Jennifer Devoe, MD, MPhil

Primary care is the foundation of the American health care system. The delivery of high-quality primary care contributes to improved health outcomes.^1,2 Patients perceive primary care as an integral aspect of the health care system and appreciate the role of primary care providers in coordinating quality care.³ In addition to coordination of care, continuity with the same health care provider is highly valued by patients.⁴ Family physicians and general practitioners (FP/GPs) play a crucial role in providing coordinated and continuous primary health care. Of Americans reporting an individual provider as their usual source of care in 1996, 62% named a family physician or a general practitioner (compared with 16% naming an internist and 15% naming a pediatrician).⁵

A potential threat to the continued reliance on this vital FP/GP workforce is physician dissatisfaction. Physician dissatisfaction affects patient satisfaction^6-8 and dissatisfied physicians can adversely influence patient behavior (eg, adherence to medical treatment),⁹ leading to a reduction in quality of care. A decrease in satisfaction among physicians can also affect access to care, since it can lead to physician attrition and higher turnover, which in turn can lead to disruption of care and inaccessibility of providers. The cost to hire a new physician is estimated to be $240,000 to $265,000.¹⁰

Dissatisfaction among today’s FP/GPs also has the potential to contribute to future shortages. The extent to which physicians voice dissatisfaction can dissuade medical school graduates from choosing careers in primary care.¹¹ Some concerns are already being raised about a decrease in the number of new doctors seeking residencies in family practice for the fourth consecutive year. Information from the National Resident Matching Program indicates that only 11.2% of US seniors matched in family practice in 2001, compared with 13.6% in 2000.¹² If this downward trend continues, it will exacerbate the problems of access to a usual source of care, especially in areas where the loss of FP/GPs will result in a drastic increase in the number of health professional shortage areas. In 1995, if FP/GPs had been removed from the 2298 US counties considered to have adequate numbers of primary care physicians, 1332 of these urban and rural counties would have been designated as shortage areas. In comparison, the simultaneous removal of internists, pediatricians, and obstetricians from these same counties would have created only 176 whole-county shortage areas.¹³

Most studies reporting physician dissatisfaction have identified high levels of physician concern over a perceived loss of autonomy.^14-17 Additionally, physicians are dissatisfied about the potential adverse effects on patient care resulting from system barriers, including restricted access for patients, increased administrative burdens for providers, and the lack of a comprehensive approach to provision of services.^14-22 These studies, however, have generally been limited to a specific geographic region^14-20 or a specialty group other than FP/GPs.^21,22 In addition, other physician dissatisfaction reports have contained only narrow analyses of how specific factors, such as income, financial incentives, or autonomy, influence satisfaction levels.^23-26

To our knowledge, the possible relationship of physician dissatisfaction with the inability to care for patients has been examined only in limited studies (eg, those that compare capitated and noncapitated care).²⁷ Moreover, few studies have systematically reviewed predictors of dissatisfaction among FP/GPs. In this paper, we report findings related to these important issues using data from a recent national survey in which more than 12,000 primary care physicians and specialists commented on their current experiences as medical professionals in the US health care system. We hypothesized that FP/GPs who identify difficulties in providing quality care to patients also report higher levels of dissatisfaction. We also expected that physician dissatisfaction would relate to access to care, particularly for very needy populations whose care is government regulated. Therefore, we ascertained the extent to which dissatisfied FP/GPs who stay in the workforce are less likely to serve the poor and the elderly by accepting new Medicaid and Medicare patients when compared with their satisfied counterparts.

Methods

Data source

Data for this study were from the Community Tracking Study (CTS) Physician Survey (1996–1997).²⁸ This survey, sponsored by the Robert Wood Johnson Foundation, was part of a major project by the Center for Studying Health System Change, a Washington, DC–based organization affiliated with Mathematica Policy Research, Inc. Information for the survey was collected from a nationally representative sample of nonfederal physicians performing direct patient care.

The sample frame of physicians was obtained from master files of the American Medical Association and the American Osteopathic Association. The sample included office-based and hospital-based physicians who spend at least 20 hours per week in direct patient care in the continental United States. Residents and fellows were excluded, as were physicians in certain specialties such as radiology, anesthesiology, and pathology. The survey followed a complex design of 60 sites supplemented by a small, independently drawn national sample.^28,29 Telephone interviews were conducted with 12,291 physicians from August 1996 to August 1997 with a 65% response rate.^30,31 The rate of nonresponse to individual survey items was very low, typically less than 3%. Primary care physicians were oversampled. The 3166 FP/GPs accounted for 44.3% of primary care doctors surveyed and 25.8% of the total sample.

Study variables

Dependent Variable: Career Satisfaction. The dependent variable for most analyses was medical career dissatisfaction. Respondents were asked, “Thinking very generally about your overall career in medicine, would you say that you are currently very satisfied, somewhat satisfied, somewhat dissatisfied, very dissatisfied, or neither satisfied or dissatisfied?” For comparative analysis, those reporting “neither satisfied or dissatisfied” were eliminated and the 4 remaining responses were collapsed into 2 categories. Physicians who reported feeling very satisfied or somewhat satisfied were classified as “satisfied”; those who reported feeling somewhat dissatisfied or very dissatisfied were classified as “dissatisfied.”

