Objectives Social preferences are widely used in economic evaluation required by regulatory agencies. In the UK, NICE requires the use of EQ-5D and its associated set of TTO preference weights for computing QALYs. The weights in question date back nearly two decades. It is reasonable to question whether they continue to represent contemporary social preferences. Were a revised set of EQ-5D weights to be produced then would this necessitate the revision of all past appraisal decisions? This paper presents the 1stphase of work designed to address that question. Methods The ICER is defined by the ratio of marginal cost (ΔC) /marginal benefit (ΔB). For a given ΔC the ICER falls as ΔB increases. For a given threshold (λ) and for a fixed incremental cost (ΔC), there is a minimum health benefit ΔBmin (given by ΔC/λ) which must be achieved to produce an ICER that comes below that threshold limit. TTO-weighted scores were computed for all 243 health states defined by the 3-level version of EQ-5D. A difference matrix was created in which D (i, j) contains the numeric difference between the ith and jth state. The number of differences below a given ΔBminwas computed for each column (health state). Threshold values were varied (£20,000-£50,000). Cost differences were varied (£500-£10,000). Results Less than 10% of health state value differences failed to meet the minimum ΔBmin of 0.0125 (ΔC =£500; λ=£20,000) indicating susceptability to changes in health state value, however this proportion rose to 57% for higher incremental costs (e. g. ΔC =£3,000). 81/243 health states account for 50% of the differences that exceed ΔBminat all tested levels of ΔC and λ. Graphical representation of these Results can be used to assess the need for reappraisal. Conclusions For higher cost interventions, relatively small differences in EQ-5D weights can generate ICERs with the propensity to reverse previous cost-effectiveness decisions.
Kind, P., & Meads, D. M. (2014). Out With The Old–In With The New: Would New Social Preference Weights For Eq-5d Inevitably Require A Reappraisal of Previous Cost-Effectiveness Determinations? Value in Health, 17(7), A439.