Abstract
PURPOSE: Analyze maternal and child predictors associated with loss to follow-up in the newborn hearing screening program at maternity hospitals in northeastern Brazil. METHODS: Retrospective cohort study, including secondary data from infants (n=604) referred to the newborn hearing screening program in two maternity hospitals for monitoring and/or diagnosis. The predictors evaluated included socioeconomic factors, such as maternal age, marital status, income, schooling, place of residence, number of children and number of prenatal visits. In addition, maternal and child health factors, such as smoking and drug intake during pregnancy, consanguinity, congenital infections, craniofacial malformations, use of ototoxic drugs, syndromes and a history of hearing loss in the family. Statistical analysis was performed based on binary logistic regression models, using the stepwise method. RESULTS: The logistic regression model containing the number of prenatal visits and the history of hearing loss in the family was significant [χ2(2) =34.271; p<0.001]. The number of prenatal visits (OR = 2.343; 95% CI = 1.626 - 3.376) and family history of hearing loss (OR = 2.167; 95% CI = 1.507 - 3.115) were significant predictors. The other predictors were not significant. CONCLUSION: The results reveal that newborns whose mothers had ≤ 5 prenatal visits and those with a family history of hearing loss increased their likelihood of loss to follow-up by 2.3 and 2.1 times, respectively. It is important to provide subsidies for public health improvements in order to help advise, guide and educate mothers, especially during prenatal care.