Drug Abuse Ontology to Harness Web-Based Data for Substance Use Epidemiology Research: Ontology Development Study

Area of Science:

• Computational linguistics
• Public Health Informatics
• Ontology Engineering

Background:

• Web-based resources and social media are increasingly vital for sharing health information.
• There is a growing interest in leveraging these platforms for epidemiological surveillance of substance use.

Purpose of the Study:

• To describe the development and application of the Drug Abuse Ontology (DAO).
• To utilize the DAO as a framework for analyzing web and social media data to inform public health and substance use research.
• To analyze user knowledge, attitudes, and behaviors related to nonmedical buprenorphine and opioid use, cannabis product trends, synthetic opioid availability, and COVID-19 social media trends.

Main Methods:

• The DAO's domain and scope were defined using the 101 ontology development methodology.
• Ontology development involved determining scope, reusing existing knowledge, enumerating terms, defining classes and properties, and creating instances.
• Ontology quality was evaluated using semantic web and natural language processing best practices.

Main Results:

• The DAO currently includes 315 classes, 31 relationships, and 814 instances.
• The ontology is flexible and accommodates new concepts, improving machine learning algorithms by reducing false alarms.
• The DAO is continually updated and applied to analyze social media and dark web marketplace data.

Conclusions:

• The DAO offers a robust framework for substance use and mental health research.
• It advances big data analytics for web-based data in substance use epidemiology.
• The ontology can be expanded and adapted for diverse public health applications.