Friend or Foe?How Primary Care Physicians Perceive Hospitalists

Background  Increased use of hospitalists is redefining the role of primary care physicians. Whether primary care physicians welcome this transition is unknown. We examined primary care physicians' perceptions of how hospitalists affect their practices, their patient relationships, and overall patient care.

Methods  A mailed survey of randomly selected general internists, general pediatricians, and family practitioners with experience with hospitalists practicing in California.

Main Outcome Measures  Physicians' self-reports of hospitalists' effects on quality of patient care and on their own practices.

Results  Seven hundred eight physicians were eligible for this study, and there was a 74% response rate. Of the 524 physicians who responded, 34% were internists, 38% were family practitioners, and 29% were pediatricians. Of the 524 respondents, 335 (64%) had hospitalists available to them and 120 (23%) were required to use hospitalists for all admissions. Physicians perceived hospitalists as increasing (41%) or not changing (44%) the overall quality of care and perceived their practice style differences as neutral or beneficial. Twenty-eight percent of primary care physicians believed that the quality of the physician-patient relationship decreased; 69% reported that hospitalists did not affect their income; 53% believed that hospitalists decreased their workload; and 50% believed that hospitalists increased practice satisfaction. In a multivariate model predicting physician perceptions, internists, physicians who attributed loss of income to hospitalists, and physicians in mandatory hospitalist systems viewed hospitalists less favorably.

Conclusions  Practicing primary care physicians have generally favorable perceptions of hospitalists' effect on patients and on their own practice satisfaction, especially in voluntary hospitalist systems that decrease the workload of primary care physicians and do not threaten their income. Primary care physicians, particularly internists, are less accepting of mandatory hospitalist systems.

Box Section Ref ID

THE NUMBER of hospitalists—physicians employed by managed care organizations and hospitals to care for hospitalized patients—is increasing in the United States, and their use may be redefining the role of the primary care physician.¹ Hospitalists threaten to transform primary care physicians into practitioners of an office-based specialty largely focused on the care of patients who are not acutely ill. Hospitalists may also be substantially changing the practice of medicine. Critics believe that hospitalist systems disrupt key elements of patient care, citing the importance of the continuity of the physician-patient relationship when patients are most ill.^2,3 They express concern that the quality of care of hospitalized patients will suffer, and posit a decrease in overall quality of care if important information is lost in the change of physicians upon admission or discharge. Opponents of hospitalist systems also note the potential for cost shifting if patients discharged early from the hospital by cost-conscious hospitalists require extra attention and additional costly testing from the primary care physician. However, proponents of hospitalist systems cite studies that show that hospitalists' care results in increased efficiency and improved quality of inpatient care.^4-6 They argue that efficiency and quality will also improve in the outpatient setting as primary care physicians concentrate their efforts on ambulatory care.^7-9

In general, primary care physicians have not participated in the decision of health care organizations to deploy hospitalists. Although they were initially introduced as voluntary programs, many health plans and hospitals are attempting to create mandatory hospitalist programs. These are frequently opposed by professional organizations representing primary care physicians.^10-12 However, little is known about how practicing primary care physicians actually perceive hospitalists.

We surveyed primary care physicians in California to investigate their experience with hospitalist systems. We examined their perceptions of how hospitalists affect their practices, their relationships with their patients, and overall patient care.

In 1998, we mailed self-administered questionnaires to primary care physicians practicing in the 13 largest urban counties in California. The survey was conducted as part of a larger study examining the impact of mandatory Medicaid managed care in these areas.^13,14 The counties studied contain 79% of California's practicing primary care physicians and 78% of the state's population. The physicians were identified from the American Medical Association's Physician Masterfile. They were eligible for inclusion if they were active in patient care, were not currently in training, and reported a primary specialty of family medicine, general practice, general internal medicine, or general pediatrics.

The study physicians were initially selected and surveyed in 1996 using a probability sample stratified by county and by physician race/ethnicity with an oversampling of nonwhite physicians. Details of the sample are given in a previous report.¹³ The sampling fractions varied by county depending on the number of primary care physicians in the county, but a minimum of 100 and a maximum of 200 physicians were selected in each county. Physicians we identified as having retired, moved out of state or died or who indicated a main practice specialty other than primary care were considered ineligible and were excluded from the survey. In the original 1996 sample, there were 1069 eligible primary care physicians with known addresses in the 13 counties, and we obtained completed responses from 759 physicians (71%). In 1998, we resurveyed all physicians who responded to the 1996 survey. Fifty-one physicians became ineligible between the 2 surveys. Data used in the present analysis are from the 708 physicians who were eligible for follow-up in 1998.

The questionnaire defined hospitalists as "physicians who work under contract with managed care organizations or hospitals and who are available to take responsibility for the care of your patients when they are hospitalized." Physicians were asked if hospitalists were available to them, and if so, whether they were required to turn over the care of some or all of their patients to the hospitalists. Those physicians who had experience with hospitalists then used a 5-point Likert-type scale (ranging from greatly increased to greatly decreased) to rate the effect of hospitalists on several aspects of patient care (overall quality of care, quality of inpatient care, patient satisfaction, amount of care needed from the primary care physician after discharge, number of tests and procedures needed after discharge, and coordination of care). They were also asked if, compared with their own inpatient practice, hospitalists reduced the patients' length of stay, consulted fewer specialists about hospitalized patients, and/or ordered fewer tests or procedures. If so, they were asked whether these practice differences compromised, improved, or had no impact on the quality of care. In addition, using a 4-point Likert-type scale (ranging from strongly agree to strongly disagree), the physicians indicated the degree to which they agreed with statements on the effect of hospitalists on the physician-patient relationship.

A second component of the questionnaire investigated the impact of hospitalists on the practices of primary care physicians. Respondents were asked to indicate the number of their patients hospitalized in a typical month, the proportion of their hospitalized patients cared for by hospitalists, and how important (on a 3-point Likert-type scale of very, somewhat, and not at all) direct involvement in the care of hospitalized patients was for their career satisfaction. Physicians who had experience with hospitalists were asked to rate on a 5-point Likert-type scale (ranging from greatly increased to greatly decreased) the impact of hospitalists on their income, workload, and practice satisfaction. They used the same scale to rate the impact of hospitalists on the quality of their relationship with their patients. The survey also included items about the characteristics of the physicians' practices, the numbers and types of patients treated, demographic characteristics of the physicians, and net income from the practice.

To simplify the presentation, responses were collapsed into directional categories of agreement or change. Statistical analyses were performed using χ² tests for bivariate comparisons of categorical data. We performed multivariate logistic regression to examine significant bivariate predictors of primary care physicians' (positive or negative) perceptions of hospitalists. Analyses generalizing to all physicians in the 13 counties studied were weighted by the inverse of the product of the sampling fraction and the participation rate to account for oversampling of nonwhite physicians and differences in response rates at the county level. All other analyses are presented with unweighted data.

Complete data were available for 524 (74%) of 708 eligible physicians. Internists, family physicians, and pediatricians each constituted approximately one third of the sample (Table 1). As expected with our sampling design, African Americans, Latinos, and Asians were overrepresented compared with their numbers in the physician population as a whole. Thirty-five percent of the respondents were solo practitioners, and 20% were employed in group/staff-model health maintenance organizations (HMOs). Approximately three quarters of primary care physicians said that inpatient medicine was either somewhat or very important to their career satisfaction.

Three hundred thirty-five primary care physicians (64%) reported having hospitalists available to them (Table 2). When weighted to adjust for our sampling design, we estimate that 62% of primary care physicians in the counties studied had hospitalists available. Availability did not differ by physician specialty, type of practice, race, or sex. Physicians who reported that inpatient medicine was very important to their career satisfaction were less likely to have hospitalists available. One hundred twenty primary care physicians were required to use hospitalists for all admissions (Table 3); this represents 36% of those who had hospitalists available and 23% of all primary care physicians. The weighted estimate for mandatory hospitalist use among primary care physicians is 22%. Primary care physicians who were employed in a group/staff-model HMO were more than twice as likely to be required to use hospitalists than those in a small office-based practice. Primary care physicians who had more than 75% of patients covered by capitated HMO commercial or private insurance were also more likely to be required to use hospitalists compared with those with fewer capitated patients. Internists were required to use hospitalists more often than the other primary care specialists: 53% of internists reported being required to use hospitalists compared with 32% of family physicians (P<.001) and 20% of pediatricians (P<.001). Of those primary care physicians who had hospitalists available to them and were not required to use them, 23% never used them and 22% always used them. Over half of the physicians voluntarily used hospitalists for fewer than 31% of their admissions.

Most primary care physicians with experience with hospitalists believed that hospitalists increased (41%) or did not affect (44%) the overall quality of patient care (Figure 1). Similar results were obtained for the quality of inpatient care. While many physicians (41%) thought that patient satisfaction was unchanged with the use of hospitalists, 23% felt that it was increased and 36% felt that it was decreased. Most primary care physicians (63%) perceived no change in the amount of care required by patients after discharge by hospitalists, although 46% perceived a drop in coordination of care between inpatient and outpatient settings under the hospitalist system.

