Automated system to improve compliance to diabetic retinopathy screening

总结 这项资助的目的是开发一个完全自动化的电子健康记录（EHR）集成软件 应用程序“DR-SCRN”，以提高患者对糖尿病视网膜病变（DR）筛查的依从性;通过预测 基于对患者自身健康数据的分析，对患者进行教育， 通知提供者患者的DR风险和筛查需求。DR-SCRN将分析现有的 EHR中患者健康数据，以识别和预测未来3年的DR风险趋势。核心创新 该项目的主要内容是：a）展示一个完全自动化和EHR集成的工具来预测患者的DR风险。B） 通过用视觉和数字数据教育患者来提高筛查依从性，c）计算DR风险 基于患者的健康数据，容易从EHR获得，d）开发软件以通知提供者患者的 筛查需求，并协助规划推荐的筛查时间表。 DR-SCRN的主要动机是通过预测患者的DR， 风险使用自己的健康数据，并使自己和他们的护理提供者意识到潜在的风险。博士 通过定期筛查和及时干预，早期发现是可以预防的。尽管DR筛查 由于糖尿病患者的眼科检查很容易获得，只有18 - 40%的糖尿病患者接受了推荐的年度眼科检查。 对DR筛查的依从性导致视力丧失，并且由于视力丧失治疗而给美国带来43亿美元的负担， 手术和生产力下降。低依从性源于患者的无知、经济限制， 缺乏接触，缺乏症状或知识，视力丧失与糖尿病有关。患者的 无知和缺乏知识是最常见的障碍，即使其他障碍被最小化。DR- SCRN试图通过促进对患者和提供者的健康教育来提高依从性 条件和潜在的风险，并提供一个无缝的系统，为患者和供应商，使 有需要的人容易获得筛查检查。 该项目的目标将通过三个具体目标实现。在目标1中，我们将开发一个 使用N=5000的回顾性数据集，基于患者健康数据计算DR风险的自动算法 从VisionQuest专有的视网膜筛查数据库中选择的糖尿病患者。在目标2中，我们将开发 外推或趋势估计算法，用于使用 N=2500名糖尿病患者的独立纵向数据库。在目标3中，我们将编写教育材料， 患者和提供者，以提高筛查依从性，提供：a）数字DR风险预测超过 未来3年，B）DR风险如何变化的直观表示，c）拟定的DR筛查计划。 在第二阶段，我们将追求三个目标：（1）大规模的纵向DR筛查临床研究， (2)使用临床研究数据验证算法和软件应用程序，（3）开发桥接- 将DR-SCRN与EHR系统集成的软件，以演示完整的应用程序原型。我们 第二阶段成功的主要衡量标准是将DR-SCRN集成到我们现有网络的EHR中 墨西哥和Retina Global的诊所已经在使用我们的自动糖尿病视网膜病变筛查， 系统

Summary The objective of this grant is to develop a fully automated, Electronic Health Record (EHR)-integrable software application, “DR-SCRN”, to improve patient’s compliance to diabetic retinopathy (DR) screening; by predicting a patient’s risk of DR based on the analysis of patient’s own health data, and educating the patient as well as notifying the provider of the patient’s DR risk and screening needs. DR-SCRN will analyze the readily available health data of patient from EHR to identify and predict a trend of DR risk over next 3 years. The core innovations of this project are: a) demonstrate a fully automated and EHR-integrated tool to predict patient’s DR risk. b) improve screening compliance by educating the patient with visual and numerical data, c) calculate DR risk based on patient’s health data, readily available from EHR, d) develop software to notify the provider of patient’s screening needs and to assist with planning of the recommended screening schedule. The main motivation for DR-SCRN is to improve patient’s compliance to DR screening by predicting their DR risk using their own health data and making themselves and their care-provider aware of the potential risks. DR is preventable with early detection through periodic screening and timely intervention. Although DR screening examinations are readily available, only 18 to 40% of diabetics undergo the recommended annual eye exam. Low compliance to DR screening results in vision loss and a $4.3 billion burden to the US due to vision loss treatment, surgery, and diminished productivity. Low compliance results from patient’s ignorance, financial constraints, lack of access, and lack of symptoms or knowledge that vision loss was associated with diabetes. Patient’s ignorance and lack of knowledge have been the most common barriers even when others are minimized. DR- SCRN attempts to improve the compliance by promoting the education of patients and providers on health condition and potential risks and providing a seamless system for both patients and providers to make the screening examinations easily accessible to the population in need. The objectives of this project will be accomplished through three specific aims. In aim 1, we will develop an automatic algorithm to calculate DR risk based on patient’s health data, using a retrospective dataset of N=5000 diabetic patients selected from VisionQuest’s proprietary retinal screening database. In aim 2, we will develop an extrapolation or trend estimation algorithm to predict future DR risk based current DR risk trend, using a separate longitudinal database of N=2500 diabetic patients. In aim 3, we will develop educational material for patients and providers to improve screening compliance, that provides: a) Numerical DR risk prediction over next 3 years, b) Visual representation of how the DR risk may change, c) Proposed DR screening schedule. Further in Phase II, we will pursue three objectives: (1) a large-scale, longitudinal DR screening clinical study, (2) validation of the algorithms and software application using the clinical study data, (3) developing a bridging- software to integrate DR-SCRN with an EHR system, to demonstrate a complete application prototype. Our primary measure of success in Phase II will be the integration of DR-SCRN into the EHR of our existing network of clinics in Mexico and at Retina Global, which are already using our automatic diabetic retinopathy screening system.