Online Program

Return to main conference page
Monday, January 6
Mon, Jan 6, 5:30 PM - 6:30 PM
Pacific D
Welcome Reception & Poster Session I

Leveraging Disparate Data Sources to Predict Risk: An Analytic Tool to Address Maternal Opioid Use Disorder (307892)

Nicole Harlaar, IBM Watson Health 
*Shannon Harrer, IBM Watson Health 
Cory Pack, IBM Watson Health 
Olivia Reding, IBM Watson Health 
Frank Yoon, IBM Watson Health 

Keywords: Predictive analytics, risk model, disparate data, pooling, ensemble, opioid use disorder

Sound policy and clinical decision making should be based on multiple sources of data. To assess health risks, decision makers must often analyze multiple levels of drivers and influence, from individual to ecological factors. However, disparate data sources are siloed and difficult to use. This work addresses this fundamental challenge by providing a predictive analytic technique that combines information from disparate data sources to help decision makers understand their data now to take action for the future. The analytic tool predicts risk of maternal opioid use disorder using a bioecological approach to population health: health is influenced by (1) cumulative effects across lifespans and generations and (2) multiple determinants at different system levels, “from cell to society.” To understand risk factors in maternal opioid use disorder, this work proposes the integration of EHR, claims, health registries, geographic measures, and client intake data for decision making. Through standardized production of analytic information, it helps the analyst or clinician identify high-risk individuals for timely and appropriate interventions and standardizations.