Architecture of research in psychiatry, 1953 to 1983.

Although there has been one report on the trends in study design in general medicine, we are aware of none for general psychiatry prior to this communication. Accordingly, articles from the American Journal of Psychiatry (N = 194) and the Archives (N = 109) were randomly sampled for the years 1953, 1963, 1973, and 1983. Two raters achieved reliability (kappa = .82) for recognizing the major types of study design (cohort, clinical trial, case control, cross sectional, case report, and review). There was a significant change in study architecture over time, with the percentage of review articles declining and the percentage of case-control and cross-sectional studies increasing. Another major finding was a large increase in use of inclusion and exclusion criteria for diagnosis in non-review article studies. The general trends are for increasingly sophisticated research designs to be used in psychiatry research. The quality of research designs in psychiatry for 1983 also compares favorably with research designs found in a respected medicine journal.