Evaluating Population-Based Colorectal Cancer Screening Beyond a Randomized Controlled Trial: A Mathematical Modelling Approach

Although the effectiveness of mass screening for colorectal cancer (CRC), stool-based tests, for example, has been demonstrated by randomized controlled trials (RCTs), whether the same benefits can be similarly observed in population-based organized service screening programs is subject to multiple factors and a complex multistate disease process. Elucidating the natural history of multistate CRC with a mathematical modelling approach can provide an opportunity to test various scenarios involved in population-based organized service screening programs. We first provide rationales and unique characteristics of the modelling approach in contrast to the traditional analysis. We then reviewed a series of stochastic models applied to elucidate the natural history of the disease and to evaluate screening programs for colorectal cancer in the literature. These models cover the traditional homogeneous Markov model, the nonhomogeneous Markov model, and the semi-Markov model. We also demonstrate how the temporal natural history of the disease modeled by the underlying stochastic processes can be applied to different scenarios, including a case-cohort sampling design for elucidating the disease course of adenoma carcinoma pathway, assessment of the efficacy of reducing malignant transformation and the effectiveness of population-based screening programs, decision analysis, and health economic decision models. A mathematical modelling approach is an efficient alternative method for evaluating a series of subsidiary issues of population-based organized service screening dispensing with a randomized controlled trial study or a complex quasi-experimental study that requires the comparator.