Visualizing the opioid overdose with a dynamic heat map to identify and predict vulnerable communities

Authors

• Justin R. Chang, VCU School of MedicineFollow
• Garrett K. Wong, VCU School of Medicine
• Chase Greco, VCU Department of Computer Science
• Yadunandan Pillai, VCU Department of Computer Science
• Mohammad A. Shahrezaei, VCU Department of Computer Science
• Melissa H. Burton, VCU Interprofessional Education and Collaborative Care
• Rob Lawrence, Richmond Ambulance Authority
• Alan Dow, VCU Interprofessional Education and Collaborative Care

Files

Abstract

Background: Opioid and heroin overdose epidemic is a public health emergency in the state of Virginia. In order to prevent overdose deaths, we need the target expertise in substance use disorders to areas with high rates of overdose. In particular, an area with an acute spike in overdoses might represent an urgent need for intervention.

Aims: The CDC urges the use of near real-time surveillance to effectively identify overdose incidence, and to coordinate community responses in the states affected by the epidemic, including Virginia. However, current opioid overdose datasets for Virginia lack adequate consistency, granularity, and temporality for rigorous data analyses and near real-time surveillance.

Methods: In this project, we modeled a collaboration between emergency medical services (EMS), public health advocates and computer scientists to increase the geodata resolution around opioid overdoses; monitor outbreaks in real-time; and produce predictive models to target interventions for vulnerable communities.

Results: EMS professionals could use this tool to funnel efforts towards these communities and intervene prior to additional overdoses. Using data obtained from the Richmond Ambulance Authority, we will create a dynamic incidence heatmap to demonstrate high resolution data visualization and machine learning opportunities.

Conclusions: We envision a no-cost, easy implementation of this approach in current EMS practice which will generate predictions of when and where first responders and substance use professionals can intervene to prevent further overdoses. Community health workers and telemedicine doctors can provide immediate care to outbreak areas, followed by peer counselors and social workers for long term follow-up.

Publication Date

2018

Keywords

opioid, overdose epidemic, interprofessional collaboration, data science, dynamic heat map, real-time surveillance, machine learning, predictive model

Disciplines

Computer Sciences | Databases and Information Systems | Data Science | Health Information Technology | Interprofessional Education | Medicine and Health Sciences | Telemedicine

Faculty Advisor/Mentor

Dr. Alan Dow

Is Part Of

VCU Graduate Research Posters