TY - JOUR T1 - What is a performance outlier? JF - BMJ Quality & Safety JO - BMJ Qual Saf SP - 95 LP - 99 DO - 10.1136/bmjqs-2015-003934 VL - 24 IS - 2 AU - David M Shahian AU - Sharon-Lise T Normand Y1 - 2015/02/01 UR - http://qualitysafety.bmj.com/content/24/2/95.abstract N2 - Healthcare performance measurement is a complex undertaking, often presenting a number of potential alternative approaches and methodological nuances. Important considerations include richness and quality of data sources; data completeness; choice of metrics and target population; sample size; patient- and provider-level data collection periods; risk adjustment; statistical methodology (eg, logistic regression vs hierarchical models); model performance, reliability and validity; and classification of outliers. Given these many considerations, as well as the absence of nationally accepted standards for provider profiling, it is not surprising that different rating organisations and methodologies may produce divergent results for the same hospitals.1–6Outlier classification, the last step in the measurement process, has particularly important ramifications. For patients, it may lead them to choose or avoid a particular provider. For providers, outlier status may positively or negatively impact referrals and reimbursement, and may influence how scarce hospital resources are deployed to address putative areas of concern. Misclassification is probably more common than generally appreciated. For example, partitioning of hospitals (eg, terciles, quartiles, quintiles, deciles) to determine outliers may lead to excessive false positives—hospitals labelled as having above or below average performance when, in fact, their results do not differ significantly from the mean based on appropriate statistical tests.7 ,8 In this issue, Paddock et al9 address a seemingly straightforward question—what precisely does it mean to be a performance outlier? Using Hospital Compare data, the authors demonstrate an apparently contradictory finding. When directly compared one to another, some individual hospitals in a given performance tier may not be statistically significantly different than individual hospitals in adjacent tiers, even when those tier assignments were made using appropriate tests of statistical significance. For instance, Paddock et al9 show that for each bottom-tier hospital, there was at least one mid-tier hospital with statistically indistinguishable performance. Among … ER -