(M-036) Pharmacometric-Pharmacoeconomic Modeling and Simulation in Atopic Dermatitis: Informing Early Drug Development Decisions for a Hypothetical New Therapeutic

Daniel Polhamus, PhD – Principal Scientist II, Group Leader, Metrum Research Group; Michelle Johnson, MBA – CEO, Metrum Research Group; Eric Anderson, MS – Senior Data Science Engineer, Metrum Research Group; Kyle Barrett, BS – Data Science Engineer, Metrum Research Group

Objectives: This study assessed the expected impact of selected target product profile (TPP) characteristics on the cost effectiveness (CE) of a hypothetical novel therapeutic in atopic dermatitis (AD) relative to a reference therapeutic, dupilumab (DU).

Methods: A pharmacometric (PM) - pharmacoeconomic (PE) model was developed to describe the PM-PE relationship for DU. The PM-PE model for hypothetical Drug X (DX) was based on the same structure, with select modifications of PM model parameters, relative to the DU reference. PM model data sources included digitized longitudinal meta-data from published studies. The PM model described longitudinal eczema area and severity index (EASI) score as a fractional decrease from baseline EASI score, including effects for placebo response, topical corticosteroids (TCS), and drug effects. The PE model was derived from a published PE analysis of dupilumab in AD and was characterized as a Markov model with transition probabilities between health states: non-responder, responder (EASI 50, EASI 75, EASI 99), and death, with each state associated with quality adjusted life years (QALYs). Simulation scenarios included variations of DX properties relative to dupilumab: increased maximum fractional decrease (Emax), shorter onset time (T50), and improved persistence of therapy (POT) for a population of mixed moderate/severe disease phenotypes. Replicate simulations were implemented in an interactive tool developed in R and Shiny, running on the Metworx platform. Results were summarized for each scenario as the difference in (DX-DU) QALYs and the probability of CE vs. willingness to pay (WTP).

Results: Improvements in Emax and T50 for DX relative to DU did improve the population mean response profile, but did not impact QALYs or probability of CE at any WTP level. Improvements in POT did result in an increase of approximately 1 QALY and improved probability of CE for DX relative to DU (30% vs 15%, respectively, at a WTP of $100,000). In order to achieve a similar probability of cost effectiveness without the effect on POT, a decrease in DX pricing of approximately 10% relative to DU would be necessary.

Conclusions: TPP factors that differentiate DX from DU on efficacy do not necessarily translate to increased QALYs or probability of CE. It may be important to consider the impact of new drug characteristics on CE when setting the TPP and in early development decision making.

Citations: [1] Blauvelt, A et al. The Lancet, 2017, 389: 10086, 2287-2303.
[2] Simpson, E et al. The New England Journal of Medicine, 2016, 375 (24): 2335-48.
[3] Silverberg et al. Annals of Allergy, Asthma, & Immunology, 2021,126:1, 40-45.
[4] Deng C et al. Journal of investigative dermatology, 2019, 139(9).