ABSTRACT
The agent-based modeling (ABM) approach comes very close to the evolving concept of “virtual reality” in which a computer environment is used to replicate the complex series of interactions on the part of individuals, which produce a wide variety of collective behavior. Issues of comorbidities often plague traditional patient segmentation schemes, often because published research or even careful analysis of International Classifications of Disease diagnosis codes in databases is insufficient to correctly categorize overlap of existing conditions. Layering in another dimension into the patient disaggregation can present significant difficulties from a data collection standpoint. ABM is well suited to address this type of situation, however, and can establish a distinct agent for each individual patient whose subsequent behavior is treated probabilistically. The Dynamic Modeling approach has typically had a foot in two different worlds—one of strategy development and one of forecasting/analytics.
