Evaluating and improving the cost-effectiveness of the implantable cardioverter—defibrillator

The implantable cardioverter defibrillator (ICD) is an expensive, widely used device for severe ventricular arrhythmias. Marginal cost-effectiveness analysis is a technique to examine the incremental cost of treatment strategy in relation to its effectiveness. In this study, we used this technique to analyze the cost-effectiveness of the ICD compared with that of electrophysiology (EP)—guided drug therapy and examined ways in which it may be improved. We analyzed Michigan Medicare discharge abstracts (1989 to 1992) and local physician visit, test, and ICD charges. Effectiveness was from 218 previously described patients with ICDs in whom the time of first event (first appropriate shock or death) was determined and presumed to represent “control” (EP-guided drug therapy) mortality. We assumed a 4-year life cycle for the ICD generator and 3.4% operative mortality and used a 5% discount to present value. Data were analyzed in a 1-month cycle Markov decision model over a 6-year horizon, and results were updated to 1993 dollars. ICD effectiveness was an increase in discounted mean life expectancy of 1.72 years. Cost-effectiveness was $31,100/year of life saved (YLS). Results were minimally or modestly sensitive to variations in preoperative mortality; resource use; consideration only of patients with ICDs who were receiving any antiarrhythmic drug or specifically amiodarone; and to a decrease in the percentage of first shocks that would equal death without the ICD until the assumed percentage decreased to <38%. At ejection fraction of <0.25 and ≥0.25, cost-effectiveness was $44,000/YLS and $27,200/YLS, respectively, and without preimplant EP study was $18,100/YLS. Cost-effectiveness of the endocardial ICD in preliminary analysis was $25,700/YLS. Sensitivities were similar to those of the epicardial ICD, as was improvement in value in patients with ejection fractions ≥0.25 and without preimplant EP study, for whom cost-effectiveness was $14,200/YLS. In conclusion, use of first-discharge data provides a method of determining ICD cost-effectiveness, which is consistent with that of other techniques. The endocardial ICD, which is now the device predominantly used, is somewhat more cost-effective than the epicardial, but becomes highly so in patients with ejection fractions ≥0.25 and in particular when hospital length of stay is reduced by elimination of the preimplantation EP test. Recent reductions in length of stay with routine implantation of the endocardial device appear also to bring cost-effectiveness to this rather favorable level.