This pilot randomized controlled trial evaluates the feasibility and preliminary efficacy of VICA, an AI-powered virtual care assistant that delivers evidence-based education, symptom monitoring, and coaching to informal caregivers of people living with Alzheimer's disease and related dementias (ADRD). Sixty caregiver-patient dyads will be randomized 1:1 to VICA or an attention-control education app (Dementia Guide Expert) for 3 months. The study will assess recruitment, VICA acceptance/usability, and engagement (Aim 1), and estimate effect sizes for reducing patient behavioral and psychological symptoms of dementia (BPSD) and caregiver distress (Aim 2).
VICA is a customized large-language-model-driven avatar that provides speech-to-speech and text conversational support grounded in the ABC (Aging Brain Care) collaborative dementia care model and the HABC-Monitor assessment. Informal caregivers randomized to VICA are encouraged to use the platform 3 times per week for approximately 20 minutes over 3 months; control participants use the Dementia Guide Expert education-only app on the same schedule. Outcomes include trial feasibility metrics (recruitment, retention, VICA acceptance via TAM-PEOU, usability via SUS, usage/engagement) and preliminary effect-size estimates for the Neuropsychiatric Inventory (NPI) total score (patient BPSD), NPI-Distress score, PHQ-9 (caregiver depression), and GAD-7 (caregiver anxiety), collected at baseline and months 1, 2, and 3.
Study Type
INTERVENTIONAL
Allocation
RANDOMIZED
Purpose
SUPPORTIVE_CARE
Masking
NONE
Enrollment
120
An AI-powered avatar delivering conversational, speech-to-speech and text-based education, resource information, and symptom-monitoring support to informal caregivers of patients with Alzheimer's disease and related dementias (ADRD), grounded in the ABC collaborative dementia care model via a retrieval-augmented generation system.
Access to Dementia Guide Expert, an existing free, education-only BPSD mobile app (ENGAGE-IL initiative, University of Illinois at Chicago), used on the same 3x/week, \~20 minute schedule for 3 months. No interactive coaching, assessment, or personalized response.
Eskanazi Health
Indianapolis, Indiana, United States
Neuropsychiatric Inventory (NPI) total score
A 12-domain caregiver-informant-rated measure of patient behavioral and psychological symptoms of dementia. Each domain is scored as frequency (1-4) times severity (1-3) if present, or 0 if absent. Domain scores are summed for a Total Score ranging 0-144; higher scores indicate more frequent and/or severe symptoms.
Time frame: Baseline, 1, 2, and 3 months
Neuropsychiatric Inventory-Distress (NPI-D) Scale
Derived from the same 12-domain NPI interview. For each domain in which a symptom is present, the caregiver rates how emotionally distressing that behavior is to them, from 0 (not at all) to 5 (very severely). Ratings are summed across domains for a total score of 0-60; higher scores indicate greater caregiver distress.
Time frame: Baseline, 1, 2, and 3 months
Assessment completion rate
Time frame: Computed at 3 months
Recruitment rate
Recruitment rate computed monthly as # approached, # agreeing, and \# consenting, tracked continuously throughout the accrual period.
Time frame: up to 15 months
Technology Acceptance Model-Perceived Ease of Use (TAM-PEOU)
A 6-item caregiver self-report measure of perceived ease of use, each item rated on a 7-point scale from Extremely Disagree to Extremely Agree. Item scores are summed for a total ranging 6-42; higher scores indicate greater perceived ease of use. Intervention-arm participants rate the VICA platform; control-arm participants rate the Dementia Guide Expert app, using parallel-worded versions of the same instrument.
Time frame: Months 1, 2, 3
System Usability Scale (SUS)
A 10-item caregiver self-report measure of perceived usability, rated on a 5-point scale from Strongly Disagree to Strongly Agree, with alternating positively- and negatively-worded items. Using standard SUS scoring, item responses are converted and summed to a total ranging 0-100; higher scores indicate better usability. Intervention-arm participants rate the VICA platform; control-arm participants rate the Dementia Guide Expert app, using parallel-worded versions of the same instrument.
Time frame: Months 1, 2, 3
VICA usage
Server-recorded log-ins, session progress/completion, and conversation content, collected from the VICA intervention arm only. Not applicable to the attention control arm, which does not use the VICA platform.
Time frame: up to 12 weeks
Patient Health Questionnaire-9 (PHQ-9)
A 9-item caregiver-reported measure of depressive symptom severity. Scored 0-27; higher scores indicate more severe depression.
Time frame: Baseline, 1, 2, and 3 months
Generalized Anxiety Disorder 7-item Scale (GAD-7)
A 7-item caregiver self-report measure of anxiety symptom severity, scored 0-3 per item and summed for a total ranging 0-21; higher scores indicate more severe anxiety.
Time frame: Baseline, 1, 2, and 3 months
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.