Leveraging deep learning and clinical notes for surveillance and prediction of intentional self-harm and suicide

利用深度学习和临床记录来监测和预测故意自残和自杀

基本信息

批准号：
10330113
负责人：
Jihad S Obeid
金额：
$ 55.18万
依托单位：
MEDICAL UNIVERSITY OF SOUTH CAROLINA
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：
2021-05-01 至 2023-04-30
项目状态：
已结题

来源：
https://reporter.nih.gov/project-details/10330113
关键词：
Address Adoption Area Artificial Intelligence Attention Categories Cause of Death Classification Clinical Code Collaborations Computer Vision Systems Data Development Disease Effectiveness Electronic Health Record Event Feeling suicidal Frequencies Future Goals Health Care Visit Health Personnel Health system Healthcare Systems Individual International Classification of Disease Codes Intervention Lead Linguistics Link Machine Learning Manuals Methods Modeling Network-based Outcome Patients Phenotype Pilot Projects Prevention Psychological Transfer Reproducibility Research Research Domain Criteria Research Personnel Risk Risk Factors Self-Injurious Behavior Semantics Site Structure Suicide Suicide attempt Suicide prevention Testing Text Training United States Work base behavior prediction deep learning deep neural network effective intervention electronic structure experience feature selection health care settings health data health record improved machine learning method mortality natural language novel outcome prediction predict clinical outcome predictive modeling prospective provider adoption screening speech recognition structured data suicidal behavior suicidal morbidity suicidal risk suicide mortality tool trustworthiness

项目摘要

PROJECT SUMMARY: Suicide is one of the leading causes of death in the United States, with more than 47,000 individuals dying by suicide each year. The identification of individuals at risk for suicide is an important step for a comprehensive approach to suicide prevention. Despite extensive research on risk factors for intentional self-harm and suicide, prospective prediction of suicide remains a difficult task with poor predictive power. Recent studies suggest that new machine learning methods applied to electronic health records (EHR) show promising results. However, more advanced computational approaches such as deep learning, have not been fully leveraged in this field, especially in the area of advanced methods for text classification of clinical notes. Our aims in this project, are to improve the phenotyping of suicidal behavior, and the prediction of future suicidal behavior and suicide deaths by integrating mortality data with EHR data and leveraging state-of-the-art natural language computational approaches. We will also investigate methods for explain ability and interpretability of the models to improve future adoption by clinicians. We will validate our models by examining reproducibility and generalizability across two health systems using similar data at both sites.

项目摘要：自杀是美国的主要死亡原因之一，超过47,000 每年因自杀而死的人。确定有自杀风险的人是全面预防自杀方法的重要步骤。尽管进行了广泛的研究关于故意自我伤害和自杀的风险因素，自杀的前瞻性预测仍然是艰巨的预测能力任务。最近的研究表明新机器学习应用于电子健康记录（EHR）的方法显示出令人鼓舞的结果。但是，更多深度学习等高级计算方法尚未完全利用该领域，尤其是在临床注释文本分类的高级方法领域。我们的该项目的目的是改善自杀行为的表型，并预测通过将死亡率数据与EHR数据整合到未来的自杀行为和自杀死亡和自杀死亡利用最新的自然语言计算方法。我们还将调查解释模型的能力和解释性的方法，以改善未来的采用临床医生。我们将通过检查整个范围的可重复性和概括性来验证我们的模型在两个站点都使用类似数据的两个卫生系统。