Fully Bayesian Unsupervised Disease Progression Modeling

We present a practical implementation of a fully unsupervised disease progression model [10]. The implementation utilizes all new components we developed for generic use in Bayesian disease progression modeling. It improves upon [10] by providing a more informative fully Bayesian approach and a faster inference algorithm. The implementation is completely built on the pyMC3 open-source library making it easy to extend the model and apply to new settings. 1 Disease Progression Models Traditionally, disease severity and progression have been assessed manually by physicians using guidelines such as the GOLD criteria for COPD [6]. These guidelines are typically based on rules applied to the patient’s biomarkers, demographics, and other data easily extracted from health records. The sub-area of machine learning called disease progression modeling (DPM) focuses on automating this process [5]. Automation leads to more accurate diagnoses and optimal treatment paths which can literally be the difference between life and death as in the case of coagulopathy patients [9]. More broadly, we expect that algorithms that learn disease progression models from electronic health records will lead to new insights on the progression of rare and difficult to stage chronic diseases, guiding both clinical practice and medical research. 2 Bayesian Models and pyMC3 Bayesian networks provide a natural framework for modeling disease progression. They allow for the flexible modeling of “hidden states” which often arise in medical scenarios where measurements are simply proxies for variables of interest. Furthermore, Bayesian posteriors provide a full description of parameters of interest as oppose to point estimates and simple confidence intervals. Several examples of Bayesian network models for disease progression exist in the literature [1, 2, 4, 7, 10]. pyMC3 is a Python module that provides a unified and comprehensive framework for fitting Bayesian models using MCMC [8]. pyMC3’s key strength is its modularity and extensibility: ran∗Research performed while interning at Evidation Health