SMILE-PD: Similarity Matching In Longitudinal Electronic Patient Data

SMILE-PD：纵向电子患者数据中的相似性匹配

• 批准号: 10799090
• 负责人: MARK G WEINER
• 金额: $ 19.07万
• 依托单位: WEILL MEDICAL COLL OF CORNELL UNIV
• 依托单位国家: 美国
• 财政年份: 2023
• 资助国家: 美国
• 起止时间: 2023-09-06 至 2025-07-31
• 项目状态: 未结题
• 来源: https://reporter.nih.gov/project-details/10799090
• 关键词: Affinity; Algorithms; All of Us Research Program; Caring; Cells; Characteristics; Chemical Structure; Chronic Disease; Clinical; Clinical Data; Comparative Effectiveness Research; Computational algorithm; Computerized Medical Record; Consumption; DNA; Data; Data Set; Data Sources; Diagnosis; Disease; Disease Outcome; Electronics; Environmental Risk Factor; Etiology; Evaluation; Generations; Genetic; Genomics; Goals; Healthcare; Heart failure; Hospitalization; Hypothyroidism; Individual; Information Resources; Investigation; Learning; Link; Location; Manuals; Measures; Medical; Mental Depression; Mental disorders; Modeling; Non-Insulin-Dependent Diabetes Mellitus; Observational Study; Outcome; Patient-Focused Outcomes; Patients; Peripheral arterial disease; Pharmaceutical Preparations; Phenotype; Population; Predictive Analytics; Predisposition; Prevention strategy; Process; Public Health Informatics; Records; Reproducibility; Research; Research Personnel; Resources; Risk; Risk Assessment; Smiling; Specific qualifier value; Time; United States National Institutes of Health; Visualization; Work; biobank; cohort; comorbidity; demographics; effectiveness evaluation; endophenotype; genome wide association study; genomic data; improved; interest; learning algorithm; learning strategy; lifestyle factors; open source; precision medicine; side effect; sound; trait; treatment strategy; vector

PROJECT SUMMARY The NIH All of Us (AoU) Research Program is a national research initiative launched by the National Institutes of Health (NIH) in May 2018 with the goal of advancing precision medicine, which aims to provide personalized medical treatment and prevention strategies based on an individual's unique genetic, environmental, and lifestyle factors. Analysis of real world data like AoU typically requires creation of cohorts that are clinically similar with the exception of the characteristic(s) being evaluated. If the creation of compared cohorts is not conducted with care, results may be confounded by systematic difference in the characteristics of patients assigned to the compared cohorts. This proposal, Similarity Matching In Longitudinal Electronic Patient Data (SMILE PD), will apply a sophisticated patient matching algorithm within an easy- to-use interface to help Investigators to create suitable analytical cohorts to enable generation of accurate, reproducible real world evidence from the rich content of the All of Us data. The project has 4 Aims : • Aim 1. Develop algorithms that learn patient similarities from AoU data in a comprehensive manner. • Aim 2: Develop approaches that can integrate drug and disease information from publicly available data sources to improve the quality of learned patient similarities. • Aim 3: Implement an interactive user interface to specify similarity matching criteria and demonstrate the learned patient similarities • Aim 4 : Apply the patient similarity matching algorithm and interface to conduct an initial assessment of the impact of mental illness and its treatment on chronic disease outcomes The patient similarity matching interface created through this proposal will enable investigators to conduct predictive analytics, risk assessments and comparative effectiveness research in a more robust and reproducible fashion.

项目概要 NIH All of Us (AoU) 研究计划是一项由美国国立卫生研究院 (NIH) 发起的国家研究计划 美国国立卫生研究院 (NIH) 于 2018 年 5 月以推进精准医学为目标， 旨在基于个体化治疗和预防策略 个人独特的遗传、环境和生活方式因素的分析，例如现实世界的数据。 AoU 通常需要创建临床上相似的队列，但以下情况除外： 如果没有进行比较队列的创建。 护理，结果可能会因患者特征的系统差异而混淆 分配给比较队列的提案是纵向电子中的相似性匹配。 患者数据 (SMILE PD) 将在一个简单的环境中应用复杂的患者匹配算法 使用界面可帮助研究人员创建合适的分析队列，从而生成 来自“我们所有人”数据丰富内容的准确、可重复的现实世界证据。 该项目有 4 个目标： • 目标 1. 开发算法，从 AoU 数据中了解患者的相似性 综合方式。 • 目标 2：开发能够整合来自以下来源的药物和疾病信息的方法： 公开可用的数据源，以提高了解患者相似性的质量。 • 目标 3：实现交互式用户界面来指定相似性匹配标准 并展示所学到的患者相似之处 • 目标 4：应用患者相似性匹配算法和界面来进行 精神疾病影响及其对慢性病治疗的初步评估 结果 通过该提案创建的患者相似性匹配界面将使研究人员能够 更深入地进行预测分析、风险评估和比较有效性研究 稳健且可复制的时尚。