Abstract
BACKGROUND: Prediction models to identify parturients who experience protracted pain, prolonged opioid use, and delayed self-assessed functional recovery are currently inadequate. METHODS: For this study, 213 nulliparous women who planned vaginal delivery were enrolled and assessed daily until they completed three outcomes: (1) pain resolution; (2) opioid cessation; and (3) self-assessed functional recovery to predelivery level. The primary composite endpoint, 'pain and opioid-free functional recovery' was the time required to reach all three endpoints. The subjects were divided into two categories (the worst (longest time) 20% and remaining 80%) for reaching the primary composite endpoint, and each individual component. Prediction models for prolonged recovery were constructed using multivariate logistic regression with demographic, obstetric, psychological, and health-related quality of life characteristics as candidate predictors. RESULTS: Labour induction (vs spontaneous labour onset) predicted the worst 20% for the primary composite endpoint in the final multivariate model. Labour induction and higher postpartum day 1 numerical rating score for pain were predictors for being in the worst 20% for both functional recovery and pain burden. Labour type, delivery type, Patient-Reported Outcomes Measurement Information System (PROMIS) anxiety score, RAND 36 Item Health Survey 1.0 (SF-36) physical health composite score, and postpartum breastfeeding success were predictive of delayed opioid cessation. CONCLUSIONS: Labour induction and elevated numerical rating score for pain are predictive of poor recovery after childbirth. Further research is necessary to determine whether modification would benefit mothers at risk for poor recovery.