Transparency, trustworthiness and usefulness in pre-eclampsia randomised controlled trials in the last three decades.

Abstract

Objectives: Over the last decades, there has been a rising number of randomised controlled trials (RCTs) on pre-eclampsia. We investigated pre-eclampsia RCTs between 1987 and 2021 and reported on trustworthiness, risk of biases, p-values, transparency, and usefulness.

Methods: We searched PubMed for RCTs containing "pre-eclampsia" or "hypertensive disorders of pregnancy" in the title between 1987 and 2021. We created a transparency, trustworthiness and usefulness checklist, and recorded the characteristics of p-value reporting, RCT characteristics including registration, journal of publication, and outcomes.

Results: We found 202 RCTs, with the median number of authors increasing from four in 1987 to 12 in 2021. Trial registration increased nearly seven-fold from 17/93 between 1987-2010 to 79/109 RCTs registered between 2011-2021. Between 2010-2021, 338 p-values were published in the abstract with a rise in non-significant p-values (significant vs non-significant: 214/338, 63% vs 124/338, 37%). The number of effect sizes published increased by 58% from 659 in 1987-2010 to 1,038 in 2011-2021. The median of combined transparency and usefulness scores of RCTs improved from 6 to 9 out of 13, between 1987-1991 and 2017-2021. From the trustworthiness checklist, trials performed worst at adequate registration (43/202, 21%). Risk of reporting, performance, and detection biases were detected in most RCTs (166/202, 82%, 131/202, 65% and 141/202, 70% respectively).

Conclusion: The reporting of non-significant p-values and effect sizes increased after mandatory trial registration in 2010, along with the transparency and trustworthiness of most trials. RCTs need to be more transparent, trustworthy, and useful to increase the value of research.