Multi-modal Health Information Technology Innovations for Precision Management of Glaucoma

项目总结/摘要 青光眼是世界上导致不可逆失明的主要原因，将影响超过1.1亿人 人到2040年。早期发现和治疗至关重要，因为症状通常不存在 直到病情恶化需要一种数据驱动的精准医学方法， 确定谁是在发展疾病的最大风险和谁是在最大的个人 迅速发展为视力丧失的风险。虽然在眼科方面取得了相当大的进展， 成像和测试，以改善青光眼监测，青光眼的精确管理， 不完整，未考虑患者同时存在的全身性疾病，并发全身性疾病 药物和治疗，以及坚持处方青光眼治疗。 了解全身状况，特别是血管状况， 高血压、嵌塞性青光眼在公共卫生中的重要性日益增加， 老年人面临的共病问题初步研究表明， 系统数据的价值，即使没有眼科终点。同样，测量药物 坚持对于指导患者咨询和参与以及避免 下游干预措施，如手术，费用高，发病率高。这些因素 对于提供更全面的青光眼管理观点很重要， 改善病人的治疗效果，但它们的研究相对不足。 我建议应用卫生信息技术（IT）的多模式进步， 这些差距，并实现以下具体目标：（1）开发基于机器学习的 使用系统电子显微镜对青光眼进展风险患者进行分类的预测模型 健康记录（EHR）数据来自不同的全国患者队列;（2）评估如何整合 来自新型智能手表家庭血压监测器的血压（BP）数据增强了预测 用于青光眼危险分层的模型，以及（3）测量青光眼药物依从性 使用创新的柔性电子传感器，以验证其在未来干预措施中的用途， 改善青光眼的依从性和临床结果。这些研究将利用国家- 大数据预测建模的最先进方法以及 传感器技术这种多方面的方法将为健康IT奠定基础 旨在改善风险分层和产生新的治疗靶点的框架 对于青光眼患者。

PROJECT SUMMARY/ABSTRACT Glaucoma is the world's leading cause of irreversible blindness and will affect >110 million people by 2040. Early detection and treatment are critical, as symptoms typically do not present until the disease is advanced. A data-driven precision medicine approach is needed to better identify individuals who are at greatest risk of developing the disease and who are at greatest risk of progressing quickly to vision loss. While there has been considerable progress in eye imaging and testing to improve glaucoma monitoring, precision management of glaucoma is incomplete without accounting for patients' co-existing systemic conditions, concurrent systemic medications and treatments, and adherence with prescribed glaucoma treatment. Understanding how systemic conditions, and specifically vascular conditions such as hypertension, impact glaucoma presents growing public health importance given the increasing co-morbidities facing aging populations. Preliminary studies have demonstrated the predictive value of systemic data, even without ophthalmic endpoints. Similarly, measuring medication adherence is important for guiding patient counseling and engagement and avoiding downstream interventions such as surgeries, which carry high cost and morbidity. These factors are important for providing a more comprehensive perspective of glaucoma management and for improving patient outcomes, yet they are relatively understudied. I propose applying multi-modal advancements in health information technology (IT) to address these gaps and achieve the following specific aims: (1) Develop machine learning-based predictive models classifying patients at risk for glaucoma progression using systemic electronic health record (EHR) data from a diverse nationwide patient cohort; (2) evaluate how integrating blood pressure (BP) data from novel smartwatch-based home BP monitors enhance predictive models for risk stratification in glaucoma, and (3) measure glaucoma medication adherence using innovative flexible electronic sensors to validate their use for future interventions aimed at improving adherence and clinical outcomes in glaucoma. These studies would leverage state- of-the-art methods in big-data predictive modeling as well as cutting-edge advancements in sensor technologies. This multi-faceted approach will build a foundation for a health IT framework geared toward improving risk stratification and generating novel therapeutic targets for glaucoma patients.