Recent changes in the financing and delivery of medical care have led researchers, policy makers, and clinicians to reassess the size and composition of the physician workforce R1563BABJDJCH–R1563BABJDFEI. Because estimates of physician workforce requirements are influenced by disease prevalence, health market economics, population growth, and other variables, they are subject to frequent change R1563BABEDDIC–R1563BABHBIEI. Physician practice patterns factor in assessments of workforce needs. Public health planners are concerned not only with the number of physicians, but with the distribution of their work effort, particularly the extent to which they serve vulnerable patient groups R1563BABBHBGD, R1563BABDAABB, R1563BABHAAFG.
In psychiatry, professional practice patterns have been shown to vary with psychiatrist age R1563BABCCGGG, gender R1563BABFDDCH, R1563BABDIIBB, race R1563BABFICJF, and professional work setting R1563BABIEGCG. There are also indications that the country in which a psychiatrist receives his or her medical school training influences the subsequent choice of practice setting R1563BABEFJAF. However, we are not aware of any published studies that have compared the clinical characteristics of patients treated by international medical graduate (IMG) psychiatrists and patients treated by U.S. medical graduate (USMG) psychiatrists.
Concerns over the increasing costs of graduate medical education R1563BABEAHAJ, secular increases in the number of IMGs entering the psychiatric workforce R1563BABCJDGC, and expansion of supply-side and capitated health care financing have fueled calls to restrict the number of IMGs. In light of these developments, we sought to examine the clinical roles of IMG psychiatrists in the delivery of psychiatric care in the United States.
Data were drawn from the 1996 National Survey of Psychiatric Practice. A detailed description of survey methods has been presented elsewhere and is summarized here R1563BABFICJF. When the survey was conducted, between May and August of 1996, it was estimated that 85% of all practicing board-eligible psychiatrists in the United States were members of APA. A systematic sample of 1,481 active APA members was selected from the population of active APA members (N=28,076). African Americans, Hispanics, and psychiatrists under 40 years of age were oversampled.
The overall response rate was 70.5% (970 of 1,375). The basic demographic characteristics, training profile, and geographical distribution of the respondents did not significantly differ from those of the nonrespondents R1563BABFICJF.
Designation of each respondent as a USMG or IMG was based on self-reported country of medical school training. Data on medical school training, board certification, and demographic characteristics of the respondents were obtained from the APA membership database. All other variables were collected from the National Survey of Psychiatric Practice.
The survey asked the psychiatrists to estimate the percentage of time they spent in different work settings. The work settings were classified as private (solo office practice, group office practice, private general hospital, private psychiatric hospital, health maintenance organization, and private clinic/outpatient facility), public (public general hospital, public psychiatric hospital, and public clinic/outpatient facility), or other (nursing home, correctional facility, and residual).
The survey also queried psychiatrists on several aspects of their workloads. This included the number of patients seen in a typical week, the total number of hours per week, and the number of direct patient care hours in a typical work week. For each respondent, we calculated the percentage of hours in direct patient care by dividing the number of hours in patient care by the total number of hours worked.
The survey also collected aggregate information concerning the demographic and clinical characteristics of the psychiatrists’ patients and the financial arrangements between the psychiatrists and their patients.
Sampling weights were developed by APA, as previously described by Zarin and colleagues R1563BABFICJF, to compensate for the oversampling of younger and minority psychiatrists. The percentages provided in this report are based on weighted estimates. A statistical adjustment developed by Potthoff and co-workers R1563BABCHBIC that tends to overcompensate for stratification artifacts was used to calculate standard errors of the survey estimates.
Many of the response variables were recorded as proportions. To circumvent the problem of heteroscedasticity (unequal variance), these values were arc sin transformed before the statistical analyses were carried out R1563BABHEJDA.
Chi-square statistics were used to test for differences in the distribution of categorical variables, and t tests were used for continuous variables. We used multiple linear regression equations to examine the relationship of clinical and practice characteristics to country of medical school training, adjusting for the effects of individual psychiatrist characteristics, i.e., age, sex, race, board certification status, and work setting. Comparisons were considered statistically significant at the 0.01 (alpha=0.01, two-tailed) level.
+
Demographic Characteristics
According to the National Survey of Psychiatric Practice, an estimated 22.3% of active psychiatrists in the United States are IMGs. The IMG psychiatrists surveyed were significantly older than the USMGs and significantly more likely to be female and to be members of a racial minority, but they were less likely to be board certified (T1).
The survey findings indicated that outpatient care was the primary clinical work setting of both the USMGs and IMGs. However, as a proportion of their clinical effort, the IMGs spent almost twice as much time as USMGs working in inpatient settings. Both the IMGs and USMGs worked primarily in the private sector. However, the IMGs devoted a larger share of their work effort to public work settings (T1).
