James S. Goodwin, MD

Readmissions of Medicare beneficiaries within 30 days of discharge are frequent and costly.[1] Concern about readmissions has prompted the Centers for Medicare & Medicaid Services (CMS) to reduce payments to hospitals with excess readmissions.[2] Research has identified a number of patient clinical and socio‐demographic factors associated with readmissions.[3] However, interventions designed to reduce readmissions have met with limited success. In a systematic review, no single intervention was regularly effective in reducing readmissions, despite the fact that interventions have targeted both predischarge, transition of care, and postdischarge processes of care.[4]

The different trajectories of care experienced by patients after hospital discharge, and their effect on risk of readmission, have been incompletely studied. Although early outpatient follow‐up after discharge is associated with lower readmission rates,[5, 6] a factor that has been minimally studied is the role of the emergency department (ED) and the ED provider in readmissions. The ED and ED providers feature prominently in the care received by patients shortly after discharge from a hospital. About a quarter of all hospitalized Medicare patients are evaluated in an ED within 30 days of discharge,[7, 8] and a majority of readmissions within 30 days of discharge are precipitated by an ED visit.[9] Hence, we asked whether when a recently discharged patient is seen in an ED, does the rate of readmission vary by ED provider and by ED facility?

We used Texas Medicare claims data to examine patients visiting the ED within 30 days of discharge from an initial hospitalization to determine if their risk of readmission varies by the ED provider caring for them and by the ED facility they visit.

METHODS

RESULTS

We identified 174,209 patients who visited an ED within 30 days of discharge from an initial hospitalization. Table 1 describes the characteristics of these patients as well as the readmission rates associated with these characteristics. The rate of readmission of our cohort of 1,627,705 discharges with or without a following ED visit was 16.2%, whereas the rate of readmission following an ED visit in our final cohort of 174,209 patients was 52.67%. This readmission rate increased with age, from 49.31% for patients between 66 and 70 years of age to 55.33% for patients older than 85 years. There were minor variations by gender and ethnicity. Patients residing in metropolitan areas or in zip codes with low education levels had higher readmission rates, as did those whose original admission was classified as emergency or those who were not discharged home.

Table 1 also presents the odds of readmission adjusted for all other factors in the table and also adjusted for clustering within ED providers in a 2‐level model. Increasing age, white race, metropolitan residence, nonhome discharge, higher severity of illness, more hospitalizations in the prior year, fewer physician visits in the prior year, and an emergency initial admission were each associated with a higher readmission rate.

We next generated estimates of readmission rates for each ED provider from the adjusted 2‐level models. Figure 1 shows the adjusted cumulative readmission rates for the 1922 ED providers. This figure shows the mean value and 95% confidence intervals of the readmission rates for each provider. Dark vertical lines indicate providers whose readmission rate differed significantly from the mean adjusted readmission rate of 52.1% for all providers. Of the ED providers, 14.2% had significantly higher readmission rates. The mean readmission rate for these 272 providers was 67.2%. Of the ED providers, 14.7% had significantly lower readmission rates. The mean readmission rate for these 283 providers was 36.8%.

Figure 1

To determine the contribution of the ED facility to the variation in readmission rates, we restricted our analysis to 48,883 patients (28.06% of our cohort) seen by 525 ED providers who were associated with only 1 facility (total of 143 facilities). Table 2 describes the unadjusted readmission rates stratified by specific characteristics of those facilities. The unadjusted readmission rate increased with the size of the associated hospital, from 47.61% for hospitals with less than 100 beds to 57.06% for hospitals with more than 400 beds. The readmission rate for nonprofit facilities was 53.81% and for for‐profit facilities was 57.39%. Facilities with no medical school affiliation had a readmission rate of 54.51%, whereas those with a major affiliation had a readmission rate of 58.72%.

With this smaller cohort, we performed 2 types of 2‐level models, where patients clustered within ED facilities and ER providers, respectively, and a 3‐level model accounting for clustering of patients within providers and of providers within facilities. From the facility‐patient 2‐level model, the variance of the ED facility was 0.2718 (95% confidence interval [CI]: 0.2083‐0.3696). From the provider‐patient 2‐level model, the variance of ED provider was 0.2532 (95% CI: 0.2166‐0.3002). However, when the 3‐level model was performed, the variance of ED provider decreased to 0.0893 (95% CI: 0.0723‐0.1132) and the variance of ED facility dropped to 0.2316 (95% CI: 0.1704‐0.3331) . This indicates 65% of the variation among ED providers was explained by the ED facility, and in contrast, 15% of the variation among ED facilities was explained by ED providers.

Table 2 also shows the adjusted odds of readmission generated from the 3‐level model. Patients receiving care in ED facilities in hospitals with more beds and in for‐profit hospitals were at higher risk for readmission. It is possible that patients seen at the ED associated with the discharging hospital had a lower risk of readmission. This finding was close to being statistically significant (P=0.051).

We repeated all the above analyses using an outcome of readmission anytime between the ED visit and 30 days after discharge from the initial hospitalization (rather than readmission on the day of or after the ED visit). All analyses produced results similar to the results presented above. For example, Figure 2 shows the adjusted cumulative readmission rates for the 1922 ED providers using this outcome. Of the ED providers, 12.8% had higher and 12.5% had lower readmission rates as compared to the mean readmission rate for all ED providers. The Spearman correlation coefficient between the rank of ED providers in immediate readmission rate (Figure 1) and readmission rate within 30 days of hospital discharge (Figure 2) was 0.94 (P<0.001).

Figure 2

DISCUSSION

This study found substantial variation in readmission rates by ED provider, despite controlling for patient clinical and sociodemographic factors. In 3‐level models, the ED facility explained a substantial part of the variation by ED provider, with patients seen at larger facilities and for‐profit facilities having higher readmission rates.

Variation among ED facilities and ED providers in readmission rates has not previously been studied. There is literature on the variation in ED facility and ED provider admission rates. As readmissions are a subset of all admissions, this literature provides context to our findings. Abualenain et al. examined admission rates for 89 ED physicians for adult patients presenting with an acute medical or surgical complaint at 3 EDs in a health system.[12] After adjusting for patient and clinical characteristics, admission rates varied from 21% to 49% among physicians and from 27% to 41% among 3 facilities. Two other studies from single hospitals have found similar variation among providers.[13, 14] The reasons for the variation among ED providers presumably relate to subjective aspects of clinical assessment and the reluctance of providers to rely solely on objective scales, even when they are available.[14, 15] Variation in admission rates among different facilities may relate to clustering of providers with similar practice styles within facilities, lack of clinical guidelines for certain conditions, as well as differences among facilities in the socioeconomic status and access to primary care of their clientele.[12, 16, 17] For example, Pines et al. have shown that ED facility admission rates are higher in communities with fewer primary care physicians per capita and are influenced by the prevailing county level admission rates.[16] Capp et al. showed persistent variation in admission rates across hospitals, despite adjusting for clinical criteria such as vital signs, chief complaints, and severity of illness.[18]

Structural differences in ED facilities may also influence the decision to admit. We found that patients visiting ED facilities in hospitals with more beds had a higher readmission rate. ED facility systems of care such as observation units or protocols are associated with lower admission rates.[19, 20] Finally, certain hospitals may actively influence the admission practice patterns of their ED providers. We noted that patients seen at for‐profit ED facilities had a greater risk of readmission. A similar finding has been described by Pines et al., who noted higher admission rates at for‐profit facilities.[16] In an extreme example, a recent Justice Department lawsuit alleged that a for‐profit hospital chain used software systems and financial incentives to ED providers to increase admissions.[21]

It is possible that the providers with low readmission rates may have inappropriately released patients who truly should have been admitted. A signal that this occurred would be if these patients were readmitted in the days after the ED visits. We examined this possibility by additionally examining readmissions occurring anytime between the ED visit until 30 days after discharge from the initial hospitalization. The results were similar to when we only included readmissions that occurred immediately following the ED visit, with a very high correlation (r=0.94) between the ranking of the ED providers by readmission rates in both circumstances. This suggests that the decisions of the ED providers with low readmission rates to admit or release from the ED were likely appropriate.

Our research has limitations. We studied patients with fee‐for‐service Medicare in a single large state in the United States over a 4‐year period. Our findings may not be generalizable to younger patient populations, other regions with different sociodemographic patterns and healthcare systems, or other time periods. We could not control for many factors that may impact the risk of readmission but are not measured in Medicare databases (eg, clinical data such as vital signs, measures of quality of transition from discharging hospital, ED provider workload). To attribute care to a single ED provider, we excluded patients who were taken care of by multiple ED providers. These patients may have different needs from our study population (eg, more complex issues and longer stays in the ED) and may bias our results.

This study provides a new direction for research and quality improvement targeting readmissions. Research should extend beyond the discharge transition and examine the entire trajectory of posthospitalization care to better understand readmissions. Based directly on this study, research could investigate the practice patterns of ED providers and systems of care at ED facilities that affect readmissions rates. Such investigation could inform quality improvement efforts to standardize care for patients in the ED.

CMS policies hold hospitals accountable for readmissions of the patients they discharge, but do not address the admission process in the ED that leads to readmissions of recently discharged patients. Given the present study, and the fact that the proportion of all hospital admissions that occur through the ED has grown to 44%,[22] consideration of the role of the ED in public policy efforts to discourage unnecessary inpatient care may be appropriate.

In summary, this study shows that a recently discharged patient's chances of being readmitted depends partly on the ED provider who evaluates them and on the ED facility at which they seek care. ED provider practice patterns and ED facility systems of care may be a target for interventions aimed at decreasing readmission rates.

Readmissions of Medicare beneficiaries within 30 days of discharge are frequent and costly.[1] Concern about readmissions has prompted the Centers for Medicare & Medicaid Services (CMS) to reduce payments to hospitals with excess readmissions.[2] Research has identified a number of patient clinical and socio‐demographic factors associated with readmissions.[3] However, interventions designed to reduce readmissions have met with limited success. In a systematic review, no single intervention was regularly effective in reducing readmissions, despite the fact that interventions have targeted both predischarge, transition of care, and postdischarge processes of care.[4]

The different trajectories of care experienced by patients after hospital discharge, and their effect on risk of readmission, have been incompletely studied. Although early outpatient follow‐up after discharge is associated with lower readmission rates,[5, 6] a factor that has been minimally studied is the role of the emergency department (ED) and the ED provider in readmissions. The ED and ED providers feature prominently in the care received by patients shortly after discharge from a hospital. About a quarter of all hospitalized Medicare patients are evaluated in an ED within 30 days of discharge,[7, 8] and a majority of readmissions within 30 days of discharge are precipitated by an ED visit.[9] Hence, we asked whether when a recently discharged patient is seen in an ED, does the rate of readmission vary by ED provider and by ED facility?

We used Texas Medicare claims data to examine patients visiting the ED within 30 days of discharge from an initial hospitalization to determine if their risk of readmission varies by the ED provider caring for them and by the ED facility they visit.

METHODS

RESULTS

We identified 174,209 patients who visited an ED within 30 days of discharge from an initial hospitalization. Table 1 describes the characteristics of these patients as well as the readmission rates associated with these characteristics. The rate of readmission of our cohort of 1,627,705 discharges with or without a following ED visit was 16.2%, whereas the rate of readmission following an ED visit in our final cohort of 174,209 patients was 52.67%. This readmission rate increased with age, from 49.31% for patients between 66 and 70 years of age to 55.33% for patients older than 85 years. There were minor variations by gender and ethnicity. Patients residing in metropolitan areas or in zip codes with low education levels had higher readmission rates, as did those whose original admission was classified as emergency or those who were not discharged home.

