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

Systematic summaries of the available evidence are a fundamental component in achieving optimal health outcomes.1 Traditional evidence hierarchies place systematic reviews and meta-analyses at their pinnacle. Meta-analyses (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 meta-analysis’, 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] meta-analysis’, where the unit of analysis is individual participants that are clustered within individual studies. IPD meta-analyses 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 IgE-mediated 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 between-study heterogeneity, allowed more sophisticated statistical analyses and could reduce research waste. Given these advantages, should IPD MA be the new norm of evidence synthesis? Table 1 details some high-level considerations for reviewers considering embarking on an aggregate MA versus IPD MA. Other factors and explanations are detailed elsewhere.3-5 Probably, the most relevant factor will be the added resource implications required with IPD MA over aggregate MA. The added time (likely a year or more), staff and costs (thousands of pounds) are all likely large enough barriers to dissuade most investigators. Guideline developers are likely to fund multiple questions using aggregate MA methods rather than focusing on their time and money, likely on fewer or only one question, for an IPD MA. Most IPD MAs are, therefore, likely to be the academic pursuit of select highly invested investigators. Such pursuits can be transformative. This is borne out by exceptional examples in other fields of medicine such as fluid resuscitation in sepsis,6 or the role of corticosteroids and IL-6 inhibitors for severe COVID-19.7 New funding mechanisms and dedicating portions of clinical trial budgets are clearly needed to reduce barriers to conducting IPD MAs and thereby allowing them to routinely inform optimal patient care. Beyond funding, trialist teams must pair with expert systematic reviewers and methodologists since IPD MA blends elements of both. 0.5 FTE statistician,2 1.0 FTE Project manager,2 2nd reviewer £60,000+ per year (estimated from Vogt et al2 and Imperial College human resources websites) Collaborator meetings Travel beyond that expected for aggregate MA Considering the value of IPD MA in reducing research waste highlights a big problem in the allergy-immunology field. Registered 5 years ago, the SCiPAD IPD MA was prospectively planned and benefitted from a large group of academic investigators that were independently conducting investigator-initiated clinical trials to answer a common question, were open to collaboration and had strong evidence synthesis methodologic support. Most studies in the field of allergy and immunology, however, are not clinical trials (i.e. on average, less robust data collection standardization, oversight and storage); some investigators may be unwilling to share (e.g. perception of competition, politics, ego, intellectual property concerns and no funding); most studies are industry-sponsored rather than investigator-initiated (i.e. industry typically owns the study participant data, not the investigator); obtaining IPD from industry for meta-analysis typically faces multiple barriers (e.g. no easy point of contact; scrutiny by company staff and lawyers; investigators often necessarily supplying entire protocol with no guarantee of confidentiality or receiving data) and is often neither timely nor successful; the previous poster child of evidence-based medicine, the systematic review, faces erosion by an explosion of duplicitous, methodologically weak, highly conflicted and uncredible systematic reviews. To achieve optimal population health, and to appropriately honour the patients and participants of research studies, evidence ecosystems must actively seek open science and a culture of sharing.8 The SCiPAD investigators’ success is emblematic of what the allergy-immunology field can do and what it needs to do more of. The authors declare the non-financial (intellectual) role as Chair, Evidence in Allergy Group, McMaster University, and developed GRADE guidelines. DC drafted the manuscript, reviewed it, and approved it.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.