The authors replicate the methods of the Harvard Medical Practice Study (1984), a seminal study of adverse events in hospitals, in this 1992 study of hospitalizations in Colorado and Utah. Using carefully controlled retrospective medical record review, the authors found an overall adverse event rate of 2.9% of hospitalizations, with approximately 30% resulting from apparent negligence. The authors outline the rates of error by medical specialty, error type, and medication class (for errors involving medications). The authors discuss the relationship of their study to the Harvard Medical Practice Study and the 1995 Quality in Australian Health Care study and briefly review the literature on improving hospital safety.