Table 1 also presents the odds of readmission adjusted for all other factors in the table and also adjusted for clustering within ED providers in a 2‐level model. Increasing age, white race, metropolitan residence, nonhome discharge, higher severity of illness, more hospitalizations in the prior year, fewer physician visits in the prior year, and an emergency initial admission were each associated with a higher readmission rate.

We next generated estimates of readmission rates for each ED provider from the adjusted 2‐level models. Figure 1 shows the adjusted cumulative readmission rates for the 1922 ED providers. This figure shows the mean value and 95% confidence intervals of the readmission rates for each provider. Dark vertical lines indicate providers whose readmission rate differed significantly from the mean adjusted readmission rate of 52.1% for all providers. Of the ED providers, 14.2% had significantly higher readmission rates. The mean readmission rate for these 272 providers was 67.2%. Of the ED providers, 14.7% had significantly lower readmission rates. The mean readmission rate for these 283 providers was 36.8%.

Figure 1

To determine the contribution of the ED facility to the variation in readmission rates, we restricted our analysis to 48,883 patients (28.06% of our cohort) seen by 525 ED providers who were associated with only 1 facility (total of 143 facilities). Table 2 describes the unadjusted readmission rates stratified by specific characteristics of those facilities. The unadjusted readmission rate increased with the size of the associated hospital, from 47.61% for hospitals with less than 100 beds to 57.06% for hospitals with more than 400 beds. The readmission rate for nonprofit facilities was 53.81% and for for‐profit facilities was 57.39%. Facilities with no medical school affiliation had a readmission rate of 54.51%, whereas those with a major affiliation had a readmission rate of 58.72%.

With this smaller cohort, we performed 2 types of 2‐level models, where patients clustered within ED facilities and ER providers, respectively, and a 3‐level model accounting for clustering of patients within providers and of providers within facilities. From the facility‐patient 2‐level model, the variance of the ED facility was 0.2718 (95% confidence interval [CI]: 0.2083‐0.3696). From the provider‐patient 2‐level model, the variance of ED provider was 0.2532 (95% CI: 0.2166‐0.3002). However, when the 3‐level model was performed, the variance of ED provider decreased to 0.0893 (95% CI: 0.0723‐0.1132) and the variance of ED facility dropped to 0.2316 (95% CI: 0.1704‐0.3331) . This indicates 65% of the variation among ED providers was explained by the ED facility, and in contrast, 15% of the variation among ED facilities was explained by ED providers.

Table 2 also shows the adjusted odds of readmission generated from the 3‐level model. Patients receiving care in ED facilities in hospitals with more beds and in for‐profit hospitals were at higher risk for readmission. It is possible that patients seen at the ED associated with the discharging hospital had a lower risk of readmission. This finding was close to being statistically significant (P=0.051).

We repeated all the above analyses using an outcome of readmission anytime between the ED visit and 30 days after discharge from the initial hospitalization (rather than readmission on the day of or after the ED visit). All analyses produced results similar to the results presented above. For example, Figure 2 shows the adjusted cumulative readmission rates for the 1922 ED providers using this outcome. Of the ED providers, 12.8% had higher and 12.5% had lower readmission rates as compared to the mean readmission rate for all ED providers. The Spearman correlation coefficient between the rank of ED providers in immediate readmission rate (Figure 1) and readmission rate within 30 days of hospital discharge (Figure 2) was 0.94 (P<0.001).

Figure 2

DISCUSSION

This study found substantial variation in readmission rates by ED provider, despite controlling for patient clinical and sociodemographic factors. In 3‐level models, the ED facility explained a substantial part of the variation by ED provider, with patients seen at larger facilities and for‐profit facilities having higher readmission rates.

Variation among ED facilities and ED providers in readmission rates has not previously been studied. There is literature on the variation in ED facility and ED provider admission rates. As readmissions are a subset of all admissions, this literature provides context to our findings. Abualenain et al. examined admission rates for 89 ED physicians for adult patients presenting with an acute medical or surgical complaint at 3 EDs in a health system.[12] After adjusting for patient and clinical characteristics, admission rates varied from 21% to 49% among physicians and from 27% to 41% among 3 facilities. Two other studies from single hospitals have found similar variation among providers.[13, 14] The reasons for the variation among ED providers presumably relate to subjective aspects of clinical assessment and the reluctance of providers to rely solely on objective scales, even when they are available.[14, 15] Variation in admission rates among different facilities may relate to clustering of providers with similar practice styles within facilities, lack of clinical guidelines for certain conditions, as well as differences among facilities in the socioeconomic status and access to primary care of their clientele.[12, 16, 17] For example, Pines et al. have shown that ED facility admission rates are higher in communities with fewer primary care physicians per capita and are influenced by the prevailing county level admission rates.[16] Capp et al. showed persistent variation in admission rates across hospitals, despite adjusting for clinical criteria such as vital signs, chief complaints, and severity of illness.[18]

Structural differences in ED facilities may also influence the decision to admit. We found that patients visiting ED facilities in hospitals with more beds had a higher readmission rate. ED facility systems of care such as observation units or protocols are associated with lower admission rates.[19, 20] Finally, certain hospitals may actively influence the admission practice patterns of their ED providers. We noted that patients seen at for‐profit ED facilities had a greater risk of readmission. A similar finding has been described by Pines et al., who noted higher admission rates at for‐profit facilities.[16] In an extreme example, a recent Justice Department lawsuit alleged that a for‐profit hospital chain used software systems and financial incentives to ED providers to increase admissions.[21]

It is possible that the providers with low readmission rates may have inappropriately released patients who truly should have been admitted. A signal that this occurred would be if these patients were readmitted in the days after the ED visits. We examined this possibility by additionally examining readmissions occurring anytime between the ED visit until 30 days after discharge from the initial hospitalization. The results were similar to when we only included readmissions that occurred immediately following the ED visit, with a very high correlation (r=0.94) between the ranking of the ED providers by readmission rates in both circumstances. This suggests that the decisions of the ED providers with low readmission rates to admit or release from the ED were likely appropriate.

Our research has limitations. We studied patients with fee‐for‐service Medicare in a single large state in the United States over a 4‐year period. Our findings may not be generalizable to younger patient populations, other regions with different sociodemographic patterns and healthcare systems, or other time periods. We could not control for many factors that may impact the risk of readmission but are not measured in Medicare databases (eg, clinical data such as vital signs, measures of quality of transition from discharging hospital, ED provider workload). To attribute care to a single ED provider, we excluded patients who were taken care of by multiple ED providers. These patients may have different needs from our study population (eg, more complex issues and longer stays in the ED) and may bias our results.

This study provides a new direction for research and quality improvement targeting readmissions. Research should extend beyond the discharge transition and examine the entire trajectory of posthospitalization care to better understand readmissions. Based directly on this study, research could investigate the practice patterns of ED providers and systems of care at ED facilities that affect readmissions rates. Such investigation could inform quality improvement efforts to standardize care for patients in the ED.

CMS policies hold hospitals accountable for readmissions of the patients they discharge, but do not address the admission process in the ED that leads to readmissions of recently discharged patients. Given the present study, and the fact that the proportion of all hospital admissions that occur through the ED has grown to 44%,[22] consideration of the role of the ED in public policy efforts to discourage unnecessary inpatient care may be appropriate.

In summary, this study shows that a recently discharged patient's chances of being readmitted depends partly on the ED provider who evaluates them and on the ED facility at which they seek care. ED provider practice patterns and ED facility systems of care may be a target for interventions aimed at decreasing readmission rates.

Readmissions of Medicare beneficiaries within 30 days of discharge are frequent and costly.[1] Concern about readmissions has prompted the Centers for Medicare & Medicaid Services (CMS) to reduce payments to hospitals with excess readmissions.[2] Research has identified a number of patient clinical and socio‐demographic factors associated with readmissions.[3] However, interventions designed to reduce readmissions have met with limited success. In a systematic review, no single intervention was regularly effective in reducing readmissions, despite the fact that interventions have targeted both predischarge, transition of care, and postdischarge processes of care.[4]

The different trajectories of care experienced by patients after hospital discharge, and their effect on risk of readmission, have been incompletely studied. Although early outpatient follow‐up after discharge is associated with lower readmission rates,[5, 6] a factor that has been minimally studied is the role of the emergency department (ED) and the ED provider in readmissions. The ED and ED providers feature prominently in the care received by patients shortly after discharge from a hospital. About a quarter of all hospitalized Medicare patients are evaluated in an ED within 30 days of discharge,[7, 8] and a majority of readmissions within 30 days of discharge are precipitated by an ED visit.[9] Hence, we asked whether when a recently discharged patient is seen in an ED, does the rate of readmission vary by ED provider and by ED facility?

We used Texas Medicare claims data to examine patients visiting the ED within 30 days of discharge from an initial hospitalization to determine if their risk of readmission varies by the ED provider caring for them and by the ED facility they visit.

METHODS

RESULTS

We identified 174,209 patients who visited an ED within 30 days of discharge from an initial hospitalization. Table 1 describes the characteristics of these patients as well as the readmission rates associated with these characteristics. The rate of readmission of our cohort of 1,627,705 discharges with or without a following ED visit was 16.2%, whereas the rate of readmission following an ED visit in our final cohort of 174,209 patients was 52.67%. This readmission rate increased with age, from 49.31% for patients between 66 and 70 years of age to 55.33% for patients older than 85 years. There were minor variations by gender and ethnicity. Patients residing in metropolitan areas or in zip codes with low education levels had higher readmission rates, as did those whose original admission was classified as emergency or those who were not discharged home.

Table 1 also presents the odds of readmission adjusted for all other factors in the table and also adjusted for clustering within ED providers in a 2‐level model. Increasing age, white race, metropolitan residence, nonhome discharge, higher severity of illness, more hospitalizations in the prior year, fewer physician visits in the prior year, and an emergency initial admission were each associated with a higher readmission rate.

We next generated estimates of readmission rates for each ED provider from the adjusted 2‐level models. Figure 1 shows the adjusted cumulative readmission rates for the 1922 ED providers. This figure shows the mean value and 95% confidence intervals of the readmission rates for each provider. Dark vertical lines indicate providers whose readmission rate differed significantly from the mean adjusted readmission rate of 52.1% for all providers. Of the ED providers, 14.2% had significantly higher readmission rates. The mean readmission rate for these 272 providers was 67.2%. Of the ED providers, 14.7% had significantly lower readmission rates. The mean readmission rate for these 283 providers was 36.8%.

Figure 1

To determine the contribution of the ED facility to the variation in readmission rates, we restricted our analysis to 48,883 patients (28.06% of our cohort) seen by 525 ED providers who were associated with only 1 facility (total of 143 facilities). Table 2 describes the unadjusted readmission rates stratified by specific characteristics of those facilities. The unadjusted readmission rate increased with the size of the associated hospital, from 47.61% for hospitals with less than 100 beds to 57.06% for hospitals with more than 400 beds. The readmission rate for nonprofit facilities was 53.81% and for for‐profit facilities was 57.39%. Facilities with no medical school affiliation had a readmission rate of 54.51%, whereas those with a major affiliation had a readmission rate of 58.72%.

