The term never event was originally coined to describe rare, devastating, and preventable events like wrong-site surgery or fatal medication errors. This definition has expanded over time to include a variety of serious adverse events; for some of them (i.e., certain health care–associated infections), the Centers for Medicare and Medicaid Services denies additional reimbursement. This article sought to determine if eight never events (mostly infectious complications of surgery) are truly preventable, by examining whether baseline patient characteristics could predict which patients would experience a never event. The authors found that incidence of most of these complications could be predicted on the basis of preexisting conditions or the specific surgical procedure performed, calling into question whether these events are truly preventable. This study exemplifies research into the "basic science" of patient safety; a prior commentary called for studies focusing on identifying truly preventable harm and developing accurate, reliable measurement standards.