Data linkage offers the potential to improve suicide risk surveillance as well as evaluate and enhance suicide prevention approaches. Data linkage could also allow for the identification of novel predictors or combinations of predictors to improve precision medicine efforts. The ability to access and analyze health care data has expanded dramatically in the past decade. The goal of this application is to leverage data from multiple sources including hospital discharges, insurance claims, death investigations, health information exchanges, and administrative data (e.g., civil status, employment status, legal system involvement, neighborhood characteristics, unemployment, gun ownership, arrests and incarceration). Aggregating these data systems will allow for novel secondary data analyses, which will enhance our understanding of risk and protective processes on suicide death in the Mid-Atlantic region.
The aims of this project include identifying and merging existing clinical data sources (e.g., medical examiner, claims, hospital discharges), geo-level data sources and other clinical data (e.g., built environment, electronic health records), and aggregated administrative data (e.g., school, incarceration, gun ownership). Our team will develop predictive models using machine learning and natural language processing algorithms. These models will be enhanced using spatial algorithms to detect the geo-clustering of suicide. The generalizability of data sources and analytic methods will be tested in the Johns Hopkins Clinical Research Network across the Mid-Atlantic region of the United States (Maryland, Pennsylvania, Delaware, Virginia, and West Virginia) to evaluate data availability, quality and data linkage feasibility. The methods of this project will be shared as a resource with other states and regions to establish a similar analytics framework to advance suicide prevention. Our team will develop national guidelines and technical support materials to facilitate data linkage and analysis based on common data sources available in most states. The project will offer essential benchmarks and measures for providers and payers to reduce suicide events in their systems.
? PUBLICHEALTHRELEVANCE Suicide is a major preventable public health challenge. Over 44,000 people in the United States died by suicide in 2015 and rates have increased 24% since 1999. This study will conduct data linkage and informatics approaches to utilize existing resources to improve suicide risk identification and prevention.