With this smaller cohort, we performed 2 types of 2‐level models, where patients clustered within ED facilities and ER providers, respectively, and a 3‐level model accounting for clustering of patients within providers and of providers within facilities. From the facility‐patient 2‐level model, the variance of the ED facility was 0.2718 (95% confidence interval [CI]: 0.2083‐0.3696). From the provider‐patient 2‐level model, the variance of ED provider was 0.2532 (95% CI: 0.2166‐0.3002). However, when the 3‐level model was performed, the variance of ED provider decreased to 0.0893 (95% CI: 0.0723‐0.1132) and the variance of ED facility dropped to 0.2316 (95% CI: 0.1704‐0.3331) . This indicates 65% of the variation among ED providers was explained by the ED facility, and in contrast, 15% of the variation among ED facilities was explained by ED providers.

Table 2 also shows the adjusted odds of readmission generated from the 3‐level model. Patients receiving care in ED facilities in hospitals with more beds and in for‐profit hospitals were at higher risk for readmission. It is possible that patients seen at the ED associated with the discharging hospital had a lower risk of readmission. This finding was close to being statistically significant (P=0.051).

We repeated all the above analyses using an outcome of readmission anytime between the ED visit and 30 days after discharge from the initial hospitalization (rather than readmission on the day of or after the ED visit). All analyses produced results similar to the results presented above. For example, Figure 2 shows the adjusted cumulative readmission rates for the 1922 ED providers using this outcome. Of the ED providers, 12.8% had higher and 12.5% had lower readmission rates as compared to the mean readmission rate for all ED providers. The Spearman correlation coefficient between the rank of ED providers in immediate readmission rate (Figure 1) and readmission rate within 30 days of hospital discharge (Figure 2) was 0.94 (P<0.001).

Figure 2

DISCUSSION

This study found substantial variation in readmission rates by ED provider, despite controlling for patient clinical and sociodemographic factors. In 3‐level models, the ED facility explained a substantial part of the variation by ED provider, with patients seen at larger facilities and for‐profit facilities having higher readmission rates.

Variation among ED facilities and ED providers in readmission rates has not previously been studied. There is literature on the variation in ED facility and ED provider admission rates. As readmissions are a subset of all admissions, this literature provides context to our findings. Abualenain et al. examined admission rates for 89 ED physicians for adult patients presenting with an acute medical or surgical complaint at 3 EDs in a health system.[12] After adjusting for patient and clinical characteristics, admission rates varied from 21% to 49% among physicians and from 27% to 41% among 3 facilities. Two other studies from single hospitals have found similar variation among providers.[13, 14] The reasons for the variation among ED providers presumably relate to subjective aspects of clinical assessment and the reluctance of providers to rely solely on objective scales, even when they are available.[14, 15] Variation in admission rates among different facilities may relate to clustering of providers with similar practice styles within facilities, lack of clinical guidelines for certain conditions, as well as differences among facilities in the socioeconomic status and access to primary care of their clientele.[12, 16, 17] For example, Pines et al. have shown that ED facility admission rates are higher in communities with fewer primary care physicians per capita and are influenced by the prevailing county level admission rates.[16] Capp et al. showed persistent variation in admission rates across hospitals, despite adjusting for clinical criteria such as vital signs, chief complaints, and severity of illness.[18]

Structural differences in ED facilities may also influence the decision to admit. We found that patients visiting ED facilities in hospitals with more beds had a higher readmission rate. ED facility systems of care such as observation units or protocols are associated with lower admission rates.[19, 20] Finally, certain hospitals may actively influence the admission practice patterns of their ED providers. We noted that patients seen at for‐profit ED facilities had a greater risk of readmission. A similar finding has been described by Pines et al., who noted higher admission rates at for‐profit facilities.[16] In an extreme example, a recent Justice Department lawsuit alleged that a for‐profit hospital chain used software systems and financial incentives to ED providers to increase admissions.[21]

It is possible that the providers with low readmission rates may have inappropriately released patients who truly should have been admitted. A signal that this occurred would be if these patients were readmitted in the days after the ED visits. We examined this possibility by additionally examining readmissions occurring anytime between the ED visit until 30 days after discharge from the initial hospitalization. The results were similar to when we only included readmissions that occurred immediately following the ED visit, with a very high correlation (r=0.94) between the ranking of the ED providers by readmission rates in both circumstances. This suggests that the decisions of the ED providers with low readmission rates to admit or release from the ED were likely appropriate.

Our research has limitations. We studied patients with fee‐for‐service Medicare in a single large state in the United States over a 4‐year period. Our findings may not be generalizable to younger patient populations, other regions with different sociodemographic patterns and healthcare systems, or other time periods. We could not control for many factors that may impact the risk of readmission but are not measured in Medicare databases (eg, clinical data such as vital signs, measures of quality of transition from discharging hospital, ED provider workload). To attribute care to a single ED provider, we excluded patients who were taken care of by multiple ED providers. These patients may have different needs from our study population (eg, more complex issues and longer stays in the ED) and may bias our results.

This study provides a new direction for research and quality improvement targeting readmissions. Research should extend beyond the discharge transition and examine the entire trajectory of posthospitalization care to better understand readmissions. Based directly on this study, research could investigate the practice patterns of ED providers and systems of care at ED facilities that affect readmissions rates. Such investigation could inform quality improvement efforts to standardize care for patients in the ED.

CMS policies hold hospitals accountable for readmissions of the patients they discharge, but do not address the admission process in the ED that leads to readmissions of recently discharged patients. Given the present study, and the fact that the proportion of all hospital admissions that occur through the ED has grown to 44%,[22] consideration of the role of the ED in public policy efforts to discourage unnecessary inpatient care may be appropriate.

In summary, this study shows that a recently discharged patient's chances of being readmitted depends partly on the ED provider who evaluates them and on the ED facility at which they seek care. ED provider practice patterns and ED facility systems of care may be a target for interventions aimed at decreasing readmission rates.

Although primary care physicians (PCPs) have traditionally treated patients in both ambulatory and hospital settings, many relinquished inpatient duties to hospitalists in recent decades.[1] Little is known about the PCPs who relinquished inpatient care duties or how the transition to the hospitalist model occurred. For example, what are the characteristics of PCPs who change? Do PCPs adopt the hospitalist model enthusiastically or cautiously? Characterizing PCPs who adopted the hospitalist model can help hospitalists understand their specialty's history and also inform health services research.

Much of the interest in the hospitalist model has been generated by studies reporting improved outcomes and lower hospital lengths of stay associated with hospitalist care.[2, 3, 4, 5] Conversely, detractors of the model point to reports of higher postacute care utilization among hospitalist patients.[6] Although these studies usually adjusted for differences among patients and hospitals, they did not account for PCP characteristics. As patients' access to PCPs and their PCP's capabilities are both plausible factors that could influence hospital length of stay (eg, decisions to complete more or less of a workup in the hospital), quality of care transitions, and postdischarge utilization, it is important to determine if PCPs who use hospitalists differ systematically from those who do not to correctly interpret health system utilization patterns that currently are attributed only to hospitalists.[7, 8]

We conducted this study to determine if observable PCP factors are associated with patients' use of hospitalists and to describe the trajectory by which PCPs referred their patients to hospitalists over time.

METHODS

RESULTS

During the 2001 through 2009 study period, between 2252 and 2848 PCPs were associated with at least 20 hospitalized beneficiaries in any single year. Among these, 1172 PCPs were associated with at least 20 hospitalized beneficiaries in every year of the study period. These 1172 PCPs were associated with 608,686 hospitalizations over the 9 years.

Table 1 presents the characteristics of the PCPs who contributed to the cross‐sectional analyses in 2001 (N=2252) and 2009 (N=2387), as well as the 1172 PCPs for whom we had data for all 9 years for the longitudinal analyses. Most PCPs were male, trained in the United States, and were board certified. The average number of Medicare patients seen by these PCPs and number of outpatient Medicare visits went up about 7% between 2001 and 2009.

Figure 1 graphs the percentage of PCPs as a function of what percent of their hospitalized patients received care from hospitalists, and how that changed from 2001 to 2009. For 70.9% of PCPs, fewer than 5% of their hospitalized patients received hospitalist care in 2001. By 2009, the percent of PCPs in this category had decreased to 15.2%. In contrast, in 2001, more than half of the patients for 2.1% of PCPs received hospitalist care, and the percent of PCPs in this category increased to 26.3% by 2009.

Figure 1

The pattern in Figure 1 shows that PCPs' use of hospitalists changed continuously and gradually over time. However, this pattern describes the PCPs as a group. When examined at the individual PCP level, different patterns emerge. Figure 2, which presents selected individual PCP's use of hospitalists over time, shows several distinct subpatterns of PCP practice behaviors. First, there are PCPs whose use of hospitalists was high in 2001 and stayed high or increased over time (eg, PCP A). There also were PCPs whose use of hospitalists stayed low over the entire study period (eg, PCP B). Finally, there were PCPs whose use of hospitalists was low in 2001 but high in 2009 (eg, PCP C). For this last group, the pattern of change in hospitalist utilization over time was discontinuous; that is, most of the increase occurred over a 1‐ or 2‐year period, instead of increasing gradually over time.

Figure 2

Among the 1172 PCPs associated with 20 hospitalized beneficiaries each year in all 9 years of the study period, group‐based trajectory modeling classified their practice patterns into 4 distinct trajectories (Figure 3). Among PCPs in group 1, more than one‐third of their hospitalized patients were cared for by hospitalists in 2001, and this increased to 60% by 2009. PCPs in groups 2 and 3 rarely used hospitalist care in 2001 but increased their use over time. The increase started early in the period for PCPs in group 2 and later for those in group 3. PCPs in group 4 were associated with little hospitalist use throughout the study period.

Figure 3

We constructed a model to describe the odds of a patient receiving care from hospitalists during the study period using patients associated with these 1172 PCPs. After adjusting for patient characteristics, the residual intraclass correlation coefficient for PCP level was 0.334, which indicates that 33.4% of the variance in whether a hospitalized patient received care from a hospitalist is explained by which PCP the patient saw. When adjusting for both patient and PCP characteristics, the overall odds of a patient receiving hospitalist care increased by 30% (95% confidence interval [CI]: 1.29‐1.30) per year from 2001 through 2009.

There were also significant interactions between year of hospitalization and several PCP characteristics. These interactions are illustrated in Table 2, which stratifies each of those PCP characteristics by 3 time periods: 2001 to 2003, 2004 to 2006, and 2007 to 2009. In all time periods, patients were more likely to receive hospitalist care if their PCP was US trained (US vs international medical graduate: odds ratio [OR]: 1.42, 95% CI: 1.19‐1.69 in 20012003; OR: 1.46, 95% CI: 1.23‐1.73 in 20072009), or specialized in family medicine (family medicine vs internal medicine: OR: 1.46, 95% CI: 1.25‐1.72 in 20012003; OR: 1.46, 95% CI: 1.25‐1.70 in 20072009). Over time, the relative odds of a patient receiving care from hospitalists decreased if their PCP was female (female vs male: OR: 1.91, 95% CI: 1.46‐2.50 in 20012003 vs OR: 1.50, 95% CI: 1.15‐1.95 in 20072009) or practiced in an urban area (largest vs smallest MSA: OR: 3.34, 95% CI: 2.72‐4.09 in 20012003; OR: 2.22, 95% CI: 1.82‐2.71 in 20072009). Although the longest‐practicing PCPs were most likely to use hospitalists in the early 2000s, this effect disappeared by 2007 to 2009 (most vs least years in practice: OR: 1.35, 95% CI: 1.06‐1.72 in 20012003 vs OR: 0.92, 95% CI: 0.73‐1.17 in 20072009).

