The goal of this study is to evaluate the feasibility, acceptability, safety and preliminary efficacy of a chatbot-based mindfulness-based stress reduction (MBSR) intervention for university students with depressive symptoms.

A rule-based MBSR chatbot will be developed and evaluated with a single-group pretest-posttest study for university students aged 18 or above in Hong Kong reporting depressive symptoms, followed by the collection of their subjective feedback. The intervention will last eight weeks. The primary clinical outcome is depression levels, with a range of secondary outcomes including stress, anxiety and mindfulness levels.

Depression is a growing concern among university students. Chatbots provide flexible, accessible, personalized psychosocial support. Delivering Mindfulness-based Stress Reduction (MBSR) sessions via chatbots may reduce depressive symptoms in university students.

This study consists of two phases. In the first phase, a rule-based automated chatbot will be developed to facilitate MBSR interventions. The second phase will employ a single-group pretest-posttest design to evaluate the feasibility, acceptability, safety and preliminary effectiveness of the chatbot-based MBSR intervention for university students with self-reported depressive symptoms. The study will be guided by the CONSORT 2010 checklist-extension for pilot and feasibility studies of information to include when reporting a pilot or feasibility trial. This study specifically targets university students in Hong Kong.

Phase One: Development of the Chatbot-based MBSR Intervention Phase one focuses on developing a chatbot-based MBSR intervention using an interactive system named Power Virtual Agents, integrated within the Microsoft Teams platform. We will first conduct a comprehensive literature review to inform the integration of chatbot conversational flows with the content of MBSR sessions and craft a comprehensive chatbot-mediated conversational flow structure to guide participants through the MBSR sessions. The sessions will be delivered in audio format (Cantonese) via the chatbot programmed to operate in Traditional Chinese. Subsequently, we will employ a logical algorithm to establish coherent flows for the intervention sessions. Each session will be programmed with specific triggers, keywords, and queries to effectively engage users. The chatbot incorporates natural language understanding capabilities to analyze user input and guide the conversation accordingly. Additionally, a lightweight database will be integrated to track and maintain user progress and facilitate continuity across the MBSR sessions. The usability of the chatbot-based MBSR intervention will be evaluated through backend data analysis from the chatbot system and qualitative feedback from users. All steps will be supervised by the authors to ensure the reliability and accuracy of the development process.

Phase Two: A Feasibility Clinical Trial Participant Recruitment Posters will be displayed across various key locations within our university. These posters include a concise overview of the study and a quick response (QR) code for assessing participant eligibility. Additionally, the study will be promoted on social media platforms such as Facebook and Instagram. Eligible participants who complete the programme will be offered coupons. The recruitment period is expected to extend for four months.

Participate Screening Eligible participants must be students aged over 18 years pursuing a bachelor degree, associate degree, or higher diploma, with self-reported depressive symptoms. The depressive symptoms are assessed pre-intervention using the 9-item Patient Health Questionnaire, with cumulative scores of 5 or higher indicating depressive symptoms. Participants must also be able to provide informed consent, read Chinese, listen to Cantonese, and ensure internet access during the study. Participants will be excluded if they had been diagnosed with a clinical psychotic condition pre-intervention or are currently involved in any mindfulness-based or other psychosocial interventions.

Sample Size According to previous single-group pre-post pilot studies, this study aims to recruit 20-40 participants to assess the feasibility, acceptability, and preliminary effectiveness of the chatbot-based MBSR programme.

Intervention Development and Content Description Recruited participants need to sign consent forms and complete baseline assessments. Instructions for interacting with the chatbot via Microsoft TEAMS® will be provided. During the intervention, participants will first undergo a single-item assessment to verify the stability of their mental states for subsequent MBSR sessions, deemed stable if rated as excellent, very good or good; They will then engage in conversations with the chatbot, responding to prompts like What is the most recent thing that made you happy? to foster interactive communication and relaxation. Subsequently, participants need to follow guided MBSR sessions and are encouraged to interact with the chatbot for at least one hour daily; all interactions will be monitored and recorded. After eight weeks, a post-intervention assessment will be conducted, followed by pre-post statistical analyses. A total of ten participants with the most and least significant changes in depression levels will be selected for an open-ended questionnaire.

