Epistemix

Epistemix is a developer of a simulation and modeling platform to measure how things spread or emerge in populations. The company's platform is designed to model realistic scenarios, compare interventions, and communicate results. The platform captures social, environmental, and behavioral elements of people to model complex social contagions and epidemics. The software assists with disease management and community health management programs and is designed for governments, medical organizations, health systems, enterprises, and related institutions.

Epistemix was founded in 2018 by Donald S. Burke, John Grefenstette, and John Cordier and is headquartered in Pittsburgh. The company was founded based on the research of Donald S. Burke, a former dean of the University of Pittsburgh Graduate School of Public Health, and John Grefenstette, a former professor of Biostatistics and Health Policy and Management. The two developed a Framework for Reconstructing Epidemiological Dynamics (FRED) to predict the progression of disease outbreaks. Epistemix's platform has been used by MITRE, RAND Corporation, University of Pittsburgh, The University of British Columbia, Resultant, and KSM Consulting.

Epistemix's platform is developed for users to create agent-based models to generate actionable insights. The platform builds computational models capable of simulating specific health impacts of competing strategies to allow them to be compared, evaluated, and communicated. The simulations are built to account for individual persons, households, neighborhoods, schools, and workplaces and include demographic characteristics, differences in population density, contact patterns, household structure, cross-immunity from prior infections, and impacts of related policies.

The platform includes simulations of realistic social dynamics, which can be difficult to model as they tend to be complicated. To achieve this, the platform simulates social determinants and specific interactions among millions of individuals and populations and also simulates the networks that connect these people and populations. Individuals in the simulation are computationally represented, simulated, and tracked to see how their behaviors may diverge depending on predispositions and social or local circumstances.

To keep the platform fast and computationally efficient, it runs on cloud computation infrastructure to allow for the scaling of production. It is developed in a straightforward modeling language, with a declarative language to make it simpler to capture complex mental models and allow the code to be reviewable by non-technical audiences. As well, the platform has a modular architecture, allowing users to add, remove, modify, and port components within and across models in a single system.