Use of a Decision-Analytic Model

Background Coronary stenting appears to provide more predictable immediate results and lower rates of restenosis than conventional balloon angioplasty for selected lesion types, but its hospital costs are significantly higher. This study was designed to evaluate the potential cost-effectiveness of Palmaz-Schatz coronary stenting relative to conventional balloon angioplasty for the treatment of patients with symptomatic, single-vessel coronary disease. Methods and Results We developed a decision-analytic model to predict quality-adjusted life expectancy and lifetime treatment costs for patients with symptomatic, single-vessel coronary disease treated by either Palmaz-Schatz stenting (PSS) or conventional angioplasty (PTCA). Estimates of the probabilities of overall procedural success (PTCA, 97%; PSS, 98%), abrupt closure requiring emergency bypass surgery (PTCA, 1.0%; PSS, 0.6%), and angiographic restenosis (PTCA, 37%; PSS, 20%) were derived from review of the literature published as of September 1993. Procedural costs were based on the true economic (ie, variable) costs of each procedure at Boston's Beth Israel Hospital. On the basis of these data, coronary stenting was estimated to result in a higher quality-adjusted life expectancy than conventional angioplasty but to incur additional costs as well. Compared with conventional angioplasty, stenting had an estimated incremental cost-effectiveness ratio of $23 600 per quality-adjusted life year gained. Although the cost-effectiveness ratio for stenting changed with variations in assumptions about the relative costs and restenosis rates, it remained less than $40 000 per qualityadjusted year of life gainedand thus was similar to many other accepted medical treatments-unless the stent angiographic restenosis rate was >23%, the angioplasty restenosis rate was <34%, or the cost of stenting (including vascular complications) exceeded that of conventional angioplasty by more than $3000. The alternative strategy of secondary stenting (initial angioplasty followed by stenting only for symptomatic restenosis) was estimated to be both less effective and less cost-effective than primary stenting over a wide range of plausible assumptions and thus does not appear to be costeffective when primary stenting is also an option. Conclusions Decision-analytic modeling can be used to evaluate the potential cost-effectiveness of new coronary interventions. Our analysis suggests that despite its higher cost, elective coronary stenting may be a reasonably cost-effective treatment for selected patients with single-vessel coronary disease. Primary stenting is unlikely to be cost-effective for lesions with a low probability of restenosis (eg, <30%) or for patients for whom the cost of stenting is expected to be much higher than usual (eg, because of a high risk of vascular complications). Given the sensitivity of the cost-effectiveness ratios to even modest variations in the relative restenosis rates and cost estimates, future studies will be necessary to determine more precisely the cost-effectiveness of coronary stenting for specific patient and lesion subsets. (Circulation. 1994;89:1859-1874.)