Family Concern for a Patient’s Health

Although patients visited their physicians for a variety of reasons, some came in to allay the fears or pestering of family members. The initiation of these visits took different forms. In most cases, patients made appointments as a result of concern expressed or pressure from a family member. In other cases, the visiting patient would ask the clinician to pressure a family member to seek care or would go ahead and schedule an appointment for him or her. These “reluctant patients” may not have otherwise come in. Although women family members most often encouraged these visits, there were also examples of concerned husbands and adult sons who prompted their family members to seek care. In the following illustration, the clinician had been prompted by the patient’s wife to talk about a particular health issue that the patient would not have otherwise brought up:

A 56-year-old man came in to review his medications for high blood pressure. The clinician asked the patient how he was doing and the patient responded, “Everything is fine.” The clinician responded, “That’s not what you wife says. She says you’re having problems with your legs.” They spent most of the visit talking about the patient’s leg problems.

Using the Family as Care Resource and Care Collaborator

In some encounters clinicians and patients discussed the roles of family members in helping patients improve health and health outcomes. Family members became care resources in a variety of ways. In some instances, the clinician asked about the possibility of involving other members of the family in care management; sometimes the patient suggested that a family member should take this role. Some family members who accompanied patients asserted themselves during visits. For example, many parents who accompanied their young children into visits asked questions and offered advice or suggestions. Clinicians also involved family members as care collaborators, asking questions and involving them in decision making about health and health care.

An adult daughter brought in her 90-year-old mother because her behavior was becoming more unpredictable and erratic. During her most recent visit to her mother’s nursing home she had witnessed her running around with no clothes. That incident, along with a recent series of falls, caused the daughter a great deal of concern. “I know when my mother acts like this, this is not really my mother,” she said. “This is not the person I know.” During most of the encounter the daughter and clinician talked about the dosage and side effects of each of her mother’s medications, gradually eliminating some of the drugs that seemed unnecessary. The clinician commented to the daughter, “We need to get some order in this.” The daughter agreed.

Family Member Receives Unscheduled Care

Sometimes the family member actually became a patient. This happened more frequently when family members accompanied the patient but also when the patient visited alone. Clinicians, patients, and family members all initiated this unplanned care. Clinicians would specifically ask about a family member who was also a patient - and would even give the patient a drug sample to take home to a spouse or child. Often, patients came in requesting medication refills or other advice about a family member’s health problem. Some family members who accompanied patients took advantage of these visits to ask about a particular health problem, to get a free sample, or to refill a medication. In the following example, the clinician saw both husband and wife, and the husband took advantage of the encounter to talk with the clinician about his own health issue and to schedule an appointment:

A husband accompanied his 79-year-old wife to her medical visit; she had shortness of breath and weakness. Near the end of the visit the husband asked about the results of his prostate-specific antigen test. Because the clinician kept family charts, the patient’s husband was able to get his results. The patient’s husband then asked the clinician if he needed to make an appointment. The clinician looked in the chart and responded, “Well, you need to have a complete physical. Make an appointment for that.”

Sometimes the accompanying person gets direct care during the visit as in the following illustration:

A mother brought her baby in for her 1-week check-up. The physician asked the mother how she was feeling. The mother mentioned that she felt a “burning sensation” after urination. After asking the mother a few specific questions, the clinician instructed mom to jump onto the table for a brief examination. The mother was diagnosed with a vaginal tear and was instructed on how to care for it.

Discussion

This paper presents a data-driven typology that shows the multiple ways in which patients’ families influence the outpatient visit in family practice. As found in previous research, family-oriented care frequently occurred through the collection or discussion of family history of illness.^4,11 Information on family history and family context frequently led to important changes in decision making and treatment for both the patient and the family. With the renewed emphasis on family history as a method of identifying patients at increased genetic risk,¹² the family history-taking skills of family physicians should become increasingly important in educating patients and their families about genetic risk.

Patients also shared important contextual information about how family relationships and family stresses were affecting their own health. They talked about the health and illness of family members that had often influenced the patient’s reason for a visit. These conversations also helped to uncover the stress-related illness, anxiety, and depression of patients and family members, sometimes representing the “real” reason for a visit. These family-oriented exchanges have been shown to positively affect the physician-patient relationship and have an impact on patient satisfaction and perceptions of quality of care.^13-15

Limitations

Despite the uniqueness of the data and the grounded analysis approach, the findings must be interpreted within the context of the study limitations. Because these data were collected by field researchers who were unaware that family context would be a focus of the analysis, it is possible that there were other patient and physician behaviors related to family issues that were not recorded. The data were sufficiently rich, however, to easily assess the effect of family knowledge on physician and patient decision making. Any unrecorded behaviors might add to but should not substantially change our conclusions. Since the patient population studied was limited to a single Midwestern state, it is possible that other populations with a different ethnic and/or racial mix might behave differently. Future research of this type should attempt to include such populations.

Conclusions

Our study demonstrates that physician knowledge of family context gained from the care of multiple family members over time improves the quality of medical decision making and may account in part for the better outcomes that have been shown to result from continuous and comprehensive care.^15-18 The current health care environment, driven by managed care does not value or encourage the long-term relationships between physicians and family members that are necessary to develop the kind of family knowledge and connectedness shown by the physicians we studied. Further research in this area should focus on outcomes in patients whose physicians are informed by family context. Such data may help convince health policymakers and legislators of the importance of continuity of physician-family relationships in the delivery of high-quality primary health care.

Acknowledgments

This study was supported by a grant from the Agency for Healthcare Research and Quality (R01 HS08776) and a Family Practice Research Center grant from the American Academy of Family Physicians. We are grateful to the physicians, staff, and patients from the 18 practices without whose participation our study would not have been possible. We also wish to thank Drs Kurt C. Stange and John G. Scott who provided helpful comments on earlier drafts of this paper. Dr Crabtree is associated with the Center for Research in Family Practice and Primary Care, Cleveland, New Brunswick, Allentown, and San Antonio.

Over the past decade health care has experienced rapid and sometimes volatile change that has affected the quality of patient care.¹ Changes in the structure and financing of health care have compromised family practice’s ability to maintain its core values of comprehensive, coordinated, and continuous care in the context of the family.² For example, the Direct Observation of Primary Care (DOPC) study found that one fourth of patients in participating northeastern Ohio practices were forced to change health care providers during a 2-year period.³ These patients reported less coordination of care, decreased continuity with their new provider, and less satisfactory interpersonal communication.

Reports from the DOPC study also found, however, that the family remains an important focus in patient care despite these disruptions in continuity and coordination. Family issues were discussed in more than 70% of patient encounters, accounting for 10% of visit time.⁴ Patients were accompanied by family members in one third of outpatient visits, providing opportunities for family involvement and even care for 18% of these "non-patients."⁵ These findings were consistent with those of a recent Canadian study that looked at the role of those people who accompany patients into the medical examination⁶ and a qualitative study of the ways the family was integrated into routine patient care.⁷

Although these and other studies confirm that the family is a salient feature in family practice, it is not known how knowledge of the family context affects the process of patient care. Data from the Prevention and Competing Demands in Primary Care Study (PCDPC) were used to answer this question through observation of visits in 18 family practices in Nebraska.

Methods

We used patient encounter data from the PCDPC practice study, an in-depth observational examination of the organizational and clinical structures and process of community-based family practices. Each of the 18 purposefully selected practices was studied using a multimethod comparative case study design that involved extensive direct observation of clinical encounters and office systems by field researchers who spent 4 weeks or more in each practice. Field researchers directly observed and dictated descriptions of approximately 30 patient encounters with each of the more than 50 clinicians and audited medical records of each of these patients. Detailed descriptive field notes documented day-to-day practice operations. Individual depth interviews with each clinician, many of the office staff, and members of the community were used to obtain different perspectives on each practice. Details of the sampling and data collection are available elsewhere in this issue.⁸

We analyzed 1600 of the 1637 outpatient encounters to determine how the family had an influence on patient care (37 visits had insufficient data for effective analysis). Each encounter was coded by one of the co-authors (S.H.) as a family-oriented visit if: (1) the patient’s family member was present in the examination room or (2) medical or health information about the patient’s family was discussed and/or addressed during the visit. Another author (D.S.M.) read each encounter using an editing organizing style⁹ by writing brief jottings about each encounter that described how the family had an impact on these medical encounters. Finally, 2 authors (S.H., D.S.M.) jointly examined these jottings to identify an organizational framework and to categorize all family-oriented care visits. The frequency of family-oriented visits was determined using SPSS for Windows (SPSS, Inc; Chicago, Ill).¹⁰ Encounter field notes were used to develop case examples to illustrate the ways the family context affected the process of patient care.

Results

Of the 1600 outpatient encounters that were analyzed, 923 (58%) were family-oriented in some way. In a total of 560 visits (35%), patients were accompanied into the examination room by at least one other person (usually a family member), while 363 (23% of total visits) included mention or discussion of the patient’s family in some way when the patients were by themselves. A large percentage of people who accompanied patients were family members (96%), indicating that such encounters were indeed “family visits.” Individuals accompanying patients into these family visits were much more likely to be women (73%), with this high percentage largely due to the high representation of mothers present during their children’s health care visits. For adult family visits, wives (29%), adult daughters (19%), and husbands (19%) most frequently accompanied patients into the examination room.

The family frequently came up in the medical visit when clinicians took patient health histories. Clinicians asked their adult patients about family history of heart disease, cancer, and diabetes, and sometimes use of alcohol, tobacco, and other drugs. Parents accompanying their children were frequently asked about their own history of disease, or they offered such information to help clinicians make a diagnosis or determine treatment for their children. For some patients, knowledge of their own family history led to their visit in the first place. For example, one patient came in to get a breast lump examined in part because she had a strong family history of breast cancer (her mother died of breast cancer). The clinician and patient explored her risk for breast cancer, and the patient was scheduled for a mammogram.

Although taking a patient’s health history was a routine part of patient care, some of these discussions led to visible changes in health care delivery. For example, during the history-taking for a 31-year-old man presenting with a constant dry cough that had lasted for 7 weeks, the physician discovered that the patient’s father and brother had died of lung cancer and his mother of a heart attack. The physician told the patient he wanted to treat him a little more aggressively because of this family history. Although this change in treatment may not have been necessary, it reflects how family-related concerns can affect patient management.

Among adults, women were more likely to be accompanied, often by their young children. When other family members accompanied adult men into the medical encounter, they were usually wives and sometimes mothers. Visits by adults older than 75 years were characterized by several factors: The patients were more likely to be men than any other visit type, and family involvement was greater than for any other adult category. Elderly patients were most likely to be accompanied by wives and daughters, many of whom served as primary caretakers. These visits were marked by frequent questioning and an exchange of information, and often care collaboration between providers and family members.

One of the most common patterns was for mothers and fathers to give and receive information about their children’s health, illness, and social context. Clinicians relied on parents to provide information about their child’s health condition, contextual information about family health history, or other familial, work, school or other environmental factors that may influence health and disease. These family encounters provided many opportunities for educating parents and providers about the important health issues of their children. Parental involvement progressively decreased through adolescence Figure 1

Qualitative analyses of patient visits identified categories of the different ways a family perspective made a difference in these patient encounters. Although many of the visits fell into a general category of family history of illness that did not appear to affect subsequent decision, 6 nonexclusive categories were identified in which a family-oriented perspective affected patient care: (1) using family social context to illuminate patient disease, illness, and health; (2) using family to discover the source of an illness; (3) discussing and managing the health and illness of family members; (4) family concern for patient’s health; (5) using the family as a care resource and care collaborator; and (6) family member receives unscheduled care. The following sections provide case examples of different ways the presence and/or incorporation of the family in patient visits can make a difference in the processes of care.

Using Family Context to Illuminate Patient Disease, Illness, and Health

In addition to discussions about the family and health history, the family social context provided important information for understanding and improving patient care. These discussions ranged from inquiring about or discussing the home and familial relationships as sources of support or stress, talking about the effects of a recent death or divorce on patient health and well-being, or determining family dynamics as they relate to a patients diagnosis or reason for visit. Patients and accompanying family members commonly initiated these discussions, but clinicians also asked about the patient’s social context, particularly when he or she knew the family. As illustrated in the following example, this contextual information can help to illuminate the patient’s “real” reason for the visit, guiding the clinician toward more productive discussions and effective patient care:

A 34-year-old man with a history of heart problems and very high cholesterol recently had stopped taking his medications. The clinician wanted to explore this further. He knew that the patient’s daughter had been killed in a car accident, so he initiated a conversation about the family. The patient admitted that he had quit taking his medications when this tragedy hit. “I just sort of gave up,” he said. “I know that I’m depressed over this.” The clinician took this opportunity to acknowledge the patient’s depression, and they talked about beginning antidepressants. He decided that medications might hamper the grieving process but encouraged the patient to talk. The clinician let the patient know that he was available any time of day.

Using Family to Discover the Source of an Illness

Clinicians frequently asked questions about the family to determine the source of a patient’s illness, for example, “Who else is sick in the family?” Patients were asked about their exposure to other family members (eg, passive smoke) and were also reminded of how they could spread disease to their families. These conversations were particularly productive when clinicians knew the entire family and both patients and clinicians could use the visit to problem-solve ways of improving the health of both patients and their family members:

A mother brought her 5-month-old infant in with complaints of a cough and runny nose. The clinician diagnosed the patient with asthmatic bronchitis and asked the mom if she smoked. The mom indicated that both she and the baby sitter smoke but not around the baby. The clinician took this opportunity to educate mom on how smoke permeates both clothing and the air. She stressed to mom that her smoking will aggravate the baby’s condition and that the baby will likely have more and longer episodes if she continued to smoke.

In this encounter, the clinician took advantage of a teachable moment to educate the parent on how her smoking was affecting her child’s health but also engaged the mother as a patient by spending time counseling her on smoking cessation.

Discussing and Managing the Health and Illness of Other Family Members

Patients often talked about the health of family members. Sometimes this came up because clinicians were also caring for other members of the patient’s family, and they wanted to know how they were doing (eg, “Is everyone else in the family well?”). In other patient encounters this arose because the patient was the primary caretaker of another family member and wanted information or support. In the following case illustration, the clinician makes recommendations that involve the husband and treats the family rather than just the patient:

A 53-year-old woman is visiting for a health maintenance visit. The patient is a breast cancer survivor who recently had a mastectomy and is currently on chemotherapy. The patient asks the clinician if she has had any experiences with women having breast cancer whom have had husbands lose interest in sex. The patient then confides that her husband has never said anything, but she senses a difference in him; he has absolutely no desire to have any sexual contact with her. The clinician listens and is very sympathetic. She encourages the patient to seek counseling for both of them and talks with her about talking with her husband about this sensitive issue.

Family Concern for a Patient’s Health

Although patients visited their physicians for a variety of reasons, some came in to allay the fears or pestering of family members. The initiation of these visits took different forms. In most cases, patients made appointments as a result of concern expressed or pressure from a family member. In other cases, the visiting patient would ask the clinician to pressure a family member to seek care or would go ahead and schedule an appointment for him or her. These “reluctant patients” may not have otherwise come in. Although women family members most often encouraged these visits, there were also examples of concerned husbands and adult sons who prompted their family members to seek care. In the following illustration, the clinician had been prompted by the patient’s wife to talk about a particular health issue that the patient would not have otherwise brought up:

A 56-year-old man came in to review his medications for high blood pressure. The clinician asked the patient how he was doing and the patient responded, “Everything is fine.” The clinician responded, “That’s not what you wife says. She says you’re having problems with your legs.” They spent most of the visit talking about the patient’s leg problems.

Using the Family as Care Resource and Care Collaborator

In some encounters clinicians and patients discussed the roles of family members in helping patients improve health and health outcomes. Family members became care resources in a variety of ways. In some instances, the clinician asked about the possibility of involving other members of the family in care management; sometimes the patient suggested that a family member should take this role. Some family members who accompanied patients asserted themselves during visits. For example, many parents who accompanied their young children into visits asked questions and offered advice or suggestions. Clinicians also involved family members as care collaborators, asking questions and involving them in decision making about health and health care.

An adult daughter brought in her 90-year-old mother because her behavior was becoming more unpredictable and erratic. During her most recent visit to her mother’s nursing home she had witnessed her running around with no clothes. That incident, along with a recent series of falls, caused the daughter a great deal of concern. “I know when my mother acts like this, this is not really my mother,” she said. “This is not the person I know.” During most of the encounter the daughter and clinician talked about the dosage and side effects of each of her mother’s medications, gradually eliminating some of the drugs that seemed unnecessary. The clinician commented to the daughter, “We need to get some order in this.” The daughter agreed.

Family Member Receives Unscheduled Care

Sometimes the family member actually became a patient. This happened more frequently when family members accompanied the patient but also when the patient visited alone. Clinicians, patients, and family members all initiated this unplanned care. Clinicians would specifically ask about a family member who was also a patient - and would even give the patient a drug sample to take home to a spouse or child. Often, patients came in requesting medication refills or other advice about a family member’s health problem. Some family members who accompanied patients took advantage of these visits to ask about a particular health problem, to get a free sample, or to refill a medication. In the following example, the clinician saw both husband and wife, and the husband took advantage of the encounter to talk with the clinician about his own health issue and to schedule an appointment:

A husband accompanied his 79-year-old wife to her medical visit; she had shortness of breath and weakness. Near the end of the visit the husband asked about the results of his prostate-specific antigen test. Because the clinician kept family charts, the patient’s husband was able to get his results. The patient’s husband then asked the clinician if he needed to make an appointment. The clinician looked in the chart and responded, “Well, you need to have a complete physical. Make an appointment for that.”

Sometimes the accompanying person gets direct care during the visit as in the following illustration:

A mother brought her baby in for her 1-week check-up. The physician asked the mother how she was feeling. The mother mentioned that she felt a “burning sensation” after urination. After asking the mother a few specific questions, the clinician instructed mom to jump onto the table for a brief examination. The mother was diagnosed with a vaginal tear and was instructed on how to care for it.

Discussion

This paper presents a data-driven typology that shows the multiple ways in which patients’ families influence the outpatient visit in family practice. As found in previous research, family-oriented care frequently occurred through the collection or discussion of family history of illness.^4,11 Information on family history and family context frequently led to important changes in decision making and treatment for both the patient and the family. With the renewed emphasis on family history as a method of identifying patients at increased genetic risk,¹² the family history-taking skills of family physicians should become increasingly important in educating patients and their families about genetic risk.

Patients also shared important contextual information about how family relationships and family stresses were affecting their own health. They talked about the health and illness of family members that had often influenced the patient’s reason for a visit. These conversations also helped to uncover the stress-related illness, anxiety, and depression of patients and family members, sometimes representing the “real” reason for a visit. These family-oriented exchanges have been shown to positively affect the physician-patient relationship and have an impact on patient satisfaction and perceptions of quality of care.^13-15

Limitations

Despite the uniqueness of the data and the grounded analysis approach, the findings must be interpreted within the context of the study limitations. Because these data were collected by field researchers who were unaware that family context would be a focus of the analysis, it is possible that there were other patient and physician behaviors related to family issues that were not recorded. The data were sufficiently rich, however, to easily assess the effect of family knowledge on physician and patient decision making. Any unrecorded behaviors might add to but should not substantially change our conclusions. Since the patient population studied was limited to a single Midwestern state, it is possible that other populations with a different ethnic and/or racial mix might behave differently. Future research of this type should attempt to include such populations.

Conclusions

Our study demonstrates that physician knowledge of family context gained from the care of multiple family members over time improves the quality of medical decision making and may account in part for the better outcomes that have been shown to result from continuous and comprehensive care.^15-18 The current health care environment, driven by managed care does not value or encourage the long-term relationships between physicians and family members that are necessary to develop the kind of family knowledge and connectedness shown by the physicians we studied. Further research in this area should focus on outcomes in patients whose physicians are informed by family context. Such data may help convince health policymakers and legislators of the importance of continuity of physician-family relationships in the delivery of high-quality primary health care.

Acknowledgments

This study was supported by a grant from the Agency for Healthcare Research and Quality (R01 HS08776) and a Family Practice Research Center grant from the American Academy of Family Physicians. We are grateful to the physicians, staff, and patients from the 18 practices without whose participation our study would not have been possible. We also wish to thank Drs Kurt C. Stange and John G. Scott who provided helpful comments on earlier drafts of this paper. Dr Crabtree is associated with the Center for Research in Family Practice and Primary Care, Cleveland, New Brunswick, Allentown, and San Antonio.

Over the past decade health care has experienced rapid and sometimes volatile change that has affected the quality of patient care.¹ Changes in the structure and financing of health care have compromised family practice’s ability to maintain its core values of comprehensive, coordinated, and continuous care in the context of the family.² For example, the Direct Observation of Primary Care (DOPC) study found that one fourth of patients in participating northeastern Ohio practices were forced to change health care providers during a 2-year period.³ These patients reported less coordination of care, decreased continuity with their new provider, and less satisfactory interpersonal communication.

Reports from the DOPC study also found, however, that the family remains an important focus in patient care despite these disruptions in continuity and coordination. Family issues were discussed in more than 70% of patient encounters, accounting for 10% of visit time.⁴ Patients were accompanied by family members in one third of outpatient visits, providing opportunities for family involvement and even care for 18% of these "non-patients."⁵ These findings were consistent with those of a recent Canadian study that looked at the role of those people who accompany patients into the medical examination⁶ and a qualitative study of the ways the family was integrated into routine patient care.⁷

Although these and other studies confirm that the family is a salient feature in family practice, it is not known how knowledge of the family context affects the process of patient care. Data from the Prevention and Competing Demands in Primary Care Study (PCDPC) were used to answer this question through observation of visits in 18 family practices in Nebraska.

Methods

We used patient encounter data from the PCDPC practice study, an in-depth observational examination of the organizational and clinical structures and process of community-based family practices. Each of the 18 purposefully selected practices was studied using a multimethod comparative case study design that involved extensive direct observation of clinical encounters and office systems by field researchers who spent 4 weeks or more in each practice. Field researchers directly observed and dictated descriptions of approximately 30 patient encounters with each of the more than 50 clinicians and audited medical records of each of these patients. Detailed descriptive field notes documented day-to-day practice operations. Individual depth interviews with each clinician, many of the office staff, and members of the community were used to obtain different perspectives on each practice. Details of the sampling and data collection are available elsewhere in this issue.⁸

We analyzed 1600 of the 1637 outpatient encounters to determine how the family had an influence on patient care (37 visits had insufficient data for effective analysis). Each encounter was coded by one of the co-authors (S.H.) as a family-oriented visit if: (1) the patient’s family member was present in the examination room or (2) medical or health information about the patient’s family was discussed and/or addressed during the visit. Another author (D.S.M.) read each encounter using an editing organizing style⁹ by writing brief jottings about each encounter that described how the family had an impact on these medical encounters. Finally, 2 authors (S.H., D.S.M.) jointly examined these jottings to identify an organizational framework and to categorize all family-oriented care visits. The frequency of family-oriented visits was determined using SPSS for Windows (SPSS, Inc; Chicago, Ill).¹⁰ Encounter field notes were used to develop case examples to illustrate the ways the family context affected the process of patient care.

Results

Of the 1600 outpatient encounters that were analyzed, 923 (58%) were family-oriented in some way. In a total of 560 visits (35%), patients were accompanied into the examination room by at least one other person (usually a family member), while 363 (23% of total visits) included mention or discussion of the patient’s family in some way when the patients were by themselves. A large percentage of people who accompanied patients were family members (96%), indicating that such encounters were indeed “family visits.” Individuals accompanying patients into these family visits were much more likely to be women (73%), with this high percentage largely due to the high representation of mothers present during their children’s health care visits. For adult family visits, wives (29%), adult daughters (19%), and husbands (19%) most frequently accompanied patients into the examination room.

The family frequently came up in the medical visit when clinicians took patient health histories. Clinicians asked their adult patients about family history of heart disease, cancer, and diabetes, and sometimes use of alcohol, tobacco, and other drugs. Parents accompanying their children were frequently asked about their own history of disease, or they offered such information to help clinicians make a diagnosis or determine treatment for their children. For some patients, knowledge of their own family history led to their visit in the first place. For example, one patient came in to get a breast lump examined in part because she had a strong family history of breast cancer (her mother died of breast cancer). The clinician and patient explored her risk for breast cancer, and the patient was scheduled for a mammogram.

Although taking a patient’s health history was a routine part of patient care, some of these discussions led to visible changes in health care delivery. For example, during the history-taking for a 31-year-old man presenting with a constant dry cough that had lasted for 7 weeks, the physician discovered that the patient’s father and brother had died of lung cancer and his mother of a heart attack. The physician told the patient he wanted to treat him a little more aggressively because of this family history. Although this change in treatment may not have been necessary, it reflects how family-related concerns can affect patient management.

Among adults, women were more likely to be accompanied, often by their young children. When other family members accompanied adult men into the medical encounter, they were usually wives and sometimes mothers. Visits by adults older than 75 years were characterized by several factors: The patients were more likely to be men than any other visit type, and family involvement was greater than for any other adult category. Elderly patients were most likely to be accompanied by wives and daughters, many of whom served as primary caretakers. These visits were marked by frequent questioning and an exchange of information, and often care collaboration between providers and family members.

One of the most common patterns was for mothers and fathers to give and receive information about their children’s health, illness, and social context. Clinicians relied on parents to provide information about their child’s health condition, contextual information about family health history, or other familial, work, school or other environmental factors that may influence health and disease. These family encounters provided many opportunities for educating parents and providers about the important health issues of their children. Parental involvement progressively decreased through adolescence Figure 1

Qualitative analyses of patient visits identified categories of the different ways a family perspective made a difference in these patient encounters. Although many of the visits fell into a general category of family history of illness that did not appear to affect subsequent decision, 6 nonexclusive categories were identified in which a family-oriented perspective affected patient care: (1) using family social context to illuminate patient disease, illness, and health; (2) using family to discover the source of an illness; (3) discussing and managing the health and illness of family members; (4) family concern for patient’s health; (5) using the family as a care resource and care collaborator; and (6) family member receives unscheduled care. The following sections provide case examples of different ways the presence and/or incorporation of the family in patient visits can make a difference in the processes of care.

Using Family Context to Illuminate Patient Disease, Illness, and Health

In addition to discussions about the family and health history, the family social context provided important information for understanding and improving patient care. These discussions ranged from inquiring about or discussing the home and familial relationships as sources of support or stress, talking about the effects of a recent death or divorce on patient health and well-being, or determining family dynamics as they relate to a patients diagnosis or reason for visit. Patients and accompanying family members commonly initiated these discussions, but clinicians also asked about the patient’s social context, particularly when he or she knew the family. As illustrated in the following example, this contextual information can help to illuminate the patient’s “real” reason for the visit, guiding the clinician toward more productive discussions and effective patient care:

A 34-year-old man with a history of heart problems and very high cholesterol recently had stopped taking his medications. The clinician wanted to explore this further. He knew that the patient’s daughter had been killed in a car accident, so he initiated a conversation about the family. The patient admitted that he had quit taking his medications when this tragedy hit. “I just sort of gave up,” he said. “I know that I’m depressed over this.” The clinician took this opportunity to acknowledge the patient’s depression, and they talked about beginning antidepressants. He decided that medications might hamper the grieving process but encouraged the patient to talk. The clinician let the patient know that he was available any time of day.

Using Family to Discover the Source of an Illness

Clinicians frequently asked questions about the family to determine the source of a patient’s illness, for example, “Who else is sick in the family?” Patients were asked about their exposure to other family members (eg, passive smoke) and were also reminded of how they could spread disease to their families. These conversations were particularly productive when clinicians knew the entire family and both patients and clinicians could use the visit to problem-solve ways of improving the health of both patients and their family members:

A mother brought her 5-month-old infant in with complaints of a cough and runny nose. The clinician diagnosed the patient with asthmatic bronchitis and asked the mom if she smoked. The mom indicated that both she and the baby sitter smoke but not around the baby. The clinician took this opportunity to educate mom on how smoke permeates both clothing and the air. She stressed to mom that her smoking will aggravate the baby’s condition and that the baby will likely have more and longer episodes if she continued to smoke.

In this encounter, the clinician took advantage of a teachable moment to educate the parent on how her smoking was affecting her child’s health but also engaged the mother as a patient by spending time counseling her on smoking cessation.

Discussing and Managing the Health and Illness of Other Family Members

Patients often talked about the health of family members. Sometimes this came up because clinicians were also caring for other members of the patient’s family, and they wanted to know how they were doing (eg, “Is everyone else in the family well?”). In other patient encounters this arose because the patient was the primary caretaker of another family member and wanted information or support. In the following case illustration, the clinician makes recommendations that involve the husband and treats the family rather than just the patient:

A 53-year-old woman is visiting for a health maintenance visit. The patient is a breast cancer survivor who recently had a mastectomy and is currently on chemotherapy. The patient asks the clinician if she has had any experiences with women having breast cancer whom have had husbands lose interest in sex. The patient then confides that her husband has never said anything, but she senses a difference in him; he has absolutely no desire to have any sexual contact with her. The clinician listens and is very sympathetic. She encourages the patient to seek counseling for both of them and talks with her about talking with her husband about this sensitive issue.

Family Concern for a Patient’s Health

Although patients visited their physicians for a variety of reasons, some came in to allay the fears or pestering of family members. The initiation of these visits took different forms. In most cases, patients made appointments as a result of concern expressed or pressure from a family member. In other cases, the visiting patient would ask the clinician to pressure a family member to seek care or would go ahead and schedule an appointment for him or her. These “reluctant patients” may not have otherwise come in. Although women family members most often encouraged these visits, there were also examples of concerned husbands and adult sons who prompted their family members to seek care. In the following illustration, the clinician had been prompted by the patient’s wife to talk about a particular health issue that the patient would not have otherwise brought up:

A 56-year-old man came in to review his medications for high blood pressure. The clinician asked the patient how he was doing and the patient responded, “Everything is fine.” The clinician responded, “That’s not what you wife says. She says you’re having problems with your legs.” They spent most of the visit talking about the patient’s leg problems.

