Comparing the Overall Result and Interaction in Aggregate Data Meta-Analysis and Individual Patient Data...

Area of Science:

• Medical research methodology
• Biostatistics
• Evidence synthesis

Background:

• Aggregate data meta-analyses (ADMAs) and individual patient data meta-analyses (IPDMAs) are crucial for synthesizing research findings.
• Understanding the concordance and differences between these two meta-analysis types is vital for accurate interpretation of evidence.

Purpose of the Study:

• To compare the agreement in overall results between ADMAs and IPDMAs.
• To assess the frequency of interaction detection in subgroup analyses within both ADMA and IPDMA.

Main Methods:

• Identified ADMA articles published before matching IPDMA articles on similar research topics.
• Defined agreement in overall results based on the direction of the effect estimate.
• Compared the number of subgroup analyses and significant interactions reported in both meta-analysis types.

Main Results:

• High agreement (91.7%) was observed in the overall effect between paired meta-analyses.
• IPDMAs conducted significantly more subgroup analyses (634 vs. 150) and identified more interactions (44 vs. 3) than ADMAs.
• A higher proportion of IPDMAs (59.3%) performed subgroup analyses compared to ADMAs (26.0%).

Conclusions:

• ADMAs generally align with IPDMAs on overall findings, particularly with improved methodology.
• IPDMAs are essential when subgroup interactions are suspected due to their superior ability to detect them.
• Methodological enhancements in ADMAs can improve their reliability, but IPDMAs remain critical for exploring complex interactions.