Collaborative Research: Mathematical Modeling of Dynamic Breast Cancer Screening

这项拨款为开发一个数学模型提供资金，以检验是否应该在绝经前和绝经后使用不同的乳房x光检查间隔。对该模型的分析将提供以下问题的见解：就死亡率风险而言，在绝经前和绝经后阶段使用不同的乳房x光检查间隔有多大价值？应该在绝经前还是绝经后更频繁地进行筛查？起始年龄/停止年龄筛查如何影响死亡风险？考虑到绝经前后乳腺癌发病率、乳腺癌侵袭性和乳房x光检查效果的动态特性，这些问题特别有趣。使用部分观察的马尔可夫决策过程模型，将评估不同的筛选政策，并构建信息权衡曲线，绘制“政策努力”与死亡风险的一些度量。注意将仅限于两种类型的政策：“常规”政策规定在患者一生中进行相同的筛查间隔，以及“两阶段”政策规定绝经前和绝经后的不同间隔。这项研究的结果将通过赋予患者权力和改善公共政策而潜在地造福社会。权衡曲线将允许女性根据个人风险偏好、经济状况、保险范围、发病率倾向等量身定制自己的乳房x光检查选择。权衡曲线将以一种既可理解又对“非科学家”有用的格式呈现筛查政策与死亡风险之间的定量关系。如果成功，这项研究的结果也可能影响美国癌症协会对乳腺癌筛查的政策建议。

This grant provides funding for the development of a mathematical model to examine whether different mammogram screening intervals should be used pre-menopause and post-menopause. Analysis of this model will provide insight into the following questions: how worthwhile is it, in terms of mortality risk, to use different mammogram intervals in the pre- and post-menopausal phases? Should screening be more frequent pre- or post-menopause? How does screening start-age/stop-age affect mortality risk? These questions are particularly interesting given the dynamic nature of breast cancer incidence, breast cancer aggression, and mammogram-efficacy between the pre- and post-menopausal phases. Using a partially observed Markov decision process model, different screening policies will be evaluated, and informative tradeoff curves that plot some measure of "policy effort" versus mortality risk will be constructed. Attention will be restricted to two types of policies: "routine" policies that prescribe the same screening interval over the life of the patient, and "two-phase" policies that prescribe one interval pre-menopause and a different interval post-menopause. The results of this research will potentially benefit society through patient-empowerment and improved public policy. The tradeoff curves will allow women to tailor their own mammogram screening choices based on individual risk preference, economic status, insurance coverage, incidence propensity, etc... The tradeoff curves will present the quantitative relationship between screening policy and mortality risk in a format that is both understandable and useful for "non-scientists." If successful, the outcomes of the proposed research may also impact American Cancer Society policy recommendations for breast cancer screening.