Using the Family as Care Resource and Care Collaborator

In some encounters clinicians and patients discussed the roles of family members in helping patients improve health and health outcomes. Family members became care resources in a variety of ways. In some instances, the clinician asked about the possibility of involving other members of the family in care management; sometimes the patient suggested that a family member should take this role. Some family members who accompanied patients asserted themselves during visits. For example, many parents who accompanied their young children into visits asked questions and offered advice or suggestions. Clinicians also involved family members as care collaborators, asking questions and involving them in decision making about health and health care.

An adult daughter brought in her 90-year-old mother because her behavior was becoming more unpredictable and erratic. During her most recent visit to her mother’s nursing home she had witnessed her running around with no clothes. That incident, along with a recent series of falls, caused the daughter a great deal of concern. “I know when my mother acts like this, this is not really my mother,” she said. “This is not the person I know.” During most of the encounter the daughter and clinician talked about the dosage and side effects of each of her mother’s medications, gradually eliminating some of the drugs that seemed unnecessary. The clinician commented to the daughter, “We need to get some order in this.” The daughter agreed.

Family Member Receives Unscheduled Care

Sometimes the family member actually became a patient. This happened more frequently when family members accompanied the patient but also when the patient visited alone. Clinicians, patients, and family members all initiated this unplanned care. Clinicians would specifically ask about a family member who was also a patient - and would even give the patient a drug sample to take home to a spouse or child. Often, patients came in requesting medication refills or other advice about a family member’s health problem. Some family members who accompanied patients took advantage of these visits to ask about a particular health problem, to get a free sample, or to refill a medication. In the following example, the clinician saw both husband and wife, and the husband took advantage of the encounter to talk with the clinician about his own health issue and to schedule an appointment:

A husband accompanied his 79-year-old wife to her medical visit; she had shortness of breath and weakness. Near the end of the visit the husband asked about the results of his prostate-specific antigen test. Because the clinician kept family charts, the patient’s husband was able to get his results. The patient’s husband then asked the clinician if he needed to make an appointment. The clinician looked in the chart and responded, “Well, you need to have a complete physical. Make an appointment for that.”

Sometimes the accompanying person gets direct care during the visit as in the following illustration:

A mother brought her baby in for her 1-week check-up. The physician asked the mother how she was feeling. The mother mentioned that she felt a “burning sensation” after urination. After asking the mother a few specific questions, the clinician instructed mom to jump onto the table for a brief examination. The mother was diagnosed with a vaginal tear and was instructed on how to care for it.

Discussion

This paper presents a data-driven typology that shows the multiple ways in which patients’ families influence the outpatient visit in family practice. As found in previous research, family-oriented care frequently occurred through the collection or discussion of family history of illness.^4,11 Information on family history and family context frequently led to important changes in decision making and treatment for both the patient and the family. With the renewed emphasis on family history as a method of identifying patients at increased genetic risk,¹² the family history-taking skills of family physicians should become increasingly important in educating patients and their families about genetic risk.

Patients also shared important contextual information about how family relationships and family stresses were affecting their own health. They talked about the health and illness of family members that had often influenced the patient’s reason for a visit. These conversations also helped to uncover the stress-related illness, anxiety, and depression of patients and family members, sometimes representing the “real” reason for a visit. These family-oriented exchanges have been shown to positively affect the physician-patient relationship and have an impact on patient satisfaction and perceptions of quality of care.^13-15

Limitations

Despite the uniqueness of the data and the grounded analysis approach, the findings must be interpreted within the context of the study limitations. Because these data were collected by field researchers who were unaware that family context would be a focus of the analysis, it is possible that there were other patient and physician behaviors related to family issues that were not recorded. The data were sufficiently rich, however, to easily assess the effect of family knowledge on physician and patient decision making. Any unrecorded behaviors might add to but should not substantially change our conclusions. Since the patient population studied was limited to a single Midwestern state, it is possible that other populations with a different ethnic and/or racial mix might behave differently. Future research of this type should attempt to include such populations.

Conclusions

Our study demonstrates that physician knowledge of family context gained from the care of multiple family members over time improves the quality of medical decision making and may account in part for the better outcomes that have been shown to result from continuous and comprehensive care.^15-18 The current health care environment, driven by managed care does not value or encourage the long-term relationships between physicians and family members that are necessary to develop the kind of family knowledge and connectedness shown by the physicians we studied. Further research in this area should focus on outcomes in patients whose physicians are informed by family context. Such data may help convince health policymakers and legislators of the importance of continuity of physician-family relationships in the delivery of high-quality primary health care.

Acknowledgments

This study was supported by a grant from the Agency for Healthcare Research and Quality (R01 HS08776) and a Family Practice Research Center grant from the American Academy of Family Physicians. We are grateful to the physicians, staff, and patients from the 18 practices without whose participation our study would not have been possible. We also wish to thank Drs Kurt C. Stange and John G. Scott who provided helpful comments on earlier drafts of this paper. Dr Crabtree is associated with the Center for Research in Family Practice and Primary Care, Cleveland, New Brunswick, Allentown, and San Antonio.

Sexual dysfunction in men is common. A large study¹ in the United States with a sample of 1410 men found a 31% 12-month prevalence of sexual dysfunction. In a small US survey² of 62 men, nearly all participants reported sexual concerns at any time during their life. One third of 789 men in a British general practice reported a current sexual problem.³ Three fourths of 78 patients surveyed in another British general practice suffered from general sexual problems; 35% reported a specific sexual dysfunction, such as premature ejaculation or erectile dysfunction.⁴

Sexual dysfunction may have organic or psychological causes. The family practice setting seems to be ideal for the evaluation and management of sexual dysfunction, because family physicians usually know the personal and family situation of their patients.^5-8Also, sexual dysfunction is sometimes related to common diseases, such as diabetes mellitus.^9,10 Medications frequently prescribed in family practice, such as antihypertensive and psychiatric drugs, may also adversely affect sexual performance.^11,12 Finally, the popularity of sildenafil (Viagra) and the accompanying publicity may encourage an increasing number of patients with erectile problems to consult their family physicians in case of sexual problems.¹³

It is important for the family physician to recognize the spectrum of sexual problems among men, since these concerns may affect their patients’ health, wellbeing, and relationships. There are only a few small studies on sexual dysfunction in family practice,^2-4,14 with limited generalizability. Most of the authors did not correlate physicians’ and patients’ views on this topic. Little is known about what men expect from their family physicians in cases of sexual disorders. We performed a survey of male patients and their physicians, focusing on the frequency and types of sexual problems in family practice, the men’s expectations of their physicians in case of sexual problems, the physicians’ perception and management of sexual problems, and the influence of physicians’ knowledge of and attitude toward sexual medicine on patients’ help-seeking behavior.

Methods

Study Population

All board-certified family physicians in the 2 districts of Hildesheim (Lower Saxony) and Heiligenstadt (Thuringia, the former East Germany) in the North of Germany were asked to take part in the study and complete a self-administered questionnaire. For one day all male patients in the participating practices 18 years or older and capable of reading the German language were asked to complete a questionnaire in a separate room of the practice and, after finishing, to put it into a box. Neither the physician nor the practice nurses had access to the box.

Questionnaires

We developed a 3-page questionnaire for family physicians and a 4-page patient questionnaire. A multiple-choice format was chosen with some room for personal comments. The patient questionnaire was pilot tested on a sample of 20 men attending a family practice; these men did not take part in the final study. The patient questionnaire had 3 parts: the quality of the patient’s relationship with the family physician with regard to sexual problems; the frequency, presumed causes, and effects of sexual dysfunction; and the patient’s view of the physician’s management of sexual problems.

In the second part of the questionnaire a list of potential sexual problems was presented. The patients were asked to indicate whether they suffered never, seldom, sometimes, often, or always from 1 or more of these problems.

The physician questionnaire, which was piloted in cooperation with 3 family physicians, focused on the frequency of sexual problems in male patients, the treatment of sexual problems in family practice, and the reasons for not talking about sexual problems with the patient.

Data were analyzed using SAS software, version 8 for Windows (SAS Institute; Cary, NC). To test for differences, we used the chi-square test, as appropriate. Correlations among ordinal variables were determined by the Spearman rank correlation.¹⁵ The units of analysis were the patient and the physician, respectively. For comparison of the patients’ and their physicians’ statements, the unit of analysis was the practice. Since not all patients answered all items of the questionnaire, we report the denominator where relevant.

Results

Physicians and Patients

Of the 46 family practices approached, 20 (43%), all single handed, gave approval to interview patients in their surgery. Seventeen of these practices were located in a major city, and 3 were in a small town. Nearly all (19/20) answered the physician questionnaire. The participating physicians were, on average, aged 49 years (range=36-61 years). When asked why they did not want to take part in the study, physicians most often said they had only a small number of patients with sexual problems in their practice (6/26) or their patients would not consent to a study dealing with sexuality (8/26). Only 21% of the women physicians addressed took part in the study, compared with 53% of the men.

A total of 307 men (84%) returned the patient questionnaire. Lack of time, feeling too sick, and unwillingness were the reasons most often given by the 57 patients who refused to participate. Demographic characteristics are shown in Table 1. The age distribution is compared with the representative ADT (Accounting Data Record) Panel of the Central Institute of Ambulatory Health Care in Germany (Zentralinstitut für die kassenärztliche Versorgung).^16,17

More than 90% of the participating men (279/307) had been consulting their particular physician for more than 1 year. According to their own assessment, 245 patients had informed their physicians about their personal, family, and social background.

Frequency of Sexual Problems

More than 90% of the male patients (251/270) reported having 1 sexual problem at least seldom. Figure 1 illustrates the frequency of sexual problems. These results were not influenced by the patients’ marital status.

Low sexual desire (73%) and premature ejaculation (66%) were dominant problems. More than one third of the respondents suffered at least sometimes from these problems Figure 2. Approximately 5% to 6% of the men reported fear of failure and erectile dysfunction as often or permanent problems. Thirty-seven men (12%) did not answer any of the questionnaire items that referred to sexual problems. These men did not differ in marital status from those who answered these items; however, on average the 37 men were older than the remainder (55 years vs 44 years). Even if we make the (rather unlikely) assumption that all patients who did not answer the sex-related items represent men without any sexual problems, more than 80% of the sample (251/307) suffered at least occasionally from a reported sexual problem. Low sexual desire and erectile problems were weakly or moderately associated with age (rho=0.18; P=.0049 and rho=0.38; P=.0001, respectively).

Half of the patients who had sexual concerns (107/201) considered occupational stress to be a cause of their problems. Pressure resulting from expectations of self (28%), difficulties in relationships (19%), and comorbid diseases (13%) were also considered common causes of their sexual dysfunction. Less common causes included medicines (9%), inhibitions resulting from strict sexual education (8%), pressure resulting from expectations of others (8%), and involuntary childlessness (3%).

As a consequence of their sexual difficulties, 46 of 257 (18%) experienced depression and depression-like symptoms, 36 (14%) sleeplessness, and 35 (14%) difficulties in their relationships. Patients’ satisfaction with their sexual lives was moderately correlated with the frequency of their sexual problems (rho = -0.33; P <.001). Patients who reported suffering sexual problems never or seldom reported satisfaction scores of 7.5 or 7.6, respectively. Patients reporting sexual problems often or always were less satisfied Figure 3. Sexual-life satisfaction and age were not correlated (rho = -0.05; P=.41)

Patients’ Expectations

Most patients considered it important to talk with their family physicians about their sexual concerns (84%). Almost half of the respondents (133/295) preferred that their physician initiate any discussions about sexuality. These expectations were not influenced by patient age (data not shown). More than two thirds of the respondents would have liked their physicians to signal his or her open-mindedness by directly addressing sexual topics during the consultation.

Management of Sexual Problems

Only 12% (34/288) of the respondents had already consulted their family physicians because of sexual problems, and most (23/34) of these patients were satisfied with the physician’s treatment (discussion about the difficulties in 56% of cases, prescription of drugs in 29%, and specialist referral in 24%).

The physicians remembered only a few consultations about sexual problems. On average, they recalled seeing 7 patients every 3 months primarily because of sexual difficulties. In an additional 6 patient contacts, sexual problems were raised during the consultation.

The physicians stated that they initiated discussion about sexual problems only sometimes (53%) or seldom (37%) during the consultation. They considered the following occasions or conditions as good opportunities to start a discussion: psychosomatic complaints (16/19; 84%), family planning (53%), questions concerning human immunodeficiency virus (47%), and diseases, such as diabetes mellitus (79%) or hypertension (63%).

In cases of sexual problems, most family physicians in our sample would have changed medication if relevant (79%) or would have referred patients to urologists (74%). According to the experience of approximately half of the physicians (9/19), sildenafil motivated more patients with sexual concerns to consult their physicians.

The physicians were asked for possible reasons for not talking about sexuality Table 2. Regarding physician-related factors, most physicians (53%) considered the lack of time a barrier to addressing sexual topics. Nearly all physicians (17/19) presumed that a feeling of shame keeps patients from talking about sexuality.

Importance of Sexual Medicine

On average, physicians rated the importance of sexual knowledge in family practice at a value of 7.2 on a 10-point Likert scale where 1 = not important at all and 10 = highly important. They rated the quality of their own knowledge 6.05.

There was a (nonsignificant) correlation between the physicians’ assumed knowledge and the patients’ wish to contact them in the case of sexual problems (rho=0.26; P=.29). Correspondingly, men were more likely to contact their physician for sexual problems if the physician assessed counseling in sexual medicine to be more important (rho=0.18; P=.47). The practice was the unit of analysis for calculating these correlations and their P values.

Discussion

Our study shows that sexual problems are widespread among male family practice patients and confirms the public health importance of this area. The high rate of men reporting sexual problems (93%) includes all patients who have or have had a sexual problem, independently of how often they suffered from it. Even if we consider the missing data (n=37) as representing patients without any sexual problem, the rate is still more than 80% and exceeds the incidence found by other authors.^3,4,14,18 Our results confirm the survey of Metz and Seifert,² in which nearly every patient had a sexual problem (97%) at some time during his life. Interestingly, this situation does not seem to be significantly different from the experience of women patients, according to a recently published survey of sexual problems among women seeking routine gynecologic care from the departments of family practice and obstetrics and gynecology.¹⁹ Nearly all the 964 women (99%) reported 1 sexual concern or more.

Health surveys reflect self-reported problems, not medically defined diseases. However, because sexual problems affect well-being and interact with other somatic and psychological complaints, family physicians should be aware that nearly all patients experience and report sexual concerns. Also, it should be recognized that 20% of patients suffer often or always from at least 1 sexual problem. These patients were significantly less satisfied with their sexual life than the remainder.

Similar to the results of other studies,^1,9,20,21 we found that erectile dysfunction and low sexual desire increase with age. However, this was only a moderate correlation and should alert physicians to be cautious about regarding erectile dysfunction as an elderly man’s complaint. According to the patients’ own assessments, many sexual problems are related to occupational stress. Therefore, the family physician, with his or her knowledge of the patient’s family and social situation, is well prepared to identify men with sexual problems. For some men of our sample, these problems may be associated with sleeplessness or depression. In these cases, the family physician should address the interaction of diseases or drugs with sexual disorders.

The participating physicians were hesitant about initiating a discussion about sexual problems. This corresponds with the results of the patient survey, as only 12% of the men had already consulted their physicians in cases of sexual concerns. Half of the physicians stated that they have had more consultations about sexual concerns since the discussion about sildenafil started. This is in line with the hypothesis of Tiefer¹³ that the availability of sildenafil could motivate an increasing number of patients suffering from sexual problems to consult their family physicians.

Family physicians considered lack of time and knowledge as the most important reasons for not taking a sexual history. In a Dutch study,²² these obstacles were also put forward by the 59% of general practitioners who did not take a sexual history even when an erectile problem was suspected. Family physicians in our study also mentioned a feeling of shame as a barrier for patients and themselves. These arguments also evolved in a survey about the management of involuntary childlessness in German family practice.²³ The interviewed physicians in that study were afraid of intrusion and inadequacy when addressing involuntary childlessness and related areas, such as sexuality, which are of an intimate and private nature. To overcome problems such as lack of knowledge or feelings of shame and inadequacy, it may be useful to give the topic of sexuality and sexual problems a higher status in basic medical education, as well as in continuing medical education. Also, key questions on how to start and maintain discussions on sexuality should be developed, evaluated, and then rehearsed.

Our results seem to indicate a correlation between patients’ consulting behavior and physicians’ attitudes and knowledge. More patients talk about sexual concerns to their physicians when the physicians assume that their own knowledge and attitude toward sexual medicine is better than that of their colleagues. This could represent a self-selection process as to patients’ preference. Because these correlations were analyzed on the practice level (as the unit of analysis), the results may not have reached statistical significance because of the low number of observations. Further studies may determine these associations more precisely.

We do not suggest that all patients suffering from sexual problems require treatment. Many of the patients in our study who reported seldom experiencing a sexual problem were just as satisfied with their sexual life as those who reported never having a problem. To date, some family physicians seem to resolve the difficulty of finding out only those sexual problems that need to be treated by neglecting the topic of sexuality during consultation. The physicians’ assumption that this attitude corresponds to the patients’ feeling of shame is not supported by our data: More than half of the patients would appreciate their physicians addressing the topic of sexuality in an active and open way, if appropriate.

Limitations

Compared with patients of the ADT Panel,^16,17 older patients, especially those older than 70 years, are under-represented in our study. This may be because this study was performed exclusively in surgeries, while the ADT Panel also included patients who were visited at home and in nursing homes. However, it should be emphasized that the response rate of 84% was excellent, so that a selection bias with regard to patients visiting the surgery is rather unlikely.

A response rate of 43% of the physicians may indicate a selection bias. Physicians who were more open-minded toward sexual problems may have been more likely to participate in our survey. Thus, sexual problems are probably an even less frequent topic during consultation than our data suggest.

Conclusions

The high level of self-reported sexual dysfunction in men and family physicians’ hesitancy to deal with this issue signals a neglected area in primary health care. Possible starting points for communication about sexuality in family practice include the prescription of drugs that trigger sexual disorders, diagnosis of conditions that are associated with sexual problems, and the patients’ personal, social, and occupational background.

Acknowledgments

Parts of our study were supported by Grant 01 KY 9605/7 from the Federal Ministry of Education and Research, Bonn, Germany. The authors would like to thank the practice nurses, physicians, and patients who participated in this survey and answered questions of a very personal nature.

Sexual dysfunction in men is common. A large study¹ in the United States with a sample of 1410 men found a 31% 12-month prevalence of sexual dysfunction. In a small US survey² of 62 men, nearly all participants reported sexual concerns at any time during their life. One third of 789 men in a British general practice reported a current sexual problem.³ Three fourths of 78 patients surveyed in another British general practice suffered from general sexual problems; 35% reported a specific sexual dysfunction, such as premature ejaculation or erectile dysfunction.⁴

Sexual dysfunction may have organic or psychological causes. The family practice setting seems to be ideal for the evaluation and management of sexual dysfunction, because family physicians usually know the personal and family situation of their patients.^5-8Also, sexual dysfunction is sometimes related to common diseases, such as diabetes mellitus.^9,10 Medications frequently prescribed in family practice, such as antihypertensive and psychiatric drugs, may also adversely affect sexual performance.^11,12 Finally, the popularity of sildenafil (Viagra) and the accompanying publicity may encourage an increasing number of patients with erectile problems to consult their family physicians in case of sexual problems.¹³

It is important for the family physician to recognize the spectrum of sexual problems among men, since these concerns may affect their patients’ health, wellbeing, and relationships. There are only a few small studies on sexual dysfunction in family practice,^2-4,14 with limited generalizability. Most of the authors did not correlate physicians’ and patients’ views on this topic. Little is known about what men expect from their family physicians in cases of sexual disorders. We performed a survey of male patients and their physicians, focusing on the frequency and types of sexual problems in family practice, the men’s expectations of their physicians in case of sexual problems, the physicians’ perception and management of sexual problems, and the influence of physicians’ knowledge of and attitude toward sexual medicine on patients’ help-seeking behavior.

Methods

Study Population

All board-certified family physicians in the 2 districts of Hildesheim (Lower Saxony) and Heiligenstadt (Thuringia, the former East Germany) in the North of Germany were asked to take part in the study and complete a self-administered questionnaire. For one day all male patients in the participating practices 18 years or older and capable of reading the German language were asked to complete a questionnaire in a separate room of the practice and, after finishing, to put it into a box. Neither the physician nor the practice nurses had access to the box.

Questionnaires

We developed a 3-page questionnaire for family physicians and a 4-page patient questionnaire. A multiple-choice format was chosen with some room for personal comments. The patient questionnaire was pilot tested on a sample of 20 men attending a family practice; these men did not take part in the final study. The patient questionnaire had 3 parts: the quality of the patient’s relationship with the family physician with regard to sexual problems; the frequency, presumed causes, and effects of sexual dysfunction; and the patient’s view of the physician’s management of sexual problems.

In the second part of the questionnaire a list of potential sexual problems was presented. The patients were asked to indicate whether they suffered never, seldom, sometimes, often, or always from 1 or more of these problems.

The physician questionnaire, which was piloted in cooperation with 3 family physicians, focused on the frequency of sexual problems in male patients, the treatment of sexual problems in family practice, and the reasons for not talking about sexual problems with the patient.

Data were analyzed using SAS software, version 8 for Windows (SAS Institute; Cary, NC). To test for differences, we used the chi-square test, as appropriate. Correlations among ordinal variables were determined by the Spearman rank correlation.¹⁵ The units of analysis were the patient and the physician, respectively. For comparison of the patients’ and their physicians’ statements, the unit of analysis was the practice. Since not all patients answered all items of the questionnaire, we report the denominator where relevant.

Results

Physicians and Patients

Of the 46 family practices approached, 20 (43%), all single handed, gave approval to interview patients in their surgery. Seventeen of these practices were located in a major city, and 3 were in a small town. Nearly all (19/20) answered the physician questionnaire. The participating physicians were, on average, aged 49 years (range=36-61 years). When asked why they did not want to take part in the study, physicians most often said they had only a small number of patients with sexual problems in their practice (6/26) or their patients would not consent to a study dealing with sexuality (8/26). Only 21% of the women physicians addressed took part in the study, compared with 53% of the men.

A total of 307 men (84%) returned the patient questionnaire. Lack of time, feeling too sick, and unwillingness were the reasons most often given by the 57 patients who refused to participate. Demographic characteristics are shown in Table 1. The age distribution is compared with the representative ADT (Accounting Data Record) Panel of the Central Institute of Ambulatory Health Care in Germany (Zentralinstitut für die kassenärztliche Versorgung).^16,17

More than 90% of the participating men (279/307) had been consulting their particular physician for more than 1 year. According to their own assessment, 245 patients had informed their physicians about their personal, family, and social background.

Frequency of Sexual Problems

More than 90% of the male patients (251/270) reported having 1 sexual problem at least seldom. Figure 1 illustrates the frequency of sexual problems. These results were not influenced by the patients’ marital status.

Low sexual desire (73%) and premature ejaculation (66%) were dominant problems. More than one third of the respondents suffered at least sometimes from these problems Figure 2. Approximately 5% to 6% of the men reported fear of failure and erectile dysfunction as often or permanent problems. Thirty-seven men (12%) did not answer any of the questionnaire items that referred to sexual problems. These men did not differ in marital status from those who answered these items; however, on average the 37 men were older than the remainder (55 years vs 44 years). Even if we make the (rather unlikely) assumption that all patients who did not answer the sex-related items represent men without any sexual problems, more than 80% of the sample (251/307) suffered at least occasionally from a reported sexual problem. Low sexual desire and erectile problems were weakly or moderately associated with age (rho=0.18; P=.0049 and rho=0.38; P=.0001, respectively).

Half of the patients who had sexual concerns (107/201) considered occupational stress to be a cause of their problems. Pressure resulting from expectations of self (28%), difficulties in relationships (19%), and comorbid diseases (13%) were also considered common causes of their sexual dysfunction. Less common causes included medicines (9%), inhibitions resulting from strict sexual education (8%), pressure resulting from expectations of others (8%), and involuntary childlessness (3%).

As a consequence of their sexual difficulties, 46 of 257 (18%) experienced depression and depression-like symptoms, 36 (14%) sleeplessness, and 35 (14%) difficulties in their relationships. Patients’ satisfaction with their sexual lives was moderately correlated with the frequency of their sexual problems (rho = -0.33; P <.001). Patients who reported suffering sexual problems never or seldom reported satisfaction scores of 7.5 or 7.6, respectively. Patients reporting sexual problems often or always were less satisfied Figure 3. Sexual-life satisfaction and age were not correlated (rho = -0.05; P=.41)

Patients’ Expectations

Most patients considered it important to talk with their family physicians about their sexual concerns (84%). Almost half of the respondents (133/295) preferred that their physician initiate any discussions about sexuality. These expectations were not influenced by patient age (data not shown). More than two thirds of the respondents would have liked their physicians to signal his or her open-mindedness by directly addressing sexual topics during the consultation.

Management of Sexual Problems

Only 12% (34/288) of the respondents had already consulted their family physicians because of sexual problems, and most (23/34) of these patients were satisfied with the physician’s treatment (discussion about the difficulties in 56% of cases, prescription of drugs in 29%, and specialist referral in 24%).

The physicians remembered only a few consultations about sexual problems. On average, they recalled seeing 7 patients every 3 months primarily because of sexual difficulties. In an additional 6 patient contacts, sexual problems were raised during the consultation.

The physicians stated that they initiated discussion about sexual problems only sometimes (53%) or seldom (37%) during the consultation. They considered the following occasions or conditions as good opportunities to start a discussion: psychosomatic complaints (16/19; 84%), family planning (53%), questions concerning human immunodeficiency virus (47%), and diseases, such as diabetes mellitus (79%) or hypertension (63%).

In cases of sexual problems, most family physicians in our sample would have changed medication if relevant (79%) or would have referred patients to urologists (74%). According to the experience of approximately half of the physicians (9/19), sildenafil motivated more patients with sexual concerns to consult their physicians.

The physicians were asked for possible reasons for not talking about sexuality Table 2. Regarding physician-related factors, most physicians (53%) considered the lack of time a barrier to addressing sexual topics. Nearly all physicians (17/19) presumed that a feeling of shame keeps patients from talking about sexuality.

Importance of Sexual Medicine

On average, physicians rated the importance of sexual knowledge in family practice at a value of 7.2 on a 10-point Likert scale where 1 = not important at all and 10 = highly important. They rated the quality of their own knowledge 6.05.

There was a (nonsignificant) correlation between the physicians’ assumed knowledge and the patients’ wish to contact them in the case of sexual problems (rho=0.26; P=.29). Correspondingly, men were more likely to contact their physician for sexual problems if the physician assessed counseling in sexual medicine to be more important (rho=0.18; P=.47). The practice was the unit of analysis for calculating these correlations and their P values.

Discussion

Our study shows that sexual problems are widespread among male family practice patients and confirms the public health importance of this area. The high rate of men reporting sexual problems (93%) includes all patients who have or have had a sexual problem, independently of how often they suffered from it. Even if we consider the missing data (n=37) as representing patients without any sexual problem, the rate is still more than 80% and exceeds the incidence found by other authors.^3,4,14,18 Our results confirm the survey of Metz and Seifert,² in which nearly every patient had a sexual problem (97%) at some time during his life. Interestingly, this situation does not seem to be significantly different from the experience of women patients, according to a recently published survey of sexual problems among women seeking routine gynecologic care from the departments of family practice and obstetrics and gynecology.¹⁹ Nearly all the 964 women (99%) reported 1 sexual concern or more.

Health surveys reflect self-reported problems, not medically defined diseases. However, because sexual problems affect well-being and interact with other somatic and psychological complaints, family physicians should be aware that nearly all patients experience and report sexual concerns. Also, it should be recognized that 20% of patients suffer often or always from at least 1 sexual problem. These patients were significantly less satisfied with their sexual life than the remainder.

Similar to the results of other studies,^1,9,20,21 we found that erectile dysfunction and low sexual desire increase with age. However, this was only a moderate correlation and should alert physicians to be cautious about regarding erectile dysfunction as an elderly man’s complaint. According to the patients’ own assessments, many sexual problems are related to occupational stress. Therefore, the family physician, with his or her knowledge of the patient’s family and social situation, is well prepared to identify men with sexual problems. For some men of our sample, these problems may be associated with sleeplessness or depression. In these cases, the family physician should address the interaction of diseases or drugs with sexual disorders.

The participating physicians were hesitant about initiating a discussion about sexual problems. This corresponds with the results of the patient survey, as only 12% of the men had already consulted their physicians in cases of sexual concerns. Half of the physicians stated that they have had more consultations about sexual concerns since the discussion about sildenafil started. This is in line with the hypothesis of Tiefer¹³ that the availability of sildenafil could motivate an increasing number of patients suffering from sexual problems to consult their family physicians.

Family physicians considered lack of time and knowledge as the most important reasons for not taking a sexual history. In a Dutch study,²² these obstacles were also put forward by the 59% of general practitioners who did not take a sexual history even when an erectile problem was suspected. Family physicians in our study also mentioned a feeling of shame as a barrier for patients and themselves. These arguments also evolved in a survey about the management of involuntary childlessness in German family practice.²³ The interviewed physicians in that study were afraid of intrusion and inadequacy when addressing involuntary childlessness and related areas, such as sexuality, which are of an intimate and private nature. To overcome problems such as lack of knowledge or feelings of shame and inadequacy, it may be useful to give the topic of sexuality and sexual problems a higher status in basic medical education, as well as in continuing medical education. Also, key questions on how to start and maintain discussions on sexuality should be developed, evaluated, and then rehearsed.