Independent Variables. The explanatory variables of primary interest were indicators to assess physicians’ perceptions of their ability to provide high-quality medical care. This determination was measured by 6 survey questions with 5-point response categories that ranged from “strongly agree” to “strongly disagree.” Table 1 presents the measures of quality of care. Physician dissatisfaction was also potentially influenced by other factors assessed in the survey. Analyses were statistically controlled for the personal, professional, and practice characteristics that appear in Table 2 to examine the relationship of physicians’ dissatisfaction with the perception of their ability to provide high-quality care.

TABLE 1
PATIENT CARE CHARACTERISTICS ASSOCIATED WITH FP/GP DISSATISFACTION (N = 3166)

Analytical strategy

Two sets of logistic regression were performed. In the first, adjusted odds ratios were derived to measure the association of personal, professional, and practice characteristics with dissatisfaction (Table 2). Next, the relationship of dissatisfaction with each of the indicators of quality of care was assessed with 6 separate multivariate logistic regression procedures (Table 1). The adjusted odds ratios in Table 1 are the products of that analysis and represent the association of perceived ability to deliver quality of care after controlling for the effects of personal, professional, and practice variables. SUDAAN software, version 7.5.3 (Research Triangle Institute, Research Triangle Park, NC), was used to conduct statistical tests and make national estimates with variance adjustment for the complex survey sample design and physician nonresponse.

TABLE 2
PERSONAL, PROFESSIONAL, AND PRACTICE-RELATED FACTORS ASSOCIATED WITH FP/GP DISSATISFACTION (N = 3166)

Results

Nearly 18% of physicians report being dissatisfied with a career in medicine. The rate of dissatisfaction among FP/GPs is similar to that for all physicians, specialists, and primary care physicians as a group. However, there is some variability among primary care specialties, with internists reporting more dissatisfaction (chi-square = 14.8, P < .01) and pediatricians (chi-square = 25.9, P < .01) reporting less dissatisfaction than FP/GPs (Table 3).

TABLE 3
EXTENT OF PHYSICIAN DISSATISFACTION

Factors associated with FP/GP dissatisfaction

Many characteristics were associated with the dissatisfaction reported by 17.6% of FP/GPs. The associated characteristics are included in 3 domains. The first 2 domains, personal/professional and practice characteristics, reveal significant factors associated with dissatisfaction (Table 2). The data in the third domain, patient care characteristics, represent results after we had statistically controlled for all factors in the first 2 (Table 1).

Personal/Professional Characteristics. A higher level of dissatisfaction was related to being older; only 10.1% of physicians younger than 35 years of age reported dissatisfaction versus 25.1% of physicians aged 55 to 64 years (odds ratio [OR] 2.46; 95% confidence interval [CI], 1.56-3.88). FP/GPs more likely to be dissatisfied were those who had osteopathic training and those who had been graduated from foreign medical schools. Levels of dissatisfaction were also higher among FP/GPs earning less than $100,000 per year.

Practice Characteristics. Physicians who fully owned their practice were more likely to express dissatisfaction with their careers than were physicians who either shared ownership or did not own their practice (OR 1.57; 95% CI, 1.11-2.21). The pattern of dissatisfaction related to gatekeeping (ie, providing permission for their patients to seek specialty care) was similar to that related to income. FP/GPs serving as gatekeepers for less than 10% or more than 30% of their patients were the most dissatisfied.

Patient Care Characteristics. After we had controlled for the effects of personal, professional, and practice characteristics, we found that FP/GP career dissatisfaction was, without exception, consistently and strongly associated with a perceived inability to provide high-quality care as assessed by physician responses to each of 6 statements (Table 1). Dissatisfied physicians were much more likely to “disagree strongly” than. to “agree strongly” with the statements about clinical freedom (OR 7.89; 95% CI, 4.86-12.83), continuity of care (OR 7.11; 95% CI, 4.90-10.33), clinical decisions free of financial penalties (OR 4.44; 95% CI, 3.13-6.31), adequacy of time with patients (OR 4.42; 95% CI, 2.84-6.87), ability to provide high-quality care (OR 4.26; 95% CI, 2.88-6.31) and sufficient communication with specialists (OR 3.57; 95% CI, 2.20-5.80). The most notable differences found between dissatisfied and satisfied FP/GPs were related to a lack of clinical freedom and difficulty maintaining continuing relationships with patients.

Physician dissatisfaction influences medicare and medicaid care

Dissatisfaction correlates with the percentage of physicians who are willing to care for Medicare and Medicaid patients. A lower percentage of dissatisfied FP/GPs are accepting all new Medicaid patients than are their satisfied counterparts (34.6% vs 43.4%; P < .01); and a higher percentage of dissatisfied FP/GPs are taking no new Medicaid patients (33.5% vs 23.7%; P < .01). Similarly, a higher percentage of dissatisfied FP/GPs are accepting no new Medicare patients (11.3% vs 8.6%; P = .04) (Table 4).

