Are Adolescents Sensitive About Sensitive Data? Exploring Student Concerns About Privacy, Confidentiality, and Data Use in Health Research.

PURPOSE: We quantitively explored adolescents' concerns about privacy, confidentiality, and data use in health research and their potential impact on the accuracy of self-report data. METHODS: We analyzed data from 17,729 secondary school students who participated in the 2023 OxWell Student Survey. The survey assessed 5 concerns about privacy, confidentiality, and data use and asked students whether these concerns impacted the accuracy of their answers. We calculated the proportions who (a) endorsed each concern and (b) reported inaccuracies associated with their concern(s). We then examined associations of concerns and self-reported inaccuracies with nonresponse and score distributions on sensitive measures of mental illness (depression/anxiety and disordered eating) and adversity (child maltreatment) using logistic regression. RESULTS: 46.0% (8,160/17,729) of students endorsed ≥1 concern, and of these, 29.2% (2,379/8,160) reported associated inaccuracies. Relative to boys, concerns were more common amongst gender diverse adolescents (adjusted odds ratio [aOR] = 5.71, 95% confidence interval [CI] 4.40-7.48), gender nondisclosing adolescents (aOR = 4.36, 95% CI 3.62-5.26), and girls (aOR = 2.52, 95% CI 2.36-2.69), with smaller differences in self-reported inaccuracies. Students with self-reported inaccuracies were significantly more likely to have nonresponse on the 3 measures of mental illness and adversity (aORs = 1.53-3.38), whilst score distributions on those measures varied substantially according to whether students reported concerns. DISCUSSION: Concerns about privacy, confidentiality, and data use were common amongst student participants, as were self-reported inaccuracies. Substantial differences in nonresponse and score distributions on sensitive measures highlight potential impacts of these concerns. Co-designing and implementing strategies to address these concerns might help to support evidence-based decision-making by improving representativeness and data quality in adolescent health research.