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/extraprofessional 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.

KEYWORDS:

• classification-tree model
• hospital environment
• low back pain
• predictive diagnosis
• risk factors

Acknowledgments

This study was performed at Sacré-Cœur Hospital (Baabda, Lebanon). The authors would like to acknowledge Sister Lamia Tamer (quality manager at the hospital) for her valuable assistance in organizing and helping with the questionnaire data. The authors also thank all the staff at the hospital who took part. The support of Dr Amal Mansour, Dean of the Faculty of Public Health, La Sagesse University is also gratefully acknowledged.