Abstract
Inconsistencies in the identification of predictors for the transition from acute low back pain (aLBP) to chronic LBP (cLBP) may be attributed to the varying definitions of aLBP used in different studies. We investigated how adopting different aLBP definitions affects the set of predictors and the risk of transition to cLBP (LBP > 3 months that caused a problem for at least half the days in the past 6 months). We leveraged data from the ongoing prospective Quebec Low Back Pain Study to compose 3 aLBP groups at baseline: nonchronic (individuals not meeting the cLBP criteria, n = 788), acute (LBP < 3 months, n = 230), and new episode (LBP < 3 months preceded by ≥3 pain-free months, n = 182). The primary outcome was the transition to cLBP at 6 months. We built predictive models within groups using the minimum redundancy maximum relevance algorithm to identify key predictors, focusing on models discrimination and calibration. Risks of transition were 35.8%, 44.3%, and 45.6%, for the nonchronic, acute, and new episode groups, respectively. Pain intensity, disability, and depression emerged as consistent predictors across definitions. The acute and new episode models, but not the nonchronic , were considered clinically useful (area under the receiver operating characteristic curve > 0.7), with the latter displaying better calibration and increased performance after adjustment to pain duration. These findings highlight the importance of standardizing aLBP definitions to improve risk stratification and targeted early interventions. Clearer definitions can enhance predictive accuracy, ensuring more effective resource allocation and preventive strategies for individuals at risk of developing chronic pain.