Completeness of reporting of randomised controlled trials including people with transient ischaemic attack (TIA) or stroke: a systematic review

Article type
Authors
Wilson B1, Burnett P2, Moher D3, Altman D4, Al-Shahi Salman R5
1University of Edinburgh
2Edinburgh Royal Infirmary
3Centre for Journalology, Clinical Epidemiology Program, Ottawa Hospital Research Institute
4Centre for Statistics in Medicine, University of Oxford
5Centre for Clinical Brain Sciences, University of Edinburgh
Abstract
Objective: To assess the adherence of stroke randomised controlled trials (RCTs) to CONSORT reporting guidelines, and investigate the factors that are associated with completeness of reporting.

Design: Two reviewers assessed the published report of the final primary results of stroke RCTs with a 10-point truncated CONSORT reporting checklist, to investigate adherence over time, univariable associations, and independent associations with total CONSORT reporting score in a multiple linear regression model.

Data sources: Random sample of stroke RCTs from the Cochrane Stroke Group’s Trial Register.

Eligibility criteria: Primary published report of the final results of transient ischaemic attack (TIA) or stroke RCTs, published in English in 1997-2016 inclusive.

Results: In 177 stroke RCTs, the mean score on the truncated CONSORT checklist was 5.8 (SD 2.2); reporting improved from 1997-2000 (4.9 SD 2.0) to 2001-2009 (5.8 SD 2.1) and to 2010-2016 (6.8 SD 2.1). A higher CONSORT score was independently associated with publication during epochs following a revision of CONSORT reporting guidelines (p