To the Editor.
—The study by the Northern New England Cardiovascular Study Group1 is a good example of the use of disease registry data for an epidemiologic experiment to assess the effectiveness of medical care improvement in changing the outcome. In this study of patients undergoing coronary artery bypass graft (CABG) surgery in 5 medical centers, the investigators compared case-fatality (disease- or procedure-specific in-hospital mortality) rates before and after a set of interventions—feedback of the outcome data to the surgeons, training of the medical and nursing staff, and learning from each other by observing the processes of CABG surgery at each site. After a 16-month intervention period, a 24% to 25% reduction in case fatality over 2 years was attributed to the interventions.1 However, the usefulness of the study findings was limited due to its weak design and nonspecific interventions.Since it was not a controlled trial, the study could not
Sheikh K. What Do We Learn From Disciplined Practice Learning?. JAMA. 1996;276(6):447. doi:10.1001/jama.1996.03540060023012