TABLE 4
RELATIONSHIP OF FP/GP DISSATISFACTION TO ACCESS FOR MEDICAID AND MEDICARE PATIENTS

Discussion

A substantial proportion of family physicians, approximately 1 in 5, were dissatisfied with their careers in 1996–1997. Associated characteristics of the dissatisfied group were older age, osteopathic training, and graduation from a foreign medical school. Neither type nor location of practice was a factor, although being a full owner of the practice was associated with greater dissatisfaction. Physicians earning less than $100,000 per year and FP/GPs for whom less than 10% or more than 30% of patients were in gatekeeping arrangements were more dissatisfied.

The strongest factors associated with dissatisfaction, however, were not personal or practice characteristics but the perceptions of family physicians about their ability to take good care of their patients. After we had controlled for personal and practice characteristics, dissatisfaction was much more likely when the family physicians felt they did not have (1) the freedom to make clinical decisions that met their patients’ needs, (2) a sufficient level of communication with specialists, (3) enough time with their patients, (4) the ability to provide high-quality patient care, (5) the freedom to make clinical decisions without financial conflicts of interest, or (6) the ability to maintain continuing relationships with their patients. More than half of FP/GPs who strongly disagreed with the statement “I have the freedom to make clinical decisions that meet my patients’ needs” were dissatisfied with their medical career.

These findings are consistent with previous findings concerning physician autonomy and the widespread backlash against constraints associated with managed care and gatekeeping. The findings draw attention from financial considerations toward clinical decision making as a critical factor in physicians’ career satisfaction. Understanding the basis of physician dissatisfaction is important because of the adverse effects of such dissatisfaction. It is difficult to imagine patients preferring to see a dissatisfied physician or to envision a visit with a dissatisfied FP/GP as superior to one with a satisfied physician. In addition, this analysis specifically demonstrates that dissatisfaction among family physicians can negatively affect groups of patients by impeding access to care for Medicaid and Medicare patients. Perhaps the key implication of these findings is the need for serious efforts to revise practice arrangements so that FP/GPs can make the best possible decisions for their patients.

Limitations

There are important limitations to our analysis. The CTS Physician Survey is cross-sectional. While we do not know whether these physicians are more or less satisfied than they were in the past, recent evidence from surveys of primary care physicians in Massachusetts suggests that dissatisfaction has increased since 1986.¹⁷ As in all surveys, responses are subject to reporting error and response bias not accounted for by statistical adjustments. Our findings are associations between variables and do not establish causal relationships.

Conclusions

The finding that family physician dissatisfaction, after study results are controlled for personal and practice variables, is associated most strongly with a perceived inability to care for patients raises significant concerns. Dissatisfaction among a large proportion of family physicians threatens the well-being of patients. Given the extent to which the US health care system relies on family physicians, understanding why these physicians are dissatisfied and responding to these problems are important. This cross-sectional snapshot of dissatisfaction among family physicians suggests that patients would benefit from strategies that support rather than disrupt their ongoing relationships with family physicians and that permit their family physician to spend enough time with them to make decisions that are not constrained by financial or other conflicts of interest.

Primary care is the foundation of the American health care system. The delivery of high-quality primary care contributes to improved health outcomes.^1,2 Patients perceive primary care as an integral aspect of the health care system and appreciate the role of primary care providers in coordinating quality care.³ In addition to coordination of care, continuity with the same health care provider is highly valued by patients.⁴ Family physicians and general practitioners (FP/GPs) play a crucial role in providing coordinated and continuous primary health care. Of Americans reporting an individual provider as their usual source of care in 1996, 62% named a family physician or a general practitioner (compared with 16% naming an internist and 15% naming a pediatrician).⁵

A potential threat to the continued reliance on this vital FP/GP workforce is physician dissatisfaction. Physician dissatisfaction affects patient satisfaction^6-8 and dissatisfied physicians can adversely influence patient behavior (eg, adherence to medical treatment),⁹ leading to a reduction in quality of care. A decrease in satisfaction among physicians can also affect access to care, since it can lead to physician attrition and higher turnover, which in turn can lead to disruption of care and inaccessibility of providers. The cost to hire a new physician is estimated to be $240,000 to $265,000.¹⁰

Dissatisfaction among today’s FP/GPs also has the potential to contribute to future shortages. The extent to which physicians voice dissatisfaction can dissuade medical school graduates from choosing careers in primary care.¹¹ Some concerns are already being raised about a decrease in the number of new doctors seeking residencies in family practice for the fourth consecutive year. Information from the National Resident Matching Program indicates that only 11.2% of US seniors matched in family practice in 2001, compared with 13.6% in 2000.¹² If this downward trend continues, it will exacerbate the problems of access to a usual source of care, especially in areas where the loss of FP/GPs will result in a drastic increase in the number of health professional shortage areas. In 1995, if FP/GPs had been removed from the 2298 US counties considered to have adequate numbers of primary care physicians, 1332 of these urban and rural counties would have been designated as shortage areas. In comparison, the simultaneous removal of internists, pediatricians, and obstetricians from these same counties would have created only 176 whole-county shortage areas.¹³

