Christine Barnett, PhD, Associate Director of Health Economics at RTI Health Solutions, proudly announces publication of Artificial Intelligence for Healthcare: Interdisciplinary Partnerships for Analytics-Driven Improvements in a Post-COVID World. This book, designed to be accessible to a diverse audience, provides an overview of interdisciplinary research partnerships that leverage artificial intelligence, industrial engineering, and operations research to tackle societal and operational problems in healthcare.
Cancer Screening Policies
In Chapter 7, Optimization of Biomarker-Based Prostate Cancer Screening Policies, coauthor Dr. Barnett discusses cancer screening, which has the potential to improve patient survival. However, cancer screening tests are imperfect, the screening process carries risks, and decisions are often made based on a single test result. Therefore, policies are necessary to optimize the potential benefits of screening.
The authors outline that many variables factor into the decision of how, if at all, a screening test should be used, including the following: test reliability, potential for harm caused by the screening process, the likelihood of a high-risk cancer, and the health of the patient. To optimize decisions that take those factors into account, the authors developed a partially observable Markov decision process (POMDP) model that maximizes quality-adjusted life-years (QALYs) while taking into account the patient’s full medical history to get a better estimate of a patient’s risk for cancer.
The authors found that the methods they describe in the chapter to develop screening policies would result in significant health benefits for patients. However, POMDPs are complex “blackbox” models, which can present challenges for implementation in practice. To address that challenge, the authors refer to Chapter 4 of the book, which covers how to integrate predictive analysis with patient decision making.
Barnett CL, Denton BT. Optimization of biomarker-based prostate cancer screening policies. In: Sze-chuan Suen, David Scheinker, Eva Enns, editors. Artificial intelligence for healthcare: interdisciplinary partnerships for analytics-driven improvements in a post-COVID world. Cambridge: United Kingdom. Cambridge University Press; 2022. p.141-58. ISBN: 9781108836739