Background: In cost-effectiveness analysis (CEA), the effects of health-care interventions on multiple health dimensions typically require consideration of both quantity and quality of life.
Objectives: To explore the impact of alternative approaches to quality-of-life adjustment using patient preferences (utilities) on the outcome of a CEA on use of tamoxifen for breast cancer risk reduction.
Research Design: A state transition Markov model tracked hypothetical cohorts of women who did or did not take 5 years of tamoxifen for breast cancer risk reduction. Incremental quality-adjusted effectiveness and cost-effectiveness ratios (ICERs) for models including and excluding a utility adjustment for menopausal symptoms were compared with each other and to a global utility model.
Subjects: Two hundred fifty-five women aged 50 and over with estimated 5-year breast cancer risk ≥1.67% participated in utility assessment interviews.
Measures: Standard gamble utilities were assessed for specified tamoxifen-related health outcomes, current health, and for a global assessment of possible outcomes of tamoxifen use.
Results: Inclusion of a utility for menopausal symptoms in the outcome-specific models substantially increased the ICER; at the threshold 5-year breast cancer risk of 1.67%, tamoxifen was dominated. When a global utility for tamoxifen was used in place of outcome-specific utilities, tamoxifen was dominated under all circumstances.
Conclusions: CEAs may be profoundly affected by the types of outcomes considered for quality-of-life adjustment and how these outcomes are grouped for utility assessment. Comparisons of ICERs across analyses must consider effects of different approaches to using utilities for quality-of-life adjustment.