Our results seem to indicate a correlation between patients’ consulting behavior and physicians’ attitudes and knowledge. More patients talk about sexual concerns to their physicians when the physicians assume that their own knowledge and attitude toward sexual medicine is better than that of their colleagues. This could represent a self-selection process as to patients’ preference. Because these correlations were analyzed on the practice level (as the unit of analysis), the results may not have reached statistical significance because of the low number of observations. Further studies may determine these associations more precisely.

We do not suggest that all patients suffering from sexual problems require treatment. Many of the patients in our study who reported seldom experiencing a sexual problem were just as satisfied with their sexual life as those who reported never having a problem. To date, some family physicians seem to resolve the difficulty of finding out only those sexual problems that need to be treated by neglecting the topic of sexuality during consultation. The physicians’ assumption that this attitude corresponds to the patients’ feeling of shame is not supported by our data: More than half of the patients would appreciate their physicians addressing the topic of sexuality in an active and open way, if appropriate.

Limitations

Compared with patients of the ADT Panel,^16,17 older patients, especially those older than 70 years, are under-represented in our study. This may be because this study was performed exclusively in surgeries, while the ADT Panel also included patients who were visited at home and in nursing homes. However, it should be emphasized that the response rate of 84% was excellent, so that a selection bias with regard to patients visiting the surgery is rather unlikely.

A response rate of 43% of the physicians may indicate a selection bias. Physicians who were more open-minded toward sexual problems may have been more likely to participate in our survey. Thus, sexual problems are probably an even less frequent topic during consultation than our data suggest.

Conclusions

The high level of self-reported sexual dysfunction in men and family physicians’ hesitancy to deal with this issue signals a neglected area in primary health care. Possible starting points for communication about sexuality in family practice include the prescription of drugs that trigger sexual disorders, diagnosis of conditions that are associated with sexual problems, and the patients’ personal, social, and occupational background.

Acknowledgments

Parts of our study were supported by Grant 01 KY 9605/7 from the Federal Ministry of Education and Research, Bonn, Germany. The authors would like to thank the practice nurses, physicians, and patients who participated in this survey and answered questions of a very personal nature.

Sexual dysfunction in men is common. A large study¹ in the United States with a sample of 1410 men found a 31% 12-month prevalence of sexual dysfunction. In a small US survey² of 62 men, nearly all participants reported sexual concerns at any time during their life. One third of 789 men in a British general practice reported a current sexual problem.³ Three fourths of 78 patients surveyed in another British general practice suffered from general sexual problems; 35% reported a specific sexual dysfunction, such as premature ejaculation or erectile dysfunction.⁴

Sexual dysfunction may have organic or psychological causes. The family practice setting seems to be ideal for the evaluation and management of sexual dysfunction, because family physicians usually know the personal and family situation of their patients.^5-8Also, sexual dysfunction is sometimes related to common diseases, such as diabetes mellitus.^9,10 Medications frequently prescribed in family practice, such as antihypertensive and psychiatric drugs, may also adversely affect sexual performance.^11,12 Finally, the popularity of sildenafil (Viagra) and the accompanying publicity may encourage an increasing number of patients with erectile problems to consult their family physicians in case of sexual problems.¹³

It is important for the family physician to recognize the spectrum of sexual problems among men, since these concerns may affect their patients’ health, wellbeing, and relationships. There are only a few small studies on sexual dysfunction in family practice,^2-4,14 with limited generalizability. Most of the authors did not correlate physicians’ and patients’ views on this topic. Little is known about what men expect from their family physicians in cases of sexual disorders. We performed a survey of male patients and their physicians, focusing on the frequency and types of sexual problems in family practice, the men’s expectations of their physicians in case of sexual problems, the physicians’ perception and management of sexual problems, and the influence of physicians’ knowledge of and attitude toward sexual medicine on patients’ help-seeking behavior.

Methods

Study Population

All board-certified family physicians in the 2 districts of Hildesheim (Lower Saxony) and Heiligenstadt (Thuringia, the former East Germany) in the North of Germany were asked to take part in the study and complete a self-administered questionnaire. For one day all male patients in the participating practices 18 years or older and capable of reading the German language were asked to complete a questionnaire in a separate room of the practice and, after finishing, to put it into a box. Neither the physician nor the practice nurses had access to the box.

Questionnaires

We developed a 3-page questionnaire for family physicians and a 4-page patient questionnaire. A multiple-choice format was chosen with some room for personal comments. The patient questionnaire was pilot tested on a sample of 20 men attending a family practice; these men did not take part in the final study. The patient questionnaire had 3 parts: the quality of the patient’s relationship with the family physician with regard to sexual problems; the frequency, presumed causes, and effects of sexual dysfunction; and the patient’s view of the physician’s management of sexual problems.

In the second part of the questionnaire a list of potential sexual problems was presented. The patients were asked to indicate whether they suffered never, seldom, sometimes, often, or always from 1 or more of these problems.

The physician questionnaire, which was piloted in cooperation with 3 family physicians, focused on the frequency of sexual problems in male patients, the treatment of sexual problems in family practice, and the reasons for not talking about sexual problems with the patient.

Data were analyzed using SAS software, version 8 for Windows (SAS Institute; Cary, NC). To test for differences, we used the chi-square test, as appropriate. Correlations among ordinal variables were determined by the Spearman rank correlation.¹⁵ The units of analysis were the patient and the physician, respectively. For comparison of the patients’ and their physicians’ statements, the unit of analysis was the practice. Since not all patients answered all items of the questionnaire, we report the denominator where relevant.

Results

Physicians and Patients

Of the 46 family practices approached, 20 (43%), all single handed, gave approval to interview patients in their surgery. Seventeen of these practices were located in a major city, and 3 were in a small town. Nearly all (19/20) answered the physician questionnaire. The participating physicians were, on average, aged 49 years (range=36-61 years). When asked why they did not want to take part in the study, physicians most often said they had only a small number of patients with sexual problems in their practice (6/26) or their patients would not consent to a study dealing with sexuality (8/26). Only 21% of the women physicians addressed took part in the study, compared with 53% of the men.

A total of 307 men (84%) returned the patient questionnaire. Lack of time, feeling too sick, and unwillingness were the reasons most often given by the 57 patients who refused to participate. Demographic characteristics are shown in Table 1. The age distribution is compared with the representative ADT (Accounting Data Record) Panel of the Central Institute of Ambulatory Health Care in Germany (Zentralinstitut für die kassenärztliche Versorgung).^16,17

More than 90% of the participating men (279/307) had been consulting their particular physician for more than 1 year. According to their own assessment, 245 patients had informed their physicians about their personal, family, and social background.

Frequency of Sexual Problems

More than 90% of the male patients (251/270) reported having 1 sexual problem at least seldom. Figure 1 illustrates the frequency of sexual problems. These results were not influenced by the patients’ marital status.

Low sexual desire (73%) and premature ejaculation (66%) were dominant problems. More than one third of the respondents suffered at least sometimes from these problems Figure 2. Approximately 5% to 6% of the men reported fear of failure and erectile dysfunction as often or permanent problems. Thirty-seven men (12%) did not answer any of the questionnaire items that referred to sexual problems. These men did not differ in marital status from those who answered these items; however, on average the 37 men were older than the remainder (55 years vs 44 years). Even if we make the (rather unlikely) assumption that all patients who did not answer the sex-related items represent men without any sexual problems, more than 80% of the sample (251/307) suffered at least occasionally from a reported sexual problem. Low sexual desire and erectile problems were weakly or moderately associated with age (rho=0.18; P=.0049 and rho=0.38; P=.0001, respectively).

Half of the patients who had sexual concerns (107/201) considered occupational stress to be a cause of their problems. Pressure resulting from expectations of self (28%), difficulties in relationships (19%), and comorbid diseases (13%) were also considered common causes of their sexual dysfunction. Less common causes included medicines (9%), inhibitions resulting from strict sexual education (8%), pressure resulting from expectations of others (8%), and involuntary childlessness (3%).

As a consequence of their sexual difficulties, 46 of 257 (18%) experienced depression and depression-like symptoms, 36 (14%) sleeplessness, and 35 (14%) difficulties in their relationships. Patients’ satisfaction with their sexual lives was moderately correlated with the frequency of their sexual problems (rho = -0.33; P <.001). Patients who reported suffering sexual problems never or seldom reported satisfaction scores of 7.5 or 7.6, respectively. Patients reporting sexual problems often or always were less satisfied Figure 3. Sexual-life satisfaction and age were not correlated (rho = -0.05; P=.41)

Patients’ Expectations

Most patients considered it important to talk with their family physicians about their sexual concerns (84%). Almost half of the respondents (133/295) preferred that their physician initiate any discussions about sexuality. These expectations were not influenced by patient age (data not shown). More than two thirds of the respondents would have liked their physicians to signal his or her open-mindedness by directly addressing sexual topics during the consultation.

Management of Sexual Problems

Only 12% (34/288) of the respondents had already consulted their family physicians because of sexual problems, and most (23/34) of these patients were satisfied with the physician’s treatment (discussion about the difficulties in 56% of cases, prescription of drugs in 29%, and specialist referral in 24%).

The physicians remembered only a few consultations about sexual problems. On average, they recalled seeing 7 patients every 3 months primarily because of sexual difficulties. In an additional 6 patient contacts, sexual problems were raised during the consultation.

The physicians stated that they initiated discussion about sexual problems only sometimes (53%) or seldom (37%) during the consultation. They considered the following occasions or conditions as good opportunities to start a discussion: psychosomatic complaints (16/19; 84%), family planning (53%), questions concerning human immunodeficiency virus (47%), and diseases, such as diabetes mellitus (79%) or hypertension (63%).

In cases of sexual problems, most family physicians in our sample would have changed medication if relevant (79%) or would have referred patients to urologists (74%). According to the experience of approximately half of the physicians (9/19), sildenafil motivated more patients with sexual concerns to consult their physicians.

The physicians were asked for possible reasons for not talking about sexuality Table 2. Regarding physician-related factors, most physicians (53%) considered the lack of time a barrier to addressing sexual topics. Nearly all physicians (17/19) presumed that a feeling of shame keeps patients from talking about sexuality.

Importance of Sexual Medicine

On average, physicians rated the importance of sexual knowledge in family practice at a value of 7.2 on a 10-point Likert scale where 1 = not important at all and 10 = highly important. They rated the quality of their own knowledge 6.05.

There was a (nonsignificant) correlation between the physicians’ assumed knowledge and the patients’ wish to contact them in the case of sexual problems (rho=0.26; P=.29). Correspondingly, men were more likely to contact their physician for sexual problems if the physician assessed counseling in sexual medicine to be more important (rho=0.18; P=.47). The practice was the unit of analysis for calculating these correlations and their P values.

Discussion

Our study shows that sexual problems are widespread among male family practice patients and confirms the public health importance of this area. The high rate of men reporting sexual problems (93%) includes all patients who have or have had a sexual problem, independently of how often they suffered from it. Even if we consider the missing data (n=37) as representing patients without any sexual problem, the rate is still more than 80% and exceeds the incidence found by other authors.^3,4,14,18 Our results confirm the survey of Metz and Seifert,² in which nearly every patient had a sexual problem (97%) at some time during his life. Interestingly, this situation does not seem to be significantly different from the experience of women patients, according to a recently published survey of sexual problems among women seeking routine gynecologic care from the departments of family practice and obstetrics and gynecology.¹⁹ Nearly all the 964 women (99%) reported 1 sexual concern or more.

Health surveys reflect self-reported problems, not medically defined diseases. However, because sexual problems affect well-being and interact with other somatic and psychological complaints, family physicians should be aware that nearly all patients experience and report sexual concerns. Also, it should be recognized that 20% of patients suffer often or always from at least 1 sexual problem. These patients were significantly less satisfied with their sexual life than the remainder.

Similar to the results of other studies,^1,9,20,21 we found that erectile dysfunction and low sexual desire increase with age. However, this was only a moderate correlation and should alert physicians to be cautious about regarding erectile dysfunction as an elderly man’s complaint. According to the patients’ own assessments, many sexual problems are related to occupational stress. Therefore, the family physician, with his or her knowledge of the patient’s family and social situation, is well prepared to identify men with sexual problems. For some men of our sample, these problems may be associated with sleeplessness or depression. In these cases, the family physician should address the interaction of diseases or drugs with sexual disorders.

The participating physicians were hesitant about initiating a discussion about sexual problems. This corresponds with the results of the patient survey, as only 12% of the men had already consulted their physicians in cases of sexual concerns. Half of the physicians stated that they have had more consultations about sexual concerns since the discussion about sildenafil started. This is in line with the hypothesis of Tiefer¹³ that the availability of sildenafil could motivate an increasing number of patients suffering from sexual problems to consult their family physicians.

Family physicians considered lack of time and knowledge as the most important reasons for not taking a sexual history. In a Dutch study,²² these obstacles were also put forward by the 59% of general practitioners who did not take a sexual history even when an erectile problem was suspected. Family physicians in our study also mentioned a feeling of shame as a barrier for patients and themselves. These arguments also evolved in a survey about the management of involuntary childlessness in German family practice.²³ The interviewed physicians in that study were afraid of intrusion and inadequacy when addressing involuntary childlessness and related areas, such as sexuality, which are of an intimate and private nature. To overcome problems such as lack of knowledge or feelings of shame and inadequacy, it may be useful to give the topic of sexuality and sexual problems a higher status in basic medical education, as well as in continuing medical education. Also, key questions on how to start and maintain discussions on sexuality should be developed, evaluated, and then rehearsed.

Our results seem to indicate a correlation between patients’ consulting behavior and physicians’ attitudes and knowledge. More patients talk about sexual concerns to their physicians when the physicians assume that their own knowledge and attitude toward sexual medicine is better than that of their colleagues. This could represent a self-selection process as to patients’ preference. Because these correlations were analyzed on the practice level (as the unit of analysis), the results may not have reached statistical significance because of the low number of observations. Further studies may determine these associations more precisely.

We do not suggest that all patients suffering from sexual problems require treatment. Many of the patients in our study who reported seldom experiencing a sexual problem were just as satisfied with their sexual life as those who reported never having a problem. To date, some family physicians seem to resolve the difficulty of finding out only those sexual problems that need to be treated by neglecting the topic of sexuality during consultation. The physicians’ assumption that this attitude corresponds to the patients’ feeling of shame is not supported by our data: More than half of the patients would appreciate their physicians addressing the topic of sexuality in an active and open way, if appropriate.

Limitations

Compared with patients of the ADT Panel,^16,17 older patients, especially those older than 70 years, are under-represented in our study. This may be because this study was performed exclusively in surgeries, while the ADT Panel also included patients who were visited at home and in nursing homes. However, it should be emphasized that the response rate of 84% was excellent, so that a selection bias with regard to patients visiting the surgery is rather unlikely.

A response rate of 43% of the physicians may indicate a selection bias. Physicians who were more open-minded toward sexual problems may have been more likely to participate in our survey. Thus, sexual problems are probably an even less frequent topic during consultation than our data suggest.

Conclusions

The high level of self-reported sexual dysfunction in men and family physicians’ hesitancy to deal with this issue signals a neglected area in primary health care. Possible starting points for communication about sexuality in family practice include the prescription of drugs that trigger sexual disorders, diagnosis of conditions that are associated with sexual problems, and the patients’ personal, social, and occupational background.

Acknowledgments

Parts of our study were supported by Grant 01 KY 9605/7 from the Federal Ministry of Education and Research, Bonn, Germany. The authors would like to thank the practice nurses, physicians, and patients who participated in this survey and answered questions of a very personal nature.

Frame and Carlson¹ published a critical review of periodic health screenings in The Journal of Family Practice in 1975. Such efforts, currently manifested in the United States Preventive Services Task Force (USPSTF), have had a profound impact on the practice of family physicians. Increased oversight of family physicians’ preventive practices by third parties and the increased medical sophistication of the public have fueled demand for preventive health services. Family physicians are experiencing increased accountability to their patients and communities for the effective delivery of preventive medical care.^2-4

A number of systems have been developed to assist physicians with the implementation of evidence-based preventive services guidelines. Over the past 15 years, numerous researchers have studied the effects of interventions using flow charts,^5-7 checklists,⁸ computer-based reminders,^9-14 and educational programs^15-19 for delivery of preventive services. Ornstein and colleagues¹⁰ found that the use of a computer-based tracking and reminder system for preventive services, together with a physician education program, modestly increased the frequency of counseling and screening services for patients. Similarly, Cheney and Ramsdell⁸ found that physicians who received an age- and sex-specific checklist in their patients’ medical charts performed significantly more age- and sex-appropriate preventive services than control physicians. But physician compliance with screening tools is often a problem.⁶ In the Cheney and Ramsdell study, only 39% of physicians used the available checklist to screen patients. Another study¹⁴ found a greater percentage of appropriate health screens by physicians using computerized reminders. However, many physicians did not make use of the available resource. Some studies found only minimal increases in screening rates with interventions such as educational seminars.¹⁹ One study⁵ used a simple flow sheet to analyze appropriate health service needs and noted that their results suggest that a flow sheet, although helpful, is not enough. Indeed, accomplishing sustained improvement in preventive services clinical outcomes is a complex process.²⁰

We investigated the effect of combining a complete set of annotated recommendations with a history and physical examination form on the rate of appropriate application of preventive service recommendations. We suspected that busy practitioners would need preventive services guidelines available at the point of care to be able to adhere closely, so we created a form that is very user-friendly. We also studied the effect of removing the screening tool after the intervention, to evaluate whether the behavioral effect was due to education or to the presence of the chart tool, a question that has not been resolved in several previous studies.

Methods

Study Setting and Instrument

Our study was conducted between June 1997 and June 1998 at a community hospital continuity clinic. Age-appropriate screening history questions, age-specific reminders, and test frequency recommendations drawn from the 1996 USPSTF were integrated into a form for documenting history and physical examinations. This Comprehensive Annotated Reminder Tool (CART) contained 8 sections: history, screening history, physical examination, screening mnemonics, laboratory screens, prophylaxis, counseling, and the assessment/plan section. CART forms were developed for patients aged 19-39, 40-64, and 65 or older. Table w1,Table w2, Table w3 3* Each of these 3-page forms contained guidelines for up to 49 preventive interventions. Twenty recommendations were evaluated in this study. Because nursing has been primarily responsible for ensuring immunization compliance in our setting, we did not expect an effect and only evaluated one immunization.

Intervention

Resident physicians were randomly assigned to a treatment group that was exposed to the CART (n=15) and a control group that used existing blank history and physical examination forms (n=16). Resident physicians from both groups precepted patients with the same faculty physicians on a regular basis. Physicians in the control group used a different area of the clinic facility from the treatment group to avoid cross-use of the CART form. The 3 time periods of the study were: pre-intervention phase (3 months), intervention phase (6 months), and post-intervention phase (3 months). The pre-intervention phase allowed for the establishment of a baseline measurement of preventive medicine practiced by physicians in both groups. In the intervention phase, the CART forms were placed on all charts of new patients seen by physicians in the experimental group. The forms were not introduced by any formal instruction. No preventive medicine lectures were given to physicians in either group during the course of the study. Finally, in the post-intervention phase the CART was no longer placed on charts.

Measures

Blinded chart reviews were performed by the principal investigator and 2 other independent reviewers on randomly selected new patients. The physicians were expected to complete a history and physical examination on each new clinic patient within the first 3 visits. Chart reviews occurred before use, during use, and after removal of the CART. They evaluated charts for screening appropriateness (not the absolute frequency of interventions). No credit was given when a preventive intervention was performed that was not indicated. A total of 300 charts were reviewed for the treatment group and 308 for the control group. The inter-rater agreement between the principal investigator and 2 other independent reviewers ranged between 93% and 98%. The median kappa statistic among 16 screening recommendations revealed very good agreement at 0.81 (all P <.042). Kappa could not be calculated for 4 of the screening recommendations, because the observed agreement of the 3 raters (100%) was equal to the expected agreement for those recommendations. All physicians were given a pretest measuring knowledge of the USPSTF recommendations at the beginning of the pre-intervention phase and a posttest near the end of the post-intervention phase. Finally, as physician learning and behavior were the focus of this study, patient adherence was not measured.

Data Analysis

Recommendations were organized into 4 categories: history, examination and laboratory, counseling, and prophylaxis. Tests of significance were calculated using a chi-square test that is designed to compare proportions among many independent samples.²¹ To control for statistical error, we used a Bonferroni adjusted P value. Because 20 screening recommendations were analyzed, a P value of .0025 (.05/20) was used as the cutoff for statistical significance.

Scores from the knowledge test were analyzed using a mixed factorial analysis of variance. The between-groups factor was group (treatment or control) and the within-subjects factor was test period (pretest or posttest). This analysis was conducted to see if the tool would lead to a group-by-test period interaction. Given the repeated use of the CART, it was thought that the treatment group might show a greater improvement in test scores than the control group.

Results

Impact of the CART

The Table 1 shows the percentages of patients who were appropriately screened during each period for the treatment and control groups. Significant increases were observed for all recommendations in the treatment group, except for occult blood and tetanus and diptheria booster. The mean absolute increase in the percentage of appropriately screened patients was 45%, 21%, and 15% for the recommendations in the history, examination and laboratory, and counseling categories, respectively. Removal of the annotated reminder during the post-intervention period brought the percentages of appropriate screenings down to baseline levels. Increases during the intervention phase were statistically significant for 16 of the 20 recommendations.

In the control group, the percentages of appropriately screened patients remained relatively stable. Although some variation is visible across the 3 study periods, there are no consistent trends among the recommendations, and none of the chi-square tests reached the Bonferroni level of significance. The Figure 1 shows the median percentage of patients screened during the baseline, intervention, and post-intervention periods. The average percentage of appropriate screenings for the clinicians using the CART increased during the intervention period and then returned to baseline levels after the tool was removed from use. The inverted V pattern repeatedly observed for the treatment group is not seen with the control group.

Screening Knowledge Scores

Only the main effect for the test period reached statistical significance. The mean test scores were 54.66 and 62.35 for clinicians at the pretest and posttest periods, respectively (F [1,27] = 23.89; P <.0005). The main effect for group was not statistically significant. The mean test scores for the treatment and control groups were 58.59 and 58.13, respectively (F [1,27] = 0.27; P <.608). The group-by-test period interaction also was not statistically significant (F [1,27] = 1.42, P=.244). Thus, although physicians had higher posttest knowledge scores than pretest scores, this increase was similar for treatment and control groups.

Discussion

CART use increased performance both in degree (percentage change) and in breadth (number of preventive services ordered appropriately). The primary outcome was the appropriateness of physician performance of preventive interventions. For example, if the patient is low-risk, then gonorrhea/chlamydia screening should not be performed, and appropriate physician behavior is to do no screening test. We did not give lectures or provide any other didactic preventive medicine training during the CART study period. But knowledge increased to a similar extent in both groups over that academic year. Nevertheless, the behavioral change seen in the treatment group was temporary; it returned to baseline in the post-intervention period. Therefore, it was the presence of the CART that improved physician behavior. The CART is comprehensive: It has prompts for up to 49 preventive recommendations. Physicians adhered significantly better to history and examination and laboratory recommendations when using the CART. The improvement in counseling was less impressive; physician behavior improved on only 4 of 6 counseling recommendations. It is likely that “form fatigue” was a factor, because the counseling recommendations were at the bottom of a page full of preventive recommendations. The difficulty of counseling, time constraints, and physician perception of the relative ineffectiveness of counseling may have also had an effect.

We believe that the primary reason for the success of the CART is immediate visual access to essential information. Computer reminders promote some behavioral change, but in other studies it was necessary to go to another computer screen for further information about a given test. Time pressure in the busy office makes this difficult. The CART gives physicians succinct information in a box adjacent to each preventive recommendation Table 1, including the recommended frequency, strength of recommendation, and appropriate groups to be screened for each intervention. Most other studies did not answer the question of whether increased knowledge over time was responsible for improved provider behavior. By removing the CART and reassessing physician behavior, we demonstrated that it was presence of the information in the CART format that was responsible for improving physician behavior.

Limitations

Our study has several limitations. We did not address the important issue of reminder systems or checklists. We only assessed the preventive services recommended by physicians during a single periodic health examination. We did not solve the problem of time constraint. It is difficult to cover every single preventive service recommendation in one visit. Our residents are less experienced and have more visit time than the average primary care provider. Finally, we only studied physician knowledge and behavior. We did not measure other important outcomes, such as patient satisfaction, adherence, morbidity, or mortality.

Conclusions

The nonspecific designation “tool” in the CART acronym implies potential future application of this concept for different subject matter and different formats. As we move inexorably toward universal electronic medical records, the CART can be adapted for electronic use. But whether on desktops, laptops, or hand-held computers, immediate visual access to information is essential. Combining the CART with computer reminder systems could yield significant improvement in patient outcomes when followed over time. Also, it would be interesting to incorporate the CART information into a computer-gathered patient history to create a partially completed history and physical examination form for provider use during the periodic health examination. This would save valuable office visit time. Many previous studies have focused on a small number of preventive services. We chose a comprehensive set of recommendations to more closely mimic the experience of primary care physicians who must cover many recommendations with patients. It is important that more studies use this approach. To maximize impact on the public health in an era of cost containment and to minimize physician error, we must enable physicians to efficiently apply a comprehensive set of preventive services in an appropriately focused manner. The CART is an effective step in the right direction.

Acknowledgments

We thank the Advocate Christ Hospital Med Fund for grant support. We would also like to thank Charles Warnell, MD, for his chart review work, and Carolyn Barsano for the literature review.

Frame and Carlson¹ published a critical review of periodic health screenings in The Journal of Family Practice in 1975. Such efforts, currently manifested in the United States Preventive Services Task Force (USPSTF), have had a profound impact on the practice of family physicians. Increased oversight of family physicians’ preventive practices by third parties and the increased medical sophistication of the public have fueled demand for preventive health services. Family physicians are experiencing increased accountability to their patients and communities for the effective delivery of preventive medical care.^2-4

A number of systems have been developed to assist physicians with the implementation of evidence-based preventive services guidelines. Over the past 15 years, numerous researchers have studied the effects of interventions using flow charts,^5-7 checklists,⁸ computer-based reminders,^9-14 and educational programs^15-19 for delivery of preventive services. Ornstein and colleagues¹⁰ found that the use of a computer-based tracking and reminder system for preventive services, together with a physician education program, modestly increased the frequency of counseling and screening services for patients. Similarly, Cheney and Ramsdell⁸ found that physicians who received an age- and sex-specific checklist in their patients’ medical charts performed significantly more age- and sex-appropriate preventive services than control physicians. But physician compliance with screening tools is often a problem.⁶ In the Cheney and Ramsdell study, only 39% of physicians used the available checklist to screen patients. Another study¹⁴ found a greater percentage of appropriate health screens by physicians using computerized reminders. However, many physicians did not make use of the available resource. Some studies found only minimal increases in screening rates with interventions such as educational seminars.¹⁹ One study⁵ used a simple flow sheet to analyze appropriate health service needs and noted that their results suggest that a flow sheet, although helpful, is not enough. Indeed, accomplishing sustained improvement in preventive services clinical outcomes is a complex process.²⁰

We investigated the effect of combining a complete set of annotated recommendations with a history and physical examination form on the rate of appropriate application of preventive service recommendations. We suspected that busy practitioners would need preventive services guidelines available at the point of care to be able to adhere closely, so we created a form that is very user-friendly. We also studied the effect of removing the screening tool after the intervention, to evaluate whether the behavioral effect was due to education or to the presence of the chart tool, a question that has not been resolved in several previous studies.

Methods

Study Setting and Instrument

Our study was conducted between June 1997 and June 1998 at a community hospital continuity clinic. Age-appropriate screening history questions, age-specific reminders, and test frequency recommendations drawn from the 1996 USPSTF were integrated into a form for documenting history and physical examinations. This Comprehensive Annotated Reminder Tool (CART) contained 8 sections: history, screening history, physical examination, screening mnemonics, laboratory screens, prophylaxis, counseling, and the assessment/plan section. CART forms were developed for patients aged 19-39, 40-64, and 65 or older. Table w1,Table w2, Table w3 3* Each of these 3-page forms contained guidelines for up to 49 preventive interventions. Twenty recommendations were evaluated in this study. Because nursing has been primarily responsible for ensuring immunization compliance in our setting, we did not expect an effect and only evaluated one immunization.

Intervention

Resident physicians were randomly assigned to a treatment group that was exposed to the CART (n=15) and a control group that used existing blank history and physical examination forms (n=16). Resident physicians from both groups precepted patients with the same faculty physicians on a regular basis. Physicians in the control group used a different area of the clinic facility from the treatment group to avoid cross-use of the CART form. The 3 time periods of the study were: pre-intervention phase (3 months), intervention phase (6 months), and post-intervention phase (3 months). The pre-intervention phase allowed for the establishment of a baseline measurement of preventive medicine practiced by physicians in both groups. In the intervention phase, the CART forms were placed on all charts of new patients seen by physicians in the experimental group. The forms were not introduced by any formal instruction. No preventive medicine lectures were given to physicians in either group during the course of the study. Finally, in the post-intervention phase the CART was no longer placed on charts.

