Joshua David Wallach
Yale School of Public Health
Co-Authors: Patrick Liu, John PA Ioannidis, Joseph S Ross, Sanket S Dhruva, Anita T Luxkaranayagam, Vasilis Vasiliou, Joshua D Wallach
Age-treatment subgroup analyses in Cochrane intervention reviews: a meta-epidemiological study
Objective: To characterize the conduct and clinical translation of age subgroup analyses in a large random sample of Cochrane intervention reviews of randomized trials (n=928). Methods: We identified and abstracted characteristics of all planned and/or conducted age subgroup analyses and searched for evidence of inclusion of statistically significant results in commonly used clinical management resources. Results: Among 928 reviews, 189 (20.4%) included plans to conduct age subgroup analyses. Most (162, 85.7%) were not conducted, commonly because of insufficient trial data. 20 (2.2%) reviews conducted a total of 65 analyses with non-overlapping subgroup levels. Eight (8/65,12.3%) reported P < 0.05 from formal interaction testing; however, none were discussed or included in common clinical practice resources. Conclusion: When subgroup analyses are performed, authors should report the findings, compare the results to previous studies, and outline any potential impact on clinical care.