To describe a managed care organization's efforts to improve value for its members by forming a coronary services network (CSN). To identify high-quality facilities for its CSN, Anthem Blue Cross and Blue Shield reviewed claims data and clinical data from hospitals that met its general quality standards. An external firm measured and risk-adjusted applicant hospitals' mortality rates. Hospitals that demonstrated superior performance were eligible to join the CSN. In 1996, 2 years after the CSN was formed, clinical outcomes of participants and new applicants were analyzed again by the same external firm. Data on more than 10,000 consecutive (all-payer) inpatients discharged after coronary bypass surgery in 1993 were collected from 16 applicant hospitals using a uniform format and data definitions. This analysis was expanded to 23 participating and applicant hospitals that discharged more than 13,000 patients who underwent either bypass surgery or coronary revascularization in 1995. We compared risk-adjusted routine length of stay (a measure of efficiency), mortality rates, and adverse outcome rates between CSN and non-CSN facilities. From 1993 to 1995, overall length of stay in the network decreased by 20%, from 12.3 to 9.8 days (P < or = 0.01) and severity-adjusted mortality rates decreased by 7.3%, from 2.9% to 2.7%. Initially, facilities outside the network had comparable efficiency but much higher mortality. However, they improved so much in both measures that their severity-adjusted mortality rate for bypass surgery in 1995 was no more than 10% higher than that of CSN hospitals. The creation of a statewide CSN that emphasized and improved the level of performance among providers ultimately benefited the carrier's managed care members. The desirability of participation was evidenced by an increase in the number of applicant hospitals over the 2 years. This may have stimulated quality improvement among competing providers in the region and among CSN facilities themselves.