Recently, a national survey among psychiatry clerkship directors in the United States revealed extensive concern about the effects of managed care on the quality of psychiatric clerkships for medical students (1). Also among their concerns was the fear that recruitment into the field will ultimately suffer. Given that recruitment into psychiatry is already at an all-time low (2) and that some projections indicate the field will fall short of the demand for psychiatrists for years to come (3), an examination of this issue is of great interest to those involved in psychiatric education. This study investigates whether the psychiatry clerkship experience at the University of Arkansas for Medical Sciences (UAMS) was related to choosing psychiatry as a specialty.
Although the choice of a medical specialty is influenced by a number of factors (4,5), the literature also suggests that the quality of the psychiatry clerkship may be the single most important modifiable medical-school influence on recruitment into psychiatry (2,6—10). A number of studies have examined the influence of the medical-student clerkship length, timing, and site on academic performance (11—13). Most of these studies to date have found that timing and length, but not site, of the clerkship were important factors on academic outcomes. Specifically, those who completed longer clerkships and those who rotated later in the academic year performed significantly better on standard examinations than students who took the clerkship earlier in the year or for shorter periods (12). Few studies have investigated whether these factors might be related to students' attitudes toward specializing in psychiatry.
The present study was undertaken in an attempt to understand more about the psychiatric clerkship experience and the importance, if any, of the timing of the psychiatric clerkship (early vs. late in the academic year) and the clerkship site on third-year medical students' interest in specializing in psychiatry and their eventually choosing to specialize in psychiatry. The hypothesis was that students who took the clerkship earlier in the year or later in the year, as opposed to mid-year, would be more likely to choose psychiatry as their specialty. We also wanted to investigate whether particular rotation sites (e.g., the type of service—such as the emergency room (ER), consultation/liaison, adult outpatient/inpatient, and child psychiatry) were associated with medical students' attitudes toward specializing in psychiatry. Of the studies that have examined the importance of rotation site or type of service, the results have been highly inconsistent. One study reported that inpatient experiences were highly favored by students during the clerkship (9), whereas other studies have reported the opposite finding (14,15).
From July 1992 to June 1997, a total of 658 third-year medical students at UAMS rotated through the junior psychiatric clerkship and also went on to graduate from UAMS. The number of students graduating from UAMS was similar for each year included in this study: In 1994, n=124; in 1995, n=135; in 1996, n=140; in 1997, n=123; and in 1998, n=136. There were eight rotation blocks per year, during which students rotated through the 6-week psychiatric clerkship: July—August, August—September, October—November, November—December, January—February, February—March, April—May, May—June. The assignment of students to rotation blocks is completed randomly, where any student would be equally likely to be assigned a particular rotation time of year.
The clerkship rotations were taken at one of four locations, including a tertiary-care university hospital, a Division of Mental Health state hospital, a Veterans Administration hospital, and a children's hospital. Each student rotated through one service at one of these locations, which were further classified according to their type of service emphasis as adult inpatient, adult outpatient, children's hospital, psychiatric emergency room (ER), or consultation/liaison. Students had the option of requesting the service they wanted (1st, 2nd, and 3rd choices) for their clinical experience. The requests were taken in order of receipt, "first-come, first-served." If there were "ties" for the requests, a disinterested person drew names. Those students not getting their first choices were given second or third choices when possible. Students who did not make requests were assigned to rotation services where remaining slots were available. The clinical experience, particularly the amount of psychiatric and medical services provided by students, varied considerably by site, whereas all didactic presentations and tests were administered to students as a group, regardless of the particular service a student was assigned.
A total of 26 preceptors provided supervision during the study time frame. All preceptors were full-time UAMS faculty members or clinical faculty members. Of these 26, a total of 21 had mentored more than 10 students during the 5-year study period. Most preceptors were assigned to the same site location for each rotation for which they were responsible during the study time frame. Because most of the preceptors taught students at only one site over the course of the study, it was not possible to adequately examine potential preceptor influences because they would be inherently confounded with site effects.
At the end of each 6-week psychiatry clerkship rotation, students completed a survey asking them to rate their interest in specializing in psychiatry before the clerkship and upon completion of the clerkship. Although it would have been preferable to have asked students their attitudes before they entered the clerkship as well as upon its completion, the present study relied on retrospective ratings of interest in psychiatry before the clerkship. Interest was rated on a 4-point Likert-type scale, with possible choices including Nonexistent (coded as 1), Weak (coded as 2), Moderate (coded as 3), and Strong (coded as 4).
