Description
A preliminary model for fatal opioid-related overdose risk prediction is used to test the feasibility of collaboration between Vanderbilt University Medical Center and the Tennessee Department of Health (TDOH). Random forests were applied to TDOH prescription, death, and hospital discharge data. Study data consisted of opioid-related overdose decedents in 2017 (n=665), and general controls (n=5,320). Discrimination and calibration performance were good (c-statistic 0.81, Brier 0.08). This partnership approach has potential to address public health challenges.
Learning Objective: Understand how collaboration between a state health department and an academic medical center can be leveraged to address the ongoing opioid epidemic.
Authors:
Christopher Puchi (Presenter)
Vanderbilt University
Ben Tyndall, Tennessee Department of Health
Melissa McPheeters, Tennessee Department of Health
Colin Walsh, Vanderbilt University Medical Center
Presentation Materials: