Hoff TH, Whitcomb WF, Williams K, Nelson JR, Cheesman RA. Characteristics and Work Experiences of Hospitalists in the United States. Arch Intern Med. 2001;161(6):851-858. doi:10.1001/archinte.161.6.851
Copyright 2001 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2001
Little is known about the personal characteristics, work-related attitudes, or professional experiences of hospitalists. In considering the future of hospital medicine as a viable career choice for physicians (primarily, internists), these issues should be examined in a systematic fashion. Learning more about hospitalists and their work can enhance dialogue about the advantages and shortcomings of such a career from the perspective of the individual physician.
A self-administered mail survey was sent to 820 hospitalists who are dues-paying members of the National Association of Inpatient Physicians and who spend 50% or more of their time doing clinical work, teaching, or research related to hospital medicine. Attitudes about topics such as job-related burnout and job satisfaction were tapped, as well as information about different professional and social experiences. The analyses were performed using descriptive statistics and analysis of variance techniques.
Analysis was based on 393 responses (48% response rate). Results show hospitalists to be a group of younger, mostly male, early-career individuals with high levels of job satisfaction and autonomy, low levels of burnout, and a long-term commitment to remaining in the role. Hospital medicine is a source of positive social and professional work experiences related to interactions with physician peers, patients and their families, and nonphysician hospital coworkers. Key components of hospitalists' jobs, practices, and workload are coalescing. However, certain developments, such as changing patterns of compensation and the enlisting of more general internists and women as hospitalists, merit further examination.
The results offer insight into the physicians who are becoming hospitalists, the jobs and settings in which they work, and how hospitalists experience their everyday work lives. Valuable baseline data are provided for assessment of attitudes, such as burnout, that should be examined regularly in this fledgling group. This study complements research looking at the performance-related outcomes of hospitalists, and it can be used by various stakeholders to better understand and assess the long-term potential of what is being proposed as a new career path.
THE FIELD of hospital medicine continues to grow. It is estimated that there are approximately 4000 physicians practicing as hospitalists in the United States. Yet, little is known about the individuals who are hospitalists, and even less is known about the sociological aspects of the role. The only systematic data on hospitalists come from a survey conducted in 1997, which provided descriptive information about the demographic, clinical, and financial aspects of hospitalist work.1 These data showed that most hospitalists were general internists, at early stages in their medical careers, men, in the role for only a short time, compensated by salary, and employed by medical groups or hospitals. Almost all were satisfied with their current positions, and more than three quarters intended to be hospitalists 3 years into the future. It is useful to update some of these descriptive findings based on our 1999 findings, as 2 years is a significant length of time, given the growth and brief existence of this fledgling specialty movement. For example, it is estimated that the number of hospitalists in the United States has doubled since 1997. More important, however, is gaining a sense for the first time of how hospitalists are experiencing their everyday roles. These experiences are reflected through key work-related attitudes such as burnout and job satisfaction, as well as perceptions related to the more social, collegial, and professional aspects of hospital medicine. None of these attitudes, except job satisfaction, have been explored in the literature. Of secondary interest related to how the field is evolving is how individuals newer to the hospitalist role compare with more-veteran hospitalists. What are the similarities between these 2 groups? What are some key differences? The present study explores these issues in detail.
An 87-item, self-administered survey was mailed during the fall of 1999 to 820 hospitalists who pay dues as members of the National Association of Inpatient Physicians (NAIP). The NAIP is the recognized professional association for hospitalists in the United States. Its mission is to be "the leading resource and voice for hospitalists and their patients" through dedication to "promoting excellence and efficiency in clinical care and advancing the practice of inpatient medicine."2 For this study, hospitalists are defined as physicians spending 50% or more of their time engaged in the practice of general hospital medicine, or in research and education related to general hospital medicine. This definition differs from that of the 1997 survey, in which hospitalists were defined as spending at least 25% of their time caring for the hospitalized patients of primary care physicians.1 However, the definition of a hospitalist used for the present study is almost identical to the recently adopted official NAIP definition.3 Thus, the individuals responding to the present survey are an accurate reflection of practitioners in the field of hospital medicine now as defined formally by its professional association. Differences between the 1997 and 1999 surveys in this regard do not necessarily render informal comparisons between the 2 sets of respondents problematic. Rather, the result is a more accurate comparison of hospitalists in 1997 and 1999 who were most committed to the field (as evidenced by their membership in its professional association) and who met the official definition of "hospitalist" put forth by the NAIP during that particular time.
