Aortocervicocephalic atherosclerotic disease and coronary artery disease share common risk factors, and patients with one condition are at high risk of harboring or developing the other.1, 2 Over the past decade, several randomized clinical trials of lipid‐lowering medications designed to reduce low‐density lipoprotein cholesterol (LDL‐C) have shown a significant decrease in the risk of coronary events and ischemic stroke among patients who have a history of or are at risk for coronary artery disease, regardless of whether serum cholesterol is elevated.3, 4 Results from more than 3000 stroke patients enrolled in the Heart Protection Study also provide evidence that aggressive lipid‐lowering therapy may prevent recurrent vascular events in individuals who have a total cholesterol level as low as 135 mg/dL and cerebrovascular disease, with or without known coronary artery disease.5

Guidelines from the National Cholesterol Evaluation Program Adult Treatment Panel (ATP) provide target LDL‐C levels for persons with atherosclerotic disease depending on the extent of their vascular risk.6 However, despite the broad dissemination of these guidelines, several published studies of patients with coronary artery disease or dyslipidemia have shown that a large proportion of patients with high vascular risk continue to be underscreened, underdiagnosed, and undertreated for dyslipidemia.79

Few studies have evaluated the quality of cholesterol management among hospitalized patients who have experienced an acute ischemic cerebrovascular event10, 11 So the data are scarce on the management of patients hospitalized for ischemic stroke or transient ischemic attack (TIA) who are, according to ATP criteria, at high risk for future coronary events and on the factors that may govern that management. Systematic reviews have suggested that incorporating a lipid profile during acute stroke presentation could assure baseline assessment and serve as a potential cue for physicians to change their behavior,12 and an American Stroke Association advisory recommends lipid treatment during hospitalization for most patients with ischemic stroke or TIA as it may increase the rate of long‐term use.13

The objectives of this study were to determine the rates of testing for and treatment of dyslipidemia according to national cholesterol guidelines among individuals hospitalized with acute ischemic stroke or TIA and to identify predictors of performance.

METHODS

The California Acute Stroke Prototype Registry (CASPR) is a Centers for Disease Controlsponsored cohort that captured detailed data on patients admitted to 11 hospitals over a 2‐year period. The methods of study have been described elsewhere.14 In brief, CASPR prospectively collected information on acute stroke care at 11 representative hospitals in 5 major population regions of California. Data were collected on diagnostic evaluation, appropriate use of treatment strategies, and disposition on discharge from the hospital. The main goal of CASPR was to pilot‐test a prototype prospective registry of acute stroke and transient ischemic attack to be used as a quality improvement tool. The study population was patients with an admitting or discharge diagnosis of suspected stroke or TIA from November 1, 2002, through January 31, 2003, and from November 1, 2003, through January 31, 2004. The human subjects review board at each participating center approved the study.

For the present analysis, data on all patients with a discharge diagnosis of ischemic stroke or TIA who were admitted during either period were included. We examined the possible association of several variables with 2 primary outcomes: (1) testing lipid profile during hospitalization (as indicated by a documented LDL‐C level) and (2) prescribing lipid‐lowering medication at discharge. In those analyses in which lipid profile testing was the outcome, no variables were considered acceptable reasons for not performing an LDL‐C assessment.

The distribution of LDL‐C levels in this portion of the cohort was determined. Patients were then categorized according to their risk for future coronary events. Patients were classified as at risk for coronary events (ACE) if they either had a documented history of myocardial infarction, coronary artery disease, or diabetes or had undergone carotid endarterectomy or carotid angioplasty/stenting during hospitalization. Criteria for initiating lipid‐lowering therapy were defined according to the ATP III guidelines,6 which were in effect during both CASPR study periods. Continuing the recommendation in ATP II, the ATP III recommendations emphasized that persons with documented coronary artery disease (CAD) receive the most aggressive lipid‐lowering treatment. But this recommendation was expanded to include patients without established CAD, whose coronary risk is equivalent to that of patients with diagnosed CAD.6

As per the ATP III guidelines, CASPR‐ACE patients were considered optimally treated if they were prescribed a lipid‐lowering agent at discharge or if their documented LDL‐C was less than 130 mg/dL. A concurrent history of liver disease, abnormal prothrombin time, life expectancy of less than 1 year, and terminal illness were each considered a valid contraindication to treatment with lipid‐lowering medication. Optimal treatment for non‐ACE patients was defined as receipt of lipid‐lowering medication at discharge or a documented LDL‐C of 160 mg/dL. The rate of optimal treatment of ACE patients was compared to that of non‐ACE patients. The ACE and non‐ACE patients were then further categorized into 1 of 4 groups according to LDL‐C level<100, 100130, 130160, and >160 mg/dLand an assessment for trend of the rate of treatment in each of the 4 categories in the ACE and non‐ACE groups was performed.

RESULTS

Data were available from the 11 CASPR hospitals for 764 patients diagnosed with either ischemic stroke or TIA. Overall, 53.4% of subjects were women, and the average age at hospitalization of 70.4 ( 15.4) years. In the cohort, 55.3% of the patients were non‐Hispanic white, 9.7% were African American, 13.4% were Hispanic, 13% were Asian, and 8.6% were classified as other. Three hundred and nine individuals (40.5% of the cohort) were classified as at risk for coronary events. Of these, 148 (47.8%) had diabetes only, and 160 (51.8%) had a history of MI, CAD, or both. One patient (0.4%) had undergone angioplasty/stenting during hospitalization but had no history of MI, CAD, or diabetes. Only 4 patients (0.52% of the entire cohort) had undergone a carotid endarterectomy or angioplasty/stenting during hospitalization. Rates of lipid assessment and optimal treatment varied widely between hospitals, but testing and treatment were correlated for each hospital. Overall, however, testing and treatment were correlated (Pearson correlation coefficient = 0.35, P < .0001). On an individual hospital level, the correlation was positive and significant for 6 hospitals, positive but not significant for 2 hospitals, and negative but not significant for 3 hospitals.

Overall, LDL‐C levels were determined in 383 patients (50.1%). The likelihood that a patient would have an LDL‐C test performed during hospitalization varied widely by hospital, ranging from 12% to 88% (P < .0001). Univariate variables significantly associated with documented LDL‐C measurement in the overall cohort at the = .10 level were diagnosis of ischemic stroke (as compared to TIA) and history of dyslipidemia (Table 1). In the CASPR cohort, 53% of the ACE subjects received a lipid profile assessment compared to 48% in the rest of the cohort (P = .14). In multivariate analysis, diagnosis of ischemic stroke and history of dyslipidemia remained significantly associated with documented LDL‐C measurement (Table 1).

Lipid‐lowering drugs were prescribed at discharge to 370 patients (48.4%); however, treatment rate varied among hospitals, from a low of 13% of patients to a high of 84% of patients (P < .0001). Univariate factors associated with a higher treatment rate at the = .10 level were diagnosis of ischemic stroke, history of stroke/TIA, history of diabetes, hypertension, history of dyslipidemia, independent ambulation at discharge, and ACE status (Table 2). Patients were less likely to receive lipid‐lowering medication if they had a history of heart failure. Fifty‐nine percent of the CASPR ACE subjects were discharged on lipid‐modifying agents compared to 42% in the rest of the cohort (P = .0006). Multivariate analyses revealed several independent predictors of treatment with lipid‐lowering medication. Diagnosis of ischemic stroke, ACE status, and history of heart failure were negative predictors (less likely to be treated), and history of dyslipidemia was a positive predictor (Table 2). Status as an academic hospital was a hospital characteristic for which a significant association was found. Academic hospitals were significantly more likely to both perform LDL profiles and administer lipid‐lowering medications at discharge than were nonacademic hospitals. This association was found in a logistic regression analysis that did not account for between‐hospital variance. However, when we used GEE analysis, which adjusted for the variance, the difference between academic and nonacademic hospitals was no longer significant.

Three of the patients with documented LDL‐C levels (0.8%) had documented contraindications to therapy. Among all those who had documented LDL‐C levels, the rate of appropriate treatment with lipid‐lowering medications was high in both the ACE and non‐ACE groups (94.6% and 98.6%, respectively; P = .02). However, because only a small number of patients did not receive optimal treatment, the odds ratio of 0.24 had a fairly wide confidence interval (95% CI = 0.06, 0.91). Although a trend toward a higher rate of treatment with increasing LDL‐C level was seen in both the ACE and non‐ACE groups, this trend was only significant for the group with non‐ACE patients (Figure 1).

Figure 1

DISCUSSION

We found that only half the patients hospitalized for ischemic stroke or TIA had LDL‐C levels tested while in the hospital, even among those identified by the ATP guidelines as at high risk for future coronary events. Our findings are in accord with those of the Coverdell Project, which evaluated key features of acute stroke care from 4 prototype registries, those in Georgia, Massachusetts, Michigan, and Ohio, finding that fewer than 40% of acute stroke patients had had lipid profiles checked during hospitalization.11 Our study also evaluated predictors for in‐hospital lipid testing and lipid‐lowering treatment during hospitalization for an acute ischemic cerebrovascular event. We found that lipid testing was correlated with treatment during stroke or TIA hospitalization, suggesting that in‐hospital lipid management is related to an overall appreciation of the importance of lipids.

Understanding the factors resulting in such underperformance is critical for improving patient care and outcomes. Lipid assessment and treatment rates varied widely between CASPR hospitals, reflecting dramatic differences in hospital practice. This finding is similar to that noted in a recent study performed in Europe10 and underscores the need to promote a more uniform approach to in‐hospital care of patients with ischemic stroke or TIA. Our study also found that ischemic stroke patients were much more likely to have their lipid level measured and to be discharged on a lipid‐lowering agent than were TIA patients. This may be so because many treating health care professionals perceive TIAs as benign events that carry a more favorable prognosis than do strokes, or it could be that the length of stay for a TIA, often shorter than that for a stroke, limited in‐hospital testing or planning for patient follow‐up.

A high proportion of non‐ACE, lipid‐tested stroke/TIA patients received lipid‐lowering drug treatment, even when their lipid levels were within the treatment range categorized as nonpharmacologic by the national guidelines. This finding could be a result of one of the goals of the primary study.15 In the primary study, the effect of standardized orders implemented during the second observational period were analyzed by comparing them to those in place during the first observational period to see if they had improved the in‐hospital stroke care process. One of the study goals was optimal discharge utilization of a lipid‐lowering agent, defined as prescription of a lipid modifier or an LDL < 100 mg/dL. There was a significant increase in the number of prescriptions for lipid modifiers at discharge after implementing the standardized orders.15 However, as this study has shown, when existing national cholesterol guidelines were strictly applied to all the patients,6 overall there was a suboptimal rate of utilization of lipid modifiers at discharge.

Lipid profile assessment during stroke admission is one of the 10 performance measures in the performance measure set of the Joint Commission on Accreditation of Healthcare Organizations (JCAHO) Stroke Disease‐Specific Care.16 Initiating therapy with lipid‐lowering agents before discharge may help to maintain continuity of care and clarify therapeutic intent, especially when a different physician is responsible for care after discharge from the hospital. Recent studies indicated that in‐hospital initiation of medication following admission for a vascular event tends to improve longer‐term patient adherence to treatment,17, 18 as well as vascular outcomes,19, 20 and is a strategy favored by the American Stroke Association.13, 21

This study had several limitations. Our definitions of dyslipidemia and of adherence to ATP III goals were based on single measurements of LDL‐C, rather than multiple determinations of lipoprotein subfractions. However, we believe that this approach parallels actual clinical practice more closely. Although LDL‐C is the most important of all the components of the lipid profile,6 because lipid subfractions other than LDL‐C were not collected in the CASPR registry, we may have misclassified a few patients. For instance, extremely high trigylceride levels can render LDL‐C levels inaccurate, and as such, not having a documented LDL‐C may not have always indicated that a lipid panel was not performed. It is also conceivable that physicians might actually have been more thorough in measuring LDL‐C, identifying contraindications to lipid‐lowering therapy, or instituting lipid‐lowering therapy than were noted in the hospital charts. However, for quality assurance purposes, what is documented is the only traceable record of what was actually asked for or done. As such, health care professionals are frequently encouraged to keep updated chart notes. This study was an assessment of in‐hospital behavior; the low utilization of lipid‐lowering agents observed may underestimate the final treatment rate, as we did not evaluate the postdischarge rate of therapy. However, recent data suggest in‐hospital prescription patterns are a major predictor of longer‐term care in the community.17, 22 Last, the CASPR investigators did not collect data on the rate of utilization of lipid agents prior to hospitalization or on the mechanisms by which the strokes and TIAs had occurred. Prehospital utilization of lipid agents has previously been revealed to influence the prescribing of lipid‐lowering agents at discharge.10 Knowledge of the mechanisms of the stroke and TIA events would have increased the number of those eligible for lipid treatment, particularly those whose events were to the result of an atherosclerotic mechanism per ATP III's more expansive definition of CHD risk equivalents, which includes carotid and other forms of clinical atherosclerotic disease.6 However, because the results of other studies that evaluated lipid management in all hospitalized stroke patients (regardless of mechanism)11, 23 or in all patients with any form of clinical atherosclerotic disease24 were in accord with those of our study, it would appear unlikely that such information would have made an overwhelming difference to our results.

In conclusion, the results of the present study suggest that considerable improvement is needed in identifying appropriate candidates among those who have had stroke or TIA and treating them with lipid‐lowering agents. Performing lipid testing in individuals hospitalized with ischemic stroke or TIA is important because it may inform the identification of persons for whom treatment should be initiated or modified. Lipid assessment during hospitalization for stroke/TIA and initiation of lipid‐lowering therapy when indicated are major management steps that all patients with ischemic cerebrovascular events should receive.

Aortocervicocephalic atherosclerotic disease and coronary artery disease share common risk factors, and patients with one condition are at high risk of harboring or developing the other.1, 2 Over the past decade, several randomized clinical trials of lipid‐lowering medications designed to reduce low‐density lipoprotein cholesterol (LDL‐C) have shown a significant decrease in the risk of coronary events and ischemic stroke among patients who have a history of or are at risk for coronary artery disease, regardless of whether serum cholesterol is elevated.3, 4 Results from more than 3000 stroke patients enrolled in the Heart Protection Study also provide evidence that aggressive lipid‐lowering therapy may prevent recurrent vascular events in individuals who have a total cholesterol level as low as 135 mg/dL and cerebrovascular disease, with or without known coronary artery disease.5

Guidelines from the National Cholesterol Evaluation Program Adult Treatment Panel (ATP) provide target LDL‐C levels for persons with atherosclerotic disease depending on the extent of their vascular risk.6 However, despite the broad dissemination of these guidelines, several published studies of patients with coronary artery disease or dyslipidemia have shown that a large proportion of patients with high vascular risk continue to be underscreened, underdiagnosed, and undertreated for dyslipidemia.79

Few studies have evaluated the quality of cholesterol management among hospitalized patients who have experienced an acute ischemic cerebrovascular event10, 11 So the data are scarce on the management of patients hospitalized for ischemic stroke or transient ischemic attack (TIA) who are, according to ATP criteria, at high risk for future coronary events and on the factors that may govern that management. Systematic reviews have suggested that incorporating a lipid profile during acute stroke presentation could assure baseline assessment and serve as a potential cue for physicians to change their behavior,12 and an American Stroke Association advisory recommends lipid treatment during hospitalization for most patients with ischemic stroke or TIA as it may increase the rate of long‐term use.13

The objectives of this study were to determine the rates of testing for and treatment of dyslipidemia according to national cholesterol guidelines among individuals hospitalized with acute ischemic stroke or TIA and to identify predictors of performance.

METHODS

The California Acute Stroke Prototype Registry (CASPR) is a Centers for Disease Controlsponsored cohort that captured detailed data on patients admitted to 11 hospitals over a 2‐year period. The methods of study have been described elsewhere.14 In brief, CASPR prospectively collected information on acute stroke care at 11 representative hospitals in 5 major population regions of California. Data were collected on diagnostic evaluation, appropriate use of treatment strategies, and disposition on discharge from the hospital. The main goal of CASPR was to pilot‐test a prototype prospective registry of acute stroke and transient ischemic attack to be used as a quality improvement tool. The study population was patients with an admitting or discharge diagnosis of suspected stroke or TIA from November 1, 2002, through January 31, 2003, and from November 1, 2003, through January 31, 2004. The human subjects review board at each participating center approved the study.

For the present analysis, data on all patients with a discharge diagnosis of ischemic stroke or TIA who were admitted during either period were included. We examined the possible association of several variables with 2 primary outcomes: (1) testing lipid profile during hospitalization (as indicated by a documented LDL‐C level) and (2) prescribing lipid‐lowering medication at discharge. In those analyses in which lipid profile testing was the outcome, no variables were considered acceptable reasons for not performing an LDL‐C assessment.

The distribution of LDL‐C levels in this portion of the cohort was determined. Patients were then categorized according to their risk for future coronary events. Patients were classified as at risk for coronary events (ACE) if they either had a documented history of myocardial infarction, coronary artery disease, or diabetes or had undergone carotid endarterectomy or carotid angioplasty/stenting during hospitalization. Criteria for initiating lipid‐lowering therapy were defined according to the ATP III guidelines,6 which were in effect during both CASPR study periods. Continuing the recommendation in ATP II, the ATP III recommendations emphasized that persons with documented coronary artery disease (CAD) receive the most aggressive lipid‐lowering treatment. But this recommendation was expanded to include patients without established CAD, whose coronary risk is equivalent to that of patients with diagnosed CAD.6

As per the ATP III guidelines, CASPR‐ACE patients were considered optimally treated if they were prescribed a lipid‐lowering agent at discharge or if their documented LDL‐C was less than 130 mg/dL. A concurrent history of liver disease, abnormal prothrombin time, life expectancy of less than 1 year, and terminal illness were each considered a valid contraindication to treatment with lipid‐lowering medication. Optimal treatment for non‐ACE patients was defined as receipt of lipid‐lowering medication at discharge or a documented LDL‐C of 160 mg/dL. The rate of optimal treatment of ACE patients was compared to that of non‐ACE patients. The ACE and non‐ACE patients were then further categorized into 1 of 4 groups according to LDL‐C level<100, 100130, 130160, and >160 mg/dLand an assessment for trend of the rate of treatment in each of the 4 categories in the ACE and non‐ACE groups was performed.

RESULTS

Data were available from the 11 CASPR hospitals for 764 patients diagnosed with either ischemic stroke or TIA. Overall, 53.4% of subjects were women, and the average age at hospitalization of 70.4 ( 15.4) years. In the cohort, 55.3% of the patients were non‐Hispanic white, 9.7% were African American, 13.4% were Hispanic, 13% were Asian, and 8.6% were classified as other. Three hundred and nine individuals (40.5% of the cohort) were classified as at risk for coronary events. Of these, 148 (47.8%) had diabetes only, and 160 (51.8%) had a history of MI, CAD, or both. One patient (0.4%) had undergone angioplasty/stenting during hospitalization but had no history of MI, CAD, or diabetes. Only 4 patients (0.52% of the entire cohort) had undergone a carotid endarterectomy or angioplasty/stenting during hospitalization. Rates of lipid assessment and optimal treatment varied widely between hospitals, but testing and treatment were correlated for each hospital. Overall, however, testing and treatment were correlated (Pearson correlation coefficient = 0.35, P < .0001). On an individual hospital level, the correlation was positive and significant for 6 hospitals, positive but not significant for 2 hospitals, and negative but not significant for 3 hospitals.

Overall, LDL‐C levels were determined in 383 patients (50.1%). The likelihood that a patient would have an LDL‐C test performed during hospitalization varied widely by hospital, ranging from 12% to 88% (P < .0001). Univariate variables significantly associated with documented LDL‐C measurement in the overall cohort at the = .10 level were diagnosis of ischemic stroke (as compared to TIA) and history of dyslipidemia (Table 1). In the CASPR cohort, 53% of the ACE subjects received a lipid profile assessment compared to 48% in the rest of the cohort (P = .14). In multivariate analysis, diagnosis of ischemic stroke and history of dyslipidemia remained significantly associated with documented LDL‐C measurement (Table 1).

Lipid‐lowering drugs were prescribed at discharge to 370 patients (48.4%); however, treatment rate varied among hospitals, from a low of 13% of patients to a high of 84% of patients (P < .0001). Univariate factors associated with a higher treatment rate at the = .10 level were diagnosis of ischemic stroke, history of stroke/TIA, history of diabetes, hypertension, history of dyslipidemia, independent ambulation at discharge, and ACE status (Table 2). Patients were less likely to receive lipid‐lowering medication if they had a history of heart failure. Fifty‐nine percent of the CASPR ACE subjects were discharged on lipid‐modifying agents compared to 42% in the rest of the cohort (P = .0006). Multivariate analyses revealed several independent predictors of treatment with lipid‐lowering medication. Diagnosis of ischemic stroke, ACE status, and history of heart failure were negative predictors (less likely to be treated), and history of dyslipidemia was a positive predictor (Table 2). Status as an academic hospital was a hospital characteristic for which a significant association was found. Academic hospitals were significantly more likely to both perform LDL profiles and administer lipid‐lowering medications at discharge than were nonacademic hospitals. This association was found in a logistic regression analysis that did not account for between‐hospital variance. However, when we used GEE analysis, which adjusted for the variance, the difference between academic and nonacademic hospitals was no longer significant.

