Psychiatry has evolved significantly over the last half-century, undergoing a shift from a largely non-evidence-based discipline to a more medically based specialty, with a sophisticated understanding of diagnostic categories and with increasing research-based evidence for the efficacy of a variety of treatments. In particular, a considerable expansion of biological treatments has occurred, as well as a growth in short-term evidence-based psychotherapies (1). These changes have been fueled in part by the growth of pharmacological research and changes in health care delivery policy, including preferential coverage for short-term treatments, and decreased coverage for long-term psychodynamic psychotherapies, which have less substantial support in the psychiatric literature (2–5).

Does psychiatry residency training reflect these changes? Historically, long-term psychoanalytical/psychodynamic psychotherapies have been the mainstay of residency training, with relatively less emphasis on biological treatments (6–7). Today, on the other hand, residency programs increasingly minimize training in psychodynamic theories, preferring to emphasize biological psychiatry and cognitive behavior therapies (8–11). A few residency programs still retain long-term psychotherapy as a major portion of training, and most programs will offer reasonably balanced or “eclectic” training experiences.

Residency programs adjust curricula periodically, not only as a response to the greater forces from outside, but also based on feedback from their own residents. This feedback, however, often comes from residents who are currently in training and therefore have not had a chance to assess empirically whether the education received during residency meets the needs of current psychiatry practice. Though most residency graduates would likely report that their training was adequate, an examination of their retrospective assessment of specific aspects of their training would be valuable, particularly if it helped identify areas of psychiatry that are underemphasized or overemphasized in current training programs.

Though a number of studies have examined the influence of training on the career choices of physicians, most published research tends to focus on medical students and their choice of specialty (12). Coryell and Wetzel (13) surveyed 378 third-year psychiatry residents’ attitudes towards psychiatry training, forms of psychotherapy, and career plans. Results suggested a dynamic-biological continuum of interest, with medical education and drug therapy rating the most popular. Training in psychoanalysis and research were ranked lowest. Other studies have examined factors influencing psychiatry residents to choose particular subspecialties (14–16). Kozlowska et al. (17) surveyed New South Wales psychiatry trainees on their training experiences but did not examine opinions about specific career niches.

Our review of the literature suggests a dearth of information about psychiatrists’ subspecialty choices, how residency training affects their choice of career, and what they consider most valuable after entering clinical practice. Garfinkel et al. (18) examined predictors of success and satisfaction in a cohort of 29 psychiatrists over a 20-year career. Their results suggested that involvement with research and practicing from an orientation other than a psychoanalytical one predicted self-perception of success, but there were no data on perceived usefulness of residency training.

More recently, Stubbe (19) surveyed 392 U.S. child and adolescent psychiatry graduates’ assessments of training experiences. Respondents felt least prepared in administration and leadership skills, medical economics and business, complex psychopharmacology, management of complicated developmental disabilities, and cognitive behavioral therapy (CBT). On the other hand, respondents felt overtrained in consultation-liaison and inpatient work.

Most of the aforementioned investigations are limited in that they do not examine the long-term impact of residency training. By characterizing the impressions of a cohort of residency graduates over 2 decades, we sought to understand whether training has evolved in a way that serves the trainees after they complete their residencies. We designed a survey to learn about the career paths of residency graduates from the Massachusetts General Hospital (MGH) program and from the combined MGH/McLean program, and to determine whether they considered their training adequate with regard to current activities. As a related question, we wished to determine whether our residency graduates believed that residency programs should emphasize certain specific areas over others. We hypothesized that residency graduates would report diverse career choices, would indicate that residency training did not closely reflect professional experiences, and would suggest several areas of emphasis to increase the value of their training.

The study was approved by the MGH Institutional Review Board. We selected subjects who had graduated from the MGH Adult Psychiatry Residency program (and the MGH-McLean combined program established in 1997) within a 20-year period (1983–2003). We obtained names and addresses of adult psychiatry residency graduates from the MGH-McLean residency training office, and where possible, verified by Internet searches. A survey (Appendix 1) with a cover letter and a stamped, addressed return envelope were mailed to each subject. We identified a total of 281 residents and sent out 134 surveys in cases where subject contact information was available. There was no remuneration for participating. Return envelopes and surveys included no personal subject identifying information, in order to protect confidentiality. Returned surveys were stored in a locked cabinet and were accessible only to the principal investigator and study staff. Data were entered into a Statview (SAS Product) database for analysis.