+
Workload Characteristics
The mean number of total hours worked by both groups of psychiatrists exceeded 40 hours per week, with the IMGs working an average of 4.2 more hours per week than the USMGs (T2). This difference remained significant after the confounding effects of psychiatrist age, sex, race, and board certification status were controlled for.
The IMGs devoted an average of 4.6 more hours to direct patient care during a typical work week than did the USMGs. Not surprisingly, the IMGs also reported seeing a significantly larger number of unduplicated patients per week than did the USMGs (T2). The amount of time devoted to patient care as a proportion of the total worked hours was also significantly higher for IMGs than USMGs, although this difference was not significant after the other psychiatrist covariates were adjusted for. All differences in workload characteristics were no longer significant after adjustment for work setting.
+
Patient Characteristics
Data on the survey respondents’ patients and on the sources of payment are shown in T3. White patients constituted a majority of both study groups’ caseloads. However, the IMGs reported a caseload mix that included on average a significantly larger percentage of Hispanic, African American, and other nonwhite patients than the caseloads of their USMG counterparts. The racial disparity in caseload composition remained significant after we controlled for race and other characteristics of the psychiatrist and for work setting.
Although a majority of patients in the caseloads of both groups of psychiatrists were under age 65, the IMGs reported caring for a proportionately larger number of patients over 65 years of age. These differences remained significant after we controlled for psychiatrist age, race, gender, and board certification but not when we controlled for work setting.
Mood disorders accounted for the largest proportion of patients treated by the IMGs and USMGs. The two groups treated similar proportions of child and adolescent disorders, alcohol-related disorders, and other substance use disorders. As compared with the USMGs, however, nearly twice the percentage of the patients seen by IMGs had a primary diagnosis of schizophrenia or other psychotic disorder. In contrast, the proportions of patients with anxiety disorders and personality disorders were larger for the USMGs than for the IMGs. All these differences remained significant after work setting and other factors were controlled for.
For both groups of psychiatrists, commercial insurance was the most common source of payment. Whereas Medicare and Medicaid each accounted for a significantly greater proportion of the income of IMG than USMG psychiatrists, the reverse was true of self-payment (T3).
The practice patterns of IMG and USMG psychiatrists differ in several important respects. IMG and USMG psychiatrists tend to work in different treatment settings, treat different groups of patients, and receive income from different payment sources. Most of these differences are not attributable to the personal demographic characteristics of the two physician groups or their work settings.
+
Demographic Characteristics of Psychiatrists
Two-thirds of IMG psychiatrists are women. This is in striking contrast to the overall population of IMGs across specialties, three-quarters of whom are men R1563BABHHJGH. Although this finding has been previously documented R1563BABFDDCH, R1563BABCJDGC, the reasons for this gender disparity remain unknown and require closer study. As previously documented, IMGs are also more ethnically heterogeneous, reflecting the diversity of their countries of origin R1563BABCJDGC.
The majority of IMGs and USMGs are board certified. However, a lower percentage of IMGs have certification than USMGs. Although board certification has increased in both groups in recent years R1563BABCJDGC the gap between the two groups may actually be increasing. We are not aware of any study that has examined the basis of these differences. Possible explanations include 1) a lower rate of pursuit of board certification by IMGs, 2) a lower rate of repeat attempts following failure of the first attempt, and 3) a lower pass rate due to language difficulty and poorer training. This is an area that deserves further study, particularly as changes in health care systems have placed greater emphasis on board certification.
+
Service Characteristics
According to the National Survey of Psychiatric Practice, the average IMG psychiatrist tends to work more hours each week and sees more patients than the average USMG psychiatrist. Like the respondents in an earlier survey R1563BABEFJAF, IMG psychiatrists were found to devote a greater percentage of their time than USMG psychiatrists to the public sector and to inpatient settings. The reasons for these disparities are also unclear. Weintraub and Book R1563BABGGCCA have suggested that one reason IMGs are more likely to work in state hospitals is because that is where they do their psychiatric residency training. Prior research R1563BABBBBDG, R1563BABDHDBG suggests that psychiatrists in public settings have relatively few opportunities to use their full range of skills, which in turn leads to lower professional satisfaction and to increased physician turnover. Women have also been found to be more likely to work in public settings R1563BABFDDCHR1563BABDIIBB. It is possible that IMGs and other minority psychiatrists have restricted opportunities regarding their choice of work setting following residency training.
Like previous investigators R1563BABBEDFI, we found that IMGs tend to treat significantly more African American and Hispanic patients than USMGs. Even after we adjusted for the race of the psychiatrist, IMGs were found to treat a larger number of ethnic minorities. This is important because, although all ethnic groups (including whites) tend to underutilize mental health services R1563BABDCADE, R1563BABJCAHF, the gap between need and use is greatest for members of ethnic and minority groups R1563BABCDHAI, R1563BABHGHBJ. While underutilization is more marked among the ethnic minority poor, it is still discernible after socioeconomic factors and insurance coverage are controlled for R1563BABHAAFG, R1563BABIBEHD.
