Objective To determine if the prevalence of schistosomiasis in children aged 9–12 years is associated with the prevalence in 5–8-year-olds and adults after preventive chemotherapy in schools or the community. Methods We combined data from four community-randomized, preventive chemotherapy trials in treatment-naïve populations in Côte d’Ivoire, Kenya and the United Republic of Tanzania during 2010–2016 according to the number of praziquantel treatments and the delivery method. Schistosoma mansoni infection was sought on two slides prepared from each participant’s first stool using the Kato–Katz technique. We assessed associations between S. mansoni prevalence in 9–12-year-olds and 5–8-year-olds and adults in the community before and after treatment using Bayesian regression models. Findings Stool samples from 47 985 5–8-year-olds, 81 077 9–12-year-olds and 20 492 adults were analysed. We found associations between the prevalence in 9–12-year-olds and that in 5–8-year-olds and adults after preventive treatment, even when only school-age children were treated. When the prevalence in 9–12-year-olds was under 10%, the prevalence in 5–8-year-olds was consistently under 10%. When the prevalence in 9–12-year-olds was under 50%, the prevalence in adults after two or four rounds of preventive chemotherapy was 10%–15% lower than before chemotherapy. Post-chemotherapy age-group associations were consistent with pre-chemotherapy associations in this analysis and previous studies. Conclusion The prevalence of S. mansoni infection in 9–12-year-olds was associated with the prevalence in other age groups and could be used to guide community treatment decisions.