In this section
@inproceedings{Plumb:2024,
author = {Plumb, W and Casale, G and Bottle, A and Liddle, A},
pages = {1220--1231},
publisher = {ACM / IEEE},
title = {Clinical pathway clustering using surrogate likelihoods and replayability validation},
year = {2024}
}
TY - CPAPER
AB - Modelling clinical pathways from Electronic Health Records (EHRs) can optimize resources and improvepatient care, but current methods for generating pathway models using clustering have limitations includingscalability and fidelity of the clusters. We propose a novel pathway modelling approach using MaximumLikelihood (ML) data clustering on Markov chain representations of clinical pathways. Our method iscalibrated to produce clusters with low inter-cluster variability across the pathways. We use machine learningwith Stochastic Radial Basis Functions (SRBF) kernels for surrogate optimization to handle non-convexityand propose an incremental optimization method to improve scalability. We also define a methodologybased on novel replayability scores to help analysts compare the fidelity of alternative clustering results.Results show that our ML method produces clusters that have higher fidelity in terms of replayability scoresthan k-means based clustering and in capturing queueing contention, which is important for bottleneckidentification in healthcare.
AU - Plumb,W
AU - Casale,G
AU - Bottle,A
AU - Liddle,A
EP - 1231
PB - ACM / IEEE
PY - 2024///
SP - 1220
TI - Clinical pathway clustering using surrogate likelihoods and replayability validation
ER -