Abstract
OBJECTIVE: This study aimed to develop a model to evaluate the risk of musculoskeletal pain (MSKP) in adolescents and identify the associated factors. METHODS: A total of 680 adolescents were surveyed, with assessments of chronic neck, back, and shoulder pain, and related factors such as mobile dependence, physical activity, sleep quality, and sociodemographics. A multivariate logistic regression model was employed, with feature selection through correlation analysis and Lasso regression, to identify significant predictors and establish a risk scale. The model's performance was evaluated using the area under the ROC curve. RESULTS: The prevalence of positive MSKP cases was 22.6% [95% CI: 19.7% - 25.9%]. Despite the self-reporting nature of the input data, the model achieved an AUC-ROC of 0.82, demonstrating good discriminatory ability. Key predictors include sleep disturbances, high mobile dependency, engagement in household chores, age 16-18, and urban residence. Girls exhibited a higher propensity for MSKP. Engagement in football was the only feature associated with a reduction of positive MSKP probability. A risk model is proposed to group the students into tertiles with low (2.3%), medium (17.4%) and high (44.7%) prevalence of MSKP. CONCLUSIONS: The MSKP Index effectively stratifies adolescent risk based on key factors, with significant associations between MSKP and sleep problems, mobile dependency, age, and gender. Regular physical activity, especially soccer, emerges as a protective factor, supporting targeted prevention strategies.