Internal house-staff records and NRMP match results were used to determine which students chose to specialize in psychiatry upon completing their fourth year of medical school. Chi-square and t-test analyses were used to examine group differences. To model the effect for rotation time-of-year, students were grouped in the following way for comparisons: those who completed the clerkship early in the academic year (July—September), during the middle of the year (October—March), and later in the year (April—June). To model the rotation sites, students were classified according to their type of service emphasis, including adult inpatient, adult outpatient, children's hospital, psychiatric ER, or consultation/liaison. A linear-regression analysis was utilized to model the effect of the clerkship rotation time-of-year and type of rotation service on post-clerkship interest ratings, controlling for pre-clerkship interest levels. A logistic-regression analysis was used to model the effect of pre- and post-clerkship interest levels, rotation time-of-year, and type of service on subsequent decisions to specialize in psychiatry.
Of the 658 students who graduated from UAMS during our 5-year study time frame, 6% chose to specialize in psychiatry. There was no significant difference in the number of students by year who went on to specialize in psychiatry, nor were there differences in pre- or post-clerkship interest levels by graduating year. Also, there was no significant difference among the various preceptors in the frequency with which their students went on to specialize in psychiatry. For the following analyses, not all totals add up to 658, since there were some students who were excluded by the statistical computer program because of missing survey data or incomplete records.
The average rating by third-year medical students of their interest in entering a psychiatry residency before the psychiatry clerkship was 1.8±0.84 (nonexistent/weak interest). Students rated their interest in specializing in psychiatry significantly higher upon completing the clerkship, with an average rating of 2.9±0.87 (moderate interest; t[554]=27.89; P<0.0001).
Students who went on to specialize in psychiatry after graduating from medical school reported an average pre-clerkship rating of 2.8±1.2 (n=36; moderate interest in specializing in psychiatry) compared with a mean rating of 1.7±0.79 (weak interest in specializing in psychiatry) for those who did not eventually specialize in psychiatry (n=578; t[516]=7.07; P<0.0001). Post-clerkship interest levels were also significantly different between these groups, with an average rating of 3.9±0.30 (strong interest) for those who eventually specialized in psychiatry upon graduation, vs. 2.8±0.81 (moderate interest) for those who did not (t[517]=7.41; P<0.0001). Both groups showed significant pre—post increases as measured by a paired t-test to examine average differences in interest levels, with a mean increase of 1.1±1.1 for those who specialized in psychiatry (t[31]=5.75; P<0.0001) and a mean increase of 1.1±0.88 for those who did not eventually specialize (t[485]=27.6; P<0.0001).
t1 presents the pre- and post- mean interest ratings for the rotation times of the year and type of services to which students were assigned. Controlling for students' pre-clerkship interest levels, the linear-regression analysis for predicting post-clerkship student ratings revealed that students who completed a rotation in an adult outpatient site were significantly more likely to have higher post-clerkship ratings (β [unstandardized regression coefficient]=0.33; P<0.01). Although not statistically significant (β [unstandardized regression coefficient]=0.27; P<0.08), there was also a trend for students who completed a rotation in the ER to have higher post-clerkship interest ratings in psychiatry. The time of year that the clerkship was completed did not significantly contribute to explaining post-clerkship interest ratings.
In the logistic-regression analysis to explain whether students chose to specialize in psychiatry or not, both the pre-clerkship interest ratings (β [unstandardized regression coefficient]=0.53; Wald[1]=4.96; P<0.03), and the post-clerkship interest ratings (β [unstandardized regression coefficient]=3.2; Wald[1]=36.7; P<0.00001) showed significant effects. For each unit of increasing change in pre-clerkship interest ratings (ranging from 1 to 4), students were 1.7 times more likely to choose psychiatry as a specialty. For each unit of increasing change in post-clerkship interest ratings, students were 24.6 times more likely to have chosen psychiatry as their specialty. The remaining factors (sites and time-of-year) were not significantly related to students' choices to specialize in psychiatry.
As in previous work (15), the surveys of the present study revealed a rather dramatic increase in the likelihood of choosing psychiatry as a specialty after the junior-year psychiatric clerkship experience. Also, as others have found, those students with a greater interest in psychiatry before the clerkship were more likely to choose psychiatry as their specialty (4). The implications from these results for psychiatric residencies would seem to point toward aggressively recruiting those students who had an initial moderate interest in psychiatry and a strong interest after completing the clerkship, since it is these students who are 25 times more likely than their peers to become psychiatrists. In the present study, this short list contains practically all of the students who desired to enter psychiatry. The meaning of these findings is explored in relation to rotation times, the clerkship setting, and student autonomy, and their relationship to the choice of psychiatry as a specialty.