Measures

Blinded chart reviews were performed by the principal investigator and 2 other independent reviewers on randomly selected new patients. The physicians were expected to complete a history and physical examination on each new clinic patient within the first 3 visits. Chart reviews occurred before use, during use, and after removal of the CART. They evaluated charts for screening appropriateness (not the absolute frequency of interventions). No credit was given when a preventive intervention was performed that was not indicated. A total of 300 charts were reviewed for the treatment group and 308 for the control group. The inter-rater agreement between the principal investigator and 2 other independent reviewers ranged between 93% and 98%. The median kappa statistic among 16 screening recommendations revealed very good agreement at 0.81 (all P <.042). Kappa could not be calculated for 4 of the screening recommendations, because the observed agreement of the 3 raters (100%) was equal to the expected agreement for those recommendations. All physicians were given a pretest measuring knowledge of the USPSTF recommendations at the beginning of the pre-intervention phase and a posttest near the end of the post-intervention phase. Finally, as physician learning and behavior were the focus of this study, patient adherence was not measured.

Data Analysis

Recommendations were organized into 4 categories: history, examination and laboratory, counseling, and prophylaxis. Tests of significance were calculated using a chi-square test that is designed to compare proportions among many independent samples.²¹ To control for statistical error, we used a Bonferroni adjusted P value. Because 20 screening recommendations were analyzed, a P value of .0025 (.05/20) was used as the cutoff for statistical significance.

Scores from the knowledge test were analyzed using a mixed factorial analysis of variance. The between-groups factor was group (treatment or control) and the within-subjects factor was test period (pretest or posttest). This analysis was conducted to see if the tool would lead to a group-by-test period interaction. Given the repeated use of the CART, it was thought that the treatment group might show a greater improvement in test scores than the control group.

Results

Impact of the CART

The Table 1 shows the percentages of patients who were appropriately screened during each period for the treatment and control groups. Significant increases were observed for all recommendations in the treatment group, except for occult blood and tetanus and diptheria booster. The mean absolute increase in the percentage of appropriately screened patients was 45%, 21%, and 15% for the recommendations in the history, examination and laboratory, and counseling categories, respectively. Removal of the annotated reminder during the post-intervention period brought the percentages of appropriate screenings down to baseline levels. Increases during the intervention phase were statistically significant for 16 of the 20 recommendations.

In the control group, the percentages of appropriately screened patients remained relatively stable. Although some variation is visible across the 3 study periods, there are no consistent trends among the recommendations, and none of the chi-square tests reached the Bonferroni level of significance. The Figure 1 shows the median percentage of patients screened during the baseline, intervention, and post-intervention periods. The average percentage of appropriate screenings for the clinicians using the CART increased during the intervention period and then returned to baseline levels after the tool was removed from use. The inverted V pattern repeatedly observed for the treatment group is not seen with the control group.

Screening Knowledge Scores

Only the main effect for the test period reached statistical significance. The mean test scores were 54.66 and 62.35 for clinicians at the pretest and posttest periods, respectively (F [1,27] = 23.89; P <.0005). The main effect for group was not statistically significant. The mean test scores for the treatment and control groups were 58.59 and 58.13, respectively (F [1,27] = 0.27; P <.608). The group-by-test period interaction also was not statistically significant (F [1,27] = 1.42, P=.244). Thus, although physicians had higher posttest knowledge scores than pretest scores, this increase was similar for treatment and control groups.

Discussion

CART use increased performance both in degree (percentage change) and in breadth (number of preventive services ordered appropriately). The primary outcome was the appropriateness of physician performance of preventive interventions. For example, if the patient is low-risk, then gonorrhea/chlamydia screening should not be performed, and appropriate physician behavior is to do no screening test. We did not give lectures or provide any other didactic preventive medicine training during the CART study period. But knowledge increased to a similar extent in both groups over that academic year. Nevertheless, the behavioral change seen in the treatment group was temporary; it returned to baseline in the post-intervention period. Therefore, it was the presence of the CART that improved physician behavior. The CART is comprehensive: It has prompts for up to 49 preventive recommendations. Physicians adhered significantly better to history and examination and laboratory recommendations when using the CART. The improvement in counseling was less impressive; physician behavior improved on only 4 of 6 counseling recommendations. It is likely that “form fatigue” was a factor, because the counseling recommendations were at the bottom of a page full of preventive recommendations. The difficulty of counseling, time constraints, and physician perception of the relative ineffectiveness of counseling may have also had an effect.

We believe that the primary reason for the success of the CART is immediate visual access to essential information. Computer reminders promote some behavioral change, but in other studies it was necessary to go to another computer screen for further information about a given test. Time pressure in the busy office makes this difficult. The CART gives physicians succinct information in a box adjacent to each preventive recommendation Table 1, including the recommended frequency, strength of recommendation, and appropriate groups to be screened for each intervention. Most other studies did not answer the question of whether increased knowledge over time was responsible for improved provider behavior. By removing the CART and reassessing physician behavior, we demonstrated that it was presence of the information in the CART format that was responsible for improving physician behavior.

Limitations

Our study has several limitations. We did not address the important issue of reminder systems or checklists. We only assessed the preventive services recommended by physicians during a single periodic health examination. We did not solve the problem of time constraint. It is difficult to cover every single preventive service recommendation in one visit. Our residents are less experienced and have more visit time than the average primary care provider. Finally, we only studied physician knowledge and behavior. We did not measure other important outcomes, such as patient satisfaction, adherence, morbidity, or mortality.

Conclusions

The nonspecific designation “tool” in the CART acronym implies potential future application of this concept for different subject matter and different formats. As we move inexorably toward universal electronic medical records, the CART can be adapted for electronic use. But whether on desktops, laptops, or hand-held computers, immediate visual access to information is essential. Combining the CART with computer reminder systems could yield significant improvement in patient outcomes when followed over time. Also, it would be interesting to incorporate the CART information into a computer-gathered patient history to create a partially completed history and physical examination form for provider use during the periodic health examination. This would save valuable office visit time. Many previous studies have focused on a small number of preventive services. We chose a comprehensive set of recommendations to more closely mimic the experience of primary care physicians who must cover many recommendations with patients. It is important that more studies use this approach. To maximize impact on the public health in an era of cost containment and to minimize physician error, we must enable physicians to efficiently apply a comprehensive set of preventive services in an appropriately focused manner. The CART is an effective step in the right direction.

Acknowledgments

We thank the Advocate Christ Hospital Med Fund for grant support. We would also like to thank Charles Warnell, MD, for his chart review work, and Carolyn Barsano for the literature review.

Frame and Carlson¹ published a critical review of periodic health screenings in The Journal of Family Practice in 1975. Such efforts, currently manifested in the United States Preventive Services Task Force (USPSTF), have had a profound impact on the practice of family physicians. Increased oversight of family physicians’ preventive practices by third parties and the increased medical sophistication of the public have fueled demand for preventive health services. Family physicians are experiencing increased accountability to their patients and communities for the effective delivery of preventive medical care.^2-4

A number of systems have been developed to assist physicians with the implementation of evidence-based preventive services guidelines. Over the past 15 years, numerous researchers have studied the effects of interventions using flow charts,^5-7 checklists,⁸ computer-based reminders,^9-14 and educational programs^15-19 for delivery of preventive services. Ornstein and colleagues¹⁰ found that the use of a computer-based tracking and reminder system for preventive services, together with a physician education program, modestly increased the frequency of counseling and screening services for patients. Similarly, Cheney and Ramsdell⁸ found that physicians who received an age- and sex-specific checklist in their patients’ medical charts performed significantly more age- and sex-appropriate preventive services than control physicians. But physician compliance with screening tools is often a problem.⁶ In the Cheney and Ramsdell study, only 39% of physicians used the available checklist to screen patients. Another study¹⁴ found a greater percentage of appropriate health screens by physicians using computerized reminders. However, many physicians did not make use of the available resource. Some studies found only minimal increases in screening rates with interventions such as educational seminars.¹⁹ One study⁵ used a simple flow sheet to analyze appropriate health service needs and noted that their results suggest that a flow sheet, although helpful, is not enough. Indeed, accomplishing sustained improvement in preventive services clinical outcomes is a complex process.²⁰

We investigated the effect of combining a complete set of annotated recommendations with a history and physical examination form on the rate of appropriate application of preventive service recommendations. We suspected that busy practitioners would need preventive services guidelines available at the point of care to be able to adhere closely, so we created a form that is very user-friendly. We also studied the effect of removing the screening tool after the intervention, to evaluate whether the behavioral effect was due to education or to the presence of the chart tool, a question that has not been resolved in several previous studies.

Methods

Study Setting and Instrument

Our study was conducted between June 1997 and June 1998 at a community hospital continuity clinic. Age-appropriate screening history questions, age-specific reminders, and test frequency recommendations drawn from the 1996 USPSTF were integrated into a form for documenting history and physical examinations. This Comprehensive Annotated Reminder Tool (CART) contained 8 sections: history, screening history, physical examination, screening mnemonics, laboratory screens, prophylaxis, counseling, and the assessment/plan section. CART forms were developed for patients aged 19-39, 40-64, and 65 or older. Table w1,Table w2, Table w3 3* Each of these 3-page forms contained guidelines for up to 49 preventive interventions. Twenty recommendations were evaluated in this study. Because nursing has been primarily responsible for ensuring immunization compliance in our setting, we did not expect an effect and only evaluated one immunization.

Intervention

Resident physicians were randomly assigned to a treatment group that was exposed to the CART (n=15) and a control group that used existing blank history and physical examination forms (n=16). Resident physicians from both groups precepted patients with the same faculty physicians on a regular basis. Physicians in the control group used a different area of the clinic facility from the treatment group to avoid cross-use of the CART form. The 3 time periods of the study were: pre-intervention phase (3 months), intervention phase (6 months), and post-intervention phase (3 months). The pre-intervention phase allowed for the establishment of a baseline measurement of preventive medicine practiced by physicians in both groups. In the intervention phase, the CART forms were placed on all charts of new patients seen by physicians in the experimental group. The forms were not introduced by any formal instruction. No preventive medicine lectures were given to physicians in either group during the course of the study. Finally, in the post-intervention phase the CART was no longer placed on charts.

Measures

Blinded chart reviews were performed by the principal investigator and 2 other independent reviewers on randomly selected new patients. The physicians were expected to complete a history and physical examination on each new clinic patient within the first 3 visits. Chart reviews occurred before use, during use, and after removal of the CART. They evaluated charts for screening appropriateness (not the absolute frequency of interventions). No credit was given when a preventive intervention was performed that was not indicated. A total of 300 charts were reviewed for the treatment group and 308 for the control group. The inter-rater agreement between the principal investigator and 2 other independent reviewers ranged between 93% and 98%. The median kappa statistic among 16 screening recommendations revealed very good agreement at 0.81 (all P <.042). Kappa could not be calculated for 4 of the screening recommendations, because the observed agreement of the 3 raters (100%) was equal to the expected agreement for those recommendations. All physicians were given a pretest measuring knowledge of the USPSTF recommendations at the beginning of the pre-intervention phase and a posttest near the end of the post-intervention phase. Finally, as physician learning and behavior were the focus of this study, patient adherence was not measured.

Data Analysis

Recommendations were organized into 4 categories: history, examination and laboratory, counseling, and prophylaxis. Tests of significance were calculated using a chi-square test that is designed to compare proportions among many independent samples.²¹ To control for statistical error, we used a Bonferroni adjusted P value. Because 20 screening recommendations were analyzed, a P value of .0025 (.05/20) was used as the cutoff for statistical significance.

Scores from the knowledge test were analyzed using a mixed factorial analysis of variance. The between-groups factor was group (treatment or control) and the within-subjects factor was test period (pretest or posttest). This analysis was conducted to see if the tool would lead to a group-by-test period interaction. Given the repeated use of the CART, it was thought that the treatment group might show a greater improvement in test scores than the control group.

Results

Impact of the CART

The Table 1 shows the percentages of patients who were appropriately screened during each period for the treatment and control groups. Significant increases were observed for all recommendations in the treatment group, except for occult blood and tetanus and diptheria booster. The mean absolute increase in the percentage of appropriately screened patients was 45%, 21%, and 15% for the recommendations in the history, examination and laboratory, and counseling categories, respectively. Removal of the annotated reminder during the post-intervention period brought the percentages of appropriate screenings down to baseline levels. Increases during the intervention phase were statistically significant for 16 of the 20 recommendations.

In the control group, the percentages of appropriately screened patients remained relatively stable. Although some variation is visible across the 3 study periods, there are no consistent trends among the recommendations, and none of the chi-square tests reached the Bonferroni level of significance. The Figure 1 shows the median percentage of patients screened during the baseline, intervention, and post-intervention periods. The average percentage of appropriate screenings for the clinicians using the CART increased during the intervention period and then returned to baseline levels after the tool was removed from use. The inverted V pattern repeatedly observed for the treatment group is not seen with the control group.

Screening Knowledge Scores

Only the main effect for the test period reached statistical significance. The mean test scores were 54.66 and 62.35 for clinicians at the pretest and posttest periods, respectively (F [1,27] = 23.89; P <.0005). The main effect for group was not statistically significant. The mean test scores for the treatment and control groups were 58.59 and 58.13, respectively (F [1,27] = 0.27; P <.608). The group-by-test period interaction also was not statistically significant (F [1,27] = 1.42, P=.244). Thus, although physicians had higher posttest knowledge scores than pretest scores, this increase was similar for treatment and control groups.

Discussion

CART use increased performance both in degree (percentage change) and in breadth (number of preventive services ordered appropriately). The primary outcome was the appropriateness of physician performance of preventive interventions. For example, if the patient is low-risk, then gonorrhea/chlamydia screening should not be performed, and appropriate physician behavior is to do no screening test. We did not give lectures or provide any other didactic preventive medicine training during the CART study period. But knowledge increased to a similar extent in both groups over that academic year. Nevertheless, the behavioral change seen in the treatment group was temporary; it returned to baseline in the post-intervention period. Therefore, it was the presence of the CART that improved physician behavior. The CART is comprehensive: It has prompts for up to 49 preventive recommendations. Physicians adhered significantly better to history and examination and laboratory recommendations when using the CART. The improvement in counseling was less impressive; physician behavior improved on only 4 of 6 counseling recommendations. It is likely that “form fatigue” was a factor, because the counseling recommendations were at the bottom of a page full of preventive recommendations. The difficulty of counseling, time constraints, and physician perception of the relative ineffectiveness of counseling may have also had an effect.

We believe that the primary reason for the success of the CART is immediate visual access to essential information. Computer reminders promote some behavioral change, but in other studies it was necessary to go to another computer screen for further information about a given test. Time pressure in the busy office makes this difficult. The CART gives physicians succinct information in a box adjacent to each preventive recommendation Table 1, including the recommended frequency, strength of recommendation, and appropriate groups to be screened for each intervention. Most other studies did not answer the question of whether increased knowledge over time was responsible for improved provider behavior. By removing the CART and reassessing physician behavior, we demonstrated that it was presence of the information in the CART format that was responsible for improving physician behavior.

Limitations

Our study has several limitations. We did not address the important issue of reminder systems or checklists. We only assessed the preventive services recommended by physicians during a single periodic health examination. We did not solve the problem of time constraint. It is difficult to cover every single preventive service recommendation in one visit. Our residents are less experienced and have more visit time than the average primary care provider. Finally, we only studied physician knowledge and behavior. We did not measure other important outcomes, such as patient satisfaction, adherence, morbidity, or mortality.

Conclusions

The nonspecific designation “tool” in the CART acronym implies potential future application of this concept for different subject matter and different formats. As we move inexorably toward universal electronic medical records, the CART can be adapted for electronic use. But whether on desktops, laptops, or hand-held computers, immediate visual access to information is essential. Combining the CART with computer reminder systems could yield significant improvement in patient outcomes when followed over time. Also, it would be interesting to incorporate the CART information into a computer-gathered patient history to create a partially completed history and physical examination form for provider use during the periodic health examination. This would save valuable office visit time. Many previous studies have focused on a small number of preventive services. We chose a comprehensive set of recommendations to more closely mimic the experience of primary care physicians who must cover many recommendations with patients. It is important that more studies use this approach. To maximize impact on the public health in an era of cost containment and to minimize physician error, we must enable physicians to efficiently apply a comprehensive set of preventive services in an appropriately focused manner. The CART is an effective step in the right direction.

Acknowledgments

We thank the Advocate Christ Hospital Med Fund for grant support. We would also like to thank Charles Warnell, MD, for his chart review work, and Carolyn Barsano for the literature review.

Patients can experience great difficulties in navigating the US medical system. They are faced with complicated decisions in a system that is often fragmented, episodic, and disease oriented.¹ As highlighted by the recent Institute of Medicine report,² the system’s complexity contributes to medical errors that harm patients. The patient with a physician family member, however, has a unique advocate for their health care.³ The physician family member has intimate knowledge of the patient, as well as an expert understanding of the system.⁴ Although previous studies have documented the conflicting roles of physician family members, we used the perceptions of these informed observers to illuminate the experiences of patients in the current system.⁵

Using a key informant interviewing technique,⁶ we solicited the chairpersons of academic departments of family medicine for their personal experiences with the health care system on the occasion when their own parents were seriously ill and required medical care. These family physicians were experts in coordination of care, continuity of care, and navigating the health care system. They were uniquely positioned to comment on the process and quality of care that their fathers received.^4,7

Our sample is unique, and the experiences of these physicians are not directly generalizable to the population at large. These physician family members, however, offered a special opportunity to observe the performance of the health care system on a personal level. We hoped that their insights would illuminate the challenges facing patients in our health care system and point to strategies that could improve care.

Methods

Using E-mail, we solicited responses from the chairs of every academic allopathic family medicine department in the United States. E-mail addresses were obtained from the national listserve of the Association of Departments of Family Medicine. The respondents were eligible to participate if either of their parents had experienced a serious or terminal illness episode within the past 5 years. Since this was a key informant analysis, we purposely sought and were satisfied with a sample of volunteers and did not pursue methods of increasing the response rate. All physicians provided verbal and written consent to participate in our study. The study was approved by the University of Washington Human Subjects Review Committee. Particular effort was taken to ensure the confidentiality of the physicians. Personal identifiers were removed from the transcripts, and the authors have been cautious to avoid reporting identifiable details of the individual cases.

One of the authors (F.M.C.) conducted in-depth, semi-structured interviews with each of the study physicians using a field-tested interview template.⁸ The instrument consisted of open-ended questions and focused on the physicians’ responses to their fathers’ care Table w1.* The interviews began with the physicians’ narratives of their fathers’ illnesses. All interviews were conducted by telephone and were audiotaped and transcribed. The initial interviews lasted 45 minutes to 1 hour.

Two of the authors (F.M.C., L.A.G.) read, coded, and labeled the transcripts for themes, using an open-coding technique. Using an iterative analysis, themes were expanded and refined during rereading of the transcripts by all 3 authors.⁹ After all themes were identified, the study physicians were re-interviewed. The second interviews ensured the reliability of the initial interviews but also served to validate and clarify themes that had emerged during the analysis.^10,11 The second interviews, lasting 15 to 20 minutes, were also audiotaped, transcribed, and analyzed.

Eight family physicians were willing to participate, met the eligibility criteria, and consented to be interviewed. Geographically, the physicians were widely distributed. The mean age of the study participants was 47.5 years (range = 43 - 54 years). Six of the physicians were men. The physicians had been in practice for an average of 19.4 years. Although they had been solicited for the illness episodes of either parent, all participants related experiences with their fathers. Of the 8 physicians’ fathers,⁵ died during or shortly after the reported illness episode.

Results

All of the physicians witnessed and reported challenges in the medical care of their fathers. Although the details of each story were quite different, there were common themes that emerged from all of the narratives Table 1. The physicians described their fathers’ need for an advocate, being compelled to act on behalf of their fathers, and an abiding inner discord about the care their fathers received.

Need for an Advocate

All of the physicians described the importance of an advocate for their fathers, someone who could navigate and coordinate his medical care.

“I think the system is so complex, that what families need are guides, people who understand the system and who can work with the individual and the family and then translate that into getting the system to work the way it needs to.” (Physician A)

As a patient’s medical care becomes increasingly complex, the advocate becomes more important and, ironically, more elusive. In many cases the responsibility for being the advocate fell to the physician family member.

“I felt that somehow I had to get in there and protect my dad, protect my family, and advocate for them. Knowing that everything that had been done was going wrong, it was hard. So the system really didn’t give me someone who I could talk to, who would understand me, understand our family, and understand the issues.” (Physician B)

“My brother’s statement was, “I don’t see how any family can go through something like this, if there’s not a family member that’s a physician.” (Physician C)

Compelled to Rescue

The physician family members expressed reluctance about taking the responsibility of being their fathers’ medical advocate. Many had strong feelings that they should not be involved with their fathers’ care.

“I knew I could not view the situation objectively. I really tried to walk that line of being just a concerned family member—but when things are so blatantly obvious, I finally couldn’t stay in the bushes anymore. I had to come out. You know, what good is all that training if you can’t help your own family?” (Physician D)

As a result of the obstacles to the optimal care of their fathers, the physicians found themselves taking an active role in their fathers’ care. There were many different challenges that compelled the physicians to intervene on behalf of their fathers. For example, they described a lack of responsiveness by providers, poor communication and confusion, a loss of continuity of care, and medical mistakes. Poor relationships between physicians and the patient were present in many of the narratives. In one case, the physician’s father suffered an acute myocardial infarction and was taken to the emergency department.

“And I said, ‘I saw the EKG. It looks like he’s had an MI. He’s been here for almost an hour. He’s not had aspirin or nitroglycerin yet.’ And the ER doc said, ‘Well, that’s what I’m in here to do.’ And I said, ‘Well then don’t spend your time trying to figure out if it’s indigestion or chest pain.’ So I’m sure I got labeled, but I could have been a layperson and known you were supposed to get aspirin and nitroglycerin. I couldn’t believe that they were taking all that time to do that.” (Physician E)

Another physician’s father, who had been sedated, awoke to find a personal keepsake missing.

“The most upsetting thing about the whole thing was [that] they took a very adversarial stance and started blaming us. ‘Well, he’s demented, so he must have thrown it away,’ or ‘In these cases it’s usually a family member who’s taken it.’ And at that point he was just devastated, and he lost hope. He said, ‘Now my cross is gone, and I’m going to die.’ (Physician F)

These physicians had a position of power and control in the system, and they were able to affect the course of medical care of their fathers. One physician’s father was hospitalized for treatment of a pericardial hemorrhage following aortic valve replacement.

“The post-op course became fairly stormy with pulmonary congestion, poor blood gases. A variety of specialties were consulted, and there were plans for a thoracentesis—this, that, and the other thing going on. Finally I couldn’t stand things much longer and I wrote 2 pages of orders. I essentially discharged him from the hospital and got him back involved with his family physician. Nobody was looking at the whole picture, and it was clear to me that I had to get him out of there.” (Physician D)

Interviewer: “Did you ever feel conflicted about intervening in that role?”

“Oh, it was terrible! On the other hand, you know at the end I was so pushed that I really felt that if I didn’t do something that he would die of iatrogenesis.” (Physician D)

Abiding Inner Discord

There was a strong sense of discord about the performance of the health care system. Invariably, these physicians observed deficiencies in the care of their fathers, and they recognized that the system was not performing or responding in the way it should.

“[You are] frustrated by the fact that you know that they can do a lot better when they really want to. I have 20 years at this place, and damn it, the least they can do is treat my father well. If this is the best they can do, what does that say about the average treatment that the average patient gets?” (Physician F)

The physicians expressed concern for the experiences of other patients. For the most part, they felt that their fathers benefited from their personal involvement. That option, however, is unavailable to most patients.

“It scares the hell out of me because what I have come to conclude is that the system’s working well for my father is the exception, rather than the rule. In the other family members that I have been involved with—my grandmother, my uncle, and most recently my mom—my level of involvement was more than it would have been otherwise, because the system was functioning so poorly. It is very scary to me, because 99% of the people accessing the systems don’t have anybody advocating for them that way.” (Physician G)

These physicians were not just malcontents within the health care system. They carried a deep ambivalence about their views of health care. They struggled to reconcile their professional pride with the imperfections in their fathers’ care. Some of the physicians questioned their own involvement in the profession and system that can produce such incongruities.

“I just feel kind of helpless in the face of what [patients] actually run into. You know, when they come back and say, ‘I had a terrible experience.’ I feel responsible for that. I feel embarrassed to be a part of a profession that doesn’t see that as something that’s important.” (Physician B)

“I think he actually got better treatment because of his family medical web of connections than most people would get. And I have some sense of discord about that. We spend so much, we have so many doctors, why is it so hard to make the system work?” (Physician A)

The persistence of their feelings of inner discord was remarkable. Months after the initial interviews and sometimes years after their fathers’ illness episodes, all the physicians expressed some degree of ongoing personal trauma, sometimes silently harboring painful doubts.

“Yeah, there’s some lingering doubt that I continue to think about. And I keep thinking that it would be harder on my mother if I pursued it than it is to leave it alone. If it were just me, I might actually go to a lawyer and ask them to request the records. I still think I wouldn’t know what I wish to know but it might give me some sense of closure on it.” (Physician E)

“What’s interesting is that what you’re looking for are larger issues and themes, but at a microlevel the value of doing this is an opportunity to at least tell the story one time. Because you don’t tell it to people who were there, and you don’t tell it to people who don’t ask you. So it remains something inside of you that is hard.” (Physician B)

Discussion

Even patients with intimate and knowledgeable advocates face challenges in receiving optimal medical care. The physicians in this study, in the unique position of being senior family physicians and concerned family members, felt strongly that patients need an informed, accountable advocate; each witnessed events and situations where such an advocate was absent when needed. Although they were initially reluctant to be involved in their fathers’ care, obstacles to optimal care compelled many of the physicians to intervene. As a result of these experiences, the physicians shared an ongoing inner discord about the performance of the health care system for all patients.

During the initial interviews, many physicians told their narratives as experiences with specific individuals and did not describe their observations as system-level issues. During the analysis, however, it became clear that the common themes reflected system-level characteristics. When the physicians were re-interviewed, there was universal agreement that the structure of the health care system contributes to poor communication by individual care providers and to medical errors. This observation focuses on the challenge of changing individual provider behavior without addressing the system within which the provider works.

We feel that our sample of 8 physicians was sufficient for a key informant analysis. The pool of potential informants was limited by their unique position and the requirement of having a parent with a recent serious illness episode. This physician sample was deliberately and purposefully selected. The stature of the respondents created a potential bias, a “VIP syndrome” for these physicians’ fathers.³ Rather than receiving excessive care, however, some of these patients received suboptimal or even antagonistic care.

As sons and daughters, it is possible that these physicians may be embittered about their fathers’ care, leading them to exaggerate or overstate their observations. The illness of a parent evokes intense emotions, but it also tends to rivet attention to the care being received.⁷ Because our informants also acted as participant-observers, it is possible that their observations lacked insight into the harm they may have caused by intervening in their fathers’ care. Despite these factors, the themes of advocacy and rescue were common to all of the physicians. The theme of abiding inner discord was strengthened by its enduring nature over time. By confirming these themes through re-interviews, we are confident that they are robust and valid for each individual as well as for the entire sample. Although these physicians’ reports of their experiences should not be considered generalizable to the population at large, they are informed expert opinions that raise serious concerns about how well the health care system is serving patients.

Our results are consistent with the burgeoning demand for improvement in our current health care system. Health care systems could affirm the continued presence of one physician who is in charge of the patient’s care and accountable to the patient and the patient’s family. Payment systems and health plan rules should not force discontinuity across different care settings. Physicians who have a relationship and previous experience with patients should be encouraged to remain involved in their care during hospitalizations. Health care begs to be rebalanced to emphasize the importance of knowing the patient at least as well as the disease process and medical technology.

We also found that these physicians’ experiences had a profound personal impact. The study physicians expressed a sense of being silenced by the system and were grateful for the relief afforded by telling their stories. This suggests that physicians and the systems they work in should create mechanisms for the discussion of troubling patient care events.

The personal experiences of these physicians hold special importance to other physicians, because they highlight the critical roles physicians are expected to play in a superior health care system. Many of the problems identified by these senior family physicians were manifest in physicians’ behaviors. Physicians should be able to express their ambivalence about problematic health care processes and encourage an environment that avoids blaming and promotes improvements. Rather than waiting for system-level change to improve health care, physicians could examine and change their own behaviors and practices.

Acknowledgments

Our paper is dedicated to Mary Lou Green, whose care at the end of her life inspired this study. The authors are indebted to Priscilla Noland and Michelle Perez for their assistance with the manuscript.

Patients can experience great difficulties in navigating the US medical system. They are faced with complicated decisions in a system that is often fragmented, episodic, and disease oriented.¹ As highlighted by the recent Institute of Medicine report,² the system’s complexity contributes to medical errors that harm patients. The patient with a physician family member, however, has a unique advocate for their health care.³ The physician family member has intimate knowledge of the patient, as well as an expert understanding of the system.⁴ Although previous studies have documented the conflicting roles of physician family members, we used the perceptions of these informed observers to illuminate the experiences of patients in the current system.⁵

Using a key informant interviewing technique,⁶ we solicited the chairpersons of academic departments of family medicine for their personal experiences with the health care system on the occasion when their own parents were seriously ill and required medical care. These family physicians were experts in coordination of care, continuity of care, and navigating the health care system. They were uniquely positioned to comment on the process and quality of care that their fathers received.^4,7

Our sample is unique, and the experiences of these physicians are not directly generalizable to the population at large. These physician family members, however, offered a special opportunity to observe the performance of the health care system on a personal level. We hoped that their insights would illuminate the challenges facing patients in our health care system and point to strategies that could improve care.

