Psychiatry has faced a serious decline in recruitment among graduates of U.S. medical schools for many years. After an increase in the number of graduates entering psychiatric residencies between 1925 and 1944, recruitment remained high until 1970. However, since the mid-1970s enrollment has generally decreased, from 5.9% of the graduating class in 1978 to 3.2% in 1994, the lowest proportion since 1929 (1). This pattern was interrupted only by a small upturn in enrollment in the mid-1980s. The decline in enrollment has become even more serious during the last decade. For instance, student matches into psychiatry between 1988 and 1994 dropped by 41%.
Numerous studies and reviews have addressed the current decline. Sierles and Taylor (1) listed various hypotheses and factors influencing specialty choice, pointing out that the decline is multifactorial in origin. The researchers argue that recruitment into psychiatry is largely affected by the type (a public school with lower tuition does better) and geographic location of the institution (the South does better). Sierles and Taylor felt that other variables, such as student attitudes and course characteristics, were less important. The multifactorial origin of the decline is unquestionable. However, the two "main" factors identified by Sierles and Taylor (1) seem to oversimplify the problem, as their conclusions no longer explain many of the differences among schools. For instance, Cornell University, one of the top recruiting schools, is private and located in the Northeast.
In 1981, Nielsen and Eaton (2) published results of a questionnaire study of the attitudes of 204 senior medical students at Georgetown (GT) and George Washington (GW), two East Coast medical schools fairly typical of American medical schools with respect to the percentage of graduates entering psychiatry. Dr. Nielsen developed a questionnaire with questions both from prior work and conventional wisdom. About half of the questions were stated in terms favorable to psychiatry. The other half were stated in negative form. Both open-ended and Likert-type questions were included. Authorization and positive introductory letters to students were provided by the deans' offices. The questionnaire mailings, sent in January and March of 1979, netted the following response rate: GW 57%, GT 59%; combined 58%. Their results suggested that most students remained generally positive about the field of psychiatry, even though it was not their career choice. A substantial number of students voiced criticisms around treatment efficacy, scientific merit, the low profile of psychiatrists during nonpsychiatric clerkships, psychiatrists not being "real docs," the low status of psychiatry, and frequent criticisms by other medical specialists and house staff. Nielsen and Eaton (2) emphasized that "strong negative socializing pressure from peers, house staff, and nonpsychiatric faculty were reported to be both common and effective in discouraging students interested in psychiatry."
These factors, together with those identified by Sierles and Taylor (1), are echoed throughout the literature on psychiatry's recruitment problems. However, despite the advantages in scientific knowledge and treatment efficacy in psychiatry in the last 20 years, recruitment continues to decline.
We studied medical students' attitudes toward and views of psychiatry in the early 1990s. Our survey instrument was partially derived from the Nielsen and Eaton survey (2) so we might observe changes in attitudes and views of psychiatry.
Students participating in psychiatric clerkships came from the following five medical schools: Albany Medical College, Boston University School of Medicine, University of New Mexico School of Medicine, University of Pittsburgh School of Medicine, and Wayne State University Detroit School of Medicine. The students were surveyed during the 1993—1994 academic year. School selection was based on willingness of each school's Director of Medical Student Education in Psychiatry to participate in this study.
A 39-item questionnaire was developed by one of the authors (R.B.). Questions surveying demographic characteristics and balanced positive and negative attitudes toward and views of psychiatry were included. Twelve of the 29 questions were taken directly from the Nielsen and Eaton questionnaire (2), or slightly modified, and 17 new questions were developed. Questions of attitudes and views were forced-choice ones—strongly agree, moderately agree, moderately disagree, strongly disagree. The questionnaire was accompanied by a letter explaining the purpose of the study and that it could be completed anonymously.
The questionnaire was distributed at the completion of the junior medical student clerkship in psychiatry. A machine-scorable answer sheet was used. According to the Human Investigation Committee at Wayne State University, written informed consent was not required, as completing this questionnaire implied consent. Participation was anonymous and voluntary, and there were no risk to the subjects.
There were missing data for many questions in a small number of subjects that is not presented in the summary table and results. Descriptive data were tabulated for each item in the questionnaire. We used the Statistical Package for the Social Sciences for statistical analysis (3).
