Resource allocation in decision support frameworks

Cost-benefit and cost-effectiveness analysis place limits on the dimensions of value that the models can incorporate. Cost-benefit analysis requires monetization of all measures of value (including life), a task sometimes deemed either difficult to accomplish or even repugnant. Cost-effectiveness analyses include health care gains in natural units (e.g., quality-adjusted life years or QALYs) rather than purely monetizing them (e.g., in dollars) and offers an efficiency perspective based on the ratio of cost per QALYs or similar health measures. These two methods use different rules for investment. Cost-benefit analysis says to invest whenever benefits exceed costs. Cost-effectiveness analysis says to invest if the intervention has a cost per QALY that meets-or is below-a designated cutoff value.

Multi-criteria frameworks expand decision analyses by considering value tradeoffs from decision makers, and then producing a synthetic measure that summarizes the performance of investment options. This evaluation is done across all chosen dimensions of value, based on the weights provided by the decision makers, but this flexibility comes at a cost. To date, no approach is widely accepted to suggest how much to invest (how to determine a budget constraint) using multi-attribute models. Moreover, there is no agreed-upon method to measure willingness to pay for incremental multi-attribute value improvements. Our paper proposes a way forward.

Based on existing dollar estimates of willingness to pay for QALYs, our concept creates a comparable cutoff for multi-criteria value measures. Our proposed method expands the acceptable cost per QALYs in proportion to how much of the total measure is accounted for by the QALY component. Agreed-upon values for cost per QALY are thus extrapolated to account for extra value created by non-QALY attributes of each intervention.

Using our proposed methods, the cost per QALY cutoff can serve as a benchmark toward creating a resource allocation cutoff in multi-criteria frameworks.