The study protocol was approved by the University at Albany, State University of New York Institutional Review Board, during the spring of 1999. To maximize survey response rates, several strategies were used. First, 2 survey mailings were conducted 3 weeks apart. A postcard reminder was sent to all nonrespondents approximately 1 month after the second survey mailing. Accompanying the surveys were 2 separate cover letters. The first was from the NAIP, signed by the 1999 copresidents (W.F.W. and J.R.N.). The second was from the Department of Health Policy, Management, and Behavior, University at Albany (signed by T.H.H. and K.W.). Each letter outlined the purpose and importance of the study from practical and academic perspectives. No incentives of any kind were used as inducements to complete and return the survey. The survey was divided into 5 major sections. In the order in which they appeared on the survey, they focused on (1) current perceptions of professional and social experiences in the hospitalist role, (2) job-related burnout, (3) motivations and realizations related to being a hospitalist, (4) specific aspects of the particular hospitalist job, and (5) demographic and nonwork (family and other) information. Several well-established psychological measures were used to examine feelings of job satisfaction, burnout, autonomy, and motivations for becoming a hospitalist.4,5 An appendix (available from the authors) lists some of the variable measures used, their operational definitions, and specific survey items. Reliabilities (Cronbach α) for all multi-item measures used in the analysis ranged from 0.60 to 0.95, which is a generally acceptable range for measures used in the behavioral sciences.6
Several questions sought information on how hospitalists experienced the sociological aspects of their work and employment setting. For example, respondents were asked the extent to which they felt recognized by important stakeholder groups for their contributions, a part of supportive and close-knit work and peer groups, and able to achieve desired relationships with patients (appendix, available from the authors). Most measures used consisted of multi-item statements that were based on a Likert scale, with responses ranging from "strongly disagree" to "strongly agree." Questions about job and personal characteristics were largely open-ended or limited to 3 or 4 response categories. Emphasis in these sections of the survey was on understanding how hospitalists' jobs and practices were structured, their compensation systems, demographic backgrounds in terms of medical training and nonwork activities, and their intent to remain in the role.
The initial sampling frame consisted of 820 individuals. From these, 393 usable surveys were obtained, for a response rate of 48%. This response rate compares favorably with those of surveys of other physician groups, and is higher than the 39% response rate of bona fide hospitalists who responded in a timely fashion to the 1997 NAIP survey.1 Overall, 399 surveys were returned in the present survey. However, 6 respondents did not meet the criterion of 50% of time spent as a hospitalist, as indicated by their response to a specific question on the survey, and thus were excluded from the analysis. An analysis of nonrespondents by physician medical specialty, ethnicity, and sex revealed minimal differences (<5% for each variable) between survey respondents and the initial 820 individuals, indicating similarity in demographic characteristics.
Results were obtained using descriptive statistics and analysis of variance techniques with commercially available software (SPSS version 8.0; SPSS Inc, Chicago, Ill). There was no multicolinearity observed among the variables. Means, SDs, and percentages were calculated for all variables included in the study. Simple means tests and 1-way analysis of variance were also used, especially in comparing newer vs more-established hospitalists. t Tests were used to determine the statistical significance of any observed differences. For the 1-way analysis of variance, a modified Bonferroni post hoc comparison was used to control for experimental error. This applies a more stringent significance test to account for all mean comparisons occurring between 2 variables with multiple categories.7 When comparing percentages rather than means across the newer vs more-established hospitalists, the χ2 test was used to assess the statistical significance of observed differences.