We compared Questions 1, 2, 3, 4, 5, 8, 9,11, 14, 17, 22, and 28 from our questionnaire with the same or opposite-worded (Question Nos. 5, 8, 28) questions from the Nielsen and Eaton study (2). An example of "opposite wording" is "psychiatric consultation…is often helpful" in our study vs. "psychiatric consultations…are only rarely helpful" in the Nielsen and Eaton study (2). The Nielsen and Eaton study also allowed for a neutral response as opposed to our study, which had a forced-choice format. Such differences make direct comparison difficult. In these situations, a variety of approaches are used to see if they converge on a common conclusion. The primary statistical analysis consisted of a chi-square test for each of the 12 questions. To accomplish this, "neutral" as a response was ignored for each question, and a chi-square test with three degrees of freedom was conducted. To preserve an overall level of significance of 0.05 here, a Bonferroni correction would adjust the level of significance to the 0.004 level. Ignoring neutral as a response is problematic if the proportion of students answering neutral is sizable (e.g., 10% or greater), as it was for 10 of the questions (this proportion ranged from 8% to 31% in the Nielsen and Eaton study).
For that reason, in the second analysis, we also compared the proportion of students agreeing (or disagreeing) with those students with other opinions for each question between the two surveys using a chi-square test with one degree of freedom. This approach includes the proportion of students answering neutral in the first survey and implicitly assumes that those students in the current survey who might have chosen neutral would instead have chosen either agree or disagree. This approach, however, has the disadvantage of lumping students who disagree or disagree strongly with those who agree on each question. Thus, the third approach was to compare the proportion of students answering either strongly agree or agree with the proportion of students answering either strongly disagree or disagree for each of the 12 questions between the two surveys using a chi-square test with one degree of freedom. To preserve an overall level of significance of 0.05, a Bonferroni correction would adjust the level of significance to 0.001.
Selected Demographic Characteristics
Data from 479 students from four medical schools were analyzed. The results from Pittsburgh were lost in the mail and and not retrievable. All students in the remaining four schools participated in the study. Student breakdown is as follows: 98 from Albany, 58 from Boston, 68 from New Mexico, and 255 from Detroit. The group consisted of 226 female (47.2%) and 251 male (52.4%) students.
To the question "How seriously are you considering psychiatry as a career choice?" 183 (38.2%) students chose "I never considered career in psychiatry."; 81 (16.9%) chose "I considered psychiatry, but I will definitely not enter it."; 121 (25.3%) chose "I have considered psychiatry, but I probably will not enter it."; 82 (17.1%) chose "I am considering psychiatry now, but I am not sure if I will enter it."; and 10 (2.1%) chose "I will definitely take a residency in psychiatry." The last number—2.1%—is close to the national average. Nationally, about 3% of medical students have been entering psychiatry in recent years. It is possible that some students in our sample, who considered psychiatry and were not sure if they would enter, later entered psychiatry.
Among those who were definitely not choosing a residency in psychiatry, students selected the following career choices: internal medicine: 96 (19.9%), family medicine: 87 (18.2%), surgery and surgical subspecialties: 71 (14.8%), pediatrics and its subspecialties: 45 (9.3%), obstetrics/gynecology: 36 (7.5%), emergency medicine: 29 (6.1%), ophthalmology: 15 (3.1%), diagnostic radiology: 12 (2.5%), neurology: 10 (2.1%), anesthesiology: 8 (1.7%), therapeutic radiology: 7 (1.5%), pathology: 6 (1.2%), dermatology: 5 (1%), and "none of the above": 29 (6.1%). Twenty-three (4.8%) students were undecided.
Attitudes and Views of Psychiatry
Results are summarized in Tables 1—6. Questions appear in different order than in the actual questionnaire.
In T1, Questions 1—3 were compared with similarly worded questions from the Nielsen and Eaton study (2). Primary statistical analysis revealed a significant difference in the sense of more positive views in our study for all three questions (Question 1: χ2=44.99, df=3, P<0.001; Question 2: χ2=15.04, df=3, P<0.002; and Question 3: χ2=16.25, df=3, P<0.001). The second approach revealed a similar trend for all three questions (χ2= 50.13, df=1, P<0.001; χ2=29,5, df=1, P<0.001; and χ2=23.32, df=1, P<0.001, respectively). P-values for Questions 2 and 3 in the third analysis were 0.035 and 0.021, respectively; however, a Bonferroni correction would adjust the level of significance to 0.001 level.
