Daily behavioral profiles and academic underperformance among university students: a cross-sectional study.

BACKGROUND: Daily routines may be associated with university students' academic outcomes, but isolated-exposure analyses can obscure how behaviors co-occur. We examined associations between learning and lifestyle behaviors and academic underperformance and explored multivariable behavioral profiles. METHODS: This cross-sectional study included 322 first- to third-year life-sciences undergraduates. The primary outcome was academic underperformance, defined as grade point average (GPA) < 2.00. Exposure-specific and primary multivariable associations were estimated using Firth penalized logistic regression. Breakfast frequency and breakfast timing were modeled separately because of structural overlap. Mixed ordinal and categorical behaviors were clustered using Gower dissimilarity and partitioning around medoids (PAM); k = 2-6 were evaluated by silhouette width and 200 repeated 80% subsamples. RESULTS: Academic underperformance occurred in 97 students (30.12%). In the primary frequency model, compared with eating breakfast every day, eating breakfast sometimes (adjusted odds ratio [aOR] = 7.37, 95% confidence interval [CI]: 1.65-70.32) and never eating breakfast (aOR = 30.89, 95% CI: 6.00-317.35) were associated with higher odds of underperformance. Gaming for 4-6 h/day (aOR = 5.58, 95% CI: 2.36-14.19) or >6 h/day (aOR = 13.70, 95% CI: 3.46-60.29), compared with <2 h/day, showed similar associations. Formal interaction tests by sex and academic year were not significant. PAM identified a Dysregulated Routine profile (n = 182) and a Structured Routine profile (n = 140); underperformance prevalence was 47.8% and 7.1%, respectively. GPA was excluded from cluster formation. CONCLUSION: Lower breakfast frequency and prolonged gaming remained associated with academic underperformance, and the two routine profiles differed markedly in academic outcomes. These exploratory, cross-sectional findings require replication in prospective cohorts before they can be translated into student-support strategies.