Context: A significant amount of attention has been devoted to the complex issue of teenage pregnancy and to programs for reducing pregnancy among adolescents. Careful evaluations of such programs are needed to ascertain what strategies will be most effective at reducing teenage pregnancy. Methods: A pretest-posttest comparison group design was used to analyze the effects of a comprehensive multicomponent school and community intervention on estimated pregnancy rates and birthrates among young people in three Kansas communities: Geary County, Franklin County and selected neighborhoods of Wichita. Results: There were high levels of program activity in all three communities during the intervention period, including teacher training and sexuality education for students. Survey respondents rated highly such project interventions as the extension of school-linked clinic hours to accommodate student schedules and support groups established in middle schools. Between 1994 and 1997, the proportions of adolescents reporting that they had ever had sex decreased significantly among all ninth and loth graders in Geary County, from 51% to 38% among females and from 63% to 43% among males. In Franklin County, more males in grades 11 and 12 reported using condoms in 1996 (55%) than had done so in 1994 (39%). Age at first intercourse remained relatively stable in Franklin and Geary counties during the intervention period. The estimated pregnancy rate among adolescents aged 14-17 decreased between 1994 and 1997 in Geary Country, while it increased in comparison areas. The estimated pregnancy rates among 14-17-year-olds decreased in both Franklin County and its comparison communities. The birthrate declined both in one target area of Wichita and in its comparison area from 1991-1993 to 1994-1996. Over the same time period, the birthrate increased in a second target area of Wichita, while it decreased in the comparison community. Conclusions: This evaluation of a comprehensive multicomponent program for adolescent pregnancy prevention contributes to our understanding of this model and its replicability in diverse communities. Ongoing program evaluation is important for developing initiatives and for refining strategies so they respond to local conditions.