A meta-analysis of an earlier multicenter, double-blind efficacy study comparing placebo, oxaprotiline and amitriptyline was performed in order to test the survival-analytical approach in modelling the onset of improvement and response to treatment with antidepressants. The sample consisted of moderately depressed male (n = 154) and female (n = 275) patients (aged 17-73), diagnosed according to DSM-III criteria for major depression. Of these, 120 were treated with oxaprotiline, 120 with amitriptyline and 189 with placebo. Efficacy criteria were Hamilton Depression (HAMD) and Anxiety (HAMA) and Zung Self-Rating scales. Up to eight ratings over a period of 40 days were available for analysis. The results showed that the sensitivity in discriminating between groups was substantially enhanced through the inclusion of drop-outs and consideration of the effect of time to withdrawal from the study due to lack of improvement. Withdrawal from the trial due to inefficacy occurred earliest under placebo (50% within the first 8 days), whereas less than 40% dropped out within the first 12 days under active treatments. The most interesting and unexpected finding of the analysis was that the time course of improvement among responders was independent of the treatment modality, and thus identical in all three groups. Specifically, the efficacy of any of the given treatments was reflected only by the total number of responders or nonresponders. Once triggered, the time course of recovery from illness becomes identical to that of spontaneous remissions as observed, for example, under placebo.(ABSTRACT TRUNCATED AT 250 WORDS)