In T2, Questions 4 and 5 were compared with the questions in the Nielsen and Eaton study (2). Question 4 was worded similarly, whereas Question 5 was worded in the opposite way. In the primary analysis, significantly more students in our study responded that they would recommend a psychiatric consultation for a family member and that psychiatric consultation for medical or surgical patients is helpful (Question 4: χ2=65.57, df=3, P<0.001; Question 5: χ2=152.13, df=3, P<0.001). The results were similar in the second approach (Question 5: χ2=76.3, df=1, P<0.001 and χ2=149.24, df=1, P<0.001). The difference for Question 5 was significant even in the third analysis (χ2=93.5, df=1, P<0.001).
In T3, Questions 8, 9, and 11 were compared with the questions in the Nielsen and Eaton study (2). The difference for Question 8 (opposite wording) was not significant in any analysis. The difference for Question 9 was significant in all three analyses (primary: χ2=47.75, df=3, P<0.001; second: χ2= 62.04, df=1, P<0.001; and third: χ2=19.34, df=1, P<0.001). Significantly more students in our study believed that psychiatrists are more qualified than psychologists and social workers. The difference for Question 11 was significant in the primary analysis and secondary analysis (primary: χ2=56.28, df=3, P<0.001; second: χ2=12.29, df=1, P<0.001). More students in our study felt that psychiatrists are too apologetic.
In T4, Question 14 was compared with the same question in the Nielsen and Eaton study (2). Significantly more students in our study felt that psychiatrists abuse their legal power (primary analysis: χ2=116.24, df=3, P<0.001; second analysis: χ2= 130.54, df=1, P<0.001; third analysis: χ2=4.27, df=1, P<0.039; however, a Bonferroni correction would adjust the level of significance to the 0.001 level).
In T5, Questions 17 and 22 were compared with similarly worded questions from the Nielsen and Eaton study (2). Significantly more students in our study felt that psychiatry had a high status among other disciplines in two analyses (primary analysis: χ2=14.52, df=3, P<0.002; second analysis: χ2= 1.72, df=1, P<0.19; and third analysis: χ2=12.75, df=1, P<0.001). More students in our study felt that would-be psychiatrists are viewed as odd; however, this difference was significant in the primary analysis (χ2=14.59, df=3, P<0.001) and second analysis (χ2=22.97, df=1, P<0.001).
In T6, Question 28 was compared with an opposite-worded question in the Nielsen and Eaton study (2). Significantly more students in our study (primary analysis: χ2=48.68, df=3, P<0.001; second analysis: χ2=12.29, df=1, P<0.001; and third analysis: χ2=44.1, df=1, P<0.001) felt that faculty at their school is respectful of psychiatry.
Presenting all of the responses for individual schools is beyond the scope of this article. Out of the 10 students who definitely decided to enter psychiatry, 4 were from Detroit (1.6% of the class), 3 from New Mexico (4.4%), 2 from Albany (2%), and 1 from Boston (1.7%). Of the 82 (17.1%) students considering psychiatry, but undecided, 47 were from Detroit (18.4%), 12 from New Mexico (17.6%), 16 from Albany (16.3%), and 7 from Boston (12.1 %).
Distribution of attitudinal answers was similar among the schools, with some exceptions. The students at the University of New Mexico had the most positive views of psychiatry, although their clerkship was the shortest (4 weeks vs. 6 or 7 weeks at the other schools). These students also felt strongest about the quality of psychiatry teaching at their medical school.
Medical students in a fairly sizable and geographically diverse sample presented fairly positive views of psychiatry. What has changed during the last 15 years? Compared with the Nielsen and Eaton study (2), the students felt more strongly that psychiatric care had progressed and that psychiatry in general was rapidly expanding and agreed less that it is still an unscientific and imprecise field. The positive view about the usefulness of a psychiatric consultation for both a family member and in general was significantly stronger in our sample.
