Most published meta-analyses were made available within two years of protocol registration

Abstract

IntroductionMeta-analyses are considered the highest form of evidence and are, therefore, important for clinical guidance. Fast publication times are desirable as they ensure that up-to-date knowledge is available for the consumers. This study aimed to determine the time from registering the protocols for meta-analyses in PROSPERO until the results were publicly available either via ePub ahead of print or journal publication. MethodsThis observational database study was reported according to the RECORD guideline. The biomedical databases PubMed and EMBASE were searched on 11th March 2020 for published meta-analyses that were registered in PROSPERO. The meta-analyses were linked to their PROSPERO registrations via their 14-digit PROSPERO registration number in the abstract. Registration date, publication date, and characteristics of reviews were extracted from PubMed, EMBASE, and PROSPERO during March 2020. ResultsWe included 2475 published meta-analyses. The median publication time was 16 months, interquartile range 10–24 months. After publication, 26% had updated the review's status in PROSPERO to published. Fewer authors were associated with shorter publication times. No difference was seen in publication times between the ‘big five’ journals and other journals. ConclusionsOf all published meta-analyses with PROSPERO registrations, 75% were published within two years of the registration of their protocol. Few authors updated the status of their PROSPERO registrations after publishing, which may make identifying abandoned reviews difficult.