Trends of Randomized Clinical Trials Citing Prior Systematic Reviews, 2007-2021
Yuanxi Jia,Bingli Li,Zhirong Yang,Fuxiao Li,Ziyi Zhao,Chang Wei,Xuhao Yang,Qianyi Jin,Di Liu,Xin Wei,Jennifer Yost,Hans Lund,Jinling Tang,Karen A. Robinson
DOI: https://doi.org/10.1001/jamanetworkopen.2023.4219
2023-01-01
JAMA Network Open
Abstract:ImportanceSystematic reviews can help to justify a new randomized clinical trial (RCT), inform its design, and interpret its results in the context of prior evidence.ObjectiveTo assess trends and factors associated with citing (a marker of the use of) prior systematic reviews in RCT reports.Design, Setting, and ParticipantsThis cross-sectional study investigated 737 Cochrane reviews assessing health interventions to identify 4003 eligible RCTs, defined as those included in an updated version but not in the first version of a Cochrane review and published 2 years after the first version of the Cochrane review was published.Main Outcomes and MeasuresThe primary outcome was the citation of prior systematic reviews, Cochrane or others, as determined by screening references of eligible RCTs. Factors that may be associated with the citation of prior systematic reviews were also examined.ResultsAmong 4003 eligible RCTs, 1241 studies (31.0%) cited Cochrane reviews, 1698 studies (42.4%) cited prior non-Cochrane reviews, and 2265 studies (56.6%) cited either type of systematic review or both; 1738 RCTs (43.4%) cited no systematic reviews. The percentage of RCTs citing prior Cochrane reviews, non-Cochrane reviews, and either or both types of review increased from 28 studies (15.3%), 46 studies (25.1%), and 65 studies (35.5%) of 183 RCTs before 2008 to 42 studies (40.8%), 65 studies (64.1%), and 73 studies (71.8%) of 102 RCTs since 2020, respectively; the annual increases were 1.9% (95% CI, 1.4%-2.3%), 3.3% (95% CI, 2.9%-3.7%), and 3.0% (95% CI, 2.5%-3.5%), respectively. The proportion of RCTs citating prior systematic reviews varied considerably across clinical specialties, ranging from 28 of 106 RCTs (26.4%) in ophthalmology to 386 of 553 RCTs (69.8%) in psychiatry (P < .001). RCTs with 100 participants or more (risk ratio [RR], 1.16; 95% CI, 1.03-1.30), nonindustry funding (RR, 1.43; 95% CI, 1.27-1.61), and authors from high-income countries (RR, 1.10; 95% CI, 1.03-1.17) were more likely to cite systematic reviews than those with fewer than 100 participants, industry funding, and authors from low- and middle-income countries, respectively. A journal requirement to cite systematic reviews was not associated with the likelihood of citing a systematic review.Conclusions and RelevanceThis study found that the citation of prior systematic reviews in RCT reports improved over time, but approximately 40% of RCTs failed to do so. These findings suggest that reference to prior evidence for initiating, designing, and reporting RCTs should be further emphasized to assure clinical relevance, improve methodological quality, and facilitate interpretation of new results.