• Lindsay Horton, Jonathan Rhodes, and Lindsay Wilson.
• 1 Division of Psychology, University of Stirling , Stirling, Scotland, United Kingdom .
• J. Neurotrauma. 2018 Sep 1; 35 (17): 2005-2014.

AbstractAs part of efforts to improve study design, the use of outcome measures in randomized controlled trials (RCTs) in traumatic brain injury (TBI) is receiving increasing attention. This review aimed to assess how clinical outcome assessments (COAs) have been used and reported in RCTs in adult TBI. Systematic literature searches were conducted to identify medium to large (n ≥ 100) acute and post-acute TBI trials published since 2000. Data were extracted independently by two reviewers using a set of structured templates. Items from the Consolidated Standards of Reporting Trials (CONSORT) 2010 Statement and CONSORT patient-reported outcomes (PROs) extension were used to evaluate reporting quality of COAs. Glasgow Outcome Scale/Extended (GOS/GOSE) data were extracted using a checklist developed specifically for the review. A total of 126 separate COAs were identified in 58 studies. The findings demonstrate heterogeneity in the use of TBI outcomes, limiting comparisons and meta-analyses of RCT findings. The GOS/GOSE was included in 39 studies, but implemented in a variety of ways, which may not be equivalent. Multi-dimensional outcomes were used in 30 studies, and these were relatively more common in rehabilitation settings. The use of PROs was limited, especially in acute study settings. Quality of reporting was variable, and key information concerning COAs was often omitted, making it difficult to know how precisely outcomes were assessed. Consistency across studies would be increased and future meta-analyses facilitated by (a) using common data elements (CDEs) recommendations for TBI outcomes and (b) following CONSORT guidelines when publishing RCTs.

11 keywords...

hide…