A Markov decision process for modeling adverse drug reactions in medication treatment of type 2 diabetes. (August 2019)

• Record Type: Journal Article
• Authors: Eghbali-Zarch, Maryam; Tavakkoli-Moghaddam, Reza; Esfahanian, Fatemeh; Azaron, Amir; Sepehri, Mohammad Mehdi
• Is Part Of: Proceedings of the Institution of Mechanical Engineers. Volume 233:Number 8(2019)
• Page Start: 793
• Page End: 811
• Publication Date: 2019-08
• Subjects: Markov decision process -- type 2 diabetes -- adverse drug reactions -- medical decision making -- endocrinology and metabolism; Biomedical engineering -- Periodicals; Medical instruments and apparatus -- Periodicals; 610.28
• DOI: 10.1177/0954411919853394
• Languages: English
• ISSNs: 0954-4119

Abstract

Type 2 diabetes has an increasing prevalence and high cost of treatment. The goal of type 2 diabetes treatment is to control patients' blood glucose level by pharmacological interventions and to prevent adverse disease-related complications. Therefore, it is important to optimize the medication treatment plans for type 2 diabetes patients to enhance the quality of their lives and to decrease the economic burden of this chronic disease. Since the treatment of type 2 diabetes relies on medication, it is vital to consider adverse drug reactions. Adverse drug reaction is undesired harmful reactions that may result from some certain medications. Therefore, a Markov decision process is developed in this article to model the medication treatment of type 2 diabetes, considering the possibility of adverse drug reaction occurring adverse drug reaction. The optimal policy of the proposed Markov decision process model is compared with clinical guidelines and existing models in the literature. Moreover, a sensitivity analysis is conducted to address the manner in which model behavior depends on model parameterization and then therapeutic insights are obtained based on the results. The satisfying results show that the model has the capability to offer an optimal treatment policy with an acceptable expected quality of life by utilizing fewer medications and provide significant implications in endocrinology and metabolism applications.