Most studies reporting physician dissatisfaction have identified high levels of physician concern over a perceived loss of autonomy.^14-17 Additionally, physicians are dissatisfied about the potential adverse effects on patient care resulting from system barriers, including restricted access for patients, increased administrative burdens for providers, and the lack of a comprehensive approach to provision of services.^14-22 These studies, however, have generally been limited to a specific geographic region^14-20 or a specialty group other than FP/GPs.^21,22 In addition, other physician dissatisfaction reports have contained only narrow analyses of how specific factors, such as income, financial incentives, or autonomy, influence satisfaction levels.^23-26

To our knowledge, the possible relationship of physician dissatisfaction with the inability to care for patients has been examined only in limited studies (eg, those that compare capitated and noncapitated care).²⁷ Moreover, few studies have systematically reviewed predictors of dissatisfaction among FP/GPs. In this paper, we report findings related to these important issues using data from a recent national survey in which more than 12,000 primary care physicians and specialists commented on their current experiences as medical professionals in the US health care system. We hypothesized that FP/GPs who identify difficulties in providing quality care to patients also report higher levels of dissatisfaction. We also expected that physician dissatisfaction would relate to access to care, particularly for very needy populations whose care is government regulated. Therefore, we ascertained the extent to which dissatisfied FP/GPs who stay in the workforce are less likely to serve the poor and the elderly by accepting new Medicaid and Medicare patients when compared with their satisfied counterparts.

Methods

Data source

Data for this study were from the Community Tracking Study (CTS) Physician Survey (1996–1997).²⁸ This survey, sponsored by the Robert Wood Johnson Foundation, was part of a major project by the Center for Studying Health System Change, a Washington, DC–based organization affiliated with Mathematica Policy Research, Inc. Information for the survey was collected from a nationally representative sample of nonfederal physicians performing direct patient care.

The sample frame of physicians was obtained from master files of the American Medical Association and the American Osteopathic Association. The sample included office-based and hospital-based physicians who spend at least 20 hours per week in direct patient care in the continental United States. Residents and fellows were excluded, as were physicians in certain specialties such as radiology, anesthesiology, and pathology. The survey followed a complex design of 60 sites supplemented by a small, independently drawn national sample.^28,29 Telephone interviews were conducted with 12,291 physicians from August 1996 to August 1997 with a 65% response rate.^30,31 The rate of nonresponse to individual survey items was very low, typically less than 3%. Primary care physicians were oversampled. The 3166 FP/GPs accounted for 44.3% of primary care doctors surveyed and 25.8% of the total sample.

Study variables

Dependent Variable: Career Satisfaction. The dependent variable for most analyses was medical career dissatisfaction. Respondents were asked, “Thinking very generally about your overall career in medicine, would you say that you are currently very satisfied, somewhat satisfied, somewhat dissatisfied, very dissatisfied, or neither satisfied or dissatisfied?” For comparative analysis, those reporting “neither satisfied or dissatisfied” were eliminated and the 4 remaining responses were collapsed into 2 categories. Physicians who reported feeling very satisfied or somewhat satisfied were classified as “satisfied”; those who reported feeling somewhat dissatisfied or very dissatisfied were classified as “dissatisfied.”

Independent Variables. The explanatory variables of primary interest were indicators to assess physicians’ perceptions of their ability to provide high-quality medical care. This determination was measured by 6 survey questions with 5-point response categories that ranged from “strongly agree” to “strongly disagree.” Table 1 presents the measures of quality of care. Physician dissatisfaction was also potentially influenced by other factors assessed in the survey. Analyses were statistically controlled for the personal, professional, and practice characteristics that appear in Table 2 to examine the relationship of physicians’ dissatisfaction with the perception of their ability to provide high-quality care.

TABLE 1
PATIENT CARE CHARACTERISTICS ASSOCIATED WITH FP/GP DISSATISFACTION (N = 3166)

Analytical strategy

Two sets of logistic regression were performed. In the first, adjusted odds ratios were derived to measure the association of personal, professional, and practice characteristics with dissatisfaction (Table 2). Next, the relationship of dissatisfaction with each of the indicators of quality of care was assessed with 6 separate multivariate logistic regression procedures (Table 1). The adjusted odds ratios in Table 1 are the products of that analysis and represent the association of perceived ability to deliver quality of care after controlling for the effects of personal, professional, and practice variables. SUDAAN software, version 7.5.3 (Research Triangle Institute, Research Triangle Park, NC), was used to conduct statistical tests and make national estimates with variance adjustment for the complex survey sample design and physician nonresponse.

