Prevalence of outcomes in children and adolescents with type 1 diabetes dependent on health-related behaviours and respective odds ratios.

aPercentage of patients with respective outcome.bP-value (χ2-Test).cOdds ratio estimated from ordinal logistic regression (proportional odds model) adjusted for age, sex and socioeconomic status, separate model for each health related behaviour.dP-value (Wald-χ2-Test).eMean Hba1c by health-related behaviour.fP-value (t-Test).gDifference estimated from linear regression adjusted for age, sex and socioeconomic status, separate model for each health related behaviour, dp-value (t-Test).hP-value (Wald-χ2-Test).iP-value (Score test for proportional odds assumption).jPhysically inactive youths were excluded.kAs in this model the proportional odds assumption was rejected, we also estimated ORs in a non-proportional odds model. The OR estimates for ≥1 and ≥2 severe hypoglycaemic events were 1.34 (95% CI 0.74; 2.44, p = 0.340) and 3.19 (95% CI 1.34; 7.44, p = 0.020), respectively.Prevalence of outcomes in children and adolescents with type 1 diabetes dependent on health-related behaviours and respective odds ratios.