Assessments

Feasibility and Acceptability and Safety： The feasibility of participant recruitment and follow-up will be assessed through three metrics: (1) the time taken to recruit the target sample size; (2) the recruitment rate and (3) the retention rate.

The acceptability of the intervention will be evaluated using the following criteria (1) adherence rate and (2) the specifically designed open-ended questionnaire for a targeted group of participants.

The safety of the intervention will be evaluated by recording any adverse events reported by participants during the study and determining whether any events are directly associated with the intervention or with participation in the study more generally.

Socio-demographics:

Participant socio-demographics will be collected using a self-designed questionnaire, including gender, marital status, current grade level, religion, faculty, year of study, living situation, and personal net monthly income.

Primary Clinical Outcome:

Depression: The depression levels of the participants will be assessed using the Patient Health Questionnaire (PHQ-9) in Chinese.

Secondary Clinical Outcomes:

Stress: The stress levels of the participants will be evaluated with the Perceived Stress Scale (PSS-14) in Chinese.

Anxiety: The anxiety levels of the participants will be measured using the Generalized Anxiety Disorder Scale (GAD-7) in Chinese.

Mindfulness: The mindfulness levels of the participants will be assessed using the Five-Facet Mindfulness Questionnaire (FFMQ-39) in Chinese.

Statistical Analysis Descriptive statistics will be used to summarize participant demographic information and total scores on different assessment tools. Statistical analysis will be conducted using Statistical Package for the Social Sciences (SPSS), version 27.0 (IBM Corp). The Shapiro-Wilk test will be employed to assess the normality of continuous data, including scores from the PHQ-9, PSS-14, GAD-7, and FFMQ. Subsequently, a t-test will be performed to evaluate changes between pre- and post-intervention scores for these measures. A chi-square test will be used to identify the severity differences regarding depression, anxiety and stress levels. A p-value of ≤.05 is considered statistically significant. Imputation will be considered for missing values. Participant responses to open-ended questions will be compiled and analyzed to gather feedback.

Ethical consideration This study will be conducted in accordance with the Declaration of Helsinki. Ethical approval for this study was granted by the Hong Kong Polytechnic University Human Subjects Ethics Review Board (No. HSEARS20221027007). Before the baseline assessments, all participants need to provide informed consent, confirming their participation and their understanding of their right to withdraw at any time without penalty. Participants will be assured of confidentiality, with their identities and data protected anonymously. Only authorized personnel have access to the data for analysis. Participants will be informed about the likelihood of encountering emotional distress during the study and will be encouraged to report any discomfort to the research team. If necessary, participants would be referred to clinical mental health services. All data collected from the chatbot will be de-identified. Usage data will be aggregated and not linked to specific participants.

Implications for Practice The pilot findings of this study will lay the groundwork for developing a more optimized chatbot intervention in future research. All data and feedback collected in this pilot study will be used to refine the application of MBSR for a future clinical trial examining its efficacy. Insights from this chatbot-based MBSR programme will also inform researchers and healthcare professionals to develop innovative digital mental health interventions to address the rising mental health challenges among university students.

Inclusion Criteria:

• University students aged over 18 years
• University students pursuing a bachelor degree, associate degree, or higher diploma in Hong Kong
• University students with self-reported depressive symptoms (The depressive symptoms are assessed pre-intervention using the 9-item Patient Health Questionnaire, with cumulative scores of 5 or higher indicating at least mild depressive symptoms.)
• Be able to provide informed consent, read Chinese, listen to Cantonese
• Ensure internet access during the study

Exclusion Criteria:

• University students who have been diagnosed with a clinical psychotic condition pre-intervention
• University students who are currently involved in any mindfulness-based or other psychosocial interventions