Learning Clinical Pathway Patterns by Hidden Markov Model

Abstract

This paper adopts Hidden Markov Models (HMMs) for discovering clinical pathways. An HMM is a stochastic probabilistic model for modeling sequential or time-series data and easily incorporating new instances to update the model. This study demonstrates the proposed framework of discovering clinical pathway patterns by HMMs. The result shows that HMMs can accurately represent clinical pathways of Normal Spontaneous Delivery. Therefore, the HMM learning process can facilitate the medical professionals' knowledge sharing and promptly maintain up-to-date clinical pathways.