In terms of PCP workload, patients of PCPs with high outpatient activity were more likely to receive hospitalists care throughout the study period, although the association had decreased by 2007 to 2009 (highest vs lowest outpatient volume: OR: 1.46, 95% CI: 1.30‐1.63 in 20012003 vs OR: 1.32, 95% CI: 1.20‐1.44 in 20072009). In contrast, PCPs with the lowest inpatient volumes became more likely to use hospitalists by the end of the study period (highest vs lowest inpatient volume: OR: 1.05, 95% CI: 0.95‐1.18 in 20012003 vs OR: 0.55, 95% CI: 0.51‐0.60 in 20072009).

The characteristics of PCPs' practice panels also were associated with patients' likelihood of receiving care from hospitalists. PCPs whose practice panels consisted of patients who were predominantly male, white, or with more outpatient comorbidities were consistently more likely to use hospitalists throughout the study period. PCPs with older patient panels were less likely to use hospitalists in 2001 to 2003, but by 2007 to 2009, they were slightly more likely to do so (oldest vs youngest average outpatient panel age: OR: 0.72, 95% CI: 0.64‐0.81 in 20012003 vs OR: 1.15, 95% CI: 1.05‐1.26 in 20072009).

CONCLUSIONS

Prior studies of the hospitalist model have shown that the likelihood of a patient receiving inpatient care from hospitalists is associated with patient characteristics, hospital characteristics, geographic region, and type of admission.[1, 16, 17] We found that PCP characteristics also predict whether patients receive care from hospitalists and that their use of hospitalists developed dynamically between 2001 to 2009. Although many factors (such as whether patients were admitted to a hospital where their PCP had admitting privileges) can influence the decision to use hospitalists, we found that over one‐third of the variance in whether a hospitalized patient received care from a hospitalist is explained by which PCP the patient saw. In showing that systemic differences exist among PCPs who use hospitalists and those who do not, our study suggests that future research on the hospitalist model should, if possible, adjust for PCP characteristics in addition to hospital and patient factors.

Although this study identifies the existence and magnitude of differences in whether or not PCPs use hospitalists, it cannot explain why the differences exist. We only can offer hypotheses. For example, our finding that PCPs with the most years of practice experience were more likely to use hospitalists in the early 2000s but not in more recent years suggests that in hospital medicine's early years, long‐practicing generalist physicians were choosing between practicing traditionalist medicine and adopting the hospitalists model, but by 2009, experienced generalist physicians had already specialized to either inpatient or outpatient settings earlier in their careers. On the other hand, the decreasing odds of urban PCPs using hospitalists may reflect a relative growth in hospitalist use in less populated areas rather than a change in urban PCPs' practice patterns.

PCPs trained in family medicine have reported less inpatient training and less comfort with providing hospital care,[18, 19] thus it is unsurprising that family physicians were more likely to refer patients to hospitalists. Although a recent study reported that family physicians' inpatient volumes remained constant, whereas those of outpatient internists declined between 2003 and 2012, the analysis used University Health Consortium data and thus reflects practice patterns in academic medical centers.[20] Our data suggest that outside of academia, family physicians have embraced the hospitalists as clinical partners.

Meltzer and Chung had previously proposed an economic model to describe the growing use of hospitalists in the United States. They posited that decisions to adopt the hospitalist model are governed by trade‐offs between coordination costs (eg, time and effort spent coordinating multiple providers across different settings) and switching costs (eg, time spent traveling between the office and the hospital or the effort of adjusting to different work settings).[16] The authors hypothesized that empirical testing of this model would show PCPs are more likely to use hospitalists if they have less available professional time (ie, work fewer hours per week), are female (due to competing demands from domestic responsibilities), have relatively few hospitalized patients, or live in areas with high traffic congestion. Our findings provide empirical evidence to support their division‐of‐labor model in showing that patients were more likely to receive hospitalist care if their PCP was female, practiced in an urban location, had higher outpatient practice volumes, or had lower inpatient volumes.

At first glance, some of our findings appear to contradict our earlier study, which showed that younger, black, male patients are more likely to receive inpatient care from hospitalists.[1] However, that study included patients regardless of whether they had a PCP. This study shows that when patients have a PCP, their PCPs are more likely to refer them to hospitalists if they are older, white, male, and have more comorbid conditions. A potential explanation for this finding is that PCPs may preferentially use hospitalists when caring for older and sicker hospitalized patients. For example, commentators often cite hospitalists' constant availability in the hospital as a valuable resource when caring for acutely ill patients.[21, 22]

Another potential explanation is that despite their preferences, PCPs who care for younger, minority patients lack access to hospitalist services. One large study of Medicare beneficiaries reported that physicians who care for black patients are less well‐trained clinically and often lack access to important clinical resources such as diagnostic imaging and nonemergency hospital admissions.[23] Similarly, international medical graduates are more likely than their US‐trained counterparts to care for underserved patients and to practice in small, independent offices.[24, 25, 26] As hospitalist groups often rely on cross‐subsidization from sources within a large healthcare organization, independent PCPs may have less access to their services when compared with PCPs in managed care organizations or large integrated groups. Viewed in this context, our findings imply that although hospitalists often care for socioeconomically vulnerable patients (eg, younger, uninsured, black men) who lack access to primary care services,[1] they also appear to share care responsibilities for more complex hospitalized patients with PCPs in more affluent communities. Further research may determine if the availability of hospitalists influences racial disparities in hospital care.

Our study has limitations. It is an observational study and thus subject to bias and confounding. As our cohort was formed using fee‐for‐service Medicare data in a single, large state, it may not be generalizable to PCPs who practice in other states, who care for a younger population, or who do not accept Medicare. Our findings also may not reflect the practice patterns of physicians‐in‐training, PCP populations with high board‐certification rates, those employed in temporary positions, or those who interrupt their practices for personal reasons, as we restricted our study to established PCPs who had been in practice long and consistently enough to be associated with 20 hospitalized patients during every year of the study. For example, the lower proportion of female PCPs in our cohort (15.6% in our study in 2009 vs 27.5% reported in a nationally representative 2008 survey[27]) may be explained by our exclusion of women who take prolonged time off for childcare duties. We also did not establish whether patient outcomes or healthcare costs differ between PCPs who adopted the hospitalist model and traditionalists. Finally, we could not examine the effect of a number of PCP factors that could plausibly influence whether or not PCPs relinquish inpatient care to hospitalists, such as their comfort with providing inpatient care, having hospital admitting privileges, having office‐based access to hospitals' electronic medical records, or the distance between their office and the hospital. However, this study lays the groundwork for future studies to explore these factors.

In summary, this study is the first, to our knowledge, to characterize PCPs who relinquished inpatient responsibilities to hospitalists. Our findings suggest that some groups of PCPs are more likely to refer patient to hospitalists, that the relationship between hospitalists and PCPs has evolved over time, and that the hospitalist model still has ample room to grow.

Although primary care physicians (PCPs) have traditionally treated patients in both ambulatory and hospital settings, many relinquished inpatient duties to hospitalists in recent decades.[1] Little is known about the PCPs who relinquished inpatient care duties or how the transition to the hospitalist model occurred. For example, what are the characteristics of PCPs who change? Do PCPs adopt the hospitalist model enthusiastically or cautiously? Characterizing PCPs who adopted the hospitalist model can help hospitalists understand their specialty's history and also inform health services research.

Much of the interest in the hospitalist model has been generated by studies reporting improved outcomes and lower hospital lengths of stay associated with hospitalist care.[2, 3, 4, 5] Conversely, detractors of the model point to reports of higher postacute care utilization among hospitalist patients.[6] Although these studies usually adjusted for differences among patients and hospitals, they did not account for PCP characteristics. As patients' access to PCPs and their PCP's capabilities are both plausible factors that could influence hospital length of stay (eg, decisions to complete more or less of a workup in the hospital), quality of care transitions, and postdischarge utilization, it is important to determine if PCPs who use hospitalists differ systematically from those who do not to correctly interpret health system utilization patterns that currently are attributed only to hospitalists.[7, 8]

We conducted this study to determine if observable PCP factors are associated with patients' use of hospitalists and to describe the trajectory by which PCPs referred their patients to hospitalists over time.

METHODS

RESULTS

During the 2001 through 2009 study period, between 2252 and 2848 PCPs were associated with at least 20 hospitalized beneficiaries in any single year. Among these, 1172 PCPs were associated with at least 20 hospitalized beneficiaries in every year of the study period. These 1172 PCPs were associated with 608,686 hospitalizations over the 9 years.

Table 1 presents the characteristics of the PCPs who contributed to the cross‐sectional analyses in 2001 (N=2252) and 2009 (N=2387), as well as the 1172 PCPs for whom we had data for all 9 years for the longitudinal analyses. Most PCPs were male, trained in the United States, and were board certified. The average number of Medicare patients seen by these PCPs and number of outpatient Medicare visits went up about 7% between 2001 and 2009.

Figure 1 graphs the percentage of PCPs as a function of what percent of their hospitalized patients received care from hospitalists, and how that changed from 2001 to 2009. For 70.9% of PCPs, fewer than 5% of their hospitalized patients received hospitalist care in 2001. By 2009, the percent of PCPs in this category had decreased to 15.2%. In contrast, in 2001, more than half of the patients for 2.1% of PCPs received hospitalist care, and the percent of PCPs in this category increased to 26.3% by 2009.

Figure 1

The pattern in Figure 1 shows that PCPs' use of hospitalists changed continuously and gradually over time. However, this pattern describes the PCPs as a group. When examined at the individual PCP level, different patterns emerge. Figure 2, which presents selected individual PCP's use of hospitalists over time, shows several distinct subpatterns of PCP practice behaviors. First, there are PCPs whose use of hospitalists was high in 2001 and stayed high or increased over time (eg, PCP A). There also were PCPs whose use of hospitalists stayed low over the entire study period (eg, PCP B). Finally, there were PCPs whose use of hospitalists was low in 2001 but high in 2009 (eg, PCP C). For this last group, the pattern of change in hospitalist utilization over time was discontinuous; that is, most of the increase occurred over a 1‐ or 2‐year period, instead of increasing gradually over time.

Figure 2

Among the 1172 PCPs associated with 20 hospitalized beneficiaries each year in all 9 years of the study period, group‐based trajectory modeling classified their practice patterns into 4 distinct trajectories (Figure 3). Among PCPs in group 1, more than one‐third of their hospitalized patients were cared for by hospitalists in 2001, and this increased to 60% by 2009. PCPs in groups 2 and 3 rarely used hospitalist care in 2001 but increased their use over time. The increase started early in the period for PCPs in group 2 and later for those in group 3. PCPs in group 4 were associated with little hospitalist use throughout the study period.

Figure 3

We constructed a model to describe the odds of a patient receiving care from hospitalists during the study period using patients associated with these 1172 PCPs. After adjusting for patient characteristics, the residual intraclass correlation coefficient for PCP level was 0.334, which indicates that 33.4% of the variance in whether a hospitalized patient received care from a hospitalist is explained by which PCP the patient saw. When adjusting for both patient and PCP characteristics, the overall odds of a patient receiving hospitalist care increased by 30% (95% confidence interval [CI]: 1.29‐1.30) per year from 2001 through 2009.

