OBJECTIVES: Stakeholders have raised concerns that use of the quality-adjusted life-year (QALY) in cost-effectiveness (CE) analysis may discriminate against patients with lower quality of life, thus restricting access to life-prolonging treatments for these patients. The equal value of life-years gained (evLYG) outcome, which replaces condition-specific utility values with a general population utility for all incremental life-years (LYs), has been proposed as an alternative. However, the specific methodology for estimating evLYGs has not been well-described. This study compares methods for estimating evLYGs for a cohort Markov model and explores extending the approach to caregiver disutilities.
METHODS: We identified two methods for estimating evLYGs for a cohort model based on when incremental LYs are accrued: only after cumulative LYs for the new technology exceed total LYs for the comparator (evLYG1) or in all years based on annual incremental LYs (evLYG2). To compare these methods, we developed a hypothetical Markov model with three health states (mild disease, severe disease, and death) using an annual cycle and a lifetime horizon. Data were assumed for natural history, treatment effect, costs, patient utilities, and caregiver disutilities. Incremental CE ratios (ICERs) for QALY, evLYG1, and evLYG2 outcomes were compared, and the analysis was extended to caregivers by assigning no disutility during incremental LYs. Sensitivity analysis assessed the relative influence of input assumptions and the generalizability of findings.
RESULTS: In our illustrative comparison, the ICERs were $62,580/QALY gained, $44,497/evLYG1 gained, and $45,767/evLYG2 gained. Extension to caregiver disutilities yielded ICERs of $38,848/evLYG1 gained and $40,212/evLYG2 gained. The difference between evLYG methods was more sensitive to the utility difference between health states than the difference between the disease and the general population.
CONCLUSIONS: This study illustrates methodological considerations for implementing the evLYG approach in cohort models and highlights the potential impact on CE results of extending the approach to other outcomes for life-prolonging treatments.