Methods

Using E-mail, we solicited responses from the chairs of every academic allopathic family medicine department in the United States. E-mail addresses were obtained from the national listserve of the Association of Departments of Family Medicine. The respondents were eligible to participate if either of their parents had experienced a serious or terminal illness episode within the past 5 years. Since this was a key informant analysis, we purposely sought and were satisfied with a sample of volunteers and did not pursue methods of increasing the response rate. All physicians provided verbal and written consent to participate in our study. The study was approved by the University of Washington Human Subjects Review Committee. Particular effort was taken to ensure the confidentiality of the physicians. Personal identifiers were removed from the transcripts, and the authors have been cautious to avoid reporting identifiable details of the individual cases.

One of the authors (F.M.C.) conducted in-depth, semi-structured interviews with each of the study physicians using a field-tested interview template.⁸ The instrument consisted of open-ended questions and focused on the physicians’ responses to their fathers’ care Table w1.* The interviews began with the physicians’ narratives of their fathers’ illnesses. All interviews were conducted by telephone and were audiotaped and transcribed. The initial interviews lasted 45 minutes to 1 hour.

Two of the authors (F.M.C., L.A.G.) read, coded, and labeled the transcripts for themes, using an open-coding technique. Using an iterative analysis, themes were expanded and refined during rereading of the transcripts by all 3 authors.⁹ After all themes were identified, the study physicians were re-interviewed. The second interviews ensured the reliability of the initial interviews but also served to validate and clarify themes that had emerged during the analysis.^10,11 The second interviews, lasting 15 to 20 minutes, were also audiotaped, transcribed, and analyzed.

Eight family physicians were willing to participate, met the eligibility criteria, and consented to be interviewed. Geographically, the physicians were widely distributed. The mean age of the study participants was 47.5 years (range = 43 - 54 years). Six of the physicians were men. The physicians had been in practice for an average of 19.4 years. Although they had been solicited for the illness episodes of either parent, all participants related experiences with their fathers. Of the 8 physicians’ fathers,⁵ died during or shortly after the reported illness episode.

Results

All of the physicians witnessed and reported challenges in the medical care of their fathers. Although the details of each story were quite different, there were common themes that emerged from all of the narratives Table 1. The physicians described their fathers’ need for an advocate, being compelled to act on behalf of their fathers, and an abiding inner discord about the care their fathers received.

Need for an Advocate

All of the physicians described the importance of an advocate for their fathers, someone who could navigate and coordinate his medical care.

“I think the system is so complex, that what families need are guides, people who understand the system and who can work with the individual and the family and then translate that into getting the system to work the way it needs to.” (Physician A)

As a patient’s medical care becomes increasingly complex, the advocate becomes more important and, ironically, more elusive. In many cases the responsibility for being the advocate fell to the physician family member.

“I felt that somehow I had to get in there and protect my dad, protect my family, and advocate for them. Knowing that everything that had been done was going wrong, it was hard. So the system really didn’t give me someone who I could talk to, who would understand me, understand our family, and understand the issues.” (Physician B)

“My brother’s statement was, “I don’t see how any family can go through something like this, if there’s not a family member that’s a physician.” (Physician C)

Compelled to Rescue

The physician family members expressed reluctance about taking the responsibility of being their fathers’ medical advocate. Many had strong feelings that they should not be involved with their fathers’ care.

“I knew I could not view the situation objectively. I really tried to walk that line of being just a concerned family member—but when things are so blatantly obvious, I finally couldn’t stay in the bushes anymore. I had to come out. You know, what good is all that training if you can’t help your own family?” (Physician D)

As a result of the obstacles to the optimal care of their fathers, the physicians found themselves taking an active role in their fathers’ care. There were many different challenges that compelled the physicians to intervene on behalf of their fathers. For example, they described a lack of responsiveness by providers, poor communication and confusion, a loss of continuity of care, and medical mistakes. Poor relationships between physicians and the patient were present in many of the narratives. In one case, the physician’s father suffered an acute myocardial infarction and was taken to the emergency department.

“And I said, ‘I saw the EKG. It looks like he’s had an MI. He’s been here for almost an hour. He’s not had aspirin or nitroglycerin yet.’ And the ER doc said, ‘Well, that’s what I’m in here to do.’ And I said, ‘Well then don’t spend your time trying to figure out if it’s indigestion or chest pain.’ So I’m sure I got labeled, but I could have been a layperson and known you were supposed to get aspirin and nitroglycerin. I couldn’t believe that they were taking all that time to do that.” (Physician E)

Another physician’s father, who had been sedated, awoke to find a personal keepsake missing.

“The most upsetting thing about the whole thing was [that] they took a very adversarial stance and started blaming us. ‘Well, he’s demented, so he must have thrown it away,’ or ‘In these cases it’s usually a family member who’s taken it.’ And at that point he was just devastated, and he lost hope. He said, ‘Now my cross is gone, and I’m going to die.’ (Physician F)

These physicians had a position of power and control in the system, and they were able to affect the course of medical care of their fathers. One physician’s father was hospitalized for treatment of a pericardial hemorrhage following aortic valve replacement.

“The post-op course became fairly stormy with pulmonary congestion, poor blood gases. A variety of specialties were consulted, and there were plans for a thoracentesis—this, that, and the other thing going on. Finally I couldn’t stand things much longer and I wrote 2 pages of orders. I essentially discharged him from the hospital and got him back involved with his family physician. Nobody was looking at the whole picture, and it was clear to me that I had to get him out of there.” (Physician D)

Interviewer: “Did you ever feel conflicted about intervening in that role?”

“Oh, it was terrible! On the other hand, you know at the end I was so pushed that I really felt that if I didn’t do something that he would die of iatrogenesis.” (Physician D)

Abiding Inner Discord

There was a strong sense of discord about the performance of the health care system. Invariably, these physicians observed deficiencies in the care of their fathers, and they recognized that the system was not performing or responding in the way it should.

“[You are] frustrated by the fact that you know that they can do a lot better when they really want to. I have 20 years at this place, and damn it, the least they can do is treat my father well. If this is the best they can do, what does that say about the average treatment that the average patient gets?” (Physician F)

The physicians expressed concern for the experiences of other patients. For the most part, they felt that their fathers benefited from their personal involvement. That option, however, is unavailable to most patients.

“It scares the hell out of me because what I have come to conclude is that the system’s working well for my father is the exception, rather than the rule. In the other family members that I have been involved with—my grandmother, my uncle, and most recently my mom—my level of involvement was more than it would have been otherwise, because the system was functioning so poorly. It is very scary to me, because 99% of the people accessing the systems don’t have anybody advocating for them that way.” (Physician G)

These physicians were not just malcontents within the health care system. They carried a deep ambivalence about their views of health care. They struggled to reconcile their professional pride with the imperfections in their fathers’ care. Some of the physicians questioned their own involvement in the profession and system that can produce such incongruities.

“I just feel kind of helpless in the face of what [patients] actually run into. You know, when they come back and say, ‘I had a terrible experience.’ I feel responsible for that. I feel embarrassed to be a part of a profession that doesn’t see that as something that’s important.” (Physician B)

“I think he actually got better treatment because of his family medical web of connections than most people would get. And I have some sense of discord about that. We spend so much, we have so many doctors, why is it so hard to make the system work?” (Physician A)

The persistence of their feelings of inner discord was remarkable. Months after the initial interviews and sometimes years after their fathers’ illness episodes, all the physicians expressed some degree of ongoing personal trauma, sometimes silently harboring painful doubts.

“Yeah, there’s some lingering doubt that I continue to think about. And I keep thinking that it would be harder on my mother if I pursued it than it is to leave it alone. If it were just me, I might actually go to a lawyer and ask them to request the records. I still think I wouldn’t know what I wish to know but it might give me some sense of closure on it.” (Physician E)

“What’s interesting is that what you’re looking for are larger issues and themes, but at a microlevel the value of doing this is an opportunity to at least tell the story one time. Because you don’t tell it to people who were there, and you don’t tell it to people who don’t ask you. So it remains something inside of you that is hard.” (Physician B)

Discussion

Even patients with intimate and knowledgeable advocates face challenges in receiving optimal medical care. The physicians in this study, in the unique position of being senior family physicians and concerned family members, felt strongly that patients need an informed, accountable advocate; each witnessed events and situations where such an advocate was absent when needed. Although they were initially reluctant to be involved in their fathers’ care, obstacles to optimal care compelled many of the physicians to intervene. As a result of these experiences, the physicians shared an ongoing inner discord about the performance of the health care system for all patients.

During the initial interviews, many physicians told their narratives as experiences with specific individuals and did not describe their observations as system-level issues. During the analysis, however, it became clear that the common themes reflected system-level characteristics. When the physicians were re-interviewed, there was universal agreement that the structure of the health care system contributes to poor communication by individual care providers and to medical errors. This observation focuses on the challenge of changing individual provider behavior without addressing the system within which the provider works.

We feel that our sample of 8 physicians was sufficient for a key informant analysis. The pool of potential informants was limited by their unique position and the requirement of having a parent with a recent serious illness episode. This physician sample was deliberately and purposefully selected. The stature of the respondents created a potential bias, a “VIP syndrome” for these physicians’ fathers.³ Rather than receiving excessive care, however, some of these patients received suboptimal or even antagonistic care.

As sons and daughters, it is possible that these physicians may be embittered about their fathers’ care, leading them to exaggerate or overstate their observations. The illness of a parent evokes intense emotions, but it also tends to rivet attention to the care being received.⁷ Because our informants also acted as participant-observers, it is possible that their observations lacked insight into the harm they may have caused by intervening in their fathers’ care. Despite these factors, the themes of advocacy and rescue were common to all of the physicians. The theme of abiding inner discord was strengthened by its enduring nature over time. By confirming these themes through re-interviews, we are confident that they are robust and valid for each individual as well as for the entire sample. Although these physicians’ reports of their experiences should not be considered generalizable to the population at large, they are informed expert opinions that raise serious concerns about how well the health care system is serving patients.

Our results are consistent with the burgeoning demand for improvement in our current health care system. Health care systems could affirm the continued presence of one physician who is in charge of the patient’s care and accountable to the patient and the patient’s family. Payment systems and health plan rules should not force discontinuity across different care settings. Physicians who have a relationship and previous experience with patients should be encouraged to remain involved in their care during hospitalizations. Health care begs to be rebalanced to emphasize the importance of knowing the patient at least as well as the disease process and medical technology.

We also found that these physicians’ experiences had a profound personal impact. The study physicians expressed a sense of being silenced by the system and were grateful for the relief afforded by telling their stories. This suggests that physicians and the systems they work in should create mechanisms for the discussion of troubling patient care events.

The personal experiences of these physicians hold special importance to other physicians, because they highlight the critical roles physicians are expected to play in a superior health care system. Many of the problems identified by these senior family physicians were manifest in physicians’ behaviors. Physicians should be able to express their ambivalence about problematic health care processes and encourage an environment that avoids blaming and promotes improvements. Rather than waiting for system-level change to improve health care, physicians could examine and change their own behaviors and practices.

Acknowledgments

Our paper is dedicated to Mary Lou Green, whose care at the end of her life inspired this study. The authors are indebted to Priscilla Noland and Michelle Perez for their assistance with the manuscript.

Patients can experience great difficulties in navigating the US medical system. They are faced with complicated decisions in a system that is often fragmented, episodic, and disease oriented.¹ As highlighted by the recent Institute of Medicine report,² the system’s complexity contributes to medical errors that harm patients. The patient with a physician family member, however, has a unique advocate for their health care.³ The physician family member has intimate knowledge of the patient, as well as an expert understanding of the system.⁴ Although previous studies have documented the conflicting roles of physician family members, we used the perceptions of these informed observers to illuminate the experiences of patients in the current system.⁵

Using a key informant interviewing technique,⁶ we solicited the chairpersons of academic departments of family medicine for their personal experiences with the health care system on the occasion when their own parents were seriously ill and required medical care. These family physicians were experts in coordination of care, continuity of care, and navigating the health care system. They were uniquely positioned to comment on the process and quality of care that their fathers received.^4,7

Our sample is unique, and the experiences of these physicians are not directly generalizable to the population at large. These physician family members, however, offered a special opportunity to observe the performance of the health care system on a personal level. We hoped that their insights would illuminate the challenges facing patients in our health care system and point to strategies that could improve care.

Methods

Using E-mail, we solicited responses from the chairs of every academic allopathic family medicine department in the United States. E-mail addresses were obtained from the national listserve of the Association of Departments of Family Medicine. The respondents were eligible to participate if either of their parents had experienced a serious or terminal illness episode within the past 5 years. Since this was a key informant analysis, we purposely sought and were satisfied with a sample of volunteers and did not pursue methods of increasing the response rate. All physicians provided verbal and written consent to participate in our study. The study was approved by the University of Washington Human Subjects Review Committee. Particular effort was taken to ensure the confidentiality of the physicians. Personal identifiers were removed from the transcripts, and the authors have been cautious to avoid reporting identifiable details of the individual cases.

One of the authors (F.M.C.) conducted in-depth, semi-structured interviews with each of the study physicians using a field-tested interview template.⁸ The instrument consisted of open-ended questions and focused on the physicians’ responses to their fathers’ care Table w1.* The interviews began with the physicians’ narratives of their fathers’ illnesses. All interviews were conducted by telephone and were audiotaped and transcribed. The initial interviews lasted 45 minutes to 1 hour.

Two of the authors (F.M.C., L.A.G.) read, coded, and labeled the transcripts for themes, using an open-coding technique. Using an iterative analysis, themes were expanded and refined during rereading of the transcripts by all 3 authors.⁹ After all themes were identified, the study physicians were re-interviewed. The second interviews ensured the reliability of the initial interviews but also served to validate and clarify themes that had emerged during the analysis.^10,11 The second interviews, lasting 15 to 20 minutes, were also audiotaped, transcribed, and analyzed.

Eight family physicians were willing to participate, met the eligibility criteria, and consented to be interviewed. Geographically, the physicians were widely distributed. The mean age of the study participants was 47.5 years (range = 43 - 54 years). Six of the physicians were men. The physicians had been in practice for an average of 19.4 years. Although they had been solicited for the illness episodes of either parent, all participants related experiences with their fathers. Of the 8 physicians’ fathers,⁵ died during or shortly after the reported illness episode.

Results

All of the physicians witnessed and reported challenges in the medical care of their fathers. Although the details of each story were quite different, there were common themes that emerged from all of the narratives Table 1. The physicians described their fathers’ need for an advocate, being compelled to act on behalf of their fathers, and an abiding inner discord about the care their fathers received.

Need for an Advocate

All of the physicians described the importance of an advocate for their fathers, someone who could navigate and coordinate his medical care.

“I think the system is so complex, that what families need are guides, people who understand the system and who can work with the individual and the family and then translate that into getting the system to work the way it needs to.” (Physician A)

As a patient’s medical care becomes increasingly complex, the advocate becomes more important and, ironically, more elusive. In many cases the responsibility for being the advocate fell to the physician family member.

“I felt that somehow I had to get in there and protect my dad, protect my family, and advocate for them. Knowing that everything that had been done was going wrong, it was hard. So the system really didn’t give me someone who I could talk to, who would understand me, understand our family, and understand the issues.” (Physician B)

“My brother’s statement was, “I don’t see how any family can go through something like this, if there’s not a family member that’s a physician.” (Physician C)

Compelled to Rescue

The physician family members expressed reluctance about taking the responsibility of being their fathers’ medical advocate. Many had strong feelings that they should not be involved with their fathers’ care.

“I knew I could not view the situation objectively. I really tried to walk that line of being just a concerned family member—but when things are so blatantly obvious, I finally couldn’t stay in the bushes anymore. I had to come out. You know, what good is all that training if you can’t help your own family?” (Physician D)

As a result of the obstacles to the optimal care of their fathers, the physicians found themselves taking an active role in their fathers’ care. There were many different challenges that compelled the physicians to intervene on behalf of their fathers. For example, they described a lack of responsiveness by providers, poor communication and confusion, a loss of continuity of care, and medical mistakes. Poor relationships between physicians and the patient were present in many of the narratives. In one case, the physician’s father suffered an acute myocardial infarction and was taken to the emergency department.

“And I said, ‘I saw the EKG. It looks like he’s had an MI. He’s been here for almost an hour. He’s not had aspirin or nitroglycerin yet.’ And the ER doc said, ‘Well, that’s what I’m in here to do.’ And I said, ‘Well then don’t spend your time trying to figure out if it’s indigestion or chest pain.’ So I’m sure I got labeled, but I could have been a layperson and known you were supposed to get aspirin and nitroglycerin. I couldn’t believe that they were taking all that time to do that.” (Physician E)

Another physician’s father, who had been sedated, awoke to find a personal keepsake missing.

“The most upsetting thing about the whole thing was [that] they took a very adversarial stance and started blaming us. ‘Well, he’s demented, so he must have thrown it away,’ or ‘In these cases it’s usually a family member who’s taken it.’ And at that point he was just devastated, and he lost hope. He said, ‘Now my cross is gone, and I’m going to die.’ (Physician F)

These physicians had a position of power and control in the system, and they were able to affect the course of medical care of their fathers. One physician’s father was hospitalized for treatment of a pericardial hemorrhage following aortic valve replacement.

“The post-op course became fairly stormy with pulmonary congestion, poor blood gases. A variety of specialties were consulted, and there were plans for a thoracentesis—this, that, and the other thing going on. Finally I couldn’t stand things much longer and I wrote 2 pages of orders. I essentially discharged him from the hospital and got him back involved with his family physician. Nobody was looking at the whole picture, and it was clear to me that I had to get him out of there.” (Physician D)

Interviewer: “Did you ever feel conflicted about intervening in that role?”

“Oh, it was terrible! On the other hand, you know at the end I was so pushed that I really felt that if I didn’t do something that he would die of iatrogenesis.” (Physician D)

Abiding Inner Discord

There was a strong sense of discord about the performance of the health care system. Invariably, these physicians observed deficiencies in the care of their fathers, and they recognized that the system was not performing or responding in the way it should.

“[You are] frustrated by the fact that you know that they can do a lot better when they really want to. I have 20 years at this place, and damn it, the least they can do is treat my father well. If this is the best they can do, what does that say about the average treatment that the average patient gets?” (Physician F)

The physicians expressed concern for the experiences of other patients. For the most part, they felt that their fathers benefited from their personal involvement. That option, however, is unavailable to most patients.

“It scares the hell out of me because what I have come to conclude is that the system’s working well for my father is the exception, rather than the rule. In the other family members that I have been involved with—my grandmother, my uncle, and most recently my mom—my level of involvement was more than it would have been otherwise, because the system was functioning so poorly. It is very scary to me, because 99% of the people accessing the systems don’t have anybody advocating for them that way.” (Physician G)

These physicians were not just malcontents within the health care system. They carried a deep ambivalence about their views of health care. They struggled to reconcile their professional pride with the imperfections in their fathers’ care. Some of the physicians questioned their own involvement in the profession and system that can produce such incongruities.

“I just feel kind of helpless in the face of what [patients] actually run into. You know, when they come back and say, ‘I had a terrible experience.’ I feel responsible for that. I feel embarrassed to be a part of a profession that doesn’t see that as something that’s important.” (Physician B)

“I think he actually got better treatment because of his family medical web of connections than most people would get. And I have some sense of discord about that. We spend so much, we have so many doctors, why is it so hard to make the system work?” (Physician A)

The persistence of their feelings of inner discord was remarkable. Months after the initial interviews and sometimes years after their fathers’ illness episodes, all the physicians expressed some degree of ongoing personal trauma, sometimes silently harboring painful doubts.

“Yeah, there’s some lingering doubt that I continue to think about. And I keep thinking that it would be harder on my mother if I pursued it than it is to leave it alone. If it were just me, I might actually go to a lawyer and ask them to request the records. I still think I wouldn’t know what I wish to know but it might give me some sense of closure on it.” (Physician E)

“What’s interesting is that what you’re looking for are larger issues and themes, but at a microlevel the value of doing this is an opportunity to at least tell the story one time. Because you don’t tell it to people who were there, and you don’t tell it to people who don’t ask you. So it remains something inside of you that is hard.” (Physician B)

Discussion

Even patients with intimate and knowledgeable advocates face challenges in receiving optimal medical care. The physicians in this study, in the unique position of being senior family physicians and concerned family members, felt strongly that patients need an informed, accountable advocate; each witnessed events and situations where such an advocate was absent when needed. Although they were initially reluctant to be involved in their fathers’ care, obstacles to optimal care compelled many of the physicians to intervene. As a result of these experiences, the physicians shared an ongoing inner discord about the performance of the health care system for all patients.

During the initial interviews, many physicians told their narratives as experiences with specific individuals and did not describe their observations as system-level issues. During the analysis, however, it became clear that the common themes reflected system-level characteristics. When the physicians were re-interviewed, there was universal agreement that the structure of the health care system contributes to poor communication by individual care providers and to medical errors. This observation focuses on the challenge of changing individual provider behavior without addressing the system within which the provider works.

We feel that our sample of 8 physicians was sufficient for a key informant analysis. The pool of potential informants was limited by their unique position and the requirement of having a parent with a recent serious illness episode. This physician sample was deliberately and purposefully selected. The stature of the respondents created a potential bias, a “VIP syndrome” for these physicians’ fathers.³ Rather than receiving excessive care, however, some of these patients received suboptimal or even antagonistic care.

As sons and daughters, it is possible that these physicians may be embittered about their fathers’ care, leading them to exaggerate or overstate their observations. The illness of a parent evokes intense emotions, but it also tends to rivet attention to the care being received.⁷ Because our informants also acted as participant-observers, it is possible that their observations lacked insight into the harm they may have caused by intervening in their fathers’ care. Despite these factors, the themes of advocacy and rescue were common to all of the physicians. The theme of abiding inner discord was strengthened by its enduring nature over time. By confirming these themes through re-interviews, we are confident that they are robust and valid for each individual as well as for the entire sample. Although these physicians’ reports of their experiences should not be considered generalizable to the population at large, they are informed expert opinions that raise serious concerns about how well the health care system is serving patients.

Our results are consistent with the burgeoning demand for improvement in our current health care system. Health care systems could affirm the continued presence of one physician who is in charge of the patient’s care and accountable to the patient and the patient’s family. Payment systems and health plan rules should not force discontinuity across different care settings. Physicians who have a relationship and previous experience with patients should be encouraged to remain involved in their care during hospitalizations. Health care begs to be rebalanced to emphasize the importance of knowing the patient at least as well as the disease process and medical technology.

We also found that these physicians’ experiences had a profound personal impact. The study physicians expressed a sense of being silenced by the system and were grateful for the relief afforded by telling their stories. This suggests that physicians and the systems they work in should create mechanisms for the discussion of troubling patient care events.

The personal experiences of these physicians hold special importance to other physicians, because they highlight the critical roles physicians are expected to play in a superior health care system. Many of the problems identified by these senior family physicians were manifest in physicians’ behaviors. Physicians should be able to express their ambivalence about problematic health care processes and encourage an environment that avoids blaming and promotes improvements. Rather than waiting for system-level change to improve health care, physicians could examine and change their own behaviors and practices.

Acknowledgments

Our paper is dedicated to Mary Lou Green, whose care at the end of her life inspired this study. The authors are indebted to Priscilla Noland and Michelle Perez for their assistance with the manuscript.

Pediatric emergencies are serious events that occur more commonly than many physicians think.¹ Recent studies have described adverse outcomes that can result when children with medical emergencies are directed to physicians’ offices by managed care plans.^2-5 However, little is known about the incidence or types of pediatric emergencies encountered or whether physicians are adequately prepared to treat them in their offices.

A few studies have estimated incidence but used different definitions of emergency, and there is considerable variation in the estimates of their occurrence. A study of pediatric practices in Vermont reported only 1 office emergency per practice per year,⁶ while a study in Connecticut reported that 82% of pediatric practices saw at least 1 child emergency per month.⁷ A Chicago study reported that 62% of pediatricians and family physicians saw at least 1 child a week who required urgent treatment or hospitalization.⁸ In a national sample, 68% of office-based pediatricians reported seeing 1 to 5 emergencies per week.⁹

The literature describing preparedness is limited principally to pediatricians and suggests that their skills in resuscitation and stabilization need improvement.^{7 ,9-12} Other than the Chicago study⁸ and one done in northern France,¹³ there are no studies of the frequency and nature of pediatric emergencies in family physicians’ offices, and little is known about how well prepared the typical family physician is to treat child emergencies. The Chicago study reported that the offices of family physicians were better stocked with resuscitation drugs than those of pediatricians but did not assess the training of physicians or their office staffs to deal with the medical emergencies of chidren.⁸ The French study is of little help in understanding the problem in the United States.¹³

In our study, we sought to understand: (1) both the incidence and variety of pediatric emergencies in the practices of family physicians and pediatricians in North Carolina; (2) how prepared physicians of each specialty were to treat such emergencies; (3) whether they believed they should provide emergency service; and (4) how important they felt it was to be prepared for pediatric emergencies.

Methods

We mailed a survey in late 1995 to a random sample of family physicians and pediatricians across North Carolina. The survey asked 25 questions about the prevalence and management of pediatric emergencies occurring in a physician’s office in 1995. The sample was randomly drawn from a database of 3184 family physicians and pediatricians licensed by the North Carolina Medical Board.

Survey Design and Procedures

The survey questions were developed by a group of family practice, pediatric, and emergency physicians. Our pilot study indicated a lower response rate for family physicians (58%) than for pediatricians (83%) and that fewer family physicians were regularly providing pediatric primary care (71% vs 100%). Accordingly, we mailed surveys to 250 family physicians and 150 pediatricians to obtain power sufficient to detect effect sizes with less than 5% error. Table 1 shows the sampling design and response rates for the 2 physician specialty groups.

Measures

The key variables are the occurrence of emergencies encountered by the practice, preparedness for emergencies through the availability of resuscitation and stabilization items, training for emergencies, and the perceived importance of providing and preparing for emergency care.

Emergency Occurrence and Preparedness

The variety of emergency occurrences was measured using 8 questions that asked whether specific pediatric emergencies were presented to the physician in the office setting during the last 12 months Table 2. Preparedness for pediatric emergencies was measured with 11 questions that asked whether specific stabilization and resuscitation items were immediately available to the physician in the office Table 3. The items were selected by experts and are regularly included in Pediatric Advanced Life Support (PALS) training. The responses allowed for this series of questions were “yes,” “no,” and “don’t know.” Answers other than “yes” were taken to mean that the item was not immediately available.

Occurrence of and preparedness for pediatric emergencies were analyzed as individual variables and as composite indexes. The composite index for the variety of emergency occurrences was created by assigning 1 point for each “yes” answer to whether a specific type of emergency occurred. This index, which we call the Emergency Occurrence Score (EOS), was an unweighted number with values ranging from 0 to 8, where a score of 8 was the greatest variety of occurrence (ie, all the types of emergencies occurred during the year). Similarly, an Emergency Preparedness Score (EPS) was created by assigning 1 point to each affirmative answer regarding the availability of specific stabilization and resuscitation items, with a range for the score of 0 to 11 (11 = optimal availability of resuscitation equipment).

Training for Emergencies

We asked 2 questions that assessed the training of physicians and office staff to deal with the emergencies of children. Specifically, we asked, “Have you taken a PALS or APLS (Advanced Pediatric Life Support) certification or instructor course in the last 2 years?” and “Has your office ever conducted a ‘mock’ or practice pediatric emergency?”

Perceived Importance of Preparedness

We measured respondents’ opinions on the importance of being prepared for a pediatric emergency and providing emergency services in the office. The specific questions were: “How important is it to you that your office or practice setting be prepared to stabilize a true pediatric emergency?” and “How important is it to you that your practice provides emergency care to critically ill children in your office?” We coded the ordinal responses for both questions on a scale from 1 through 4 (4 = very important).

Analysis

Our analysis contrasted family physicians and pediatricians. Statistical significance for all contrasts was set, a priori, at P less than .05. Cross-tabular analyses with chi-squares were used to compare percentage responses by specialty on the individual variables measuring occurrence and preparedness, mock emergencies, and PALS training. We also used a chi-square to examine the relationship between preparedness and PALS training. Independent-sample t tests were used to compare means by specialty on the 2 composite indices (EOS and EPS). Independent-sample t tests were also used to compare mean values for the questions on the believed importance of providing care to critically ill children and being prepared to stabilize them.

Two post-hoc hypotheses were suggested. One was whether family physicians and pediatricians differed on their knowledge of the availability of resuscitation items (ie, the number of “don’t know” responses). Another was whether physicians who reported seeing patients with respiratory emergencies had oxygen and a mask with which to administer it. Respiratory emergencies were defined as respiratory or cardiac arrest, foreign body in airway, asthma flare, and moderate to severe croup.

Results

Occurrence of Pediatric Emergencies

Table 2 presents the reported occurrence of pediatric emergencies in the office during the course of a year. Pediatric emergencies involving respiratory problems or cardiac arrest were seen in 4% of all practices, 2% in those of family physicians, and 7% in those of pediatricians. This difference was not statistically significant. Similar percentages (approximately 11%) of family physicians and pediatricians saw at least 1 child with a foreign body in the airway.

Each of the other 6 conditions were seen in a third or more of all family physician practices in the course of a year and in three fifths or more of pediatric practices. Asthma flares were the most commonly seen condition, occurring in 93% of both family practice and pediatric offices. More than 90% of both family physicians and pediatricians encountered at least 1 respiratory emergency. Approximately two thirds of the physicians in both specialties (64% of family physicians and 71% of pediatricians) commonly encountered allergic reactions.

There was no pediatric emergency condition that family physicians were statistically more likely to encounter than pediatricians. Pediatricians were significantly more likely to see severe croup (75% vs 48%), seizure (62% vs 33%), dehydration (85% vs 64%), and serious febrile illnesses (82% vs 65%). The average pediatric practice saw at least 4 of each of these child emergencies in the course of a year.