There were also significant interactions between year of hospitalization and several PCP characteristics. These interactions are illustrated in Table 2, which stratifies each of those PCP characteristics by 3 time periods: 2001 to 2003, 2004 to 2006, and 2007 to 2009. In all time periods, patients were more likely to receive hospitalist care if their PCP was US trained (US vs international medical graduate: odds ratio [OR]: 1.42, 95% CI: 1.19‐1.69 in 20012003; OR: 1.46, 95% CI: 1.23‐1.73 in 20072009), or specialized in family medicine (family medicine vs internal medicine: OR: 1.46, 95% CI: 1.25‐1.72 in 20012003; OR: 1.46, 95% CI: 1.25‐1.70 in 20072009). Over time, the relative odds of a patient receiving care from hospitalists decreased if their PCP was female (female vs male: OR: 1.91, 95% CI: 1.46‐2.50 in 20012003 vs OR: 1.50, 95% CI: 1.15‐1.95 in 20072009) or practiced in an urban area (largest vs smallest MSA: OR: 3.34, 95% CI: 2.72‐4.09 in 20012003; OR: 2.22, 95% CI: 1.82‐2.71 in 20072009). Although the longest‐practicing PCPs were most likely to use hospitalists in the early 2000s, this effect disappeared by 2007 to 2009 (most vs least years in practice: OR: 1.35, 95% CI: 1.06‐1.72 in 20012003 vs OR: 0.92, 95% CI: 0.73‐1.17 in 20072009).

In terms of PCP workload, patients of PCPs with high outpatient activity were more likely to receive hospitalists care throughout the study period, although the association had decreased by 2007 to 2009 (highest vs lowest outpatient volume: OR: 1.46, 95% CI: 1.30‐1.63 in 20012003 vs OR: 1.32, 95% CI: 1.20‐1.44 in 20072009). In contrast, PCPs with the lowest inpatient volumes became more likely to use hospitalists by the end of the study period (highest vs lowest inpatient volume: OR: 1.05, 95% CI: 0.95‐1.18 in 20012003 vs OR: 0.55, 95% CI: 0.51‐0.60 in 20072009).

The characteristics of PCPs' practice panels also were associated with patients' likelihood of receiving care from hospitalists. PCPs whose practice panels consisted of patients who were predominantly male, white, or with more outpatient comorbidities were consistently more likely to use hospitalists throughout the study period. PCPs with older patient panels were less likely to use hospitalists in 2001 to 2003, but by 2007 to 2009, they were slightly more likely to do so (oldest vs youngest average outpatient panel age: OR: 0.72, 95% CI: 0.64‐0.81 in 20012003 vs OR: 1.15, 95% CI: 1.05‐1.26 in 20072009).

CONCLUSIONS

Prior studies of the hospitalist model have shown that the likelihood of a patient receiving inpatient care from hospitalists is associated with patient characteristics, hospital characteristics, geographic region, and type of admission.[1, 16, 17] We found that PCP characteristics also predict whether patients receive care from hospitalists and that their use of hospitalists developed dynamically between 2001 to 2009. Although many factors (such as whether patients were admitted to a hospital where their PCP had admitting privileges) can influence the decision to use hospitalists, we found that over one‐third of the variance in whether a hospitalized patient received care from a hospitalist is explained by which PCP the patient saw. In showing that systemic differences exist among PCPs who use hospitalists and those who do not, our study suggests that future research on the hospitalist model should, if possible, adjust for PCP characteristics in addition to hospital and patient factors.

Although this study identifies the existence and magnitude of differences in whether or not PCPs use hospitalists, it cannot explain why the differences exist. We only can offer hypotheses. For example, our finding that PCPs with the most years of practice experience were more likely to use hospitalists in the early 2000s but not in more recent years suggests that in hospital medicine's early years, long‐practicing generalist physicians were choosing between practicing traditionalist medicine and adopting the hospitalists model, but by 2009, experienced generalist physicians had already specialized to either inpatient or outpatient settings earlier in their careers. On the other hand, the decreasing odds of urban PCPs using hospitalists may reflect a relative growth in hospitalist use in less populated areas rather than a change in urban PCPs' practice patterns.

PCPs trained in family medicine have reported less inpatient training and less comfort with providing hospital care,[18, 19] thus it is unsurprising that family physicians were more likely to refer patients to hospitalists. Although a recent study reported that family physicians' inpatient volumes remained constant, whereas those of outpatient internists declined between 2003 and 2012, the analysis used University Health Consortium data and thus reflects practice patterns in academic medical centers.[20] Our data suggest that outside of academia, family physicians have embraced the hospitalists as clinical partners.

Meltzer and Chung had previously proposed an economic model to describe the growing use of hospitalists in the United States. They posited that decisions to adopt the hospitalist model are governed by trade‐offs between coordination costs (eg, time and effort spent coordinating multiple providers across different settings) and switching costs (eg, time spent traveling between the office and the hospital or the effort of adjusting to different work settings).[16] The authors hypothesized that empirical testing of this model would show PCPs are more likely to use hospitalists if they have less available professional time (ie, work fewer hours per week), are female (due to competing demands from domestic responsibilities), have relatively few hospitalized patients, or live in areas with high traffic congestion. Our findings provide empirical evidence to support their division‐of‐labor model in showing that patients were more likely to receive hospitalist care if their PCP was female, practiced in an urban location, had higher outpatient practice volumes, or had lower inpatient volumes.

At first glance, some of our findings appear to contradict our earlier study, which showed that younger, black, male patients are more likely to receive inpatient care from hospitalists.[1] However, that study included patients regardless of whether they had a PCP. This study shows that when patients have a PCP, their PCPs are more likely to refer them to hospitalists if they are older, white, male, and have more comorbid conditions. A potential explanation for this finding is that PCPs may preferentially use hospitalists when caring for older and sicker hospitalized patients. For example, commentators often cite hospitalists' constant availability in the hospital as a valuable resource when caring for acutely ill patients.[21, 22]

Another potential explanation is that despite their preferences, PCPs who care for younger, minority patients lack access to hospitalist services. One large study of Medicare beneficiaries reported that physicians who care for black patients are less well‐trained clinically and often lack access to important clinical resources such as diagnostic imaging and nonemergency hospital admissions.[23] Similarly, international medical graduates are more likely than their US‐trained counterparts to care for underserved patients and to practice in small, independent offices.[24, 25, 26] As hospitalist groups often rely on cross‐subsidization from sources within a large healthcare organization, independent PCPs may have less access to their services when compared with PCPs in managed care organizations or large integrated groups. Viewed in this context, our findings imply that although hospitalists often care for socioeconomically vulnerable patients (eg, younger, uninsured, black men) who lack access to primary care services,[1] they also appear to share care responsibilities for more complex hospitalized patients with PCPs in more affluent communities. Further research may determine if the availability of hospitalists influences racial disparities in hospital care.

Our study has limitations. It is an observational study and thus subject to bias and confounding. As our cohort was formed using fee‐for‐service Medicare data in a single, large state, it may not be generalizable to PCPs who practice in other states, who care for a younger population, or who do not accept Medicare. Our findings also may not reflect the practice patterns of physicians‐in‐training, PCP populations with high board‐certification rates, those employed in temporary positions, or those who interrupt their practices for personal reasons, as we restricted our study to established PCPs who had been in practice long and consistently enough to be associated with 20 hospitalized patients during every year of the study. For example, the lower proportion of female PCPs in our cohort (15.6% in our study in 2009 vs 27.5% reported in a nationally representative 2008 survey[27]) may be explained by our exclusion of women who take prolonged time off for childcare duties. We also did not establish whether patient outcomes or healthcare costs differ between PCPs who adopted the hospitalist model and traditionalists. Finally, we could not examine the effect of a number of PCP factors that could plausibly influence whether or not PCPs relinquish inpatient care to hospitalists, such as their comfort with providing inpatient care, having hospital admitting privileges, having office‐based access to hospitals' electronic medical records, or the distance between their office and the hospital. However, this study lays the groundwork for future studies to explore these factors.

In summary, this study is the first, to our knowledge, to characterize PCPs who relinquished inpatient responsibilities to hospitalists. Our findings suggest that some groups of PCPs are more likely to refer patient to hospitalists, that the relationship between hospitalists and PCPs has evolved over time, and that the hospitalist model still has ample room to grow.

Although primary care physicians (PCPs) have traditionally treated patients in both ambulatory and hospital settings, many relinquished inpatient duties to hospitalists in recent decades.[1] Little is known about the PCPs who relinquished inpatient care duties or how the transition to the hospitalist model occurred. For example, what are the characteristics of PCPs who change? Do PCPs adopt the hospitalist model enthusiastically or cautiously? Characterizing PCPs who adopted the hospitalist model can help hospitalists understand their specialty's history and also inform health services research.

Much of the interest in the hospitalist model has been generated by studies reporting improved outcomes and lower hospital lengths of stay associated with hospitalist care.[2, 3, 4, 5] Conversely, detractors of the model point to reports of higher postacute care utilization among hospitalist patients.[6] Although these studies usually adjusted for differences among patients and hospitals, they did not account for PCP characteristics. As patients' access to PCPs and their PCP's capabilities are both plausible factors that could influence hospital length of stay (eg, decisions to complete more or less of a workup in the hospital), quality of care transitions, and postdischarge utilization, it is important to determine if PCPs who use hospitalists differ systematically from those who do not to correctly interpret health system utilization patterns that currently are attributed only to hospitalists.[7, 8]

We conducted this study to determine if observable PCP factors are associated with patients' use of hospitalists and to describe the trajectory by which PCPs referred their patients to hospitalists over time.

METHODS

RESULTS

During the 2001 through 2009 study period, between 2252 and 2848 PCPs were associated with at least 20 hospitalized beneficiaries in any single year. Among these, 1172 PCPs were associated with at least 20 hospitalized beneficiaries in every year of the study period. These 1172 PCPs were associated with 608,686 hospitalizations over the 9 years.

Table 1 presents the characteristics of the PCPs who contributed to the cross‐sectional analyses in 2001 (N=2252) and 2009 (N=2387), as well as the 1172 PCPs for whom we had data for all 9 years for the longitudinal analyses. Most PCPs were male, trained in the United States, and were board certified. The average number of Medicare patients seen by these PCPs and number of outpatient Medicare visits went up about 7% between 2001 and 2009.

Figure 1 graphs the percentage of PCPs as a function of what percent of their hospitalized patients received care from hospitalists, and how that changed from 2001 to 2009. For 70.9% of PCPs, fewer than 5% of their hospitalized patients received hospitalist care in 2001. By 2009, the percent of PCPs in this category had decreased to 15.2%. In contrast, in 2001, more than half of the patients for 2.1% of PCPs received hospitalist care, and the percent of PCPs in this category increased to 26.3% by 2009.

