Cost-Effectiveness of Cardiac Rehabilitation in Patients with Coronary Artery Disease: A Meta-Analysis.

ABSTRACT

BACKGROUND: Medical costs associated with cardiovascular disease are increasing considerably worldwide; therefore, an efficacious, cost-effective therapy which allows the effective use of medical resources is vital. There have been few economic evaluations of cardiac rehabilitation (CR), especially meta-analyses of medical cost versus patient outcome. METHODS: The target population in this meta-analysis included convalescent and comprehensive CR patients with coronary artery disease (CAD), the status most commonly observed postmyocardial infarction (MI). Here, we evaluated medical costs, quality-adjusted life year (QALY), cost-effectiveness, mortality, and life year (LY). Regarding cost-effectiveness analysis, we analyzed medical costs per QALY, medical costs per LY, and the incremental cost-utility ratio (ICUR). We then examined the differences in effects for the 2 treatment arms (CR vs. usual care (UC)) using the risk ratio (RR) and standardized mean difference (SMD). RESULTS: We reviewed 59 studies and identified 5 studies that matched our selection criteria. In total, 122,485 patients were included in the analysis. Meta-analysis results revealed that the CR arm significantly improved QALY (SMD -1.78; 95% confidence interval (CI) -2.69, -0.87) compared with UC. Although medical costs tended to be higher in the CR arm compared to the UC arm (SMD 0.02; 95% CI -0.08, 0.13), cost/QALY was significantly improved in the CR arm compared with the UC arm (SMD -0.31; 95% CI -0.53, -0.09). The ICURs for the studies (4 RCTs and 1 model analysis) were as follows -48,327.6 USD/QALY; -5,193.8 USD/QALY (dominant, CR is cheaper and more effective than UC); and 4,048.0 USD/QALY, 17,209.4 USD/QALY, and 26,888.7 USD/QALY (<50,000 USD/QALY, CR is costlier but more effective than UC), respectively. Therefore, there were 2 dominant and 3 effective results. CONCLUSIONS: While there are some limitations, primarily regarding data sources, our results suggest that CR is potentially cost-effective.