BACKGROUND: Because insulin dosing requires optimization of glycemic control, it is important to use a single metric of benefit and risk to determine best insulin dosing practices. We retrospectively applied a multiplicative clinical utility index (CUI) to a study of LY2605541 (Eli Lilly and Company, Indianapolis, IN), a novel, long-acting basal insulin. MATERIALS AND METHODS: A CUI was developed to transform the multidimensional problem of assessing benefit/risk of multiple dosing algorithms into a single decision-making metric to evaluate two LY2605541 dosing algorithms relative to the insulin glargine (GL) dosing algorithm. The CUI was applied to data in a 12-week, open-label, Phase 2 trial in patients with type 2 diabetes mellitus who were randomized to one of two dosing algorithms for LY2605541 (LY1 or LY2) or GL (algorithm similar to LY1). The CUI was created (via expert input) by weighing the relative benefit/risk of four components (glycosylated hemoglobin [HbA1c], percentage of patients with HbA1c ≤ 7%, hypoglycemia rate, and time to steady-state dose); individual utility values were multiplied to compute CUI values for LY1 and LY2 relative to GL, and bootstrap samples were used to determine variability. RESULTS: The mean CUI was 0.82 for LY1 and 1.35 for LY2. Based on 3,000 bootstrap samples, there was a 48% probability of LY2 performing better than LY1 and a 28% probability of LY1 performing better than LY2. CONCLUSIONS: CUI methodology, and in particular this CUI, is a useful tool for comparing dosing algorithms. Based on this CUI, LY2 is likely to be a better dosing algorithm than LY1.
RCT Entities:
BACKGROUND: Because insulin dosing requires optimization of glycemic control, it is important to use a single metric of benefit and risk to determine best insulin dosing practices. We retrospectively applied a multiplicative clinical utility index (CUI) to a study of LY2605541 (Eli Lilly and Company, Indianapolis, IN), a novel, long-acting basal insulin. MATERIALS AND METHODS: A CUI was developed to transform the multidimensional problem of assessing benefit/risk of multiple dosing algorithms into a single decision-making metric to evaluate two LY2605541 dosing algorithms relative to the insulinglargine (GL) dosing algorithm. The CUI was applied to data in a 12-week, open-label, Phase 2 trial in patients with type 2 diabetes mellitus who were randomized to one of two dosing algorithms for LY2605541 (LY1 or LY2) or GL (algorithm similar to LY1). The CUI was created (via expert input) by weighing the relative benefit/risk of four components (glycosylated hemoglobin [HbA1c], percentage of patients with HbA1c ≤ 7%, hypoglycemia rate, and time to steady-state dose); individual utility values were multiplied to compute CUI values for LY1 and LY2 relative to GL, and bootstrap samples were used to determine variability. RESULTS: The mean CUI was 0.82 for LY1 and 1.35 for LY2. Based on 3,000 bootstrap samples, there was a 48% probability of LY2 performing better than LY1 and a 28% probability of LY1 performing better than LY2. CONCLUSIONS: CUI methodology, and in particular this CUI, is a useful tool for comparing dosing algorithms. Based on this CUI, LY2 is likely to be a better dosing algorithm than LY1.