Three of the patients with documented LDL‐C levels (0.8%) had documented contraindications to therapy. Among all those who had documented LDL‐C levels, the rate of appropriate treatment with lipid‐lowering medications was high in both the ACE and non‐ACE groups (94.6% and 98.6%, respectively; P = .02). However, because only a small number of patients did not receive optimal treatment, the odds ratio of 0.24 had a fairly wide confidence interval (95% CI = 0.06, 0.91). Although a trend toward a higher rate of treatment with increasing LDL‐C level was seen in both the ACE and non‐ACE groups, this trend was only significant for the group with non‐ACE patients (Figure 1).

Figure 1

DISCUSSION

We found that only half the patients hospitalized for ischemic stroke or TIA had LDL‐C levels tested while in the hospital, even among those identified by the ATP guidelines as at high risk for future coronary events. Our findings are in accord with those of the Coverdell Project, which evaluated key features of acute stroke care from 4 prototype registries, those in Georgia, Massachusetts, Michigan, and Ohio, finding that fewer than 40% of acute stroke patients had had lipid profiles checked during hospitalization.11 Our study also evaluated predictors for in‐hospital lipid testing and lipid‐lowering treatment during hospitalization for an acute ischemic cerebrovascular event. We found that lipid testing was correlated with treatment during stroke or TIA hospitalization, suggesting that in‐hospital lipid management is related to an overall appreciation of the importance of lipids.

Understanding the factors resulting in such underperformance is critical for improving patient care and outcomes. Lipid assessment and treatment rates varied widely between CASPR hospitals, reflecting dramatic differences in hospital practice. This finding is similar to that noted in a recent study performed in Europe10 and underscores the need to promote a more uniform approach to in‐hospital care of patients with ischemic stroke or TIA. Our study also found that ischemic stroke patients were much more likely to have their lipid level measured and to be discharged on a lipid‐lowering agent than were TIA patients. This may be so because many treating health care professionals perceive TIAs as benign events that carry a more favorable prognosis than do strokes, or it could be that the length of stay for a TIA, often shorter than that for a stroke, limited in‐hospital testing or planning for patient follow‐up.

A high proportion of non‐ACE, lipid‐tested stroke/TIA patients received lipid‐lowering drug treatment, even when their lipid levels were within the treatment range categorized as nonpharmacologic by the national guidelines. This finding could be a result of one of the goals of the primary study.15 In the primary study, the effect of standardized orders implemented during the second observational period were analyzed by comparing them to those in place during the first observational period to see if they had improved the in‐hospital stroke care process. One of the study goals was optimal discharge utilization of a lipid‐lowering agent, defined as prescription of a lipid modifier or an LDL < 100 mg/dL. There was a significant increase in the number of prescriptions for lipid modifiers at discharge after implementing the standardized orders.15 However, as this study has shown, when existing national cholesterol guidelines were strictly applied to all the patients,6 overall there was a suboptimal rate of utilization of lipid modifiers at discharge.

Lipid profile assessment during stroke admission is one of the 10 performance measures in the performance measure set of the Joint Commission on Accreditation of Healthcare Organizations (JCAHO) Stroke Disease‐Specific Care.16 Initiating therapy with lipid‐lowering agents before discharge may help to maintain continuity of care and clarify therapeutic intent, especially when a different physician is responsible for care after discharge from the hospital. Recent studies indicated that in‐hospital initiation of medication following admission for a vascular event tends to improve longer‐term patient adherence to treatment,17, 18 as well as vascular outcomes,19, 20 and is a strategy favored by the American Stroke Association.13, 21

This study had several limitations. Our definitions of dyslipidemia and of adherence to ATP III goals were based on single measurements of LDL‐C, rather than multiple determinations of lipoprotein subfractions. However, we believe that this approach parallels actual clinical practice more closely. Although LDL‐C is the most important of all the components of the lipid profile,6 because lipid subfractions other than LDL‐C were not collected in the CASPR registry, we may have misclassified a few patients. For instance, extremely high trigylceride levels can render LDL‐C levels inaccurate, and as such, not having a documented LDL‐C may not have always indicated that a lipid panel was not performed. It is also conceivable that physicians might actually have been more thorough in measuring LDL‐C, identifying contraindications to lipid‐lowering therapy, or instituting lipid‐lowering therapy than were noted in the hospital charts. However, for quality assurance purposes, what is documented is the only traceable record of what was actually asked for or done. As such, health care professionals are frequently encouraged to keep updated chart notes. This study was an assessment of in‐hospital behavior; the low utilization of lipid‐lowering agents observed may underestimate the final treatment rate, as we did not evaluate the postdischarge rate of therapy. However, recent data suggest in‐hospital prescription patterns are a major predictor of longer‐term care in the community.17, 22 Last, the CASPR investigators did not collect data on the rate of utilization of lipid agents prior to hospitalization or on the mechanisms by which the strokes and TIAs had occurred. Prehospital utilization of lipid agents has previously been revealed to influence the prescribing of lipid‐lowering agents at discharge.10 Knowledge of the mechanisms of the stroke and TIA events would have increased the number of those eligible for lipid treatment, particularly those whose events were to the result of an atherosclerotic mechanism per ATP III's more expansive definition of CHD risk equivalents, which includes carotid and other forms of clinical atherosclerotic disease.6 However, because the results of other studies that evaluated lipid management in all hospitalized stroke patients (regardless of mechanism)11, 23 or in all patients with any form of clinical atherosclerotic disease24 were in accord with those of our study, it would appear unlikely that such information would have made an overwhelming difference to our results.

In conclusion, the results of the present study suggest that considerable improvement is needed in identifying appropriate candidates among those who have had stroke or TIA and treating them with lipid‐lowering agents. Performing lipid testing in individuals hospitalized with ischemic stroke or TIA is important because it may inform the identification of persons for whom treatment should be initiated or modified. Lipid assessment during hospitalization for stroke/TIA and initiation of lipid‐lowering therapy when indicated are major management steps that all patients with ischemic cerebrovascular events should receive.

Aortocervicocephalic atherosclerotic disease and coronary artery disease share common risk factors, and patients with one condition are at high risk of harboring or developing the other.1, 2 Over the past decade, several randomized clinical trials of lipid‐lowering medications designed to reduce low‐density lipoprotein cholesterol (LDL‐C) have shown a significant decrease in the risk of coronary events and ischemic stroke among patients who have a history of or are at risk for coronary artery disease, regardless of whether serum cholesterol is elevated.3, 4 Results from more than 3000 stroke patients enrolled in the Heart Protection Study also provide evidence that aggressive lipid‐lowering therapy may prevent recurrent vascular events in individuals who have a total cholesterol level as low as 135 mg/dL and cerebrovascular disease, with or without known coronary artery disease.5

Guidelines from the National Cholesterol Evaluation Program Adult Treatment Panel (ATP) provide target LDL‐C levels for persons with atherosclerotic disease depending on the extent of their vascular risk.6 However, despite the broad dissemination of these guidelines, several published studies of patients with coronary artery disease or dyslipidemia have shown that a large proportion of patients with high vascular risk continue to be underscreened, underdiagnosed, and undertreated for dyslipidemia.79

Few studies have evaluated the quality of cholesterol management among hospitalized patients who have experienced an acute ischemic cerebrovascular event10, 11 So the data are scarce on the management of patients hospitalized for ischemic stroke or transient ischemic attack (TIA) who are, according to ATP criteria, at high risk for future coronary events and on the factors that may govern that management. Systematic reviews have suggested that incorporating a lipid profile during acute stroke presentation could assure baseline assessment and serve as a potential cue for physicians to change their behavior,12 and an American Stroke Association advisory recommends lipid treatment during hospitalization for most patients with ischemic stroke or TIA as it may increase the rate of long‐term use.13

The objectives of this study were to determine the rates of testing for and treatment of dyslipidemia according to national cholesterol guidelines among individuals hospitalized with acute ischemic stroke or TIA and to identify predictors of performance.

METHODS

The California Acute Stroke Prototype Registry (CASPR) is a Centers for Disease Controlsponsored cohort that captured detailed data on patients admitted to 11 hospitals over a 2‐year period. The methods of study have been described elsewhere.14 In brief, CASPR prospectively collected information on acute stroke care at 11 representative hospitals in 5 major population regions of California. Data were collected on diagnostic evaluation, appropriate use of treatment strategies, and disposition on discharge from the hospital. The main goal of CASPR was to pilot‐test a prototype prospective registry of acute stroke and transient ischemic attack to be used as a quality improvement tool. The study population was patients with an admitting or discharge diagnosis of suspected stroke or TIA from November 1, 2002, through January 31, 2003, and from November 1, 2003, through January 31, 2004. The human subjects review board at each participating center approved the study.

For the present analysis, data on all patients with a discharge diagnosis of ischemic stroke or TIA who were admitted during either period were included. We examined the possible association of several variables with 2 primary outcomes: (1) testing lipid profile during hospitalization (as indicated by a documented LDL‐C level) and (2) prescribing lipid‐lowering medication at discharge. In those analyses in which lipid profile testing was the outcome, no variables were considered acceptable reasons for not performing an LDL‐C assessment.

The distribution of LDL‐C levels in this portion of the cohort was determined. Patients were then categorized according to their risk for future coronary events. Patients were classified as at risk for coronary events (ACE) if they either had a documented history of myocardial infarction, coronary artery disease, or diabetes or had undergone carotid endarterectomy or carotid angioplasty/stenting during hospitalization. Criteria for initiating lipid‐lowering therapy were defined according to the ATP III guidelines,6 which were in effect during both CASPR study periods. Continuing the recommendation in ATP II, the ATP III recommendations emphasized that persons with documented coronary artery disease (CAD) receive the most aggressive lipid‐lowering treatment. But this recommendation was expanded to include patients without established CAD, whose coronary risk is equivalent to that of patients with diagnosed CAD.6

As per the ATP III guidelines, CASPR‐ACE patients were considered optimally treated if they were prescribed a lipid‐lowering agent at discharge or if their documented LDL‐C was less than 130 mg/dL. A concurrent history of liver disease, abnormal prothrombin time, life expectancy of less than 1 year, and terminal illness were each considered a valid contraindication to treatment with lipid‐lowering medication. Optimal treatment for non‐ACE patients was defined as receipt of lipid‐lowering medication at discharge or a documented LDL‐C of 160 mg/dL. The rate of optimal treatment of ACE patients was compared to that of non‐ACE patients. The ACE and non‐ACE patients were then further categorized into 1 of 4 groups according to LDL‐C level<100, 100130, 130160, and >160 mg/dLand an assessment for trend of the rate of treatment in each of the 4 categories in the ACE and non‐ACE groups was performed.

RESULTS

Data were available from the 11 CASPR hospitals for 764 patients diagnosed with either ischemic stroke or TIA. Overall, 53.4% of subjects were women, and the average age at hospitalization of 70.4 ( 15.4) years. In the cohort, 55.3% of the patients were non‐Hispanic white, 9.7% were African American, 13.4% were Hispanic, 13% were Asian, and 8.6% were classified as other. Three hundred and nine individuals (40.5% of the cohort) were classified as at risk for coronary events. Of these, 148 (47.8%) had diabetes only, and 160 (51.8%) had a history of MI, CAD, or both. One patient (0.4%) had undergone angioplasty/stenting during hospitalization but had no history of MI, CAD, or diabetes. Only 4 patients (0.52% of the entire cohort) had undergone a carotid endarterectomy or angioplasty/stenting during hospitalization. Rates of lipid assessment and optimal treatment varied widely between hospitals, but testing and treatment were correlated for each hospital. Overall, however, testing and treatment were correlated (Pearson correlation coefficient = 0.35, P < .0001). On an individual hospital level, the correlation was positive and significant for 6 hospitals, positive but not significant for 2 hospitals, and negative but not significant for 3 hospitals.

Overall, LDL‐C levels were determined in 383 patients (50.1%). The likelihood that a patient would have an LDL‐C test performed during hospitalization varied widely by hospital, ranging from 12% to 88% (P < .0001). Univariate variables significantly associated with documented LDL‐C measurement in the overall cohort at the = .10 level were diagnosis of ischemic stroke (as compared to TIA) and history of dyslipidemia (Table 1). In the CASPR cohort, 53% of the ACE subjects received a lipid profile assessment compared to 48% in the rest of the cohort (P = .14). In multivariate analysis, diagnosis of ischemic stroke and history of dyslipidemia remained significantly associated with documented LDL‐C measurement (Table 1).

Lipid‐lowering drugs were prescribed at discharge to 370 patients (48.4%); however, treatment rate varied among hospitals, from a low of 13% of patients to a high of 84% of patients (P < .0001). Univariate factors associated with a higher treatment rate at the = .10 level were diagnosis of ischemic stroke, history of stroke/TIA, history of diabetes, hypertension, history of dyslipidemia, independent ambulation at discharge, and ACE status (Table 2). Patients were less likely to receive lipid‐lowering medication if they had a history of heart failure. Fifty‐nine percent of the CASPR ACE subjects were discharged on lipid‐modifying agents compared to 42% in the rest of the cohort (P = .0006). Multivariate analyses revealed several independent predictors of treatment with lipid‐lowering medication. Diagnosis of ischemic stroke, ACE status, and history of heart failure were negative predictors (less likely to be treated), and history of dyslipidemia was a positive predictor (Table 2). Status as an academic hospital was a hospital characteristic for which a significant association was found. Academic hospitals were significantly more likely to both perform LDL profiles and administer lipid‐lowering medications at discharge than were nonacademic hospitals. This association was found in a logistic regression analysis that did not account for between‐hospital variance. However, when we used GEE analysis, which adjusted for the variance, the difference between academic and nonacademic hospitals was no longer significant.

Three of the patients with documented LDL‐C levels (0.8%) had documented contraindications to therapy. Among all those who had documented LDL‐C levels, the rate of appropriate treatment with lipid‐lowering medications was high in both the ACE and non‐ACE groups (94.6% and 98.6%, respectively; P = .02). However, because only a small number of patients did not receive optimal treatment, the odds ratio of 0.24 had a fairly wide confidence interval (95% CI = 0.06, 0.91). Although a trend toward a higher rate of treatment with increasing LDL‐C level was seen in both the ACE and non‐ACE groups, this trend was only significant for the group with non‐ACE patients (Figure 1).

Figure 1

DISCUSSION

We found that only half the patients hospitalized for ischemic stroke or TIA had LDL‐C levels tested while in the hospital, even among those identified by the ATP guidelines as at high risk for future coronary events. Our findings are in accord with those of the Coverdell Project, which evaluated key features of acute stroke care from 4 prototype registries, those in Georgia, Massachusetts, Michigan, and Ohio, finding that fewer than 40% of acute stroke patients had had lipid profiles checked during hospitalization.11 Our study also evaluated predictors for in‐hospital lipid testing and lipid‐lowering treatment during hospitalization for an acute ischemic cerebrovascular event. We found that lipid testing was correlated with treatment during stroke or TIA hospitalization, suggesting that in‐hospital lipid management is related to an overall appreciation of the importance of lipids.

Understanding the factors resulting in such underperformance is critical for improving patient care and outcomes. Lipid assessment and treatment rates varied widely between CASPR hospitals, reflecting dramatic differences in hospital practice. This finding is similar to that noted in a recent study performed in Europe10 and underscores the need to promote a more uniform approach to in‐hospital care of patients with ischemic stroke or TIA. Our study also found that ischemic stroke patients were much more likely to have their lipid level measured and to be discharged on a lipid‐lowering agent than were TIA patients. This may be so because many treating health care professionals perceive TIAs as benign events that carry a more favorable prognosis than do strokes, or it could be that the length of stay for a TIA, often shorter than that for a stroke, limited in‐hospital testing or planning for patient follow‐up.

A high proportion of non‐ACE, lipid‐tested stroke/TIA patients received lipid‐lowering drug treatment, even when their lipid levels were within the treatment range categorized as nonpharmacologic by the national guidelines. This finding could be a result of one of the goals of the primary study.15 In the primary study, the effect of standardized orders implemented during the second observational period were analyzed by comparing them to those in place during the first observational period to see if they had improved the in‐hospital stroke care process. One of the study goals was optimal discharge utilization of a lipid‐lowering agent, defined as prescription of a lipid modifier or an LDL < 100 mg/dL. There was a significant increase in the number of prescriptions for lipid modifiers at discharge after implementing the standardized orders.15 However, as this study has shown, when existing national cholesterol guidelines were strictly applied to all the patients,6 overall there was a suboptimal rate of utilization of lipid modifiers at discharge.

Lipid profile assessment during stroke admission is one of the 10 performance measures in the performance measure set of the Joint Commission on Accreditation of Healthcare Organizations (JCAHO) Stroke Disease‐Specific Care.16 Initiating therapy with lipid‐lowering agents before discharge may help to maintain continuity of care and clarify therapeutic intent, especially when a different physician is responsible for care after discharge from the hospital. Recent studies indicated that in‐hospital initiation of medication following admission for a vascular event tends to improve longer‐term patient adherence to treatment,17, 18 as well as vascular outcomes,19, 20 and is a strategy favored by the American Stroke Association.13, 21

This study had several limitations. Our definitions of dyslipidemia and of adherence to ATP III goals were based on single measurements of LDL‐C, rather than multiple determinations of lipoprotein subfractions. However, we believe that this approach parallels actual clinical practice more closely. Although LDL‐C is the most important of all the components of the lipid profile,6 because lipid subfractions other than LDL‐C were not collected in the CASPR registry, we may have misclassified a few patients. For instance, extremely high trigylceride levels can render LDL‐C levels inaccurate, and as such, not having a documented LDL‐C may not have always indicated that a lipid panel was not performed. It is also conceivable that physicians might actually have been more thorough in measuring LDL‐C, identifying contraindications to lipid‐lowering therapy, or instituting lipid‐lowering therapy than were noted in the hospital charts. However, for quality assurance purposes, what is documented is the only traceable record of what was actually asked for or done. As such, health care professionals are frequently encouraged to keep updated chart notes. This study was an assessment of in‐hospital behavior; the low utilization of lipid‐lowering agents observed may underestimate the final treatment rate, as we did not evaluate the postdischarge rate of therapy. However, recent data suggest in‐hospital prescription patterns are a major predictor of longer‐term care in the community.17, 22 Last, the CASPR investigators did not collect data on the rate of utilization of lipid agents prior to hospitalization or on the mechanisms by which the strokes and TIAs had occurred. Prehospital utilization of lipid agents has previously been revealed to influence the prescribing of lipid‐lowering agents at discharge.10 Knowledge of the mechanisms of the stroke and TIA events would have increased the number of those eligible for lipid treatment, particularly those whose events were to the result of an atherosclerotic mechanism per ATP III's more expansive definition of CHD risk equivalents, which includes carotid and other forms of clinical atherosclerotic disease.6 However, because the results of other studies that evaluated lipid management in all hospitalized stroke patients (regardless of mechanism)11, 23 or in all patients with any form of clinical atherosclerotic disease24 were in accord with those of our study, it would appear unlikely that such information would have made an overwhelming difference to our results.

In conclusion, the results of the present study suggest that considerable improvement is needed in identifying appropriate candidates among those who have had stroke or TIA and treating them with lipid‐lowering agents. Performing lipid testing in individuals hospitalized with ischemic stroke or TIA is important because it may inform the identification of persons for whom treatment should be initiated or modified. Lipid assessment during hospitalization for stroke/TIA and initiation of lipid‐lowering therapy when indicated are major management steps that all patients with ischemic cerebrovascular events should receive.

In the nearly 10 years since Bob Wachter and Lee Goldman coined the word hospitalist,1 it has been inspiring to see the dramatic growth of this specialty and with it, the growth in membership of the Society of Hospital Medicine.

Over the same period, the health care system has made progress toward ensuring that it provides the safest, highest‐quality health care possible.

In my mind, the two phenomena are related. The Society of Hospital Medicine, along with the hospitalist field more generally, has played a critical role in promoting the use of evidence‐based care, improved teamwork, and health information technology, which can make a significant difference in the care patients receive in the hospital. Similarly, the mission of the Agency for Healthcare Research and Quality (AHRQ) is to improve the quality, safety, efficiency, and effectiveness of health care for all Americans. So, both of our organizations are working to create positive change that will improve the health and health care of all patients.

As a research agency, we support studies, systematic reviews, and evaluations that help to build the foundation of evidence for health care. However, our work goes beyond simply conducting, supporting, and disseminating health services research. At its heart, our mission is helping the health care system translate research into improved practice and policy. We do not see research as an end in itself but rather a vehicle to improve health care and health. We achieve our goals by working with our public‐ and private‐sector partners to translate the research we support and conduct into knowledge and information that can be used immediately to improve health care for all Americans.

Health Information Technology

This commentary features AHRQ's quality‐related initiatives, including promoting the use of health information technology to improve quality and safety, providing the tools to assess health care quality, and expanding training to promote quality improvement in local communities. Many of these tools are ideal for hospitalists to use in their mission to ensure high‐quality care and an efficient and thorough handoff at discharge.

AHRQ is at the leading edge of President Bush's vision of a health care system that harnesses the power of health information technology (IT) to improve quality. AHRQ has invested more than $166 million in more than 100 projects to promote the use of health IT, with a special focus on rural hospitals and communities. These projects will enable providers to improve patient safety and reduce medication errors by eliminating handwritten prescriptions, help to ensure that important information follows patients as they move among health care settings, and reduce duplicative and unnecessary testing.

As part of this investment, AHRQ has awarded multiyear contracts totaling nearly $30 million to Colorado, Delaware, Indiana, Rhode Island, Tennessee, and Utah to help in the development of statewide networks that are secure, ensure privacy, and make information more accessible. Participants in the networks include major purchasers of health care, public and private payers, hospitals, ambulatory care facilities, home health care providers, and long‐term care providers.

