OBJECTIVES : Different health economic (HE) models are used to simulate real-world practice. The choice of model depends on the disease arearesearch question, and model objectives. In cancer screening tests, two common issues are that the risk of cancer changes dramatically over the patient’s lifetime and that cancer screening protocols are usually a combination of two tests (eg, initial screening and follow-up testing). There are limitations to properly addressing these two issues in HE models. The present study aims to investigate which type of HE models can appropriately simulate the real-world practice of cancer screening in ovarian cancer, colorectal cancer (CRC), and prostate cancer.

METHODS : A targeted search of economic studies in screening tests was conducted for ovarian cancer, CRC, and prostate cancer. The economic modeling data from manuscripts published between 2015 and 2021 were extracted, and the limitations and benefits of different types of HE models were evaluated.

RESULTS : Forty-six studies met the inclusion criteria. Partitioned survival models (PSMs) were not used in any included studies. Model types were decision tree (4.3%), Markov (56.5%), and microsimulation (39.1%). Microsimulation models, including discrete event simulation, were commonly used in screening modelling for CRC and prostate cancer but not ovarian cancer.

CONCLUSIONS : Decision tree and Markov models are useful when a decision problem involves a continuous risk level over time, but when the risk of tumor varies over time, these types of models cannot properly capture this effect since they consider a constant risk of malignancy for all included individuals. PSM has a limitation in modelling conditions with multiple health states; consequently, none of the included studies used this method. Microsimulation models can effectively capture changes over time; moreover, they can precisely simulate the interaction between two tests in multimodal screening methods.