Methodological considerations for observational studies of acute kidney injury using existing data sources.

Acute kidney injury (AKI), as defined by small, often reversible changes in kidney function, has recently been recognized as an important complication in hospitalized patients, and has been consistently associated with prolonged hospital length of stay, increased associated costs and short-term mortality. Research studies on the epidemiology of AKI must address a number of unique methodological challenges, which have the potential to impact study results and validity. This review explores several methodological issues relevant to the design and conduct of observational studies that employ preexisting laboratory, administrative or research databases and that examine AKI as an outcome or an exposure. We discuss how methodological decisions may affect study results, particularly as they relate to selection bias, misclassification and confounding. Highlighting these areas may facilitate the design of studies of high methodological rigor that advance our understanding of AKI.