AbstractBackground:Coronary angioplasty (PTCA) offers improved symptom control over medical treatment in patients with stable angina and single‐vessel disease. However, it is uncertain if PTCA is more cost‐effective. Cost‐shifting could also influence the provision of PTCA.Methods:Data from the only randomised trial comparing PTCA to medical therapy (ACME study) were used with costs from an Australian teaching hospital to estimate the costs and freedom from angina in 100 patients over three years. The incremental cost‐effectiveness of PTCA, and the potential for cost‐shifting were also examined.Results:Although the total cost of treating 100 patients over three years with PTCA ($678,978) was higher than a medical strategy ($631,078), PTCA was more cost‐effective ($10,930versus$12,682 per patient free of angina). The incremental cost‐effectiveness of PTCA ($3875 per extra patient free of angina) was also substantially less than the cost of the medical strategy. These should be considered crude estimates as they were based on limited data on resource use. The hospital could reduce costs by pursuing a medical strategy, but 54% of the savings would result from shifting the cost of treating patients to the Federal Government and patients. By performing PTCA on privately insured rather than Medicare patients, die hospital could shift $29,876 per 100 patients to the Federal government and private insurance funds.Conclusions:From society's perspective, PTCA may be more cost‐effective than a medical strategy. However, the financial interests of the hospital are best served by limiting PTCA or restricting PTCA to privately insured patients. Cost‐shifting may have a major impact on the provision of PTCA. The costs of providing medical services need to be weighed against the cost of n