A Markov Decision Process Approach to Estimate the Risk of Obesity Related Cancers

Abstract

Around 13% of the world’s adult population was obese in 2016 and the prevalence of obesity increased at a significant rate in the last decade (Deitel in Obes Surg 13:329–330, 2003). One of the health consequences of obesity is an increased cancer risk. In this study, we model obesity levels based on BMI, cancer, and death using a Markov decision process model in order to observe the effect of obesity on cancer and mortality risks. The objective of the model is total discounted quality adjusted life years and we simulate an individual’s lifetime from 20 to 70 years by sex. Actions available to the decision makers are no intervention and bariatric surgery. Bariatric surgery is one of the effective clinical prevention methods of obesity and it is particularly recommended for morbidly obese patients. However, it is also associated with increased mortality risk. Our model aims to observe this complex dynamic between obesity, cancer and mortality risks and bariatric surgery. We parametrize the model using randomized clinical trials and published literature and obtain the optimal policy by sex. Our results suggest that obese patients for all obesity levels should undergo bariatric surgery to improve their health outcomes and to decrease cancer risk. This study has the potential to provide guidance to the obese individuals when considering bariatric surgery and it could be further enhanced by the addition of other health outcomes of obesity to the model.