Completeness of reporting in abstracts of randomized controlled trials and factors associated with complete reporting: a meta-research study

The CONSORT statement for reporting abstracts of randomized controlled trials (RCTs) provides a list of essential items to report to ensure reliable interpretation. We evaluated the completeness of reporting of RCT abstracts published in high-impact-factor journals and examined factors associated with adherence to the CONSORT for abstracts. On March 5, 2024, we searched PubMed for abstracts of RCTs published in the top five journals (ISI Web of Knowledge 2023) in 22 medical fields, including Medicine, General & Internal, from July 1, 2023, to February 29, 2024. We assessed the completeness of reporting of the 15 items of the CONSORT for abstracts in duplicate via a standardized data extraction form. The primary outcome was the mean number of items adequately reported. The prespecified factors potentially associated with adherence to CONSORT for abstracts included study-specific factors (income category of the country, funding type, and sample size) and journal-specific factors (type of journal [specialized vs. general medicine], CONSORT endorsement, abstract word count limit, and structured abstract). Among 1267 citations, 610 abstracts published in 82 distinct journals were included. The mean (SD) number of items adequately reported was 10.1 (2.5). Among the 610 abstracts included, the quality of reporting varied across items; some items were adequately reported (97% (n=594) for participants, 100% (n=610) for interventions, 96% (n=585) for objectives) and others were poorly reported (29% (n=175) for trial design, 36% (n=221) for funding sources, 32% (n=197) for blinding and 13% (n=81) for randomization). The multivariate linear-mixed effects model showed that a sample size above 100, the CONSORT endorsement and Medicine, General & Internal journals were associated with a higher mean number of items adequately reported. Reporting of RCT abstracts remains incomplete particularly for key items including trial design, random sequence generation, blinding, and funding source disclosure. CONSORT endorsement, publication in a general medical journal and larger sample size were associated with better adherence to CONSORT for abstracts.