Feasibility and Acceptability of an LLM-based Chatbot for Family Caregivers: Evaluation Study

ConditionCaregivers
Trial statusNot yet recruiting
Trial phaseNot applicable
Trial typeInterventional
Biological sexAll
Age18+
SponsorUniversity of Michigan

About this trial

This study evaluates the feasibility and acceptability of using the artificial intelligence (AI) chatbot mobile application (an LLM-based Chatbot) to support the well-being of informal care partners of individuals with traumatic brain injury (TBI), dementia, or Huntington disease (HD). Over an 8-week intervention period, up to 60 care partners will engage with the application (app), spending four weeks focusing on physical health and four weeks focusing on mental health. Participants will complete surveys at the midpoint and end of the intervention period. Participants will also complete a semi-structured interview to discuss participants experience using an LLM-based Chatbot.

The study team hypothesizes that care partners will find the chatbot acceptable and feasible to use in daily life, and that there will be low attrition and high completion rates of the study.

Eligibility criteria

This trial does not accept healthy volunteers

Qualifiers

Age 18 years or older

Able to independently provide informed consent

Able to read, speak, and understand English

Have an Apple (iOS) smartphone, tablet, or laptop, have access to the internet and text messages, and be willing to use these resources for the study, including downloading and using the AI chatbot app (Earkick)

Disqualifiers

Anyone not meeting inclusion criteria above

Professional, paid caregivers

Anyone at the discretion of the Principal investigator (PI) that would be a preclusion of safe or meaningful participation

Anyone not meeting inclusion criteria

Trial design

Design model

Single group

Treatments tested in this trial

  • LLM-based Chatbot

    Other intervention

    Participants will download the app with the study team's help and set it up. The participants will select topics of interest to focus on during the 8-week use of the LLM-based Chatbot. Participants will choose one mental health related topic to focus on from weeks 1-4 and one physical health related topic to focus on from weeks 5-8. During these 4 week cycles the participants will be guided to engage with the app's "Chat" feature, which offers conversational support through interactions with the chatbot. Additionally, participants will complete surveys at baseline, the midpoint, and end of the 8-week intervention period. Participants will also complete a semi-structured interview to discuss their experience using the LLM-based Chatbot.

Treatment groups

60 Participants
are divided into 1 treatment group
Group A: Earkick appExperimental treatment 1 intervention

Trial outcomes

Primary outcomes

1

Feasibility and Acceptability as measured by the Feasibility Questionnaire

Feasibility Questionnaire (items are scaled from 1 to 5 to indicate level of agreement, where "1" indicates "strong disagreement" and "5" indicates "strong agreement"). Feasibility and Acceptability as measured by ≥80% of participants indicating that care partners either "Agree" or "Strongly Agree" that the different study elements are feasible and acceptable.

Time frame
8 weeks (post intervention)

Secondary outcomes

1

Attrition as measured by the percent of participants completing the study

The study team expects ≥80% of participants to complete the study.

Time frame
Baseline to post intervention interview (approximately 8-12 weeks)
2

Adherence as measured by the percentage of missing data over the course of the study

Percent missing data of total expected data for each person and on average will also be calculated to characterize the feasibility of this approach (i.e., the study team expects ≥80% completion rates for the midpoint and end-of-study assessments).

Time frame
Baseline to 8 weeks

Other outcomes

Sponsors and contacts

Click on the lead sponsor to view all of their trials.

This trial is not recruiting at the moment. You can still explore other options: