The need for quality assurance in vascular surgery can be deduced from the variability in unruptured abdominal aneurysm operative death rates in a group of patients large enough that factors influencing mortality rates other than quality of care can be controlled. Operative mortality rate for 3570 patients undergoing unruptured abdominal aortic aneurysm repair was determined for all non-Veterans Administration surgeons and hospitals in New York State from 1985 to 1987. The average annual number of aneurysm operations per surgeon was 3.6, and per hospital it was 10.2. Unruptured aneurysm repair mortality for surgeons performing 1 to 5 aneurysm operations per year was 10% whereas for surgeons performing more than 26 aneurysm operations per year it was 6% (p less than 0.0001). Unruptured aneurysm repair mortality for hospitals performing 1 to 5 aneurysm operations per year was 14% and for hospitals performing more than 38 aneurysm operations per year it was 5% (p less than 0.0001). Even when these mortality rates were adjusted for differences in patient age, severity of illness, secondary diagnoses and admission status, significant mortality rate differences persisted: 9% versus 4% for low and high volume surgeons, respectively (p less than 0.001), and 12% versus 5% for low and high volume hospitals, respectively (p less than 0.001). Surgeons who performed more than 7 aneurysm operations per year devoted more of their practice to aortic (11%) and vascular operations (52%) than did surgeons who performed 7 or fewer aneurysm operations per year (2% and 19%, respectively).(ABSTRACT TRUNCATED AT 250 WORDS)