We hypothesized that medical students were more likely to choose psychiatry as their specialty if their junior-year psychiatry rotation was either early or late in the year. This idea was based on the rationale that early exposure to a discipline would influence a student to use the early clinical experience as a standard by which to judge subsequent clinical experiences. We also predicted that late-year exposure to a positive clinical experience would influence the selection of a specialty because of the imminent selection of senior electives. However, the data did not confirm these hypotheses. It seems that students' pre- and post-clerkship interest levels had much more influence on specialty selection than did the rotation timing. This finding suggests that medical students' interest in specializing in psychiatry is likely to be based on opinions they formed before and during the clerkship.
Also of note in this study was the finding that students who completed the clerkship in an outpatient setting were significantly more likely to have reported a higher interest in psychiatry after the clerkship than students rotating through other settings. Although not significant, there was also a trend for students who completed a rotation in the ER to have higher post-clerkship interest ratings in psychiatry; although the rotation setting appears to be related to students' post-clerkship interest ratings, it was not significantly related to choosing psychiatry as a specialty. When examined as a whole, previous research findings on the influence of the type of clerkship site have been primarily mixed; nonetheless, some studies have produced similar findings in terms of the clerkship site. Werkman et al. (14) also found that medical students gave the highest ratings for outpatient clinic and ER clerkship rotations. They suggested that this result may be due to the direct patient involvement and primary responsibility for patient care afforded by these rotations.
Although we did not measure the degree of patient involvement or primary responsibility for patient care afforded at each site for this study, anecdotally, our results are consistent with Werkman et al.'s suggestion (14), since the psychiatric clerkship at UAMS has a high degree of student autonomy, particularly for the rotations that showed the most positive change (the ER and adult outpatient services). For the inpatient services, students are typically required to report directly to a resident or attending physician after each patient they see—or, in some cases, such as the consultation/liaison service, psychiatric residents or attending physicians will accompany medical students, whereas the ER and outpatient services allow students to have more autonomy to make decisions related to patient care. The degree of autonomy may be important in bringing "moderately interested" students to the "highly interested" level.
Alternatively, perhaps the weaker response on inpatient rotations reflects student attitudes about severity of illness. Particularly since the advent of managed care, only the sickest patients are admitted for hospital-level care. Also, shorter lengths of hospital stays may prevent students from seeing the benefit of treatment on an inpatient unit. Although we do not have reason to believe that different preceptors accounted for the interest differences on various services, it is also possible that the unique quality of the learning experience or the charisma of different preceptors at various sites, for example, could have affected students' attitudes. More research, replicating our results and exploring potential reasons for these findings, will be needed to fully explain interest changes in psychiatry among medical students.
Other medical schools may vary considerably with regard to the clerkship structure or other factors, thus limiting the potential generalizability of these findings. For example, Weissman et al. (3) has shown that public medical schools in the southeastern United States have the highest "interest—change ratios" during medical school of any region in the United States. One explanation that has been offered for these results is that the choice of a medical specialty is probably greatly affected by the context of local economic conditions in which students plan to pursue postgraduate training and careers. The ready availability of jobs (e.g., great rural needs and a relative shortage of psychiatrists in the UAMS region) may limit the generalizability of these results, since some regions of the United States do not have such opportunity.
Other limitations of this study that should be noted include the way that "pre-" interest levels in psychiatry were measured. Ideally, it would have been best to have asked students about their "pre-" interest levels before they entered the clerkship, as opposed to using a retrospective "pre-" assessment of interest. Also, the assignment of students to rotation sites was not completely random, but rather, students had a modest amount of input with regard to the psychiatry service rotation to which they were assigned. Also, these results reflect the experience of students at a single medical school located in the southeastern United States.
Despite these limitations, the results have implications in terms of medical student education and recruitment into psychiatry. The present results indicate that psychiatric learning experiences in adult outpatient locations are most favored by medical students. Also, of the factors examined in this study, the strongest predictor of eventual specialization in psychiatry was post-clerkship attitudes, whereas the rotation site and time-of-year were not associated with the choice of psychiatry as a specialty. Because the psychiatric clerkship experience is clearly one of the most important elements of recruitment that educators can control, psychiatric residency programs will likely benefit by focusing on high-quality psychiatric teaching with an emphasis on safe, autonomous patient care.
The authors thank the medical students who completed surveys, Dr. Geoff Curran for his helpful statistical review, and Dr. Sidney H. Weissman for his consultation about the current status of research in this area.