In addition, AHRQ created the AHRQ National Resource Center for Health Information Technology (http://healthit.ahrq.gov) as a focus for technical assistance, information sharing, and collaboration. The resource center site provides emerging lessons from the field, a knowledge library with links to more than 5000 health IT information resources, an evaluation toolkit to help those implementing health IT projects, a summary of key topics, plus other resources pointing to current health IT activities, funding opportunities, and other information.

Effective Health Care Program

However, as we all know, health IT is not a magic bullet or the sole answer to the quality and safety problems facing the American health care system. It is a means to an end. Although health IT has the great potential to deliver evidence to clinicians, patients, and other health care decision makers when they need it, one challenge is to ensure the evidence base is readily available.

To that end, AHRQ's new Effective Health Care Program, authorized under Section 1013 of the Medicare Prescription Drug, Improvement, and Modernization Act (MMA) of 2003, is conducting research with a focus on outcomes, comparative clinical effectiveness, and appropriateness of pharmaceuticals, devices, and health care services. At press time, AHRQ had released two effectiveness reports, Comparative Effectiveness of Management Strategies for Gastroesophageal Reflux Disease2 and Effectiveness of Noninvasive Diagnostic Tests for Breast Abnormalities.3

The AHRQ Effective Health Care Program takes three approaches to research on the comparative effectiveness of different treatments and clinical practices:

• Review and synthesize knowledge. AHRQ's Evidence‐Based Practice Centers systematically review published and unpublished scientific evidence to develop evidence reports.

• Promote and generate knowledge. A new AHRQ‐supported research network called DEcIDE (Developing Evidence to Inform Decisions about Effectiveness) conducts accelerated practical studies of new scientific evidence and analytic tools.

• Compile the findings and translate knowledge. The John M. Eisenberg Clinical Decisions and Communications Science Center compiles the research results into a variety of useful formats for stakeholders.

Interested readers should go to the Effective Health Care Web site, www.effectivehealthcare.ahrq.gov, to read more about AHRQ's Effective Health Care Program and to see work in progress, suggest topics for research, or comment on research questions and draft reports.

Patient Safety and Quality

Since 2001, AHRQ has been the leading funder of patient safety research, and I am proud that our $165 million patient safety research program is bearing fruit.

For example, Bob Wachterin his spare timeand his team at the University of California, San Francisco, under contract to AHRQ, developed AHRQ's Patient Safety Network, which can be found at http://www.psnet.ahrq.gov. AHRQ PSNet is a national Web‐based portal featuring the latest news and essential resources on patient safety. Included in the Patient Safety Network is AHRQ's Web M&M site (http://www.webmm.ahrq.gov/), an anonymous forum where clinicians post new cases of medical errors for discussion. The site also includes expert commentaries on how to think through such cases, identifying problem areas and potential solutions. Together, PSNet and Web M&M receive more than 100,000 visits each month.

Readers of JHM also will be interested in AHRQ's Hospital Survey on Patient Safety Culture, which we released in partnership with Premier, Inc., the Department of Defense (DoD), and the American Hospital Association. The survey can be used to evaluate employees' attitudes about patient safety in their facilities or within specific units. It addresses a critical aspect of patient safety improvement: measuring organizational conditions that can lead to adverse events and patient harm. The survey, which is being used in the DoD's medical facilities, is available at http://www.ahrq.gov/qual/hospculture. Premier and the AHA are strongly encouraging their members to adopt the Hospital Survey on Patient Safety Culture if they are not already using a survey as part of their safety improvement efforts.

In another quality initiative, AHRQ also has developed a series of software tools that can help hospitals gauge the quality of care they provide. AHRQ's Prevention Quality Indicators allow hospitals to detect potentially avoidable hospital admissions for illnesses that can be effectively treated with high‐quality, community‐based primary care.

Another tool is the Inpatient Quality Indicators, 29 measures that can be used to help hospitals identify potential problem areas and to provide a proxy measure of hospital quality of care. The Patient Safety Indicators can help hospitals enhance their performance by quickly detecting potential medical errors in patients who have undergone medical or surgical care. Staff can then investigate to determine whether the problems detected by the indicators were caused by potentially preventable medical errors or have some other explanation.

Building on its long track record of developing surveys to gauge consumers' experiences in the health care system, AHRQ has developed H‐CAHPS, a survey tool that hospitals, employers, states, and others can use to assess the perceptions of hospital patients about the quality of the care they receive. This information is designed to help patients, their employers, and other purchasers make informed decisions and give hospitals feedback they can use to improve care. The Centers for Medicare & Medicaid Services, in partnership with the nation's major hospital trade groups, is using H‐CAHPS as part of their collaborative Hospital Quality Alliance to develop comparative information about hospitals.

In the future, AHRQ plans to create other surveys, including Ambulatory CAHPS, In‐Center Hemodialysis CAHPS, and Nursing Home CAHPS.

AHRQ also is now working in partnership with the Department of Veterans Affairs to train the third class of state and hospital teams participating in the Patient Safety Improvement Corps. The program was created because states asked us for help in areas such as conducting effective investigations of reports of medical errors and developing interventions and changes in standard clinical practice. When trained, the teams return to their local communities armed with the knowledge to improve patient safety.

National Health Care Quality and Disparities Reports

Finally, in January 2006, AHRQ released the third annual National Healthcare Quality Report (NHQR) and the National Healthcare Disparities Report. These reports provide data on the quality of health care and disparities in the use of health care services associated with patient characteristics, including race, ethnicity, income, education, and area of residence.45 In March 2006, AHRQ released a Web‐based tool called State Snapshots for states to use in measuring health care quality. The State Snapshots provides quick and easy access to the many measures and tables of the 2005 NHQR and also provides trend data that can help in the understanding of the quality of health care in individual states, including strengths, weaknesses, and opportunities for improvement. The reports and the State Snapshots are available on AHRQ's QualityTools Web site at http://www.qualitytools.ahrq.gov/.

I hope this article has provided a glimpse into the quality improvement initiatives and activities supported by AHRQ. For ongoing information on these activities, I urge readers to go to AHRQ's Web site and sign up for our electronic newsletter (http://www.ahrq.gov/news/ahrqlist.htm) and our patient safety newsletter (http://www.ahrq.gov/qual/ptsflist.htm).

AHRQ's aim is to make doing the right thing the easy thing to do for the health care system. We look forward to working with readers of the Journal of Hospital Medicine to achieve that goal.

In the nearly 10 years since Bob Wachter and Lee Goldman coined the word hospitalist,1 it has been inspiring to see the dramatic growth of this specialty and with it, the growth in membership of the Society of Hospital Medicine.

Over the same period, the health care system has made progress toward ensuring that it provides the safest, highest‐quality health care possible.

In my mind, the two phenomena are related. The Society of Hospital Medicine, along with the hospitalist field more generally, has played a critical role in promoting the use of evidence‐based care, improved teamwork, and health information technology, which can make a significant difference in the care patients receive in the hospital. Similarly, the mission of the Agency for Healthcare Research and Quality (AHRQ) is to improve the quality, safety, efficiency, and effectiveness of health care for all Americans. So, both of our organizations are working to create positive change that will improve the health and health care of all patients.

As a research agency, we support studies, systematic reviews, and evaluations that help to build the foundation of evidence for health care. However, our work goes beyond simply conducting, supporting, and disseminating health services research. At its heart, our mission is helping the health care system translate research into improved practice and policy. We do not see research as an end in itself but rather a vehicle to improve health care and health. We achieve our goals by working with our public‐ and private‐sector partners to translate the research we support and conduct into knowledge and information that can be used immediately to improve health care for all Americans.

Health Information Technology

This commentary features AHRQ's quality‐related initiatives, including promoting the use of health information technology to improve quality and safety, providing the tools to assess health care quality, and expanding training to promote quality improvement in local communities. Many of these tools are ideal for hospitalists to use in their mission to ensure high‐quality care and an efficient and thorough handoff at discharge.

AHRQ is at the leading edge of President Bush's vision of a health care system that harnesses the power of health information technology (IT) to improve quality. AHRQ has invested more than $166 million in more than 100 projects to promote the use of health IT, with a special focus on rural hospitals and communities. These projects will enable providers to improve patient safety and reduce medication errors by eliminating handwritten prescriptions, help to ensure that important information follows patients as they move among health care settings, and reduce duplicative and unnecessary testing.

As part of this investment, AHRQ has awarded multiyear contracts totaling nearly $30 million to Colorado, Delaware, Indiana, Rhode Island, Tennessee, and Utah to help in the development of statewide networks that are secure, ensure privacy, and make information more accessible. Participants in the networks include major purchasers of health care, public and private payers, hospitals, ambulatory care facilities, home health care providers, and long‐term care providers.

In addition, AHRQ created the AHRQ National Resource Center for Health Information Technology (http://healthit.ahrq.gov) as a focus for technical assistance, information sharing, and collaboration. The resource center site provides emerging lessons from the field, a knowledge library with links to more than 5000 health IT information resources, an evaluation toolkit to help those implementing health IT projects, a summary of key topics, plus other resources pointing to current health IT activities, funding opportunities, and other information.

Effective Health Care Program

However, as we all know, health IT is not a magic bullet or the sole answer to the quality and safety problems facing the American health care system. It is a means to an end. Although health IT has the great potential to deliver evidence to clinicians, patients, and other health care decision makers when they need it, one challenge is to ensure the evidence base is readily available.

To that end, AHRQ's new Effective Health Care Program, authorized under Section 1013 of the Medicare Prescription Drug, Improvement, and Modernization Act (MMA) of 2003, is conducting research with a focus on outcomes, comparative clinical effectiveness, and appropriateness of pharmaceuticals, devices, and health care services. At press time, AHRQ had released two effectiveness reports, Comparative Effectiveness of Management Strategies for Gastroesophageal Reflux Disease2 and Effectiveness of Noninvasive Diagnostic Tests for Breast Abnormalities.3

The AHRQ Effective Health Care Program takes three approaches to research on the comparative effectiveness of different treatments and clinical practices:

• Review and synthesize knowledge. AHRQ's Evidence‐Based Practice Centers systematically review published and unpublished scientific evidence to develop evidence reports.

• Promote and generate knowledge. A new AHRQ‐supported research network called DEcIDE (Developing Evidence to Inform Decisions about Effectiveness) conducts accelerated practical studies of new scientific evidence and analytic tools.

• Compile the findings and translate knowledge. The John M. Eisenberg Clinical Decisions and Communications Science Center compiles the research results into a variety of useful formats for stakeholders.

Interested readers should go to the Effective Health Care Web site, www.effectivehealthcare.ahrq.gov, to read more about AHRQ's Effective Health Care Program and to see work in progress, suggest topics for research, or comment on research questions and draft reports.

Patient Safety and Quality

Since 2001, AHRQ has been the leading funder of patient safety research, and I am proud that our $165 million patient safety research program is bearing fruit.

For example, Bob Wachterin his spare timeand his team at the University of California, San Francisco, under contract to AHRQ, developed AHRQ's Patient Safety Network, which can be found at http://www.psnet.ahrq.gov. AHRQ PSNet is a national Web‐based portal featuring the latest news and essential resources on patient safety. Included in the Patient Safety Network is AHRQ's Web M&M site (http://www.webmm.ahrq.gov/), an anonymous forum where clinicians post new cases of medical errors for discussion. The site also includes expert commentaries on how to think through such cases, identifying problem areas and potential solutions. Together, PSNet and Web M&M receive more than 100,000 visits each month.

Readers of JHM also will be interested in AHRQ's Hospital Survey on Patient Safety Culture, which we released in partnership with Premier, Inc., the Department of Defense (DoD), and the American Hospital Association. The survey can be used to evaluate employees' attitudes about patient safety in their facilities or within specific units. It addresses a critical aspect of patient safety improvement: measuring organizational conditions that can lead to adverse events and patient harm. The survey, which is being used in the DoD's medical facilities, is available at http://www.ahrq.gov/qual/hospculture. Premier and the AHA are strongly encouraging their members to adopt the Hospital Survey on Patient Safety Culture if they are not already using a survey as part of their safety improvement efforts.

In another quality initiative, AHRQ also has developed a series of software tools that can help hospitals gauge the quality of care they provide. AHRQ's Prevention Quality Indicators allow hospitals to detect potentially avoidable hospital admissions for illnesses that can be effectively treated with high‐quality, community‐based primary care.

Another tool is the Inpatient Quality Indicators, 29 measures that can be used to help hospitals identify potential problem areas and to provide a proxy measure of hospital quality of care. The Patient Safety Indicators can help hospitals enhance their performance by quickly detecting potential medical errors in patients who have undergone medical or surgical care. Staff can then investigate to determine whether the problems detected by the indicators were caused by potentially preventable medical errors or have some other explanation.

Building on its long track record of developing surveys to gauge consumers' experiences in the health care system, AHRQ has developed H‐CAHPS, a survey tool that hospitals, employers, states, and others can use to assess the perceptions of hospital patients about the quality of the care they receive. This information is designed to help patients, their employers, and other purchasers make informed decisions and give hospitals feedback they can use to improve care. The Centers for Medicare & Medicaid Services, in partnership with the nation's major hospital trade groups, is using H‐CAHPS as part of their collaborative Hospital Quality Alliance to develop comparative information about hospitals.

In the future, AHRQ plans to create other surveys, including Ambulatory CAHPS, In‐Center Hemodialysis CAHPS, and Nursing Home CAHPS.

AHRQ also is now working in partnership with the Department of Veterans Affairs to train the third class of state and hospital teams participating in the Patient Safety Improvement Corps. The program was created because states asked us for help in areas such as conducting effective investigations of reports of medical errors and developing interventions and changes in standard clinical practice. When trained, the teams return to their local communities armed with the knowledge to improve patient safety.

National Health Care Quality and Disparities Reports

Finally, in January 2006, AHRQ released the third annual National Healthcare Quality Report (NHQR) and the National Healthcare Disparities Report. These reports provide data on the quality of health care and disparities in the use of health care services associated with patient characteristics, including race, ethnicity, income, education, and area of residence.45 In March 2006, AHRQ released a Web‐based tool called State Snapshots for states to use in measuring health care quality. The State Snapshots provides quick and easy access to the many measures and tables of the 2005 NHQR and also provides trend data that can help in the understanding of the quality of health care in individual states, including strengths, weaknesses, and opportunities for improvement. The reports and the State Snapshots are available on AHRQ's QualityTools Web site at http://www.qualitytools.ahrq.gov/.

I hope this article has provided a glimpse into the quality improvement initiatives and activities supported by AHRQ. For ongoing information on these activities, I urge readers to go to AHRQ's Web site and sign up for our electronic newsletter (http://www.ahrq.gov/news/ahrqlist.htm) and our patient safety newsletter (http://www.ahrq.gov/qual/ptsflist.htm).

AHRQ's aim is to make doing the right thing the easy thing to do for the health care system. We look forward to working with readers of the Journal of Hospital Medicine to achieve that goal.

In the nearly 10 years since Bob Wachter and Lee Goldman coined the word hospitalist,1 it has been inspiring to see the dramatic growth of this specialty and with it, the growth in membership of the Society of Hospital Medicine.

Over the same period, the health care system has made progress toward ensuring that it provides the safest, highest‐quality health care possible.

In my mind, the two phenomena are related. The Society of Hospital Medicine, along with the hospitalist field more generally, has played a critical role in promoting the use of evidence‐based care, improved teamwork, and health information technology, which can make a significant difference in the care patients receive in the hospital. Similarly, the mission of the Agency for Healthcare Research and Quality (AHRQ) is to improve the quality, safety, efficiency, and effectiveness of health care for all Americans. So, both of our organizations are working to create positive change that will improve the health and health care of all patients.

As a research agency, we support studies, systematic reviews, and evaluations that help to build the foundation of evidence for health care. However, our work goes beyond simply conducting, supporting, and disseminating health services research. At its heart, our mission is helping the health care system translate research into improved practice and policy. We do not see research as an end in itself but rather a vehicle to improve health care and health. We achieve our goals by working with our public‐ and private‐sector partners to translate the research we support and conduct into knowledge and information that can be used immediately to improve health care for all Americans.

Health Information Technology

This commentary features AHRQ's quality‐related initiatives, including promoting the use of health information technology to improve quality and safety, providing the tools to assess health care quality, and expanding training to promote quality improvement in local communities. Many of these tools are ideal for hospitalists to use in their mission to ensure high‐quality care and an efficient and thorough handoff at discharge.

AHRQ is at the leading edge of President Bush's vision of a health care system that harnesses the power of health information technology (IT) to improve quality. AHRQ has invested more than $166 million in more than 100 projects to promote the use of health IT, with a special focus on rural hospitals and communities. These projects will enable providers to improve patient safety and reduce medication errors by eliminating handwritten prescriptions, help to ensure that important information follows patients as they move among health care settings, and reduce duplicative and unnecessary testing.

As part of this investment, AHRQ has awarded multiyear contracts totaling nearly $30 million to Colorado, Delaware, Indiana, Rhode Island, Tennessee, and Utah to help in the development of statewide networks that are secure, ensure privacy, and make information more accessible. Participants in the networks include major purchasers of health care, public and private payers, hospitals, ambulatory care facilities, home health care providers, and long‐term care providers.

In addition, AHRQ created the AHRQ National Resource Center for Health Information Technology (http://healthit.ahrq.gov) as a focus for technical assistance, information sharing, and collaboration. The resource center site provides emerging lessons from the field, a knowledge library with links to more than 5000 health IT information resources, an evaluation toolkit to help those implementing health IT projects, a summary of key topics, plus other resources pointing to current health IT activities, funding opportunities, and other information.

Effective Health Care Program

However, as we all know, health IT is not a magic bullet or the sole answer to the quality and safety problems facing the American health care system. It is a means to an end. Although health IT has the great potential to deliver evidence to clinicians, patients, and other health care decision makers when they need it, one challenge is to ensure the evidence base is readily available.

To that end, AHRQ's new Effective Health Care Program, authorized under Section 1013 of the Medicare Prescription Drug, Improvement, and Modernization Act (MMA) of 2003, is conducting research with a focus on outcomes, comparative clinical effectiveness, and appropriateness of pharmaceuticals, devices, and health care services. At press time, AHRQ had released two effectiveness reports, Comparative Effectiveness of Management Strategies for Gastroesophageal Reflux Disease2 and Effectiveness of Noninvasive Diagnostic Tests for Breast Abnormalities.3

The AHRQ Effective Health Care Program takes three approaches to research on the comparative effectiveness of different treatments and clinical practices:

• Review and synthesize knowledge. AHRQ's Evidence‐Based Practice Centers systematically review published and unpublished scientific evidence to develop evidence reports.

• Promote and generate knowledge. A new AHRQ‐supported research network called DEcIDE (Developing Evidence to Inform Decisions about Effectiveness) conducts accelerated practical studies of new scientific evidence and analytic tools.

• Compile the findings and translate knowledge. The John M. Eisenberg Clinical Decisions and Communications Science Center compiles the research results into a variety of useful formats for stakeholders.

Interested readers should go to the Effective Health Care Web site, www.effectivehealthcare.ahrq.gov, to read more about AHRQ's Effective Health Care Program and to see work in progress, suggest topics for research, or comment on research questions and draft reports.

Patient Safety and Quality

Since 2001, AHRQ has been the leading funder of patient safety research, and I am proud that our $165 million patient safety research program is bearing fruit.

For example, Bob Wachterin his spare timeand his team at the University of California, San Francisco, under contract to AHRQ, developed AHRQ's Patient Safety Network, which can be found at http://www.psnet.ahrq.gov. AHRQ PSNet is a national Web‐based portal featuring the latest news and essential resources on patient safety. Included in the Patient Safety Network is AHRQ's Web M&M site (http://www.webmm.ahrq.gov/), an anonymous forum where clinicians post new cases of medical errors for discussion. The site also includes expert commentaries on how to think through such cases, identifying problem areas and potential solutions. Together, PSNet and Web M&M receive more than 100,000 visits each month.

Readers of JHM also will be interested in AHRQ's Hospital Survey on Patient Safety Culture, which we released in partnership with Premier, Inc., the Department of Defense (DoD), and the American Hospital Association. The survey can be used to evaluate employees' attitudes about patient safety in their facilities or within specific units. It addresses a critical aspect of patient safety improvement: measuring organizational conditions that can lead to adverse events and patient harm. The survey, which is being used in the DoD's medical facilities, is available at http://www.ahrq.gov/qual/hospculture. Premier and the AHA are strongly encouraging their members to adopt the Hospital Survey on Patient Safety Culture if they are not already using a survey as part of their safety improvement efforts.

In another quality initiative, AHRQ also has developed a series of software tools that can help hospitals gauge the quality of care they provide. AHRQ's Prevention Quality Indicators allow hospitals to detect potentially avoidable hospital admissions for illnesses that can be effectively treated with high‐quality, community‐based primary care.

Another tool is the Inpatient Quality Indicators, 29 measures that can be used to help hospitals identify potential problem areas and to provide a proxy measure of hospital quality of care. The Patient Safety Indicators can help hospitals enhance their performance by quickly detecting potential medical errors in patients who have undergone medical or surgical care. Staff can then investigate to determine whether the problems detected by the indicators were caused by potentially preventable medical errors or have some other explanation.

Building on its long track record of developing surveys to gauge consumers' experiences in the health care system, AHRQ has developed H‐CAHPS, a survey tool that hospitals, employers, states, and others can use to assess the perceptions of hospital patients about the quality of the care they receive. This information is designed to help patients, their employers, and other purchasers make informed decisions and give hospitals feedback they can use to improve care. The Centers for Medicare & Medicaid Services, in partnership with the nation's major hospital trade groups, is using H‐CAHPS as part of their collaborative Hospital Quality Alliance to develop comparative information about hospitals.

