J Clin Periodontol. 2021 Mar 21. doi: 10.1111/jcpe.13451. Online ahead of print.
ABSTRACT
BACKGROUND: Outcome discrepancies between protocols and respective publications represent a concerning bias. The purpose of this study was to assess the prevalence of selective outcome reporting (SOR) in root coverage randomized clinical trials (RCTs).
METHODS: Published root coverage RCTs (July 2005 to March 2020) were included if a corresponding protocol could be identified in a public registry. Discrepancies between protocol and its correspondent publication(s) were compared regarding primary and secondary outcomes and other study characteristics. Associations between trial characteristics and SOR were evaluated.
RESULTS: Forty four studies (54 publications) were included. The majority of studies (77.3%) were retrospectively registered. SOR was frequent (40.9% of trials) and consisted of primary outcome downgrade (22.7%); secondary outcome upgrade (11.4%); new primary outcome introduced in publication (25%); protocol primary outcome omitted from publication (13.6%) and discrepancy in primary outcome timing (18.2%). SOR was unclear in 20.5% of studies and favoured statistical significance in 12 studies (27.3%). SOR was significantly associated with study significance (p < 0.001) and unclear outcome definition in the publication (p < 0.001). Only a third (32.8%) of primary outcomes were completely defined.
CONCLUSIONS: The present study identified high prevalence of SOR in root coverage RCTs.
PMID:33745136 | DOI:10.1111/jcpe.13451
