Collaborative Research: Optimal Design of a Teleretinal Screening Program for At-Risk Patients

该奖项将通过研究远程视网膜筛查的有效使用和糖尿病视网膜病变综合筛查系统的设计，为促进国民健康和福利做出贡献。糖尿病视网膜病变是美国最常见的糖尿病性眼病，也是美国成年人失明的主要原因。虽然糖尿病相关的视力丧失大多可以通过早期发现和治疗来预防，但全国筛查率明显低于预期，社会经济地位较低的少数民族患者获得筛查的机会受到不成比例的限制。该奖项支持对筛查系统设计的基本理解，该系统利用远程视网膜成像技术来减少差异并增加人口水平的筛查。预计该项目将通过确定具有成本效益的筛查政策和最佳筛查地点，对少数民族和代表性不足的患者的健康和福利产生特别影响。该奖项将资助一名从事跨学科研究的博士生，并为眼科、远程医疗和公共卫生领域的操作研究人员和医疗保健专业人员创建教育材料。本研究将在双层优化框架下对综合筛查系统设计进行建模：(i)将领导者问题建模为概率最大覆盖定位问题，以确定大型县卫生系统中远程终端筛查设施的最佳位置；（ii）将追随者问题建模为单个患者的筛查决策，并将其建模为部分可观察马尔可夫决策过程，以最大化健康效益和患者依从性。由于双层框架的计算挑战，该研究将通过分析低层问题的解决方案结构和价值函数来开发适当的近似，从而实现简单的解决方案形式，并开发基于地理的分布式启发式方法，以近似位置决策和患者级筛查决策之间的相互作用。该研究有望在涉及随机和顺序决策的许多低级问题的双层优化方面取得进展。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

This award will contribute to the advancement of the national health and welfare by studying effective use of teleretinal screening and design of an integrated screening system for diabetic retinopathy. Diabetic retinopathy is the most common diabetic eye disease in the US and the leading cause of blindness in American adults. While diabetes-related vision loss is mostly preventable with early detection and treatment, national screening rates are significantly lower than desired, with disproportionately limited access for minority patients with a low socioeconomic status. This award supports a fundamental understanding of screening system design that utilizes teleretinal imaging technology to reduce the disparities and increase population-level screening. This project is expected to have a particular impact on the health and welfare of minority and underrepresented patients by identifying cost-effective screening policies and optimal screening locations. This award will fund a doctoral student in inter-disciplinary research and create educational materials for both operations researchers and healthcare professionals in ophthalmology, telehealth, and public health.This research will model an integrated screening system design in a bilevel optimization framework: (i) the leader's problem is modeled as a probabilistic maximal covering location problem to identify optimal locations of teleretinal screening facilities in a large county health system, and (ii) the followers' problems address screening decisions for individual patients and are modeled as partially observable Markov decision processes to maximize both health benefit and patient adherence. Because of the computational challenges of the bilevel framework, the research will develop appropriate approximations by analyzing solution structures and value functions of the lower-level problems that enable simple solution forms and developing geography-based, distributed heuristics that approximate the interaction between the location decisions and patient-level screening decisions. The research is expected to lead to advances in bilevel optimization where there are many lower-level problems involving stochastic and sequential decision making.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.