Evidence from a multicentre randomized controlled study

Publisher Copyright: Copyright © 2022 Valente de Almeida, Correa and Gonçalves.Objectives: We measure the impacts of an intersectoral intervention tackling adolescent substance use implemented between 2017 and 2019 in a tri-border region of Brazil, Paraguay, and Argentina. Methods: The intervention involved 23 institutions from different sectors and 880 adolescents, equally split between randomly selected treatment and control classes across institutions. Treatment group students were involved in the co-development of activities to tackle substance use within their institutions. Both treatment and control group students benefited from the activities developed and implemented from the second year of the intervention. We use difference-in-differences models to measure the impacts of participation in the co-development of the activities on alcohol, tobacco, and cannabis consumption. Results: Adolescents involved in the co-development of activities are 8 pp less likely to consume tobacco and cannabis, and 13 pp less likely to consume alcohol (p (Formula presented.) 0.01), compared to those who only participate in the activities. Among cannabis users, frequent consumption is also reduced by the intervention. Peer frequency of consumption is strongly associated with individual consumption. Conclusion: Co-development of activities by the subjects themselves can be key to decreasing substance use in this very crucial stage of life, especially if the institutions and the implementers are familiar with the area and subjects of the intervention.publishersversionpublishe