Usability and functional validation of artificial intelligence‐enabled technology services in dementia care using a controlled trial in clinical practice

Abstract

AbstractBackgroundMany community settings lack emotional, social, and educational resources to support persons living with dementia (PLWD) and their families, which contribute to the stressors associated with caregiving, termed "caregiver load." This paper describes the impact of artificial intelligence (AI)‐enabled personalized assistive and decision support services on improving dementia care management and reducing caregiver load. We present a study that assesses AI’s usability and functionality in the outpatient setting and is built off of two prior pilots, with each pilot expanding in its scope of technology provided, target population, and caregiving needs.Objectives: The study’s primary outcome is the functionality and usability of AI services, collectively called NeuViCare, as measured by the System Usability Scale (SUS) from caregivers. Secondary outcomes include: 1) reduction in caregiver load as measured by Oberst Caregiving Burden Scale (OCBS); and 2) increase of caregivers’ knowledge as measured by Dementia knowledge assessment scale (DKAS).MethodsThis one‐arm, open‐labeled trial is recruiting ten patients with MOCA scores of less than 26 and impaired activities of daily living and their one designated caregiver. Caregivers must have a smartphone, tablet, or laptop to access NeuViCare and to utilize all four of its HIPAA and FHIR compliant services for three months: 1) Task‐Support Service 2) Resource Advisor 3) Community; and 4) Knowledge Advisor. Periodic touchpoints with caregiver participants are in place as a safety measure.ResultsThe first pilot’s target population was aides from a national home care agency. The survey‐study reported NeuViCare's user experience to be “very positive” and “easy to use”; the quality of its information to be “clear, concise and with good depth”; and, 3) its value to be “useful for less experienced caregivers”. The second pilot focused on family caregivers from Mumbai. This study showed that Knowledge Advisor provided 97% correct answers. This third pilot study targets informal caregivers, and we hypothesize that mean SUS scores will be significantly associated with reduced OCBS scores.ConclusionEffectiveness of AI technology in reducing caregiver load will lead to a scalable, cost‐effective solution that can be made available to caregivers everywhere.