Under the aegis of curricular reform, our institution reduced the length of the psychiatry clinical clerkship from 4 week to 3 weeks, making it the shortest psychiatry clerkship in North America. A 3-week elective was created for those students who desired additional exposure to psychiatry.

Medical educators—especially those in psychiatry—have debated for many years the appropriate lengths for core clerkships, with debate focusing on the minimal number of weeks required to acquire essential clinical skills and knowledge. In response to the declining lengths of many American psychiatry clerkships and the lack of standardization of length, the Association of Directors of Medical Student Education in Psychiatry (ADMSEP) published a position statement in 2006, stating that a rotation “must be at least 6 weeks in length or longer” to develop the skills necessary to recognize and appropriately manage psychiatric illnesses (1). Although both median and average American clerkship lengths are 6 weeks, individual lengths range from 3 to 8 weeks.

Although investigators disagree about whether shelf exams measure clinical skills, they are the only nearly-universal objective measure of psychiatric knowledge. Limited published reports reveal no consistent association between clerkship length and shelf-exam score for 6- to 8-week clerkships (2, 3). When rotation lengths are shorter, clerkship directors have been found to weigh PSE scores more than clinical performance (4), suggesting that reducing psychiatry clerkship length may result in a greater emphasis on standardized test scores than on interviewing and communication skills, crucial components in psychiatric diagnosis. Although shorter psychiatry clerkships do not necessarily influence psychiatry recruitment (5), future nonpsychiatrists in a shorter psychiatry clerkship get less exposure to psychiatric issues and diagnoses affecting patients in all medical and surgical specialties.

In light of this data and the ultra-short psychiatry clerkship at our institution, we grew concerned about the adverse impact this curricular reform might be having on our graduates’ psychiatry knowledge. We sought to compare differences in our students’ performances on the required psychiatry shelf exam (PSE), based on clerkship length.

We collected PSE scores on all medical students who completed the third-year psychiatry core clerkship between the years 1998 and 2010. We correlated this data with the weeks of psychiatry clerkship: 3, 4, or 6 weeks, experienced by the student before taking the exam. For the 3-week clerkship, students spent their time on a single inpatient psychiatry service: Consultation–Liaison (C/L), Medical Psychiatry, or Psychiatry Acute Care. For the 4-week clerkship, students could choose to spend the entire time on one service or select two services on which to spend 2 weeks each (including child psychiatry). Students in both 3- and 4-week-long clerkships took their PSE at the end of that block. Students who chose an elective ended up with 6 weeks of psychiatry. They were a self-selected group with a clear interest in psychiatry; their elective and core rotations were potentially separated by up to a year, and the student could elect to take the exam at any time during that period.

We compiled PSE scores from all Mayo medical students who took the exam over a 12-year period. We divided the students into three groups, based on clerkship length (6-week core/elective [N=14], 4-week clerkship [N=478], and 3-week clerkship [N=51]). We then determined the mean exam score for each group, and compared averages across the three groups by one-way ANOVA. Since the overall ANOVA F-test was statistically significant, we next performed two-sample t-tests between each pair of means. We used JMP Version 8 and SAS Version 9.1 (SAS Institute Inc., Cary, NC) programs for all data analysis. No adjustments were made for multiple comparisons. The Mayo Foundation Institutional Review Board granted study approval.

Mean shelf-exam scores were 81.5 (95% CI [CI]: 76.6–86.4; 6 weeks), 75.3 (95% CI: 74.4–76.1; 4 weeks), and 75.5 (95% CI: 73.0–77.9; 3 weeks). The p value for the overall ANOVA (testing the null hypothesis that the mean score is equal across all three duration groups) is significant at 0.04, demonstrating that there is a significant difference between the three groups. Results showed statistically significant reductions in exam scores in comparing the 6-week group to the 3-week group (p=0.03) and the 4-week group (p=0.01). The mean exam score in the 3-week group was slightly higher than in the 4-week group, but this difference was not statistically significant. We compared MCAT scores for the three groups, using students for which we had these data, and found that differences between the groups did not achieve statistical significance.

Our results show that, in our institution, students with shorter clerkship lengths had significantly lower PSE scores than those with longer clerkships. However, a difference of only 1 week did not result in a statistically significant performance reduction. Our results are similar to those of both Gary and Rosevear (2) and Case and colleagues (3). The former authors found no significant difference when clerkship length was shortened by 1 week, whereas the latter authors confirmed a significant PSE score reduction with a 2-week decrease in clerkship length (2, 3). Of note, the static shelf-score average at our medical school over the past 12 years has not kept up with the national mean, which has risen from 75.7 to 78.2 between 2007 and 2010 on a test scaled to a mean of 70 in 1993–1994.

A potential shortcoming of our findings is that clerkship length may not be the only variable influencing exam performance. Documented variables include exam preparation methods (6), number of patients cared for (7), rotation timing in the academic year (8), and didactic preparation as a specific part of the clerkship. On the other hand, clerkship setting (VA, non-VA, inpatient, C/L, or outpatient) has not been found to significantly influence shelf-exam score (9, 10). In our institution, students may have scored higher on the PSE if they had more interest in psychiatry, as reflected in their having chosen to do an additional 3-week elective. Thus, the 6-week group may have scored higher either because they had more time to study or because they were the group with the greatest interest in the specialty. Psychiatry clerkship directors weight the PSE objective data variably or not at all in grading. Of 111 clerkship directors surveyed, 69% factor the score into the final rotation grade, counting it, on average, for 31% of the overall grade (11). Directors of shorter clerkships tend to place more weight on the PSE (4).

General limitations of our study reflected in PSE scores could stem from students’ having had different teachers or different teaching methods over the years, as well as differences among the students themselves. Specific limitations include 1) the small size of the 3- and 4-week groups; 2) the use of a convenience sample, rather than a randomized assignment of students to varying clerkship lengths; 3) variations in teachers, teaching methods, and curricula across the 12 years of the study; and 4) changes in the test itself. Our findings may not be generalizable to all psychiatry clerkships.

In conclusion, in response to recent changes in our medical school’s curriculum that resulted in the Psychiatry core clerkship’s being reduced from 4 to 3 weeks, with an optional 6-week combined core/elective rotation, we compared PSE scores of students before and after this change. Students who completed 6 weeks in psychiatry had significantly higher PSE scores than students who completed either 3 or 4 weeks. Students completing 4 weeks in psychiatry had slightly lower scores than those with 3 weeks, though the difference did not achieve significance. Our findings suggest that clerkship lengths shorter than the ADMSEP-recommended 6 weeks lead to a significant reduction in psychiatry knowledge as reflected by the PSE score.