On the role of cost-effectiveness thresholds in healthcare priority setting

In the past few years, empirical estimates of the marginal cost at which health care produces a quality-adjusted life year (QALY, k) have begun to emerge. In theory, these estimates could be used as cost-effectiveness thresholds by health-maximizing decision makers, but prioritization decisions in practice often include other considerations than just efficiency. Pharmaceutical reimbursement in Sweden is one such example, where the reimbursement authority (TLV) uses a threshold range to give priority to disease severity and rarity. In this paper, we argue that estimates of k should not be used to inform threshold ranges. Instead, they are better used directly in health technology assessment (HTA) to quantify how much health is forgone when a new technology is funded in place of other healthcare services. Using a recent decision made by TLV as a case, we show that an estimate of k for Sweden implies that reimbursement meant forgoing 8.6 QALYs for every QALY that was gained. Reporting cost-effectiveness evidence as QALYs forgone per QALY gained has several advantages: (i) it frames the decision as assigning an equity weight to QALYs gained, which is more transparent about the trade-off between equity and efficiency than determining a monetary cost per QALY threshold, (ii) it makes it less likely that decision makers neglect taking the opportunity cost of reimbursement into account by making it explicit, and (iii) it helps communicate the reason for sometimes denying reimbursement in a way that might be less objectionable to the public than current practice.