Date of Award

Spring 5-25-2019

Document Type


Degree Name

Master of Science (MS)

First Advisor

Mahboubeh Madadi


Colorectal cancer (CRC) is one of the deadliest types of cancer in the US due to its high incidence and mortality rates. Detection of CRC in the early stages through available screening tests increases the patient's survival chances. In this study, we investigate the cost-effectiveness of a wide variety of multi-modal CRC screening policies. More specifically, we develop a Monte Carlo simulation framework to model the CRC natural history and preventive interventions. Age-specific and size-specific progression rates of adenomatous polyps are estimated using an innovative active learning method. Specifically, we develop a decision tree model to estimate size-specific and age-specific adenoma progression and regression rates. Compared to traditional methods, the proposed calibration process expedites the searching of the model parameter space significantly. CRC age-specific incidence rates and CRC stage distribution are the two output measures used in the calibration process. Seventy-eight CRC screening policies are applied to a cohort of U.S. male population using the simulation model and compared in terms of expected Quality Adjusted Life Years (QALY) and costs. Eleven policies are identified as efficient frontier policies. Among these 9 are identified as cost- effective at the willingness to pay (WTP) threshold of $50,000. Fecal Occult Blood Test (FOBT) biennially in conjunction with one time Colonoscopy at 60, FOBT biennially along with one time Colonoscopy at 50, Fecal Immunochemical Test (FIT) biennially in conjunction with two times Flexible Sigmoidoscopy (FS) at 60 and 65. FIT biennially with one time Colonoscopy at 65, Colonoscopy at 50, 60 and 70, FOBT biennially along with two times Colonoscopy at 55 and 65, FOBT annually with 2 times FS at 70 and 75, FOBT annually in conjunction with FS at 50 and 55, and FIT biennially along with FS every 5 years are the nine identified cost-effective policies.