The application of a classification-tree model for predicting low back pain prevalence among hospital staff

Abstract

Low back pain (LBP) is a widespread musculoskeletal condition that frequently occurs in the working-age population (including hospital staff). This study proposes a classification-tree model to predict LBP risk levels in Sacré-Cœur Hospital, Lebanon (as a case study-236 chosen staffs) using various predictor individual and occupational factors. The developed tree model explained 80% of variance in LBP risk levels using standing hours/day (90% in relative importance), job status/sitting hours per day (80% each), body mass index (71%), working days/week (63%), domestic activity hours/week (36%), weight (35%), job dissatisfaction/sitting on ergonomic chairs (30% each), height (28%), gender (27%), sufficient break time (26%), using handling techniques/age (25% each), job stress (24%), marital status/wearing orthopedic insoles/extra professional activity (22% each), practicing prevention measures (20%), children care hours/week (16%), and type of sport activity/sports hours per week, car sitting, and fear of changing work due to LBP (15% each). The overall accuracy of this predictive tree once compared with actual subjects was estimated to be 77%. The proposed tree model can be used by expert physicians in their decision-making for LBP diagnosis among hospital staff.