Data on dengue incidence in South-eastern Brazil, 2014-2018.

Data from the routine surveillance systems have been extensively used to estimate the incidence of dengue. However, routine surveillance data frequently underestimate the diseases' incidence. Underreporting of dengue cases is related to the varying spectrum of its clinical presentation, with a large proportion of mild and asymptomatic infections, to its unspecific signs and symptoms, to the limitations of access to health care, and to the performance of the surveillance system itself [1-3]. In order to obtain accurate figures on dengue incidence, a cohort of children and adolescents was set up and followed during four years. The incidence of reported cases was used as a reference for the sample size calculation, which was stratified by age groups. A two-stage procedure was used to select the participants: census tracts were randomly selected, and within each one, a pre-determined number of children of each age group was randomly selected. The parents or legal guardians of the participating children and adolescents provided a written informed consent. In the first home visit, they responded to a questionnaire containing data on socio-demographic characteristics, housing, access to water, sewage, and garbage collection. Also, during the first visit a blood sample of the participating child/adolescent was collected for dengue baseline serology. Beginning in the week after the enrolment, the parent or legal guardian that was designated in the first visit received weekly phone calls for fever surveillance. If the child/adolescent had fever during the week, a nurse was dispatched to the family's home to collect more detailed data on the fever episode and collect a blood sample for dengue diagnosis (IgG, IgM, NS1 and PCR). If the dengue diagnosis was confirmed, a medical appointment was scheduled, and another blood sample for confirmatory tests was collected. It was also agreed that in every anniversary of their participation, they would receive another visit for a blood collection for dengue serology, regardless if they had a fever episode or a confirmed dengue diagnosis during the previous year. This article contains the description of the cohort's dataset. It is associated with the article published in Acta Tropica, under the title "A cohort study to assess the incidence of dengue, Brazil, 2014-2018" [4]. The associated article focused on the seroprevalence and incidence of dengue, and explored some associations between both outcomes and some explanatory variables.