Effectiveness of antenatal corticosteroids at term: Can we trust the data that 'inform' us?. (June 2021)

• Record Type: Journal Article
• Authors: Mol, Ben W.; Li, Wentao; Lai, Shimona; Stock, Sarah
• Is Part Of: European journal of obstetrics, gynecology, and reproductive biology. Volume 261(2021)
• Page Start: 144
• Page End: 147
• Publication Date: 2021-06
• Subjects: Antenatal corticosteroids -- Data integrity; Obstetrics -- Periodicals; Gynecology -- Periodicals; Reproductive health -- Periodicals; Gynecology -- Periodicals; Obstetrics -- Periodicals; Reproduction -- Periodicals; Obstétrique -- Périodiques; Gynécologie -- Périodiques; Reproduction -- Périodiques; Verloskunde; Gynaecologie; Voortplanting (biologie); Gynecology; Obstetrics; Reproduction; Electronic journals; Periodicals; Electronic journals; 618.05
• DOI: 10.1016/j.ejogrb.2021.04.031
• Languages: English
• ISSNs: 0301-2115

Abstract

Abstract: Randomized controlled trials (RCTs) are a cornerstone for the assessment of the effectiveness of interventions. Appropriate randomization, design, sample size, statistical analyses, and conduct that reduces the risk of bias, enhance the chance they will deliver true research findings. The credibility of RCTs is difficult to assess without objective evidence of compliance with Good Clinical Practice standards. Remarkably, no mechanisms are in place both in the initial peer review process and during meta-analysis to assess these, and little guidance on how to assess data where research integrity cannot be confirmed (e.g. where data originated from a setting without established infrastructure or an era preceding current standards). We describe the case of the use of antenatal steroids. When these drugs are used in early preterm birth, their benefits outweigh the harms. However, later in pregnancy, and specifically at term, this balance is less clear. We describe that the four randomised clinical trials that inform clinical practice through the Cochrane meta-analysis, for various reasons, lack clear governance which makes it difficult to verify provenance and reliability of the data. We conclude that transparency and assessment of data credibility need to be inbuilt both at the time of publication and at the time of meta-analysis. This will drive up standards and encourage appropriate interpretation of results and the context from which they were derived.