A Two-stage Stochastic Programming Model to Determine the Optimal Screening Strategy for Colorectal Cancer

Screening for colorectal cancer (CRC) is an effective way to drastically reduce the impact of the disease or prevent it altogether. This paper presents a stochastic mathematical programming model to determine the optimal screening strategy for CRC of a given population. The objective of the model is to maximize the expected quality adjusted life years an individual would gain by following the optimum screening strategy. The model incorporates the uncertainty of CRC progression through the use of the time taken to progress to the various stages of the disease. The data to model the uncertainty of the progression of CRC within an individual was obtained from a continuous time simulation. The solution of the stochastic programming model for the average-risk male population yielded an expected gain of 0.2384 quality-adjusted life years with three colonoscopies.