Questions covered a number of areas, including career choices made since graduation; the extent to which graduates felt their training reflected their current professional duties; and ways in which graduates felt their training could have been improved with regard to preparation for postresidency activities. Specific questions evaluating training and areas of preferred emphasis were ranked on a scale of 1 to 6, where 1 represented more positive feelings about residency and particular didactic topics, and 6 represented more negative feelings about training and specific topics.

Responses were analyzed by descriptive statistics. To determine whether opinions varied with the amount of time passed since graduation from residency, we divided respondents into three cohorts. Comparisons between cohorts were made using the Mann-Whitney U test; comparisons within cohorts were made using the Wilcoxon signed-ranks test. Statistical significance was set as p<0.05.

Sixty-six subjects (49%) returned surveys. Respondents were divided into three cohorts based on the decade of graduation (1980s, 1990s, 2000s). The MGH-McLean group (2000s) was designated as one cohort, in view of the programmatic change involved. There were comparable numbers of respondents from the 1980s and 1990s, so setting breakpoints by decade of graduation provided cohorts large enough for statistical comparison and coincided with major changes in residency training. Three respondents did not include graduation year data and were excluded from comparisons between cohorts.

Cohort 1 (1980s; N=20; 35% female) trained largely under the psychoanalytical perspective, with growing psychopharmacology training and minimal research training. The majority of their seminars focused on process and psychoanalytical thinking. No seminars or formal training in CBT were provided.

Cohort 2 (1990s; N=27; 44% female) underwent training that was progressively influenced by the changes in Residency Review Committee requirements emphasizing more CBT and more specialty (e.g., addictions, geriatrics) and core competency-based training. In particular, the MGH program developed a clinical psychopharmacology training module providing a half-day per week of structured clinical training, supervision, conferences, and diagnostic evaluations. This modular program continues today as a significant component of the residency’s outpatient training. Simultaneously, the residency structure went from requiring longitudinal outpatient training that included adult, child, and group experiences throughout PGY-3 and PGY-4, to a more elective PGY-4. This change was implemented to optimize the individuality of residents and give them an opportunity to use their fourth postgraduate year to develop leadership skills and prepare for fellowship training and specialty careers.

Cohort 3 (2000s; N=16; 56% female) comprised graduates of the merged MGH-McLean residency program (which began in 1997, graduating the first class in 2000). The core clinical and didactic curriculum resulting from the merger offered fewer advanced psychoanalytical seminars and balanced psychodynamic training with CBT, neuroscience, and psychopharmacology. The latest RRC requirements were fully implemented. The program’s didactic curriculum underwent a series of changes from year to year whose final impact by 2000 was an overall increase in time devoted to both didactics and clinical supervision. The increases reflected additional CBT and psychopharmacology training and less advanced psychoanalytical teaching. Exact hours were not quantified.

Mean age of respondents overall (N=66), was 44 (SD=7) years. Forty respondents (61%) pursued fellowship training following residency. Thirty-six respondents (55%) described their current job as comprising a variety of “clinical” activities; 17 (26%) described their job as “mixed” (including research, clinical, and administrative duties); the remaining job descriptions (19%) collectively included positions that involved teaching only, child or adult clinical psychiatry only, child or adult research only, and unspecified research only. Regarding specific job activities, psychopharmacology was the most common (N=52; 79%), followed by teaching (N=42; 64%), supervision (N=33; 50%), research (N=31; 47%), administration (N=29; 44%), psychodynamic therapy (face to face therapy, usually once or twice a week, using dynamic principles) (N=29; 44%), and supportive therapy (N=25; 38%). Twelve respondents (18%) practiced CBT or a related psychotherapy. Least common activities included geriatric psychiatry (N=1; 2%), addiction psychiatry (N=5; 8%), and psychoanalysis (the traditional practice using the couch, free association, multiple appointments per week, and prerequisite psychoanalytical training for the practitioner) (N=7; 11%). These trends were generally consistent across all three cohorts. Cohort 1 had a greater proportion involved in administration (65%) and supportive therapy (55%) compared with Cohort 2 (33% and 37%, respectively) and Cohort 3 (38% and 25%, respectively); Cohort 3 had a greater proportion involved in research (75%) and inpatient care (44%) compared with Cohort 1 (35% and 10%, respectively) and Cohort 2 (41% and 7%, respectively). The percent distribution of all activities is summarized in Table 1.