When asked to compare the inpatient practice style of hospitalists with their own, many primary care physicians who had experience with hospitalists did not perceive a difference (Figure 2). The major difference noted was that more than half of the primary care physicians believed that hospitalists decreased the length of a patient's stay. The differences in practice style between hospitalists and primary care physicians were perceived by most primary care physicians to be either neutral or beneficial to the quality of patient care.

The majority of primary care physicians with experience with hospitalists reported that hospitalists had no effect on their income (69%), decreased their workload (53%), and increased their practice satisfaction (50%) (Figure 3). While most primary care physicians reported that hospitalists had not changed the quality of their relationships with patients, 28% reported that the quality of their relationships with patients had decreased.

The majority of primary care physicians somewhat or strongly agreed with the statement that the "hospitalist system upsets my patients because they would prefer that I take care of them," and that the hospitalist system "undermines my patient's confidence in me" (63% and 69%, respectively). Forty-one percent agreed that the hospitalist system "allows me to participate in the major decisions about the management of my patients while hospitalized."

There were considerable differences in the perceptions of hospitalists according to physician specialty (Figure 4). Internists had more consistently negative views of hospitalists than either family physicians or pediatricians: for example, 51% of internists perceived hospitalists as decreasing patient satisfaction, compared with 35% of family physicians (P = .01) and 20% of pediatricians (P<.001). Pediatricians tended to report more favorable perceptions of hospitalists than family physicians or internists: for example, 53% of pediatricians perceived hospitalists as increasing the overall quality of patient care, compared with 44% of family physicians (P = .20) and 29% of internists (P<.001). Similar differences among the specialties were seen in response to questions on the effect of hospitalists on the coordination of care and the amount of care needed by patients after discharge (data not shown).

To determine whether physician specialty was independently predictive of perceptions of hospitalists, we performed a multivariate analysis controlling for physician age, sex, race/ethnicity, volume of hospitalized patients, hospitalist-associated changes in income and workload, self-rating of the importance of inpatient medicine to career satisfaction, and whether the hospitalist system was mandatory or voluntary (Table 4). After controlling for these factors, internists continued to have significantly more negative views of hospitalists' effect on quality of care (odds ratio [OR], 5.0; 95% confidence interval [CI], 1.6-5.6) and patient satisfaction (OR, 3.2; 95% CI, 1.5-6.9) than pediatricians. Family physicians tended to be more positive than internists toward hospitalists (OR, 2.1; 95% CI, 0.9-5.1), but the difference was not statistically significant. Those primary care physicians who attributed a loss of income to hospitalist systems (OR, 0.2; 95% CI, 0.1-0.4) were also significantly less likely to perceive a beneficial or neutral effect of hospitalists on quality of care or patient satisfaction. Primary care physicians were more likely to report that hospitalists decreased their own practice satisfaction if they considered the practice of inpatient medicine important to their career satisfaction. They also reported a decrease in their own practice satisfaction if their income decreased or if their workload increased as a result of hospitalists, or if they worked in a mandatory hospitalist system. Asians were more likely than members of other racial/ethnic groups to report that hospitalists increased their practice satisfaction. As physicians working in group/staff-model HMOs may be more used to working with hospitalists, we repeated the multivariate analyses excluding those physicians. For physicians working in office practices, mandatory hospitalist systems were strongly associated with negative perceptions of the effect of hospitalists on quality of care, patient satisfaction, and physician practice satisfaction.

As of 1998, primary care physicians in California's largest urban counties had extensive experience with hospitalists. More than 62% used hospitalists and 22% (weighted estimate) were required to admit all their patients to hospitalists.

Hospitalists are generally well accepted by primary care physicians. Half of our respondents indicated that the use of hospitalists actually increased their practice satisfaction, while only 17% reported a decrease in practice satisfaction. This reported increase in practice satisfaction and the largely favorable perceptions of hospitalists have occurred mainly in voluntary systems, in which hospitalists typically decrease the primary care physicians' workload without affecting income. Mandatory hospitalist systems were rated less favorably by respondents. An increase in the mandatory model might be expected to be associated with decreased primary care physician practice satisfaction. In voluntary systems, only 22% of primary care physicians with hospitalists available used them for all their patients, which suggests that primary care physicians value the option to care for some of their hospitalized patients. The strong association between loss of income due to hospitalists and negative perceptions of hospitalists is not surprising. About a quarter of primary care physicians attributed income loss to hospitalists. Cost shifting from the inpatient to the outpatient setting and loss of inpatient fees are wellsprings of unfavorable perceptions of hospitalists among primary care physicians.

