The association between digital wellbeing application data and sleep quality in medical faculty students: a cross-sectional study

Digital device use among university students has raised concerns about adverse effects on sleep, but most evidence relies on self-reported screen-time estimates, which are prone to recall bias. We investigated the association between objectively logged screen-time data and sleep quality in medical students, and the sociodemographic and behavioural correlates of poor sleep. A cross-sectional analytical study was carried out at a Turkish medical faculty during the 2022–2023 academic year, with data collected between February and October 2023. A total of 215 students from all six years were enrolled. Sociodemographic data, the Pittsburgh Sleep Quality Index (PSQI), and seven-day per-application screen-time data extracted from each participant’s smartphone digital-wellbeing module were collected under researcher supervision. Multivariable linear and logistic regression models were fitted, with pre-specified sensitivity analyses. Mean daily screen time was 5.0 ± 2.0 h (34.6 ± 14.6 h weekly). The mean global PSQI score was 7.0 ± 3.4, and 60.5% of participants were poor sleepers. Mean PSQI scores were higher in third-year students, dormitory residents, recent ex-smokers and alcohol users. Weekly screen time was not associated with PSQI score, whereas monthly mobile-data allowance was ( r = 0.193, p = 0.004). In the multivariable linear model (R² = 0.181), higher PSQI scores were associated with longer daily WhatsApp use (β = 0.142, p = 0.027), alcohol consumption (β = 0.271, p < 0.001), lower perceived income and dormitory accommodation. In a parallel logistic model (PSQI > 5), alcohol consumption (OR 3.05, 95% CI 1.55–6.00) and non-dormitory accommodation remained significant, whereas WhatsApp use was borderline (OR 1.57, 95% CI 0.99–2.50, p = 0.056) and did not persist after full adjustment. In this cohort, six of every ten medical students reported poor sleep quality. Total smartphone screen time was not associated with sleep quality in any model. Alcohol consumption and dormitory accommodation were robustly associated with impaired sleep, whereas the association with messaging-application use was modest and did not survive full adjustment, and should be regarded as hypothesis-generating. Sleep- and digital-hygiene counselling in undergraduate medical curricula may nonetheless be warranted.