Using Social Media Data to Identify Potential Candidates for Drug Repurposing: A Feasibility Study

Area of Science:

• Pharmacology
• Computational Biology
• Social Media Analytics

Background:

• Drug repurposing offers a promising avenue for drug development amidst declining novel drug success rates.
• Traditionally, new drug indications are discovered serendipitously; however, systematic approaches are now feasible.
• Patient-generated data from social media provides insights into medication side effects and benefits.

Purpose of the Study:

• To evaluate the feasibility of leveraging social media patient reviews for identifying drug repurposing candidates.
• To explore novel methods for drug discovery through analysis of real-world patient experiences.

Main Methods:

• Extracted patient reviews for 180 medications from WebMD.
• Employed dictionary-based and machine learning techniques to identify disease names within reviews.
• Utilized public resources to filter known indications and adverse effects, followed by manual review and a rule-based system to detect beneficial effects.

Main Results:

• Identified 2,178 and 6,171 disease names using dictionary-based and machine learning systems, respectively, from 64,616 patient comments.
• Cataloged 10 common patterns patients use to describe beneficial medication effects.
• Pinpointed five potential drug repurposing candidates after manual validation of identified beneficial effects.

Conclusions:

• This study pioneers the use of social media data for drug repurposing candidate identification.
• Demonstrated that even a basic rule-based system can effectively identify beneficial effect mentions in patient commentary.
• Social media data holds significant potential for advancing drug repurposing strategies.