Advancing Maryland's Statewide Suicide Data Warehouse to Improve Individual and Population-level Mortality Prediction and Prevention

❖项目总结 2018年，NIMH资助开发了马里兰州自杀数据仓库(MSDW)。开发MSDW的目的 是将一组不同的数据源联系起来，并使用不同层次的风险因素来预测自杀死亡 人口。目前，MSDW包括2012至2017年间马里兰州约500多万居民的数据。MSDW包含 医院和紧急出院、健康信息交换(HIE)数据、商业保险索赔、法医 数据、来自5个选定健康系统的电子健康记录(EHR)数据，以及健康数据的基于地点的社会决定因素 涵盖住房、就业、教育、收入和犯罪等多个领域。 这项研究将利用MSDW链接的独特数据类型来改善可推广的自杀死亡预测 患者群体。我们建议的研究将进一步扩展基础数据并开发以下前沿使用案例 MSDW.这项研究的目标是取得一些成果：(1)从所有现有的MSDW数据源获取更新数据； (2)使用MSDW中存储的大量数据源来推进自杀死亡率的预测，同时控制 数据质量和潜在的模型偏差；(3)探索MSDW的各种使用案例，例如无监督的自杀集群 死亡案例；以及，(4)与外部调查人员合作，验证MSDW得出的自杀预测模型 死亡。

❖ PROJECT SUMMARY In 2018, NIMH funded the development of the Maryland Suicide Data Warehouse (MSDW). The aim of developing MSDW was to link a diverse set of data sources and use various layers of risk factors to predict suicide death on a generalizable population. Currently, MSDW includes data on ~5+ million Maryland residents spanning from 2012 to 2017. MSDW contains hospital and emergency discharges, health information exchange (HIE) data, commercial insurance claims, medical examiner data, electronic health record (EHR) data from 5 select health systems, and place-based social determinants of health data covering various domains such as housing, employment, education, income and crime. This study will leverage the unique data types linked by MSDW to improve suicide death predictions among generalizable patient populations. Our proposed study will further expand the underlying data and develop cutting-edge use cases of MSDW. The study aims to achieve a number of outcomes: (1) Acquire refreshed data from all existing MSDW data sources; (2) Use the wide array of data sources stored in MSDW to advance the prediction of suicide mortality, while controlling for data quality and potential model bias; (3) Explore various uses cases of the MSDW such as unsupervised clustering of suicide mortality cases; and, (4) Collaborate with external investigators to validate the MSDW-derived predictive models of suicide death.