Methods and Outcomes Reporting in Exercise-Based Cardiovascular Rehabilitation Research: a cross-sectional analysis

Abstract

AbstractThe poor reporting quality of methods and outcomes is relatively recognized in the biomedical field. Its prevalence and implications have been studied in the cardiovascular rehabilitation literature but not so extensively in exercise-based trials. Our main objective was to cross-sectionally estimate the prevalence of both methodological and outcome reporting items in CVR trials with EBI. We also searched for associations (secondary outcomes) between (1) the effect size reported and the direction of the primary outcome, as well as (2) associations with the frequency of Spin. We cross-sectionally screened the sample of eligible trials dated between 2017 and 2021, and then collected the prevalence of methodological and outcome characteristics, independent and blinded manner. Our study shows that there was an insufficient reporting of methods and outcomes. Also, studies reporting effect size measures had a lower chance of Spin. The primary outcome effect size was not reported in 35% of the studies SES. However, more than 2/3 of the sample (69%) had a statement in the discussion or conclusion sections mentioning clinical relevance or meaningful benefit of the statistically significant results. Selective outcome reporting has important implications for translating science into practice, once not so threatens the validity of an intervention effectiveness, but also frustrates the use of its evidence in meta-analyses.What is new?Our study shows that randomized controlled trials with cardiovascular rehabilitation based on exercise insufficiently reported the various methods and outcomes characteristics.Although nearly 70% studies had stated its outcomes as clinically meaningful within our sample, about 41%of the studies clearly stated the primary outcome confidence intervals.More than half of the sample presented at least one spin in the results section, and studies reporting effect size measures had a lower chance of Spin.40% of the studies within our sample did not report a priori sample size calculation, with 1/4 not stating the number of randomized subjects that could meet the intended power.We did not find any associations regarding the direction of the results (positive or negative) and the prevalence of spin, contrary to what have been found in the literature.