Personalized Risk Predictions with Deep Learning Methods in the Presence of Missing and Biased Electronic Health Record Data

在存在缺失和有偏差的电子健康记录数据的情况下，利用深度学习方法进行个性化风险预测

• 批准号: 10646324
• 负责人: Padhraic Smyth
• 金额: $ 33.07万
• 依托单位: NEW YORK UNIVERSITY SCHOOL OF MEDICINE
• 依托单位国家: 美国
• 财政年份: 2021
• 资助国家: 美国
• 起止时间: 2021-08-06 至 2025-05-31
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10646324
• 关键词: Age; Algorithms; Benchmarking; Biological Markers; Calibration; Cardiovascular Diseases; Characteristics; Chronic Disease; Clinical; Clinical Data; Clinical Medicine; Clinical Research; Computers; Computing Methodologies; Data; Data Set; Development; Diagnosis; Disease; Disease Outcome; Disparate; Drug Prescriptions; Electronic Health Record; Frequencies; Gender; Health; Health Surveys; Healthcare; Hospitals; Individual; Institution; Interview; Link; Medical; Methodology; Methods; Modeling; Neural Network Simulation; New York; New York City; Non-Insulin-Dependent Diabetes Mellitus; Outcome; Patients; Physical assessment; Population; Probability; Procedures; Race; Records; Reproducibility; Research; Research Personnel; Retirement; Risk; Risk Factors; Sample Size; Sampling; Scientist; Site; Software Tools; Statistical Methods; Surveys; System; Universities; Validation; Variant; Visit; cohort; data modeling; data standards; deep learning; demographics; electronic health data; experience; flexibility; health care service utilization; improved; innovation; interest; learning strategy; maltreatment; patient population; patient subsets; personalized risk prediction; population based; population health; predictive modeling; recurrent neural network; risk prediction; risk prediction model; web app

Abstract Since 2010, clinical medicine has benefited from a rapid surge of clinical research on chronic diseases using data from electronic health records (EHRs). EHRs are appealing because they can offer large sample sizes, timely information, and a wealth of clinical information beyond that obtained from either health surveys or administrative data. However, while millions of patient records are included in large EHR records, they are not population-representative random samples, a constraint that potentially biases inferences based on such data and, therefore, has limited their utility for population health research. EHR data typically contain multiple types of biases, particularly: 1) sampling inclusion bias: EHR data only include information on patients visiting participating medical systems, and they primarily capture data when patients are ill. Even among populations with a particular disease, patients represented in EHRs tend to over-represent individuals who are sicker and have higher health care utilization; 2) sampling frequency bias: the numbers of patients’ encounters and features in EHRs are at various frequencies and these frequencies correlate with both patients’ characteristics and outcomes; and 3) institution bias: EHR samples of any hospital reflect the characteristics of patients population served by that specific hospital. Consequently, EHR-based risk prediction models will have 1) biases in risk factor selection and estimation for population inferences; 2) disparate mistreatment (unfairness) in terms of variation in a model’s prediction accuracy across patient subgroups (such as gender, race, and age) with various sampling inclusion probabilities or frequencies; 3) biased prediction model to reflect characteristics of patients served by the local hospitals. We propose to develop: 1) effective sample-weighting method to correct biases in risk factor selection and estimation for population inferences (Aim 1), 2) flexible deep learning method for EHR personalized risk prediction with fairness criteria (Aim 2); and 3) innovative calibration method to improve reproducibility of EHR-based risk models between institutions (Aim 3). We will predict risk of subsequent incident cardiovascular disease (CVD) in patients with type 2 diabetes (T2DM) as a demonstration of methodology development. Broader use of these methods will be generally applicable to other diseases outcomes and population of interest. To develop and validate these methods, we propose to analyze three unique datasets: 1) the New York University Langone Health EHR data (NYU-CDRN, 2009 to now) including demographics, vitals, diagnoses, lab results, prescriptions, and procedures; 2) the New York City Clinical Data Research Network (NYC-CDRN)—an EHR network comprising 20 NYC healthcare institutions, including the NYU-CDRN, with longitudinally linked data on >12 million patient encounters under a Common Data Model, and 3) the Health and Retirement Survey (HRS, begun in 1992 and ongoing), as a benchmark population- based cohort, that has nationally representative health interview data for over 20 years, as well as biomarkers, physical assessment information, prescription drug data, and claims linkages.

抽象的 自2010年以来，临床医学受益于慢性病临床研究的快速增长。 来自电子健康记录（EHR）的数据很有吸引力，因为它们可以提供大样本量， 及时的信息以及丰富的临床信息，超出了从健康调查或健康调查中获得的信息 然而，虽然大型 EHR 记录中包含数以百万计的患者记录，但它们并未包含在内。 具有代表性的随机样本，这是一种可能使基于此类数据的推论产生偏差的约束 因此，电子病历数据通常包含多种类型，因此限制了它们在人口健康研究中的效用。 偏差，特别是：1）抽样包含偏差：EHR 数据仅包含就诊患者的信息 参与的医疗系统，它们主要在患者生病时收集数据，甚至在人群中也是如此。 对于某种特定的疾病，电子病历中所代表的患者往往会过多地代表病情较重且病情较重的患者。 具有较高的医疗保健利用率；2）抽样频率偏差：患者就诊的数量和 EHR 中的特征具有不同的频率，这些频率与患者的特征相关 和结果；3) 机构偏差：任何医院的 EHR 样本都反映患者的特征 该特定医院所服务的人群，基于 EHR 的风险预测模型将具有 1) 风险因素选择和群体推断估计存在偏差；2) 不同的虐待（不公平） 模型在患者亚组（例如性别、种族和年龄）之间的预测准确性的差异 具有不同的采样包含概率或频率；3）有偏差的预测模型来反映特征 我们建议开发：1）有效的样本加权方法。 纠正风险因素选择和总体推断估计中的偏差（目标 1），2）灵活的深度学习 具有公平性标准的 EHR 个性化风险预测方法（目标 2）和 3）创新校准方法； 提高机构之间基于 EHR 的风险模型的可重复性（目标 3）。 以 2 型糖尿病 (T2DM) 患者随后发生的心血管疾病 (CVD) 为例 这些方法的更广泛应用将普遍适用于其他疾病。 为了开发和验证这些方法，我们建议分析三种方法。 独特的数据集：1) 纽约大学 Langone Health EHR 数据（NYU-CDRN，2009 年至今），包括 人口统计、生命体征、诊断、实验室结果、处方和程序 2) 纽约市临床数据； 研究网络 (NYC-CDRN) — 由 20 家纽约市医疗机构组成的 EHR 网络，其中包括 NYU-CDRN 在通用数据模型下拥有超过 1200 万患者就诊的纵向关联数据， 3) 健康与退休调查（HRS，1992 年开始并持续进行），作为基准人口- 基于队列，拥有 20 多年来具有全国代表性的健康访谈数据以及生物标志物， 身体评估信息、处方药数据和索赔链接。