TABLE 2
PERSONAL, PROFESSIONAL, AND PRACTICE-RELATED FACTORS ASSOCIATED WITH FP/GP DISSATISFACTION (N = 3166)

Results

Nearly 18% of physicians report being dissatisfied with a career in medicine. The rate of dissatisfaction among FP/GPs is similar to that for all physicians, specialists, and primary care physicians as a group. However, there is some variability among primary care specialties, with internists reporting more dissatisfaction (chi-square = 14.8, P < .01) and pediatricians (chi-square = 25.9, P < .01) reporting less dissatisfaction than FP/GPs (Table 3).

TABLE 3
EXTENT OF PHYSICIAN DISSATISFACTION

Factors associated with FP/GP dissatisfaction

Many characteristics were associated with the dissatisfaction reported by 17.6% of FP/GPs. The associated characteristics are included in 3 domains. The first 2 domains, personal/professional and practice characteristics, reveal significant factors associated with dissatisfaction (Table 2). The data in the third domain, patient care characteristics, represent results after we had statistically controlled for all factors in the first 2 (Table 1).

Personal/Professional Characteristics. A higher level of dissatisfaction was related to being older; only 10.1% of physicians younger than 35 years of age reported dissatisfaction versus 25.1% of physicians aged 55 to 64 years (odds ratio [OR] 2.46; 95% confidence interval [CI], 1.56-3.88). FP/GPs more likely to be dissatisfied were those who had osteopathic training and those who had been graduated from foreign medical schools. Levels of dissatisfaction were also higher among FP/GPs earning less than $100,000 per year.

Practice Characteristics. Physicians who fully owned their practice were more likely to express dissatisfaction with their careers than were physicians who either shared ownership or did not own their practice (OR 1.57; 95% CI, 1.11-2.21). The pattern of dissatisfaction related to gatekeeping (ie, providing permission for their patients to seek specialty care) was similar to that related to income. FP/GPs serving as gatekeepers for less than 10% or more than 30% of their patients were the most dissatisfied.

Patient Care Characteristics. After we had controlled for the effects of personal, professional, and practice characteristics, we found that FP/GP career dissatisfaction was, without exception, consistently and strongly associated with a perceived inability to provide high-quality care as assessed by physician responses to each of 6 statements (Table 1). Dissatisfied physicians were much more likely to “disagree strongly” than. to “agree strongly” with the statements about clinical freedom (OR 7.89; 95% CI, 4.86-12.83), continuity of care (OR 7.11; 95% CI, 4.90-10.33), clinical decisions free of financial penalties (OR 4.44; 95% CI, 3.13-6.31), adequacy of time with patients (OR 4.42; 95% CI, 2.84-6.87), ability to provide high-quality care (OR 4.26; 95% CI, 2.88-6.31) and sufficient communication with specialists (OR 3.57; 95% CI, 2.20-5.80). The most notable differences found between dissatisfied and satisfied FP/GPs were related to a lack of clinical freedom and difficulty maintaining continuing relationships with patients.

Physician dissatisfaction influences medicare and medicaid care

Dissatisfaction correlates with the percentage of physicians who are willing to care for Medicare and Medicaid patients. A lower percentage of dissatisfied FP/GPs are accepting all new Medicaid patients than are their satisfied counterparts (34.6% vs 43.4%; P < .01); and a higher percentage of dissatisfied FP/GPs are taking no new Medicaid patients (33.5% vs 23.7%; P < .01). Similarly, a higher percentage of dissatisfied FP/GPs are accepting no new Medicare patients (11.3% vs 8.6%; P = .04) (Table 4).

TABLE 4
RELATIONSHIP OF FP/GP DISSATISFACTION TO ACCESS FOR MEDICAID AND MEDICARE PATIENTS

Discussion

A substantial proportion of family physicians, approximately 1 in 5, were dissatisfied with their careers in 1996–1997. Associated characteristics of the dissatisfied group were older age, osteopathic training, and graduation from a foreign medical school. Neither type nor location of practice was a factor, although being a full owner of the practice was associated with greater dissatisfaction. Physicians earning less than $100,000 per year and FP/GPs for whom less than 10% or more than 30% of patients were in gatekeeping arrangements were more dissatisfied.

The strongest factors associated with dissatisfaction, however, were not personal or practice characteristics but the perceptions of family physicians about their ability to take good care of their patients. After we had controlled for personal and practice characteristics, dissatisfaction was much more likely when the family physicians felt they did not have (1) the freedom to make clinical decisions that met their patients’ needs, (2) a sufficient level of communication with specialists, (3) enough time with their patients, (4) the ability to provide high-quality patient care, (5) the freedom to make clinical decisions without financial conflicts of interest, or (6) the ability to maintain continuing relationships with their patients. More than half of FP/GPs who strongly disagreed with the statement “I have the freedom to make clinical decisions that meet my patients’ needs” were dissatisfied with their medical career.

