Reporting quality of trial abstracts-improved yet suboptimal: A systematic review and meta-analysis

We conducted a systematic review and meta-analysis of literature to determine if the publication of the Consolidated Standards of Reporting Trials (CONSORT) abstract guideline in 2008 was followed by change in reporting quality of randomized controlled trial (RCT) abstracts.

Evaluations were included if they compared reporting quality of RCT abstracts before and after the publication of CONSORT-abstract guideline. The literature search was performed (January 2008 to April 2017) in Medline (Ovid), EMbase, CINAHL plus and Cochrane methodologies register. We assessed study validity with a special validity tool, adapted from a previous Cochrane review.

Initial search identified 4142 articles, of which total 10 evaluations including 5184 abstracts were included. Total 22 outcomes related to individual items of CONSORT-abstract guideline were assessed, and 14 showed significant effect sizes favoring CONSORT-abstract guideline. Despite significant effect size, the overall post-CONSORT reporting (PCR) was suboptimal for ten items: title (RR = 1.40, 95% CI 1.23 to 1.59, PCR = 53.4%), participants (RR = 1.58, 95% CI 1.11 to 2.26, PCR = 24.5%), primary outcome (RR = 1.12, 95% CI 1.02 to 1.23, PCR = 65%), blinding (RR = 2.13, 95% CI 1.20 to 3.76, PCR = 13.9%), trial status (RR = 1.81, 95% 1.39 to 2.35, PCR = 10.6%), numbers analyzed (RR = 1.51, 95% CI 1.15 to 1.98, PCR = 26.5%), outcome (RR = 1.40, 95% 1.05 to 1.86, PCR = 21.9%), effect size and precision (RR = 1.59, 95% CI 1.15 to 2.19, PCR = 58.9%), harms (RR = 1.24, 95% CI 1.04 to 1.48, PCR = 41.8%), trial registration (RR = 2.02, 95% CI 1.63 to 2.50, PCR = 33.8%). Three items with favorable effect size in addition had wide CIs: randomization (RR = -4.28, 95% CI 1.56 to 11.75, PCR = -3.3%), allocation concealment (RR = -19.89, 95% CI 1.54 to 256.69, PCR = -5.7%), and funding (RR = -22.61, 95% CI 8.13 to 62.67, PCR = -11.32%).

The change in reporting quality of RCT abstracts is far from satisfactory, as evidenced by suboptimal post-CONSORT rates and wide CIs of effect sizes for majority of improved items. Mere publication of CONSORT-abstract guideline, without strict endorsement has failed to translate into good quality abstracts.