Detecting Drug Safety Signals from Social Media
Detecting adverse drug reactions from analyzing social media and consumer health content
Researchers in Drexel’s College of Computing & Informatics have developed a system that extracts consumer health data from social media, such as the MedHelp website message boards. An analysis tool detects the association between prescribed drugs and side effects or adverse drug reactions, along with the measurement correlation strength. Because the current method of monitoring drug safety is through a self-reporting system to the FDA, the amount of data received by the regulatory agency is limited. By expanding monitoring to online patient-driven sites, more information can be readily captured. In related work, the researchers have developed algorithms for identifying off-label drug uses and drug repositioning based on online consumer data and constructing a heterogeneous healthcare network. The software can be deployed as a standalone application or directly on the web.
Applications
- Discover adverse drug reactions
- Data mining for off-label uses of prescribed drugs
- Drug repositioning to find new uses for existing drugs on market
- Continued investigations of drug safety and efficacy
Advantages
- Monitor adverse drug reactions that are not reported to FDA through social media
- Expand dataset for post-marketing surveillance
- Pathway for clinical innovation by pharmaceutical companies to find new uses for prescription drugs
- Reduce cost of initial research to reposition developed drugs for new disease targets
Intellectual Property and Development Status
Copyright
References
Yang C.C. and Zhao M. Determining associations with word embedding in heterogeneous network for detecting off-label drug uses. Proceedings of IEEE International Conference on Healthcare Informatics, 2017.
Zhao M. and Yang C.C. Automated off-label drug use detection from user generated content. Proceedings of the ACM Conference on Bioinformatics, Computational Biology, and Health Informatics, 2017.