OBJECTIVES: Patients with unmet expectations for the time required to experience improvement (or sustained improvement) with treatment may prematurely discontinue otherwise effective treatment. We assessed median time to successful sexual intercourse (SSI) with sildenafil and placebo. METHODS: Successful sexual intercourse was assessed in an 8-week, double-blind, placebo-controlled trial of sildenafil treatment in men with erectile dysfunction. SSI for each episode of sexual activity was defined by patient report and by two events: 1) the subject attempted sexual intercourse and 2) the subject had an erection lasting long enough for successful intercourse. Median time to SSI was calculated for two types of events: 1) time to the first SSI (transient event) and 2) stable SSI (initial improvement sustained for the remainder of the double-blind 56-day study, defined as at least 50% of sexual intercourses being successful until the end of the study). Log-rank tests compared time-to-event curves between treatment groups. RESULTS: Of 288 men enrolled, 95 received placebo, 99 received sildenafil 100 mg, and 94 received sildenafil 50 mg. Median time from the start of the study to achieve initial SSI was 3 days for each group of sildenafil-treated patients (50 mg and 100 mg), compared with 13 days for placebo-treated patients (P<0.001). (Note that not everyone attempted sexual intercourse on the first day of treatment.) Median time to achieve stable (sustained) SSI was 5 days (50 mg) and 3 days  (100 mg) for sildenafil-treated patients versus 55 days for placebo-treated patients (P<0.001). CONCLUSIONS: This novel application of time-to-event analysis provides useful timeframes for when patients can expect to see initial and then stable improvements with sildenafil treatment. Median times to initial SSI and sustained SSI were significantly shorter for sildenafil-treated vs placebo-treated men.