A multi-objective constrained POMDP model for breast cancer screening

Breast cancer is a common and deadly disease, but it is often curable when diagnosed early. While most countries have large-scale screening programs, there is no consensus on a single globally accepted policy for breast cancer screening. The complex nature of the disease; limited availability of screening methods such as mammography, magnetic resonance imaging (MRI), and ultrasound screening; and public health policies all factor into the development of screening policies. Resource availability concerns necessitate the design of policies which conform to a budget, a problem which can be modelled as a constrained partially observable Markov decision process (CPOMDP). In this study, we propose a multi-objective CPOMDP model for breast cancer screening with two objectives: minimize the lifetime risk of dying due to breast cancer and maximize the quality-adjusted life years. Additionally, we consider an expanded action space which allows for screening methods beyond mammography. Each action has a unique impact on quality-adjusted life years and lifetime risk, as well as a unique cost. Our results reveal the Pareto frontier of optimal solutions for average and high risk patients at different budget levels, which can be used by decision makers to set policies in practice.