A decision-analysis model was developed to estimate the effects and costs of alternative initial management strategies for women presenting with dysuria and pyuria. We compared days of morbidity and direct medical costs associated with single-dose and multiple-dose regimens of amoxicillin and trimethoprim-sulfamethoxazole and examined the cost-effectiveness of doing an initial urine culture. We used varying assumptions for prevalence of etiologic agents, treatment efficacy, frequency of side effects, and duration of symptoms. Single-dose regimens were preferable to multiple-dose regimens of either drug, and trimethoprim-sulfamethoxazole was preferable to amoxicillin. Single-dose trimethoprim-sulfamethoxazole therapy resulted in the fewest expected symptom-days (2.7) and the lowest expected cost (+54). The advantage of single-dose strategies in minimizing expected symptom-days resulted largely from the threefold to fourfold increase in the incidence of side effects reported with multiple-dose therapy. Obtaining an initial urine culture in all patients reduced expected symptom-days by about 10% but increased expected cost by about 40%.