Related Experiment Video
Updated: Aug 6, 2026
04:19
A Computer-Based Platform for Aiding Clinicians in Eating Disorder Analysis and Diagnosis
Published on: May 10, 2022
User-Reported Issues With Mental Health Apps: Machine-Assisted Topic Analysis of Social Media Posts
Jack Bolter1, Trisevgeni Papakonstantinou2, Paulina Bondaronek3
1Department of Psychology, School of Social Sciences, University of Westminster, London, United Kingdom.
JMIR Mhealth and Uhealth
|July 21, 2026
Summary
Mental health app users report issues like poor guidance, technical problems, and negative emotional responses. Analyzing social media comments with machine learning reveals user experience challenges in digital mental health tools.
Area of Science:
- Digital Health Interventions
- Mental Healthcare Technology
- Social Media Data Analysis
Background:
- Mobile mental health apps are increasingly popular for user support.
- Understanding user-reported issues is crucial for app safety and effectiveness.
- Existing research often relies on researcher analysis, not direct user feedback.
Purpose of the Study:
- To identify user experience issues in mental health apps through user comments.
- To evaluate a machine learning approach (structural topic modeling) for analyzing large datasets.
- To combine machine learning with human analysis for deeper insights.
Main Methods:
- Collected 79,703 posts from X (formerly Twitter) related to five popular mental health apps.
- Applied structural topic modeling (STM) and human qualitative analysis to user-generated posts.
- Filtered for negative sentiment, resulting in 19,603 relevant posts for analysis.
Main Results:
- Identified eight key user experience themes from the analyzed posts.
- Major themes included guidance shortcomings, technical difficulties, and emotional responses to app-affiliated celebrities.
- Negative impacts of sleep self-monitoring were also reported as a distinct issue.
Conclusions:
- Combining STM with qualitative analysis effectively identifies user-reported issues in mental health apps.
- These issues are often linked to negative user outcomes.
- This approach offers a rapid method for analyzing large-scale social media data for digital health interventions.