These findings are consistent with previous findings concerning physician autonomy and the widespread backlash against constraints associated with managed care and gatekeeping. The findings draw attention from financial considerations toward clinical decision making as a critical factor in physicians’ career satisfaction. Understanding the basis of physician dissatisfaction is important because of the adverse effects of such dissatisfaction. It is difficult to imagine patients preferring to see a dissatisfied physician or to envision a visit with a dissatisfied FP/GP as superior to one with a satisfied physician. In addition, this analysis specifically demonstrates that dissatisfaction among family physicians can negatively affect groups of patients by impeding access to care for Medicaid and Medicare patients. Perhaps the key implication of these findings is the need for serious efforts to revise practice arrangements so that FP/GPs can make the best possible decisions for their patients.

Limitations

There are important limitations to our analysis. The CTS Physician Survey is cross-sectional. While we do not know whether these physicians are more or less satisfied than they were in the past, recent evidence from surveys of primary care physicians in Massachusetts suggests that dissatisfaction has increased since 1986.¹⁷ As in all surveys, responses are subject to reporting error and response bias not accounted for by statistical adjustments. Our findings are associations between variables and do not establish causal relationships.

Conclusions

The finding that family physician dissatisfaction, after study results are controlled for personal and practice variables, is associated most strongly with a perceived inability to care for patients raises significant concerns. Dissatisfaction among a large proportion of family physicians threatens the well-being of patients. Given the extent to which the US health care system relies on family physicians, understanding why these physicians are dissatisfied and responding to these problems are important. This cross-sectional snapshot of dissatisfaction among family physicians suggests that patients would benefit from strategies that support rather than disrupt their ongoing relationships with family physicians and that permit their family physician to spend enough time with them to make decisions that are not constrained by financial or other conflicts of interest.

Primary care is the foundation of the American health care system. The delivery of high-quality primary care contributes to improved health outcomes.^1,2 Patients perceive primary care as an integral aspect of the health care system and appreciate the role of primary care providers in coordinating quality care.³ In addition to coordination of care, continuity with the same health care provider is highly valued by patients.⁴ Family physicians and general practitioners (FP/GPs) play a crucial role in providing coordinated and continuous primary health care. Of Americans reporting an individual provider as their usual source of care in 1996, 62% named a family physician or a general practitioner (compared with 16% naming an internist and 15% naming a pediatrician).⁵

A potential threat to the continued reliance on this vital FP/GP workforce is physician dissatisfaction. Physician dissatisfaction affects patient satisfaction^6-8 and dissatisfied physicians can adversely influence patient behavior (eg, adherence to medical treatment),⁹ leading to a reduction in quality of care. A decrease in satisfaction among physicians can also affect access to care, since it can lead to physician attrition and higher turnover, which in turn can lead to disruption of care and inaccessibility of providers. The cost to hire a new physician is estimated to be $240,000 to $265,000.¹⁰

Dissatisfaction among today’s FP/GPs also has the potential to contribute to future shortages. The extent to which physicians voice dissatisfaction can dissuade medical school graduates from choosing careers in primary care.¹¹ Some concerns are already being raised about a decrease in the number of new doctors seeking residencies in family practice for the fourth consecutive year. Information from the National Resident Matching Program indicates that only 11.2% of US seniors matched in family practice in 2001, compared with 13.6% in 2000.¹² If this downward trend continues, it will exacerbate the problems of access to a usual source of care, especially in areas where the loss of FP/GPs will result in a drastic increase in the number of health professional shortage areas. In 1995, if FP/GPs had been removed from the 2298 US counties considered to have adequate numbers of primary care physicians, 1332 of these urban and rural counties would have been designated as shortage areas. In comparison, the simultaneous removal of internists, pediatricians, and obstetricians from these same counties would have created only 176 whole-county shortage areas.¹³

Most studies reporting physician dissatisfaction have identified high levels of physician concern over a perceived loss of autonomy.^14-17 Additionally, physicians are dissatisfied about the potential adverse effects on patient care resulting from system barriers, including restricted access for patients, increased administrative burdens for providers, and the lack of a comprehensive approach to provision of services.^14-22 These studies, however, have generally been limited to a specific geographic region^14-20 or a specialty group other than FP/GPs.^21,22 In addition, other physician dissatisfaction reports have contained only narrow analyses of how specific factors, such as income, financial incentives, or autonomy, influence satisfaction levels.^23-26

To our knowledge, the possible relationship of physician dissatisfaction with the inability to care for patients has been examined only in limited studies (eg, those that compare capitated and noncapitated care).²⁷ Moreover, few studies have systematically reviewed predictors of dissatisfaction among FP/GPs. In this paper, we report findings related to these important issues using data from a recent national survey in which more than 12,000 primary care physicians and specialists commented on their current experiences as medical professionals in the US health care system. We hypothesized that FP/GPs who identify difficulties in providing quality care to patients also report higher levels of dissatisfaction. We also expected that physician dissatisfaction would relate to access to care, particularly for very needy populations whose care is government regulated. Therefore, we ascertained the extent to which dissatisfied FP/GPs who stay in the workforce are less likely to serve the poor and the elderly by accepting new Medicaid and Medicare patients when compared with their satisfied counterparts.

