Year: 2019
Type: dissertation
Abstract: Healthcare processes are complex and may vary considerably among the same cohort of patients. \nProcess mining techniques play a significant role in automating the construction of healthcare \nmodels using a system's event log. An event log is a data type that records any event that \noccurs within the process. It is a basic element of any information system and has three main \ncomponents: process instance id, event and time when an event has occurred. Using ordinary \ntechniques of process mining in healthcare produces `spaghetti-like' models which are difficult \nto understand and thus have little value. Previously published studies have highlighted the importance \nof event abstraction which is considered as a central tool for reducing complexity and \nimproving efficiency. Although studies have successfully improved the understandability of process \nmodels, they have generally relied on involvement from a domain expert. Untangling these \n`spaghetti-like' models with the help of domain experts can be expensive and time-consuming. \nMachine learning techniques such as Hidden Markov Model (HMM) has been used for modelling \nsequential data for a long time. State transition modelling has also been explored by process \nmining research and is advocated for sequence clustering purposes where a model is trained \nover a group of sequences and then used to evaluate if a process instance is more likely to be \ngenerated from this model or not. However, state transition models can also be utilised for detecting \nhidden processes which can be used subsequently for process abstraction. In this thesis, \nwe aim to address healthcare process complexity using unsupervised abstraction. We adopt \nan unsupervised method for detecting hidden processes using HMM and the Viterbi algorithm. \nThe method in this research includes eight stages; event logs extraction, preprocessing, learning, \ndecoding, optimisation, selection, visualisation and lastly model evaluation. One of the main \ncontributions of this research is the design of two different types of process model optimisation \nwhich are strict and soft optimisations. Models that are selected by the proposed optimisation \naddress the limitations of other standard metrics that can be used for model selection in \nHMM such as Bayesian Information criteria (BIC). Two different real healthcare data sources \nare used in this research namely the Medical Information Mart for Intensive Care (MIMIC-III) \nfrom Boston, USA and the Patients Pathway Manager (PPM) from Leeds, UK. Models are \ntrained using the MIMIC-III medical event log and then tested using the PPM dataset to be \nevaluated later by a domain expert. Three breast cancer case studies that range in complexity \nare extracted. The results of our method have significantly improved model complexity \nand provided a conceptually valid abstraction for several care patterns. Promising results are \ndemonstrated in the improvement of the precision and fitness of the abstracted models. The \nabstracted models can then be used as a middle step for bringing structure to unstructured \nprocesses which helps in finding cohorts of patients based on similar healthcare processes. The \nhealthcare processes of a cohort of patients can then be modelled using any process mining tool \nwhere their process similarity could not be captured in the complex models.
Author: Amirah Alharbi
Language: English
Subfield: Management Information Systems
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W2995131913