Construction and validation of a musculoskeletal disease risk prediction model for underground coal miners

To understand the prevalence among underground coal miners of musculoskeletal disorders (MSDs), analyze the risk factors affecting MSDs, and develop and validate a risk prediction model for the development of MSDs.

MSD questionnaires were used to investigate the prevalence of work-related musculoskeletal disorders among 860 underground coal miners in Xinjiang. The Chinese versions of the Effort-Reward Imbalance Questionnaire (ERI), the Burnout Scale (MBI), and the Self-Rating Depression Inventory (SDS) were used to investigate the occupational mental health status of underground coal miners. The R4.1.3 software cart installation package was applied to randomly divide the study subjects into a 1:1 training set and validation set, screen independent predictors using single- and multi-factor regression analysis, and draw personalized nomogram graph prediction models based on regression coefficients. Subject work characteristic (ROC) curves, calibration (Calibrate) curves, and decision curves (DCA) were used to analyze the predictive value of each variable on MSDs and the net benefit.

(1) The prevalence of MSDs was 55.3%, 51.2%, and 41.9% since joining the workforce, in the past year, and in the past week, respectively; the highest prevalence was in the lower back (45.8% vs. 38.8% vs. 33.7%) and the lowest prevalence was in the hips and buttocks (13.3% vs. 11.4% vs. 9.1%) under different periods. (2) Underground coal miners: the mean total scores of occupational stress, burnout, and depression were 1.55 ± 0.64, 51.52 ± 11.53, and 13.83 ± 14.27, respectively. (3) Univariate regression revealed a higher prevalence of MSDs in those older than 45 years (49.5%), length of service > 15 years (56.4%), annual income <$60,000 (79.1%), and moderate burnout (43.2%). (4) Binary logistic regression showed that the prevalence of MSDs was higher for those with 5-20 years of service (OR = 0.295, 95% CI: 0.169-0.513), >20 years of service (OR = 0.845, 95% CI: 0.529-1.350), annual income ≥$60,000 (OR = 1.742, 95% CI: 1.100-2.759), and severe burnout (OR = 0.284, 95% CI: 0.109-0.739), and that these were independent predictors of the occurrence of MSDs among workers in underground coal mine operations (p <  0.05). (5) The areas under the ROC curve for the training and validation sets were 0.665 (95% CI: 0.615-0.716) and 0.630 (95% CI: 0.578-0.682), respectively, indicating that the model has good predictive ability; the calibration plots showed good agreement between the predicted and actual prevalence of the model; and the DCA curves suggested that the predictive value of this nomogram model for MSDs was good.

The prevalence of MSDs among workers working underground in coal mines was high, and the constructed nomogram showed good discriminatory ability and optimal accuracy.