Family physicians saw fewer children with medical emergencies than did pediatricians. The mean EOS score was 3.8 for family physicians versus 4.9 for pediatricians (P <.001).

Emergency Preparedness

Table 3 shows emergency preparedness, as measured by the immediate availability of resuscitation items. The items are listed in increasing order of availability.

Ninety-six percent of all practices reported that they had oxygen immediately available; 88% had resuscitation drugs; and 86% had intravenous (IV) fluids. However, 27% had no appropriate way of administering IV fluids or drugs (no child-sized IV catheter or intraosseous needle). Family physicians were significantly less likely than pediatricians to have child-sized IV catheters, a pediatric bag-valve mask, an oral/nasal pediatric airway, or suction and pediatric catheters. Family physicians were also less likely to have available an intraosseous needle, a pediatric laryngoscope and endotracheal tube, a Broselow tape, or continuous pulse oximetry. In regard to having resucitation items, both family physicians and pediatricians were certain about the availability of oxygen, IV fluids, and continuous pulse oximetry. No family physicians or pediatricians responded “don’t know” on any of these items. Physicians were most uncertain about whether they had a Broselow tape (34% chose the “don’t know” responses) and an intraosseous needle (17% chose “don’t know”). Family physicians were significantly less likely than pediatricians to know whether they had a Broselow tape (43% vs 20%, respectively), an intraosseus needle (24% vs 5%, respectively), and suction and pediatric catheters (14% vs 5%, respectively).

Of those who encountered respiratory emergencies, a fifth of the physicians had no child-sized mask for administering oxygen. Family physicians were less likely to have one than pediatricians (30% did not vs 4%, respectively).

Family physicians were substantially less prepared for child emergencies, according to the inventory of immediately available items. The mean EPS was 5.7 for family physicians versus 8.6 for pediatricians (P <.001). Four-fifths of all physician offices had never conducted a mock or practice pediatric emergency. Only 6% of family physician offices had conducted a mock emergency versus 40% of pediatric offices (P <.001). Approximately one third of all physicians had taken a PALS training course during the previous 2 years. Less than a fifth of family physicians reported taking PALS training, compared with half of the pediatricians (P <.01).

There is a positive relationship between PALS training and preparedness. Those who had PALS training were significantly more likely to have a Broeslow tape available than those who did not (58% vs 21%, P <.001) and also were more likely to have an intraosseous needle (62% vs 24%, P <.001). Those who had PALS training were more than twice as likely to have practiced for a pediatric emergency (30% vs 13%, P <.01).

Beliefs About Preparing for Pediatric Emergencies in the Office

Family physicians and pediatricians differed in opinion on the importance of providing emergency care to critically ill children in their offices. Family physicians thought it less important to provide this service than did pediatricians. Family physicians rated providing care to critically ill children in their offices between “not very important” and “somewhat important.” The mean importance score for family physicians was 2.8 versus 3.5 for pediatricians (P <.001), with 4.0 being “very important.”

Physicians of both specialties attached less importance to providing care to critically ill children than being prepared to do so but differed by specialty on this dimension as well. Pediatricians thought it was more important to be prepared for these events (mean for family physicians = 3.4 vs 3.7 for pediatricians, P <.05), but the difference may not be clinically relevant. Each rated being prepared to stabilize a true pediatric emergency between “somewhat important” and “very important.”

Discussion

Although our study was conducted in only one state, we know of no reason the North Carolina experience would be unique. In contrast to the few previous studies, we asked about specific types of emergencies. We found that family physicians encounter a smaller variety of office pediatric emergencies than do pediatricians, but none of these events are unusual for either, and many are common for both. Only respiratory or cardiac arrest might be considered rare for family physicians (annually experienced by less than 5% of practices), but the 2% of family physicians in our sample would equal 40 cases per year in North Carolina. One in 10 family physicians treated a child with a foreign body in the airway, which would be more than 200 occurrences for family practice offices yearly in the state.

A third of those family physicians who regularly treat children can expect to encounter at least 1 pediatric seizure in their office during the year. Two thirds or more are likely to encounter children with seriously high fevers, dehydration, allergic reactions, or asthma attacks. Concern about office preparedness for these emergencies is warranted.

Our data suggest that family physicians’ practices may not be appropriately equipped. Having an oxygen tank is of little use if there is no mask to administer it. We do not recommend intubation in the office setting. Most emergency medicine experts agree that the key to airway management in out-of-hospital settings is positive-pressure ventilation through use of a bag-valve-mask, but the correct size is required. Similarly, there must be quick and sure means to administer fluids and IV drugs in the proper doses. The intraosseous needle may be the safest and fastest way, but less than 1 in 7 family physicians had this item. Doses for children vary greatly and quick information aids (such as the Broselow tape, which indicates drug dose and device size according to body length) may not be available. Familiarity and skill with these 2 items could be acquired in PALS training.

It is puzzling that family physicians think it less important to provide emergency care to critically ill children in their offices than do pediatricians. It may be that they think it is not appropriate for them to treat such emergencies or that such emergencies do not happen often enough to be of concern. Perhaps most of their data are anecdotal, and they simply have seen little information on the incidence, variety, or severity of pediatric office emergencies. Might they generalize their proficiency with adult emergencies to those of children? Some may believe that training in Advanced Cardiac Life Support provides sufficient skill to handle pediatric emergencies, but the techniques, doses, and device sizes for infants and young children are quite different from those for older children, adolescents, and adults.

Our study provides evidence that family physicians and pediatricians should assess the likelihood of children with serious medical emergencies presenting in their offices. As Trachtenbarg¹⁴ recently suggested, they should evaluate their own unique patient mix and be prepared accordingly. Both family physicians and pediatricians should evaluate the likelihood of seeing more pediatric emergencies because of managed care. Our study may help them judge their need for continuing training. Many states have developed specific training programs through the Emergency Medical Services for Children (EMS-C) program,* funded by the federal Maternal and Child Health Bureau. PALS training courses are commonly available. Our current and future work involves promoting and conducting such training.

Conclusions

Though no primary care office can be as well equipped as a hospital emergency department, the cost of basic equipment should not be prohibitive. It can cost less than $600 to prepare an office.¹⁵ In terms of time, PALS training is a 2-day course with refreshers every 2 years, and a mock emergency in the office can take less than an hour. Emergencies are by no means as common as ear infections; however, they do occur, and the consequences of being unprepared are serious. Family physicians need many skills. These data demonstrate that preparedness to stabilize and resuscitate children in emergency office encounters should be part of their repertoire.

Acknowledgments

Our study was funded by the North Carolina Office of Emergency Medical Services, Division of Facilities Services, Department of Health and Human Services, through a grant from the US Department of Health and Human Services. We thank Bob Bailey, Director of the North Carolina Office of Emergency Medical Services, for his assistance and support. We thank also Zoe Yetman (East Carolina University) for her attention to detail in managing the survey and Matthew Curry and Kevin Gross for reviewing the final manuscript.

Related Resources

• Emergency Medical Services for Children
• Maintained by the EMSC program, which is primarily supported and jointly administered by the U.S. Department of Health and Human Services’ Health Resources and Services Administration and the U.S. Department of Transportation’s National Highway Traffic Safety Administration. The site offers information on education and training, national and state activities, funding opportunities, and family education resources for injury control and prevention. http://www.ems-c.org
• American Academy of Pediatrics‹a comparison of APLS and PALS training and telephone numbers for both programs. " target="_blank">http://www.aap.org/profed/nrp/aplscom.htm>
• The components of the 14-hour APLS course may be found at " target="_blank">http://www.aap.org/profed/nrp/aplsccreq.htm>
• American Academy of Family Physicians (AAFP)‹Pediatric Advanced Life Support: A Review of the AHA Recommendations by Inis Jane Bardella, M.D. This article summarizes information from the PALS program but does not serve as a replacement for completing a PALS course. http://www.aafp.org/afp/991015ap/1743.html
• AAFP CME page‹Location and dates of PALS courses http://www.aafp.org/afp/calendar/cmecours.html

Pediatric emergencies are serious events that occur more commonly than many physicians think.¹ Recent studies have described adverse outcomes that can result when children with medical emergencies are directed to physicians’ offices by managed care plans.^2-5 However, little is known about the incidence or types of pediatric emergencies encountered or whether physicians are adequately prepared to treat them in their offices.

A few studies have estimated incidence but used different definitions of emergency, and there is considerable variation in the estimates of their occurrence. A study of pediatric practices in Vermont reported only 1 office emergency per practice per year,⁶ while a study in Connecticut reported that 82% of pediatric practices saw at least 1 child emergency per month.⁷ A Chicago study reported that 62% of pediatricians and family physicians saw at least 1 child a week who required urgent treatment or hospitalization.⁸ In a national sample, 68% of office-based pediatricians reported seeing 1 to 5 emergencies per week.⁹

The literature describing preparedness is limited principally to pediatricians and suggests that their skills in resuscitation and stabilization need improvement.^{7 ,9-12} Other than the Chicago study⁸ and one done in northern France,¹³ there are no studies of the frequency and nature of pediatric emergencies in family physicians’ offices, and little is known about how well prepared the typical family physician is to treat child emergencies. The Chicago study reported that the offices of family physicians were better stocked with resuscitation drugs than those of pediatricians but did not assess the training of physicians or their office staffs to deal with the medical emergencies of chidren.⁸ The French study is of little help in understanding the problem in the United States.¹³

In our study, we sought to understand: (1) both the incidence and variety of pediatric emergencies in the practices of family physicians and pediatricians in North Carolina; (2) how prepared physicians of each specialty were to treat such emergencies; (3) whether they believed they should provide emergency service; and (4) how important they felt it was to be prepared for pediatric emergencies.

Methods

We mailed a survey in late 1995 to a random sample of family physicians and pediatricians across North Carolina. The survey asked 25 questions about the prevalence and management of pediatric emergencies occurring in a physician’s office in 1995. The sample was randomly drawn from a database of 3184 family physicians and pediatricians licensed by the North Carolina Medical Board.

Survey Design and Procedures

The survey questions were developed by a group of family practice, pediatric, and emergency physicians. Our pilot study indicated a lower response rate for family physicians (58%) than for pediatricians (83%) and that fewer family physicians were regularly providing pediatric primary care (71% vs 100%). Accordingly, we mailed surveys to 250 family physicians and 150 pediatricians to obtain power sufficient to detect effect sizes with less than 5% error. Table 1 shows the sampling design and response rates for the 2 physician specialty groups.

Measures

The key variables are the occurrence of emergencies encountered by the practice, preparedness for emergencies through the availability of resuscitation and stabilization items, training for emergencies, and the perceived importance of providing and preparing for emergency care.

Emergency Occurrence and Preparedness

The variety of emergency occurrences was measured using 8 questions that asked whether specific pediatric emergencies were presented to the physician in the office setting during the last 12 months Table 2. Preparedness for pediatric emergencies was measured with 11 questions that asked whether specific stabilization and resuscitation items were immediately available to the physician in the office Table 3. The items were selected by experts and are regularly included in Pediatric Advanced Life Support (PALS) training. The responses allowed for this series of questions were “yes,” “no,” and “don’t know.” Answers other than “yes” were taken to mean that the item was not immediately available.

Occurrence of and preparedness for pediatric emergencies were analyzed as individual variables and as composite indexes. The composite index for the variety of emergency occurrences was created by assigning 1 point for each “yes” answer to whether a specific type of emergency occurred. This index, which we call the Emergency Occurrence Score (EOS), was an unweighted number with values ranging from 0 to 8, where a score of 8 was the greatest variety of occurrence (ie, all the types of emergencies occurred during the year). Similarly, an Emergency Preparedness Score (EPS) was created by assigning 1 point to each affirmative answer regarding the availability of specific stabilization and resuscitation items, with a range for the score of 0 to 11 (11 = optimal availability of resuscitation equipment).

Training for Emergencies

We asked 2 questions that assessed the training of physicians and office staff to deal with the emergencies of children. Specifically, we asked, “Have you taken a PALS or APLS (Advanced Pediatric Life Support) certification or instructor course in the last 2 years?” and “Has your office ever conducted a ‘mock’ or practice pediatric emergency?”

Perceived Importance of Preparedness

We measured respondents’ opinions on the importance of being prepared for a pediatric emergency and providing emergency services in the office. The specific questions were: “How important is it to you that your office or practice setting be prepared to stabilize a true pediatric emergency?” and “How important is it to you that your practice provides emergency care to critically ill children in your office?” We coded the ordinal responses for both questions on a scale from 1 through 4 (4 = very important).

Analysis

Our analysis contrasted family physicians and pediatricians. Statistical significance for all contrasts was set, a priori, at P less than .05. Cross-tabular analyses with chi-squares were used to compare percentage responses by specialty on the individual variables measuring occurrence and preparedness, mock emergencies, and PALS training. We also used a chi-square to examine the relationship between preparedness and PALS training. Independent-sample t tests were used to compare means by specialty on the 2 composite indices (EOS and EPS). Independent-sample t tests were also used to compare mean values for the questions on the believed importance of providing care to critically ill children and being prepared to stabilize them.

Two post-hoc hypotheses were suggested. One was whether family physicians and pediatricians differed on their knowledge of the availability of resuscitation items (ie, the number of “don’t know” responses). Another was whether physicians who reported seeing patients with respiratory emergencies had oxygen and a mask with which to administer it. Respiratory emergencies were defined as respiratory or cardiac arrest, foreign body in airway, asthma flare, and moderate to severe croup.

Results

Occurrence of Pediatric Emergencies

Table 2 presents the reported occurrence of pediatric emergencies in the office during the course of a year. Pediatric emergencies involving respiratory problems or cardiac arrest were seen in 4% of all practices, 2% in those of family physicians, and 7% in those of pediatricians. This difference was not statistically significant. Similar percentages (approximately 11%) of family physicians and pediatricians saw at least 1 child with a foreign body in the airway.

Each of the other 6 conditions were seen in a third or more of all family physician practices in the course of a year and in three fifths or more of pediatric practices. Asthma flares were the most commonly seen condition, occurring in 93% of both family practice and pediatric offices. More than 90% of both family physicians and pediatricians encountered at least 1 respiratory emergency. Approximately two thirds of the physicians in both specialties (64% of family physicians and 71% of pediatricians) commonly encountered allergic reactions.

There was no pediatric emergency condition that family physicians were statistically more likely to encounter than pediatricians. Pediatricians were significantly more likely to see severe croup (75% vs 48%), seizure (62% vs 33%), dehydration (85% vs 64%), and serious febrile illnesses (82% vs 65%). The average pediatric practice saw at least 4 of each of these child emergencies in the course of a year.

Family physicians saw fewer children with medical emergencies than did pediatricians. The mean EOS score was 3.8 for family physicians versus 4.9 for pediatricians (P <.001).

Emergency Preparedness

Table 3 shows emergency preparedness, as measured by the immediate availability of resuscitation items. The items are listed in increasing order of availability.

Ninety-six percent of all practices reported that they had oxygen immediately available; 88% had resuscitation drugs; and 86% had intravenous (IV) fluids. However, 27% had no appropriate way of administering IV fluids or drugs (no child-sized IV catheter or intraosseous needle). Family physicians were significantly less likely than pediatricians to have child-sized IV catheters, a pediatric bag-valve mask, an oral/nasal pediatric airway, or suction and pediatric catheters. Family physicians were also less likely to have available an intraosseous needle, a pediatric laryngoscope and endotracheal tube, a Broselow tape, or continuous pulse oximetry. In regard to having resucitation items, both family physicians and pediatricians were certain about the availability of oxygen, IV fluids, and continuous pulse oximetry. No family physicians or pediatricians responded “don’t know” on any of these items. Physicians were most uncertain about whether they had a Broselow tape (34% chose the “don’t know” responses) and an intraosseous needle (17% chose “don’t know”). Family physicians were significantly less likely than pediatricians to know whether they had a Broselow tape (43% vs 20%, respectively), an intraosseus needle (24% vs 5%, respectively), and suction and pediatric catheters (14% vs 5%, respectively).

Of those who encountered respiratory emergencies, a fifth of the physicians had no child-sized mask for administering oxygen. Family physicians were less likely to have one than pediatricians (30% did not vs 4%, respectively).

Family physicians were substantially less prepared for child emergencies, according to the inventory of immediately available items. The mean EPS was 5.7 for family physicians versus 8.6 for pediatricians (P <.001). Four-fifths of all physician offices had never conducted a mock or practice pediatric emergency. Only 6% of family physician offices had conducted a mock emergency versus 40% of pediatric offices (P <.001). Approximately one third of all physicians had taken a PALS training course during the previous 2 years. Less than a fifth of family physicians reported taking PALS training, compared with half of the pediatricians (P <.01).

There is a positive relationship between PALS training and preparedness. Those who had PALS training were significantly more likely to have a Broeslow tape available than those who did not (58% vs 21%, P <.001) and also were more likely to have an intraosseous needle (62% vs 24%, P <.001). Those who had PALS training were more than twice as likely to have practiced for a pediatric emergency (30% vs 13%, P <.01).

Beliefs About Preparing for Pediatric Emergencies in the Office

Family physicians and pediatricians differed in opinion on the importance of providing emergency care to critically ill children in their offices. Family physicians thought it less important to provide this service than did pediatricians. Family physicians rated providing care to critically ill children in their offices between “not very important” and “somewhat important.” The mean importance score for family physicians was 2.8 versus 3.5 for pediatricians (P <.001), with 4.0 being “very important.”

Physicians of both specialties attached less importance to providing care to critically ill children than being prepared to do so but differed by specialty on this dimension as well. Pediatricians thought it was more important to be prepared for these events (mean for family physicians = 3.4 vs 3.7 for pediatricians, P <.05), but the difference may not be clinically relevant. Each rated being prepared to stabilize a true pediatric emergency between “somewhat important” and “very important.”

Discussion

Although our study was conducted in only one state, we know of no reason the North Carolina experience would be unique. In contrast to the few previous studies, we asked about specific types of emergencies. We found that family physicians encounter a smaller variety of office pediatric emergencies than do pediatricians, but none of these events are unusual for either, and many are common for both. Only respiratory or cardiac arrest might be considered rare for family physicians (annually experienced by less than 5% of practices), but the 2% of family physicians in our sample would equal 40 cases per year in North Carolina. One in 10 family physicians treated a child with a foreign body in the airway, which would be more than 200 occurrences for family practice offices yearly in the state.

A third of those family physicians who regularly treat children can expect to encounter at least 1 pediatric seizure in their office during the year. Two thirds or more are likely to encounter children with seriously high fevers, dehydration, allergic reactions, or asthma attacks. Concern about office preparedness for these emergencies is warranted.

Our data suggest that family physicians’ practices may not be appropriately equipped. Having an oxygen tank is of little use if there is no mask to administer it. We do not recommend intubation in the office setting. Most emergency medicine experts agree that the key to airway management in out-of-hospital settings is positive-pressure ventilation through use of a bag-valve-mask, but the correct size is required. Similarly, there must be quick and sure means to administer fluids and IV drugs in the proper doses. The intraosseous needle may be the safest and fastest way, but less than 1 in 7 family physicians had this item. Doses for children vary greatly and quick information aids (such as the Broselow tape, which indicates drug dose and device size according to body length) may not be available. Familiarity and skill with these 2 items could be acquired in PALS training.

It is puzzling that family physicians think it less important to provide emergency care to critically ill children in their offices than do pediatricians. It may be that they think it is not appropriate for them to treat such emergencies or that such emergencies do not happen often enough to be of concern. Perhaps most of their data are anecdotal, and they simply have seen little information on the incidence, variety, or severity of pediatric office emergencies. Might they generalize their proficiency with adult emergencies to those of children? Some may believe that training in Advanced Cardiac Life Support provides sufficient skill to handle pediatric emergencies, but the techniques, doses, and device sizes for infants and young children are quite different from those for older children, adolescents, and adults.

Our study provides evidence that family physicians and pediatricians should assess the likelihood of children with serious medical emergencies presenting in their offices. As Trachtenbarg¹⁴ recently suggested, they should evaluate their own unique patient mix and be prepared accordingly. Both family physicians and pediatricians should evaluate the likelihood of seeing more pediatric emergencies because of managed care. Our study may help them judge their need for continuing training. Many states have developed specific training programs through the Emergency Medical Services for Children (EMS-C) program,* funded by the federal Maternal and Child Health Bureau. PALS training courses are commonly available. Our current and future work involves promoting and conducting such training.

Conclusions

Though no primary care office can be as well equipped as a hospital emergency department, the cost of basic equipment should not be prohibitive. It can cost less than $600 to prepare an office.¹⁵ In terms of time, PALS training is a 2-day course with refreshers every 2 years, and a mock emergency in the office can take less than an hour. Emergencies are by no means as common as ear infections; however, they do occur, and the consequences of being unprepared are serious. Family physicians need many skills. These data demonstrate that preparedness to stabilize and resuscitate children in emergency office encounters should be part of their repertoire.

Acknowledgments

Our study was funded by the North Carolina Office of Emergency Medical Services, Division of Facilities Services, Department of Health and Human Services, through a grant from the US Department of Health and Human Services. We thank Bob Bailey, Director of the North Carolina Office of Emergency Medical Services, for his assistance and support. We thank also Zoe Yetman (East Carolina University) for her attention to detail in managing the survey and Matthew Curry and Kevin Gross for reviewing the final manuscript.

Related Resources

• Emergency Medical Services for Children
• Maintained by the EMSC program, which is primarily supported and jointly administered by the U.S. Department of Health and Human Services’ Health Resources and Services Administration and the U.S. Department of Transportation’s National Highway Traffic Safety Administration. The site offers information on education and training, national and state activities, funding opportunities, and family education resources for injury control and prevention. http://www.ems-c.org
• American Academy of Pediatrics‹a comparison of APLS and PALS training and telephone numbers for both programs. " target="_blank">http://www.aap.org/profed/nrp/aplscom.htm>
• The components of the 14-hour APLS course may be found at " target="_blank">http://www.aap.org/profed/nrp/aplsccreq.htm>
• American Academy of Family Physicians (AAFP)‹Pediatric Advanced Life Support: A Review of the AHA Recommendations by Inis Jane Bardella, M.D. This article summarizes information from the PALS program but does not serve as a replacement for completing a PALS course. http://www.aafp.org/afp/991015ap/1743.html
• AAFP CME page‹Location and dates of PALS courses http://www.aafp.org/afp/calendar/cmecours.html

Pediatric emergencies are serious events that occur more commonly than many physicians think.¹ Recent studies have described adverse outcomes that can result when children with medical emergencies are directed to physicians’ offices by managed care plans.^2-5 However, little is known about the incidence or types of pediatric emergencies encountered or whether physicians are adequately prepared to treat them in their offices.

A few studies have estimated incidence but used different definitions of emergency, and there is considerable variation in the estimates of their occurrence. A study of pediatric practices in Vermont reported only 1 office emergency per practice per year,⁶ while a study in Connecticut reported that 82% of pediatric practices saw at least 1 child emergency per month.⁷ A Chicago study reported that 62% of pediatricians and family physicians saw at least 1 child a week who required urgent treatment or hospitalization.⁸ In a national sample, 68% of office-based pediatricians reported seeing 1 to 5 emergencies per week.⁹

The literature describing preparedness is limited principally to pediatricians and suggests that their skills in resuscitation and stabilization need improvement.^{7 ,9-12} Other than the Chicago study⁸ and one done in northern France,¹³ there are no studies of the frequency and nature of pediatric emergencies in family physicians’ offices, and little is known about how well prepared the typical family physician is to treat child emergencies. The Chicago study reported that the offices of family physicians were better stocked with resuscitation drugs than those of pediatricians but did not assess the training of physicians or their office staffs to deal with the medical emergencies of chidren.⁸ The French study is of little help in understanding the problem in the United States.¹³

In our study, we sought to understand: (1) both the incidence and variety of pediatric emergencies in the practices of family physicians and pediatricians in North Carolina; (2) how prepared physicians of each specialty were to treat such emergencies; (3) whether they believed they should provide emergency service; and (4) how important they felt it was to be prepared for pediatric emergencies.

Methods

We mailed a survey in late 1995 to a random sample of family physicians and pediatricians across North Carolina. The survey asked 25 questions about the prevalence and management of pediatric emergencies occurring in a physician’s office in 1995. The sample was randomly drawn from a database of 3184 family physicians and pediatricians licensed by the North Carolina Medical Board.

Survey Design and Procedures

The survey questions were developed by a group of family practice, pediatric, and emergency physicians. Our pilot study indicated a lower response rate for family physicians (58%) than for pediatricians (83%) and that fewer family physicians were regularly providing pediatric primary care (71% vs 100%). Accordingly, we mailed surveys to 250 family physicians and 150 pediatricians to obtain power sufficient to detect effect sizes with less than 5% error. Table 1 shows the sampling design and response rates for the 2 physician specialty groups.

Measures

The key variables are the occurrence of emergencies encountered by the practice, preparedness for emergencies through the availability of resuscitation and stabilization items, training for emergencies, and the perceived importance of providing and preparing for emergency care.

Emergency Occurrence and Preparedness

The variety of emergency occurrences was measured using 8 questions that asked whether specific pediatric emergencies were presented to the physician in the office setting during the last 12 months Table 2. Preparedness for pediatric emergencies was measured with 11 questions that asked whether specific stabilization and resuscitation items were immediately available to the physician in the office Table 3. The items were selected by experts and are regularly included in Pediatric Advanced Life Support (PALS) training. The responses allowed for this series of questions were “yes,” “no,” and “don’t know.” Answers other than “yes” were taken to mean that the item was not immediately available.

Occurrence of and preparedness for pediatric emergencies were analyzed as individual variables and as composite indexes. The composite index for the variety of emergency occurrences was created by assigning 1 point for each “yes” answer to whether a specific type of emergency occurred. This index, which we call the Emergency Occurrence Score (EOS), was an unweighted number with values ranging from 0 to 8, where a score of 8 was the greatest variety of occurrence (ie, all the types of emergencies occurred during the year). Similarly, an Emergency Preparedness Score (EPS) was created by assigning 1 point to each affirmative answer regarding the availability of specific stabilization and resuscitation items, with a range for the score of 0 to 11 (11 = optimal availability of resuscitation equipment).

Training for Emergencies

We asked 2 questions that assessed the training of physicians and office staff to deal with the emergencies of children. Specifically, we asked, “Have you taken a PALS or APLS (Advanced Pediatric Life Support) certification or instructor course in the last 2 years?” and “Has your office ever conducted a ‘mock’ or practice pediatric emergency?”

Perceived Importance of Preparedness

We measured respondents’ opinions on the importance of being prepared for a pediatric emergency and providing emergency services in the office. The specific questions were: “How important is it to you that your office or practice setting be prepared to stabilize a true pediatric emergency?” and “How important is it to you that your practice provides emergency care to critically ill children in your office?” We coded the ordinal responses for both questions on a scale from 1 through 4 (4 = very important).

Analysis

Our analysis contrasted family physicians and pediatricians. Statistical significance for all contrasts was set, a priori, at P less than .05. Cross-tabular analyses with chi-squares were used to compare percentage responses by specialty on the individual variables measuring occurrence and preparedness, mock emergencies, and PALS training. We also used a chi-square to examine the relationship between preparedness and PALS training. Independent-sample t tests were used to compare means by specialty on the 2 composite indices (EOS and EPS). Independent-sample t tests were also used to compare mean values for the questions on the believed importance of providing care to critically ill children and being prepared to stabilize them.

Two post-hoc hypotheses were suggested. One was whether family physicians and pediatricians differed on their knowledge of the availability of resuscitation items (ie, the number of “don’t know” responses). Another was whether physicians who reported seeing patients with respiratory emergencies had oxygen and a mask with which to administer it. Respiratory emergencies were defined as respiratory or cardiac arrest, foreign body in airway, asthma flare, and moderate to severe croup.

Results

Occurrence of Pediatric Emergencies

Table 2 presents the reported occurrence of pediatric emergencies in the office during the course of a year. Pediatric emergencies involving respiratory problems or cardiac arrest were seen in 4% of all practices, 2% in those of family physicians, and 7% in those of pediatricians. This difference was not statistically significant. Similar percentages (approximately 11%) of family physicians and pediatricians saw at least 1 child with a foreign body in the airway.

Each of the other 6 conditions were seen in a third or more of all family physician practices in the course of a year and in three fifths or more of pediatric practices. Asthma flares were the most commonly seen condition, occurring in 93% of both family practice and pediatric offices. More than 90% of both family physicians and pediatricians encountered at least 1 respiratory emergency. Approximately two thirds of the physicians in both specialties (64% of family physicians and 71% of pediatricians) commonly encountered allergic reactions.

There was no pediatric emergency condition that family physicians were statistically more likely to encounter than pediatricians. Pediatricians were significantly more likely to see severe croup (75% vs 48%), seizure (62% vs 33%), dehydration (85% vs 64%), and serious febrile illnesses (82% vs 65%). The average pediatric practice saw at least 4 of each of these child emergencies in the course of a year.

Family physicians saw fewer children with medical emergencies than did pediatricians. The mean EOS score was 3.8 for family physicians versus 4.9 for pediatricians (P <.001).

Emergency Preparedness

Table 3 shows emergency preparedness, as measured by the immediate availability of resuscitation items. The items are listed in increasing order of availability.

