Epidural analgesia during labour, routinely or on request: a cost-effectiveness analysis

Objective: To assess the cost-effectiveness of routine labour epidural analgesia (EA), from a societal perspective, as compared with labour analgesia on request. Study design: Women delivering of a singleton in cephalic presentation beyond 36 + 0 weeks' gestation were randomly allocated to routine labour EA or analgesia on request in one university and one non university teaching hospital in the Netherlands. Costs included all medical, non-medical and indirect costs from randomisation to 6 weeks postpartum. Effectiveness was defined as a non-operative, spontaneous vaginal delivery without EA-related maternal adverse effects. Incremental cost-effectiveness ratio (ICER) was defined as the ratio of the difference in costs and the difference in effectiveness between both groups. Data were analysed according to intention to treat and divided into a base case analysis and a sensitivity analysis. Results: Total delivery costs in the routine EA group (n = 233) were higher than in the labour on request group (n = 255) (difference -(sic)322, 95% CI -(sic)60 to (sic)355) due to more medication costs (including EA), a longer stay in the labour ward, and more operations including caesarean sections. Total postpartum hospital costs in the routine EA group were lower (difference -(sic)344, 95% CI -(sic)1338 to (sic)621) mainly due to less neonatal admissions (difference 472, 95% CI -(sic) 1297 to (sic) 331), whereas total postpartum home and others costs were comparable (difference -(sic)20, 95% CI -(sic) 267 to (sic)248, and -(sic) 1, 95% CI -(sic)67 to (sic)284, respectively). As a result, the overall mean costs per woman were comparable between the routine EA group and the analgesia on request group ((sic)8.708 and (sic)8.710, respectively, mean difference 2, 95% CI -(sic)1.012 to (sic)916). Routine labour EA resulted in more deliveries with maternal adverse effects, nevertheless the ICER remained low ((sic)8; bootstrap 95% CI -(sic)6.120 to (sic) 8.659). The cost-effectiveness acceptability curve indicated a low probability that routine EA is costeffective. Conclusion: Routine labour EA generates comparable costs as analgesia on request, but results in more operative deliveries and more EA-related maternal adverse effects. Based on cost-effectiveness, no preference can be given to routine labour EA as compared with analgesia on request.