COMPASS: A Family-Centered, Culturally Responsive Conversational AI Assistant to Support ADRD Caregivers and Family Care Networks in Real-Time

NIH RePORTER · NIH · R43 · $499,996 · view on reporter.nih.gov ↗

Abstract

Project Summary The majority of people living with dementia (PLWD) in the U.S. are cared for at home by family or friends without formal training. The challenges of providing ADRD care, dealing with challenging ADRD behaviors, and coordinating with the family care network causes significant mental, physical, and emotional burden for caregivers, especially for minoritized ADRD caregivers who face additional adverse health outcomes and cultural barriers in accessing support. Commercial solutions are not scalable for affordable use (i.e., remote care with ADRD experts) or lack personalization and cultural responsiveness (i.e., mobile or web -based caregiver apps with “generic” information). Support groups pose scheduling, geographic, and stigma barriers and do not offer real-time expert feedback or evidence-based skills. Without 24/7 expert support, caregivers are left to search through the Internet for “generic” solutions. There is a significant unmet need for culturally relevant, accessible, and scalable personalized ADRD caregiving guidance and coordination support to address the real-time needs of diverse ADRD caregivers, PLWD, and families. We are developing a family- centered, conversational AI assistant called COMPASS to support ADRD caregivers in real-time with 1) personalized, culturally relevant ADRD caregiving guidance and 2) assistive coordination within their family care network. Our novel technology leverages advances in culturally responsive AI to understand diverse caregiver queries and personalize responses and delivery to each caregiver and PLWD’s personal/cultural context. A virtual avatar tailored to the caregiver’s preferences will be used to communicate guidance in areas consistent with existing and validated multicomponent ADRD care interventions to assist with the daily challenges of ADRD caregiving and communication within the family care network. With caregiver approval, COMPASS can help share actionable information to family members in the network. In Phase 1, our 12-month goals are to develop a prototype of COMPASS and test its initial usability/feasibility for racially and ethnically diverse AD/ADRD family caregivers. By providing real-time support and building a care community, COMPASS aligns with the NIA’s goals to improve access and adherence to evidence-based caregiving support while reducing caregiver burden and enabling effective care for the PLWD.

Key facts

NIH application ID
11008802
Project number
1R43AG087829-01A1
Recipient
BEAVER HEALTH, INC.
Principal Investigator
Emily S. Wang
Activity code
R43
Funding institute
NIH
Fiscal year
2024
Award amount
$499,996
Award type
1
Project period
2024-09-20 → 2026-08-31