JMIR Publications

ABSTRACT

Background:

Chat-based online platforms have been shown to be effective intervention tools for mental health support. However, user attrition may limit their capacity to provide meaningful support and hinder the validation of their clinical benefits. Factors contributing to the continued use or non-use of chat-based platforms must be investigated to address barriers to engagement.

Objective:

This study investigates user engagement with a chat-based mental health platform. Specifically, it aims to identify the factors that contribute to the initiation and subsequent initiation of sessions with mental health professionals (MHPs) through the chat platform. The findings may inform the development, refinement and implementation of chat-based mental health interventions with improved user retention.

Methods:

To investigate user engagement, the study employed a prospective longitudinal design. A total of 119 students were enrolled in sequential batches, with each batch observed for a three-month period. During this period, data on chat-session activity and interactions with MHPs, who are licensed clinical psychologists, were automatically recorded and collected through the SAGIP mobile app. Each MHP handled no more than 10 students at any given time. Data pertaining to participants’ well-being, mental health, demographic information and experiences with the chat intervention were collected through surveys and interviews. Of the 119 participants, 32 were interviewed, and four of them initiated a subsequent session. A binomial mixed-effects generalized linear model was used to identify factors associated with the initiation and subsequent initiation of sessions with MHPs through the online mental health chat platform. Descriptive analyses were conducted to further examine these factors and their relationships with user engagement. In addition, thematic analysis was applied to the qualitative text data to explain other factors that may influence user engagement.

Results:

Based on the binomial mixed-effects generalized linear model, the number of user messages sent within the current session emerged as a key significant factor associated with user initiation in that same session. Meanwhile, the number of fast responses (i.e., responses provided within five minutes) users received during their current session was a significant factor associated with subsequent initiation. The thematic analysis further revealed three critical factors influencing user engagement: the availability of mental health professionals, therapeutic alliance, and usability of the platform.

Conclusions:

This study combined a binomial mixed-effects generalized linear model with thematic analysis that provided a comprehensive understanding of user-initiated engagement with a mental health chat-based platform. The findings suggest that users are more likely to initiate sessions when they have concerns, experiences, or needs they wish to share, and are more likely to continue using the platform when they receive prompt and useful responses. In addition, platform usability emerged as an important factor influencing engagement. Clinical Trial: WMSU REOC Code: 2025-EF-0379