Maternal Morbidity and Mortality: Risk Factors, Early Detection and Personalized Intervention

项目摘要 在美国，从2011年到2014年，有7,208例孕产妇死亡，这一趋势逐年恶化（CDC， 2016）。除了700多人死亡外，至少有50，000名妇女经历了危及生命的并发症， 每年。根据CDC（2019）的数据，每死亡一人，就有70多名妇女遭受可避免的创伤， 怀孕引起的并发症。 Medstar健康研究所和Invaryant公司提出了心脏风险评估 孕妇和产后妇女的评估工具;该工具直接从患者那里自动更新 医疗记录;可穿戴设备;和患者调查。成功意味着女性在性别方面的颠覆性改善， 健康拟议的研究涉及三个关键要素：技术、数据和健康的社会决定因素 （SDOH）使用患者位置的地理空间映射。 技术：研究技术应基于Invaryant Health Platform（IHP），该技术 使用专有AI从医疗记录、可穿戴设备和其他来源自动摄取数据 基于互操作性的技术称为网格复杂方法交换（Mesh-CMX）。我们建议使用 IHP结合新的原型级技术，即“人人享有健康成果” 经验-心血管风险评估技术（HOPE-CAT）和不变的机器学习 技术来监测病人，基于信号，范围外的“绊网”，和母亲的趋势， 需要医疗干预健康数据。通过将概念验证扩展到早期商业版本 将其集成到IHP中，IHP将自动更新患者医疗信息的变化， 根据以往的记录，我们将为母亲及其提供者提供一个“预警”系统。 数据：本研究将使用MedStar的分析平台（MAP）对患者的医疗记录进行测试， 超过500万独特患者的登记。该工具随后将有可能被利用， 为MedStar和Cerner医院系统提供9000万份医疗记录， 资格标准包括性别、年龄、种族、妊娠和医疗结果。第二阶段 该项目将采用回顾性研究（本赠款申请）的结果，并在 Medstar医院系统，以验证在“现实世界的设置”的结果的有效性。使用我们 专有的人工智能技术，我们将比较每个母亲的进展与一组回顾性数据， 增强诊断并向护理人员和患者提供实时反馈。 地理空间映射：将患者的医疗记录和健康信息映射到他们的社会环境， 这对了解影响不同地区母亲健康的潜在社会结构至关重要。

PROJECT ABSTRACT In the U.S., from 2011 to 2014, 7,208 maternal fatalities, with the trend worsening year-on-year (CDC, 2016). In addition to the 700+ fatalities, at least 50,000 woman experienced life-threatening complications, annually. According to CDC (2019) for every fatality, 70 more women suffer avoidable, traumatic complications as a result of pregnancy. Medstar Health Research Institute and Invaryant, Inc, propose the evaluation of a cardiac risk assessment tool for pregnant and postpartum women; this tool updates automatically directly from patient medical records; wearable devices; and patient surveys. Success implies disruptive improvement in women's health. Proposed research involves three key elements: technology, data, and social determinants of health (SDOH) using geospatial mapping of patient locations. Technology: Study technology shall be based on the Invaryant Health Platform (IHP), a technology that automatically ingests data, from medical records, wearable devices, and other sources using proprietary AI based interoperability technology called Mesh-Complex Method Exchange (Mesh-CMX). We propose using the IHP in conjunction with novel prototype-level technology, namely Healthy Outcomes for all Pregnancy Experiences-Cardiovascular-risk Assessment Technology (HOPE-CAT) and the Invaryant machine learning technologies to monitor the patient, based on signals, out-of-range “trip-wires”, and trends in the mother's health data that merit medical intervention. By extending the proof of concept into an early commercial version of the software, and integrating it to the IHP which will automatically update changes in the patient's medical record, we will provide an “early warning” system for mothers and their providers. Data: The study will be tested on patients' medical records using the MedStar's Analytics Platform (MAP), a registry of over 5 Million unique patients. The tool will subsequently have the potential to be leveraged to over 90 million medical records for the MedStar and Cerner hospital systems, distilled down to meet specific eligibility criteria including, gender, age, race, pregnancy and medical outcomes. A second phase of this project would take the findings from the retrospective study (this grant request) and use the technology within the Medstar hospital system, to validate the efficacy of the findings in a “real-world setting”. Using our proprietary AI technology, we will compare each mother's progress against a cohort of retrospective data to enhance diagnostics and provide real-time feedback to caregivers and patients. Geospatial mapping: Mapping the patient medical record and the health information to their social setting is vital for understanding the underlying social constructs that affect the health of mothers in different regions.