Analysis of non-prospective trial registration in clinical trials submitted to The BMJ: observational study

Abstract

Objectives: To identify variables associated with non-prospective registration of clinical trials submitted to The BMJ; examine deficiencies in registration, disclosure of deficiencies, and subsequent publication status of such trials; and assess the authors' claims of prospective registration.

Design: Observational study.

Setting: The BMJ, London.

Population: 239 of 287 submissions to The BMJ (2019-23) reporting clinical trial results as defined by the International Committee of Medical Journal Editors (ICMJE) and flagged by editors as potentially not prospectively registered, and 239 trials prospectively registered in an ICJME accepted registry from a randomised list of all clinical trial submissions in the same period.

Main outcome measures: Study outcomes included non-prospective registration in an ICJME accepted registry and, for non-prospectively registered trials, deficiencies in registration deficiencies in registration (retrospective in an ICJME accepted registry, registration in a non-ICJME accepted registry, or no registration), delay in registration of retrospectively registered trials, and subsequent publication status. Time to publication and disclosure of registration deficiencies on publication were also assessed, along with the journal's impact factor and whether the journal claimed adherence to the ICMJE registration recommendations.

Results: Reduced odds of non-prospective registration in an ICMJE accepted registry were associated with larger sample sizes (101-500 v 1-100; odds ratio 0.43, 95% confidence interval 0.22 to 0.84), corresponding authors from Oceania (reference: Europe; 0.35, 0.14 to 0.82), a greater number of authors (10 v 5; 0.71, 0.59 to 0.87), mention of Consolidated Standards Of Reporting Trials (0.22, 0.06 to 0.67), more recent submissions (2021-23 v 2019-20; 0.63, 0.42 to 0.95), and presence of funding (eg, non-profit v no funding or no information; 0.20, 0.09 to 0.41). Higher odds were observed for corresponding authors based in Asia (reference: Europe; 1.75, 1.07 to 2.89). Of 176 trials not prospectively registered in an ICMJE accepted registry submitted between 2019 and 2021, 146 (83%) were retrospectively registered (median delay 193 days), 23 (13%) were unregistered, and seven (4%) were registered in a non-ICJME accepted registry. Most (155 trials, 88%) were subsequently published; 138 (89%) of these appeared in journals with an impact factor (median 5.39) and 96 (62%) in journals claiming adherence to the ICMJE recommendations. The median time from initial submission to The BMJ to publication was 12 months. Only about one sixth explicitly disclosed registration deficiencies on publication. Of 72 authors responding about prospective registration on submission, 60 (83%) incorrectly claimed adherence.

Conclusions: Many trials rejected by The BMJ for non-prospective registration in an ICMJE accepted registry were later published in high impact journals claiming adherence to the ICMJE recommendations, often without disclosure of registration deficiencies.

Fig 1

Forest plot of variables associated with non-prospective registration in an ICMJE accepted registry (multivariable model). CI=confidence interval; CONSORT=Consolidated Standards Of Reporting Trials; ICMJE=International Committee of Medical Journal Editors