Approximately three quarters of respondents were men trained in general internal medicine (Table 1). The mean age was 40 years. However, almost one third of the group were 35 years or younger, and fewer than 10% were 50 years or older. More than one quarter of hospitalists had completed their residencies within the past 5 years. Almost half had been in hospitalist roles for 2 years or less, compared with 15% who had been in the role for more than 5 years. Time spent in the present job was similar to an individual's total tenure in the hospitalist role, as most hospitalists had been in their current jobs for 2 years or less. Job stability is discerned, however, in the fact that respondents who had been hospitalists for more than 3 years also had been in their current job for a mean 4½ years. Respondents spent the bulk of their weekly work time in the hospitalist role. Forty percent spent 100% of their monthly work time (not including time when they were on call) in patient care as hospitalists, while another 22% spent between 90% and 100% of their weekly time in patient care. Thus, just less than three quarters of the group spent almost all of their work time engaged in patient care, as opposed to a combination of patient care, research, and education related to hospital medicine, as put forth by the NAIP in its definition of a hospitalist. The most strongly held motivation among respondents for becoming a hospitalist was the intellectual stimulation and enjoyment that came from practicing hospital medicine. Secondary motivations included the chance to be involved in an emerging specialty and being able to have a job with predictable hours and lifestyle. Respondents indicated that limited alternative employment opportunities were not a driving force in their decision to become a hospitalist.
Hospitals and medical groups employed more than two thirds of respondents (Table 2). Only about 1 in 10 hospitalists classified himself or herself as self-employed. Regardless of who the employer was, mean practice sizes were similar, except for hospitalist practices in managed care organizations, which were significantly larger than practices in any of the other settings (P<.05 for all comparisons between managed care organizations and other settings). The mean duration of practice establishment was slightly less than 4 years, regardless of the type of employer (hospital or medical group). In terms of workload and scheduling, most respondents worked in a system in which hospitalists performed daily patient rounds and were available off-hours by beeper paging. Shift scheduling, in which a hospitalist is on site around-the-clock, was less prevalent. Half of all respondents said that they spent more than 20 hours per week on call. Specific workloads for hospitalists were investigated by asking them how many new admissions and consultations they experienced per month. Four of 10 stated that they personally admitted or consulted on more than 60 new patients per month. For 2 of 10, this figure was 90 or more per month. Although neither of these workload aspects varied significantly by the major types of employer, self-employed hospitalists appeared to spend more hours per week on call and to admit more patients per month than did hospitalists who worked as employees in medical groups, hospitals, or managed care organizations (Table 3).
The mean annual compensation for the respondent group was approximately $141 000. However, only 20% of respondents said that they earned less than $120 000 per year as hospitalists. The biggest difference in compensation was seen between male and female respondents, with the former earning a mean $22 000 more per year ($146 400 vs $124 600; P<.001). In addition, hospitalists who were compensated through a fixed salary generally earned between $10 000 and $30 000 less per year than did individuals who were paid through other means, such as salary plus a performance incentive, capitation, or fee-for-service. Self-employed hospitalists earned significantly more annually than did individuals working as employees in hospitals, medical groups, or managed care organizations (P<.01 for all differences). These differences ranged from $28 000 to $34 000 per year. All of these compensation differences (men vs women, fixed salary vs other means of compensation, and self-employed vs employee) persisted even after controlling for several aspects of workload, such as number of new admissions and consults per month, hours spent on call per week, and time spent practicing hospital medicine per week. Thus, the compensation differences did not appear to be a function of differences in workload between the various groups. Notably, almost three quarters of respondents reported that none of their annual compensation was linked to financial incentives. Slightly more than 3% reported more than a quarter of their income was tied to such incentives. Furthermore, although those whose income was linked to financial incentives earned a mean $11 000 more per year than did others, this difference was not statistically significant (P= .06). These findings imply that financial incentives do not at present create large compensation differentials among hospitalist physicians.
A high degree of job satisfaction was endemic to survey respondents. More than half could be classified as "highly satisfied" and another third as "satisfied" (Table 4). Related to the prevalence of job satisfaction, the survey group had high perceived levels of clinical autonomy, widespread beliefs that their initial expectations for entering the field had already been met, and an absence of job-related burnout. Only 13% of respondents could be classified as burned out. However, an additional 25% of responding hospitalists indicated levels of physical, emotional, and mental exhaustion that placed them in the "at risk" category. Only 1 in 5 respondents thought that their initial expectations for the hospitalist role had not yet been met. Fewer than 3% indicated an unsatisfactory level of autonomy in their everyday work. Given the widespread prevalence of these positive work-related attitudes, it is no surprise that almost two thirds of respondents said that they would still be doing hospitalist work 7 years from now.
