PDB23 BODY MASS INDEX ASSOCIATED UTILITY APPROACHES IN THE IQVIA CORE DIABETES MODEL

Abstract

Changes in body mass index (BMI) is one of the key drivers of cost-effectiveness in diabetes modeling via its impact on utility. Most models are using a fixed disutility per point increase of BMI. However, it can be questioned whether decreasing BMI has the same impact on utility in overweighed and non-overweighed patients. IQVIA Core Diabetes Model (CDM) version 9.5 includes two disutility approaches: 1) fixed disutility per unit gain of BMI above a BMI of 25kg/m2 (already available in version 9.0); 2) polynomial model based on a regression analysis with a variable disutility based on the current BMI (Soltoft F, 2009). The aim of this study was to assess the impact of the two approaches on QALYs and ICER. The observational EDGE study comparing metformin+vildagliptin (M+V) versus metformin+sulphonylurea (M+S) was used as a base case. Basal insulin rescue therapy was applied to both arms when at an HbA1c threshold level of 7.5% was reached. Baseline characteristics and treatments effects from the published EDGE study were applied. HbA1c reductions and change in body weight were applied -0.99% -1.19% and -0.3kg -1.6kg for M+S and M+V respectively. Fixed disutility per point of BMI was the NICE approved CDM default value (0.0061). UK costs (2018) were used and annual discounting of 3.5% was applied on costs and QALYs. With the polynomial approach, the lifetime QALYs were 8.147 and 8.119 for M+V and M+S respectively. When using the static approach, QALY were 7.940 and 7.896 respectively. With the new approach the ICER increased from 12,061 GBP/QALY to 19,655 GBP/QALY. The polynomial approach is more conservative resulting in higher total QALY in both arms, lower incrementals and thus higher ICERs compared to the static approach.