In the future, AHRQ plans to create other surveys, including Ambulatory CAHPS, In‐Center Hemodialysis CAHPS, and Nursing Home CAHPS.

AHRQ also is now working in partnership with the Department of Veterans Affairs to train the third class of state and hospital teams participating in the Patient Safety Improvement Corps. The program was created because states asked us for help in areas such as conducting effective investigations of reports of medical errors and developing interventions and changes in standard clinical practice. When trained, the teams return to their local communities armed with the knowledge to improve patient safety.

National Health Care Quality and Disparities Reports

Finally, in January 2006, AHRQ released the third annual National Healthcare Quality Report (NHQR) and the National Healthcare Disparities Report. These reports provide data on the quality of health care and disparities in the use of health care services associated with patient characteristics, including race, ethnicity, income, education, and area of residence.45 In March 2006, AHRQ released a Web‐based tool called State Snapshots for states to use in measuring health care quality. The State Snapshots provides quick and easy access to the many measures and tables of the 2005 NHQR and also provides trend data that can help in the understanding of the quality of health care in individual states, including strengths, weaknesses, and opportunities for improvement. The reports and the State Snapshots are available on AHRQ's QualityTools Web site at http://www.qualitytools.ahrq.gov/.

I hope this article has provided a glimpse into the quality improvement initiatives and activities supported by AHRQ. For ongoing information on these activities, I urge readers to go to AHRQ's Web site and sign up for our electronic newsletter (http://www.ahrq.gov/news/ahrqlist.htm) and our patient safety newsletter (http://www.ahrq.gov/qual/ptsflist.htm).

AHRQ's aim is to make doing the right thing the easy thing to do for the health care system. We look forward to working with readers of the Journal of Hospital Medicine to achieve that goal.

As these conversations illustrate, discussing preferences regarding resuscitation is a challenging and important task for physicians. Understanding patients' wishes at the end of life allows clinicians to provide the type of care patients want, to avoid unwanted interventions, and to promote patient autonomy and dignity. Despite the frequency with which physicians have these conversations, they typically fall short when discussing code status with patients. Clinicians fail to address patients' concerns, monopolize conversations, and commonly misunderstand patients' resuscitation preferences.13 Often these discussions do not occur at all; more than 70% of seriously ill patients have never discussed advance directives with their physician.2, 4 The multicenter SUPPORT study, which demonstrated serious problems in the care of seriously ill hospitalized patients, documented that only 47% of physicians knew when their patients preferred do‐not‐resuscitate status.5

Hospitalists frequently conduct resuscitation discussions. Patients who are admitted to the hospital are usually seriously ill, and hospitalists need to assess rapidly whether an individual patient would want a resuscitation attempt if he or she had a cardiopulmonary arrest in the hospital. They need to build trust and rapport quickly with patients they have never met. Despite this challenge, hospitalists are in a good position to discuss resuscitation preferences.6 Patients may be more willing to discuss these issues in inpatient rather than outpatient settings because their acute deterioration may encourage self‐reflection.6 Furthermore, the time and productivity pressures of office practice often make it difficult for primary care physicians, who often know the patient and family best, to address advance directives. Although studies have documented that patients are interested in talking to their primary physicians about these matters,7, 8 these conversations do not occur with regularity. Preliminary research has raised the possibility that cancer patients may actually prefer to discuss these issues in an inpatient setting with a hospital‐based provider rather than with their oncologists.9 Studies have not addressed the question of whether patients with diseases other than cancer would prefer to discuss these issues with a hospitalist or their outpatient subspecialist.

Given that more than half of all Americans die in hospitals, hospitalists care for many people who are terminally ill and will need to assess preferences for cardiopulmonary resuscitation (CPR) and other treatments. Hospitalists need to be competent and compassionate in their approach to patients and their families. In this articles we review clinician barriers to holding these conversations, offer a variety of approaches to enhance these discussions, and review communication techniques that can be used to improve understanding.

Clinician Barriers to Discussing Resuscitation

Clinicians' own barriers may lead to infrequent and inadequate conversations about resuscitation. Understanding these barriers may allow providers to overcome them and facilitate better and more frequent communication. A discussion of patient barriers is beyond the scope of this article.

Laying Groundwork for the Discussion

The decision regarding resuscitation should be seen in the context of the patient's goals and values and overall health status. To address code status effectively, it is imperative first to elicit the person's view of his/her illness and then gently correct any misunderstandings. A patient who thinks her/his life will go on indefinitely may feel no need to consider her/his own mortality or alternatives to full resuscitation status. Alternatively, a patient who senses his/her mortality may have already thought about resuscitation and have clear preferences. A key first step in the conversation is to understand a patient's values and goals and comprehension of his/her illness. As in the second discussion above, a clinician can begin a discussion by saying, Tell me what you understand about your illness.

Discussing prognosis with patients can be difficult as physicians struggle with uncertainty. In the most comprehensive study to date of prognostication, physicians overestimated patient survival on average by a factor of 5.31 Nonetheless, there are compelling reasons to discuss prognosis. Failure to do so often results in patients spending their last days in the hospital receiving more aggressive treatments than they might choose if they understood their prognoses.32 Further, patients are denied the opportunity to address issues of life closure, such as spending time with family, thinking about legacy, and settling financial affairs. Physicians also fear they will take away hope with prognostication and believe patients expect greater accuracy than they can provide.33

Physicians can improve their prognostication skills by considering patients' functional status and clinical signs and symptoms and by using validated scales. The Karnofsky Performance Score (KPS) and the Eastern Cooperative Oncology Group (ECOG) Performance Status have been shown to correlate with survival,3440 and the Palliative Prognostic Score (PaP) has been validated in both cancer and noncancer patients.41, 4446 The PaP uses a combination of the KPS, clinical signs and symptoms, and the clinician's clinical prediction of survival. In addition, clinicians can be honest with patients about prognostic uncertainties and give prognoses in ranges, such as days to weeks, or weeks to months.

How to Broach the Subject of Resuscitation

In the first scenario presented in this article, the hospitalist says, There's something I need to talk with you about that I talk with all of my patients about. Although many clinicians begin resuscitation discussions this way, the question is problematic because often it is untrue. Most clinicians do not discuss code status with all patients who are admitted to the hospital. A better option would be to say, When I take care of patients with advanced cancer, I like to talk with them about their wishes regarding resuscitation. Is that all right? Better yet would be to ask a general question such as As you look ahead, what worries you most? or As you look to the future, what do you hope for? These and other useful questions appear in Table 1. These questions allow patients to bring up their concerns, show that the clinician cares about them, and often segue into a discussion of patients' hopes and worries about their own death and dying process. These questions often allow patients to bring up important goals that bear directly on the issue of resuscitation. For example, in the second scenario at the beginning of the article, the patient says she wants to have quality time with her family at home and does not want life‐sustaining technologies. Such a patient may not want resuscitation. When discussing CPR, it can also be helpful to state explicitly that resuscitation is used when a patient has died, rather than to use euphemisms such as, If your heart were to stop. The clinician can ask explicitly, If you were to die, would you want? There are other strategies for introducing the subject of resuscitation if these questions do not work. If a patient seems uncomfortable with the conversation, the clinician can address this discomfort directly by saying, This conversation seems to make you uncomfortable. Other strategies for exploring these issues include inquiring if the person has ever discussed resuscitation with his/her family or another physician, or asking if anyone else in the family has been very sick. Additionally, clinicians can ask questions about surrogate decision making. If you were to get so sick that you were unable to make decisions for yourself, who would you want to make them for you? This can then lead into a discussion of whether the patient has spoken to that person about his/her preferences, and if so, what was the content of that conversation. Another useful question is, Is there any state that would be worse than death for you? This question focuses on outcomes and allows the physician to put the issue of resuscitation into perspective for a patient.

Knowing the Facts about Cardiopulmonary Resuscitation

In general, about 1 in 7 patients who have undergone CPR in the hospital survive to hospital discharge. Two literature reviews, from 1989 and 1998, of hospitalized patients who underwent CPR in the hospital reported surprisingly similar statistics. Immediate survival in both series was 41%, and survival to hospital discharge was 13%‐14%. Factors that increased survival included arrest due to coronary artery disease or drug overdose/adverse reaction, and location of arrest in the intensive care unit (ICU). Factors that decreased survival included sepsis at time of arrest, cancer, whether localized or metastatic, dementia, acute renal failure, bed‐bound status, and residence in a nursing home. Neither age nor sex was associated with survival.47, 48 In a meta‐analysis of outcome studies of CPR in metastatic cancer patients, none of 117 patients survived to hospital discharge.49

Most people get their information about CPR survival rates from the mass media, which portray CPR as a very successful procedure. For instance, on television, the sole source of information on CPR for most patients, the rate of survival to hospital discharge is 64%, much higher than the 1314% cited in the medical literature.50 Thus, it is no surprise that a patient with metastatic cancer or another life‐limiting illness would assume a positive outcome with CPR.

Knowledge of the facts about CPR survival rates is key when a physician discusses code status with a patient because these data influence patients' decisions. Patients who have a realistic understanding of their chances are less likely to prefer resuscitation.51, 52 Offering patients information about the success rate for CPR in their particular situation is critical in helping them reach a decision consistent with their values and goals. This information can be given quantitatively or qualitatively depending on the clinician assessment of what the patient would prefer.

Avoiding Stark Dichotomies in Resuscitation Discussions

In clinical practice there are 3 ways in which physicians can present resuscitation decisions as black and white and ignore the shades of gray. First, clinicians may present the choice between resuscitation and do not resuscitate (DNR) as a choice between life and death. In the first scenario above, the physician states If your heart were to stop, do you want us to do everything to bring you back? implying he would be able to save the patient's life with resuscitation attempts. When discussing resuscitation, clinicians should avoid language that suggests such a stark dichotomy. The reality is that most patients die despite resuscitation attempts. In fact, a patient is actually choosing between certain death (without resuscitation) and likely death (if resuscitation is attempted). For a patient with a serious, life‐limiting illness, it may be more effective to frame the conversation in terms of how that person envisions the end of life, and not whether death will eventually occur.

Second, clinicians and patients sometimes equate DNR with doing nothing or giving up. Clinicians fail to discuss other treatment options or alternatives, such as continuing ongoing aggressive medical treatments with DNR status or pursuing palliative care. Performing resuscitation is equated with activity and treatment, whereas withholding resuscitation is seen as passivity and giving up. To the patient, this can feel like abandonment, as if the doctor is withholding a treatment and not offering anything in its place. Examples of positive phrases that demonstrate the physician will continue to offer excellent care include: We will continue maximal medical therapy. However, if you die despite these treatments, we will let you die peacefully and won't attempt to revive you; We'll continue the intravenous antibiotics, but we won't plan to move you to the ICU if things worsen53; and We will work hard to treat your pain and other symptoms and to get you home. In addition, hospitalists must ensure, when signing out to physician colleagues, nurses, respiratory therapists, and others, that DNR orders are not overinterpreted to mean no treatment. Although a DNR order states that in the event of a cardiac arrest, no attempt at resuscitation will be made, it should have no bearing on other appropriate and desired interventions, including antibiotics, chemotherapy, and artificial nutrition; treatment for pain and other symptoms; and even monitoring in an ICU. This misunderstanding of DNR status is common among health care providers and has led many to argue that DNR orders should be part of a more comprehensive treatment plan that outlines where the patient's wishes for treatment fall on the spectrum from otherwise aggressive measures to comfort care.54, 55 Physicians who have a clear understanding of a DNR order will be able to reassure a patient that they will continue to receive desired care, but that if the patient dies, no attempt to resuscitate will be made.

Third, it is important to remind patients who choose full resuscitation status that additional decisions will need to be made if resuscitation is successful. Most patients who survive cardiopulmonary arrest end up worse off clinically and spend time in an ICU with life‐support measures in place, such as mechanical ventilation and vasopressors. Even if they survive, there will likely be a period during which they are unable to speak for themselves. This situation puts the burden of decision making on their surrogates or an appointed durable power of attorney for health care (DPOA‐HC). It is important to ask patients ahead of time whether there are conditions under which they would not want ongoing life‐sustaining measures. For example, a person might opt for discontinuation of life‐support measures if the physician and family agreed that there was only a minimal chance of cognitive recovery existed and that ongoing support was only prolonging inevitable death. To clarify the patient's wishes in this situation, you might ask, Are there conditions that would be worse than death? Encouraging the patient to share his/her wishes in this situation with a surrogate or DPOA‐HC will help to ensure those wishes are respected.

Communication Techniques

When discussing advance directives, it is important to give patients the chance to describe their life goals and their values to establish a context for understanding the role of life‐sustaining treatments. One useful method to elicit these goals and values is to ask open‐ended questions, followed by periods of silence so the patient has time to express himself/herself. In the second scenario, above, the physician used open‐ended questions in several instances: Tell me what you understand about your illness and As you look ahead, what worries you most? Tulsky and colleagues documented that medical residents spoke 76% of the time in discussions with patients about code status.30 In an ideal case, this ratio should be reversed or at least be even, allowing patients the time to explain their thoughts. Acknowledging patients' emotions by stating simply, You seem [angry/sad/perplexed], and waiting for an answer can help patients feel they are being understood. Making empathic statements is another powerful communication technique that conveys understanding.56 Examples include, That must be really sad for you, and It must be frightening to be in so much pain. As noted, silence can also be a powerful tool. Clinicians tend to be uncomfortable with silence and so fill the gaps with words. Allowing for silence enables patients to digest what they have heard, encourages them to continue speaking, and shows them the clinician wants to hear what they have to say. When giving information about any medical issues and especially about CPR, it is important to explain concepts in lay terms and to avoid medical jargon.57 Additionally, nonverbal communication techniques such as making eye contact, head nodding, and leaning in toward the patient all help in communicating engagement in the conversation. Having the conversation in a quiet and private place and sitting at the same level as the patient or family is also important. It is always a good idea to check in with patients to assess their understanding. Simply asking, Do you have any questions about what I said? or Does that all make sense? gives patients the opportunity to ask for clarification. Attempting to summarize what a patient has said can also help to clarify misunderstandings. Useful phrases include, Let me see if I've gotten this right or I want to make sure that I understood what you're telling me, followed by the clinician's synopsis of important points discussed.58 A summary of important communication techniques can be found in Table 2.

There are 2 additional statements that can be very useful when patients and families are struggling with the reality of severe illness and are still hoping for longer life and cure. The first is the I wish statement, in which the clinician allies himself or herself with the patient's or family's wishes by stating, I wish it were different. I wish there were more chemotherapy we could give you that would make a difference.59 Occasionally, when tension is developing because the clinician does not believe an intervention is warranted but the patient desires it nonetheless, the I wish statement can be a powerful way of realigning with the patient. For example, responding to a patient who says, I want chemo to cure my cancer, with the statement There is no chemotherapy to help you can seem antagonistic. In contrast, saying, I wish there were a treatment that would make your cancer go away aligns the clinician with the patient and supports the patient. Another advantage of I wish statements is that they are truethe clinician does wish that there was an effective treatment and would gladly provide it if available. In general, I wish statements are more effective than I'm sorry statements, because the latter can be interpreted as the clinician taking responsibility for the situation. When a provider says, I'm sorry to give you this bad news, the patient may feel the need to say, That's OK. On the other hand, saying, I'm so sorry that your mother is dying, is very human and unlikely to be interpreted as the clinician taking responsibility for the death.

A similar technique is to use the statement Hope for the best, prepare for the worst when speaking with patients and families.60 For example, a physician could say, I really wish your mother could get better, and we should still hope for that; at the same time, we need to prepare for what will happen if she doesn't get better. Once again, this phrase both allows the patient or family to continue hoping things will improve and the clinician to support this hope, while simultaneously beginning the process of planning for the more likely outcome. Over time, the patient and family often move toward accepting that the patient is dying. Finally, trying to help the patient or family maintain hope in the face of illness and death is challenging but important: If your mother can't get better, are there other things you can hope for? Helping to identify tangible and realistic goals, such as being free from pain, seeing an important family member one last time, or getting home can provide hope at a difficult time.

Giving a Recommendation

Most patients with serious illness and their families want help making complex and ethically charged decisions. When clinicians ask patients to make decisions unilaterally, patients often feel anxious, sometimes for weeks.61 Families are often paralyzed when faced with the very difficult decision of whether to withdraw life‐sustaining interventions from a family member with an advanced terminal illness. Even if they understand on an intellectual level that ceasing to provide potentially curative or life‐prolonging therapies is the best choice, they are not yet able to accept this decision on an emotional level and ultimately may feel responsible and then guilty for the patient's death. Physicians need to carry some of the burden of making these difficult decisions. One way to relieve family members of some of this guilt is to recommend a plan of care based on substituted judgment, that is, on what the patient said she or he wanted or what the family thinks the patient would have wanted.6264 In addition, clinicians should use their medical expertise, experience, and understanding of the situation to make recommendations. The patient or family can then accept or reject the physician's advice, which maintains patient autonomy, yet not have to explicitly instruct the clinician to withdraw or limit life‐sustaining interventions.

The preceding discussion and recommendations can guide scenarios like those presented at the beginning of this article. In the second conversation, the clinician had just told the patient that he could help her to achieve her goal of pain control and of returning and staying home.

CONCLUSIONS

The decision about cardiopulmonary resuscitation is part of a larger conversation with a patient about how she or he wants to spend the rest of his/her life. Importantly, the decision should be made in context, rather than in isolation. Given the understanding that develops between physician and patient in the conversation above, it is not necessary to describe all the specific treatments that occur during CPR because the physician has already established that the patient does not want to return to the hospital, and she understands that she has a terminal condition and is dying. Through exploring a patient's goals and values, a clinician can discover a patient's preferences for care generally and come up with a comprehensive plan that addresses the particular individual's medical, social, and emotional needs. For physicians, few interventions are as important or rewarding as relieving patients' suffering and helping them attain their goals during and at the end of life.

As these conversations illustrate, discussing preferences regarding resuscitation is a challenging and important task for physicians. Understanding patients' wishes at the end of life allows clinicians to provide the type of care patients want, to avoid unwanted interventions, and to promote patient autonomy and dignity. Despite the frequency with which physicians have these conversations, they typically fall short when discussing code status with patients. Clinicians fail to address patients' concerns, monopolize conversations, and commonly misunderstand patients' resuscitation preferences.13 Often these discussions do not occur at all; more than 70% of seriously ill patients have never discussed advance directives with their physician.2, 4 The multicenter SUPPORT study, which demonstrated serious problems in the care of seriously ill hospitalized patients, documented that only 47% of physicians knew when their patients preferred do‐not‐resuscitate status.5

Hospitalists frequently conduct resuscitation discussions. Patients who are admitted to the hospital are usually seriously ill, and hospitalists need to assess rapidly whether an individual patient would want a resuscitation attempt if he or she had a cardiopulmonary arrest in the hospital. They need to build trust and rapport quickly with patients they have never met. Despite this challenge, hospitalists are in a good position to discuss resuscitation preferences.6 Patients may be more willing to discuss these issues in inpatient rather than outpatient settings because their acute deterioration may encourage self‐reflection.6 Furthermore, the time and productivity pressures of office practice often make it difficult for primary care physicians, who often know the patient and family best, to address advance directives. Although studies have documented that patients are interested in talking to their primary physicians about these matters,7, 8 these conversations do not occur with regularity. Preliminary research has raised the possibility that cancer patients may actually prefer to discuss these issues in an inpatient setting with a hospital‐based provider rather than with their oncologists.9 Studies have not addressed the question of whether patients with diseases other than cancer would prefer to discuss these issues with a hospitalist or their outpatient subspecialist.

Given that more than half of all Americans die in hospitals, hospitalists care for many people who are terminally ill and will need to assess preferences for cardiopulmonary resuscitation (CPR) and other treatments. Hospitalists need to be competent and compassionate in their approach to patients and their families. In this articles we review clinician barriers to holding these conversations, offer a variety of approaches to enhance these discussions, and review communication techniques that can be used to improve understanding.

Clinician Barriers to Discussing Resuscitation

Clinicians' own barriers may lead to infrequent and inadequate conversations about resuscitation. Understanding these barriers may allow providers to overcome them and facilitate better and more frequent communication. A discussion of patient barriers is beyond the scope of this article.

Laying Groundwork for the Discussion

The decision regarding resuscitation should be seen in the context of the patient's goals and values and overall health status. To address code status effectively, it is imperative first to elicit the person's view of his/her illness and then gently correct any misunderstandings. A patient who thinks her/his life will go on indefinitely may feel no need to consider her/his own mortality or alternatives to full resuscitation status. Alternatively, a patient who senses his/her mortality may have already thought about resuscitation and have clear preferences. A key first step in the conversation is to understand a patient's values and goals and comprehension of his/her illness. As in the second discussion above, a clinician can begin a discussion by saying, Tell me what you understand about your illness.

