RWD106 PREDICTION OF RARE DISEASE MEDICATION TRANSITION PATTERNS BASED ON AN INTERPRETABLE MACHINE LEARNING MODEL
Jingyi Li, Danyang Wei, Xuanqi Qiao, Hongfei Gu, Hu Min · Value in Health Regional Issues · 2025
real-world data (RWD) sources for AD in Singapore for conducting evidence generation studies.Methods: A Markov model was developed to project AD progression and assess the lifetime cost-utility of lecanemab plus standard of care (SoC) compared to SoC alone, using data from the Phase III Clarity AD study.The model incorporates unit costs, mortality rates, and natural history data sourced from published literature and validated by local clinicians in Singapore.It also accounts for treatment discontinuation as patients progress to moderate AD or require nursing home care.Costs and benefits are discounted at a rate of 3.0%, and a lifetime horizon is assumed.Additionally, an AI-powered systematic literature review leveraging large language models (LLMs) was conducted on academic publications from 2014 to 2024 in PubMed to identify RWD sources for AD in Singapore.Results: Treatment with lecanemab was associated with an increase in qualityadjusted life years over the model time horizon as well as incurring incremental costs for both the MCI due to AD and mild AD dementia populations, with the incremental cost-utility ratio less than 1x GDP per capita.Taking a societal perspective reduced the incremental cost-utility ratio (ICUR) for lecanemab plus SoC versus SoC alone.However, Singapore significantly lacks real-world data for AD.Conclusions: Lecanemab offers a cost-effective strategy for early AD treatment, delaying disease progression and offering societal benefits.Additional AD RWD would better support the value of AD interventions in Singapore.