Semi-supervised Approaches to Denoising Electronic Health Records Data for Risk Prediction

用于风险预测的电子健康记录数据去噪半监督方法

基本信息

批准号：
10453558
负责人：
TIANXI CAI
金额：
$ 33.74万
依托单位：
HARVARD SCHOOL OF PUBLIC HEALTH
依托单位国家：
美国
项目类别：
财政年份：
2021
资助国家：
美国
起止时间：
2021-08-01 至 2025-04-30
项目状态：
未结题

来源：
https://reporter.nih.gov/project-details/10453558
关键词：
Address Adoption Aftercare Algorithms Biological Clinical Clinical Data Clinical Investigator Clinical Trials Code Cohort Studies Colon Carcinoma Colorectal Cancer Complex Computer software Computerized Medical Record Data Data Set Development Dimensions Disease Disease Outcome Disease Progression Electronic Health Record Face Factor X Genes Goals Gold Healthcare Healthcare Systems Heterogeneity Human Inflammatory Inflammatory Bowel Diseases Label Learning Malignant Neoplasms Massachusetts Measurement Medical Genetics Medical Records Meta-Analysis Methodology Methods Modeling Oncology Outcome Patients Population Procedures Registries Research Research Activity Research Proposals Risk Risk Estimate Risk Factors Site Source Specimen Statistical Algorithm Statistical Methods Supervision System TNF gene Target Populations Testing Training Translational Research Validation adalimumab antagonist base biobank burden of illness clinical application clinical practice cohort colon cancer risk cost cost effective deep learning denoising disorder risk electronic data genomic data high dimensionality improved individualized medicine infliximab large datasets learning strategy machine learning method machine learning model multiple data sources novel outcome prediction patient population patient privacy personalized management personalized predictions personalized risk prediction precision medicine predictive modeling programs repository risk prediction risk prediction model study population supervised learning tool transfer learning treatment response user-friendly

项目摘要

Project Summary While clinical trials remain a critical source for oncology research, their study findings may not be gener- alizable to the real world due to the restricted patient population. In recent years, due to the increasing adoption of electronic health records (EHR) and the linkage of EHR with specimen bio-repositories and other research registries, integrated large datasets now exist as a new source for translational research. These integrated datasets open opportunities for developing accurate EHR-based prediction models for disease progression and treatment response, which can be easily incorporated into clinical practice. These models can also be contrasted with models derived from clinical trials, bridging the gap between clinical trials and the real world. However, efficiently deriving and evaluating personalized prediction models using such real world data (RWD) remains challenging due to practical and methodological obstacles. For example, validated outcome information from EHR, such as development of colon cancer and 1-year treatment response, requires laborious medical record review and hence is often not readily available for research. Naive use of error prone surrogates of the outcome, such as billing codes or procedure codes, as the true outcome may greatly hamper the power of EHR studies and produce biased results. Semi-supervised risk prediction methods, leveraging noisy surrogates and a small amount of human annotations on the outcome, may greatly improve the utility of EHR for precision medicine research. Deriving a precise estimate of the risk model becomes even more challenging when the number of candidate features is not small relative to the number of annotated outcomes. Another major challenge with EHR risk modeling lies in the transportability. Complex machine learning models trained in one EHR system often attain low accuracy in another EHR system, due to the heterogeneity in the patient population and healthcare system. Transfer learning methods that can automatically adjust model developed for one EHR cohort to better fit to another EHR cohort is of great value. Synthesizing information from multiple data sources can improve the quality of evidence. However, meta analyzing EHR from multiple EHR cohorts faces an additional challenge due to patient privacy. We address these challenges by developing semi-supervised risk prediction methods with high dimensional predictions in Aim 1; semi-supervised transfer learning methods to enable risk prediction modeling in target populations with no gold standard labels uted learin Aim 2; and distributed learning methods for high dimensional predictive modeling in Aim.

项目摘要尽管临床试验仍然是肿瘤学研究的关键来源，但他们的研究结果可能并非产生。由于患者人数受到限制，可与现实世界相称。近年来，由于采用的越来越多电子健康记录（EHR）以及EHR与标本生物重新构件和其他研究的联系注册表，集成的大数据集现在是转化研究的新来源。这些整合数据集开放机会开发基于EHR的疾病进展的准确预测模型和治疗反应，可以很容易地将其纳入临床实践中。这些模型也可以是与临床试验得出的模型形成鲜明对比，弥合了临床试验与现实世界之间的差距。但是，使用此类现实世界数据（RWD）有效地得出和评估个性化预测模型由于实用和方法论上的障碍，仍然具有挑战性。例如，验证结果 EHR的信息，例如结肠癌的发展和1年治疗反应，需要因此，经常不容易获得研究的病历审查，因此通常不容易获得研究。幼稚地使用易犯错结果的替代物，例如计费代码或程序代码，因为真正的结果可能会大大妨碍 EHR研究的力量并产生有偏见的结果。半监督风险预测方法，利用关于结果的吵闹代孕和少量人类注释可能会大大改善 EHR用于精确医学研究。得出对风险模型的确切估计变得更加相对于注释结果的数量，候选功能的数量并不小时，具有挑战性。 EHR风险建模的另一个主要挑战在于运输能力。复杂的机器学习模型在一个EHR系统中受过训练通常在另一个EHR系统中获得较低的精度，因为该系统的异质性患者人群和医疗保健系统。可以自动调整模型的转移学习方法为一个EHR队列开发以更好地适合另一个EHR队列。综合信息来自多个数据源可以提高证据质量。但是，元分析来自多个的EHR EHR队列由于患者隐私而面临另一个挑战。我们通过发展来应对这些挑战 AIM 1中具有高维预测的半监督风险预测方法；半监督转移学习方法可以在目标种群中启用风险预测建模，而没有黄金标签标签 Learin Aim 2；以及针对AIM中高维预测建模的分布式学习方法。