Prevalência e qualidade dos protocolos de registro de ensaios clínicos na fisioterapia

Abstract

OBJECTIVE: To investigate the completeness of clinical trial registration and the extent of selective reporting of outcomes in randomized trials of physical therapy interventions published in 2019 and to compare them with equivalent data from 2009. METHODS: A random sample of 200 randomized trials was obtained from the Physiotherapy Evidence Database (PEDro) platform. Data about registration were obtained from each of the randomized trials through their registered protocol and through direct contact with the authors and were compared with the data available from the published report. We evaluated the percentages of registered trials; of prospective registration (trials registered prior to the start of enrollment); unambiguous primary outcomes (definition and time point of assessment clearly defined in the registered protocol); adequate registration (prospective registration with an unambiguous primary outcome); inconsistency of outcomes (a mismatch between the published report and registered protocol); selective reporting of outcome (only registry entries with unambiguous outcomes are taken into account); whether clinical trials published in journals of higher impact factor had a higher prevalence of registries in contrast to the general sample; and which factors were associated with clinical trial registration. RESULTS: The proportion of clinical trials registered in 2019 was 126/200 (63%) versus 67/200 (34%) in 2009 (Relative Risk – RR 1.88, 95% CI 1.51 to 2.35). Randomized trials published in high-impact factor journals had a significantly higher prevalence of registration, 100% in 2019 (RR 1.58, 95% CI 1.42 to 1.76) than the general sample from the same year. In 2019, the prospective registry had a prevalence of 35/200 (18%) in contrast to 12/200 (6%) in 2009 (RR 2.91, 95% CI 1.56 to 5.45). Unambiguous primary outcomes were present in 60 clinical trials in 2019. Adequate registration occurred in 15/200 (8%) of the 2019 sample, compared to 5/200 (3%) in 2009 (RR 3.00, 95% CI 1.11 to 8.10). In 2019, 44 clinical trials (73%), out of 60 that allowed the evaluation, evidenced the selective reporting of data, against 23/49 (47%) in 2009 (RR 1.56, 95% CI 1.12 to 2.18). CONCLUSION: The registration of clinical trials of physical therapy interventions has increased in the last decade, but it is still not adequately performed. Therefore, the benefits of clinical trial registration are not yet being fully achieved. These data indicate that a complementary effort is still needed for clinical trials of physical therapy interventions to be performed correctly.