A fully automated conversational agent for promoting mental well-being: A pilot RCT using mixed methods

Area of Science:

• Digital mental health
• Psychological interventions
• Human-computer interaction

Background:

• Fully automated self-help interventions offer cost-effective mental health promotion but often suffer from poor adherence.
• Conversational agents can enhance adherence by mimicking human therapeutic support.

Purpose of the Study:

• To evaluate the effectiveness and adherence of a smartphone app delivering positive psychology and CBT strategies via an automated chatbot (Shim) to a non-clinical population.
• To explore user experiences and perceptions of interacting with the Shim chatbot.

Main Methods:

• A randomized controlled trial with 28 participants assigned to either the chatbot intervention (n=14) or a wait-list control group (n=14).
• Intervention duration was 2 weeks, with assessments of psychological well-being (FS) and perceived stress (PSS-10).
• Qualitative data were collected to explore user experiences.

Main Results:

• Participants adhering to the intervention (n=13) demonstrated significant improvements in psychological well-being and reductions in perceived stress compared to the control group (small to large effect sizes).
• High engagement was observed, with an average app open ratio of 17.71 times over the intervention period, exceeding rates in other studies.
• Qualitative findings identified novel themes, including the moderating role of the chatbot's format.

Conclusions:

• The Shim chatbot intervention shows promise as an effective and engaging digital self-help tool for mental health promotion.
• High adherence rates validate the potential of this approach, suggesting replication with larger samples and active control groups.
• Developing automated, engaging, and effective digital interventions is crucial for public health initiatives.