Disentangling the Complex Landscape of Infectious Diseases: How Can Modelers Best Choose the Level of Complexity That Increases Both Investment in Technologies and Model Interpretability?
Author(s)
Moderator: Anuj Mubayi, PhD, IQVIA, Tempe, AZ, USA
Panelists: Martin Meltzer, PhD, Health Economics and Modeling Unit, Center for Disease Control and Prevention, Atlanta, GA, USA; Pedro Nascimento de Lima, PhD, Rand Corporation, Wake Forest, NC, USA; Mindy Cheng, PhD, Vir Biotechnology, Alamo, CA, USA
ISSUE: Understanding complexity in healthcare via disease system has the potential to reduce decision and treatment uncertainty. The complexity associated with assessment of a health technology are linked via two different processes: the first through adaptive implementation of a technology having multiple targets, and the other through dynamical heterogeneity in infection transmission for which the technology is sought for. Transmission dynamics models of infectious diseases may play a greater role in addressing these aspects. However, the tendency of many health economics models of interventions is to limit focus on the ecology of the disease system. These models are deficient by not modeling the way the system reacts to the technology. Complex models may overfit the real-world data. Conversely, complexity can be important to include when uncertain factors are central to a disease process. How to choose the right level of ecological complexity for modeling an infectious disease? Should models capture how the disease system reacts to an intervention? What is needed to evaluate the quality of such complex models?
OVERVIEW: This panel will discuss the choice of modeling complexity and broadening access to the transmission dynamics models for infectious diseases—through ecological mechanisms—considering models as R&D products developed by innovators for a payer and policymaker consumers. Dr Mubayi will moderate and provide context for the discussion, outline parallels between model development and the gaps in health technologies evaluations. Dr Lima will provide insights on key ecological drivers for modeling efforts for which evidence may be readily available and types of modeling techniques, and Dr Meltzer will argue the need to address “market failures” in modeling, drawing on the experience of the economic value projects, an important step to understand the trade-off between model accuracy and interpretability. Panelists will provide brief 8-minute presentations on their perspectives, followed by moderated discussion and audience questions.
Conference/Value in Health Info
Code
225
Topic
Methodological & Statistical Research