Discussing prognosis with patients can be difficult as physicians struggle with uncertainty. In the most comprehensive study to date of prognostication, physicians overestimated patient survival on average by a factor of 5.31 Nonetheless, there are compelling reasons to discuss prognosis. Failure to do so often results in patients spending their last days in the hospital receiving more aggressive treatments than they might choose if they understood their prognoses.32 Further, patients are denied the opportunity to address issues of life closure, such as spending time with family, thinking about legacy, and settling financial affairs. Physicians also fear they will take away hope with prognostication and believe patients expect greater accuracy than they can provide.33

Physicians can improve their prognostication skills by considering patients' functional status and clinical signs and symptoms and by using validated scales. The Karnofsky Performance Score (KPS) and the Eastern Cooperative Oncology Group (ECOG) Performance Status have been shown to correlate with survival,3440 and the Palliative Prognostic Score (PaP) has been validated in both cancer and noncancer patients.41, 4446 The PaP uses a combination of the KPS, clinical signs and symptoms, and the clinician's clinical prediction of survival. In addition, clinicians can be honest with patients about prognostic uncertainties and give prognoses in ranges, such as days to weeks, or weeks to months.

How to Broach the Subject of Resuscitation

In the first scenario presented in this article, the hospitalist says, There's something I need to talk with you about that I talk with all of my patients about. Although many clinicians begin resuscitation discussions this way, the question is problematic because often it is untrue. Most clinicians do not discuss code status with all patients who are admitted to the hospital. A better option would be to say, When I take care of patients with advanced cancer, I like to talk with them about their wishes regarding resuscitation. Is that all right? Better yet would be to ask a general question such as As you look ahead, what worries you most? or As you look to the future, what do you hope for? These and other useful questions appear in Table 1. These questions allow patients to bring up their concerns, show that the clinician cares about them, and often segue into a discussion of patients' hopes and worries about their own death and dying process. These questions often allow patients to bring up important goals that bear directly on the issue of resuscitation. For example, in the second scenario at the beginning of the article, the patient says she wants to have quality time with her family at home and does not want life‐sustaining technologies. Such a patient may not want resuscitation. When discussing CPR, it can also be helpful to state explicitly that resuscitation is used when a patient has died, rather than to use euphemisms such as, If your heart were to stop. The clinician can ask explicitly, If you were to die, would you want? There are other strategies for introducing the subject of resuscitation if these questions do not work. If a patient seems uncomfortable with the conversation, the clinician can address this discomfort directly by saying, This conversation seems to make you uncomfortable. Other strategies for exploring these issues include inquiring if the person has ever discussed resuscitation with his/her family or another physician, or asking if anyone else in the family has been very sick. Additionally, clinicians can ask questions about surrogate decision making. If you were to get so sick that you were unable to make decisions for yourself, who would you want to make them for you? This can then lead into a discussion of whether the patient has spoken to that person about his/her preferences, and if so, what was the content of that conversation. Another useful question is, Is there any state that would be worse than death for you? This question focuses on outcomes and allows the physician to put the issue of resuscitation into perspective for a patient.

Knowing the Facts about Cardiopulmonary Resuscitation

In general, about 1 in 7 patients who have undergone CPR in the hospital survive to hospital discharge. Two literature reviews, from 1989 and 1998, of hospitalized patients who underwent CPR in the hospital reported surprisingly similar statistics. Immediate survival in both series was 41%, and survival to hospital discharge was 13%‐14%. Factors that increased survival included arrest due to coronary artery disease or drug overdose/adverse reaction, and location of arrest in the intensive care unit (ICU). Factors that decreased survival included sepsis at time of arrest, cancer, whether localized or metastatic, dementia, acute renal failure, bed‐bound status, and residence in a nursing home. Neither age nor sex was associated with survival.47, 48 In a meta‐analysis of outcome studies of CPR in metastatic cancer patients, none of 117 patients survived to hospital discharge.49

Most people get their information about CPR survival rates from the mass media, which portray CPR as a very successful procedure. For instance, on television, the sole source of information on CPR for most patients, the rate of survival to hospital discharge is 64%, much higher than the 1314% cited in the medical literature.50 Thus, it is no surprise that a patient with metastatic cancer or another life‐limiting illness would assume a positive outcome with CPR.

Knowledge of the facts about CPR survival rates is key when a physician discusses code status with a patient because these data influence patients' decisions. Patients who have a realistic understanding of their chances are less likely to prefer resuscitation.51, 52 Offering patients information about the success rate for CPR in their particular situation is critical in helping them reach a decision consistent with their values and goals. This information can be given quantitatively or qualitatively depending on the clinician assessment of what the patient would prefer.

Avoiding Stark Dichotomies in Resuscitation Discussions

In clinical practice there are 3 ways in which physicians can present resuscitation decisions as black and white and ignore the shades of gray. First, clinicians may present the choice between resuscitation and do not resuscitate (DNR) as a choice between life and death. In the first scenario above, the physician states If your heart were to stop, do you want us to do everything to bring you back? implying he would be able to save the patient's life with resuscitation attempts. When discussing resuscitation, clinicians should avoid language that suggests such a stark dichotomy. The reality is that most patients die despite resuscitation attempts. In fact, a patient is actually choosing between certain death (without resuscitation) and likely death (if resuscitation is attempted). For a patient with a serious, life‐limiting illness, it may be more effective to frame the conversation in terms of how that person envisions the end of life, and not whether death will eventually occur.

Second, clinicians and patients sometimes equate DNR with doing nothing or giving up. Clinicians fail to discuss other treatment options or alternatives, such as continuing ongoing aggressive medical treatments with DNR status or pursuing palliative care. Performing resuscitation is equated with activity and treatment, whereas withholding resuscitation is seen as passivity and giving up. To the patient, this can feel like abandonment, as if the doctor is withholding a treatment and not offering anything in its place. Examples of positive phrases that demonstrate the physician will continue to offer excellent care include: We will continue maximal medical therapy. However, if you die despite these treatments, we will let you die peacefully and won't attempt to revive you; We'll continue the intravenous antibiotics, but we won't plan to move you to the ICU if things worsen53; and We will work hard to treat your pain and other symptoms and to get you home. In addition, hospitalists must ensure, when signing out to physician colleagues, nurses, respiratory therapists, and others, that DNR orders are not overinterpreted to mean no treatment. Although a DNR order states that in the event of a cardiac arrest, no attempt at resuscitation will be made, it should have no bearing on other appropriate and desired interventions, including antibiotics, chemotherapy, and artificial nutrition; treatment for pain and other symptoms; and even monitoring in an ICU. This misunderstanding of DNR status is common among health care providers and has led many to argue that DNR orders should be part of a more comprehensive treatment plan that outlines where the patient's wishes for treatment fall on the spectrum from otherwise aggressive measures to comfort care.54, 55 Physicians who have a clear understanding of a DNR order will be able to reassure a patient that they will continue to receive desired care, but that if the patient dies, no attempt to resuscitate will be made.

Third, it is important to remind patients who choose full resuscitation status that additional decisions will need to be made if resuscitation is successful. Most patients who survive cardiopulmonary arrest end up worse off clinically and spend time in an ICU with life‐support measures in place, such as mechanical ventilation and vasopressors. Even if they survive, there will likely be a period during which they are unable to speak for themselves. This situation puts the burden of decision making on their surrogates or an appointed durable power of attorney for health care (DPOA‐HC). It is important to ask patients ahead of time whether there are conditions under which they would not want ongoing life‐sustaining measures. For example, a person might opt for discontinuation of life‐support measures if the physician and family agreed that there was only a minimal chance of cognitive recovery existed and that ongoing support was only prolonging inevitable death. To clarify the patient's wishes in this situation, you might ask, Are there conditions that would be worse than death? Encouraging the patient to share his/her wishes in this situation with a surrogate or DPOA‐HC will help to ensure those wishes are respected.

Communication Techniques

When discussing advance directives, it is important to give patients the chance to describe their life goals and their values to establish a context for understanding the role of life‐sustaining treatments. One useful method to elicit these goals and values is to ask open‐ended questions, followed by periods of silence so the patient has time to express himself/herself. In the second scenario, above, the physician used open‐ended questions in several instances: Tell me what you understand about your illness and As you look ahead, what worries you most? Tulsky and colleagues documented that medical residents spoke 76% of the time in discussions with patients about code status.30 In an ideal case, this ratio should be reversed or at least be even, allowing patients the time to explain their thoughts. Acknowledging patients' emotions by stating simply, You seem [angry/sad/perplexed], and waiting for an answer can help patients feel they are being understood. Making empathic statements is another powerful communication technique that conveys understanding.56 Examples include, That must be really sad for you, and It must be frightening to be in so much pain. As noted, silence can also be a powerful tool. Clinicians tend to be uncomfortable with silence and so fill the gaps with words. Allowing for silence enables patients to digest what they have heard, encourages them to continue speaking, and shows them the clinician wants to hear what they have to say. When giving information about any medical issues and especially about CPR, it is important to explain concepts in lay terms and to avoid medical jargon.57 Additionally, nonverbal communication techniques such as making eye contact, head nodding, and leaning in toward the patient all help in communicating engagement in the conversation. Having the conversation in a quiet and private place and sitting at the same level as the patient or family is also important. It is always a good idea to check in with patients to assess their understanding. Simply asking, Do you have any questions about what I said? or Does that all make sense? gives patients the opportunity to ask for clarification. Attempting to summarize what a patient has said can also help to clarify misunderstandings. Useful phrases include, Let me see if I've gotten this right or I want to make sure that I understood what you're telling me, followed by the clinician's synopsis of important points discussed.58 A summary of important communication techniques can be found in Table 2.

There are 2 additional statements that can be very useful when patients and families are struggling with the reality of severe illness and are still hoping for longer life and cure. The first is the I wish statement, in which the clinician allies himself or herself with the patient's or family's wishes by stating, I wish it were different. I wish there were more chemotherapy we could give you that would make a difference.59 Occasionally, when tension is developing because the clinician does not believe an intervention is warranted but the patient desires it nonetheless, the I wish statement can be a powerful way of realigning with the patient. For example, responding to a patient who says, I want chemo to cure my cancer, with the statement There is no chemotherapy to help you can seem antagonistic. In contrast, saying, I wish there were a treatment that would make your cancer go away aligns the clinician with the patient and supports the patient. Another advantage of I wish statements is that they are truethe clinician does wish that there was an effective treatment and would gladly provide it if available. In general, I wish statements are more effective than I'm sorry statements, because the latter can be interpreted as the clinician taking responsibility for the situation. When a provider says, I'm sorry to give you this bad news, the patient may feel the need to say, That's OK. On the other hand, saying, I'm so sorry that your mother is dying, is very human and unlikely to be interpreted as the clinician taking responsibility for the death.

A similar technique is to use the statement Hope for the best, prepare for the worst when speaking with patients and families.60 For example, a physician could say, I really wish your mother could get better, and we should still hope for that; at the same time, we need to prepare for what will happen if she doesn't get better. Once again, this phrase both allows the patient or family to continue hoping things will improve and the clinician to support this hope, while simultaneously beginning the process of planning for the more likely outcome. Over time, the patient and family often move toward accepting that the patient is dying. Finally, trying to help the patient or family maintain hope in the face of illness and death is challenging but important: If your mother can't get better, are there other things you can hope for? Helping to identify tangible and realistic goals, such as being free from pain, seeing an important family member one last time, or getting home can provide hope at a difficult time.

Giving a Recommendation

Most patients with serious illness and their families want help making complex and ethically charged decisions. When clinicians ask patients to make decisions unilaterally, patients often feel anxious, sometimes for weeks.61 Families are often paralyzed when faced with the very difficult decision of whether to withdraw life‐sustaining interventions from a family member with an advanced terminal illness. Even if they understand on an intellectual level that ceasing to provide potentially curative or life‐prolonging therapies is the best choice, they are not yet able to accept this decision on an emotional level and ultimately may feel responsible and then guilty for the patient's death. Physicians need to carry some of the burden of making these difficult decisions. One way to relieve family members of some of this guilt is to recommend a plan of care based on substituted judgment, that is, on what the patient said she or he wanted or what the family thinks the patient would have wanted.6264 In addition, clinicians should use their medical expertise, experience, and understanding of the situation to make recommendations. The patient or family can then accept or reject the physician's advice, which maintains patient autonomy, yet not have to explicitly instruct the clinician to withdraw or limit life‐sustaining interventions.

The preceding discussion and recommendations can guide scenarios like those presented at the beginning of this article. In the second conversation, the clinician had just told the patient that he could help her to achieve her goal of pain control and of returning and staying home.

CONCLUSIONS

The decision about cardiopulmonary resuscitation is part of a larger conversation with a patient about how she or he wants to spend the rest of his/her life. Importantly, the decision should be made in context, rather than in isolation. Given the understanding that develops between physician and patient in the conversation above, it is not necessary to describe all the specific treatments that occur during CPR because the physician has already established that the patient does not want to return to the hospital, and she understands that she has a terminal condition and is dying. Through exploring a patient's goals and values, a clinician can discover a patient's preferences for care generally and come up with a comprehensive plan that addresses the particular individual's medical, social, and emotional needs. For physicians, few interventions are as important or rewarding as relieving patients' suffering and helping them attain their goals during and at the end of life.

As these conversations illustrate, discussing preferences regarding resuscitation is a challenging and important task for physicians. Understanding patients' wishes at the end of life allows clinicians to provide the type of care patients want, to avoid unwanted interventions, and to promote patient autonomy and dignity. Despite the frequency with which physicians have these conversations, they typically fall short when discussing code status with patients. Clinicians fail to address patients' concerns, monopolize conversations, and commonly misunderstand patients' resuscitation preferences.13 Often these discussions do not occur at all; more than 70% of seriously ill patients have never discussed advance directives with their physician.2, 4 The multicenter SUPPORT study, which demonstrated serious problems in the care of seriously ill hospitalized patients, documented that only 47% of physicians knew when their patients preferred do‐not‐resuscitate status.5

Hospitalists frequently conduct resuscitation discussions. Patients who are admitted to the hospital are usually seriously ill, and hospitalists need to assess rapidly whether an individual patient would want a resuscitation attempt if he or she had a cardiopulmonary arrest in the hospital. They need to build trust and rapport quickly with patients they have never met. Despite this challenge, hospitalists are in a good position to discuss resuscitation preferences.6 Patients may be more willing to discuss these issues in inpatient rather than outpatient settings because their acute deterioration may encourage self‐reflection.6 Furthermore, the time and productivity pressures of office practice often make it difficult for primary care physicians, who often know the patient and family best, to address advance directives. Although studies have documented that patients are interested in talking to their primary physicians about these matters,7, 8 these conversations do not occur with regularity. Preliminary research has raised the possibility that cancer patients may actually prefer to discuss these issues in an inpatient setting with a hospital‐based provider rather than with their oncologists.9 Studies have not addressed the question of whether patients with diseases other than cancer would prefer to discuss these issues with a hospitalist or their outpatient subspecialist.

Given that more than half of all Americans die in hospitals, hospitalists care for many people who are terminally ill and will need to assess preferences for cardiopulmonary resuscitation (CPR) and other treatments. Hospitalists need to be competent and compassionate in their approach to patients and their families. In this articles we review clinician barriers to holding these conversations, offer a variety of approaches to enhance these discussions, and review communication techniques that can be used to improve understanding.

Clinician Barriers to Discussing Resuscitation

Clinicians' own barriers may lead to infrequent and inadequate conversations about resuscitation. Understanding these barriers may allow providers to overcome them and facilitate better and more frequent communication. A discussion of patient barriers is beyond the scope of this article.

Laying Groundwork for the Discussion

The decision regarding resuscitation should be seen in the context of the patient's goals and values and overall health status. To address code status effectively, it is imperative first to elicit the person's view of his/her illness and then gently correct any misunderstandings. A patient who thinks her/his life will go on indefinitely may feel no need to consider her/his own mortality or alternatives to full resuscitation status. Alternatively, a patient who senses his/her mortality may have already thought about resuscitation and have clear preferences. A key first step in the conversation is to understand a patient's values and goals and comprehension of his/her illness. As in the second discussion above, a clinician can begin a discussion by saying, Tell me what you understand about your illness.

Discussing prognosis with patients can be difficult as physicians struggle with uncertainty. In the most comprehensive study to date of prognostication, physicians overestimated patient survival on average by a factor of 5.31 Nonetheless, there are compelling reasons to discuss prognosis. Failure to do so often results in patients spending their last days in the hospital receiving more aggressive treatments than they might choose if they understood their prognoses.32 Further, patients are denied the opportunity to address issues of life closure, such as spending time with family, thinking about legacy, and settling financial affairs. Physicians also fear they will take away hope with prognostication and believe patients expect greater accuracy than they can provide.33

Physicians can improve their prognostication skills by considering patients' functional status and clinical signs and symptoms and by using validated scales. The Karnofsky Performance Score (KPS) and the Eastern Cooperative Oncology Group (ECOG) Performance Status have been shown to correlate with survival,3440 and the Palliative Prognostic Score (PaP) has been validated in both cancer and noncancer patients.41, 4446 The PaP uses a combination of the KPS, clinical signs and symptoms, and the clinician's clinical prediction of survival. In addition, clinicians can be honest with patients about prognostic uncertainties and give prognoses in ranges, such as days to weeks, or weeks to months.

How to Broach the Subject of Resuscitation

In the first scenario presented in this article, the hospitalist says, There's something I need to talk with you about that I talk with all of my patients about. Although many clinicians begin resuscitation discussions this way, the question is problematic because often it is untrue. Most clinicians do not discuss code status with all patients who are admitted to the hospital. A better option would be to say, When I take care of patients with advanced cancer, I like to talk with them about their wishes regarding resuscitation. Is that all right? Better yet would be to ask a general question such as As you look ahead, what worries you most? or As you look to the future, what do you hope for? These and other useful questions appear in Table 1. These questions allow patients to bring up their concerns, show that the clinician cares about them, and often segue into a discussion of patients' hopes and worries about their own death and dying process. These questions often allow patients to bring up important goals that bear directly on the issue of resuscitation. For example, in the second scenario at the beginning of the article, the patient says she wants to have quality time with her family at home and does not want life‐sustaining technologies. Such a patient may not want resuscitation. When discussing CPR, it can also be helpful to state explicitly that resuscitation is used when a patient has died, rather than to use euphemisms such as, If your heart were to stop. The clinician can ask explicitly, If you were to die, would you want? There are other strategies for introducing the subject of resuscitation if these questions do not work. If a patient seems uncomfortable with the conversation, the clinician can address this discomfort directly by saying, This conversation seems to make you uncomfortable. Other strategies for exploring these issues include inquiring if the person has ever discussed resuscitation with his/her family or another physician, or asking if anyone else in the family has been very sick. Additionally, clinicians can ask questions about surrogate decision making. If you were to get so sick that you were unable to make decisions for yourself, who would you want to make them for you? This can then lead into a discussion of whether the patient has spoken to that person about his/her preferences, and if so, what was the content of that conversation. Another useful question is, Is there any state that would be worse than death for you? This question focuses on outcomes and allows the physician to put the issue of resuscitation into perspective for a patient.

Knowing the Facts about Cardiopulmonary Resuscitation

In general, about 1 in 7 patients who have undergone CPR in the hospital survive to hospital discharge. Two literature reviews, from 1989 and 1998, of hospitalized patients who underwent CPR in the hospital reported surprisingly similar statistics. Immediate survival in both series was 41%, and survival to hospital discharge was 13%‐14%. Factors that increased survival included arrest due to coronary artery disease or drug overdose/adverse reaction, and location of arrest in the intensive care unit (ICU). Factors that decreased survival included sepsis at time of arrest, cancer, whether localized or metastatic, dementia, acute renal failure, bed‐bound status, and residence in a nursing home. Neither age nor sex was associated with survival.47, 48 In a meta‐analysis of outcome studies of CPR in metastatic cancer patients, none of 117 patients survived to hospital discharge.49

Most people get their information about CPR survival rates from the mass media, which portray CPR as a very successful procedure. For instance, on television, the sole source of information on CPR for most patients, the rate of survival to hospital discharge is 64%, much higher than the 1314% cited in the medical literature.50 Thus, it is no surprise that a patient with metastatic cancer or another life‐limiting illness would assume a positive outcome with CPR.

Knowledge of the facts about CPR survival rates is key when a physician discusses code status with a patient because these data influence patients' decisions. Patients who have a realistic understanding of their chances are less likely to prefer resuscitation.51, 52 Offering patients information about the success rate for CPR in their particular situation is critical in helping them reach a decision consistent with their values and goals. This information can be given quantitatively or qualitatively depending on the clinician assessment of what the patient would prefer.

Avoiding Stark Dichotomies in Resuscitation Discussions

In clinical practice there are 3 ways in which physicians can present resuscitation decisions as black and white and ignore the shades of gray. First, clinicians may present the choice between resuscitation and do not resuscitate (DNR) as a choice between life and death. In the first scenario above, the physician states If your heart were to stop, do you want us to do everything to bring you back? implying he would be able to save the patient's life with resuscitation attempts. When discussing resuscitation, clinicians should avoid language that suggests such a stark dichotomy. The reality is that most patients die despite resuscitation attempts. In fact, a patient is actually choosing between certain death (without resuscitation) and likely death (if resuscitation is attempted). For a patient with a serious, life‐limiting illness, it may be more effective to frame the conversation in terms of how that person envisions the end of life, and not whether death will eventually occur.

