Methodological Rigor and Quality of Reporting of Clinical Trials Published with Physical Activity Interventions: A Report from the Strengthening the Evidence in Exercise Sciences Initiative (SEES Initiative)

Abstract

Background In response to the growing number of published randomized clinical trials (RCTs), efforts have been made to promote transparent and reproducible research practices. However, limited evidence exists to demonstrate the level of adherence to established reporting and methodological standards. This study aimed to evaluate the methodological standards and quality of reporting in RCTs of physical activity (PA) interventions. We aimed to present Strengthening the Evidence in Exercise Sciences Initiative (SEES Initiative) 2020 assessments results. Methods RCTs of PA advice or exercise interventions published in 2020 were selected. Monthly searches were conducted on PubMed/MEDLINE targeting six top-tier exercise science journals. Assessments were conducted by two independent authors, based on 44 items originally from CONSORT and TIDieR reporting guidelines. These items were divided into seven domains: transparency, completeness, participants, intervention, rigor methodology, outcomes and critical analysis. Descriptive analysis was performed using absolute and relative frequencies, and exploratory analysis was done by comparing proportions using the X2^test (α = 0.05). Results Out of 1,766 RCTs evaluated for eligibility, 53 were included. The median adherence to recommended items included among the studies was 30, ranging from 18 (41%) to 44 (100%) items in individual study assessments. The items presenting full adherence among the studies were related to intervention description, justification, outcome measurement, effect sizes, and statistical analysis. The least reported item referred to mentioning unplanned modifications during trials, appearing in 6 (11.3%) studies. Out of the 53 RCTs, 36 (67.9%) reported having a registration, and these studies demonstrated higher adherence to the assessed items compared to non-registered studies. Conclusions In general, items related to critical analysis were more thoroughly described, such as hypotheses and statistical analyses. However, aspects more related to transparency aspects, such as protocol registrations/modifications and intervention descriptions, were suboptimally reported. Finally, this study highlights the need to promote the existing resources concerning quality of reporting and transparent research practices for investigators and editors in the exercise sciences discipline.