Can Mental Health Chatbots Help Chronic Disease Populations?

Evaluating the Usefulness of an Artificial Intelligence (A.I.) Mental Health Chatbot in a Healthcare Setting

Past research has shown that patients suffering from chronic health conditions tend to experience high levels of negative mental health symptoms (e.g., depression). The purpose of the current study is to evaluate whether an artificial intelligence (A.I.) mental health chatbot can be used to reduce negative mental health symptoms within this population. A minimum of 60 individuals with a chronic health condition (diabetes or arthritis) will be recruited. Participants will be randomly assigned to either a treatment group or a control group. Those assigned to the treatment group will use the mental health chatbot Wysa (Touchkin eServices, Bangalore) over a period of four weeks. Those assigned to the control group will receive no chatbot. Participants will complete measures of depression, anxiety, stress, and life satisfaction via Qualtrics at the outset of the study, two weeks into the study, and four weeks into the study (i.e., the final assessment point). Results from the treatment and control groups will be compared using ANOVA models. Participants in the treatment group will also be asked to complete some open-ended questions about their experiences with the chatbot program. A subset of participants from the treatment group may be asked to complete optional phone or video interviews to gain a better understanding of their experiences. Results will provide insight into the usefulness of chatbot programs for reducing negative mental health symptoms among patients with a chronic health condition. Results may also be used to inform policy decisions about the use of these programs for healthcare delivery, and to provide practical insight into how these programs can be best integrated into healthcare settings.

Study Overview

• Status: Completed
• Conditions: Depression; Diabetes; Arthritis; Stress; Anxiety
• Intervention / Treatment: Device: Wysa

Detailed Description

People dealing with chronic health conditions are susceptible to mental health issues such as depression and anxiety. Providing conventional mental health services to all of these individuals is not practical given the limited resources of the healthcare system. Artificial intelligence (A.I.) mental health chatbots may be an accessible and cost-effective means by which people can receive some degree of mental health support while they cope with their conditions. These automated programs act as a source of virtual support, talking with individuals and providing them with therapeutic exercises to improve their mental wellbeing. Several chatbots have been designed to deliver interventions based on popular psychological therapies (e.g., Wysa, Woebot, and Tess). Research has shown that these programs can reduce symptoms of depression, anxiety, and stress in nonclinical populations. However, the effectiveness of these programs has not been tested in chronic disease populations.

The purpose of the current research is to gain a better understanding of the usefulness of mental health chatbots for chronic disease populations. This research will be guided by two fundamental objectives: (1) to determine whether a mental health chatbot can reduce or prevent negative mental health symptoms in individuals who are dealing with a chronic health condition, and (2) to learn more about how individuals with a chronic health condition view these programs, particularly in terms of their potential benefits or drawbacks when used in healthcare settings. This research will focus on two specific chronic disease populations that are prone to elevated levels of mental health symptoms: people with arthritis and diabetes.

Participants will be recruited through social media channels (including online groups), newspaper advertisements, and emails and newsletters from relevant organizations (e.g., the Arthritis Society, Diabetes Canada). After volunteering to participate, participants will set up a phone or video conferencing call with the primary investigator to orient them into the study. Participants will be randomly assigned to either a treatment group or control group. Those assigned to the treatment group will download the mental health chatbot Wysa (Touchkin eServices, Bangalore) on their smartphones. They will interact with the chatbot a minimum of two times per week over a period of four weeks, with each interaction lasting a minimum of five minutes. Participants assigned to the control group will receive no chatbot (i.e., they will be in a no-treatment control group).

Regardless of their group assignment, participants will complete online materials via Qualtrics at the outset of the study, two weeks into the study, and four weeks into the study (i.e., the final assessment point). At the outset of the study, participants will fill out an informed consent form, a demographic questionnaire, and four psychological assessments tools: measures of depression, anxiety, stress, and life satisfaction. Two weeks into the study, participants will complete the four psychological assessment tools a second time. Four weeks into the study, participants will complete the four assessment tools a final time, and those in the treatment group will be presented with a post-study questionnaire that contains qualitative questions regarding their experiences with the chatbot. Participants in both groups will be presented with a debriefing form providing more information about the study. Those in the control group will be given the opportunity to download and use the chatbot at this point.

After the data from the four-week study are analyzed, a subset of participants from the treatment group may be asked to complete optional phone or video interviews to gain more insight into their experiences with and opinions on the chatbot program. Approximately 15 to 20 participants will be sought for the interviews. The questions for these interviews will be developed based on the collective results from the quantitative and qualitative analysis described above.

• Study Type: Interventional
• Enrollment (Actual): 79
• Phase: Not Applicable

Contacts and Locations

Study Locations

• Canada
  • New Brunswick
    • Saint John, New Brunswick, Canada, E2L4L5
      • Centre for Research in Integrated Care, University of New Brunswick

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: 14 years to 61 years (Adult, Older Adult)
• Accepts Healthy Volunteers: No

Description

Inclusion Criteria:

• Participants must have a diagnosis of diabetes (type 1 or type 2 diabetes) or arthritis (osteoarthritis, rheumatoid arthritis, or another type of arthritis).
• Participants must have a phone with an active Internet connection.

