To predict opioid consumption and pain intensity after the index cesarean delivery, we tested a hypothesis that opioid consumption after the previous cesarean delivery of the same patient can predict the opioid consumption after the index cesarean delivery. We further tested a secondary hypothesis that the pain scores after the previous cesarean delivery can predict the pain scores after the index cesarean delivery.
This is a retrospective cohort study of 470 women who underwent both previous and index cesarean deliveries at a single institution from January 2011 to June 2019. To predict the opioid consumption (primary outcome) and average pain scores (on 11-point numeric rating scale) after their index cesarean delivery, we used a linear regression model incorporating only the opioid consumption and average pain scores after the previous cesarean delivery, respectively (unadjusted models). Demographic and obstetric variables were then added as predictors (adjusted models). The bootstrap was used to compare these models with respect to proportion of variance of the outcome accounted for (R2).
Unadjusted models were weakly predictive of opioid consumption (R2 = 0.268; 95% confidence interval [CI], 0.146–0.368) and average pain scores (R2 = 0.176; 95% CI, 0.057–0.250). An adjusted model for opioid consumption was weakly predictive (R2 = 0.363; 95% CI, 0.208–0.478), but an adjusted model for average pain scores was not predictive of the outcomes (R2 = 0.070; 95% CI, −0.143 to 0.219). Adjusted models failed to explain variances of opioid consumption and average pain scores significantly better than unadjusted models (P = .099 and P = .141, respectively).
Opioid consumption and pain scores after women’s previous cesarean delivery only explain 27% of variance of opioid consumption and 18% of variance of their pain after their index cesarean delivery. Therefore, previous cesarean delivery analgesic metrics are not robust enough to be used as clinically applicable predictors for index delivery.