Figure 1

The pattern in Figure 1 shows that PCPs' use of hospitalists changed continuously and gradually over time. However, this pattern describes the PCPs as a group. When examined at the individual PCP level, different patterns emerge. Figure 2, which presents selected individual PCP's use of hospitalists over time, shows several distinct subpatterns of PCP practice behaviors. First, there are PCPs whose use of hospitalists was high in 2001 and stayed high or increased over time (eg, PCP A). There also were PCPs whose use of hospitalists stayed low over the entire study period (eg, PCP B). Finally, there were PCPs whose use of hospitalists was low in 2001 but high in 2009 (eg, PCP C). For this last group, the pattern of change in hospitalist utilization over time was discontinuous; that is, most of the increase occurred over a 1‐ or 2‐year period, instead of increasing gradually over time.

Figure 2

Among the 1172 PCPs associated with 20 hospitalized beneficiaries each year in all 9 years of the study period, group‐based trajectory modeling classified their practice patterns into 4 distinct trajectories (Figure 3). Among PCPs in group 1, more than one‐third of their hospitalized patients were cared for by hospitalists in 2001, and this increased to 60% by 2009. PCPs in groups 2 and 3 rarely used hospitalist care in 2001 but increased their use over time. The increase started early in the period for PCPs in group 2 and later for those in group 3. PCPs in group 4 were associated with little hospitalist use throughout the study period.

Figure 3

We constructed a model to describe the odds of a patient receiving care from hospitalists during the study period using patients associated with these 1172 PCPs. After adjusting for patient characteristics, the residual intraclass correlation coefficient for PCP level was 0.334, which indicates that 33.4% of the variance in whether a hospitalized patient received care from a hospitalist is explained by which PCP the patient saw. When adjusting for both patient and PCP characteristics, the overall odds of a patient receiving hospitalist care increased by 30% (95% confidence interval [CI]: 1.29‐1.30) per year from 2001 through 2009.

There were also significant interactions between year of hospitalization and several PCP characteristics. These interactions are illustrated in Table 2, which stratifies each of those PCP characteristics by 3 time periods: 2001 to 2003, 2004 to 2006, and 2007 to 2009. In all time periods, patients were more likely to receive hospitalist care if their PCP was US trained (US vs international medical graduate: odds ratio [OR]: 1.42, 95% CI: 1.19‐1.69 in 20012003; OR: 1.46, 95% CI: 1.23‐1.73 in 20072009), or specialized in family medicine (family medicine vs internal medicine: OR: 1.46, 95% CI: 1.25‐1.72 in 20012003; OR: 1.46, 95% CI: 1.25‐1.70 in 20072009). Over time, the relative odds of a patient receiving care from hospitalists decreased if their PCP was female (female vs male: OR: 1.91, 95% CI: 1.46‐2.50 in 20012003 vs OR: 1.50, 95% CI: 1.15‐1.95 in 20072009) or practiced in an urban area (largest vs smallest MSA: OR: 3.34, 95% CI: 2.72‐4.09 in 20012003; OR: 2.22, 95% CI: 1.82‐2.71 in 20072009). Although the longest‐practicing PCPs were most likely to use hospitalists in the early 2000s, this effect disappeared by 2007 to 2009 (most vs least years in practice: OR: 1.35, 95% CI: 1.06‐1.72 in 20012003 vs OR: 0.92, 95% CI: 0.73‐1.17 in 20072009).

In terms of PCP workload, patients of PCPs with high outpatient activity were more likely to receive hospitalists care throughout the study period, although the association had decreased by 2007 to 2009 (highest vs lowest outpatient volume: OR: 1.46, 95% CI: 1.30‐1.63 in 20012003 vs OR: 1.32, 95% CI: 1.20‐1.44 in 20072009). In contrast, PCPs with the lowest inpatient volumes became more likely to use hospitalists by the end of the study period (highest vs lowest inpatient volume: OR: 1.05, 95% CI: 0.95‐1.18 in 20012003 vs OR: 0.55, 95% CI: 0.51‐0.60 in 20072009).

The characteristics of PCPs' practice panels also were associated with patients' likelihood of receiving care from hospitalists. PCPs whose practice panels consisted of patients who were predominantly male, white, or with more outpatient comorbidities were consistently more likely to use hospitalists throughout the study period. PCPs with older patient panels were less likely to use hospitalists in 2001 to 2003, but by 2007 to 2009, they were slightly more likely to do so (oldest vs youngest average outpatient panel age: OR: 0.72, 95% CI: 0.64‐0.81 in 20012003 vs OR: 1.15, 95% CI: 1.05‐1.26 in 20072009).

CONCLUSIONS

Prior studies of the hospitalist model have shown that the likelihood of a patient receiving inpatient care from hospitalists is associated with patient characteristics, hospital characteristics, geographic region, and type of admission.[1, 16, 17] We found that PCP characteristics also predict whether patients receive care from hospitalists and that their use of hospitalists developed dynamically between 2001 to 2009. Although many factors (such as whether patients were admitted to a hospital where their PCP had admitting privileges) can influence the decision to use hospitalists, we found that over one‐third of the variance in whether a hospitalized patient received care from a hospitalist is explained by which PCP the patient saw. In showing that systemic differences exist among PCPs who use hospitalists and those who do not, our study suggests that future research on the hospitalist model should, if possible, adjust for PCP characteristics in addition to hospital and patient factors.

Although this study identifies the existence and magnitude of differences in whether or not PCPs use hospitalists, it cannot explain why the differences exist. We only can offer hypotheses. For example, our finding that PCPs with the most years of practice experience were more likely to use hospitalists in the early 2000s but not in more recent years suggests that in hospital medicine's early years, long‐practicing generalist physicians were choosing between practicing traditionalist medicine and adopting the hospitalists model, but by 2009, experienced generalist physicians had already specialized to either inpatient or outpatient settings earlier in their careers. On the other hand, the decreasing odds of urban PCPs using hospitalists may reflect a relative growth in hospitalist use in less populated areas rather than a change in urban PCPs' practice patterns.

PCPs trained in family medicine have reported less inpatient training and less comfort with providing hospital care,[18, 19] thus it is unsurprising that family physicians were more likely to refer patients to hospitalists. Although a recent study reported that family physicians' inpatient volumes remained constant, whereas those of outpatient internists declined between 2003 and 2012, the analysis used University Health Consortium data and thus reflects practice patterns in academic medical centers.[20] Our data suggest that outside of academia, family physicians have embraced the hospitalists as clinical partners.

Meltzer and Chung had previously proposed an economic model to describe the growing use of hospitalists in the United States. They posited that decisions to adopt the hospitalist model are governed by trade‐offs between coordination costs (eg, time and effort spent coordinating multiple providers across different settings) and switching costs (eg, time spent traveling between the office and the hospital or the effort of adjusting to different work settings).[16] The authors hypothesized that empirical testing of this model would show PCPs are more likely to use hospitalists if they have less available professional time (ie, work fewer hours per week), are female (due to competing demands from domestic responsibilities), have relatively few hospitalized patients, or live in areas with high traffic congestion. Our findings provide empirical evidence to support their division‐of‐labor model in showing that patients were more likely to receive hospitalist care if their PCP was female, practiced in an urban location, had higher outpatient practice volumes, or had lower inpatient volumes.

At first glance, some of our findings appear to contradict our earlier study, which showed that younger, black, male patients are more likely to receive inpatient care from hospitalists.[1] However, that study included patients regardless of whether they had a PCP. This study shows that when patients have a PCP, their PCPs are more likely to refer them to hospitalists if they are older, white, male, and have more comorbid conditions. A potential explanation for this finding is that PCPs may preferentially use hospitalists when caring for older and sicker hospitalized patients. For example, commentators often cite hospitalists' constant availability in the hospital as a valuable resource when caring for acutely ill patients.[21, 22]

Another potential explanation is that despite their preferences, PCPs who care for younger, minority patients lack access to hospitalist services. One large study of Medicare beneficiaries reported that physicians who care for black patients are less well‐trained clinically and often lack access to important clinical resources such as diagnostic imaging and nonemergency hospital admissions.[23] Similarly, international medical graduates are more likely than their US‐trained counterparts to care for underserved patients and to practice in small, independent offices.[24, 25, 26] As hospitalist groups often rely on cross‐subsidization from sources within a large healthcare organization, independent PCPs may have less access to their services when compared with PCPs in managed care organizations or large integrated groups. Viewed in this context, our findings imply that although hospitalists often care for socioeconomically vulnerable patients (eg, younger, uninsured, black men) who lack access to primary care services,[1] they also appear to share care responsibilities for more complex hospitalized patients with PCPs in more affluent communities. Further research may determine if the availability of hospitalists influences racial disparities in hospital care.

Our study has limitations. It is an observational study and thus subject to bias and confounding. As our cohort was formed using fee‐for‐service Medicare data in a single, large state, it may not be generalizable to PCPs who practice in other states, who care for a younger population, or who do not accept Medicare. Our findings also may not reflect the practice patterns of physicians‐in‐training, PCP populations with high board‐certification rates, those employed in temporary positions, or those who interrupt their practices for personal reasons, as we restricted our study to established PCPs who had been in practice long and consistently enough to be associated with 20 hospitalized patients during every year of the study. For example, the lower proportion of female PCPs in our cohort (15.6% in our study in 2009 vs 27.5% reported in a nationally representative 2008 survey[27]) may be explained by our exclusion of women who take prolonged time off for childcare duties. We also did not establish whether patient outcomes or healthcare costs differ between PCPs who adopted the hospitalist model and traditionalists. Finally, we could not examine the effect of a number of PCP factors that could plausibly influence whether or not PCPs relinquish inpatient care to hospitalists, such as their comfort with providing inpatient care, having hospital admitting privileges, having office‐based access to hospitals' electronic medical records, or the distance between their office and the hospital. However, this study lays the groundwork for future studies to explore these factors.

In summary, this study is the first, to our knowledge, to characterize PCPs who relinquished inpatient responsibilities to hospitalists. Our findings suggest that some groups of PCPs are more likely to refer patient to hospitalists, that the relationship between hospitalists and PCPs has evolved over time, and that the hospitalist model still has ample room to grow.

Continuity of care is considered by many physicians to be of critical importance in providing high‐quality patient care. Most of the research to date has focused on continuity in outpatient primary care. Research on outpatient continuity of care has been facilitated by the fact that a number of measurement tools for outpatient continuity exist.1 Outpatient continuity of care has been linked to better quality of life scores,2 lower costs,3 and less emergency room use.4 As hospital medicine has taken on more and more of the responsibility of inpatient care, primary care doctors have voiced concerns about the impact of hospitalists on overall continuity of care5 and the quality of the doctorpatient relationship.6

Recently, continuity of care in the hospital setting has also received attention. When the Accreditation Council for Graduate Medical Education (ACGME) first proposed restrictions to resident duty hours, the importance of continuity of inpatient care began to be debated in earnest in large part because of the increase in hand‐offs which accompanies discontinuity.7, 8 A recent study of hospitalist communication documented that as many as 13% of hand‐offs at the time of service changes are judged as incomplete by the receiving physician. These incomplete hand‐offs were more likely to be associated with uncertainty regarding the plan of care, as well as perceived near misses or adverse events.9 In addition, several case reports and studies suggest that systems with less continuity may have poorer outcomes.7, 1015

Continuity in the hospital setting is likely to be important for several reasons. First, the acuity of a patient's problem during a hospitalization is likely greater than during an outpatient visit. Thus the complexity of information to be transferred between physicians during a hospital stay is correspondingly greater. Second, the diagnostic uncertainty surrounding many admissions leads to complex thought processes that may be difficult to recreate when handing off patient care to another physician. Finally, knowledge of a patient's hospital course and the likely trajectory of care is facilitated by firsthand knowledge of where the patient has been. All this information can be difficult to distill into a brief sign‐out to another physician who assumes care of the patient.