Most individuals also appeared to be having favorable experiences in the hospitalist role. More than half of all respondents believed that they worked within a hospitalist group containing high levels of camaraderie and solidarity, while 75% believed that their immediate work group (multidisciplinary group of nurses, social workers, and other individuals working on a particular ward or hospital floor) functioned effectively and cooperatively in providing care to patients. With respect to their patients, almost all respondents believed that they were able to provide them with complete care within the inpatient setting. Complete care was defined by statements such as "I feel I can treat the whole patient in my hospitalist job" and "I can address a broad range of problems when treating patients in my hospitalist job." Furthermore, 4 of 5 respondents felt properly recognized for their work by patients and the families of patients. This recognition involved "expressions of gratitude" and "getting credit" for the services provided during the course of their everyday work.
Some informal comparisons can be drawn between these data and those of the 1997 survey conducted by Lindenauer et al.1 Similarities include respondents' mean young age, early to mid medical career points, newness to the hospitalist role, high levels of satisfaction, and intent to remain in the field for at least 3 years into the future (Table 5). In addition, the mean compensation is similar between the 1997 and 1999 respondents, although the 1997 group's was slightly higher. This is despite the fact that the 1997 survey had a broader definition of hospitalist that included physicians spending less time (as little as 25%) practicing inpatient medicine. A comparison of the 2 sets of survey results shows several emerging differences among the 1999 respondents, including a growing differential between male and female hospitalist compensation, larger percentages of the field composed of general internists and female physicians, more individuals employed by hospitals, and a declining prevalence of financial incentives tied to compensation among the 1999 survey respondents. Direct comparisons between the 2 sets of survey results should be viewed with caution, given their different definitions of a hospitalist. However, these observations are useful because they focus our attention on potentially important areas to explore in more detail as this field evolves further.
Evolution of the hospitalist field was also examined within the 1999 survey by comparing characteristics of individuals newer to the role with those of more-established hospitalists (Table 6). Three years was selected as the cutoff point for defining new and more-established hospitalists for 2 reasons: (1) individuals in the role for less than 3 years arguably define the present and near-future expansion of the field of hospital medicine and (2) a 3-year cutoff yielded 2 relatively equivalent groups in terms of numbers of respondents, which facilitated performing means tests on various comparison variables. Perhaps the most meaningful difference observed is that hospitalists within the field for less than 3 years earned significantly less than did their counterparts with longer tenure (mean difference of $14 000; P<.001). This was despite the fact that the 2 groups, on average, were similar in numbers of hours worked, types of employment, workload, and hours per week on call. Although the shorter-tenured group had a larger percentage of female hospitalists and individuals compensated through fixed salary (2 variables associated with decreased compensation), controlling for both did not render insignificant the difference observed between the 2 tenure groups. One might also reasonably assume that more-established hospitalists might get paid more because they have been with their employer significantly longer (a mean of almost 3 years longer) than those in the field for 3 years or less. However, an examination of bivariate correlations revealed no association between positional tenure and compensation among respondents that could account for the observed difference.
Several other differences were noted between newer and more-established hospitalists. Eight percent more of the shorter-tenured group are women, although this difference is not statistically significant (P= .09). In addition, 9% more of the newcomers are general internists (P = .048). Regarding work-related attitudes, physicians in the hospitalist role for less than 3 years report mean lower levels of clinical autonomy in their work (P = .048), although high levels of job satisfaction and low levels of burnout exist among both groups of hospitalists. A smaller percentage of the shorter-tenured group also indicated that they would still be doing hospitalist work more than 7 years into the future (P = .03). However, despite this difference, almost half of each group indicated that they would still be hospitalists in 10 years. Thus, longevity in the field is a common desire of newer and more-established hospitalists.
The results presented herein illuminate key characteristics of hospitalists. They also provide the first systematic information, to our knowledge, about how these individuals experience the professional and social aspects of their work. In general, most hospitalists responding to the survey were at early stages in their medical careers, men, trained in general internal medicine, and relatively new to the field of hospital medicine. One third of the respondent group was 35 years or younger, which is slightly more than the 25% of general internists nationally who are 35 years or younger.8 Most respondents provided full-time patient care in their hospitalist role, conformed to a daily patient rounding schedule, practiced within small groups of hospitalists, worked as employees, were compensated through straight salary or salary with a performance incentive, and did not face pressures of financial incentives in their everyday work. The group, in general, was satisfied with their work and experienced high levels of autonomy and low levels of burnout.