Individuals over 65 years of age are also particularly underserved with respect to mental health services R1563BABHJBJE. IMGs spend on average 35% more time working with the elderly than USMGs. As the population ages, the need for psychiatrists with experience treating older patients is certain to grow. IMGs, with a larger presence in the public sector and inpatient settings than USMGs, are likely to deliver an increasingly large share of the psychiatric care to this age group.
The data from the National Survey of Psychiatric Practice do not include gross or net physician income. However, we found that IMGs are more likely than USMGs to obtain their income from Medicaid and Medicare and less likely to receive patient self-payment, a pattern similar to that of early-career psychiatrists R1563BABCCGGG. Previous research also indicates that Medicaid is an important source of income for IMG psychiatrists R1563BABFACCA. The combination of work setting, patient diagnoses, and payment sources suggests that IMGs are currently taking care of a proportionately higher number of the most severely ill patients.
+
Health Policy Implications
Having practice patterns that differ substantially from those of USMGs, IMGs contribute in a unique way to the delivery of mental health care in the United States. Each IMG sees a large number of patients and provides a disproportionate share of the care in the public sector and inpatient psychiatric settings. Many of the patients treated by IMGs are chronically ill and economically disadvantaged. Others are members of underserved populations by virtue of their ethnicity or age. Individuals from all of these groups tend to have fragile support systems and may have relatively complex mental health care needs. They may also have less propensity to seek mental health services R1563BABHAAFG.
Recognizing the barriers to care of vulnerable populations, the federal government has established different programs to facilitate their access to care. Medicare and Medicaid decrease financial barriers to the elderly and low-income individuals. Other programs encourage physicians to work in underserved areas, either through selective admission to medical school or through loan forgiveness programs R1563BABGFDFG, R1563BABFEEDA.
Increasing concerns with a perceived oversupply of physicians have stimulated proposals to limit the number of residency slots that will be open to IMGs. Those policy proposals are based, in part, on previous research showing that USMGs and IMGs have similar patterns of geographic distribution after residency R1563BABBHBGD. Thus, it has been assumed that the two groups play similar roles in the delivery of care. However, data from the National Survey of Psychiatric Practice reveal that the patient populations differ in several key demographic and clinical variables, even after adjustment for work setting. Our data are consistent with the view that further restrictions on the entry of IMGs may adversely affect the availability of psychiatrists to serve certain minorities or currently underserved groups.
From the perspective of the physician workforce, IMGs devote a higher percentage of their time to the public sector, which has traditionally had difficulty in retaining psychiatrists R1563BABEAJGD. A decrease in the number of IMGs may further affect retention of mental health professionals and the availability of services provided in this sector.
The National Survey of Psychiatric Practice is limited in several ways. The data are self-reported, so their validity depends on the extent to which psychiatrists can accurately report their practice patterns. A second and related limitation is the absence of information on specific visits or patients. A third limitation is related to the representativeness of the sample. While the sample represents APA members, it does not necessarily represent non-APA psychiatrists, who as a group are less likely to be board eligible or board certified and are more likely to be IMGs R1563BABFICJF. However, these limitations should be considered in relation to the dearth of information available on the practice patterns of IMG psychiatrists.
Data from the National Survey of Psychiatric Practice indicate that IMG and USMG psychiatrists have very different practice patterns. IMGs play a distinct role in the delivery of mental health care in the United States that is focused on the treatment of publicly insured, socioeconomically disadvantaged groups and the most severely ill. We hope that these findings will enlarge the debate over access of IMGs to United States psychiatric residency training programs by focusing attention on the different public health functions performed by IMG and USMG psychiatrists. Further research is needed to examine whether the general findings of this study apply to other medical specialties.
Received Feb. 26, 1998; revision received July 20, 1998; accepted Aug. 25, 1998. From the Department of Psychiatry, Columbia University/New York State Psychiatric Institute; the Department of Psychiatry, Saint Vincent’s Hospital, New York; and the APA Office of Research. Address reprint requests to Dr. Blanco, New York State Psychiatric Institute, Box 69, 1051 Riverside Dr., New York, NY 10032; cb255@columbia.edu (e-mail). Supported in part by a grant from the John D. and Catherine T. MacArthur Foundation. The authors thank the APA members who participated in this study and the APA staff who worked on the project: Deborah A. Zarin, M.D. (principal investigator), Joyce C. West, M.P.P., Ana P. Suarez, M.P.H., and Brennan D. Peterson, B.S.