PMS55 - THE EFFECTS OF MORTALITY ASSUMPTIONS ON THE COST EFFECTIVENESS OF A NEW GOUT TREATMENT

Evaluate the effects of various mortality assumptions on the expected survival and cost effectiveness, using a gout NICE-submission Markov-cohort cost-effectiveness model for adding lesinurad to allopurinol after patients fail to reach target serum uric acid (sUA) level on allopurinol alone. Several publications have quantified association between hyperuricaemia in gout and mortality. In the gout model quality adjusted life years are gained by improvements in flares, tophi, and sUA in addition to increased survival due to lower sUA. Two sources of mortality ratios were used to derive mortality rates by sUA ranges. The extent of mortality due to sUA reduction from treatment is unknown, so reduced mortality proportions considered modifiable were applied to the ratios. The use of gender specific ratios was also evaluated. Several scenarios varying mortality ratios, derived from two sources estimated total survival, and the cost effectiveness of lesinurad treatment. ICERs varied from £11,167 to £16,041. Incremental costs were essentially unchanged. The lowest mortality estimate used a US study and 50% modifiable mortality while the higher used 34% modifiable excess mortality with Scottish mortality data that found a smaller effect of hyperuricaemia and a large difference in the effect between genders. Compared to the 50% modifiable scenario there is only a 7% reduction in life-years but it accounts for 45% of the difference in QALYs between the two treatments. In the 34% gender specific estimate these are reduced to 2% and 20%. If no modifiable excess mortality is modeled, the ICER is £20,074. Neither the percentage modifiable nor the study data source have a huge effect on the ICER. It does not seem credible that there is no mortality benefit achieved by reducing sUA. The effect of gender on mortality in hyperuricaemia merits further study, but a treatment effect on overall mortality seems warranted.