Outcomes among participants vs nonparticipants of randomized trials during pregnancy: a systematic review and meta-analysis.

ABSTRACT

OBJECTIVE: This study aimed to evaluate the outcomes among individuals who were eligible and approached for participation in a randomized controlled trial during pregnancy, comparing those who enrolled with those who declined participation. DATA SOURCES MEDLINE, Scopus, CINAHL, the Cochrane Library, and Ovid were searched from study inception to May 2022. STUDY ELIGIBILITY CRITERIA This study included all obstetrical randomized controlled trials that reported clinical outcomes for both participants and nonparticipants. METHODS: The primary outcome captured the presence of morbidity. It was a composite of the primary outcome of each study comparing the participant arm with the nonparticipant arm. If a primary outcome was not clearly defined, a surrogate was developed on the basis of the core outcomes for the clinical condition studied. The risk of bias was assessed with the Newcastle-Ottawa Scale. Subgroup analyses for relevant obstetrical and neonatal outcomes were performed. The summary comparisons were reported as odds ratios with 95% confidence intervals computed using random-effects meta-analysis with heterogeneity evaluated using the I2 statistic. A funnel plot was used to examine publication bias, and there was no asymmetry. RESULTS: After reviewing more than 1100 abstracts, 17 obstetrical randomized controlled trials (103,610, with 26,293 participants and 77,317 nonparticipants) met our inclusion criteria and were analyzed. Of note, 9 studies were not rated as high quality, primarily for failing to control for confounding factors. Trial interventions were categorized as antepartum (n=11), intrapartum (n=5), or postpartum (n=1). Overall, participants in obstetrical randomized controlled trials had no difference in outcomes compared with nonparticipants (n=17 odds ratio, 0.88; 95% confidence interval, 0.52-1.49; I2=90%). Moreover, there was no difference seen when only randomized controlled trials that reported a primary outcome were included (n=12 odds ratio, 0.76; 95% confidence interval, 0.38-0.1.49; I2=93%). In addition, there was no difference noted in the subgroup where the randomized controlled trial intervention was not available to nonparticipants (n=7 odds ratio, 0.91; 95% confidence interval, 0.45-1.85; I2=68%). CONCLUSION: Pregnant individuals who agreed to participate in randomized controlled trials had no difference in outcomes compared with those who decline participation. Pregnant individuals should be provided with this reassuring information when offered to participate in a randomized controlled trial. Moreover, this information may improve patient recruitment into randomized controlled trials.