Objective: We compared the efficacy of bladder training, pelvic muscle exercise with biofeedback-assisted instruction, and combination therapy, on urinary incontinence in women. The primary hypothesis was that combination therapy would be the most effective in reducing incontinent episodes. Study Design: A randomized clinical trial with three treatment groups was conducted in gynecologic practices at two university medical centers. Two hundred and four women diagnosed with genuine stress incontinence (n = 145) and/or detrusor instability (n = 59) received a 12-week intervention program (6 weekly office visits and 6 weeks of mail/telephone contact) with immediate and 3-month follow-up. Outcome variables included number of incontinent episodes, quality of life, perceived improvement, and satisfaction. Data analyses consisted of analysis of covariance using baseline values as covariates and χ² tests. Results: The combination therapy group had significantly fewer incontinent episodes, better quality of life, and greater treatment satisfaction immediately after treatment. No differences among groups were observed 3 months later. Women with genuine stress incontinence had greater improvement in life impact, and those with detrusor instability had less symptom distress at the immediate follow-up; otherwise, no differences were noted by diagnosis, incontinence severity, or treatment site. Conclusions: Combination therapy had the greatest immediate efficacy in the management of female urinary incontinence regardless of urodynamic diagnosis. However, each of the 3 interventions had similar effects 3 months after treatment. Results suggest that the specific treatment may not be as important as having a structured intervention program with education, counseling, and frequent patient contact. (Am J Obstet Gynecol 1998;179:999-1007.)