Ninety-six percent of all practices reported that they had oxygen immediately available; 88% had resuscitation drugs; and 86% had intravenous (IV) fluids. However, 27% had no appropriate way of administering IV fluids or drugs (no child-sized IV catheter or intraosseous needle). Family physicians were significantly less likely than pediatricians to have child-sized IV catheters, a pediatric bag-valve mask, an oral/nasal pediatric airway, or suction and pediatric catheters. Family physicians were also less likely to have available an intraosseous needle, a pediatric laryngoscope and endotracheal tube, a Broselow tape, or continuous pulse oximetry. In regard to having resucitation items, both family physicians and pediatricians were certain about the availability of oxygen, IV fluids, and continuous pulse oximetry. No family physicians or pediatricians responded “don’t know” on any of these items. Physicians were most uncertain about whether they had a Broselow tape (34% chose the “don’t know” responses) and an intraosseous needle (17% chose “don’t know”). Family physicians were significantly less likely than pediatricians to know whether they had a Broselow tape (43% vs 20%, respectively), an intraosseus needle (24% vs 5%, respectively), and suction and pediatric catheters (14% vs 5%, respectively).

Of those who encountered respiratory emergencies, a fifth of the physicians had no child-sized mask for administering oxygen. Family physicians were less likely to have one than pediatricians (30% did not vs 4%, respectively).

Family physicians were substantially less prepared for child emergencies, according to the inventory of immediately available items. The mean EPS was 5.7 for family physicians versus 8.6 for pediatricians (P <.001). Four-fifths of all physician offices had never conducted a mock or practice pediatric emergency. Only 6% of family physician offices had conducted a mock emergency versus 40% of pediatric offices (P <.001). Approximately one third of all physicians had taken a PALS training course during the previous 2 years. Less than a fifth of family physicians reported taking PALS training, compared with half of the pediatricians (P <.01).

There is a positive relationship between PALS training and preparedness. Those who had PALS training were significantly more likely to have a Broeslow tape available than those who did not (58% vs 21%, P <.001) and also were more likely to have an intraosseous needle (62% vs 24%, P <.001). Those who had PALS training were more than twice as likely to have practiced for a pediatric emergency (30% vs 13%, P <.01).

Beliefs About Preparing for Pediatric Emergencies in the Office

Family physicians and pediatricians differed in opinion on the importance of providing emergency care to critically ill children in their offices. Family physicians thought it less important to provide this service than did pediatricians. Family physicians rated providing care to critically ill children in their offices between “not very important” and “somewhat important.” The mean importance score for family physicians was 2.8 versus 3.5 for pediatricians (P <.001), with 4.0 being “very important.”

Physicians of both specialties attached less importance to providing care to critically ill children than being prepared to do so but differed by specialty on this dimension as well. Pediatricians thought it was more important to be prepared for these events (mean for family physicians = 3.4 vs 3.7 for pediatricians, P <.05), but the difference may not be clinically relevant. Each rated being prepared to stabilize a true pediatric emergency between “somewhat important” and “very important.”

Discussion

Although our study was conducted in only one state, we know of no reason the North Carolina experience would be unique. In contrast to the few previous studies, we asked about specific types of emergencies. We found that family physicians encounter a smaller variety of office pediatric emergencies than do pediatricians, but none of these events are unusual for either, and many are common for both. Only respiratory or cardiac arrest might be considered rare for family physicians (annually experienced by less than 5% of practices), but the 2% of family physicians in our sample would equal 40 cases per year in North Carolina. One in 10 family physicians treated a child with a foreign body in the airway, which would be more than 200 occurrences for family practice offices yearly in the state.

A third of those family physicians who regularly treat children can expect to encounter at least 1 pediatric seizure in their office during the year. Two thirds or more are likely to encounter children with seriously high fevers, dehydration, allergic reactions, or asthma attacks. Concern about office preparedness for these emergencies is warranted.

Our data suggest that family physicians’ practices may not be appropriately equipped. Having an oxygen tank is of little use if there is no mask to administer it. We do not recommend intubation in the office setting. Most emergency medicine experts agree that the key to airway management in out-of-hospital settings is positive-pressure ventilation through use of a bag-valve-mask, but the correct size is required. Similarly, there must be quick and sure means to administer fluids and IV drugs in the proper doses. The intraosseous needle may be the safest and fastest way, but less than 1 in 7 family physicians had this item. Doses for children vary greatly and quick information aids (such as the Broselow tape, which indicates drug dose and device size according to body length) may not be available. Familiarity and skill with these 2 items could be acquired in PALS training.

It is puzzling that family physicians think it less important to provide emergency care to critically ill children in their offices than do pediatricians. It may be that they think it is not appropriate for them to treat such emergencies or that such emergencies do not happen often enough to be of concern. Perhaps most of their data are anecdotal, and they simply have seen little information on the incidence, variety, or severity of pediatric office emergencies. Might they generalize their proficiency with adult emergencies to those of children? Some may believe that training in Advanced Cardiac Life Support provides sufficient skill to handle pediatric emergencies, but the techniques, doses, and device sizes for infants and young children are quite different from those for older children, adolescents, and adults.

Our study provides evidence that family physicians and pediatricians should assess the likelihood of children with serious medical emergencies presenting in their offices. As Trachtenbarg¹⁴ recently suggested, they should evaluate their own unique patient mix and be prepared accordingly. Both family physicians and pediatricians should evaluate the likelihood of seeing more pediatric emergencies because of managed care. Our study may help them judge their need for continuing training. Many states have developed specific training programs through the Emergency Medical Services for Children (EMS-C) program,* funded by the federal Maternal and Child Health Bureau. PALS training courses are commonly available. Our current and future work involves promoting and conducting such training.

Conclusions

Though no primary care office can be as well equipped as a hospital emergency department, the cost of basic equipment should not be prohibitive. It can cost less than $600 to prepare an office.¹⁵ In terms of time, PALS training is a 2-day course with refreshers every 2 years, and a mock emergency in the office can take less than an hour. Emergencies are by no means as common as ear infections; however, they do occur, and the consequences of being unprepared are serious. Family physicians need many skills. These data demonstrate that preparedness to stabilize and resuscitate children in emergency office encounters should be part of their repertoire.

Acknowledgments

Our study was funded by the North Carolina Office of Emergency Medical Services, Division of Facilities Services, Department of Health and Human Services, through a grant from the US Department of Health and Human Services. We thank Bob Bailey, Director of the North Carolina Office of Emergency Medical Services, for his assistance and support. We thank also Zoe Yetman (East Carolina University) for her attention to detail in managing the survey and Matthew Curry and Kevin Gross for reviewing the final manuscript.

Related Resources

• Emergency Medical Services for Children
• Maintained by the EMSC program, which is primarily supported and jointly administered by the U.S. Department of Health and Human Services’ Health Resources and Services Administration and the U.S. Department of Transportation’s National Highway Traffic Safety Administration. The site offers information on education and training, national and state activities, funding opportunities, and family education resources for injury control and prevention. http://www.ems-c.org
• American Academy of Pediatrics‹a comparison of APLS and PALS training and telephone numbers for both programs. " target="_blank">http://www.aap.org/profed/nrp/aplscom.htm>
• The components of the 14-hour APLS course may be found at " target="_blank">http://www.aap.org/profed/nrp/aplsccreq.htm>
• American Academy of Family Physicians (AAFP)‹Pediatric Advanced Life Support: A Review of the AHA Recommendations by Inis Jane Bardella, M.D. This article summarizes information from the PALS program but does not serve as a replacement for completing a PALS course. http://www.aafp.org/afp/991015ap/1743.html
• AAFP CME page‹Location and dates of PALS courses http://www.aafp.org/afp/calendar/cmecours.html

The 1990s were a time of remarkable upheaval in US health care. The input of physicians and consumers was marginalized in a market-driven system guided by insurers, employers, and government. Well-intentioned managed care concepts that many physicians would endorse were less evident than managed care business practices that sparked rancor and frustration, as well as creating time and income pressures.

Although 80% of physicians believe that the current health care system requires fundamental change,^1,2 there is little agreement about what form that change should take. For example, 25%³ to 40%⁴ of physicians in statewide surveys and 66% of students entering one medical school system⁵ endorse a single-payer national health plan. Despite the lack of consensus about what direction reform should take, there is growing skepticism about the current model. For example, only 21% of surveyed internal medicine residents believe that corporate managed care is the best model for our health care system.⁶ Among students, residents, faculty members, and deans of medical schools, 52% favor a single-payer system, while 22% favor a corporate managed care system.⁷

Coincident with marked changes in American health care, the popular press have reported deteriorating morale and career satisfaction among physicians.^8,9 For family physicians there are mixed but generally positive data about career satisfaction. Survey data report that only 52% of young family physicians would definitely again choose to become a physician,10 yet from 70%10 to 74%¹¹ would again choose family practice as a specialty. In a 1996 survey⁴ (at a time when both family practice residency applicants and salaries for new graduates peaked), 82% of family physicians reported satisfaction with their careers; 75% stated that they would again select family practice as a specialty; and 65% reported satisfaction with their compensation.⁴

Since 1996, annual surveys of the family practice residents in our own training program have shown a steady decline in the number who would go to medical school again (from 73% to 50% between 1996 and 1999). We suspected that our residents were reacting to changes and uncertainties similar to those faced by all family physicians. This prompted us to survey practicing family physicians about their experiences with and attitudes toward their practice and the US health care system.

Methods

We designed a 41-item questionnaire that focused on family physicians’ observations related to corporate managed health care and personal perception of their own well-being and career. Family physicians were asked to rate each item on a 5-point scale (where 1=strongly disagree, and 5=strongly agree). Additional items focused on demographic information, nature of practice, presence of a primary relationship, presence of children in the home, political inclination, extent of managed care participation, and both the nature and perceived adequacy of compensation. The final questionnaire was the product of pretesting and feedback by 6 of the family practice faculty members from our residency training program and consultation with researchers.

We selected a random national sample of 800 practicing members of the American Academy of Family Physicians (AAFP) in June 1999. A 6-page survey was mailed to each family physician, accompanied by a cover letter requesting participation and a postcard by which a summary of the final survey results could be requested. We sent a second mailing 5 weeks later to those who had not yet responded.

We examined the data from all 41 items that were rated on a Likert scale using principal component factor analysis and varimax rotation with Kaiser normalization.¹² The 4 factors of highest reliability (a >0.650) and highest eigenvalue were chosen post hoc as dependent variables (ie, attitudes about managed care, satisfaction with career, workload stress, and satisfaction with personal life) and analyzed by stepwise regression using all remaining factors and personal and practice variables as independent variables.

Results

A total of 361 completed surveys were returned; 16 surveys were returned as undeliverable (response rate=46%). The mean physician age was 46 years (range=27-77 years); 74% were men, 26% were women. Most were white/non-Hispanic; only 1.1% were Hispanic, 1.7%, African-American, and 4.8%, Asian-American. Practice characteristics are shown in Table 1. The characteristics of these respondents were very similar to those reported for the AAFP membership in December 2000.¹³ The exceptions were that minority family physicians were underrepresented among respondents to our survey and that respondents averaged nearly 5 more years in practice than the AAFP membership mean.

Among this sample of family physicians, 51% received all of their income by salary. Only 10% reported that more than half their income was capitated; 72% still received no income through capitation. The respondents reported working a mean of 51 hours per week; however, 33% reported averaging more than 60 hours, and 7% averaged more than 70 hours. They had spent a mean of 15.3 years in practice (range = 1 to 51 years).

Only 55% of respondents felt they were appropriately compensated financially. Family physicians who were satisfied with their financial compensation reported more favorable views of managed care (F=10.55; df=1; P <.001), greater career satisfaction (F=27.07; df=1; P <.0001) and lower workload stress (F=10.42; df=1; P <.001).

Regarding their personal lives, 91% of these family physicians reported they were currently involved in a relationship, and 65% had children in their home. Politically, 30% identified themselves as democrats; 46%, republicans; and 24%, independents.

The responses to the questionnaire are shown in Table 2,Table 2b,Table 2c,. The results of stepwise regression models are presented in Table 3*Table w2.

Attitudes About Managed Care

The responses to items composing this factor show that this sample of family physicians was sharply critical of corporate managed care. For example, only 7% agreed or strongly agreed that corporate managed care is the best way to provide the health care America needs at a cost that society can afford.

The regression analyses summarized in Table 3 suggest that overall attitudes about managed care were predicted primarily by a cluster of 3 factors that tap into other concerns about managed care, reform, and accountability. Also, positive attitudes about managed care were best predicted by greater career satisfaction, a willingness to let others worry about the business of medicine, support of hospitalists, having children at home, and democratic political inclinations.

Satisfaction with Career

The response to items composing this factor suggests evidence of discontent among a significant minority of these family physicians. For example, knowing what they know now, if they could start over 63% would go to medical school again, and 66% would choose family practice again. Overall, 59% said they were satisfied with their professional lives.

In regression analyses, family physicians who were retrospectively satisfied with residency training, felt that they were appropriately compensated for their current work, and were satisfied with their personal lives were most likely to be satisfied with their careers. Both a positive attitude about managed care and reform-mindedness each independently predicted career satisfaction.

A decreased workload stress factor predicted increased career satisfaction; however, longer work hours were reported by those more satisfied with their careers. Women reported less satisfaction with their careers. Given their covariance, it is reasonable to view these variables as a cluster that together predicts 55% of the variance in career-satisfaction ratings.

Workload Stress

The response to items composing this factor suggests that a significant number of respondents felt their practice was very stressful. For example, only 9% reported that they are not overwhelmed by paperwork and administrative requirements, and 48% said that they had personally experienced more stress-related symptoms within the past year. Regression analyses suggested that greater workload stress was significantly predicted by dissatisfaction with either career or personal life, longer work hours, being female, and greater reform-mindedness.

Satisfaction with Personal Life

In their personal lives, 85% of these family physicians felt supported by a primary relationship, their family, or a social network. Healthy lifestyle habits that might help to buffer stress are reported by 52%. Overall, 71% said they were satisfied with their personal lives. As provided in Table 3, regression analyses showed that greater workload stress and longer work hours each predicted less satisfaction with personal life. Overall career satisfaction and being in a primary relationship predicted greater satisfaction with personal life. Having children in the home did not predict greater satisfaction with personal life.

Discussion

The optimism and career satisfaction reported among the majority of family physicians during the mid-1990s may have eroded in subsequent years. In this national survey in 1999, 63% would again choose to go to medical school, and 66% would again choose family practice. Only 55% were satisfied that they are receiving appropriate financial compensation, and 31% stated that they were “tired of being a doctor” and worried that they were “burning out.”

Although it is unlikely that financial incentives are a primary motivator for the decision to choose family practice, satisfaction with financial compensation was significantly associated with greater career satisfaction and lower work stress.

Nearly half of these family physicians reported that they had personally experienced a recent increase in stress-related symptoms. Many did not feel in control of the stressors they faced in daily practice, and most were overwhelmed by the paperwork and administrative requirements. It is no surprise that career satisfaction and workload stress are inversely related. However, it is less clear whether escalating work stress usually erodes career satisfaction or whether those who are less enamored with their career choice are more likely to experience practice demands as stressful.

Higher reported work hours were associated with greater career satisfaction as well as with higher perceived workload stress and less satisfaction with personal life. Although family physicians working the most hours may be doing so because they love what they are doing, long hours take their toll. Forty-four percent of respondents said they were working so many hours that their health or relationships were adversely affected. Several respondents wrote in the margins that they had opted for 50% to 80% of full time, because it was the only way to make practice tolerable. However, such a work arrangement often constitutes a full-time job by most standards, but for reduced pay.

Only 12% of the respondents felt that residency training had provided good training in practice administration. Despite recent efforts to create more relevant practice management curricula, ratings of this item did not vary significantly by years in practice, suggesting that most new graduates still do not feel prepared for the business of medicine.

Previous research has clearly shown that increased demands coupled with loss of control and a lack of safe outlets for frustration can lead to increased stress.¹⁴ This is the situation in which many of today’s family physicians find themselves. Residency training programs and state professional associations can play a significant role by speaking with candor about physician stress; integrating programs on stress management, career goals, and values clarification; and encouraging advocacy for constructive change of the health care system. Training programs must contemplate which changes are likely to maximize the number of graduates who become successful, satisfied practitioners in the current health care environment.

This sample of family physicians was sharply critical of a health care system driven by corporate managed care. Only 7% of family physicians agreed that corporate managed care is the best way to provide the health care America needs at a cost society can afford. Only 20% believed that managed care had accomplished improvements in both accountability and quality of care. Forty-two percent of these family physicians reported bad patient outcomes they perceived as attributable to managed care business processes. Despite the improved access touted by behavioral health carve-outs, 76% of family physicians report that obtaining good-quality behavioral or mental health treatment for their patients has gotten more difficult as such corporate entities have dominated that market.

There was virtual consensus for specific reforms. More than 85% of these family physicians agreed that all Americans should have access to redress for damages when managed care agents inappropriately delay or deny treatment and that managed care medical directors should be accountable to state boards of physician quality assurance for their decisions about patient care. Only 36% unequivocally supported a government-administered universal national health plan; however, an additional 21% who were equivocal suggested that growing discontent with the status quo could shift the balance to a majority in support.

Limitations

The numerous limitations inherent to our study include the use of a newly constructed questionnaire with design flaws, a modest response rate, the likelihood that respondents were different than nonrespondents, and poor representation of minority respondents. Comparison with data from other studies is limited by differences in study design and sample. These cross-sectional regression analyses allow only statistical prediction of associations, not cause-and-effect conclusions. This survey cannot be generalized to other physician specialties, and it is not appropriate to assume that family physicians are more discontent than other physicians, other professionals, or other workers.

Conclusions

This survey conveys a sobering glimpse of the prevalence of discontent among family physicians. Given the daily effects of upheaval in our health care system, perhaps we should be pleased that 2 out of 3 family physicians would still choose the same path if they had to do it over again. However, the next decade may answer whether our health care system can thrive if there is continued erosion of the support and confidence of skilled, committed, and reasonably content primary care physicians.

Acknowledgments

Financial assistance was provided by Medstar Research Institute. Thanks to both Ms. Michelle Appel and Medstar Research Institute for statistical consultation. The opinions expressed are the views of the authors and are not to be construed as reflecting the views of Medstar Health or the Public Health Service.

The 1990s were a time of remarkable upheaval in US health care. The input of physicians and consumers was marginalized in a market-driven system guided by insurers, employers, and government. Well-intentioned managed care concepts that many physicians would endorse were less evident than managed care business practices that sparked rancor and frustration, as well as creating time and income pressures.

Although 80% of physicians believe that the current health care system requires fundamental change,^1,2 there is little agreement about what form that change should take. For example, 25%³ to 40%⁴ of physicians in statewide surveys and 66% of students entering one medical school system⁵ endorse a single-payer national health plan. Despite the lack of consensus about what direction reform should take, there is growing skepticism about the current model. For example, only 21% of surveyed internal medicine residents believe that corporate managed care is the best model for our health care system.⁶ Among students, residents, faculty members, and deans of medical schools, 52% favor a single-payer system, while 22% favor a corporate managed care system.⁷

Coincident with marked changes in American health care, the popular press have reported deteriorating morale and career satisfaction among physicians.^8,9 For family physicians there are mixed but generally positive data about career satisfaction. Survey data report that only 52% of young family physicians would definitely again choose to become a physician,10 yet from 70%10 to 74%¹¹ would again choose family practice as a specialty. In a 1996 survey⁴ (at a time when both family practice residency applicants and salaries for new graduates peaked), 82% of family physicians reported satisfaction with their careers; 75% stated that they would again select family practice as a specialty; and 65% reported satisfaction with their compensation.⁴

Since 1996, annual surveys of the family practice residents in our own training program have shown a steady decline in the number who would go to medical school again (from 73% to 50% between 1996 and 1999). We suspected that our residents were reacting to changes and uncertainties similar to those faced by all family physicians. This prompted us to survey practicing family physicians about their experiences with and attitudes toward their practice and the US health care system.

Methods

We designed a 41-item questionnaire that focused on family physicians’ observations related to corporate managed health care and personal perception of their own well-being and career. Family physicians were asked to rate each item on a 5-point scale (where 1=strongly disagree, and 5=strongly agree). Additional items focused on demographic information, nature of practice, presence of a primary relationship, presence of children in the home, political inclination, extent of managed care participation, and both the nature and perceived adequacy of compensation. The final questionnaire was the product of pretesting and feedback by 6 of the family practice faculty members from our residency training program and consultation with researchers.

We selected a random national sample of 800 practicing members of the American Academy of Family Physicians (AAFP) in June 1999. A 6-page survey was mailed to each family physician, accompanied by a cover letter requesting participation and a postcard by which a summary of the final survey results could be requested. We sent a second mailing 5 weeks later to those who had not yet responded.

We examined the data from all 41 items that were rated on a Likert scale using principal component factor analysis and varimax rotation with Kaiser normalization.¹² The 4 factors of highest reliability (a >0.650) and highest eigenvalue were chosen post hoc as dependent variables (ie, attitudes about managed care, satisfaction with career, workload stress, and satisfaction with personal life) and analyzed by stepwise regression using all remaining factors and personal and practice variables as independent variables.

Results

A total of 361 completed surveys were returned; 16 surveys were returned as undeliverable (response rate=46%). The mean physician age was 46 years (range=27-77 years); 74% were men, 26% were women. Most were white/non-Hispanic; only 1.1% were Hispanic, 1.7%, African-American, and 4.8%, Asian-American. Practice characteristics are shown in Table 1. The characteristics of these respondents were very similar to those reported for the AAFP membership in December 2000.¹³ The exceptions were that minority family physicians were underrepresented among respondents to our survey and that respondents averaged nearly 5 more years in practice than the AAFP membership mean.

Among this sample of family physicians, 51% received all of their income by salary. Only 10% reported that more than half their income was capitated; 72% still received no income through capitation. The respondents reported working a mean of 51 hours per week; however, 33% reported averaging more than 60 hours, and 7% averaged more than 70 hours. They had spent a mean of 15.3 years in practice (range = 1 to 51 years).

Only 55% of respondents felt they were appropriately compensated financially. Family physicians who were satisfied with their financial compensation reported more favorable views of managed care (F=10.55; df=1; P <.001), greater career satisfaction (F=27.07; df=1; P <.0001) and lower workload stress (F=10.42; df=1; P <.001).

Regarding their personal lives, 91% of these family physicians reported they were currently involved in a relationship, and 65% had children in their home. Politically, 30% identified themselves as democrats; 46%, republicans; and 24%, independents.

The responses to the questionnaire are shown in Table 2,Table 2b,Table 2c,. The results of stepwise regression models are presented in Table 3*Table w2.

Attitudes About Managed Care

The responses to items composing this factor show that this sample of family physicians was sharply critical of corporate managed care. For example, only 7% agreed or strongly agreed that corporate managed care is the best way to provide the health care America needs at a cost that society can afford.

The regression analyses summarized in Table 3 suggest that overall attitudes about managed care were predicted primarily by a cluster of 3 factors that tap into other concerns about managed care, reform, and accountability. Also, positive attitudes about managed care were best predicted by greater career satisfaction, a willingness to let others worry about the business of medicine, support of hospitalists, having children at home, and democratic political inclinations.

Satisfaction with Career

The response to items composing this factor suggests evidence of discontent among a significant minority of these family physicians. For example, knowing what they know now, if they could start over 63% would go to medical school again, and 66% would choose family practice again. Overall, 59% said they were satisfied with their professional lives.

In regression analyses, family physicians who were retrospectively satisfied with residency training, felt that they were appropriately compensated for their current work, and were satisfied with their personal lives were most likely to be satisfied with their careers. Both a positive attitude about managed care and reform-mindedness each independently predicted career satisfaction.

A decreased workload stress factor predicted increased career satisfaction; however, longer work hours were reported by those more satisfied with their careers. Women reported less satisfaction with their careers. Given their covariance, it is reasonable to view these variables as a cluster that together predicts 55% of the variance in career-satisfaction ratings.

Workload Stress

The response to items composing this factor suggests that a significant number of respondents felt their practice was very stressful. For example, only 9% reported that they are not overwhelmed by paperwork and administrative requirements, and 48% said that they had personally experienced more stress-related symptoms within the past year. Regression analyses suggested that greater workload stress was significantly predicted by dissatisfaction with either career or personal life, longer work hours, being female, and greater reform-mindedness.

Satisfaction with Personal Life

In their personal lives, 85% of these family physicians felt supported by a primary relationship, their family, or a social network. Healthy lifestyle habits that might help to buffer stress are reported by 52%. Overall, 71% said they were satisfied with their personal lives. As provided in Table 3, regression analyses showed that greater workload stress and longer work hours each predicted less satisfaction with personal life. Overall career satisfaction and being in a primary relationship predicted greater satisfaction with personal life. Having children in the home did not predict greater satisfaction with personal life.

Discussion

The optimism and career satisfaction reported among the majority of family physicians during the mid-1990s may have eroded in subsequent years. In this national survey in 1999, 63% would again choose to go to medical school, and 66% would again choose family practice. Only 55% were satisfied that they are receiving appropriate financial compensation, and 31% stated that they were “tired of being a doctor” and worried that they were “burning out.”

Although it is unlikely that financial incentives are a primary motivator for the decision to choose family practice, satisfaction with financial compensation was significantly associated with greater career satisfaction and lower work stress.

Nearly half of these family physicians reported that they had personally experienced a recent increase in stress-related symptoms. Many did not feel in control of the stressors they faced in daily practice, and most were overwhelmed by the paperwork and administrative requirements. It is no surprise that career satisfaction and workload stress are inversely related. However, it is less clear whether escalating work stress usually erodes career satisfaction or whether those who are less enamored with their career choice are more likely to experience practice demands as stressful.

Higher reported work hours were associated with greater career satisfaction as well as with higher perceived workload stress and less satisfaction with personal life. Although family physicians working the most hours may be doing so because they love what they are doing, long hours take their toll. Forty-four percent of respondents said they were working so many hours that their health or relationships were adversely affected. Several respondents wrote in the margins that they had opted for 50% to 80% of full time, because it was the only way to make practice tolerable. However, such a work arrangement often constitutes a full-time job by most standards, but for reduced pay.

Only 12% of the respondents felt that residency training had provided good training in practice administration. Despite recent efforts to create more relevant practice management curricula, ratings of this item did not vary significantly by years in practice, suggesting that most new graduates still do not feel prepared for the business of medicine.

Previous research has clearly shown that increased demands coupled with loss of control and a lack of safe outlets for frustration can lead to increased stress.¹⁴ This is the situation in which many of today’s family physicians find themselves. Residency training programs and state professional associations can play a significant role by speaking with candor about physician stress; integrating programs on stress management, career goals, and values clarification; and encouraging advocacy for constructive change of the health care system. Training programs must contemplate which changes are likely to maximize the number of graduates who become successful, satisfied practitioners in the current health care environment.

This sample of family physicians was sharply critical of a health care system driven by corporate managed care. Only 7% of family physicians agreed that corporate managed care is the best way to provide the health care America needs at a cost society can afford. Only 20% believed that managed care had accomplished improvements in both accountability and quality of care. Forty-two percent of these family physicians reported bad patient outcomes they perceived as attributable to managed care business processes. Despite the improved access touted by behavioral health carve-outs, 76% of family physicians report that obtaining good-quality behavioral or mental health treatment for their patients has gotten more difficult as such corporate entities have dominated that market.

There was virtual consensus for specific reforms. More than 85% of these family physicians agreed that all Americans should have access to redress for damages when managed care agents inappropriately delay or deny treatment and that managed care medical directors should be accountable to state boards of physician quality assurance for their decisions about patient care. Only 36% unequivocally supported a government-administered universal national health plan; however, an additional 21% who were equivocal suggested that growing discontent with the status quo could shift the balance to a majority in support.

Limitations

The numerous limitations inherent to our study include the use of a newly constructed questionnaire with design flaws, a modest response rate, the likelihood that respondents were different than nonrespondents, and poor representation of minority respondents. Comparison with data from other studies is limited by differences in study design and sample. These cross-sectional regression analyses allow only statistical prediction of associations, not cause-and-effect conclusions. This survey cannot be generalized to other physician specialties, and it is not appropriate to assume that family physicians are more discontent than other physicians, other professionals, or other workers.

Conclusions

This survey conveys a sobering glimpse of the prevalence of discontent among family physicians. Given the daily effects of upheaval in our health care system, perhaps we should be pleased that 2 out of 3 family physicians would still choose the same path if they had to do it over again. However, the next decade may answer whether our health care system can thrive if there is continued erosion of the support and confidence of skilled, committed, and reasonably content primary care physicians.

Acknowledgments

Financial assistance was provided by Medstar Research Institute. Thanks to both Ms. Michelle Appel and Medstar Research Institute for statistical consultation. The opinions expressed are the views of the authors and are not to be construed as reflecting the views of Medstar Health or the Public Health Service.

The 1990s were a time of remarkable upheaval in US health care. The input of physicians and consumers was marginalized in a market-driven system guided by insurers, employers, and government. Well-intentioned managed care concepts that many physicians would endorse were less evident than managed care business practices that sparked rancor and frustration, as well as creating time and income pressures.

Although 80% of physicians believe that the current health care system requires fundamental change,^1,2 there is little agreement about what form that change should take. For example, 25%³ to 40%⁴ of physicians in statewide surveys and 66% of students entering one medical school system⁵ endorse a single-payer national health plan. Despite the lack of consensus about what direction reform should take, there is growing skepticism about the current model. For example, only 21% of surveyed internal medicine residents believe that corporate managed care is the best model for our health care system.⁶ Among students, residents, faculty members, and deans of medical schools, 52% favor a single-payer system, while 22% favor a corporate managed care system.⁷

Coincident with marked changes in American health care, the popular press have reported deteriorating morale and career satisfaction among physicians.^8,9 For family physicians there are mixed but generally positive data about career satisfaction. Survey data report that only 52% of young family physicians would definitely again choose to become a physician,10 yet from 70%10 to 74%¹¹ would again choose family practice as a specialty. In a 1996 survey⁴ (at a time when both family practice residency applicants and salaries for new graduates peaked), 82% of family physicians reported satisfaction with their careers; 75% stated that they would again select family practice as a specialty; and 65% reported satisfaction with their compensation.⁴

Since 1996, annual surveys of the family practice residents in our own training program have shown a steady decline in the number who would go to medical school again (from 73% to 50% between 1996 and 1999). We suspected that our residents were reacting to changes and uncertainties similar to those faced by all family physicians. This prompted us to survey practicing family physicians about their experiences with and attitudes toward their practice and the US health care system.

