EUVIMEDDescoperă EUVIMED
Open access

Forging Online Community Among People in Recovery From Substance Use: Natural Language Processing and Deep-Learning Analysis of The Phoenix App User-Generated Data

Danny Valdez iD, Katie M Heinrich iD, Beth Collinson iD, Aspen Streetman iD, Zach Sloan iD

JMIR mHealth and uHealth · 2025

DOI: 10.2196/68438 ↗

At a glance

Journal article in JMIR mHealth and uHealth, 2025. Written by Danny Valdez and 4 co-authors. Institution: Department of Applied Health Science, Indiana University, Bloomington, IN, United States. Major MEDLINE topics: Substance-Related Disorders, Natural Language Processing, Mobile Applications, Deep Learning. 55 references. The full text is freely available in PubMed Central.

RezumatSubjectsDate bibliograficeCiteRelated studies

Abstract (original language)

Background

Mobile apps are powerful tools for promoting and sustaining healthy behaviors, including supporting diverse recovery pathways from substance use, including alcohol use disorder. Indeed, prior research strongly supports the notion that social connection through mobile platforms, supplemented by an in-person interaction, is vital in helping individuals strengthen their recovery and improve overall well-being and mental health. However, research into the digital footprints of mobile app users, as a strategy to assess app usage experiences in a recovery context, is lacking.

Objective

This study utilizes a dataset from The Phoenix app, a social media platform specifically designed for individuals impacted by substance use, including those in or seeking recovery, to identify core uses of the app, including how it is leveraged by members from a thematic and emotional valence context.

Methods

We applied natural language processing and deep learning methods to analyze a random sample of 19,685 posts. Analyses included the Bidirectional Encoder Representation from Transformers topic modeling tool to generate themes and a Valence Aware Dictionary and Sentiment Reasoner sentiment analysis to approximate emotional tone and mood from posts ranging from highly negative (−0.99) to highly positive (0.99).

Results

After removing duplicate and nonsensical posts, we retained a final sample size of 17,617 posts. Bidirectional Encoder Representation from Transformers topic modeling tool identified 10 topics (coherence score=0.48) within 2 overarching themes: (1) those related to engaging app members through in-person and online interactions (7 topics) and (2) as a forum to discuss more serious topics pertaining to substance use and mental health recovery (3 topics). Overall, the topics revealed a distinct and recurring theme of community support. Valence Aware Dictionary and Sentiment Reasoner sentiment analysis was 0.44 (SD 0.42), indicating highly positive posts, with only 429 (2.4%) being highly negative.

Conclusions

The study findings broadly show positive uses of The Phoenix app as a tool for social connections and community among people in recovery from substance use. With the high positive sentiment of posts, the app was distinct from other social media platforms (eg, X, Reddit, Facebook), which often feature a mix of highly positive and highly negative posts. Additional research is needed to confirm these results using a larger dataset and with comparative analysis of other recovery forums to contribute to the understanding of social media’s role and function in changing health-related behaviors.

Subjects

MeSH terms (MEDLINE)

Substance-Related DisordersNatural Language ProcessingMobile ApplicationsDeep LearningHumansSocial Media

Author keywords

communitymHealthmental healthmobile healthsocial supportsubstance use

Authors and institutions

  • Department of Applied Health Science, Indiana University, Bloomington, IN, United States.
  • Department of Research and Evaluation, The Phoenix, 54 Newmarket Sq, Boston, MA, 02118, United States, 1 8084579525.
  • Department of Pediatrics, Vanderbilt University Medical Center, Nashville, TN, United States.
  • The Phoenix, Boston, MA, United States.

ORCID profiles: Danny Valdez, Katie M Heinrich, Beth Collinson, Aspen Streetman, Zach Sloan

Cite

APA

Valdez, D., Heinrich, K. M., Collinson, B., Streetman, A., & Sloan, Z. (2025). Forging Online Community Among People in Recovery From Substance Use: Natural Language Processing and Deep-Learning Analysis of The Phoenix App User-Generated Data. JMIR mHealth and uHealth. https://doi.org/10.2196/68438

Vancouver

Valdez D, Heinrich KM, Collinson B, Streetman A, Sloan Z. Forging Online Community Among People in Recovery From Substance Use: Natural Language Processing and Deep-Learning Analysis of The Phoenix App User-Generated Data. JMIR mHealth and uHealth. 2025. doi:10.2196/68438

BibTeX
@article{Valdez2025,
  author = {Valdez, Danny and Heinrich, Katie M and Collinson, Beth and Streetman, Aspen and Sloan, Zach},
  title = {Forging Online Community Among People in Recovery From Substance Use: Natural Language Processing and Deep-Learning Analysis of The Phoenix App User-Generated Data},
  journal = {JMIR mHealth and uHealth},
  year = {2025},
  doi = {10.2196/68438}
}
Start for free →