The utility of double-blind continuation data in assessing long-term efficacy was evaluated in an analysis of data from the development program of a new antidepressant, nefazodone. The benefit of nefazodone therapy was examined during continuation treatment of patients who improved during the 6-8 week acute phase of efficacy studies. Discontinuation for lack of efficacy was used as an indicator of relapse. Pooled long-term data from the extension phases of placebo-controlled, acute efficacy trials were analyzed by the Kaplan-Meier method to estimate survival curves for time to discontinuation. Both nefazodone (p < .01) and imipramine (p < .05) were more effective than placebo during long-term continuation therapy. The findings demonstrate the value of double-blind continuation data in the evaluation of long-term drug benefit and provide an early assessment of nefazodone's effectiveness in continuation treatment of major depression.