Significantly more students in our study felt that psychiatrists are more qualified to treat mental illness than are other mental health professionals (78 vs. 49%), and there is higher level of respect for psychiatry among other medical school faculty. Even though the percentage of students in our study who felt that the status of psychiatry among other medical disciplines was low, it was significantly higher than in Nielsen and Eaton's (2) study (15% vs. 4%). The students also were significantly more inclined to believe that psychiatrists are too apologetic when teaching psychiatry, that would-be psychiatrists are odd, and that psychiatrists abuse their legal power. Our results on psychiatrists as clear thinkers, on perceived income, on not being encouraged to pursue psychiatry as a career, and the stigmatization of psychiatry itself were similar to the Nielsen and Eaton results.
Overall, the attitudes and views in all four medical schools were fairly positive and similar with some exceptions (e.g., the students at New Mexico were more positive in general).
Only 2.1% of the students chose a residency in psychiatry. Thus, any meaningful interpretation of the data relative to psychiatry as a career choice was impossible. However, the number of students entering psychiatry continues to decline despite the fact that attitudes and views of psychiatry are generally more positive than they were 15 years ago. Their opinions do not appear to play an important role in students' career choice decisions.
Our results are cross-sectional and do not answer another important question. How does the psychiatric clerkship experience influence views, attitudes, and career choices? Most authors believe the clerkship experience is the most important influence on recruitment (1). However, this matter has not been adequately studied. The length of a clerkship may not be an influential factor in a student's decision about their career choice of psychiatry. The students in New Mexico (with the shortest clerkship) had the highest percentage of students definitely taking a residency position in psychiatry (4.4%).
We used questions similar to those used by Nielsen and Eaton (2) and by Yager et al. (4), and obtained similar results. Forced-choice questions provided a clearer idea of student perceptions. As our questionnaire was strictly anonymous and administered at the end of a clerkship experience, we believe that the answers truly reflect students' attitudes and views. However, as we questioned students at the end of their clinical experience, it is possible that they were most likely to identify with psychiatrists and also most likely to attribute their negative attitudes to other specialties. Further limitations of our study include the limited number of schools sampled and their relative lack of representativeness.
What are the implications of our findings? The image of psychiatry among medical students is fairly positive and improving. Psychiatry is viewed as a valid, progress-making branch of medicine and as a helping discipline. However, psychiatrists are perceived as having a low status among other disciplines, with lower income compared with other physicians. A significant portion of medical students still feel uncomfortable with mentally ill patients.
We can only speculate about the decline of interest in psychiatry among U.S. medical students. The most frequently quoted reasons for this decline are 1) psychiatry has become more biological, hence more conventional and more like other medical specialties, making it less glamorous for students contemplating careers as psychotherapists and psychoanalysts (1); 2) psychology, social work, and psychiatry's competition for college students (1); 3) students' preference for high-income specialties as well as the fact that the rate of increase in a psychiatrist's income is less than that for other physicians (1); and 4) psychiatry's failure in its promise to change society through medicine. The image of the psychiatric practice is changing. In the past, the kind of student who went into psychiatry was the one interested in the psychosocial aspects of medicine and hoped to practice psychotherapy. Currently, with the consequences of managed care, psychiatrists are frequently viewed as psychopharmacologists primarily, with psychologists and social workers doing the psychotherapy.
It is also important to point out the fairly high negative ratings of psychiatry in our questionnaire revolving around career and personal rewards. Those items highlighted the stigma attached to psychiatry by society itself, and by other disciplines, that may interfere with recruitment despite the overwhelmingly positive student perceptions of psychiatry. Efforts to increase recruitment may be doomed to failure if societal and peer pressures discourage entry into the discipline.
Furthermore, the interpretation of our findings can only be fully appreciated in the larger context of psychiatric manpower. For years, we have not been able to answer the question "Are we training too many psychiatrists?" (5) If we are, declining interest may not be such a burning issue. Yet if we do not train enough psychiatrists or if the number of psychiatrists in training is just about right, then the continuous decline of interest in the midst of an improving image for our field presents a serious problem.
Further systematic research should then focus on issues such as 1) the medical school admission process and its relevance for recruitment into psychiatry, 2) what factors influence a student's choice of specialty during medical school, and 3) longitudinal studies of students undecided about psychiatry and students with a definite interest or disinterest in psychiatry.
This paper was presented at the Annual Meeting of the Association of Directors of Medical Student Education in Psychiatry, Santa Fe, NM, June 13—15, 1996.
The authors thank Frederick S. Sierles, M.D., for his invaluable help and advice and Cynthia Arfken, Ph.D., for help with statistical analysis.