Objective: This study was conducted to investigate whether the quality of antidepressant drug trials has improved during the last 40 years.
Method: A sample of 314 randomized clinical trials published between 1962 and 1998 was analysed.
Results: From 1962 to 1970 the median number of patients per trial was 56 (range 24-137), from 1971 to 1980 was 50 (10-211) and from 1981 to 1990 was 51 (20-314). In the last 8 years the median sample size increased to 100 patients (20-1002). Trials had a median duration of 4 weeks in the first two decades of publication, and a median duration of 6 weeks in the following two decades. Patient selection criteria have become increasingly sophisticated and the median number of efficacy measures increased in the last 4 decades from 1 to 4.
Conclusion: Stringent selection criteria and sophisticated outcome assessment tend to exclude typical patients from randomized controlled trials and made it more difficult to follow many patients in the long term.