Cn1 - Validating Survival Extrapolations in First Line Treatment of Renal Cell Carcinoma Using Literature-Based Conditional Survival Probabilities

Abstract

perspective. METHODS: The IQVIA Core Diabetes Model (CDM) was calibrated to reproduce the outcomes from the EMPA-REG OUTCOME trial. Baseline characteristics and observed effects on physiological parameters (HbA1c, BMI, blood pressure, lipids) were used as inputs. Network meta-analysis provided the relative risks for cardiovascular outcomes with empagliflozin versus sitagliptin and saxagliptin. HbA1c progression for all arms was projected based on the EMPA-REG OUTCOME trial. The effects of the CVOTs were applied up until treatment switch (reaching HbA1c of 8.5%) after which, the UKPDS82 risk equations predicted events based on physiological parameters. Basal-bolus rescue therapy was assumed after reaching an HbA1c of 8.5%. UK event costs and quality of life data were taken from literature. Drug costs were from the British National Formulary and Monthly Index of Medical Specialities. Discounting of 3.5% was applied. RESULTS: CDM projected 5.923, 5.462, and 5.324 qualityadjusted life-years (QALYs) and 51,629GBP, 48,306GBP and 48,885GBP total lifetime costs for empagliflozin, sitagliptin and saxagliptin, respectively. The incremental cost-effectiveness ratio of empagliflozin versus sitagliptin and saxagliptin was 7,209GBP/QALY and 4,582GBP/QALY, respectively. One-way and probabilistic sensitivity analyses showed robustness of the results. CONCLUSIONS: The results of this cost effectiveness analysis suggested that empagliflozin + SoC was costeffective compared to sitagliptin + SoC and saxagliptin + SoC at a willingness to pay threshold of 20,000GBP/QALY.

Predicting Survival for Chimeric Antigen Receptor T-Cell Therapy: A Validation of Survival Models Using Follow-Up Data From ZUMA-1

“…Ouwens et al 16 focused instead on the "area-under-the-curve" to compare different models, equivalent to estimating restricted mean survival time, and also presented plots of the estimated hazard function to inform model selection. Klijn et al 17 compared estimates of the conditional survival between models as an alternative means of judging model fit. Nevertheless, there is no established consensus on how to choose the best-fitting models, in spite of there being a large number of studies concerning how to undertake survival analysis for HTA.…”