Methods

Data source

Data for this study were from the Community Tracking Study (CTS) Physician Survey (1996–1997).²⁸ This survey, sponsored by the Robert Wood Johnson Foundation, was part of a major project by the Center for Studying Health System Change, a Washington, DC–based organization affiliated with Mathematica Policy Research, Inc. Information for the survey was collected from a nationally representative sample of nonfederal physicians performing direct patient care.

The sample frame of physicians was obtained from master files of the American Medical Association and the American Osteopathic Association. The sample included office-based and hospital-based physicians who spend at least 20 hours per week in direct patient care in the continental United States. Residents and fellows were excluded, as were physicians in certain specialties such as radiology, anesthesiology, and pathology. The survey followed a complex design of 60 sites supplemented by a small, independently drawn national sample.^28,29 Telephone interviews were conducted with 12,291 physicians from August 1996 to August 1997 with a 65% response rate.^30,31 The rate of nonresponse to individual survey items was very low, typically less than 3%. Primary care physicians were oversampled. The 3166 FP/GPs accounted for 44.3% of primary care doctors surveyed and 25.8% of the total sample.

Study variables

Dependent Variable: Career Satisfaction. The dependent variable for most analyses was medical career dissatisfaction. Respondents were asked, “Thinking very generally about your overall career in medicine, would you say that you are currently very satisfied, somewhat satisfied, somewhat dissatisfied, very dissatisfied, or neither satisfied or dissatisfied?” For comparative analysis, those reporting “neither satisfied or dissatisfied” were eliminated and the 4 remaining responses were collapsed into 2 categories. Physicians who reported feeling very satisfied or somewhat satisfied were classified as “satisfied”; those who reported feeling somewhat dissatisfied or very dissatisfied were classified as “dissatisfied.”

Independent Variables. The explanatory variables of primary interest were indicators to assess physicians’ perceptions of their ability to provide high-quality medical care. This determination was measured by 6 survey questions with 5-point response categories that ranged from “strongly agree” to “strongly disagree.” Table 1 presents the measures of quality of care. Physician dissatisfaction was also potentially influenced by other factors assessed in the survey. Analyses were statistically controlled for the personal, professional, and practice characteristics that appear in Table 2 to examine the relationship of physicians’ dissatisfaction with the perception of their ability to provide high-quality care.

TABLE 1
PATIENT CARE CHARACTERISTICS ASSOCIATED WITH FP/GP DISSATISFACTION (N = 3166)

Analytical strategy

Two sets of logistic regression were performed. In the first, adjusted odds ratios were derived to measure the association of personal, professional, and practice characteristics with dissatisfaction (Table 2). Next, the relationship of dissatisfaction with each of the indicators of quality of care was assessed with 6 separate multivariate logistic regression procedures (Table 1). The adjusted odds ratios in Table 1 are the products of that analysis and represent the association of perceived ability to deliver quality of care after controlling for the effects of personal, professional, and practice variables. SUDAAN software, version 7.5.3 (Research Triangle Institute, Research Triangle Park, NC), was used to conduct statistical tests and make national estimates with variance adjustment for the complex survey sample design and physician nonresponse.

TABLE 2
PERSONAL, PROFESSIONAL, AND PRACTICE-RELATED FACTORS ASSOCIATED WITH FP/GP DISSATISFACTION (N = 3166)

Results

Nearly 18% of physicians report being dissatisfied with a career in medicine. The rate of dissatisfaction among FP/GPs is similar to that for all physicians, specialists, and primary care physicians as a group. However, there is some variability among primary care specialties, with internists reporting more dissatisfaction (chi-square = 14.8, P < .01) and pediatricians (chi-square = 25.9, P < .01) reporting less dissatisfaction than FP/GPs (Table 3).

TABLE 3
EXTENT OF PHYSICIAN DISSATISFACTION

Factors associated with FP/GP dissatisfaction

Many characteristics were associated with the dissatisfaction reported by 17.6% of FP/GPs. The associated characteristics are included in 3 domains. The first 2 domains, personal/professional and practice characteristics, reveal significant factors associated with dissatisfaction (Table 2). The data in the third domain, patient care characteristics, represent results after we had statistically controlled for all factors in the first 2 (Table 1).

Personal/Professional Characteristics. A higher level of dissatisfaction was related to being older; only 10.1% of physicians younger than 35 years of age reported dissatisfaction versus 25.1% of physicians aged 55 to 64 years (odds ratio [OR] 2.46; 95% confidence interval [CI], 1.56-3.88). FP/GPs more likely to be dissatisfied were those who had osteopathic training and those who had been graduated from foreign medical schools. Levels of dissatisfaction were also higher among FP/GPs earning less than $100,000 per year.

