OBJECTIVES: We sought evidence of DREAMS' impact on uptake of services and sexual risk among adolescent-girls-and-young-women (AGYW). DESIGN: Cohorts of AGYW aged 13-22 years were randomly selected in 2017-2018 and followed-up to 2019; 1081 in Nairobi, Kenya;1171 in Gem, western Kenya;and 2184 in uMkhanyakude, South Africa. METHODS: Outcomes were knowledge of HIV status, condomless sex (past 12 months), lifetime partners, transactional sex (past 12 months), and awareness and use of condoms and pre-exposure-prophylaxis (PrEP). Using a causal inference framework, we estimated the proportions with each outcome if all vs. none were DREAMS invitees by 2018. RESULTS: Among AGYW followed up in 2019, the percentage invited to DREAMS by 2018 was 74, 57, and 53% in Nairobi, Gem, and uMkhanyakude, respectively. By 2018, the estimated percentages of AGYW who would know their HIV status, comparing the scenarios that all vs. none were DREAMS invitees, were 86 vs. 56% in Nairobi, 80 vs. 68% in Gem, and 56 vs. 49% in uMkhanyakude. By 2019, awareness of condoms and PrEP was high among DREAMS invitees, but recent participation in condom promotion activities was less than 50% and recent PrEP use was around 0-10%. In Gem, there was evidence of a reduction attributable to DREAMS in condomless sex, and among younger AGYW in the number of lifetime partners;in Nairobi evidence of a reduction in condomless sex among sexually active older AGYW;and in uMkhanya-kude no evidence that DREAMS changed these outcomes. CONCLUSION: Alongside sustaining high levels of knowledge of HIV status, more is needed to link AGYW into prevention methods such as PrEP and condoms.Comprehensive HIV prevention promotes safer sexual partnerships, but poverty, social norms, and inequalities limit AGYW's prevention choices.