DPM360: New Additions to Advanced Disease Progression Modeling

Abstract

Disease Progression Modeling $ (DPM)^1 $ aims to characterize the progression of a disease and its comorbidities overtime using a wide range of analytics models. Typical approaches include predictive $ modeling^2 $ , time-to-event estimation $ estimation {3,4,5,6,7} $, and state-based modeling $ {8,9,10} $ for key disease-related events. DPM has applications throughout the healthcare ecosystem, from providers, to payers, and pharmaceutical companies. But the complexity of building effective DPM models can be a roadblock for their rapid experimentation and adoption when adopting cutting-edge algorithms. Some of this is addressed by standardization of data model and tooling for data analysis and cohort selection. However, there are still unmet needs to facilitate the development of advanced machine learning techniques such as recent deep learning and probabilistic modeling. To address this, we have been developing Disease Progression Modeling Workbench 360(DPM360) as an opensource project. DPM360 is an easy-to-install system to help research and development of DPM models. It manages the entire modeling life cycle, from data analysis (e.g, cohort identification) to machine learning algorithm development and prototyping. While we showed general features of DPM360 and predictive analysis in the past OHDSI event, we now demonstrate advanced modeling capability including OHDISI data tooling, and extensible training framework which exploits recent achievements of time-to-event estimation, and state-based modeling.