Effect Modification Analyses in Individual Participant Data Meta-Analyses: A Systematic Review

Area of Science:

• Biostatistics
• Epidemiology
• Clinical Trials

Background:

• Effect modification analyses are crucial for understanding subgroup effects in meta-analyses.
• The prespecification and consistent reporting of these analyses in individual participant data meta-analyses (IPDMAs) remain unclear.

Purpose of the Study:

• To investigate the planning of effect modification analyses in IPDMA protocols.
• To evaluate the concordance between planned effect modification analyses in protocols and their reporting in published IPDMA reports.

Main Methods:

• Systematic review of IPDMA protocols and published reports from Medline, Embase, Cochrane, and PROSPERO.
• Paired reviewers independently extracted data on planning and reporting characteristics of effect modification analyses.
• Comparison of matched protocol-report pairs for concordance in prespecification.

Main Results:

• Most protocols (94.4%) planned effect modification analyses, but few specified direction (3.6%) or handled continuous variables consistently.
• A high proportion of reports (89.8%) included effect modification analyses, yet only 56.4% reported prespecified ones.
• Concordance between protocols and reports was poor (19.3%), with frequent omission of planned and addition of unplanned analyses.

Conclusions:

• Protocols and reports in IPDMAs often fail to prespecify effect modification directions and handle continuous variables appropriately.
• There is poor concordance between planned and reported effect modification analyses, with frequent omissions and additions.
• The findings highlight a need for improved standardization and reporting practices in IPDMAs.