Including Social and Behavioral Determinants in Predictive Models: Trends, Challenges, and Opportunities

Abstract

In an era of accelerated health information technology capability, health care organizations increasingly use digital data to predict outcomes such as emergency department use, hospitalizations, and health care costs. This trend occurs alongside a growing recognition that social and behavioral determinants of health (SBDH) influence health and medical care use. Consequently, health providers and insurers are starting to incorporate new SBDH data sources into a wide range of health care prediction models, although existing models that use SBDH variables have not been shown to improve health care predictions more than models that use exclusively clinical variables. In this viewpoint, we review the rationale behind the push to integrate SBDH data into health care predictive models and explore the technical, strategic, and ethical challenges faced as this process unfolds across the United States. We also offer several recommendations to overcome these challenges to reach the promise of SBDH predictive analytics to improve health and reduce health care disparities.

Measuring the Value of a Practical Text Mining Approach to Identify Patients With Housing Issues in the Free-Text Notes in Electronic Health Record: Findings of a Retrospective Cohort Study

“…This information would help care managers, care coordinators, and social workers to tailor specific social interventions and/or conducting referrals to community-based social services organizations (52,53). Clinicians can also utilize such information to explore the underlying housing issues at the point of care, and population health experts might use this information to better predict utilization rates associated with such patient population (54).…”

Moving Beyond Simple Risk Prediction: Segmenting Patient Populations Using Consumer Data

“…Other challenges of using consumer data include reproducibility and analytical challenges. Predictive models developed by the private sector are not shared publicly, therefore cannot be replicated by other researchers to ensure accuracy, validity and potential model bias (34). Additionally, researchers should be cautious when selecting the analytical approaches when it comes to the inclusion of marketing data to predict health outcomes.…”

“…Concurrent with patient segmentation models developed by researchers, many predictive models based on SDOH have been developed by health payers and analytics companies. Most often these models are proprietary hence not available for review and scrutiny (34). For instance, a nonprofit health insurance company used consumer data to develop a segmentation model to make informed adjustments to its Medicare marketing efforts (35).…”

A scoping review of knowledge authoring tools used for developing computerized clinical decision support systems

“… 80–82 Individual and neighborhood level SDOH data are increasingly used in the clinical decision-making process to improve outcomes and reduce utilization. 83 , 84 However, none of the reviewed CKATs mentioned how such unstandardized information would be encoded and integrated into the knowledge creation process.…”