To determine the effect of repeated, annual, age-targeted therapy on prevalence and intensity of Schistosoma haematobium infection in an endemic area, we treated all available, infected, school-age children (n = 2, 493) in the Msambweni area of Coast Province, Kenya with a randomized protocol of oral metrifonate (10 mg/kg for three doses each year) or praziquantel therapy (40 mg/kg as a single dose each year) for a period of one to three years. During 1984-1987, 1, 101 children completed three years of therapy, 550 received two years, and 842 received a single year. Annual followup revealed significant long-term suppression of S. haematobium infection in the targeted school-age population. Both cross-sectional analysis and study of individual outcomes suggested maximal suppression of infection after two years of therapy. Suppression lasted more than two years after cessation of treatment, and was associated with reduced community transmission (gauged by decreased prevalence among new study entrants and decreasing negative-to-positive conversion on annual parasitologic examinations). Comparison of metrifonate and praziquantel outcomes indicated greater suppression of infection and longer infection-free intervals for some subgroups given praziquantel. We conclude that annual population-based therapy targeted to schoolchildren has direct and indirect beneficial effects for endemic communities. In some specific situations, repeat therapy may not suppress transmission, and reduced drug efficacy may be observed after one to three years, suggesting the need for additional non-drug control measures in highly endemic villages.