Characterizing Opioid Overdoses Using Emergency Medical Services Data : A Case Definition Algorithm Enhanced by...

Area of Science:

• Public Health
• Data Science
• Emergency Medicine

Background:

• The opioid overdose epidemic necessitates accurate tracking of nonfatal overdoses.
• Existing methods for identifying opioid overdoses in emergency medical services (EMS) data have limitations.

Purpose of the Study:

• To develop an innovative and flexible case definition for opioid overdose using North Carolina EMS data.
• To improve the accuracy of identifying opioid overdose cases in EMS encounters.

Main Methods:

• Utilized de-identified North Carolina EMS encounter data (2010-2015) for patients over 12 years old.
• Developed a rule-based algorithm incorporating expert knowledge and text mining of EMS narratives.
• Applied a machine-learning classification tree model to predict opioid overdose cases based on chief concern.

Main Results:

• The novel case definition achieved a 90.0% positive predictive value, outperforming a previous definition (82.7%).
• Identified a significant increase in unresponsive opioid overdoses from 3412 (2010) to 7194 (2015).
• Observed a 1.7-fold increase in the monthly rate of opioid overdoses and an 83% prevalence of naloxone use in these cases.

Conclusions:

• Machine learning combined with substantive knowledge algorithms can create effective case definitions for opioid overdose in EMS data.
• The developed methodology offers a flexible approach applicable to other states' EMS data.
• This improved case definition enhances the ability to track and respond to the opioid overdose epidemic.