Exclusion Criteria:

• Participants must not be receiving ongoing treatment from a mental health professional.
• Participants must not be using a mental health chatbot (i.e., prior to the study).
• Participants must not have started or experienced a dosage change in a psychopharmacological drug within the previous month.

Study Plan

How is the study designed?

Design Details

• Primary Purpose: Supportive Care
• Allocation: Randomized
• Interventional Model: Parallel Assignment
• Masking: Single

Number of Arms

2

Arms and Interventions

Participant Group / Arm	Intervention / Treatment
No Intervention: Control Group
Experimental: Treatment Group (Wysa)	Device: Wysa; Wysa is an AI-based mental health chatbot. It uses evidence-based techniques to help users build mental resilience skills and improve their mental health.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Changes in Depression as Measured by the Patient Health Questionnaire (PHQ-9)	The Patient Health Questionnaire (PHQ-9) contains nine items, each of which is rated on a scale of 0 to 3. Ratings are summed for a composite score; the minimum score is 0 and the maximum score is 27. Higher scores indicate higher levels of depression.	Participants are assessed at baseline, two weeks into the study, and four weeks into the study.
Changes in Anxiety as Measured by the Generalized Anxiety Disorder Scale (GAD-7)	The Generalized Anxiety Disorder Scale (GAD-7) contains seven items, each of which is rated on a scale of 0 to 3. Ratings are summed for a composite score; the minimum score is 0 and the maximum score is 21. Higher scores indicate higher levels of anxiety.	Participants are assessed at baseline, two weeks into the study, and four weeks into the study.
Changes in Stress as Measured by the Perceived Stress Scale (PSS-10)	The Perceived Stress Scale (PSS-10) contains ten items, each of which is rated on a scale of 0 to 4. Ratings are summed for a composite score; the minimum score is 0 and the maximum score is 40. Higher scores indicate higher levels of stress.	Participants are assessed at baseline, two weeks into the study, and four weeks into the study.

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Changes in Life Satisfaction as Measured by the Satisfaction With Life Scale	The Satisfaction with Life Scale contains five items, each of which is rated on a scale of 1 to 7. Scores are summed for a composite score; the minimum score is 5 and the maximum score is 35. Higher scores indicate higher levels of life satisfaction.	Participants are assessed at baseline, two weeks into the study, and four weeks into the study.

Collaborators and Investigators

• Sponsor: Luke MacNeill
• Collaborators: New Brunswick Innovation Foundation; Touchkin eServices Private Limited
• Investigators: Principal Investigator: Luke MacNeill, PhD, Centre for Research in Integrated Care, University of New Brunswick

Publications and helpful links

General Publications

• Fitzpatrick KK, Darcy A, Vierhile M. Delivering Cognitive Behavior Therapy to Young Adults With Symptoms of Depression and Anxiety Using a Fully Automated Conversational Agent (Woebot): A Randomized Controlled Trial. JMIR Ment Health. 2017 Jun 6;4(2):e19. doi: 10.2196/mental.7785.
• Fulmer R, Joerin A, Gentile B, Lakerink L, Rauws M. Using Psychological Artificial Intelligence (Tess) to Relieve Symptoms of Depression and Anxiety: Randomized Controlled Trial. JMIR Ment Health. 2018 Dec 13;5(4):e64. doi: 10.2196/mental.9782.
• Inkster B, Sarda S, Subramanian V. An Empathy-Driven, Conversational Artificial Intelligence Agent (Wysa) for Digital Mental Well-Being: Real-World Data Evaluation Mixed-Methods Study. JMIR Mhealth Uhealth. 2018 Nov 23;6(11):e12106. doi: 10.2196/12106.
• Ly KH, Ly AM, Andersson G. A fully automated conversational agent for promoting mental well-being: A pilot RCT using mixed methods. Internet Interv. 2017 Oct 10;10:39-46. doi: 10.1016/j.invent.2017.10.002. eCollection 2017 Dec.
• Clarke DM, Currie KC. Depression, anxiety and their relationship with chronic diseases: a review of the epidemiology, risk and treatment evidence. Med J Aust. 2009 Apr 6;190(S7):S54-60. doi: 10.5694/j.1326-5377.2009.tb02471.x.

Study record dates

Study Major Dates

• Study Start (Actual): October 1, 2021
• Primary Completion (Actual): October 31, 2022
• Study Completion (Actual): October 31, 2022

Study Registration Dates

• First Submitted: November 2, 2020
• First Submitted That Met QC Criteria: November 2, 2020
• First Posted (Actual): November 9, 2020

Study Record Updates

• Last Update Posted (Actual): August 26, 2024
• Last Update Submitted That Met QC Criteria: July 31, 2024
• Last Verified: July 1, 2024

More Information

Terms related to this study

• Keywords: Depression; Diabetes; Arthritis; Anxiety; Stress; Chatbot; Conversational Agent
• Additional Relevant MeSH Terms: Behavioral Symptoms; Joint Diseases; Musculoskeletal Diseases; Depression; Arthritis
• Other Study ID Numbers: REB 036-2020

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: NO

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No