Framing Helper Therapy to Support User Engagement: Causal Evidence from a Public Deployment of a Mental Health Support Text Messaging Program

Paper Title

Framing Helper Therapy to Support User Engagement: Causal Evidence from a Public Deployment of a Mental Health Support Text Messaging Program

Publication Info

• Topic area: Digital mental health interventions and user engagement strategies
• Keywords: Helper therapy, digital mental health, user engagement, text messaging, framing effects, peer support, self-focused motivation, linguistic analysis, BIPOC, LGBTQ+

Background and Problem

• Problem / challenge: Digital mental health (DMH) tools face challenges in sustaining user engagement, particularly for effortful tasks like helper therapy exercises. The best way to frame these exercises to motivate participation remains unclear.
• Significance: Increasing engagement in DMH tools can help address mental health treatment gaps, especially for underserved populations, by promoting accessible and scalable interventions.
• Motivation and related work: Prior studies highlight the benefits of helper therapy for both helpers and recipients but note the difficulty in motivating helpers to participate. Existing research lacks rigorous evaluation of framing strategies and their impact on engagement, particularly in real-world settings and among diverse user groups.

Solution

• Proposed approach: The study investigates how different framings of a helper therapy writing task—focused on either self-benefit or benefit to others—affect user engagement in a publicly deployed DMH program called Small Steps.
• Novelty:
  1. Causal analysis of framing effects on user engagement in a real-world DMH intervention with 3,817 participants.
  2. Examination of how user characteristics, including BIPOC and LGBTQ+ identities, influence framing effectiveness.
  3. Analysis of linguistic properties of user-generated support messages under different framings.
  4. Insights into tailoring DMH interventions to diverse user identities.
• Procedure and key techniques:
  • Participants were randomized into different framing conditions: three rationale framings (self-focused, other-focused, no rationale) and three target peer framings (support someone depressed, someone like them, or a friend).
  • Engagement was measured by whether users wrote a support message.
  • Statistical methods included two-proportion z-tests for causal inference and linguistic analysis using LIWC and EmoLex lexicons.

Results

• Concrete findings:
  • Overall, 19.5% of users engaged by writing support messages, with higher engagement among LGBTQ+ (20.3%) and BIPOC (21.6%) users.
  • Self-focused rationale increased engagement rates by 4.6% (95% CI [1.6, 7.7]) compared to other-focused rationale.
  • Among LGBTQ+ users, self-focused rationale increased engagement by 8.6% (95% CI [2.5, 14.5]) over other-focused rationale.
  • Among BIPOC users, the “someone like you” target peer framing increased engagement by 9.5% (95% CI [3.4, 15.5]) over “someone depressed.”
• Advantage over baselines:
  • Self-focused framings consistently outperformed other-focused or no-rationale framings in motivating engagement.
  • The “someone like you” target peer framing elicited more positive and trust-related language and less negative emotion language compared to “someone depressed.”
• Experiments / evaluation:
  • Conducted on a sample of 3,817 users from the Small Steps program, which delivers an 8-week DMH intervention.
  • Engagement was measured as the primary outcome, with linguistic properties of messages analyzed as secondary outcomes.
  • Subgroup analyses focused on LGBTQ+ and BIPOC users.
• Limitations and future work:
  • Sample skewed toward women (75%) and LGBTQ+ (28%), limiting generalizability.
  • Focused on short-term engagement; longitudinal effects remain unexplored.
  • Helper therapy was analyzed from the helper’s perspective; future work should examine message impact on recipients.
  • Limited personalization of prompts; future research could explore adaptive or user-selected framings.

Summary

This study provides causal evidence that self-focused framings in helper therapy exercises significantly increase user engagement in a DMH text messaging program, with even larger effects observed among LGBTQ+ and BIPOC users. The findings suggest that framing tasks as benefiting the user themselves, rather than others, can motivate participation and elicit more positive and trust-related language. These results have practical implications for designing personalized and inclusive DMH interventions that better support diverse populations. Future research should explore longitudinal effects, message impact on recipients, and adaptive personalization of framings.