Using mathematical modelling to inform policy making for bowel cancer early diagnosis Seminar

The early diagnosis of cancer is a topic with the potential for large impacts in terms of saving both money and lives. Understanding of cancer progression rates is limited and the use of clinical trial to determine optimum screening strategies is often unfeasible. Hence, mathematical modelling can be extremely useful to inform policy making in this area. Unknowns in bowel cancer screening (disease progression rates, symptomatic presentation rates, and screening test characteristics) were estimated using a mathematical model. The technical details of these methods are described including: model calibration, Bayesian approach, Metropolis Hastings algorithm. Further model development to relax the assumption of constant screening test characteristics is also explored. A summary of results from several studies using the ScHARR bowel cancer screening model are presented including: Bowel cancer screening options appraisal; Bowel cancer early awareness campaign assessment; Bowel cancer chemoprevention HTA; Screening age range expansion. The conclusions of a study looking at differences between outcomes in the first two round of the screening program are also presented. Mathematical modelling can provide data to inform policy making in early cancer diagnosis. Challenges of modelling in this area are discussed including: differences between model predictions and real world observations; and modelling several interventions simultaneously. Additional data availability is expected to allow more sophisticated models to be developed in the future.