Prioritizing Discordant Chronic Comorbidities and Predicting the Medication Using Machine Learning

Date of Award


Degree Name

M.C.S. (Master of Computer Science)


Department of Computer Science


Tom Ongwere


The aim of this thesis is to design a tool that predicts an optimal medication combination for patients with Discordant Chronic Comorbidities (DCCs). This tool allows a patient to select a combination of diseases they have and the major treatment concerns (such as cost of medication, interactions with existing medication, weight gain, etc.) and then uses machine learning algorithms to recommend optimal treatment regimen. Patients with DCCs experience constantly changing treatment needs that require them to schedule numerous appointments with their care team, navigate complex care structure and coordination between various clinical specialists to ensure quality of their care and treatment. However, with the complex and constantly changing needs of patients with DCCs, it is difficult for the patients and providers to prioritize the medications and understand the impact caused by one treatment on concurrent treatments. We believe that predicting an optional treatment regimen for DCCs would; a) aid the healthcare practitioners in making complex decisions, b) guide the patients to stay aware of the alternative medications, c) and prepare patients in having informed and engaging discussions with the care team about their medication options. To design this tool, we first explored how healthcare professionals include unique patients’ preferences when deciding on a medication plan for their patients. To do this exploration, we conducted a survey, targeting healthcare professionals. We used the processed data from that survey to train the commonly used machine learning models (including Random Forest, K-nearest neighbors, AdaBoost and XGBoost). We created benchmarks based on the performance of each model and designed an iterative user interface that captures patients’ treatment concerns and returns the optimal medication regimen.


Machine learning, Chronic diseases, Discordant chronic comorbidities, prediction, multiclass multi-output classification

Rights Statement

Copyright © 2023, Author