Estimating EQ-5D weights from other trial outcome measures for use in cost-effectiveness studies: an example in patients with frozen shoulder

摘要

Background Health-related quality of life (HRQoL) weights, such as EQ-5D, needed for conducting economic evaluations in frozen shoulder populations are currently lacking. The present study examines the feasibility of using mapping methodology to estimate HRQoL, obtain quality-adjusted life-years (QALYs) and undertake a cost-effectiveness analysis to help inform decision-making.Methods Patient-level data from a randomized trial in a similar population were used to undertake a mapping exercise evaluating a variety of regression modelling techniques. The Short-Form-36 component summary scores and pain visual analogue scale scores were mapped to EQ-5D. The best fit model was selected and the predicted weights were used for estimating QALYs and cost-effectiveness results. Results The selected model explained 41% (/?2) of the variation in EQ-5D weights and the estimates.for interventions ranged from 0.5025 to 0.6619 at baseline and from 0.6903 to 0.7600 at 3 months. The QALYs (gained) ranged from 0.1518 to 0.1715 and the estimates of cost-effectiveness were up to ￡400 per QALY for steroid injection compared to physical therapy with steroid injection.Discussion Few studies in frozen shoulder populations include HRQoL weights. We have shown that mapping can potentially take advantage of datasets in similar populations and estimate HRQoL outcomes and subsequently cost-effectiveness. The results may be uncertain but can be helpful in informing the discussion around cost-effectiveness and future areas of research.