Second, clinicians and patients sometimes equate DNR with doing nothing or giving up. Clinicians fail to discuss other treatment options or alternatives, such as continuing ongoing aggressive medical treatments with DNR status or pursuing palliative care. Performing resuscitation is equated with activity and treatment, whereas withholding resuscitation is seen as passivity and giving up. To the patient, this can feel like abandonment, as if the doctor is withholding a treatment and not offering anything in its place. Examples of positive phrases that demonstrate the physician will continue to offer excellent care include: We will continue maximal medical therapy. However, if you die despite these treatments, we will let you die peacefully and won't attempt to revive you; We'll continue the intravenous antibiotics, but we won't plan to move you to the ICU if things worsen53; and We will work hard to treat your pain and other symptoms and to get you home. In addition, hospitalists must ensure, when signing out to physician colleagues, nurses, respiratory therapists, and others, that DNR orders are not overinterpreted to mean no treatment. Although a DNR order states that in the event of a cardiac arrest, no attempt at resuscitation will be made, it should have no bearing on other appropriate and desired interventions, including antibiotics, chemotherapy, and artificial nutrition; treatment for pain and other symptoms; and even monitoring in an ICU. This misunderstanding of DNR status is common among health care providers and has led many to argue that DNR orders should be part of a more comprehensive treatment plan that outlines where the patient's wishes for treatment fall on the spectrum from otherwise aggressive measures to comfort care.54, 55 Physicians who have a clear understanding of a DNR order will be able to reassure a patient that they will continue to receive desired care, but that if the patient dies, no attempt to resuscitate will be made.

Third, it is important to remind patients who choose full resuscitation status that additional decisions will need to be made if resuscitation is successful. Most patients who survive cardiopulmonary arrest end up worse off clinically and spend time in an ICU with life‐support measures in place, such as mechanical ventilation and vasopressors. Even if they survive, there will likely be a period during which they are unable to speak for themselves. This situation puts the burden of decision making on their surrogates or an appointed durable power of attorney for health care (DPOA‐HC). It is important to ask patients ahead of time whether there are conditions under which they would not want ongoing life‐sustaining measures. For example, a person might opt for discontinuation of life‐support measures if the physician and family agreed that there was only a minimal chance of cognitive recovery existed and that ongoing support was only prolonging inevitable death. To clarify the patient's wishes in this situation, you might ask, Are there conditions that would be worse than death? Encouraging the patient to share his/her wishes in this situation with a surrogate or DPOA‐HC will help to ensure those wishes are respected.

Communication Techniques

When discussing advance directives, it is important to give patients the chance to describe their life goals and their values to establish a context for understanding the role of life‐sustaining treatments. One useful method to elicit these goals and values is to ask open‐ended questions, followed by periods of silence so the patient has time to express himself/herself. In the second scenario, above, the physician used open‐ended questions in several instances: Tell me what you understand about your illness and As you look ahead, what worries you most? Tulsky and colleagues documented that medical residents spoke 76% of the time in discussions with patients about code status.30 In an ideal case, this ratio should be reversed or at least be even, allowing patients the time to explain their thoughts. Acknowledging patients' emotions by stating simply, You seem [angry/sad/perplexed], and waiting for an answer can help patients feel they are being understood. Making empathic statements is another powerful communication technique that conveys understanding.56 Examples include, That must be really sad for you, and It must be frightening to be in so much pain. As noted, silence can also be a powerful tool. Clinicians tend to be uncomfortable with silence and so fill the gaps with words. Allowing for silence enables patients to digest what they have heard, encourages them to continue speaking, and shows them the clinician wants to hear what they have to say. When giving information about any medical issues and especially about CPR, it is important to explain concepts in lay terms and to avoid medical jargon.57 Additionally, nonverbal communication techniques such as making eye contact, head nodding, and leaning in toward the patient all help in communicating engagement in the conversation. Having the conversation in a quiet and private place and sitting at the same level as the patient or family is also important. It is always a good idea to check in with patients to assess their understanding. Simply asking, Do you have any questions about what I said? or Does that all make sense? gives patients the opportunity to ask for clarification. Attempting to summarize what a patient has said can also help to clarify misunderstandings. Useful phrases include, Let me see if I've gotten this right or I want to make sure that I understood what you're telling me, followed by the clinician's synopsis of important points discussed.58 A summary of important communication techniques can be found in Table 2.

There are 2 additional statements that can be very useful when patients and families are struggling with the reality of severe illness and are still hoping for longer life and cure. The first is the I wish statement, in which the clinician allies himself or herself with the patient's or family's wishes by stating, I wish it were different. I wish there were more chemotherapy we could give you that would make a difference.59 Occasionally, when tension is developing because the clinician does not believe an intervention is warranted but the patient desires it nonetheless, the I wish statement can be a powerful way of realigning with the patient. For example, responding to a patient who says, I want chemo to cure my cancer, with the statement There is no chemotherapy to help you can seem antagonistic. In contrast, saying, I wish there were a treatment that would make your cancer go away aligns the clinician with the patient and supports the patient. Another advantage of I wish statements is that they are truethe clinician does wish that there was an effective treatment and would gladly provide it if available. In general, I wish statements are more effective than I'm sorry statements, because the latter can be interpreted as the clinician taking responsibility for the situation. When a provider says, I'm sorry to give you this bad news, the patient may feel the need to say, That's OK. On the other hand, saying, I'm so sorry that your mother is dying, is very human and unlikely to be interpreted as the clinician taking responsibility for the death.

A similar technique is to use the statement Hope for the best, prepare for the worst when speaking with patients and families.60 For example, a physician could say, I really wish your mother could get better, and we should still hope for that; at the same time, we need to prepare for what will happen if she doesn't get better. Once again, this phrase both allows the patient or family to continue hoping things will improve and the clinician to support this hope, while simultaneously beginning the process of planning for the more likely outcome. Over time, the patient and family often move toward accepting that the patient is dying. Finally, trying to help the patient or family maintain hope in the face of illness and death is challenging but important: If your mother can't get better, are there other things you can hope for? Helping to identify tangible and realistic goals, such as being free from pain, seeing an important family member one last time, or getting home can provide hope at a difficult time.

Giving a Recommendation

Most patients with serious illness and their families want help making complex and ethically charged decisions. When clinicians ask patients to make decisions unilaterally, patients often feel anxious, sometimes for weeks.61 Families are often paralyzed when faced with the very difficult decision of whether to withdraw life‐sustaining interventions from a family member with an advanced terminal illness. Even if they understand on an intellectual level that ceasing to provide potentially curative or life‐prolonging therapies is the best choice, they are not yet able to accept this decision on an emotional level and ultimately may feel responsible and then guilty for the patient's death. Physicians need to carry some of the burden of making these difficult decisions. One way to relieve family members of some of this guilt is to recommend a plan of care based on substituted judgment, that is, on what the patient said she or he wanted or what the family thinks the patient would have wanted.6264 In addition, clinicians should use their medical expertise, experience, and understanding of the situation to make recommendations. The patient or family can then accept or reject the physician's advice, which maintains patient autonomy, yet not have to explicitly instruct the clinician to withdraw or limit life‐sustaining interventions.

The preceding discussion and recommendations can guide scenarios like those presented at the beginning of this article. In the second conversation, the clinician had just told the patient that he could help her to achieve her goal of pain control and of returning and staying home.

CONCLUSIONS

The decision about cardiopulmonary resuscitation is part of a larger conversation with a patient about how she or he wants to spend the rest of his/her life. Importantly, the decision should be made in context, rather than in isolation. Given the understanding that develops between physician and patient in the conversation above, it is not necessary to describe all the specific treatments that occur during CPR because the physician has already established that the patient does not want to return to the hospital, and she understands that she has a terminal condition and is dying. Through exploring a patient's goals and values, a clinician can discover a patient's preferences for care generally and come up with a comprehensive plan that addresses the particular individual's medical, social, and emotional needs. For physicians, few interventions are as important or rewarding as relieving patients' suffering and helping them attain their goals during and at the end of life.

Most people believe the term hospitalist first appeared in the literature in the August 15, 1996, issue of the New England Journal of Medicine (NEJM). That issue carried an article that Lee Goldman and I wrote titled The Emerging Role of Hospitalists in the American Health Care System.1 But the term was actually coined about a year earlier, in an article I wrote for our University of California, San Francisco (UCSF), residents' newsletter, the Medical Residents' Progress Note (MRPN), circulation about 180. In that article, I mused about a new model of care in which separate physicians assumed the role of caring for inpatients in place of patients' primary care doctors. Several peopleboth residents and facultyapproached me soon after the MRPN article was published and said, I read your articleyou should really buff it up and send it to a real journal. (By the way, when you publish a scholarly article, people generally say, I saw your article, rather than I read your article). This prompting led me to polish up the piece, with Lee Goldman's able assistance, and send it to the NEJM.

Although people often introduce me today as the guy who invented hospitalists (to which I typically respond, yeah, just like Al Gore invented the Internet), I did no such thingI merely kept my eyes and ears open, spotted the trend early, and gave it a name that stuck. In the mid‐1990s, the California market was being besieged by managed care, which was seeking new ways to cut hospital utilization and costs. In 1994, the huge Kaiser Permanente system decided to reorganize its hospital care around a cadre of hospital‐based specialists (HBSs), essentially dichotomizing the roles of inpatient and outpatient physicians. (Interestingly, Kaiser's main motivations were to improve outpatient satisfaction by assuring constant availability of primary care physicians and to create a vehicle to promote inpatient quality improvement activities, not necessarily to improve inpatient efficiency.) Around the same time, I read reports in throwaway magazines about Park Nicollet in Minneapolis and the Scripps Clinic in La Jolla, California, doing the same thing. Then one day I heard that a talented young UCSF faculty member was leaving our VA system to take a job as the inpatient manager at a local community teaching hospital. A few weeks later, I took him out to lunchI was intrigued by this new role and wanted to better understand it. As he described it to me over sandwiches, it made all the sense in the world, and the seeds of the MRPNand later NEJMarticle was planted.

I have always had an abiding interest in the notion of valuea fundamental belief that our system is inexorably becoming one in which health care choices and competition will be based on demonstrable quality, safety, the patient's experience, and cost rather than on tradition, impression, and proximity. As I began thinking about hospital care, it seemed likely this new modeldichotomizing the roles of inpatient and outpatient doctors such that the former could be constantly available and become an expert in inpatient clinical care and hospital microsystemswould provide more value than the traditional structure, both in community settings (replacing the single primary care doctor managing both inpatients and outpatients) and the academic setting (replacing the traditional one‐month‐a‐year ward attending).

At the time I was thinking all this through, a new chairman of our department of medicine arrived from Harvard. Lee Goldman, who virtually invented the field of clinical epidemiology, came to UCSF with a powerful vision that matched mineto transform training and clinical care to improve both value and education. Lee had been a resident at UCSF 20 years earlier and returned in 1995 to an inpatient service whose structure and culture had barely changed over a generation. Lee (who, to my great chagrin, recently left UCSF to become Columbia's medical school dean, and who does not have the term good enough in his vocabulary) sat down with me and articulated his vision for a new type of academic inpatient model, led by faculty who cared for inpatients and taught trainees hospital medicine for a living. This was entirely in sync with my thoughts, and so we set out to build it.

Reaction to both the New England Journal of Medicine article and our vision for an academic hospitalist service was swift and negative. One letter to the NEJM said it all:

Our response to this and the other letters emphasized the need for evidence:

My father, a retired businessman living in Florida, brought the controversy to an even finer point a year later. I met this guy playing tennis today, he told me on the phone one day. And he's heard of you! I listened for the heartwarming sounds of fatherly pride, but none were forthcoming. He hates you, he added.

Our attempts to build an academic hospitalist program generated other concerns. Many faculty enjoyed serving as ward attendings and worried about being kicked off the wards (although many privately told me that they knew their time was up and were grateful for a way to exit with dignity.) One world‐famous faculty member bolted out of his seat during the Q&A period after my medical grand rounds at his institution in 1997. How will the house staff learn anything if we don't allow them to learn from their mistakes? he huffed. (I told him that I was flying cross‐country the next day, and I'll be really pissed off if my pilot is there to learn from his mistakes.) Our residents also worried terribly about losing their autonomy, having these bright young attendings breathing down our backs. Everyone worried about where the resources to pay for the program would come from.

At UCSF, our strategy was to reassure everyone that we would be measuring the impact of the new model in terms of cost, quality, patient satisfaction, and education. By making clear that the results of this research would guide further change (and that we were willing to end the experiment if it turned out negatively), the faculty and house staff largely suspended their disbelief for the first year. That study4 would demonstrate impressive cost savings with no adverse impact on quality and patient satisfaction and a hint of improved resident satisfaction (later proven more conclusively5), allowing us to expand the program over time and to make the argument for ongoing medical center support of the new model.

Just as Lee Goldman's arrival at UCSF in 1995 was a remarkable and crucial bit of serendipity, my partnership with Dr. Win Whitcomb and Dr. John Nelson was every bit as important for the growth of the movement nationally. John, at that time a young internist in Gainesville, Florida, had been practicing as a hospitalist (though it wasn't called that) since completing his internal medicine residency in the later 1980s. He had hooked up with Win, another young internist who had left a private practice job to begin a hospital‐based practice at Mercy Medical Center in Springfield, Massachusetts. Together the two of them had begun to network with the handful of physicians around the country who were practicing in this new model. But they needed a larger megaphone, both to let other hospitalists know about each other, and to make hospitals and systems more aware of this new model of care.

John tells the story of pulling the August 15, 1996, issue of the NEJM out of his mailbox, seeing my article, and literally running into to his house to tell his wife that his practice had finally been discovered. John's thoughtful exuberance is one of the reasons for the growth of our field, and he did something that is uniquely Johncalling the author of an article that piqued his interest to discuss its contents, something he'd been doing for years. At that point, Lee Goldman was an internationally known leader in internal medicine; as chair of a major academic department, he had several layers of administrative assistants running interference when he received cold calls. I, on the other hand, ran a sleepy medical service and had little to do other than to answer calls and to respond to this new thing called e‐mail. John didn't know that; in his experience, first authors of articles in major journals were nearly always too busy to answer calls from country docs like him. So he tried Lee Goldman first but failed to get through. Win, on the other hand, decided to call me and had no problem getting through immediately. We hit it off like we'd been buddies for decades, sharing our instinctive recognition that that we were at the cutting edge of a new specialty. In what, in retrospect, seems like an extraordinary amount of hubris, we essentially divided up the world, asking the question: what does an emerging specialty need in order to be successful? I'm reminded of one of my favorite parts of the brilliant dialogue by Mel Brooks and Carl Reiner, The 2000 Year Old Man. Brooks, playing the title part, describes his relationship with Joan of Arc (What a cutie, he gushes) to Reiner (playing the interviewer), and how Joan's mission got in the way of their ardor. She used to say to me, Ive got to save France,' says the 2000 Year Old Man. I said, Look, Ive got to wash up, you save France, I'll see you later' That was usWin and John agreed to focus on building a new professional society and on networking with community‐based hospitalists, while I emphasized the academic side of things: organizing meetings, developing training programs, publishing a textbook, and launching a research agenda.

The first national gathering of hospitalists was astonishing. In the spring of 1997, I hosted what I thought would be a small hospital medicine CME meeting at a Holiday Inn in San Francisco in a seedy part of town. I expected about 50 people to attend and was shocked to see 3 times that (plus several homeless people who wandered into the sessions). Most remarkably, at the end of day 1, following 8 hours of clinical lectures, Win, John, and I asked the attendees if anybody wanted to stay a while and discuss the possibility of forming a new society. To our amazement, virtually everybody stayedmore than 100 people! Would anybody be willing to contribute some money to get this started? asked John, expecting nothing. And people began passing $20 bills up to the front of the room. That was the moment we knew we were onto something very bigthe atmosphere was electric, the enthusiasm easily palpable.

We initially called the new society the National Association of Inpatient Physicians (NAIP), as the name hospitalist was still very controversial, and many thought it would not have legsthe term inpatient physician was believed to be more inclusive. NAIP rapidly reached a crucial turning point. Our few hundred dollars in dues collections and ad revenues lived in Win's shoebox in Massachusetts (and later in a checking account opened by Ron Angus in Texas), and Win, John, and I were keeping databases of hospitalists on our computers and the backs of napkins. It was clear we needed to either create a full‐fledged infrastructure or partner with an organization that could help us. I approached Hal Sox, now the editor of the Annals of Internal Medicine but at that time president of the American College of Physicians (and an old fellowship mentor of mine), about the possibility of NAIP establishing a formal relationship with the ACP. Hal was reluctant at first, noting many ACP members were pretty strongly against the idea of hospitalists. In one of many acts of brinksmanship, I told him we would need to look for other partners if ACP did not get over its ambivalence and embrace our new field. To his credit and to the credit particularly of Dr. Walt McDonald, ACP's executive director at the time, both recognized the potential growth of this new field and worked through the internal politics to offer us an affiliation. However, we found their initial offerto become the Section on Hospital Medicine within the ACPunattractive. Wanting to be a full‐fledged independent organization that enjoyed a relationship with the College, we proposed a relationship that would link us and allow ACP to support our infrastructure, but that allowed us to retain independent decision making, governance, and budget. John, in his charming Southern drawl, described our position to an early gathering of about 100 hospitalists at a NAIP meeting in San Diego. Their offer would have them up here, and we'd be down there, he said, his hands depicting an obvious hierarchy, with us on the bottom. But we insisted on being equal partners, he said, with his hands on the same plane. I turned to Win, sitting next to me in the audience, and whispered something like, Yeah, equalexcept for the small fact that they have 120,000 members and we have 87. Nevertheless, they agreed, and our relationship has been incredibly positive for hospitalists, and I believe for the ACP as well.

The rest, as they say, is history. The society, renamed the Society of Hospital Medicine in April 2003, has thrived under the leadership of a strong series of boards, a wonderful staff, and a charismatic and highly effective CEO, Dr. Larry Wellikson. We successfully navigated the many early challenges and took advantage of key opportunities. In this regard, I consider our 3 most important decisions and actions to be: 1) creating a body of research that demonstrated, in an evidence‐based way, that the theoretical promise of the field was real6 (it was this research that led hospitals to embrace the field more vigorously and that justified the crucial support that most hospitals provide their hospitalist programs); 2) vigorously pushing back against managed care‐based hospitalist models that had begun to force primary care physicians to hand their patients off to hospitalists against their will (NAIP's first policy pronouncement was to come out strongly against such mandatory models, which seemed counterintuitive to some but which markedly decreased our vulnerability to being tagged as a cost‐cutting vehicle of managed care); and 3) linking ourselves as strongly as possible with the growing quality and safety movements. When the IOM reports on medical errors7 and later quality8 were published, we immediately saw in the new agendas a tremendous opportunity to brand hospitalists as indispensable leaders of quality and safety in hospitalsanother key rationale for hospitalists' value proposition and another reason for hospitals and policymakers to support the young field.

Looking back at the 1996 New England Journal of Medicine article, I am struck by both the number of things I got right (even a blind squirrel) and the number that I did not anticipate or got wrong. Lee and I thought that many hospitalists would be subspecialists who would focus on hospital medicine for only part of their work. This was true early on, but the field has evolved to be more of a generalist endeavor (although recently there have emerged laborists, neurology hospitalists, and even surgical hospitalists). I probably could have anticipated the growth of the field in pediatrics, but it certainly was not on my radar screen until years later.9 I did not count on the work hours of house staff being regulated; even if I had, I'm not sure I would have fully recognized how the need to create nonteaching services would turbo‐charge the growth of the hospitalist field in teaching hospitals. The one mild disappointment: I anticipated stronger evidence by now of the field's salutary impact on safety and quality. The effort to study and hopefully demonstrate such improvements should be a major focus for the next 510 years. Finally, although I thought the field would grow rapidly, I did not anticipate that a decade later there would be 15,000 hospitalists nationally or 24 in my group at UCSF. I also did not guess that an April 2006 Medline search of hospitalist would find 561 articles or that a Google search of hospitalist would yield 689,000 entries (hell, there was no Google to search in 1996!).

As I reflect back on the last decade, I am humbled by the remarkable work I have seen from hospitalists around the country and grateful for the wonderful friendships I have enjoyed with my colleagues in our new field. I am even more convinced of the fundamental accuracy of my underlying premise: the U.S. health care system will increasingly embrace models, strategies, and providers who can demonstrably improve the value of care. I have no doubt thatcollectivelyAmerican hospitalists have saved tens of thousands of lives, prevented tens of thousands of errors, orchestrated tens of thousands of good deaths, comforted tens of thousands of families, and saved billions of dollars. It is an ongoing legacy that gives me considerable pride and joy.

Most people believe the term hospitalist first appeared in the literature in the August 15, 1996, issue of the New England Journal of Medicine (NEJM). That issue carried an article that Lee Goldman and I wrote titled The Emerging Role of Hospitalists in the American Health Care System.1 But the term was actually coined about a year earlier, in an article I wrote for our University of California, San Francisco (UCSF), residents' newsletter, the Medical Residents' Progress Note (MRPN), circulation about 180. In that article, I mused about a new model of care in which separate physicians assumed the role of caring for inpatients in place of patients' primary care doctors. Several peopleboth residents and facultyapproached me soon after the MRPN article was published and said, I read your articleyou should really buff it up and send it to a real journal. (By the way, when you publish a scholarly article, people generally say, I saw your article, rather than I read your article). This prompting led me to polish up the piece, with Lee Goldman's able assistance, and send it to the NEJM.

