Journal Name: Production and Operations Management, Forthcoming
Hypertension has not been well studied by operations researchers from a clinical decision support perspective. Moreover, little personalized (i.e. patient-centric) guidance is available regarding the number and combination of antihypertensive medications. To fill this gap, we develop a Markov Decision Process (MDP) to characterize the optimal sequence (and combination) of antihypertensive medications under the standard medication dose. Our model is patient-centric as it takes into account a set of relevant patient characteristics such as age, gender, blood pressure level, smoking habits, diabetes status, and cholesterol level. Based on a set of intuitive assumptions, we prove that our model yields a series of structured optimal policies. Having calibrated our model based on real data and medical literature, we analyze these optimal policies and discuss their insights to the real practice. We also compare the benefits, in terms of quality adjusted life expectancy, QALE, obtained from our results with those obtained from British Hypertension Society (BHS) guideline.