Hospitalists were perceived as improving or having a neutral effect on the quality of care of hospitalized patients and on the overall quality of patient care. Primary care physicians are well positioned to observe the effects of hospitalist systems on patients in their practices. Our results may reassure critics that, at least in their current incarnation, hospitalist systems do not harm patients. Even when asked to compare hospitalists with themselves, few primary care physicians felt that hospitalist-associated differences in practice styles—discharging patients earlier or ordering fewer tests and consultations—harmed the quality of patient care.

Nearly a third of primary care physicians reported a decrease in the quality of the physician-patient relationship. This was anticipated by some critics of hospitalists, who decried the loss of increased intimacy and trust often established between physician and patient during an acute illness.¹⁵ Some proponents of hospitalist systems believe that physicians can still make "social calls" and otherwise participate in the care of their hospitalized patients, thereby mitigating the effect of their absence during a critical time.¹⁶ However, our data suggest that most primary care physicians feel excluded from major decision making about the treatment of their hospitalized patients. This decrease in the quality of the physician-patient relationship may be the most significant detrimental effect of hospitalists on primary care physicians—and on patients.

A decrease in patient satisfaction was noted by more than a third of primary care physicians. As we did not survey patients, it is difficult to know how to interpret this potentially worrisome finding. Patient satisfaction with hospitalist systems should be studied from the perspective of the patient, as should the widely held belief of primary care physicians that patients prefer their primary physician to treat them while hospitalized.

Even after controlling for other explanatory variables, internists tended to be less likely to perceive hospitalists as improving the quality of care and more likely to believe that patient satisfaction and coordination of care are decreased. One possible explanation is that internists, who still train predominantly in hospital settings, are more invested than family physicians or pediatricians in caring for their hospitalized patients. Alternatively, the hospitalist system, with its obligatory discontinuity, may be less successful at caring for the patients with multiple medical problems often encountered in the practice of general internal medicine. For these patients, the absence of continuity may impair inpatient decision making, discharge planning, and overall care.

Our study is a survey of physician beliefs in California in 1998. It may not reflect the experience and attitudes of primary care physicians in other states. Physician attitudes may change as hospitalist systems evolve. While our results are dependent on physicians' self-reports, we believe that primary care physicians are well situated to report on issues that impact their own practices. Moreover, the results appear consistent with other studies that have examined outcomes and practice styles in hospitalist systems.^4,5

Hospitalist systems represent another step in the process of increasing specialization and fragmentation of medical care.¹⁷ Similar to disease carve-out programs and other service changes driven by the financial interests of hospitals and managed care organizations, hospitalists are increasingly common despite relatively little knowledge of their impact on care. Whether hospitalist systems will benefit patients and primary care physicians as they continue to evolve toward mandatory systems is unclear.¹⁸ Our data suggest that they currently do not harm the majority of either patients or primary care physicians and may benefit a substantial proportion of both.

Most European nations long ago delegated inpatient care to hospital-based physicians, with primary care physicians (mainly general practitioners) performing only ambulatory care. Many primary care physicians in the United States seem ready to accept this European model of patient care, despite some concerns about the effect on the patient-physician relationship and coordination of care. However, general internists in the United States appear more reluctant than family physicians or pediatricians to relinquish their traditional role in inpatient care and to accept redefinition exclusively as ambulatory care practitioners. It remains to be seen whether patients are well or ill served by their reticence. What seems clearer is that primary care physicians are less accepting of mandatory systems that force them to hand over their patients at the hospital door.

Accepted for publication May 19, 2000.

This study was supported by grant R03HS09557-01 from the Agency for Health Care Policy and Research, Rockville, Md. Drs Bindman and Grumbach were Robert Wood Johnson Foundation Generalist Physician Faculty Scholars during the performance of this project.

We thank Dennis Keane, MPH, Deborah Jaffe, BA, and Melissa Ehman, MPH, for their invaluable research assistance and Amy Markowitz and Nora Goldschlager, MD, for their editorial assistance and helpful comments.

Corresponding author: Alicia Fernandez, MD, San Francisco General Hospital, Bldg 90/Ward 95, 1001 Potrero Ave, San Francisco, CA 94110 (e-mail: aliciaf@itsa.ucsf.edu).