Umeå University's logo

Aims: Patient characteristics and treatment setting are potential predictors of premature dropout from lifestyle interventions, but their relative importance is unknown.

Methods: From the quality registry of the unit for behavioral medicine, Umeå University hospital, we identified 2589 patients who had been enrolled in a multimodal lifestyle intervention for cardiometabolic risk reduction between 2006 and 2015. Baseline characteristics predicting dropout before 1-year follow-up were selected by a stepwise logistic regression algorithm.

Results: Better physical health and older age predicted full participation, with odds ratios for premature dropout (ORs) of.44 (95% confidence interval (CI).31-.63), and.47 (95% CI.34-.65) in the highest compared to the lowest quartile, respectively. Odds of premature dropout were also lower among female participants,.71 (95% CI.58-.89). Premature dropout was predicted by higher BMI, snuffing tobacco, and smoking, with ORs of 1.53 (95% CI 1.13-2.08) in the highest compared to the lowest quartile of BMI, 1.37 (95% CI 1.03-1.81) comparing snuff user with non-users and 2.53 (95% CI 1.79-3.61) comparing smokers with non-smokers. Odds ratio for premature dropout among inpatients compared with outpatients was.84 (95% CI.68-1.04).

Conclusion: Higher risk at baseline predicts premature dropout.