The cost-effectiveness of cardiac monitoring in breast cancer patients who have received cardiotoxic therapies

It has been known that anthracycline-based chemotherapy has the potential to cause cardiac dysfunction in breast cancer patients; however, recently evidence has shown that the addition of trastuzumab increases this risk. The study objective was to compare the cost-effectiveness of monitoring for cardiotoxicity with B-type natriuretic peptide (BNP), multi-gated acquisition scanning (MUGA), echocardiography (ECHO) or no monitoring from a payer’s prospective. Cost-effectiveness was compared between alternatives using an incremental cost-effectiveness ratio with outcomes of 1) quality-adjusted life-years and 2) percentage of patients diagnosed with each monitoring strategy. Costs estimates (in 2010 U.S. Dollars) of each strategy (obtained from the Center for Medicare and Medicaid Services website [www.cms.gov]) included the cost of the test, cost of treating heart failure once discovered (which includes medications, routine office visits, medication management) and the cost of potential acute care (which includes emergency department visits and hospitalizations). Estimates for the probabilities of heart failure development, disease progression, need for acute care, and mortality, as well as utility estimates for all disease stages were obtained from published literature. A 15-year time-frame was used with a 3% discount rate for both costs and QALYs. In the base-case analysis, the average costs and QALYs for monitoring patients were $10,062/ 6.92 QALY, $13,627/4.22 QALY, $14,739/ 6.61 QALY and $15,656/ 6.49 QALY for BNP, No Monitoring, ECHO and MUGA respectively. When comparing all alternatives to BNP, the ICER values were negative, indicating that BNP was the dominant monitoring strategy. Percent detection was similar between the three monitoring methods [21-22 % for HER-2(-) and 30-31% for HER-2(+) patients]. Again BNP was dominant over the other monitoring strategies. Sensitivity analyses were robust to changes in discount rate, probability of patients testing HER-2 (+), probability of patients being diagnosed in an asymptomatic stage, incidence of cardiac dysfunction in patients receiving anthracycline therapy ± trastuzumab and estimate of disutility associated with additional testing. A probabilistic sensitivity analysis conducted via Monte Carlo simulation led to the same conclusion as the base-case analysis; BNP was the dominant strategy over all monitoring alternatives.