Extended Abstract: Health expenditures for diabetes and its complications totals 376 billion dollar (USD) and is expected to exceed 490 USD billion by 2030. Developed world countries count for 80% and US for 52% of these total health expenditures. Type 2 diabetes is a chronic disease with long term complications as blindness, renal failure and increased risk for stroke and myocardial infraction. There are several studies related to prediction of diabetes type 2. Most famous models and widely used are the IRIC, QDScore, DESIR. All models seek to be aligned with age, BMI and waist circumference as variables, however, the performance of each model varies between countries, age, sex, and adiposity. One of the methods that have been used in the past to define the risks and performance of several treatments for diabetes type 2, is Markov chain models constructed for a heterogeneous subscriber population, and used to examine the long-run effects of particular utilization patterns on disease functioning.
Systems Engineering (SE) is an approach combining the advantage of model-based systems descriptions, based on modular components, integrated with sophisticated tradeoff analysis and design space exploration, to design and analyze the performance of complex systems in many domains, ranging from engineering, to economics, to enterprises, and most recently to healthcare systems and processes [17]. In fact, the recent PCAST report [17] recommends SE as a critical methodology for accelerating improvements in healthcare systems towards higher quality and lower costs. However, there are very few studies using this modern and interdisciplinary approach to healthcare. The present paper describes our research towards developing such a framework for the case of diabetes type 2, that will allow all stakeholders of healthcare to assess the benefits of specific policies, technologies, treatments etc., with respect to several metrics, including economic metrics. Further we propose that modern SE methodologies are uniquely capable to evaluate the interrelationships between healthcare, information technology and economics. We offer the present study as a prototypical example.
The purpose of the present study is to classify population groups based on diabetic risk, formulate models of structure and behavior, set requirements for treatment performance and construct states (for various relevant processes, including the disease, patients, treatments, technologies). Every state will have different probabilities of disease progression, cost function and health performance. Finally, the results of every state will be used to construct and validate the Markov chain model. From the Markov chain model, for every step, outputs of the model will be total cost, risk, health performance. Also every state will generate a different sum total of cost, health performance and risk. Every state shows different progression of the disease that will need different cost function (combination of treatment and medicine), health function (Qaly). Each step represents the evolution (progression) of the model through time.