Abstract
BACKGROUND: Low- and middle-income countries (LMICs) face a lack of epidemiological data. The development of high-quality surveys is a key research priority in countries such as Brazil. Our aim is to discuss the difficulties in conducting a longitudinal epidemiological survey in a pilot study of a school-based sample in São Paulo. METHODS: Data came from a cohort of school-attending adolescents in two neighborhoods with different levels of urbanicity in São Paulo. Students born in 2002 and in the 7th grade during 2014 were recruited from nine public schools. Adolescents and caregivers were interviewed separately at baseline and at one year follow-up, using several instruments, including the Schedule for Affective Disorders and Schizophrenia for School-Age Children/Present and Lifetime Version (K-SADS-PL). RESULTS: Achieving unbiased sampling, keeping an updated register of participants' contact information, using a full clinical interview without an algorithm for its scoring, and maintaining a highly-trained research team were among the difficulties faced. CONCLUSION: Working closely with community leaders, organizing group efforts to perform interviews, using a short, easy to understand instrument and providing some reward for participants were identified as alternatives to dealing with these difficulties, useful not only in Brazil, but also in other LMICs.