pointsSeven existing criteria help clinicians assess the credibility of putative subgroup effects on a continuum from "highly plausible" to "extremely unlikely" We suggest four additional criteria: subgroup definition on the basis of baseline characteristics, independence of the subgroup effect, a priori specification of the direction of the subgroup effect, and consistency across related outcomes We propose a re-structured checklist of items addressing study design, analysis, and context

I nvestigators who conduct randomized controlled trials (RCTs) and meta-analyses of RCTs often perform analyses of effect modification to assess whether intervention effects might vary by another variable such as age, disease severity or, in a meta-analysis, study setting or year of study. 1-14 The terminology varies; Box 1 presents the alternatives currently in use. Investigators sometimes make claims that an effect modification is present. Literature surveys suggest that 14%-26% of RCTs and meta-analyses emphasize at least 1 potential effect modification in their abstract or discussion. 4-9,11 The interest in effect modification is understandable: if patients with differing characteristics respond differently to the same intervention, the overall effect estimate is misleading for some, if not all, patients. Identifying situations in which true variation in effects exist is important, and the notion of tailoring therapy to patients has enormous appeal. Moreover, the opportunities for analyzing effect modification grow with the increasing number of newly developed diagnostic and genomic markers. However, mistaken claims of effect modification may compromise optimal patient care, and many claims of effect modification have subsequently proved spurious. 15-24 Applying a mistaken claim of effect modification may cause harms through administration of ineffective treatment or may lead to patients' being denied beneficial therapies, and will likely increase health care costs. Numerous theoretical analyses and simulation studies show that the fundamental reason for misleading claims of effect RESEARCH Development of the Instrument to assess the Credibility of Effect Modification Analyses (ICEMAN) in randomized controlled trials and meta-analyses