Related Experiment Video
Updated: Aug 1, 2025
08:53
Integrating Computerized Linguistic and Social Network Analyses to Capture Addiction Recovery Capital in an Online Community
Published on: May 31, 2019
5.2K
Unsupervised Machine Learning to Detect and Characterize Barriers to Pre-exposure Prophylaxis Therapy: Multiplatform
Qing Xu1,2, Matthew C Nali1,2,3, Tiana McMann1,2,3
1S-3 Research San Diego, CA United States.
JMIR Infodemiology
|April 28, 2023
Summary
Pre-exposure prophylaxis (PrEP) barriers for racial, ethnic, and sexual minorities were identified using social media data. Key barriers include lack of knowledge, access issues, and social stigma, impacting HIV prevention efforts.
Area of Science:
- Public Health
- Infodemiology
- Machine Learning
Background:
- HIV infection remains a significant public health challenge for racial and ethnic minority groups.
- Pre-exposure prophylaxis (PrEP) is a highly effective HIV prevention method when used as prescribed.
- Understanding PrEP experiences and barriers among minority populations is crucial for effective implementation.
Purpose of the Study:
- To identify and characterize barriers to PrEP uptake and adherence using big data and machine learning.
- To analyze user experiences and attitudes regarding PrEP among racial, ethnic, and sexual minority groups.
- To inform public health strategies for broader PrEP adoption in vulnerable communities.
Main Methods:
- Data mining of social media platforms (Twitter, YouTube, Tumblr, Instagram, Reddit) using PrEP and HIV-related keywords.
- Unsupervised machine learning for data analysis, followed by manual annotation using deductive coding.
- Identification and thematic categorization of user-generated posts discussing PrEP barriers.
Main Results:
- Analysis of 522,430 social media posts revealed 785 posts related to PrEP barriers.
- Barriers were categorized into provider (1.7%), patient (72.6%), and community (21.1%) levels.
- Key barriers included lack of knowledge, access issues (insurance, prescriptions, COVID-19), and adherence challenges (side effects, stigma).
- 40.8% of posts involved users identifying as racial/ethnic or sexual minorities, highlighting specific concerns.
Conclusions:
- Social media data provides valuable insights into objective and subjective barriers to PrEP initiation, access, and adherence.
- User-generated content highlights specific challenges faced by racial, ethnic, and sexual minority groups in PrEP utilization.
- Findings can inform targeted health promotion and regulatory science to improve PrEP access and adherence in at-risk communities.