In the current study, we sought to examine the trends over time in continuity of inpatient care. We chose patients likely to be cared for by general internists: those hospitalized for chronic obstructive pulmonary disease (COPD), pneumonia, and congestive heart failure (CHF). The general internists caring for patients in the hospital could be the patient's primary care physician (PCP), a physician covering for the patient's PCP, a physician assigned at admission by the hospital, or a hospitalist. Our goals were to describe the current level of continuity of care in the hospital setting, to examine whether continuity has changed over time, and to determine factors affecting continuity of care.

Methods

We used a 5% national sample of claims data from Medicare beneficiaries for the years 19962006.16 This included Medicare enrollment files, Medicare Provider Analysis and Review (MEDPAR) files, Medicare Carrier files, and Provider of Services (POS) files.17, 18

Results

Between 1996 and 2006, 528,453 patients hospitalized for COPD, pneumonia, and CHF received care by a generalist physician during their hospital stay. Of these, 64.3% were seen by one generalist physician, 26.9% by two generalist physicians, and 8.8% by three or more generalist physicians during hospitalization.

Figure 1 shows the percentage of all patients seen by 1, 2, and 3 or more generalist physicians between 1996 and 2006. The percentage of patients receiving care from one generalist physician declined from 70.7% in 1996 to 59.4% in 2006 (P < 0.001). During the same period, the percentage of patients receiving care from 3 or more generalist physicians increased from 6.5% to 10.7% (P < 0.001). Similar trends were seen for each of the 3 conditions. There was a decrease in overall length of stay during this period, from a mean of 5.7 to 4.9 days (P < 0.001). The increase in the number of generalist physicians providing care during the hospital stay did not correspond to an increase in total number of visits during the hospitalization. The average number of daily visits from a generalist physician was 0.94 (0.30) in 1996 and 0.96 (0.35) in 2006.

Figure 1

Table 1 presents the percentage of patients receiving care from 1, 2, and 3 or more generalist physicians during hospitalization stratified by patient and hospital characteristics. Older adults, females, non‐Hispanic whites, those with higher socioeconomic status, and those with more comorbidities were more likely to receive care by multiple generalist physicians. There was also large variation by geographic region, metropolitan area size, and hospital characteristics. All of these differences were significant at the P < 0.0001 level.

Table 2 presents the results of a multivariable analysis of factors independently associated with experiencing continuity of care. In this analysis, continuity of care was defined as receiving inpatient care from one generalist physician (vs two or more). In the unadjusted models, the odds of experiencing continuity of care decreased by 5.5% per year from 1996 through 2006, and this decrease did not substantially change after adjusting for all other variables (4.8% yearly decrease). Younger patients, females, black patients, and those with low socioeconomic status were slightly more likely to experience continuity of care. As expected, patients admitted on weekends, emergency admissions, and those with intensive care unit (ICU) stays were less likely to experience continuity. There were marked geographic variations in continuity, with continuity approximately half as likely in New England as in the South. Continuity was greatest in smaller metropolitan areas versus rural and large metropolitan areas. Hospital size and teaching status produced only minor variation.

In Table 2 we also show that patients with an established PCP and those who received care from a hospitalist in the hospital were substantially less likely to experience continuity of care. There are several possible interpretations for that finding. For example, it might be that patients admitted to a hospitalist service were likely to see multiple hospitalists. Alternatively, the decreased continuity associated with hospitalists could reflect the fact that some patients cared for predominantly by non‐hospitalists may have seen a hospitalist on call for a sudden change in health status. To further explore these possible explanatory pathways, we constructed three new cohorts: 1) patients receiving all their care from non‐hospitalists, 2) patients receiving all their care from hospitalists, and 3) patients seen by both. As shown in Table 3, in patients seen by non‐hospitalists only, the mean number of generalist physicians seen during hospitalization was slightly greater than in patients cared for only by hospitalists.

We also tested for interactions in Table 2 between admission year and other factors. There was a significant interaction between admission year and having an identifiable PCP in the year prior to admission (Table 2). The odds of experiencing continuity of care decreased more rapidly for patients who did not have a PCP (5.5% per year; 95% CI: 5.2%5.8%) than for those who had one (4.3% per year; 95% CI: 4.1%4.6%).

Discussion

We conducted this study to better understand the degree to which hospitalized patients experience discontinuity of care within a hospital stay and to determine which patients are most likely to experience discontinuity. In our study, we specifically chose admission conditions that would likely be followed primarily by generalist physicians. We found that, over the past decade, discontinuity of care for hospitalized patients has increased substantially, as indicated by the proportion of patients taken care of by more than one generalist physician during a single hospital stay. This occurred even though overall length of stay was decreasing in this same period.

It is perhaps not surprising that inpatient continuity of care has been decreasing in the past 10 years. Outpatient practices are becoming busier, and more doctors are practicing in large group practices, which could lead to several different physicians in the same practice rounding on a hospitalized patient. We have previously demonstrated that hospitalists are caring for an increasing number of patients over this same time period,21 so another possibility is that hospitalist services are being used more often because of this heavy outpatient workload. Our analyses allowed us to test the hypothesis that having hospitalists involved in patient care increases discontinuity.

At first glance, it appears that being cared for by hospitalists may result in worse continuity of care. However, closer scrutiny of the data reveals that the discontinuity ascribed to the hospitalists in the multivariable model appears to be an artifact of defining the hospitalist variable as having been seen by any hospitalist during the hospital stay. This would include patients who saw a hospitalist in addition to their PCP or another non‐hospitalist generalist. When we compared hospitalist‐only care to other generalist care, we could not detect a difference in discontinuity. We know that generalist visits per day to patients has not substantially increased over time, so this discontinuity trend is not explained by having visits by both a hospitalist and the PCP. Therefore, this combination of findings suggests that the increased discontinuity associated with having a hospitalist involved in patient care is likely the result of system issues rather than hospitalist care per se. In fact, patients seem to experience slightly better continuity when they see only hospitalists as opposed to only non‐hospitalists.

What types of systems issues might lead to this finding? Generalists in most settings could choose to involve a hospitalist at any point in the patient's hospital stay. This could occur because of a change in patient acuity requiring the involvement of hospitalists who are present in the hospital more. It is also possible that hospitalists' schedules are created to maximize inpatient continuity of care with individual hospitalists. Even though hospitalists clearly work shifts, the 7 on, 7 off model22 likely results in patients seeing the same physician each day until the switch day. This is in contrast to outpatient primary care doctors whose concentration may be on maintaining continuity within their practice.

As the field of hospital medicine was emerging, many internal medicine physicians from various specialties were concerned about the impact of hospitalists on patient care. In one study, 73% of internal medicine physicians who were not hospitalists thought that hospitalists would worsen continuity of care.23 Primary care and subspecialist internal medicine physicians also expressed the concern that hospitalists could hurt their own relationships with patients,6 presumably because of lost continuity between the inpatient and outpatient settings. However, this fear seems to diminish once hospitalist programs are implemented and primary care doctors have experience with them.23 Our study suggests that the decrease in continuity that has occurred since these studies were published is not likely due to the emergence of hospital medicine, but rather due to other factors that influence who cares for hospitalized patients.

This study had some limitations. Length of stay is an obvious mediator of number of generalist physicians seen. Therefore, the sickest patients are likely to have both a long length of stay and low continuity. We adjusted for this in the multivariable modeling. In addition, given that this study used a large database, certain details are not discernable. For example, we chose to operationalize discontinuity as visits from multiple generalists during a single hospital stay. That is not a perfect definition, but it does represent multiple physicians directing the care of a patient. Importantly, this does not appear to represent continuity with one physician with extra visits from another, as the total number of generalist visits per day did not change over time. It is also possible that patients in the non‐hospitalist group saw physicians only from a single practice, but those details are not included in the database. Finally, we cannot tell what type of hand‐offs were occurring for individual patients during each hospital stay. Despite these disadvantages, using a large database like this one allows for detection of fairly small differences that could still be clinically important.

In summary, hospitalized patients appear to experience less continuity now than 10 years ago. However, the hospitalist model does not appear to play a role in this discontinuity. It is worth exploring in more detail why patients would see both hospitalists and other generalists. This pattern is not surprising, but may have some repercussions in terms of increasing the number of hand‐offs experienced by patients. These could lead to problems with patient safety and quality of care. Future work should explore the reasons for this discontinuity and look at the relationship between inpatient discontinuity outcomes such as quality of care and the doctorpatient relationship.

Continuity of care is considered by many physicians to be of critical importance in providing high‐quality patient care. Most of the research to date has focused on continuity in outpatient primary care. Research on outpatient continuity of care has been facilitated by the fact that a number of measurement tools for outpatient continuity exist.1 Outpatient continuity of care has been linked to better quality of life scores,2 lower costs,3 and less emergency room use.4 As hospital medicine has taken on more and more of the responsibility of inpatient care, primary care doctors have voiced concerns about the impact of hospitalists on overall continuity of care5 and the quality of the doctorpatient relationship.6

Recently, continuity of care in the hospital setting has also received attention. When the Accreditation Council for Graduate Medical Education (ACGME) first proposed restrictions to resident duty hours, the importance of continuity of inpatient care began to be debated in earnest in large part because of the increase in hand‐offs which accompanies discontinuity.7, 8 A recent study of hospitalist communication documented that as many as 13% of hand‐offs at the time of service changes are judged as incomplete by the receiving physician. These incomplete hand‐offs were more likely to be associated with uncertainty regarding the plan of care, as well as perceived near misses or adverse events.9 In addition, several case reports and studies suggest that systems with less continuity may have poorer outcomes.7, 1015

Continuity in the hospital setting is likely to be important for several reasons. First, the acuity of a patient's problem during a hospitalization is likely greater than during an outpatient visit. Thus the complexity of information to be transferred between physicians during a hospital stay is correspondingly greater. Second, the diagnostic uncertainty surrounding many admissions leads to complex thought processes that may be difficult to recreate when handing off patient care to another physician. Finally, knowledge of a patient's hospital course and the likely trajectory of care is facilitated by firsthand knowledge of where the patient has been. All this information can be difficult to distill into a brief sign‐out to another physician who assumes care of the patient.

In the current study, we sought to examine the trends over time in continuity of inpatient care. We chose patients likely to be cared for by general internists: those hospitalized for chronic obstructive pulmonary disease (COPD), pneumonia, and congestive heart failure (CHF). The general internists caring for patients in the hospital could be the patient's primary care physician (PCP), a physician covering for the patient's PCP, a physician assigned at admission by the hospital, or a hospitalist. Our goals were to describe the current level of continuity of care in the hospital setting, to examine whether continuity has changed over time, and to determine factors affecting continuity of care.

Methods

We used a 5% national sample of claims data from Medicare beneficiaries for the years 19962006.16 This included Medicare enrollment files, Medicare Provider Analysis and Review (MEDPAR) files, Medicare Carrier files, and Provider of Services (POS) files.17, 18

Results

Between 1996 and 2006, 528,453 patients hospitalized for COPD, pneumonia, and CHF received care by a generalist physician during their hospital stay. Of these, 64.3% were seen by one generalist physician, 26.9% by two generalist physicians, and 8.8% by three or more generalist physicians during hospitalization.

