Individual‐patient data and aggregate evidence syntheses and the future of allergy‐immunology research

Abstract

Systematic summaries of the available evidence are a fundamental component in achieving optimal health outcomes.1 Traditional evidence hierarchies place systematic reviews and metaanalyses at their pinnacle. Metaanalyses (MA) can be subdivided into two analytic approaches: those that primarily combine existing published data using the values reported in individual studies, called ‘aggregate data metaanalysis’, where individual trials are a kind of unit of analysis; and those that seek to combine the raw study data from multiple studies, called ‘individual patient data [IPD] metaanalysis’, where the unit of analysis is individual participants that are clustered within individual studies. IPD metaanalyses have been claimed to be the ‘gold standard’ of evidence synthesis. What are the merits of IPD MA and why are investigators not doing more of them? In this issue, Van Vogt, Cro and colleagues, representing the Skincare interventions for the prevention of atopic dermatitis (SCiPAD) collaboration leadership, report a comparison of aggregate MA vs IPD MA of skin care interventions, primarily moisturizers (emollients), vs standard care for the prevention of atopic dermatitis and IgEmediated food allergy in infants.2 Smartly planned, excellently done, spectacularly interpreted and impactfully informative, they report similar effect estimates using both analytic approaches, and the IPD approach better addressed the betweenstudy heterogeneity, allowed more sophisticated statistical analyses and could