The 13% of respondents who were classified as "burned out" based on their survey results is less than the 40% to 60% seen in fields such as critical care and emergency medicine.9- 11 However, the relative newness of many physicians to the hospitalist role suggests that a comparison with burnout in similar specialties should be reserved until the field has achieved more longevity, as a whole. Although a group with many newcomers, more than three quarters of hospitalists already believed that their original expectations for assuming the role had been met or exceeded. Their positive attitudes about work are consistent with the finding that almost two thirds of respondents intended to be practicing hospital medicine 7 years from the time of the survey. In addition, hospitalists also reported having favorable social and professional experiences in the field. Perhaps surprisingly, almost all respondents believed they could provide complete care to their patients. This meant having the chance to treat a broad range of problems that extended beyond the major acute problems presented by patients. Most respondents also perceived a high degree of solidarity with their local practice group and appropriate recognition by patients and their families for the work that they do.
There are limitations to the study. First, those surveyed were members of the NAIP and not the entire universe of practicing hospitalists in the United States. Undoubtedly, there are many physicians who could be classified as hospitalists who do not belong to the NAIP. Whether these individuals look at, think about, or experience their roles differently from NAIP members is a valid empirical question. Even within the NAIP, this survey included only those individuals who were paying members and who fit our working definition of a hospitalist. Members of the NAIP who did not complete the survey may also differ from those who did. Despite these limitations, the present survey includes those physicians who are committed to a career in hospital medicine. It is impossible to include all those considering themselves hospitalists, given that there is no systematic way to identify them for purposes of completing and returning a survey. Thus, this survey represents, to our knowledge, the most updated and systematic source of information on hospitalists and their work experiences.
A second limitation involves reliance in the survey on physician self-reporting for several job-related questions related to workload, scheduling, type of employer, and compensation. Ideally, it would be better to obtain this information directly from employers or employee records because individuals sometimes cannot accurately classify their work situations. Yet, when surveying a large number of people, this is virtually impossible. Also, some questions on the survey, because of the limited response categories, may not have captured all of the relevant variations. However, the survey questions were carefully developed through discussions between us, 2 of whom (W.F.W. and J.R.N.) are practicing hospitalists, and several other hospitalists. In addition, questions that measure various social and professional experiences, such as the level of occupational or work group solidarity, rely on hospitalists' perceptions instead of purely objective assessments. Yet, the psychological research that has developed many of these concepts offers strong evidence that it is individuals' perceptions of situations, more than anything else, that shape their attitudes and behaviors.12 Finally, the survey did not ask about the types of patients hospitalists serve (such as Medicaid vs private insurance), the manner in which the particular hospitalist practice is used by the hospital or medical group, or whether hospitalists are substituting for or complementing other physicians already caring for inpatients. This type of information could be important insofar as it is linked with how hospitalists experience their jobs, including their levels of satisfaction and burnout, as well as their perceptions of coworker support.
The findings of this study suggest that hospital medicine remains appealing to younger physicians, drawing particularly from those trained in general internal medicine.13 These individuals perceive the role to be one offering an opportunity to engage in work that stimulates them intellectually, as well as to obtain a well-paying job and a balanced lifestyle. Positive attitudes such as job satisfaction remain as strongly entrenched among the 1999 group as they were among the 1997 group. In addition, job burnout does not appear to be a problem, although there is a moderate percentage of hospitalists who could be classified as "at risk" for burnout. Additional research should identify the types of factors that could cause burnout. To our knowledge, this is the first examination of burnout among hospitalists, and it offers baseline data for future investigation. Finally, our findings did not support the suggestion that hospitalists are subject to negative work experiences because hospital medicine is affected by the need for greater organizational control over hospital admissions and lengths of stay. Why hospitalists appear to have such favorable relationships with coworkers, physician colleagues, and patients and their families, despite the inherent tension associated with a clinical field driven at least in part by managed care imperatives, merits further study.
