Positive correlation exists for patient’s social determinant of health risk factor and their use of acute care services (ER visit and hospitalization) especially in veteran population. We predict no hospitalization 95.8% of time, ER visit 78.8% of time based on correlating social determinants with acute care utilization. The ability to predict these using machine learning would guide interventions towards a large patient population to prevent ER visits or hospital admissions.

Learning Objective: Learning Objectives
1) Understand that patients at high risk for acute care utilization (ER visits and Hospitalization) can be predicted based on social determinants of health factors.
2) Formulate an approach to identify groups of patients at high risk or ER visit or Hospitalization from specific social factors for who care coordination and access to resources may be valuable.
3) Understanding how leveraging technology, such as machine learning and statistical modeling, can allow for healthcare providers to keep patients out of the hospital by proactively addressing their issues.

Authors:

Soham Shah (Presenter)
University of Arizona Banner Health

Peter Nguyen, University of Arizona Banner Health
Manisha Shah, Mayo Clinic
Hamed Abbaszadegan, University of Arizona Banner Health

Presentation Materials:

• Abstract/Manuscript - 1