Implementing a bundle for evidence-based cesarean delivery may not be as beneficial as expected: a multicenter, pre- and post-study.

BACKGROUND

Standardization of procedures improves outcomes. Though systematic reviews have summarized the evidence-based steps of cesarean delivery, their bundled implementation has not been investigated.

OBJECTIVE

In this preimplementation and postimplementation trial, we sought to ascertain if bundled evidence-based steps of cesarean delivery, compared with the surgeon's preference, improve outcomes.

STUDY DESIGN

A Standards for Reporting Implementation Studies compliant, multicenter preimplementation and postimplementation trial at 4 teaching hospitals was conducted. The preimplementation period consisted of cesarean delivery done on the basis of the physicians' preferences for 3 months; educational intervention (eg, didactics, badge cards, posters, video) occurred in the fourth month. Cesarean deliveries in the postimplementation period employed the bundled evidence-based steps. A preplanned 10% randomized audit of both groups assessed adherence and uptake of evidence-based steps. The primary outcome was composite maternal morbidity, which included estimated blood loss ≥1000 mL, blood transfusion, endometritis, postpartum fever, wound complications, sepsis, thrombosis, intensive care unit admission, hysterectomy, or death. The secondary outcome was composite neonatal morbidity, and some of its components were a 5-min Apgar score <7, positive pressure oxygen use, hypoglycemia, or sepsis. A priori Bayesian sample size calculation indicated 700 cesarean deliveries in each group were needed to demonstrate a 20% relative reduction (from 15% to 12%) of composite maternal morbidity with 75% certainty. Bayesian logistic regression with neutral priors was used to calculate the likelihood of net improvement in adjusted relative risk with 95% credible intervals.

RESULTS

A total of 1425 consecutive cesarean deliveries (721 in preimplementation and 704 in postimplementation group) were examined. Audited data indicated that the baseline evidence-based steps utilization rate during the preimplementation period was 79%; after the implementation of bundled evidence-based steps of cesarean delivery, the audited adherence was 89%-an uptake of 10.0% of the evidence-based steps. In 4 aspects, the maternal characteristics differed significantly in the preimplementation and postimplementation periods: race/ethnicity, hypertensive disorder, and the relative contribution of the 4 centers to the cohorts and the gestational age at delivery, but the indications for cesarean delivery and whether its duration was less or greater than 60 minutes did not. The rates of composite maternal morbidity in the preimplementation and postimplementation groups were 26% and 22%, respectively (adjusted relative risk, 0.88 [95% credible intervals, 0.73-1.04]), with a 94 % Bayesian probability of a reduction in composite maternal morbidity. The composite maternal morbidity occurred in 37% of the preimplementation and 41% of the postimplementation group (adjusted relative risk, 1.12 [95% credible intervals, 0.98-1.39]), with a 95% Bayesian probability of worsening in composite maternal morbidity. When composite maternal morbidity was segregated by preterm (<37 weeks) and term (≥37 weeks) cesarean delivery, the improvement in maternal outcomes persisted; when composite maternal morbidity was segregated by gestational age subgroups, the potential for worsening neonatal outcomes persisted as well.

CONCLUSION

Standardization of the evidence-based bundled steps of cesarean delivery resulted in a modest reduction of the composite maternal outcome; however, a paradoxical increase in neonatal composite morbidity was noted. Although individual evidence-based steps may be of value while awaiting additional intervention trials, a formal bundling of such steps is currently not recommended.