This survey of systematic reviews evaluating physiotherapy interventions was conducted to estimate the proportion of reviews that have a registered protocol, compare the methodological quality of registered and unregistered reviews, and calculate the prevalence of outcome reporting bias in registered reviews. A random sample of 150 systematic reviews published in 2015, written in English, Italian, Portuguese or Spanish, and indexed in the PEDro evidence resource were evaluated. Protocol registration was determined by searching the full-text of the reviews, searching protocol registries (PROSPERO and Cochrane Database of Systematic Reviews), and contacting authors. Two independent raters evaluated methodological quality using the AMSTAR checklist and extracted data about the methods used from both the reviews and registered protocols. Only 19% (n=29) of reviews were registered. Registered reviews demonstrated significantly higher methodological quality (median 8/11) than unregistered reviews (median 5/11). One-third (n=9) of the registered reviews demonstrated discrepancies between the protocol and review results, with no evidence that such discrepancies were applied to favor the statistical significance of the intervention (relative risk 1.16; 95% confidence interval 0.63 to 2.12). A low proportion of physiotherapy systematic reviews are registered. The registered systematic reviews showed high methodological quality without evidence of outcome reporting bias. Further strategies should be implemented to encourage registration.
Oliveira CB, et al. A low proportion of systematic reviews in physical therapy are registered: a survey of 150 published systematic reviews. Braz J Phys Ther 2018;22(3):177-83