Although people often introduce me today as the guy who invented hospitalists (to which I typically respond, yeah, just like Al Gore invented the Internet), I did no such thingI merely kept my eyes and ears open, spotted the trend early, and gave it a name that stuck. In the mid‐1990s, the California market was being besieged by managed care, which was seeking new ways to cut hospital utilization and costs. In 1994, the huge Kaiser Permanente system decided to reorganize its hospital care around a cadre of hospital‐based specialists (HBSs), essentially dichotomizing the roles of inpatient and outpatient physicians. (Interestingly, Kaiser's main motivations were to improve outpatient satisfaction by assuring constant availability of primary care physicians and to create a vehicle to promote inpatient quality improvement activities, not necessarily to improve inpatient efficiency.) Around the same time, I read reports in throwaway magazines about Park Nicollet in Minneapolis and the Scripps Clinic in La Jolla, California, doing the same thing. Then one day I heard that a talented young UCSF faculty member was leaving our VA system to take a job as the inpatient manager at a local community teaching hospital. A few weeks later, I took him out to lunchI was intrigued by this new role and wanted to better understand it. As he described it to me over sandwiches, it made all the sense in the world, and the seeds of the MRPNand later NEJMarticle was planted.

I have always had an abiding interest in the notion of valuea fundamental belief that our system is inexorably becoming one in which health care choices and competition will be based on demonstrable quality, safety, the patient's experience, and cost rather than on tradition, impression, and proximity. As I began thinking about hospital care, it seemed likely this new modeldichotomizing the roles of inpatient and outpatient doctors such that the former could be constantly available and become an expert in inpatient clinical care and hospital microsystemswould provide more value than the traditional structure, both in community settings (replacing the single primary care doctor managing both inpatients and outpatients) and the academic setting (replacing the traditional one‐month‐a‐year ward attending).

At the time I was thinking all this through, a new chairman of our department of medicine arrived from Harvard. Lee Goldman, who virtually invented the field of clinical epidemiology, came to UCSF with a powerful vision that matched mineto transform training and clinical care to improve both value and education. Lee had been a resident at UCSF 20 years earlier and returned in 1995 to an inpatient service whose structure and culture had barely changed over a generation. Lee (who, to my great chagrin, recently left UCSF to become Columbia's medical school dean, and who does not have the term good enough in his vocabulary) sat down with me and articulated his vision for a new type of academic inpatient model, led by faculty who cared for inpatients and taught trainees hospital medicine for a living. This was entirely in sync with my thoughts, and so we set out to build it.

Reaction to both the New England Journal of Medicine article and our vision for an academic hospitalist service was swift and negative. One letter to the NEJM said it all:

Our response to this and the other letters emphasized the need for evidence:

My father, a retired businessman living in Florida, brought the controversy to an even finer point a year later. I met this guy playing tennis today, he told me on the phone one day. And he's heard of you! I listened for the heartwarming sounds of fatherly pride, but none were forthcoming. He hates you, he added.

Our attempts to build an academic hospitalist program generated other concerns. Many faculty enjoyed serving as ward attendings and worried about being kicked off the wards (although many privately told me that they knew their time was up and were grateful for a way to exit with dignity.) One world‐famous faculty member bolted out of his seat during the Q&A period after my medical grand rounds at his institution in 1997. How will the house staff learn anything if we don't allow them to learn from their mistakes? he huffed. (I told him that I was flying cross‐country the next day, and I'll be really pissed off if my pilot is there to learn from his mistakes.) Our residents also worried terribly about losing their autonomy, having these bright young attendings breathing down our backs. Everyone worried about where the resources to pay for the program would come from.

At UCSF, our strategy was to reassure everyone that we would be measuring the impact of the new model in terms of cost, quality, patient satisfaction, and education. By making clear that the results of this research would guide further change (and that we were willing to end the experiment if it turned out negatively), the faculty and house staff largely suspended their disbelief for the first year. That study4 would demonstrate impressive cost savings with no adverse impact on quality and patient satisfaction and a hint of improved resident satisfaction (later proven more conclusively5), allowing us to expand the program over time and to make the argument for ongoing medical center support of the new model.

Just as Lee Goldman's arrival at UCSF in 1995 was a remarkable and crucial bit of serendipity, my partnership with Dr. Win Whitcomb and Dr. John Nelson was every bit as important for the growth of the movement nationally. John, at that time a young internist in Gainesville, Florida, had been practicing as a hospitalist (though it wasn't called that) since completing his internal medicine residency in the later 1980s. He had hooked up with Win, another young internist who had left a private practice job to begin a hospital‐based practice at Mercy Medical Center in Springfield, Massachusetts. Together the two of them had begun to network with the handful of physicians around the country who were practicing in this new model. But they needed a larger megaphone, both to let other hospitalists know about each other, and to make hospitals and systems more aware of this new model of care.

John tells the story of pulling the August 15, 1996, issue of the NEJM out of his mailbox, seeing my article, and literally running into to his house to tell his wife that his practice had finally been discovered. John's thoughtful exuberance is one of the reasons for the growth of our field, and he did something that is uniquely Johncalling the author of an article that piqued his interest to discuss its contents, something he'd been doing for years. At that point, Lee Goldman was an internationally known leader in internal medicine; as chair of a major academic department, he had several layers of administrative assistants running interference when he received cold calls. I, on the other hand, ran a sleepy medical service and had little to do other than to answer calls and to respond to this new thing called e‐mail. John didn't know that; in his experience, first authors of articles in major journals were nearly always too busy to answer calls from country docs like him. So he tried Lee Goldman first but failed to get through. Win, on the other hand, decided to call me and had no problem getting through immediately. We hit it off like we'd been buddies for decades, sharing our instinctive recognition that that we were at the cutting edge of a new specialty. In what, in retrospect, seems like an extraordinary amount of hubris, we essentially divided up the world, asking the question: what does an emerging specialty need in order to be successful? I'm reminded of one of my favorite parts of the brilliant dialogue by Mel Brooks and Carl Reiner, The 2000 Year Old Man. Brooks, playing the title part, describes his relationship with Joan of Arc (What a cutie, he gushes) to Reiner (playing the interviewer), and how Joan's mission got in the way of their ardor. She used to say to me, Ive got to save France,' says the 2000 Year Old Man. I said, Look, Ive got to wash up, you save France, I'll see you later' That was usWin and John agreed to focus on building a new professional society and on networking with community‐based hospitalists, while I emphasized the academic side of things: organizing meetings, developing training programs, publishing a textbook, and launching a research agenda.

The first national gathering of hospitalists was astonishing. In the spring of 1997, I hosted what I thought would be a small hospital medicine CME meeting at a Holiday Inn in San Francisco in a seedy part of town. I expected about 50 people to attend and was shocked to see 3 times that (plus several homeless people who wandered into the sessions). Most remarkably, at the end of day 1, following 8 hours of clinical lectures, Win, John, and I asked the attendees if anybody wanted to stay a while and discuss the possibility of forming a new society. To our amazement, virtually everybody stayedmore than 100 people! Would anybody be willing to contribute some money to get this started? asked John, expecting nothing. And people began passing $20 bills up to the front of the room. That was the moment we knew we were onto something very bigthe atmosphere was electric, the enthusiasm easily palpable.

We initially called the new society the National Association of Inpatient Physicians (NAIP), as the name hospitalist was still very controversial, and many thought it would not have legsthe term inpatient physician was believed to be more inclusive. NAIP rapidly reached a crucial turning point. Our few hundred dollars in dues collections and ad revenues lived in Win's shoebox in Massachusetts (and later in a checking account opened by Ron Angus in Texas), and Win, John, and I were keeping databases of hospitalists on our computers and the backs of napkins. It was clear we needed to either create a full‐fledged infrastructure or partner with an organization that could help us. I approached Hal Sox, now the editor of the Annals of Internal Medicine but at that time president of the American College of Physicians (and an old fellowship mentor of mine), about the possibility of NAIP establishing a formal relationship with the ACP. Hal was reluctant at first, noting many ACP members were pretty strongly against the idea of hospitalists. In one of many acts of brinksmanship, I told him we would need to look for other partners if ACP did not get over its ambivalence and embrace our new field. To his credit and to the credit particularly of Dr. Walt McDonald, ACP's executive director at the time, both recognized the potential growth of this new field and worked through the internal politics to offer us an affiliation. However, we found their initial offerto become the Section on Hospital Medicine within the ACPunattractive. Wanting to be a full‐fledged independent organization that enjoyed a relationship with the College, we proposed a relationship that would link us and allow ACP to support our infrastructure, but that allowed us to retain independent decision making, governance, and budget. John, in his charming Southern drawl, described our position to an early gathering of about 100 hospitalists at a NAIP meeting in San Diego. Their offer would have them up here, and we'd be down there, he said, his hands depicting an obvious hierarchy, with us on the bottom. But we insisted on being equal partners, he said, with his hands on the same plane. I turned to Win, sitting next to me in the audience, and whispered something like, Yeah, equalexcept for the small fact that they have 120,000 members and we have 87. Nevertheless, they agreed, and our relationship has been incredibly positive for hospitalists, and I believe for the ACP as well.

The rest, as they say, is history. The society, renamed the Society of Hospital Medicine in April 2003, has thrived under the leadership of a strong series of boards, a wonderful staff, and a charismatic and highly effective CEO, Dr. Larry Wellikson. We successfully navigated the many early challenges and took advantage of key opportunities. In this regard, I consider our 3 most important decisions and actions to be: 1) creating a body of research that demonstrated, in an evidence‐based way, that the theoretical promise of the field was real6 (it was this research that led hospitals to embrace the field more vigorously and that justified the crucial support that most hospitals provide their hospitalist programs); 2) vigorously pushing back against managed care‐based hospitalist models that had begun to force primary care physicians to hand their patients off to hospitalists against their will (NAIP's first policy pronouncement was to come out strongly against such mandatory models, which seemed counterintuitive to some but which markedly decreased our vulnerability to being tagged as a cost‐cutting vehicle of managed care); and 3) linking ourselves as strongly as possible with the growing quality and safety movements. When the IOM reports on medical errors7 and later quality8 were published, we immediately saw in the new agendas a tremendous opportunity to brand hospitalists as indispensable leaders of quality and safety in hospitalsanother key rationale for hospitalists' value proposition and another reason for hospitals and policymakers to support the young field.

Looking back at the 1996 New England Journal of Medicine article, I am struck by both the number of things I got right (even a blind squirrel) and the number that I did not anticipate or got wrong. Lee and I thought that many hospitalists would be subspecialists who would focus on hospital medicine for only part of their work. This was true early on, but the field has evolved to be more of a generalist endeavor (although recently there have emerged laborists, neurology hospitalists, and even surgical hospitalists). I probably could have anticipated the growth of the field in pediatrics, but it certainly was not on my radar screen until years later.9 I did not count on the work hours of house staff being regulated; even if I had, I'm not sure I would have fully recognized how the need to create nonteaching services would turbo‐charge the growth of the hospitalist field in teaching hospitals. The one mild disappointment: I anticipated stronger evidence by now of the field's salutary impact on safety and quality. The effort to study and hopefully demonstrate such improvements should be a major focus for the next 510 years. Finally, although I thought the field would grow rapidly, I did not anticipate that a decade later there would be 15,000 hospitalists nationally or 24 in my group at UCSF. I also did not guess that an April 2006 Medline search of hospitalist would find 561 articles or that a Google search of hospitalist would yield 689,000 entries (hell, there was no Google to search in 1996!).

As I reflect back on the last decade, I am humbled by the remarkable work I have seen from hospitalists around the country and grateful for the wonderful friendships I have enjoyed with my colleagues in our new field. I am even more convinced of the fundamental accuracy of my underlying premise: the U.S. health care system will increasingly embrace models, strategies, and providers who can demonstrably improve the value of care. I have no doubt thatcollectivelyAmerican hospitalists have saved tens of thousands of lives, prevented tens of thousands of errors, orchestrated tens of thousands of good deaths, comforted tens of thousands of families, and saved billions of dollars. It is an ongoing legacy that gives me considerable pride and joy.

Most people believe the term hospitalist first appeared in the literature in the August 15, 1996, issue of the New England Journal of Medicine (NEJM). That issue carried an article that Lee Goldman and I wrote titled The Emerging Role of Hospitalists in the American Health Care System.1 But the term was actually coined about a year earlier, in an article I wrote for our University of California, San Francisco (UCSF), residents' newsletter, the Medical Residents' Progress Note (MRPN), circulation about 180. In that article, I mused about a new model of care in which separate physicians assumed the role of caring for inpatients in place of patients' primary care doctors. Several peopleboth residents and facultyapproached me soon after the MRPN article was published and said, I read your articleyou should really buff it up and send it to a real journal. (By the way, when you publish a scholarly article, people generally say, I saw your article, rather than I read your article). This prompting led me to polish up the piece, with Lee Goldman's able assistance, and send it to the NEJM.

Although people often introduce me today as the guy who invented hospitalists (to which I typically respond, yeah, just like Al Gore invented the Internet), I did no such thingI merely kept my eyes and ears open, spotted the trend early, and gave it a name that stuck. In the mid‐1990s, the California market was being besieged by managed care, which was seeking new ways to cut hospital utilization and costs. In 1994, the huge Kaiser Permanente system decided to reorganize its hospital care around a cadre of hospital‐based specialists (HBSs), essentially dichotomizing the roles of inpatient and outpatient physicians. (Interestingly, Kaiser's main motivations were to improve outpatient satisfaction by assuring constant availability of primary care physicians and to create a vehicle to promote inpatient quality improvement activities, not necessarily to improve inpatient efficiency.) Around the same time, I read reports in throwaway magazines about Park Nicollet in Minneapolis and the Scripps Clinic in La Jolla, California, doing the same thing. Then one day I heard that a talented young UCSF faculty member was leaving our VA system to take a job as the inpatient manager at a local community teaching hospital. A few weeks later, I took him out to lunchI was intrigued by this new role and wanted to better understand it. As he described it to me over sandwiches, it made all the sense in the world, and the seeds of the MRPNand later NEJMarticle was planted.

I have always had an abiding interest in the notion of valuea fundamental belief that our system is inexorably becoming one in which health care choices and competition will be based on demonstrable quality, safety, the patient's experience, and cost rather than on tradition, impression, and proximity. As I began thinking about hospital care, it seemed likely this new modeldichotomizing the roles of inpatient and outpatient doctors such that the former could be constantly available and become an expert in inpatient clinical care and hospital microsystemswould provide more value than the traditional structure, both in community settings (replacing the single primary care doctor managing both inpatients and outpatients) and the academic setting (replacing the traditional one‐month‐a‐year ward attending).

At the time I was thinking all this through, a new chairman of our department of medicine arrived from Harvard. Lee Goldman, who virtually invented the field of clinical epidemiology, came to UCSF with a powerful vision that matched mineto transform training and clinical care to improve both value and education. Lee had been a resident at UCSF 20 years earlier and returned in 1995 to an inpatient service whose structure and culture had barely changed over a generation. Lee (who, to my great chagrin, recently left UCSF to become Columbia's medical school dean, and who does not have the term good enough in his vocabulary) sat down with me and articulated his vision for a new type of academic inpatient model, led by faculty who cared for inpatients and taught trainees hospital medicine for a living. This was entirely in sync with my thoughts, and so we set out to build it.

Reaction to both the New England Journal of Medicine article and our vision for an academic hospitalist service was swift and negative. One letter to the NEJM said it all:

Our response to this and the other letters emphasized the need for evidence:

My father, a retired businessman living in Florida, brought the controversy to an even finer point a year later. I met this guy playing tennis today, he told me on the phone one day. And he's heard of you! I listened for the heartwarming sounds of fatherly pride, but none were forthcoming. He hates you, he added.

Our attempts to build an academic hospitalist program generated other concerns. Many faculty enjoyed serving as ward attendings and worried about being kicked off the wards (although many privately told me that they knew their time was up and were grateful for a way to exit with dignity.) One world‐famous faculty member bolted out of his seat during the Q&A period after my medical grand rounds at his institution in 1997. How will the house staff learn anything if we don't allow them to learn from their mistakes? he huffed. (I told him that I was flying cross‐country the next day, and I'll be really pissed off if my pilot is there to learn from his mistakes.) Our residents also worried terribly about losing their autonomy, having these bright young attendings breathing down our backs. Everyone worried about where the resources to pay for the program would come from.

At UCSF, our strategy was to reassure everyone that we would be measuring the impact of the new model in terms of cost, quality, patient satisfaction, and education. By making clear that the results of this research would guide further change (and that we were willing to end the experiment if it turned out negatively), the faculty and house staff largely suspended their disbelief for the first year. That study4 would demonstrate impressive cost savings with no adverse impact on quality and patient satisfaction and a hint of improved resident satisfaction (later proven more conclusively5), allowing us to expand the program over time and to make the argument for ongoing medical center support of the new model.

Just as Lee Goldman's arrival at UCSF in 1995 was a remarkable and crucial bit of serendipity, my partnership with Dr. Win Whitcomb and Dr. John Nelson was every bit as important for the growth of the movement nationally. John, at that time a young internist in Gainesville, Florida, had been practicing as a hospitalist (though it wasn't called that) since completing his internal medicine residency in the later 1980s. He had hooked up with Win, another young internist who had left a private practice job to begin a hospital‐based practice at Mercy Medical Center in Springfield, Massachusetts. Together the two of them had begun to network with the handful of physicians around the country who were practicing in this new model. But they needed a larger megaphone, both to let other hospitalists know about each other, and to make hospitals and systems more aware of this new model of care.

John tells the story of pulling the August 15, 1996, issue of the NEJM out of his mailbox, seeing my article, and literally running into to his house to tell his wife that his practice had finally been discovered. John's thoughtful exuberance is one of the reasons for the growth of our field, and he did something that is uniquely Johncalling the author of an article that piqued his interest to discuss its contents, something he'd been doing for years. At that point, Lee Goldman was an internationally known leader in internal medicine; as chair of a major academic department, he had several layers of administrative assistants running interference when he received cold calls. I, on the other hand, ran a sleepy medical service and had little to do other than to answer calls and to respond to this new thing called e‐mail. John didn't know that; in his experience, first authors of articles in major journals were nearly always too busy to answer calls from country docs like him. So he tried Lee Goldman first but failed to get through. Win, on the other hand, decided to call me and had no problem getting through immediately. We hit it off like we'd been buddies for decades, sharing our instinctive recognition that that we were at the cutting edge of a new specialty. In what, in retrospect, seems like an extraordinary amount of hubris, we essentially divided up the world, asking the question: what does an emerging specialty need in order to be successful? I'm reminded of one of my favorite parts of the brilliant dialogue by Mel Brooks and Carl Reiner, The 2000 Year Old Man. Brooks, playing the title part, describes his relationship with Joan of Arc (What a cutie, he gushes) to Reiner (playing the interviewer), and how Joan's mission got in the way of their ardor. She used to say to me, Ive got to save France,' says the 2000 Year Old Man. I said, Look, Ive got to wash up, you save France, I'll see you later' That was usWin and John agreed to focus on building a new professional society and on networking with community‐based hospitalists, while I emphasized the academic side of things: organizing meetings, developing training programs, publishing a textbook, and launching a research agenda.

The first national gathering of hospitalists was astonishing. In the spring of 1997, I hosted what I thought would be a small hospital medicine CME meeting at a Holiday Inn in San Francisco in a seedy part of town. I expected about 50 people to attend and was shocked to see 3 times that (plus several homeless people who wandered into the sessions). Most remarkably, at the end of day 1, following 8 hours of clinical lectures, Win, John, and I asked the attendees if anybody wanted to stay a while and discuss the possibility of forming a new society. To our amazement, virtually everybody stayedmore than 100 people! Would anybody be willing to contribute some money to get this started? asked John, expecting nothing. And people began passing $20 bills up to the front of the room. That was the moment we knew we were onto something very bigthe atmosphere was electric, the enthusiasm easily palpable.

We initially called the new society the National Association of Inpatient Physicians (NAIP), as the name hospitalist was still very controversial, and many thought it would not have legsthe term inpatient physician was believed to be more inclusive. NAIP rapidly reached a crucial turning point. Our few hundred dollars in dues collections and ad revenues lived in Win's shoebox in Massachusetts (and later in a checking account opened by Ron Angus in Texas), and Win, John, and I were keeping databases of hospitalists on our computers and the backs of napkins. It was clear we needed to either create a full‐fledged infrastructure or partner with an organization that could help us. I approached Hal Sox, now the editor of the Annals of Internal Medicine but at that time president of the American College of Physicians (and an old fellowship mentor of mine), about the possibility of NAIP establishing a formal relationship with the ACP. Hal was reluctant at first, noting many ACP members were pretty strongly against the idea of hospitalists. In one of many acts of brinksmanship, I told him we would need to look for other partners if ACP did not get over its ambivalence and embrace our new field. To his credit and to the credit particularly of Dr. Walt McDonald, ACP's executive director at the time, both recognized the potential growth of this new field and worked through the internal politics to offer us an affiliation. However, we found their initial offerto become the Section on Hospital Medicine within the ACPunattractive. Wanting to be a full‐fledged independent organization that enjoyed a relationship with the College, we proposed a relationship that would link us and allow ACP to support our infrastructure, but that allowed us to retain independent decision making, governance, and budget. John, in his charming Southern drawl, described our position to an early gathering of about 100 hospitalists at a NAIP meeting in San Diego. Their offer would have them up here, and we'd be down there, he said, his hands depicting an obvious hierarchy, with us on the bottom. But we insisted on being equal partners, he said, with his hands on the same plane. I turned to Win, sitting next to me in the audience, and whispered something like, Yeah, equalexcept for the small fact that they have 120,000 members and we have 87. Nevertheless, they agreed, and our relationship has been incredibly positive for hospitalists, and I believe for the ACP as well.

