Cost-effectiveness Analysis of Breast Cancer Screening Using Mammography in Singapore: A Modeling Study

Abstract

Background: Limited research is available on the cost-effectiveness of breast cancer screening programs in Asian countries. We evaluated the cost-effectiveness of Singapore's national mammography screening program, implemented in 2002, recommending annual screening between ages 40 and 49 and biennial screening between ages 50 and 69, and alternative screening scenarios taking into account important country-specific factors. Methods: We used national data from Singapore in the MIcro-simulation SCreening ANalysis-Fatal diameter (MISCAN-Fadia) model to simulate 302 screening scenarios for 10 million women born between 1910 and 1969. Screening scenarios varied by starting and ending age, screening interval, and attendance. Outcome measures included life-years gained (LYG), breast cancer deaths averted, false positives, overdiagnosis, qualityadjusted life years (QALY), costs (in 2002 Singapore dollars; S$), and incremental cost-effectiveness ratios (ICER). Costs and effects were calculated and discounted with 3% using a health care provider's perspective. Results: Singapore's current screening program at observed attendance levels required 54,158 mammograms per 100,000 women, yielded 1,054 LYG, and averted 57 breast cancer deaths. At attendance rates >= 50%, the current program was near the efficiency frontier. Most scenarios on the efficiency frontier started screening at age 40. The ICERs of the scenarios on the efficiency frontiers ranged between S$10,186 and S$56,306/QALY, which is considered cost-effective at a willingness-to-pay threshold of S$70,000/QALY gained. Conclusions: Singapore's current screening program lies near the efficiency frontier, and starting screening at age 40 or 45 is costeffective. Furthermore, enhancing screening attendance rates would increase benefits while maintaining cost-effectiveness. Impact: Screening all women at age 40 or 45 is cost-efficient in Singapore, and a policy change may be considered.