Among the 66 respondents, the median satisfaction score with residency training as general preparation for current practice was high on the six-point scale, as was satisfaction with training as specific preparation for current activities (Table 2). Overall, respondents reported that training had strongly influenced their ultimate choice of specialty (Table 2). Mann-Whitney U test comparisons between the different cohorts revealed no significant difference in responses in these three areas (p>0.05 for all comparisons), suggesting no significant impact on response by decade of training, though Cohort 1 was noted for a higher satisfaction with their training as preparation for current activities (Table 2).

When asked what particular areas were inadequately covered during residency, CBT (N=7), administration (N=8), and research (N=6) were most strongly noted as deficient by the sample as a whole and in each cohort. Additional deficiencies were reported for child psychiatry in Cohort 1 (N=2), neuroimaging and neuropsychiatry in Cohort 2 (N=2), and psychodynamic therapy in Cohort 3 (N=2). Topics considered excessively emphasized during residency included psychodynamic therapy (N=10), inpatient psychiatry (N=7), and psychopharmacology (N=5). In the cohort analysis, other topics considered excessively emphasized included emergency psychiatry in Cohort 1 (N=2) and community psychiatry in Cohort 2 (N=2).

Various forms of continuing education were sought out by 46 respondents (70%) to remedy perceived deficits in residency training. Psychodynamic therapy (N=6) and research (N=5) were the most popular forms of continuing education. Additional areas of postgraduate training included CBT (N=3), supervision (N=3), child/adolescent psychiatry (N=3), unspecified conferences (N=3), psychoanalysis (N=2), consultation-liaison (N=2), addiction (N=2), and neuroimaging/neuropsychiatry (N=2). Areas receiving one endorsement each included psychopharmacology, supportive therapy, developmental disorders, electroconvulsive therapy, women’s mental health, and adult attention deficit hyperactivity disorder. Nine respondents did not specify the type of additional training pursued.

When asked to design a residency program with regard to specific training areas they wanted emphasized, the respondents indicated that psychopharmacology was far and away the most desired topic, both for clinical and didactic training. School psychiatry and psychoanalysis were the least popular topics, but psychodynamic therapy training and CBT training were ranked relatively high (Table 2). Similar trends were observed for all three cohorts, with psychopharmacology emerging as the most popular topic, a more moderate interest in CBT and psychodynamic therapy training (though Cohort 1 gave CBT a median score of 1 in didactic and clinical training and Cohort 3 gave psychodynamic therapy a 1 for clinical training), and relatively weak interest in psychoanalysis. To determine whether relative emphasis on different areas of training was significant, the Wilcoxon signed-ranks test was performed for each cohort with psychopharmacology as the standard comparator, given that it was the most popular topic. Because of the minimal differences between median scores for didactic and clinical, these categories were collapsed into one category for the analysis.

The sample significantly preferred psychopharmacology over psychoanalysis (z=−4.58, p<0.0001). The preference was strongest in Cohort 1 (z=−2.70, p=0.007), Cohort 2 (z=−2.50, p=0.012), and Cohort 3 (z=−2.52, p=0.012) (Table 2). We found a significant preference for psychopharmacology over psychodynamic therapy training in Cohort 3 (z=−2.03, p=0.042) but no significant preference in Cohorts 1 and 2 (p>0.05 for each). All cohorts expressed a significant preference for psychopharmacology training over training in behavioral therapies (p<0.05 for all).

Among the sample as a whole, we found a significant preference for CBT training over psychoanalysis (z=−4.06, p<0.001). This trend held for all three cohorts (p<0.05 for all). We did not find a significant preference for CBT over psychodynamic therapy training among the sample as a whole (z=−0.85, p=0.394). Examination of cohorts revealed a significant preference for psychodynamic training over CBT in Cohort 2 (z=−2.47, p=0.014) but not in Cohorts 1 and 3 (p>0.05 for each).