Figure 1 shows the percentage of all patients seen by 1, 2, and 3 or more generalist physicians between 1996 and 2006. The percentage of patients receiving care from one generalist physician declined from 70.7% in 1996 to 59.4% in 2006 (P < 0.001). During the same period, the percentage of patients receiving care from 3 or more generalist physicians increased from 6.5% to 10.7% (P < 0.001). Similar trends were seen for each of the 3 conditions. There was a decrease in overall length of stay during this period, from a mean of 5.7 to 4.9 days (P < 0.001). The increase in the number of generalist physicians providing care during the hospital stay did not correspond to an increase in total number of visits during the hospitalization. The average number of daily visits from a generalist physician was 0.94 (0.30) in 1996 and 0.96 (0.35) in 2006.

Figure 1

Table 1 presents the percentage of patients receiving care from 1, 2, and 3 or more generalist physicians during hospitalization stratified by patient and hospital characteristics. Older adults, females, non‐Hispanic whites, those with higher socioeconomic status, and those with more comorbidities were more likely to receive care by multiple generalist physicians. There was also large variation by geographic region, metropolitan area size, and hospital characteristics. All of these differences were significant at the P < 0.0001 level.

Table 2 presents the results of a multivariable analysis of factors independently associated with experiencing continuity of care. In this analysis, continuity of care was defined as receiving inpatient care from one generalist physician (vs two or more). In the unadjusted models, the odds of experiencing continuity of care decreased by 5.5% per year from 1996 through 2006, and this decrease did not substantially change after adjusting for all other variables (4.8% yearly decrease). Younger patients, females, black patients, and those with low socioeconomic status were slightly more likely to experience continuity of care. As expected, patients admitted on weekends, emergency admissions, and those with intensive care unit (ICU) stays were less likely to experience continuity. There were marked geographic variations in continuity, with continuity approximately half as likely in New England as in the South. Continuity was greatest in smaller metropolitan areas versus rural and large metropolitan areas. Hospital size and teaching status produced only minor variation.

In Table 2 we also show that patients with an established PCP and those who received care from a hospitalist in the hospital were substantially less likely to experience continuity of care. There are several possible interpretations for that finding. For example, it might be that patients admitted to a hospitalist service were likely to see multiple hospitalists. Alternatively, the decreased continuity associated with hospitalists could reflect the fact that some patients cared for predominantly by non‐hospitalists may have seen a hospitalist on call for a sudden change in health status. To further explore these possible explanatory pathways, we constructed three new cohorts: 1) patients receiving all their care from non‐hospitalists, 2) patients receiving all their care from hospitalists, and 3) patients seen by both. As shown in Table 3, in patients seen by non‐hospitalists only, the mean number of generalist physicians seen during hospitalization was slightly greater than in patients cared for only by hospitalists.

We also tested for interactions in Table 2 between admission year and other factors. There was a significant interaction between admission year and having an identifiable PCP in the year prior to admission (Table 2). The odds of experiencing continuity of care decreased more rapidly for patients who did not have a PCP (5.5% per year; 95% CI: 5.2%5.8%) than for those who had one (4.3% per year; 95% CI: 4.1%4.6%).

Discussion

We conducted this study to better understand the degree to which hospitalized patients experience discontinuity of care within a hospital stay and to determine which patients are most likely to experience discontinuity. In our study, we specifically chose admission conditions that would likely be followed primarily by generalist physicians. We found that, over the past decade, discontinuity of care for hospitalized patients has increased substantially, as indicated by the proportion of patients taken care of by more than one generalist physician during a single hospital stay. This occurred even though overall length of stay was decreasing in this same period.

It is perhaps not surprising that inpatient continuity of care has been decreasing in the past 10 years. Outpatient practices are becoming busier, and more doctors are practicing in large group practices, which could lead to several different physicians in the same practice rounding on a hospitalized patient. We have previously demonstrated that hospitalists are caring for an increasing number of patients over this same time period,21 so another possibility is that hospitalist services are being used more often because of this heavy outpatient workload. Our analyses allowed us to test the hypothesis that having hospitalists involved in patient care increases discontinuity.

At first glance, it appears that being cared for by hospitalists may result in worse continuity of care. However, closer scrutiny of the data reveals that the discontinuity ascribed to the hospitalists in the multivariable model appears to be an artifact of defining the hospitalist variable as having been seen by any hospitalist during the hospital stay. This would include patients who saw a hospitalist in addition to their PCP or another non‐hospitalist generalist. When we compared hospitalist‐only care to other generalist care, we could not detect a difference in discontinuity. We know that generalist visits per day to patients has not substantially increased over time, so this discontinuity trend is not explained by having visits by both a hospitalist and the PCP. Therefore, this combination of findings suggests that the increased discontinuity associated with having a hospitalist involved in patient care is likely the result of system issues rather than hospitalist care per se. In fact, patients seem to experience slightly better continuity when they see only hospitalists as opposed to only non‐hospitalists.

What types of systems issues might lead to this finding? Generalists in most settings could choose to involve a hospitalist at any point in the patient's hospital stay. This could occur because of a change in patient acuity requiring the involvement of hospitalists who are present in the hospital more. It is also possible that hospitalists' schedules are created to maximize inpatient continuity of care with individual hospitalists. Even though hospitalists clearly work shifts, the 7 on, 7 off model22 likely results in patients seeing the same physician each day until the switch day. This is in contrast to outpatient primary care doctors whose concentration may be on maintaining continuity within their practice.

As the field of hospital medicine was emerging, many internal medicine physicians from various specialties were concerned about the impact of hospitalists on patient care. In one study, 73% of internal medicine physicians who were not hospitalists thought that hospitalists would worsen continuity of care.23 Primary care and subspecialist internal medicine physicians also expressed the concern that hospitalists could hurt their own relationships with patients,6 presumably because of lost continuity between the inpatient and outpatient settings. However, this fear seems to diminish once hospitalist programs are implemented and primary care doctors have experience with them.23 Our study suggests that the decrease in continuity that has occurred since these studies were published is not likely due to the emergence of hospital medicine, but rather due to other factors that influence who cares for hospitalized patients.

This study had some limitations. Length of stay is an obvious mediator of number of generalist physicians seen. Therefore, the sickest patients are likely to have both a long length of stay and low continuity. We adjusted for this in the multivariable modeling. In addition, given that this study used a large database, certain details are not discernable. For example, we chose to operationalize discontinuity as visits from multiple generalists during a single hospital stay. That is not a perfect definition, but it does represent multiple physicians directing the care of a patient. Importantly, this does not appear to represent continuity with one physician with extra visits from another, as the total number of generalist visits per day did not change over time. It is also possible that patients in the non‐hospitalist group saw physicians only from a single practice, but those details are not included in the database. Finally, we cannot tell what type of hand‐offs were occurring for individual patients during each hospital stay. Despite these disadvantages, using a large database like this one allows for detection of fairly small differences that could still be clinically important.

In summary, hospitalized patients appear to experience less continuity now than 10 years ago. However, the hospitalist model does not appear to play a role in this discontinuity. It is worth exploring in more detail why patients would see both hospitalists and other generalists. This pattern is not surprising, but may have some repercussions in terms of increasing the number of hand‐offs experienced by patients. These could lead to problems with patient safety and quality of care. Future work should explore the reasons for this discontinuity and look at the relationship between inpatient discontinuity outcomes such as quality of care and the doctorpatient relationship.

Continuity of care is considered by many physicians to be of critical importance in providing high‐quality patient care. Most of the research to date has focused on continuity in outpatient primary care. Research on outpatient continuity of care has been facilitated by the fact that a number of measurement tools for outpatient continuity exist.1 Outpatient continuity of care has been linked to better quality of life scores,2 lower costs,3 and less emergency room use.4 As hospital medicine has taken on more and more of the responsibility of inpatient care, primary care doctors have voiced concerns about the impact of hospitalists on overall continuity of care5 and the quality of the doctorpatient relationship.6

Recently, continuity of care in the hospital setting has also received attention. When the Accreditation Council for Graduate Medical Education (ACGME) first proposed restrictions to resident duty hours, the importance of continuity of inpatient care began to be debated in earnest in large part because of the increase in hand‐offs which accompanies discontinuity.7, 8 A recent study of hospitalist communication documented that as many as 13% of hand‐offs at the time of service changes are judged as incomplete by the receiving physician. These incomplete hand‐offs were more likely to be associated with uncertainty regarding the plan of care, as well as perceived near misses or adverse events.9 In addition, several case reports and studies suggest that systems with less continuity may have poorer outcomes.7, 1015

Continuity in the hospital setting is likely to be important for several reasons. First, the acuity of a patient's problem during a hospitalization is likely greater than during an outpatient visit. Thus the complexity of information to be transferred between physicians during a hospital stay is correspondingly greater. Second, the diagnostic uncertainty surrounding many admissions leads to complex thought processes that may be difficult to recreate when handing off patient care to another physician. Finally, knowledge of a patient's hospital course and the likely trajectory of care is facilitated by firsthand knowledge of where the patient has been. All this information can be difficult to distill into a brief sign‐out to another physician who assumes care of the patient.

In the current study, we sought to examine the trends over time in continuity of inpatient care. We chose patients likely to be cared for by general internists: those hospitalized for chronic obstructive pulmonary disease (COPD), pneumonia, and congestive heart failure (CHF). The general internists caring for patients in the hospital could be the patient's primary care physician (PCP), a physician covering for the patient's PCP, a physician assigned at admission by the hospital, or a hospitalist. Our goals were to describe the current level of continuity of care in the hospital setting, to examine whether continuity has changed over time, and to determine factors affecting continuity of care.

Methods

We used a 5% national sample of claims data from Medicare beneficiaries for the years 19962006.16 This included Medicare enrollment files, Medicare Provider Analysis and Review (MEDPAR) files, Medicare Carrier files, and Provider of Services (POS) files.17, 18

Results

Between 1996 and 2006, 528,453 patients hospitalized for COPD, pneumonia, and CHF received care by a generalist physician during their hospital stay. Of these, 64.3% were seen by one generalist physician, 26.9% by two generalist physicians, and 8.8% by three or more generalist physicians during hospitalization.

Figure 1 shows the percentage of all patients seen by 1, 2, and 3 or more generalist physicians between 1996 and 2006. The percentage of patients receiving care from one generalist physician declined from 70.7% in 1996 to 59.4% in 2006 (P < 0.001). During the same period, the percentage of patients receiving care from 3 or more generalist physicians increased from 6.5% to 10.7% (P < 0.001). Similar trends were seen for each of the 3 conditions. There was a decrease in overall length of stay during this period, from a mean of 5.7 to 4.9 days (P < 0.001). The increase in the number of generalist physicians providing care during the hospital stay did not correspond to an increase in total number of visits during the hospitalization. The average number of daily visits from a generalist physician was 0.94 (0.30) in 1996 and 0.96 (0.35) in 2006.

Figure 1

Table 1 presents the percentage of patients receiving care from 1, 2, and 3 or more generalist physicians during hospitalization stratified by patient and hospital characteristics. Older adults, females, non‐Hispanic whites, those with higher socioeconomic status, and those with more comorbidities were more likely to receive care by multiple generalist physicians. There was also large variation by geographic region, metropolitan area size, and hospital characteristics. All of these differences were significant at the P < 0.0001 level.