Methods

We designed a 41-item questionnaire that focused on family physicians’ observations related to corporate managed health care and personal perception of their own well-being and career. Family physicians were asked to rate each item on a 5-point scale (where 1=strongly disagree, and 5=strongly agree). Additional items focused on demographic information, nature of practice, presence of a primary relationship, presence of children in the home, political inclination, extent of managed care participation, and both the nature and perceived adequacy of compensation. The final questionnaire was the product of pretesting and feedback by 6 of the family practice faculty members from our residency training program and consultation with researchers.

We selected a random national sample of 800 practicing members of the American Academy of Family Physicians (AAFP) in June 1999. A 6-page survey was mailed to each family physician, accompanied by a cover letter requesting participation and a postcard by which a summary of the final survey results could be requested. We sent a second mailing 5 weeks later to those who had not yet responded.

We examined the data from all 41 items that were rated on a Likert scale using principal component factor analysis and varimax rotation with Kaiser normalization.¹² The 4 factors of highest reliability (a >0.650) and highest eigenvalue were chosen post hoc as dependent variables (ie, attitudes about managed care, satisfaction with career, workload stress, and satisfaction with personal life) and analyzed by stepwise regression using all remaining factors and personal and practice variables as independent variables.

Results

A total of 361 completed surveys were returned; 16 surveys were returned as undeliverable (response rate=46%). The mean physician age was 46 years (range=27-77 years); 74% were men, 26% were women. Most were white/non-Hispanic; only 1.1% were Hispanic, 1.7%, African-American, and 4.8%, Asian-American. Practice characteristics are shown in Table 1. The characteristics of these respondents were very similar to those reported for the AAFP membership in December 2000.¹³ The exceptions were that minority family physicians were underrepresented among respondents to our survey and that respondents averaged nearly 5 more years in practice than the AAFP membership mean.

Among this sample of family physicians, 51% received all of their income by salary. Only 10% reported that more than half their income was capitated; 72% still received no income through capitation. The respondents reported working a mean of 51 hours per week; however, 33% reported averaging more than 60 hours, and 7% averaged more than 70 hours. They had spent a mean of 15.3 years in practice (range = 1 to 51 years).

Only 55% of respondents felt they were appropriately compensated financially. Family physicians who were satisfied with their financial compensation reported more favorable views of managed care (F=10.55; df=1; P <.001), greater career satisfaction (F=27.07; df=1; P <.0001) and lower workload stress (F=10.42; df=1; P <.001).

Regarding their personal lives, 91% of these family physicians reported they were currently involved in a relationship, and 65% had children in their home. Politically, 30% identified themselves as democrats; 46%, republicans; and 24%, independents.

The responses to the questionnaire are shown in Table 2,Table 2b,Table 2c,. The results of stepwise regression models are presented in Table 3*Table w2.

Attitudes About Managed Care

The responses to items composing this factor show that this sample of family physicians was sharply critical of corporate managed care. For example, only 7% agreed or strongly agreed that corporate managed care is the best way to provide the health care America needs at a cost that society can afford.

The regression analyses summarized in Table 3 suggest that overall attitudes about managed care were predicted primarily by a cluster of 3 factors that tap into other concerns about managed care, reform, and accountability. Also, positive attitudes about managed care were best predicted by greater career satisfaction, a willingness to let others worry about the business of medicine, support of hospitalists, having children at home, and democratic political inclinations.

Satisfaction with Career

The response to items composing this factor suggests evidence of discontent among a significant minority of these family physicians. For example, knowing what they know now, if they could start over 63% would go to medical school again, and 66% would choose family practice again. Overall, 59% said they were satisfied with their professional lives.

In regression analyses, family physicians who were retrospectively satisfied with residency training, felt that they were appropriately compensated for their current work, and were satisfied with their personal lives were most likely to be satisfied with their careers. Both a positive attitude about managed care and reform-mindedness each independently predicted career satisfaction.

A decreased workload stress factor predicted increased career satisfaction; however, longer work hours were reported by those more satisfied with their careers. Women reported less satisfaction with their careers. Given their covariance, it is reasonable to view these variables as a cluster that together predicts 55% of the variance in career-satisfaction ratings.

Workload Stress

The response to items composing this factor suggests that a significant number of respondents felt their practice was very stressful. For example, only 9% reported that they are not overwhelmed by paperwork and administrative requirements, and 48% said that they had personally experienced more stress-related symptoms within the past year. Regression analyses suggested that greater workload stress was significantly predicted by dissatisfaction with either career or personal life, longer work hours, being female, and greater reform-mindedness.

Satisfaction with Personal Life

In their personal lives, 85% of these family physicians felt supported by a primary relationship, their family, or a social network. Healthy lifestyle habits that might help to buffer stress are reported by 52%. Overall, 71% said they were satisfied with their personal lives. As provided in Table 3, regression analyses showed that greater workload stress and longer work hours each predicted less satisfaction with personal life. Overall career satisfaction and being in a primary relationship predicted greater satisfaction with personal life. Having children in the home did not predict greater satisfaction with personal life.

Discussion

The optimism and career satisfaction reported among the majority of family physicians during the mid-1990s may have eroded in subsequent years. In this national survey in 1999, 63% would again choose to go to medical school, and 66% would again choose family practice. Only 55% were satisfied that they are receiving appropriate financial compensation, and 31% stated that they were “tired of being a doctor” and worried that they were “burning out.”

Although it is unlikely that financial incentives are a primary motivator for the decision to choose family practice, satisfaction with financial compensation was significantly associated with greater career satisfaction and lower work stress.

Nearly half of these family physicians reported that they had personally experienced a recent increase in stress-related symptoms. Many did not feel in control of the stressors they faced in daily practice, and most were overwhelmed by the paperwork and administrative requirements. It is no surprise that career satisfaction and workload stress are inversely related. However, it is less clear whether escalating work stress usually erodes career satisfaction or whether those who are less enamored with their career choice are more likely to experience practice demands as stressful.

Higher reported work hours were associated with greater career satisfaction as well as with higher perceived workload stress and less satisfaction with personal life. Although family physicians working the most hours may be doing so because they love what they are doing, long hours take their toll. Forty-four percent of respondents said they were working so many hours that their health or relationships were adversely affected. Several respondents wrote in the margins that they had opted for 50% to 80% of full time, because it was the only way to make practice tolerable. However, such a work arrangement often constitutes a full-time job by most standards, but for reduced pay.

Only 12% of the respondents felt that residency training had provided good training in practice administration. Despite recent efforts to create more relevant practice management curricula, ratings of this item did not vary significantly by years in practice, suggesting that most new graduates still do not feel prepared for the business of medicine.

Previous research has clearly shown that increased demands coupled with loss of control and a lack of safe outlets for frustration can lead to increased stress.¹⁴ This is the situation in which many of today’s family physicians find themselves. Residency training programs and state professional associations can play a significant role by speaking with candor about physician stress; integrating programs on stress management, career goals, and values clarification; and encouraging advocacy for constructive change of the health care system. Training programs must contemplate which changes are likely to maximize the number of graduates who become successful, satisfied practitioners in the current health care environment.

This sample of family physicians was sharply critical of a health care system driven by corporate managed care. Only 7% of family physicians agreed that corporate managed care is the best way to provide the health care America needs at a cost society can afford. Only 20% believed that managed care had accomplished improvements in both accountability and quality of care. Forty-two percent of these family physicians reported bad patient outcomes they perceived as attributable to managed care business processes. Despite the improved access touted by behavioral health carve-outs, 76% of family physicians report that obtaining good-quality behavioral or mental health treatment for their patients has gotten more difficult as such corporate entities have dominated that market.

There was virtual consensus for specific reforms. More than 85% of these family physicians agreed that all Americans should have access to redress for damages when managed care agents inappropriately delay or deny treatment and that managed care medical directors should be accountable to state boards of physician quality assurance for their decisions about patient care. Only 36% unequivocally supported a government-administered universal national health plan; however, an additional 21% who were equivocal suggested that growing discontent with the status quo could shift the balance to a majority in support.

Limitations

The numerous limitations inherent to our study include the use of a newly constructed questionnaire with design flaws, a modest response rate, the likelihood that respondents were different than nonrespondents, and poor representation of minority respondents. Comparison with data from other studies is limited by differences in study design and sample. These cross-sectional regression analyses allow only statistical prediction of associations, not cause-and-effect conclusions. This survey cannot be generalized to other physician specialties, and it is not appropriate to assume that family physicians are more discontent than other physicians, other professionals, or other workers.

Conclusions

This survey conveys a sobering glimpse of the prevalence of discontent among family physicians. Given the daily effects of upheaval in our health care system, perhaps we should be pleased that 2 out of 3 family physicians would still choose the same path if they had to do it over again. However, the next decade may answer whether our health care system can thrive if there is continued erosion of the support and confidence of skilled, committed, and reasonably content primary care physicians.

Acknowledgments

Financial assistance was provided by Medstar Research Institute. Thanks to both Ms. Michelle Appel and Medstar Research Institute for statistical consultation. The opinions expressed are the views of the authors and are not to be construed as reflecting the views of Medstar Health or the Public Health Service.

Neonatal circumcision is one of the most common surgical procedures performed in the United States.^1,2 Neonates clearly perceive pain during this procedure;³ however, many physicians still do not offer analgesia or anesthesia. The pain of neonatal circumcision has measurable physiologic consequences (eg, pulse oximetry).⁴ Also, this early painful experience may have sustained effects on the neonate some detectable months into the future.^5-9

Previous studies have demonstrated dorsal penile nerve block (DPNB), subcutaneous ring block, and some topical anesthetic formulations (eg, eutectic mixture of local anesthetics cream) to be effective.^10-16 The barriers to offering analgesia or anesthesia are related to a variety of factors, such as new skill acquisition, fear of complications, or inconvenience of technique.^17-19 The circumcision policy statement issued by the American Academy of Pediatrics in 1999 acknowledged that analgesia was safe and effective and that adequate pain relief should be provided.²⁰

Other investigators have linked the use of concentrated sugar solutions with decreased pain activity in neonates undergoing mildly painful procedures, including circumcision. The administration of glucose for pain relief is thought to be because of the induction of endorphin production. Two mL of 12%, 25%, and 50% sugar solutions before heel stick blood collection caused a 50% decrease in the total crying time for infants, compared with those who received only sterile water.^21,22 The improved pain tolerance is blocked by the administration of naloxone.^21-23 However, none of this work has directly compared the use of an oral sugar solution with an established analgesic or anesthetic technique. We directly compared the analgesic properties of sterile water (placebo), concentrated oral glucose (50% dextrose solution [D50]), and DPNB.

Methods

We undertook a randomized double-blind placebo-controlled clinical trial to test the hypothesis that 2 mL of D50 would reduce the total crying time during the circumcision by 50%. One group received placebo (sterile water); a second group was given D50; and the third group received a DPNB. A power analysis with an a of 0.05 and a b of 0.80 revealed that 16 patients in each of the 3 arms would be adequate to detect a 50% difference in the percentage of the procedure time the infant was crying. After approval by the Human Subjects Review Committee, a total of 71 patients were enrolled from November 1, 1996, through March 13, 1997.

Subjects were chosen from all male live births at Naval Hospital Bremerton during the study period. Exclusion criteria included any high-risk characteristics as determined by the medical staff. Of 232 male children born during the study interval, 162 underwent circumcision, and 71 were included in our study after obtaining informed consent for the procedure and participation in the study. Fifty parents refused participation. Forty-one neonates were not included because of the unavailability of the investigators.

The primary outcome for this study was the percentage of the procedure time the infant spent crying. Secondary outcomes are the percentage change in heart rate from baseline, the percentage of oxygen saturation, and the modified behavioral pain scale (MBPS) score at 30-second intervals Table 1.

The infants were randomized (by computer modeling) to 1 of the 3 arms of the study: 24 received 2 mL of sterile water orally; 24 underwent DPNB; and 23 received 2 mL of D50 orally. The DPNB was carried out in the usual manner. The base of the penis in all patients was covered with a sterile 2 × 2 gauze pad secured with tape to obscure evidence of a DPNB. All circumcisions were performed using a Gomco clamp (Allied Healthcare Products, Inc; St. Louis, Mo) and begun between 2 and 6 minutes of the pain relief intervention.

We determined pulse oximetry and heart rate at baseline (after pain relief intervention but before beginning circumcision) and at 1-minute intervals during the procedure. Total procedure time and crying time were assessed later by viewing a videotape of the infant. Pain behavior scores were also independently determined at 30-second intervals by each investigator during videotape reviews using the MBPS.²⁴

Continuous variables were compared using analysis of variance (ANOVA) and the Student t test. We compared pain behavior scale determinations at each 30-second mark with a 3-way ANOVA using time as repeated measures and a Scheffe test for pair-wise comparisons between groups.

Results

The baseline values are shown in Table 2. The DPNB group differed at baseline with respect to the D50 group in time since last fed and heart rate. No other significant differences were noted.

Table 3 shows the data obtained from the circumcision procedures. Mean heart rate differed significantly among the 3 groups, with the D50 group having the highest mean heart rate. However, the percentage increase in heart rate did not differ between the D50 group and the placebo group. No other differences were noted between the D50 group and the placebo group. Mean pulse oximetry measurements and percentage of crying time also differed significantly between the DPNB group and the other 2 groups.

The data from the modified behavioral pain scale confirmed the findings from the crying time, heart rate, and pulse oximetry measurements. The 3-way ANOVA test showed that the MBPS score varied significantly by patient group and by the time when it was measured (P <.001). The Scheffe test for pair-wise comparisons showed that the DPNB group exhibited significantly less pain behavior than either the placebo or dextrose groups (P <.001).

Discussion

Administration of concentrated dextrose solution before circumcision does not offer adequate analgesia. For all pain-related measures (heart rate, pulse oximetry, crying time, and MBPS), there were no statistically significant differences detected between the placebo group and the D50 group.

The baseline differences between the 3 groups studied include the time since last feeding and heart rate. The time since last feeding does not seem to be clinically significant. The higher heart rate in the D50 group may be a result of the high glucose load. The findings of the experiment should not be affected by these differences.

The DPNB group showed significant differences in heart rate, pulse oximetry, crying time, and pain scale scores compared with the other 2 groups during circumcision. This fact reinforces that the study size was adequately powered to detect differences in these parameters and that DPNB provides pain relief for neonatal circumcision. No comparison was done using a topical agent or the circumferential ring block, although previous research has shown these methods to be similar, or in some cases superior, to DPNB.¹⁴

Our results differ from those of Herschel and colleagues,²³ who found a benefit from sucrose administered through a pacifier. Their study compared changes in heart rate, oxygen saturation, and pain-related behavior among 3 study groups: no treatment, DPNB, and delivery of a 50% sucrose solution through a nipple held in place during the procedure. They found significant improvement in pain measures with the sucrose pacifier group. A potential flaw in this study is use of a nipple to deliver the sucrose. The suckling action of the infant on the pacifier alone may have produced analgesia. Our study eliminates the potential effects of suckling on pain-related behavior and physiologic responses.

Limitations

One drawback to our study concerns the use of D50 versus a concentrated sucrose solution. We chose D50 because of its ready availability, compared with a nonstandard solution like sucrose, and because in animal studies the specific type of oral solution used (eg, milk, various sugars, fats) did not alter the measured analgesic effect.²⁵ Further study using a concentrated sucrose solution should be considered before dismissing this form of analgesia entirely.

Conclusions

Future study comparing concentrated sucrose, DPNB, or superficial ring block, and topical local anesthetics would further clarify the issues brought up by our study. Also, the antinociceptive properties of suckling during circumcision should be evaluated. For now, the readily available concentrated glucose solutions, such as 50% dextrose, do not offer any advantage over placebo in relieving the pain associated with neonatal circumcision, and are inferior to a DPNB.

Neonatal circumcision is one of the most common surgical procedures performed in the United States.^1,2 Neonates clearly perceive pain during this procedure;³ however, many physicians still do not offer analgesia or anesthesia. The pain of neonatal circumcision has measurable physiologic consequences (eg, pulse oximetry).⁴ Also, this early painful experience may have sustained effects on the neonate some detectable months into the future.^5-9

Previous studies have demonstrated dorsal penile nerve block (DPNB), subcutaneous ring block, and some topical anesthetic formulations (eg, eutectic mixture of local anesthetics cream) to be effective.^10-16 The barriers to offering analgesia or anesthesia are related to a variety of factors, such as new skill acquisition, fear of complications, or inconvenience of technique.^17-19 The circumcision policy statement issued by the American Academy of Pediatrics in 1999 acknowledged that analgesia was safe and effective and that adequate pain relief should be provided.²⁰

Other investigators have linked the use of concentrated sugar solutions with decreased pain activity in neonates undergoing mildly painful procedures, including circumcision. The administration of glucose for pain relief is thought to be because of the induction of endorphin production. Two mL of 12%, 25%, and 50% sugar solutions before heel stick blood collection caused a 50% decrease in the total crying time for infants, compared with those who received only sterile water.^21,22 The improved pain tolerance is blocked by the administration of naloxone.^21-23 However, none of this work has directly compared the use of an oral sugar solution with an established analgesic or anesthetic technique. We directly compared the analgesic properties of sterile water (placebo), concentrated oral glucose (50% dextrose solution [D50]), and DPNB.

Methods

We undertook a randomized double-blind placebo-controlled clinical trial to test the hypothesis that 2 mL of D50 would reduce the total crying time during the circumcision by 50%. One group received placebo (sterile water); a second group was given D50; and the third group received a DPNB. A power analysis with an a of 0.05 and a b of 0.80 revealed that 16 patients in each of the 3 arms would be adequate to detect a 50% difference in the percentage of the procedure time the infant was crying. After approval by the Human Subjects Review Committee, a total of 71 patients were enrolled from November 1, 1996, through March 13, 1997.

Subjects were chosen from all male live births at Naval Hospital Bremerton during the study period. Exclusion criteria included any high-risk characteristics as determined by the medical staff. Of 232 male children born during the study interval, 162 underwent circumcision, and 71 were included in our study after obtaining informed consent for the procedure and participation in the study. Fifty parents refused participation. Forty-one neonates were not included because of the unavailability of the investigators.

The primary outcome for this study was the percentage of the procedure time the infant spent crying. Secondary outcomes are the percentage change in heart rate from baseline, the percentage of oxygen saturation, and the modified behavioral pain scale (MBPS) score at 30-second intervals Table 1.

The infants were randomized (by computer modeling) to 1 of the 3 arms of the study: 24 received 2 mL of sterile water orally; 24 underwent DPNB; and 23 received 2 mL of D50 orally. The DPNB was carried out in the usual manner. The base of the penis in all patients was covered with a sterile 2 × 2 gauze pad secured with tape to obscure evidence of a DPNB. All circumcisions were performed using a Gomco clamp (Allied Healthcare Products, Inc; St. Louis, Mo) and begun between 2 and 6 minutes of the pain relief intervention.

We determined pulse oximetry and heart rate at baseline (after pain relief intervention but before beginning circumcision) and at 1-minute intervals during the procedure. Total procedure time and crying time were assessed later by viewing a videotape of the infant. Pain behavior scores were also independently determined at 30-second intervals by each investigator during videotape reviews using the MBPS.²⁴

Continuous variables were compared using analysis of variance (ANOVA) and the Student t test. We compared pain behavior scale determinations at each 30-second mark with a 3-way ANOVA using time as repeated measures and a Scheffe test for pair-wise comparisons between groups.

Results

The baseline values are shown in Table 2. The DPNB group differed at baseline with respect to the D50 group in time since last fed and heart rate. No other significant differences were noted.

Table 3 shows the data obtained from the circumcision procedures. Mean heart rate differed significantly among the 3 groups, with the D50 group having the highest mean heart rate. However, the percentage increase in heart rate did not differ between the D50 group and the placebo group. No other differences were noted between the D50 group and the placebo group. Mean pulse oximetry measurements and percentage of crying time also differed significantly between the DPNB group and the other 2 groups.

The data from the modified behavioral pain scale confirmed the findings from the crying time, heart rate, and pulse oximetry measurements. The 3-way ANOVA test showed that the MBPS score varied significantly by patient group and by the time when it was measured (P <.001). The Scheffe test for pair-wise comparisons showed that the DPNB group exhibited significantly less pain behavior than either the placebo or dextrose groups (P <.001).

Discussion

Administration of concentrated dextrose solution before circumcision does not offer adequate analgesia. For all pain-related measures (heart rate, pulse oximetry, crying time, and MBPS), there were no statistically significant differences detected between the placebo group and the D50 group.

The baseline differences between the 3 groups studied include the time since last feeding and heart rate. The time since last feeding does not seem to be clinically significant. The higher heart rate in the D50 group may be a result of the high glucose load. The findings of the experiment should not be affected by these differences.

The DPNB group showed significant differences in heart rate, pulse oximetry, crying time, and pain scale scores compared with the other 2 groups during circumcision. This fact reinforces that the study size was adequately powered to detect differences in these parameters and that DPNB provides pain relief for neonatal circumcision. No comparison was done using a topical agent or the circumferential ring block, although previous research has shown these methods to be similar, or in some cases superior, to DPNB.¹⁴

Our results differ from those of Herschel and colleagues,²³ who found a benefit from sucrose administered through a pacifier. Their study compared changes in heart rate, oxygen saturation, and pain-related behavior among 3 study groups: no treatment, DPNB, and delivery of a 50% sucrose solution through a nipple held in place during the procedure. They found significant improvement in pain measures with the sucrose pacifier group. A potential flaw in this study is use of a nipple to deliver the sucrose. The suckling action of the infant on the pacifier alone may have produced analgesia. Our study eliminates the potential effects of suckling on pain-related behavior and physiologic responses.

Limitations

One drawback to our study concerns the use of D50 versus a concentrated sucrose solution. We chose D50 because of its ready availability, compared with a nonstandard solution like sucrose, and because in animal studies the specific type of oral solution used (eg, milk, various sugars, fats) did not alter the measured analgesic effect.²⁵ Further study using a concentrated sucrose solution should be considered before dismissing this form of analgesia entirely.

Conclusions

Future study comparing concentrated sucrose, DPNB, or superficial ring block, and topical local anesthetics would further clarify the issues brought up by our study. Also, the antinociceptive properties of suckling during circumcision should be evaluated. For now, the readily available concentrated glucose solutions, such as 50% dextrose, do not offer any advantage over placebo in relieving the pain associated with neonatal circumcision, and are inferior to a DPNB.

Neonatal circumcision is one of the most common surgical procedures performed in the United States.^1,2 Neonates clearly perceive pain during this procedure;³ however, many physicians still do not offer analgesia or anesthesia. The pain of neonatal circumcision has measurable physiologic consequences (eg, pulse oximetry).⁴ Also, this early painful experience may have sustained effects on the neonate some detectable months into the future.^5-9

Previous studies have demonstrated dorsal penile nerve block (DPNB), subcutaneous ring block, and some topical anesthetic formulations (eg, eutectic mixture of local anesthetics cream) to be effective.^10-16 The barriers to offering analgesia or anesthesia are related to a variety of factors, such as new skill acquisition, fear of complications, or inconvenience of technique.^17-19 The circumcision policy statement issued by the American Academy of Pediatrics in 1999 acknowledged that analgesia was safe and effective and that adequate pain relief should be provided.²⁰

Other investigators have linked the use of concentrated sugar solutions with decreased pain activity in neonates undergoing mildly painful procedures, including circumcision. The administration of glucose for pain relief is thought to be because of the induction of endorphin production. Two mL of 12%, 25%, and 50% sugar solutions before heel stick blood collection caused a 50% decrease in the total crying time for infants, compared with those who received only sterile water.^21,22 The improved pain tolerance is blocked by the administration of naloxone.^21-23 However, none of this work has directly compared the use of an oral sugar solution with an established analgesic or anesthetic technique. We directly compared the analgesic properties of sterile water (placebo), concentrated oral glucose (50% dextrose solution [D50]), and DPNB.

Methods

We undertook a randomized double-blind placebo-controlled clinical trial to test the hypothesis that 2 mL of D50 would reduce the total crying time during the circumcision by 50%. One group received placebo (sterile water); a second group was given D50; and the third group received a DPNB. A power analysis with an a of 0.05 and a b of 0.80 revealed that 16 patients in each of the 3 arms would be adequate to detect a 50% difference in the percentage of the procedure time the infant was crying. After approval by the Human Subjects Review Committee, a total of 71 patients were enrolled from November 1, 1996, through March 13, 1997.

Subjects were chosen from all male live births at Naval Hospital Bremerton during the study period. Exclusion criteria included any high-risk characteristics as determined by the medical staff. Of 232 male children born during the study interval, 162 underwent circumcision, and 71 were included in our study after obtaining informed consent for the procedure and participation in the study. Fifty parents refused participation. Forty-one neonates were not included because of the unavailability of the investigators.

The primary outcome for this study was the percentage of the procedure time the infant spent crying. Secondary outcomes are the percentage change in heart rate from baseline, the percentage of oxygen saturation, and the modified behavioral pain scale (MBPS) score at 30-second intervals Table 1.

The infants were randomized (by computer modeling) to 1 of the 3 arms of the study: 24 received 2 mL of sterile water orally; 24 underwent DPNB; and 23 received 2 mL of D50 orally. The DPNB was carried out in the usual manner. The base of the penis in all patients was covered with a sterile 2 × 2 gauze pad secured with tape to obscure evidence of a DPNB. All circumcisions were performed using a Gomco clamp (Allied Healthcare Products, Inc; St. Louis, Mo) and begun between 2 and 6 minutes of the pain relief intervention.

We determined pulse oximetry and heart rate at baseline (after pain relief intervention but before beginning circumcision) and at 1-minute intervals during the procedure. Total procedure time and crying time were assessed later by viewing a videotape of the infant. Pain behavior scores were also independently determined at 30-second intervals by each investigator during videotape reviews using the MBPS.²⁴

Continuous variables were compared using analysis of variance (ANOVA) and the Student t test. We compared pain behavior scale determinations at each 30-second mark with a 3-way ANOVA using time as repeated measures and a Scheffe test for pair-wise comparisons between groups.

Results

The baseline values are shown in Table 2. The DPNB group differed at baseline with respect to the D50 group in time since last fed and heart rate. No other significant differences were noted.

Table 3 shows the data obtained from the circumcision procedures. Mean heart rate differed significantly among the 3 groups, with the D50 group having the highest mean heart rate. However, the percentage increase in heart rate did not differ between the D50 group and the placebo group. No other differences were noted between the D50 group and the placebo group. Mean pulse oximetry measurements and percentage of crying time also differed significantly between the DPNB group and the other 2 groups.

The data from the modified behavioral pain scale confirmed the findings from the crying time, heart rate, and pulse oximetry measurements. The 3-way ANOVA test showed that the MBPS score varied significantly by patient group and by the time when it was measured (P <.001). The Scheffe test for pair-wise comparisons showed that the DPNB group exhibited significantly less pain behavior than either the placebo or dextrose groups (P <.001).

Discussion

Administration of concentrated dextrose solution before circumcision does not offer adequate analgesia. For all pain-related measures (heart rate, pulse oximetry, crying time, and MBPS), there were no statistically significant differences detected between the placebo group and the D50 group.

The baseline differences between the 3 groups studied include the time since last feeding and heart rate. The time since last feeding does not seem to be clinically significant. The higher heart rate in the D50 group may be a result of the high glucose load. The findings of the experiment should not be affected by these differences.

The DPNB group showed significant differences in heart rate, pulse oximetry, crying time, and pain scale scores compared with the other 2 groups during circumcision. This fact reinforces that the study size was adequately powered to detect differences in these parameters and that DPNB provides pain relief for neonatal circumcision. No comparison was done using a topical agent or the circumferential ring block, although previous research has shown these methods to be similar, or in some cases superior, to DPNB.¹⁴

Our results differ from those of Herschel and colleagues,²³ who found a benefit from sucrose administered through a pacifier. Their study compared changes in heart rate, oxygen saturation, and pain-related behavior among 3 study groups: no treatment, DPNB, and delivery of a 50% sucrose solution through a nipple held in place during the procedure. They found significant improvement in pain measures with the sucrose pacifier group. A potential flaw in this study is use of a nipple to deliver the sucrose. The suckling action of the infant on the pacifier alone may have produced analgesia. Our study eliminates the potential effects of suckling on pain-related behavior and physiologic responses.

Limitations

One drawback to our study concerns the use of D50 versus a concentrated sucrose solution. We chose D50 because of its ready availability, compared with a nonstandard solution like sucrose, and because in animal studies the specific type of oral solution used (eg, milk, various sugars, fats) did not alter the measured analgesic effect.²⁵ Further study using a concentrated sucrose solution should be considered before dismissing this form of analgesia entirely.

Conclusions

Future study comparing concentrated sucrose, DPNB, or superficial ring block, and topical local anesthetics would further clarify the issues brought up by our study. Also, the antinociceptive properties of suckling during circumcision should be evaluated. For now, the readily available concentrated glucose solutions, such as 50% dextrose, do not offer any advantage over placebo in relieving the pain associated with neonatal circumcision, and are inferior to a DPNB.