New Study Utilizing AI Highlights Stigma Around Medications for Opioid Use Disorder

A new study published in the Journal of Addiction Medicine demonstrates how artificial intelligence (AI) can be an effective tool in community-engaged research by enabling rapid analysis of complex qualitative data that otherwise requires extensive time to review and code manually. In this study, researchers used AI to analyze community coalition meeting minutes from the HEALing Communities Study (HCS) to identify key themes and how they intersected with decisions when selecting evidence-based practices (EBPs) to reduce overdose deaths.

Researchers utilized various AI tools, such as Natural Language Processing (NLP), Machine Learning (ML), Large Language Models (LLMs), and ChatGPT Enterprise, to analyze minutes from 127 audio-recorded coalition meetings from 13 New York communities participating in the HCS. The study explored how often stigma was discussed in relation to EBP implementation, including medications for opioid use disorder (MOUD), overdose education and naloxone distribution (OEND), safer opioid prescribing practices, and stigma-reducing communication campaigns—how those discussions varied across counties, and how race and ethnicity shaped the conversation.

Key Findings

• Stigma was more frequently discussed when coalitions focused on MOUD use and delivery, compared to other EBPs, such as OEND or safer prescribing. Researchers believe this may be due to persistent misconceptions that MOUD is simply substituting one drug for another.

• Counties with larger racial and ethnic minority populations were more likely to discuss stigma when talking about  EBP implementation.

• Stigma-related conversations were 57% more likely to occur when racial or ethnic disparities were mentioned.

• In several counties, coalition members noted that stigma within pharmacies and health care settings posed barriers to MOUD access.

• How stigma and disparities were discussed varied widely between counties, shaped by factors such as local demographics, coalition priorities, and community comfort with addressing racial inequities.

  
  

Dr. Redonna Chandler, Chair of the Scientific Advisory Board at the Addiction Policy Forum, emphasized the importance of this work: “Stigma remains a huge barrier to support for medications for opioid use disorder. AI tools can be used to effectively explore large datasets to understand the levels of stigma held by key stakeholders, how stigma is discussed, and the impact.”

This study reveals that stigma is a widespread, contextually nuanced barrier in implementing EBPs, particularly in communities with higher racial/ethnic diversity. Many of the HCS coalitions proposed strategies to counter stigma, including locally tailored communications campaigns, personal storytelling, and involvement of people with lived experience in campaign design.

By identifying themes, such as stigma and EBP implementation barriers in real time, AI may be a promising tool to support faster, informed decision-making and allow researchers and practitioners to adapt interventions based on emerging insights from the community.

“Our findings show that stigma is not just a barrier to substance use treatment, but that it is shaped by race, equity, and local community dynamics. By using artificial intelligence (AI) to analyze hundreds of hours of coalition meeting discussions from the HEALing Communities Study in New York State, we were able to uncover these patterns after the study’s completion,” said Dr. Nabila El-Bassel, lead author of the paper.

“Looking ahead, the same AI tools can be applied in real time during implementation of studies on stigma to generate faster insights, allowing researchers to ensure they are amplifying community voices, and letting people with lived experience guide how we reduce stigma and expand access to lifesaving treatments.”

Read More

The Article

El-Bassel, N., David, J. L., Aragundi, E., Walters, S. T., Wu, E., Gilbert, L., Chandler, R., Hunt, T., Frye, V., Campbell, A. N. C., Goddard-Erich, D. A., Chen, M., Davé, P., Benjamin, S. N., Lounsbury, D., Sabounchi, N., Aggarwal, M., Feaster, D., Huang, T., & Zheng, T. (2025). Artificial intelligence and stigma in addiction research: Insights from the HEALing Communities Study coalition meetings. Journal of Addiction Medicine. Advance online publication. https://doi.org/10.1097/ADM.0000000000001534