Practice Characteristics. Physicians who fully owned their practice were more likely to express dissatisfaction with their careers than were physicians who either shared ownership or did not own their practice (OR 1.57; 95% CI, 1.11-2.21). The pattern of dissatisfaction related to gatekeeping (ie, providing permission for their patients to seek specialty care) was similar to that related to income. FP/GPs serving as gatekeepers for less than 10% or more than 30% of their patients were the most dissatisfied.

Patient Care Characteristics. After we had controlled for the effects of personal, professional, and practice characteristics, we found that FP/GP career dissatisfaction was, without exception, consistently and strongly associated with a perceived inability to provide high-quality care as assessed by physician responses to each of 6 statements (Table 1). Dissatisfied physicians were much more likely to “disagree strongly” than. to “agree strongly” with the statements about clinical freedom (OR 7.89; 95% CI, 4.86-12.83), continuity of care (OR 7.11; 95% CI, 4.90-10.33), clinical decisions free of financial penalties (OR 4.44; 95% CI, 3.13-6.31), adequacy of time with patients (OR 4.42; 95% CI, 2.84-6.87), ability to provide high-quality care (OR 4.26; 95% CI, 2.88-6.31) and sufficient communication with specialists (OR 3.57; 95% CI, 2.20-5.80). The most notable differences found between dissatisfied and satisfied FP/GPs were related to a lack of clinical freedom and difficulty maintaining continuing relationships with patients.

Physician dissatisfaction influences medicare and medicaid care

Dissatisfaction correlates with the percentage of physicians who are willing to care for Medicare and Medicaid patients. A lower percentage of dissatisfied FP/GPs are accepting all new Medicaid patients than are their satisfied counterparts (34.6% vs 43.4%; P < .01); and a higher percentage of dissatisfied FP/GPs are taking no new Medicaid patients (33.5% vs 23.7%; P < .01). Similarly, a higher percentage of dissatisfied FP/GPs are accepting no new Medicare patients (11.3% vs 8.6%; P = .04) (Table 4).

TABLE 4
RELATIONSHIP OF FP/GP DISSATISFACTION TO ACCESS FOR MEDICAID AND MEDICARE PATIENTS

Discussion

A substantial proportion of family physicians, approximately 1 in 5, were dissatisfied with their careers in 1996–1997. Associated characteristics of the dissatisfied group were older age, osteopathic training, and graduation from a foreign medical school. Neither type nor location of practice was a factor, although being a full owner of the practice was associated with greater dissatisfaction. Physicians earning less than $100,000 per year and FP/GPs for whom less than 10% or more than 30% of patients were in gatekeeping arrangements were more dissatisfied.

The strongest factors associated with dissatisfaction, however, were not personal or practice characteristics but the perceptions of family physicians about their ability to take good care of their patients. After we had controlled for personal and practice characteristics, dissatisfaction was much more likely when the family physicians felt they did not have (1) the freedom to make clinical decisions that met their patients’ needs, (2) a sufficient level of communication with specialists, (3) enough time with their patients, (4) the ability to provide high-quality patient care, (5) the freedom to make clinical decisions without financial conflicts of interest, or (6) the ability to maintain continuing relationships with their patients. More than half of FP/GPs who strongly disagreed with the statement “I have the freedom to make clinical decisions that meet my patients’ needs” were dissatisfied with their medical career.

These findings are consistent with previous findings concerning physician autonomy and the widespread backlash against constraints associated with managed care and gatekeeping. The findings draw attention from financial considerations toward clinical decision making as a critical factor in physicians’ career satisfaction. Understanding the basis of physician dissatisfaction is important because of the adverse effects of such dissatisfaction. It is difficult to imagine patients preferring to see a dissatisfied physician or to envision a visit with a dissatisfied FP/GP as superior to one with a satisfied physician. In addition, this analysis specifically demonstrates that dissatisfaction among family physicians can negatively affect groups of patients by impeding access to care for Medicaid and Medicare patients. Perhaps the key implication of these findings is the need for serious efforts to revise practice arrangements so that FP/GPs can make the best possible decisions for their patients.

Limitations

There are important limitations to our analysis. The CTS Physician Survey is cross-sectional. While we do not know whether these physicians are more or less satisfied than they were in the past, recent evidence from surveys of primary care physicians in Massachusetts suggests that dissatisfaction has increased since 1986.¹⁷ As in all surveys, responses are subject to reporting error and response bias not accounted for by statistical adjustments. Our findings are associations between variables and do not establish causal relationships.

Conclusions

The finding that family physician dissatisfaction, after study results are controlled for personal and practice variables, is associated most strongly with a perceived inability to care for patients raises significant concerns. Dissatisfaction among a large proportion of family physicians threatens the well-being of patients. Given the extent to which the US health care system relies on family physicians, understanding why these physicians are dissatisfied and responding to these problems are important. This cross-sectional snapshot of dissatisfaction among family physicians suggests that patients would benefit from strategies that support rather than disrupt their ongoing relationships with family physicians and that permit their family physician to spend enough time with them to make decisions that are not constrained by financial or other conflicts of interest.