Abstract 9876: Optimizing Public Naloxone Kit Locations Through Mathematical Modeling

Introduction: Use of bystander-administered naloxone may lead to improved likelihood of recovery from opioid overdose. We sought to determine the accessibility of public access naloxone kits on nearby opioid overdose incidents if placed at public transit stops, compared to placing kits outside pharmacies or with existing public access automated external defibrillators (PADs). Methods: We included all incidents in Metro Vancouver, British Columbia responded to by British Columbia Emergency Health Services coded as a drug overdose with naloxone administered on-scene (Dec. 2014 to Aug. 2020). We geo-coded all public transit bus stops and used a mathematical optimization model to select bus stops where publicly accessible naloxone kits could be placed to maximize accessibility (defined as ≤100 m walking distance) to opioid overdoses. We evaluated accessibility on out-of-sample OHCAs using five-fold cross validation and compared against two baseline policies: placing publicly accessible naloxone kits at all pharmacies identified by the College of Pharmacists of British Columbia, and placing kits at all PADs identified by the British Columbia AED Registry. Statistical analysis was conducted using McNemar’s test. Results: We identified 14,318 opioid overdoses, 8,972 bus stops, 736 pharmacies, and 425 PADs. Accessibility of public naloxone kits for opioid overdose locations was 5.1% when placed at all pharmacies and 3.5% when placed with all existing PADs. Optimized naloxone kit placement using bus stops as candidate locations resulted in significantly higher accessibility than both pharmacy and PAD-based placement at 14.8% with 10 optimized locations (P Conclusion: Optimizing placement of public access naloxone kits at select public transit locations can provide significantly higher accessibility to opioid overdose locations compared to placement at pharmacies or at existing PAD locations.