Reporting of funding and conflicts of interest improved from preprints to peer-reviewed publications of biomedical research.

BACKGROUND AND OBJECTIVES: To assess changes in the reporting of funding and conflicts of interest (COI) in biomedical research between preprint server publications and their corresponding versions in peer-reviewed journals. METHODS: We selected preprint servers publishing exclusively biomedical research. From these, we screened articles by order of publication date and identified 200 preprints first published in 2020 with subsequent versions in peer-reviewed journals. We judged eligibility and extracted data about authorship, funding, and COI in duplicate and independently. We performed descriptive statistics. RESULTS: A quarter of the studies added at least one author to the peer-reviewed version. Most studies reported funding in both versions (87%), and a quarter of these added at least one funder to the peer-reviewed version. Eighteen studies (9%) reported funding only in the peer-reviewed version. A majority of studies reported COI in both versions (69%) and 5% of these had authors reporting more COI in the peer-reviewed version. A minority of studies (23%) reported COI only in the peer-reviewed version. None of the studies justified any changes in authorship, funding, or COI. CONCLUSION: Reporting of funding and COI improved in peer-reviewed versions. However, substantive percentages of studies added authors, funders, and COI disclosures in their peer-reviewed versions.

Policies of biomedical preprint servers on conflicts of interest, authorship, and research integrity lacked important details.

OBJECTIVES: To assess the policies of biomedical preprint servers on the reporting of funding, conflict of interest (COI), author contributions, and research integrity. STUDY DESIGN AND SETTING: We generated a list of potentially eligible preprint servers then judged their eligibility in duplicate and independently. Between July and September 2021, we extracted information from the websites of these servers in duplicate and independently and through a simulated submission. RESULTS: We included 37 preprint servers. A variable percentage of these servers had policies for reporting of funding (43%), disclosure of COI (78%), specification of authors' contributions (54%), and research integrity (76%). While 97% of the COI policies required authors to disclose their own interests, only 7% addressed interests related to family members. About a third of authors' contribution policies (30%) did not specify the types of contributions to report. While a majority of research integrity polices described screening checks for plagiarism (71%), a minority described screening checks for fabrication (39%), falsification (36%), and image manipulation (32%). CONCLUSION: Less than half of biomedical preprint servers have policies on reporting on funding. Policies related to COI disclosure, reporting of author contributions, and research integrity lack important details.