Some of the differences observed between the 1997 and 1999 respondents invite closer scrutiny, although they cannot yet be called "trends." Comparisons between the 1997 and 1999 groups and between the newer and more-established respondents in 1999 suggest that hospital medicine may be attracting more general internists than in the past. As hospital medicine continues to evolve, will it become even more appealing to general internists? Attracting disproportionate numbers of general internists (as opposed to other primary care physicians or subspecialists) may limit the clinical diversity of the field and tend to stereotype it as the domain of one particular medical specialty. This might have implications for the kinds of work hospitalists perform, the types of patients cared for, and the manner in which best practices and quality clinical outcomes in the field are defined. In internal medicine, this could potentially increase tension and divisiveness within the specialty.14 In addition, the field appears to be attracting female physicians at a higher rate than in the past. However, the 26% representation of women in the surveyed group is less than the 34% of general internists nationally who are women.8 Thus, the increased presence of women in hospital medicine may simply reflect the fact that more general internists are tending to choose the field. In any event, the observed pay differential between male and female hospitalists in this study should be explored in further detail, especially because men and women appear to be working in similar types of jobs with similar workloads.
At least one other key difference in compensation should be examined closely. Individuals newer to the field in 1999 earned a mean of almost $14 000 less than did their more-established counterparts. This pay difference between shorter- and longer-tenured individuals does not appear to be the result of empirical differences alone, such as the longer-tenured group's having been in their current jobs longer and earning more because of work experience or seniority, being more frequently on call, working for different types of employers, practicing longer hours, or comprising higher percentages of male hospitalists, who earn more. It may be a product of larger economic forces. For example, as various roles develop within competitive industries, such as health care, there can be a decrease in individual compensation as the number of positions available stabilizes or shrinks, concomitant with an increase in the number of individuals available to fill these positions. Differences in compensation method (straight salary vs salary plus performance incentives) may also be relevant. Compared with other specialties, the mean annual compensation for the 1999 survey respondents is $6000 less than that for general internal medicine, $9000 more than general and family practice, $21 000 more than pediatrics, and more than $50 000 less than emergency medicine.8
This study illuminates the characteristics and experiences of the group of physicians called "hospitalists." Its unique examination of their experiences, particularly those that reflect the sociological aspects of their everyday work, provides a necessary complement to recent studies15,16 looking at performance-related outcomes of hospitalists in areas such as productivity, quality of care, and patient satisfaction. Rather than examining performance, this study has focused on the people who practice hospital medicine and their work lives. In future research, however, this type of information should be integrated with information on outcomes to provide a complete picture of the hospitalist's work experience, including the individuals doing the work, the organizations that employ them, and the patients served. It could also be combined with information on how other physicians view and act toward hospitalists, because relationships with nonhospitalist physician colleagues remain an important variable speaking to the long-term role of these individuals in the medical workplace. Hospitalists' personal experiences could also be compared and contrasted with patient experiences. For example, although it is valuable to know that hospitalists feel recognized for their work by patients and their families, it would be useful to ask how patients see their relationships with the hospitalists who serve them. For example, do they view these relationships as fundamentally different from relationships they may have with their primary care physician? A key issue that needs further study is whether the hospitalist disrupts, in any way, existing relationships between patients and their primary care physicians.
The information presented in this study provides baseline data that can be compared with the findings of future surveys, which should be performed to capture formally the early history of the hospitalist movement. These results can be used by various stakeholders to better understand and assess the long-term potential of what is proposed as a new career path in medicine.
Accepted for publication October 18, 2000.
This research was supported in part by the National Association of Inpatient Physicians (NAIP), Philadelphia, Pa, and the University at Albany, State University of New York.
We thank the staff at NAIP headquarters, without whose assistance we would have not been able to conduct the survey. We also thank Peter Lindenauer, MD, of Baystate Medical Center, Springfield, Mass, who provided valuable feedback and suggestions regarding aspects of survey design and content.
Corresponding author and reprints: Timothy J. Hoff, PhD, Department of Health Policy, Management, and Behavior, University at Albany, State University of New York, 1 University Pl, Rensselaer, NY 12144 (e-mail: email@example.com).