Systematic reviews with published protocols compared to those without: more effort, older search

Abstract

Objective: To explore trends in published protocols of systematic reviews (SRs) and to analyze how SRs with published protocols differ from those without.

Study design and setting: We searched PubMed up to December 31, 2016 to identify SR protocols. We also searched for the corresponding SR for each protocol published in 2012 and 2013 and matched this with an SR without published protocol by year and journal.

Results: The number of protocols published increased from 42 in 2012 to 404 in 2016; 125 were published in 2012 and 2013. One-third of SRs remained unpublished after 3-5 years. We included 80 SRs with protocols and 80 controls. SRs with protocols reported their methods more comprehensively than their controls, but their median time from search to submission was longer (325 vs. 122 days; P < 0.001). Almost two-thirds of the SRs with protocols and about 10% of the controls could be found in the International Prospective Register of Systematic Reviews (PROSPERO).

Conclusion: Time from search to submission was longer for SRs with published protocols, while at the same time SRs with published protocols were better elaborated and reported. As quality, transparency, and currency are cornerstones of SRs, we suggest critically discussing the current practice of publishing SR protocols.

Keywords: Methodology; PRISMA; PROSPERO; Protocol; Research reporting; Systematic review.