Unlearning psychiatry: a cohort effect in the training environment.

The purposes of this study were to investigate the relationship between 1) residents' psychiatric performance and their primary care training environment; 2) residents' psychiatric performance and their year of training; and 3) the utility of feedback from the General Health Questionnaire (GHQ) and the year of training. Resident physicians (N = 32) were assigned to experimental and control groups; feedback of GHQ results constituted the intervention in the experimental site. Baseline data, as well as post intervention data, were collected on a total of 1040 patient encounters. Residents' psychiatric performance was measured by notation of a mental disorder in the Assessment portion of the SOAP note. Results showed that control group residents achieved superior performance at pretest and, in spite of a two-fold increase in the number of psychiatric diagnoses following GHQ feedback, experimental site posttest diagnoses did not significantly exceed control site frequency. In addition, experimental site first-year residents outperformed second-year residents, who likewise outperformed third-year residents, in recognizing their patient's mental distress. Differences in patient morbidity, over-diagnosis by first-year residents, and failure to have acquired psychiatric skills when a first-year resident were all tested and rejected as explanations for the unexpected inverse relationship between performance and years in training. It was therefore concluded that integration of psychiatric knowledge and skills into a primary care practice may be profoundly (and negatively) influenced by factors in the residents' training environment.