Severity-Adjusted Probability of Being Cost Effective.

Background: In the context of priority setting, a differential cost-effectiveness threshold can be used to reflect a higher societal willingness to pay for quality-adjusted life-year gains in the worse off. However, uncertainty in the estimate of severity can lead to problems when evaluating the outcomes of cost-effectiveness analyses.

Objectives: This study standardizes the assessment of severity, integrates its uncertainty with the uncertainty in cost-effectiveness results and provides decision makers with a new estimate: the severity-adjusted probability of being cost effective.

Methods: Severity is expressed in proportional and absolute shortfall and estimated using life tables and country-specific EQ-5D values. We use the three severity-based cost-effectiveness thresholds (€20.000, €50.000 and €80.000, per QALY) adopted in The Netherlands. We exemplify procedures of integrating uncertainty with a stylized example of a hypothetical oncology treatment.

Results: Applying our methods, taking into account the uncertainty in the cost-effectiveness results and in the estimation of severity identifies the likelihood of an intervention being cost effective when there is uncertainty about the appropriate severity-based cost-effectiveness threshold.

Conclusions: Higher willingness-to-pay thresholds for severe diseases are implemented in countries to reflect societal concerns for an equitable distribution of resources. However, the estimates of severity are uncertain, patient populations are heterogeneous, and this can be accounted for with the severity-adjusted probability of being cost effective proposed in this study. The application to the Netherlands suggests that not adopting the new method could result in incorrect decisions in the reimbursement of new health technologies.