The rest, as they say, is history. The society, renamed the Society of Hospital Medicine in April 2003, has thrived under the leadership of a strong series of boards, a wonderful staff, and a charismatic and highly effective CEO, Dr. Larry Wellikson. We successfully navigated the many early challenges and took advantage of key opportunities. In this regard, I consider our 3 most important decisions and actions to be: 1) creating a body of research that demonstrated, in an evidence‐based way, that the theoretical promise of the field was real6 (it was this research that led hospitals to embrace the field more vigorously and that justified the crucial support that most hospitals provide their hospitalist programs); 2) vigorously pushing back against managed care‐based hospitalist models that had begun to force primary care physicians to hand their patients off to hospitalists against their will (NAIP's first policy pronouncement was to come out strongly against such mandatory models, which seemed counterintuitive to some but which markedly decreased our vulnerability to being tagged as a cost‐cutting vehicle of managed care); and 3) linking ourselves as strongly as possible with the growing quality and safety movements. When the IOM reports on medical errors7 and later quality8 were published, we immediately saw in the new agendas a tremendous opportunity to brand hospitalists as indispensable leaders of quality and safety in hospitalsanother key rationale for hospitalists' value proposition and another reason for hospitals and policymakers to support the young field.

Looking back at the 1996 New England Journal of Medicine article, I am struck by both the number of things I got right (even a blind squirrel) and the number that I did not anticipate or got wrong. Lee and I thought that many hospitalists would be subspecialists who would focus on hospital medicine for only part of their work. This was true early on, but the field has evolved to be more of a generalist endeavor (although recently there have emerged laborists, neurology hospitalists, and even surgical hospitalists). I probably could have anticipated the growth of the field in pediatrics, but it certainly was not on my radar screen until years later.9 I did not count on the work hours of house staff being regulated; even if I had, I'm not sure I would have fully recognized how the need to create nonteaching services would turbo‐charge the growth of the hospitalist field in teaching hospitals. The one mild disappointment: I anticipated stronger evidence by now of the field's salutary impact on safety and quality. The effort to study and hopefully demonstrate such improvements should be a major focus for the next 510 years. Finally, although I thought the field would grow rapidly, I did not anticipate that a decade later there would be 15,000 hospitalists nationally or 24 in my group at UCSF. I also did not guess that an April 2006 Medline search of hospitalist would find 561 articles or that a Google search of hospitalist would yield 689,000 entries (hell, there was no Google to search in 1996!).

As I reflect back on the last decade, I am humbled by the remarkable work I have seen from hospitalists around the country and grateful for the wonderful friendships I have enjoyed with my colleagues in our new field. I am even more convinced of the fundamental accuracy of my underlying premise: the U.S. health care system will increasingly embrace models, strategies, and providers who can demonstrably improve the value of care. I have no doubt thatcollectivelyAmerican hospitalists have saved tens of thousands of lives, prevented tens of thousands of errors, orchestrated tens of thousands of good deaths, comforted tens of thousands of families, and saved billions of dollars. It is an ongoing legacy that gives me considerable pride and joy.

Self‐check‐in kiosks started to appear in airports in the late 1990s, and within a few years, they seem to have become ubiquitous in the airline industry. Today, almost 70% of business travelers use them, and other sectors of the travel industry are beginning to experiment with the technology.1 Compared to this innovation in the airline industry, adoption of computerized physician order entry (CPOE) in U.S. hospitals, first pioneered in the early 1970s,2, 3 has taken a much more leisurely pace. Despite numerous studies documenting its benefits,47 promotion by prominent national patient safety advocacy groups such as LeapFrog,8 and numerous guides on best adoption practices.912 fewer than 10% of U.S. hospitals have fully adopted this technology.13 Moreover, as Lindenauer et al.14 pointed out, most hospitals that have successfully implemented CPOE are academic medical centers that rely on house staff to enter orders. With notable exceptions,3 adoption of CPOE in community hospitals where attending physicians write most orders remains anemic.

Although an increasing number of scholarly articles has documented the reasons for this slow rate of adoption even in hospitals that have the resources to invest in this technology, much of that research is based on expert opinion and case studies.11, 1519 In this context, Lindenauer et al.14 should be commended for using empirical evidence to delineate the predictors of adoption. Lindenauer et al. found that physicians who trained in hospitals with CPOE were more likely to be frequent users of CPOE in their new environment. Although the analysis did not account for possible confounding such as employment status of the physician, this result does confirm the conventional wisdom that physicians‐in‐training are more malleable and that residency is an important opportunity to expose physicians to safety technologies. If this finding is borne out by further research, it would bode well for the adoption of CPOE, as many physicians are trained in academic institutions, which are more likely to have CPOE,20 and almost all physicians spend part of their training in a VA hospital, which has uniformly adopted CPOE. Similarly, Lindenauer et al. found that physicians who use computers for personal purposes are more likely to be frequent users of CPOE. Given the increasingly ubiquitous use of computers in all spheres of life, time is on the side of increasing acceptance of CPOE.

However, a closer examination of the data presented by Lindenauer et al. raises several concerns. First, the substantial number of infrequent users across all demographic subgroups and clinical disciplines, even among users who were exposed to CPOE during training or those who used computers regularly for personal purposes, highlights the absence of shortcuts to the universal acceptance of CPOE. Second, whereas 63% of surveyed physicians believed that CPOE would reduce the incidence of medication errors and 71% believed that CPOE would prevent aspects of care from slipping through the cracks, only 42% of the surveyed physicians were frequent users of CPOE. This implies that even when physicians believe in the safety and quality benefits of CPOE, that belief alone may not be sufficient to convince all of them to adopt this technology wholeheartedly; other factors such as speed, ease of use, and training are likely important prerequisites. Third, although 66% of orders placed in person at the 2 study hospitals were entered through CPOE, acceptance of this technology, as measured by Lindenauer et al, was moderate at both institutions. This suggests that even when organizations have reached the 70% threshold set by Leapfrog as the proportion of orders placed in CPOE that qualifies as full implementation, they may continue to face resistance to full acceptance of the technology.

Compared to their academic counterparts, community hospitals face additional hurdles as they implement CPOE. Not only does their smaller size make it difficult to achieve economies of scale, they are also at a disadvantage because of the relationship the community hospital has with its physicians. Unlike physicians‐in‐training in academic medical centers, physicians in community hospitals function as largely autonomous agents over whom the hospital administration has little control. Although these physicians and their hospitals share the common goals of patient safety and quality, the financial incentives for the adoption of CPOE are often misaligned. For example, a recent cost benefit analysis21 showed the enormous potential for hospitals to cut costs if physicians fully adopt a CPOE system with rich decision support features. However, those savings typically accrue to the hospital, not to the physicians who use the system. Assuming the typical learning curve that accompanies the use of any new technology, physicians in community hospitals may have little incentive to invest the time to learn to use the system efficiently.

So what can be done to overcome these seemingly formidable barriers to full adoption of CPOE? Emerging research, which has so far largely focused on CPOE implementation at academic hospitals, suggests there is no silver bullet. Instead, it has taught us how the complex interplay among vendor capability, organizational behavior, clinician work flow, and implementation strategy determines the success or failure of adoption.11, 17, 18, 22 Although physician characteristics will play a role in determining whether an individual adopts this technology, local factors such as the presence of champions, governance model for the project, support for staff throughout the process, and relationship between administration and physicians are likely important determinants of success at both academic and community hospitals. In addition, organizations that embark on CPOE implementation need to understand the enormity of the task at hand and must devote not only sufficient financial but also human capital over time.11, 18 In the words of a chief medical information officer, Implementing CPOE should not be thought of as an event, but a long‐term commitment.

Beyond following proposed best practices for the implementation of CPOE, community hospitals may need to adopt additional strategies to address their unique challenges. Given the misalignment of incentives for physicians' use of CPOE, leadership in community hospitals must be particularly skilled at articulating the benefits of CPOE to physicians. These benefits include not only decreased professional liability from improved patient safety and better quality of care, but also fewer pharmacy callbacks, remote access, and rapid ordering through order sets. Hospitals may also want to elicit support from physicians early by empowering them to create order sets for their disciplines. Mechanisms for hospitals and physicians to engage in mutual cost‐sharing arrangements may provide addition opportunities for hospitals to entice physicians to adopt the technology. Finally, and of particular interest to the readership of this journal, as hospitalists become more prevalent and take care of an increasing proportion of hospitalized patients,23 they are often ideal candidates to lead the implementation of CPOE in community hospitals. Because hospitalists spend most of their time in the hospital, they are often in the best position to get fully trained on CPOE, to define their own order sets, and to redesign care processes in order to take full advantage of CPOE capabilities. In addition, as many hospitalists are directly employed or supported by the hospital, their goals for quality, safety, and efficiency are usually better aligned with those of the hospital.

The stakes involved in implementing CPOE are high. Hospitals invest enormous sums of money in these systems, and many will not have the financial or political capital to attempt a second implementation after an initial failure. In addition, as recent research has pointed out,24 inappropriate implementation strategies may lead to delays in essential care and direct patient harm. In many ways, the complex task of implementing CPOE is not unlike other endeavors in patient care, where optimal outcomes require sound knowledge and reliable processes and where disaster can strike for lack of attention to detail or common sense. If Hippocrates were alive today, he might have this to say about CPOE implementation: Life is short, the art long, opportunity fleeting, experience treacherous, judgment difficult.

Self‐check‐in kiosks started to appear in airports in the late 1990s, and within a few years, they seem to have become ubiquitous in the airline industry. Today, almost 70% of business travelers use them, and other sectors of the travel industry are beginning to experiment with the technology.1 Compared to this innovation in the airline industry, adoption of computerized physician order entry (CPOE) in U.S. hospitals, first pioneered in the early 1970s,2, 3 has taken a much more leisurely pace. Despite numerous studies documenting its benefits,47 promotion by prominent national patient safety advocacy groups such as LeapFrog,8 and numerous guides on best adoption practices.912 fewer than 10% of U.S. hospitals have fully adopted this technology.13 Moreover, as Lindenauer et al.14 pointed out, most hospitals that have successfully implemented CPOE are academic medical centers that rely on house staff to enter orders. With notable exceptions,3 adoption of CPOE in community hospitals where attending physicians write most orders remains anemic.

Although an increasing number of scholarly articles has documented the reasons for this slow rate of adoption even in hospitals that have the resources to invest in this technology, much of that research is based on expert opinion and case studies.11, 1519 In this context, Lindenauer et al.14 should be commended for using empirical evidence to delineate the predictors of adoption. Lindenauer et al. found that physicians who trained in hospitals with CPOE were more likely to be frequent users of CPOE in their new environment. Although the analysis did not account for possible confounding such as employment status of the physician, this result does confirm the conventional wisdom that physicians‐in‐training are more malleable and that residency is an important opportunity to expose physicians to safety technologies. If this finding is borne out by further research, it would bode well for the adoption of CPOE, as many physicians are trained in academic institutions, which are more likely to have CPOE,20 and almost all physicians spend part of their training in a VA hospital, which has uniformly adopted CPOE. Similarly, Lindenauer et al. found that physicians who use computers for personal purposes are more likely to be frequent users of CPOE. Given the increasingly ubiquitous use of computers in all spheres of life, time is on the side of increasing acceptance of CPOE.

However, a closer examination of the data presented by Lindenauer et al. raises several concerns. First, the substantial number of infrequent users across all demographic subgroups and clinical disciplines, even among users who were exposed to CPOE during training or those who used computers regularly for personal purposes, highlights the absence of shortcuts to the universal acceptance of CPOE. Second, whereas 63% of surveyed physicians believed that CPOE would reduce the incidence of medication errors and 71% believed that CPOE would prevent aspects of care from slipping through the cracks, only 42% of the surveyed physicians were frequent users of CPOE. This implies that even when physicians believe in the safety and quality benefits of CPOE, that belief alone may not be sufficient to convince all of them to adopt this technology wholeheartedly; other factors such as speed, ease of use, and training are likely important prerequisites. Third, although 66% of orders placed in person at the 2 study hospitals were entered through CPOE, acceptance of this technology, as measured by Lindenauer et al, was moderate at both institutions. This suggests that even when organizations have reached the 70% threshold set by Leapfrog as the proportion of orders placed in CPOE that qualifies as full implementation, they may continue to face resistance to full acceptance of the technology.

Compared to their academic counterparts, community hospitals face additional hurdles as they implement CPOE. Not only does their smaller size make it difficult to achieve economies of scale, they are also at a disadvantage because of the relationship the community hospital has with its physicians. Unlike physicians‐in‐training in academic medical centers, physicians in community hospitals function as largely autonomous agents over whom the hospital administration has little control. Although these physicians and their hospitals share the common goals of patient safety and quality, the financial incentives for the adoption of CPOE are often misaligned. For example, a recent cost benefit analysis21 showed the enormous potential for hospitals to cut costs if physicians fully adopt a CPOE system with rich decision support features. However, those savings typically accrue to the hospital, not to the physicians who use the system. Assuming the typical learning curve that accompanies the use of any new technology, physicians in community hospitals may have little incentive to invest the time to learn to use the system efficiently.

So what can be done to overcome these seemingly formidable barriers to full adoption of CPOE? Emerging research, which has so far largely focused on CPOE implementation at academic hospitals, suggests there is no silver bullet. Instead, it has taught us how the complex interplay among vendor capability, organizational behavior, clinician work flow, and implementation strategy determines the success or failure of adoption.11, 17, 18, 22 Although physician characteristics will play a role in determining whether an individual adopts this technology, local factors such as the presence of champions, governance model for the project, support for staff throughout the process, and relationship between administration and physicians are likely important determinants of success at both academic and community hospitals. In addition, organizations that embark on CPOE implementation need to understand the enormity of the task at hand and must devote not only sufficient financial but also human capital over time.11, 18 In the words of a chief medical information officer, Implementing CPOE should not be thought of as an event, but a long‐term commitment.

Beyond following proposed best practices for the implementation of CPOE, community hospitals may need to adopt additional strategies to address their unique challenges. Given the misalignment of incentives for physicians' use of CPOE, leadership in community hospitals must be particularly skilled at articulating the benefits of CPOE to physicians. These benefits include not only decreased professional liability from improved patient safety and better quality of care, but also fewer pharmacy callbacks, remote access, and rapid ordering through order sets. Hospitals may also want to elicit support from physicians early by empowering them to create order sets for their disciplines. Mechanisms for hospitals and physicians to engage in mutual cost‐sharing arrangements may provide addition opportunities for hospitals to entice physicians to adopt the technology. Finally, and of particular interest to the readership of this journal, as hospitalists become more prevalent and take care of an increasing proportion of hospitalized patients,23 they are often ideal candidates to lead the implementation of CPOE in community hospitals. Because hospitalists spend most of their time in the hospital, they are often in the best position to get fully trained on CPOE, to define their own order sets, and to redesign care processes in order to take full advantage of CPOE capabilities. In addition, as many hospitalists are directly employed or supported by the hospital, their goals for quality, safety, and efficiency are usually better aligned with those of the hospital.

The stakes involved in implementing CPOE are high. Hospitals invest enormous sums of money in these systems, and many will not have the financial or political capital to attempt a second implementation after an initial failure. In addition, as recent research has pointed out,24 inappropriate implementation strategies may lead to delays in essential care and direct patient harm. In many ways, the complex task of implementing CPOE is not unlike other endeavors in patient care, where optimal outcomes require sound knowledge and reliable processes and where disaster can strike for lack of attention to detail or common sense. If Hippocrates were alive today, he might have this to say about CPOE implementation: Life is short, the art long, opportunity fleeting, experience treacherous, judgment difficult.

Self‐check‐in kiosks started to appear in airports in the late 1990s, and within a few years, they seem to have become ubiquitous in the airline industry. Today, almost 70% of business travelers use them, and other sectors of the travel industry are beginning to experiment with the technology.1 Compared to this innovation in the airline industry, adoption of computerized physician order entry (CPOE) in U.S. hospitals, first pioneered in the early 1970s,2, 3 has taken a much more leisurely pace. Despite numerous studies documenting its benefits,47 promotion by prominent national patient safety advocacy groups such as LeapFrog,8 and numerous guides on best adoption practices.912 fewer than 10% of U.S. hospitals have fully adopted this technology.13 Moreover, as Lindenauer et al.14 pointed out, most hospitals that have successfully implemented CPOE are academic medical centers that rely on house staff to enter orders. With notable exceptions,3 adoption of CPOE in community hospitals where attending physicians write most orders remains anemic.

Although an increasing number of scholarly articles has documented the reasons for this slow rate of adoption even in hospitals that have the resources to invest in this technology, much of that research is based on expert opinion and case studies.11, 1519 In this context, Lindenauer et al.14 should be commended for using empirical evidence to delineate the predictors of adoption. Lindenauer et al. found that physicians who trained in hospitals with CPOE were more likely to be frequent users of CPOE in their new environment. Although the analysis did not account for possible confounding such as employment status of the physician, this result does confirm the conventional wisdom that physicians‐in‐training are more malleable and that residency is an important opportunity to expose physicians to safety technologies. If this finding is borne out by further research, it would bode well for the adoption of CPOE, as many physicians are trained in academic institutions, which are more likely to have CPOE,20 and almost all physicians spend part of their training in a VA hospital, which has uniformly adopted CPOE. Similarly, Lindenauer et al. found that physicians who use computers for personal purposes are more likely to be frequent users of CPOE. Given the increasingly ubiquitous use of computers in all spheres of life, time is on the side of increasing acceptance of CPOE.

However, a closer examination of the data presented by Lindenauer et al. raises several concerns. First, the substantial number of infrequent users across all demographic subgroups and clinical disciplines, even among users who were exposed to CPOE during training or those who used computers regularly for personal purposes, highlights the absence of shortcuts to the universal acceptance of CPOE. Second, whereas 63% of surveyed physicians believed that CPOE would reduce the incidence of medication errors and 71% believed that CPOE would prevent aspects of care from slipping through the cracks, only 42% of the surveyed physicians were frequent users of CPOE. This implies that even when physicians believe in the safety and quality benefits of CPOE, that belief alone may not be sufficient to convince all of them to adopt this technology wholeheartedly; other factors such as speed, ease of use, and training are likely important prerequisites. Third, although 66% of orders placed in person at the 2 study hospitals were entered through CPOE, acceptance of this technology, as measured by Lindenauer et al, was moderate at both institutions. This suggests that even when organizations have reached the 70% threshold set by Leapfrog as the proportion of orders placed in CPOE that qualifies as full implementation, they may continue to face resistance to full acceptance of the technology.

Compared to their academic counterparts, community hospitals face additional hurdles as they implement CPOE. Not only does their smaller size make it difficult to achieve economies of scale, they are also at a disadvantage because of the relationship the community hospital has with its physicians. Unlike physicians‐in‐training in academic medical centers, physicians in community hospitals function as largely autonomous agents over whom the hospital administration has little control. Although these physicians and their hospitals share the common goals of patient safety and quality, the financial incentives for the adoption of CPOE are often misaligned. For example, a recent cost benefit analysis21 showed the enormous potential for hospitals to cut costs if physicians fully adopt a CPOE system with rich decision support features. However, those savings typically accrue to the hospital, not to the physicians who use the system. Assuming the typical learning curve that accompanies the use of any new technology, physicians in community hospitals may have little incentive to invest the time to learn to use the system efficiently.

So what can be done to overcome these seemingly formidable barriers to full adoption of CPOE? Emerging research, which has so far largely focused on CPOE implementation at academic hospitals, suggests there is no silver bullet. Instead, it has taught us how the complex interplay among vendor capability, organizational behavior, clinician work flow, and implementation strategy determines the success or failure of adoption.11, 17, 18, 22 Although physician characteristics will play a role in determining whether an individual adopts this technology, local factors such as the presence of champions, governance model for the project, support for staff throughout the process, and relationship between administration and physicians are likely important determinants of success at both academic and community hospitals. In addition, organizations that embark on CPOE implementation need to understand the enormity of the task at hand and must devote not only sufficient financial but also human capital over time.11, 18 In the words of a chief medical information officer, Implementing CPOE should not be thought of as an event, but a long‐term commitment.

Beyond following proposed best practices for the implementation of CPOE, community hospitals may need to adopt additional strategies to address their unique challenges. Given the misalignment of incentives for physicians' use of CPOE, leadership in community hospitals must be particularly skilled at articulating the benefits of CPOE to physicians. These benefits include not only decreased professional liability from improved patient safety and better quality of care, but also fewer pharmacy callbacks, remote access, and rapid ordering through order sets. Hospitals may also want to elicit support from physicians early by empowering them to create order sets for their disciplines. Mechanisms for hospitals and physicians to engage in mutual cost‐sharing arrangements may provide addition opportunities for hospitals to entice physicians to adopt the technology. Finally, and of particular interest to the readership of this journal, as hospitalists become more prevalent and take care of an increasing proportion of hospitalized patients,23 they are often ideal candidates to lead the implementation of CPOE in community hospitals. Because hospitalists spend most of their time in the hospital, they are often in the best position to get fully trained on CPOE, to define their own order sets, and to redesign care processes in order to take full advantage of CPOE capabilities. In addition, as many hospitalists are directly employed or supported by the hospital, their goals for quality, safety, and efficiency are usually better aligned with those of the hospital.

The stakes involved in implementing CPOE are high. Hospitals invest enormous sums of money in these systems, and many will not have the financial or political capital to attempt a second implementation after an initial failure. In addition, as recent research has pointed out,24 inappropriate implementation strategies may lead to delays in essential care and direct patient harm. In many ways, the complex task of implementing CPOE is not unlike other endeavors in patient care, where optimal outcomes require sound knowledge and reliable processes and where disaster can strike for lack of attention to detail or common sense. If Hippocrates were alive today, he might have this to say about CPOE implementation: Life is short, the art long, opportunity fleeting, experience treacherous, judgment difficult.