UMass Amherst and REOFTech to Train AI to Support the Relationship Between Alzheimer’s Disease Patients and Their Caregivers The platform is poised to help caregivers better manage their patients’ behaviors relying on off-the-shelf wearable health tracker

Original Article

Researchers from the University of Massachusetts Institute for Applied Life Sciences (IALS) and technology and consulting company REOFTech have been awarded $280,000 to conduct a pilot study to develop a new at-home technology platform to help patients and their caretakers manage Alzheimer’s disease and related dementias (ADRD). 

Funded by the Massachusetts AI and Technology Center for Connected Care in Aging and Alzheimer’s Disease (MassAITC), the study will use a combination of wearable technologies and a suite of smart home sensors to capture data from 22 households of patients with these conditions. This data will train an AI model to determine the actions that increase or alleviate agitation in patients as well as the strategies that support or hurt caretaker wellbeing. 

“We are gaining a better understanding of the complexities involved in caring for individuals with Alzheimer's disease and other dementias. The influence a dementia patient has on a family caregiver’s well-being, and vice versa, can be significant and should not be underestimated. Structuring care to address the needs of all household members often leads to the best outcomes,” explains Michael Busa, clinical professor at IALS and director of the Center for Human Health & Performance. “To support this approach, we are exploring how technology and machine learning can reveal more about the interconnected well-being of caregivers and patients. Our goal is to develop a system that helps patients, family caregivers and medical teams create effective treatment plans and services tailored to the entire household.”

“We’re in the process of figuring out how to manage patient agitation and caregiver well-being from a dyadic standpoint,” says G. Antonio Sosa-Pascual, CEO of REOFTech, a technology company developing digital health tools. “Our theory is that this relationship is symbiotic, and that through small interventions based on data, we can have a positive impact on health outcomes.”

Yet, with dementia’s diverse and changing symptoms, the patient-caretaker relationship is a challenging one. “Our customer discovery led us to gain a deep understanding of the caregiver’s needs and experiences,” he continues. “One caregiver told us that, every day, it seemed like they were dealing with a new patient.”

The researchers’ vision beyond this pilot study is to expand on REOFTech’s existing patient information management system, EC-Safety®, a proprietary electronic health record that uses validated tools to capture clinical data, such as cognitive decline, medical/social history, and depression. It is available in Spanish and English with versions for clinicians and home caretakers.

In its most mature form, beyond the scope of this pilot study, the intelligent service will create an AI Dynamic Care Plan© that responds to the situation at hand by having AI assess relevant behavioral or physiological markers related to agitative (or other) events captured from off-the-shelf smart home technology and wearables, explains Busa. 

In order to realize this ultimate concept, the researchers will use EC-Safety as well as data collected from wearable watches (such as Fitbits), other smart home sensors and surveys. They are implementing a no audio, no video strategy because they are confident that they can get sufficient data without compromising peoples’ feelings of privacy. 

Image

Data Core, part of the Center for Data Science in the Manning College of Information and Computer Sciences, will then use this data to train the initial AI model to find what behaviors and patient/caretaker outcomes are the most closely related. 

This research is being conducted with OPAPA, an Alzheimer’s disease care organization in San Juan, Puerto Rico, in part because Puerto Rico has a higher prevalence of Alzheimer’s and related dementias that are associated with higher mortality rates than the U.S. at-large.

“End-user representation at all stages of product design and testing are critically important to building tools that patients and care teams want and will use while also mitigating bias,” says David Paquette, associate director of research for MassAITC. “REOFTech, based in San Juan, PR, has been engaging with and developing tools for outpatient dementia clinics in Puerto Rico for the past four years and is ideally positioned to make an impact on dementia care in the Latinx community.” He also mentions that, given the high Puerto Rican population in Holyoke, research in Western Massachusetts could be a future direction for this work.

Another aim of the research is to address the fact that family caregivers are largely unpaid for their work.

“One of the great motivators of this project, at least for me, is the ‘helplessness’ that families face from all angles,” says Sosa-Pascual. “The big helplessness is of course that there’s no cure. But second is that there haven’t been many reimbursement mechanisms. [This is] a way to empower the caregiver to take some kind of control over their life in this journey.” This platform would provide a mechanism for caretakers to be paid for their currently unpaid labor through Medicare and Medicaid.

This grant was one of 11 pilot programs awarded through MassAITC, totaling $2.3 million.