Quality of trial reporting: not great, but getting better slowly

Last week, updated recommendations for reporting randomized trials — the CONSORT 2010 guidelines — were published in PLoS Medicine and other medical journals (BMJ, Lancet, Obstetrics and Gynecology, Annals of Internal Medicine, Open Medicine, Journal of Clinical Epidemiology, BMC Medicine, and Trials).

However, a recently published research study describes the poor quality of reporting of many randomized trials in the medical literature, despite 14 years having passed since the original development and endorsement of the CONSORT guidelines. In the study, a group of researchers from the Universities of Oxford and Toronto (many of whom have also been involved in CONSORT) compared reporting quality between two series of randomized controlled trials published in December 2000 and December 2006. Unsurprisingly, most of the 1135 trials were published in specialist journals, and had  small sample sizes – with a median of 80 participants in each trial. The study demonstrated poor uptake of CONSORT, with fewer than half (44%) of the trials published in December 2006 appearing in CONSORT-endorsing journals. The researchers noted improvements between 2000 and 2006 in several aspects of reporting quality, for example in detailing the trial primary outcome measure (risk ratio for improvement 1.18 (95% CI 1.04 – 1.33); describing sample size calculation (RR 1.66 (95% CI 1.40 – 1.95); in reporting methods of randomization sequence generation (RR 1.62 (95% CI 1.32 – 1.97), and in reporting allocation concealment (RR 1.40 (95% CI 1.11 – 1.76). Reporting quality for specific items was found to be better in CONSORT-endorsing versus non-endorsing journals.

The authors concluded by saying that “the quality of reporting remains well below an acceptable level”, and propose that more journals should endorse and mandate use of the CONSORT tools.

This entry was posted in General, Media, Policy and tagged , , , . Bookmark the permalink.

Comments are closed.