In the last decade the emerging activities in the field of quality and safety improvement in health care are resulting in new practices of governing medicine and posing challenges to prevailing notions of what it means to be a good doctor, patient, manager, or even health-care system. Following the seminal work of Wennberg and Gittelsohn (1973), who analyzed high variation in treatment patterns in neighboring communities in New En gland that could not be explained by clinical differences or "case mix" (Wennberg 1984), and based on reported adherence rates to clinical guidelines by medical professionals of approximately 50 percent, institutions for healthcare improvement have created an awareness among patients, policy makers, and clinicians that receiving treatment is actually a risky business.1 Health care is not simply a domain where patients are cured or cared for: it is a field full of dangers, and hospitals in particular are increasingly seen as risky places.