An understanding of the experience of residency graduates can yield valuable information about the adequacy of their training and may help residency programs determine what particular areas and skills should be emphasized. Previous investigators have proposed new guidelines for psychiatry training based on theoretical frameworks (20), but none on truly empirical data. To our knowledge, our survey represents the first attempt to establish the interests and views of adult psychiatry graduates with regard to what skills and topics their training emphasized and what they believe training should emphasize.

Overall, our residency graduates from 1983 to 2003 considered their training good preparation for the “real world” of practice and, contrary to our expectations, felt that educational topics emphasized were generally appropriate to their current activities. Responses from the sample as a whole and within each of the three cohorts suggested that psychopharmacology training should be strongly emphasized during residency. This is in line with past investigations (13) and certainly not surprising, given the speed with which new psychotropic medications are being developed and marketed. Administration and research were widely performed by our respondents and often felt to be deficient in residency training. These results echo previous complaints about the lack of emphasis on administrative skills during residency (19). Interestingly, neither administration nor research skills were emphasized by our sample as a whole when designing their “ideal” residency program. The high number of graduates who pursued fellowships after completing their residency was not surprising because our program tends to attract residents with an interest in academic careers.

The issue of psychotherapy training was of particular interest to us, given the apparently diminishing emphasis on psychodynamic and psychoanalytic instruction in U.S. residency programs in favor of biological treatments and cognitive behavior therapies (8–11). Our survey showed a significant rejection of psychoanalysis relative to psychopharmacology and CBT in all cohorts, most strongly in Cohort 1. Why might this be? The older Cohort 1 received more emphasis on psychoanalysis relative to psychopharmacology and CBT during training, and may feel that the latter disciplines should be emphasized more in the current environment that compensates CBT and psychopharmacology better than psychodynamic/psychoanalytic therapies. Conversely, the younger Cohorts 2 and 3 had less exposure to psychoanalysis than Cohort 1, and may not consider it important relative to other skills.

Our results appear consistent with those of Garfinkel et al. (18), which revealed an avoidance of psychoanalytic perspectives in their sample. Notwithstanding our subjects’ reticence towards psychoanalysis, all cohorts gave strong support for the continued inclusion of psychodynamic therapy in residency training. Psychodynamic therapy training was not rejected relative to CBT or psychopharmacology training, and even received a significant endorsement over CBT in Cohort 2. This cohort entered training when CBT was gaining prominence, and perhaps felt they received too little exposure to psychodynamics relative to CBT. These findings suggest that despite the growing emphasis on short-term therapies and biological treatments, psychodynamic therapy continues to be valued by residency graduates.

Our study has several limitations. We surveyed graduates of one particular residency program, which limits the generalizability of our results. The response rate was lower than hoped, limiting the robustness of the findings. Although we cannot confirm whether all graduates received surveys, the Internet facilitates verification of physician contact information; we therefore believe that the majority received the surveys. Respondents may have different qualities compared with nonrespondents, and this may bias their responses. The relatively small size of Cohort 3 compared to the other two cohorts may also limit the usefulness of statistical comparisons. Retrospective bias is another concern. Older graduates predictably reported a wider variety of past occupations than younger graduates. Many younger respondents may not yet have found their “right” niche, and therefore may not have a full perspective about the value of their training, compared with more seasoned respondents. On the other hand, it may be difficult for older graduates to clearly separate the value of their learning from residency versus postresidency training and CME activities. Responses may also be influenced by the niche area of each respondent. An endorsed area of preference may suggest respondent skill in this area and the belief that others should be skilled as well. Alternatively, it may reflect deficiency in this area and the desire for more training.

Our questionnaire is not yet formally assessed for reliability and validity and has some inherent limitations. For example, it does not differentiate between types of practice settings or address the impact of managed care on responses from practitioners in different settings. We may have overlooked certain categories of occupations and further training, but respondents were allowed to write in additional categories.

Given the rapidly changing nature of psychiatry and the health care system, this study could pave the way for more rigorous investigations of psychiatry graduates that may inform residency training committees with regard to future educational directions. For example, the need for continued emphasis on psychopharmacology training is clear, as is greater attention to training in research and administration. Our results suggest that investigations of this type are viable, inexpensive, and potentially useful for curriculum development. This work needs to be replicated by other investigators and should be targeted nationwide to psychiatrists from different residency programs in order to characterize generalizable trends among psychiatrists. Such studies are currently under development.