A somewhat rare use of artificial intelligence in senior living is helping an operator maximize limited resources by predicting the severity of COVID-19 symptoms in residents to determine who should be tested more frequently and who will benefit the most from the vaccine when it becomes available.
With Oakland, CA-based machine learning algorithm developer Dascena, Bloomfield, NJ-based Juniper Communities has launched an algorithm-driven COVID-19 smart surveillance testing program for senior living.
The program is important as vaccination is rolling out slowly across the country, Juniper founder and CEO Lynne Katzmann, Ph.D., told McKnight’s Senior Living. The company has long-term care communities in Colorado, New Jersey and Pennsylvania.
The first of Juniper’s vaccination clinics in an assisted living community, in New Jersey, will occur Jan. 20, she said. Two of Juniper’s first vaccination clinics in skilled nursing occurred Dec. 29 and Saturday.
“What this algorithm does is enable us to predict who of our residents we should test twice a week, to make sure that they are not likely to get COVID” and need hospitalization, Katzmann said.
Juniper already is testing employees twice a week, she said. Because of a previous cheek swab validation study the company did with Dascena, nasal swabs no longer are necessary for the company’s tests of workers and residents, “which is really important, particularly in memory care,” she said.
Using AI to improve workflow
Juniper and Dascena also worked together to develop a documentation and reporting portal and associated workflow
“It’s not just the test that is the expense. The test carries with it a whole host of staffing-related indirect costs that no one really talks about,” Katzmann said. “So we worked with Dascena to streamline the documentation process to effectively automate the process involved in documenting and taking tests.”
In a preliminary study, the new portal and process cut administrative time in half, saving “significantly” on indirect costs associated with testing, according to Juniper.
Even after vaccines are administered, COVID-19 testing frequency will continue for a month, Katzmann said, because the vaccines are not 100% effective and people can be asymptomatic transmitters of the disease. The company has made COVID-19 vaccination a condition of employment.
Offering lifestyle-based health solutions
And Juniper plans to expand its use of the algorithm, she said.
“Our plan for the rest of the year is to look at things that are more predictive in terms of lifestyle and diet and things like that, to help people make the best choices for staying healthy,” Katzmann said.
Using data on preferences and lifestyle choices, as well as vital statistics and information about chronic conditions, she said, Juniper will be able to identify the “best path” for each resident. “Obviously, we can’t control the choices, but we can offer solutions. We can offer a pathway,” she said.
Juniper is developing a new program, Catalyst, for rollout later this year. Katzmann called it “a much broader lifestyle management program” than its current Connect4Life care coordination program.
An independent analysis released in 2017 found that the Connect4Life program resulted in 50% lower inpatient hospitalization rates, more than 80% lower readmission rates and 15% lower emergency department use for Juniper residents compared with Medicare beneficiaries who did not live in Juniper communities but had similar conditions.
“Connect4Life brought together healthcare providers and integrated them with the work we do in the [senior living] community,” Katmannn said. “Catalyst will be a much broader version of that that will look at activities and food and a variety of different pieces of the puzzle that enable people to live a healthy life.” Using AI.
Growing use of AI
Juniper’s relationship with Dascena represents a somewhat uncommon but growing use of artificial intelligence in senior living. Other senior living operators and associations previously have used AI to help address workforce challenges.
Waltham, MA-based Benchmark Senior Living and McLean, VA-based Sunrise Senior Living, for instance, worked with a company called Arena to improve employee retention. And in April, Argentum and the American Seniors Housing Association announced the Senior Living FastMatch resource, which uses artificial intelligence by Arena to try to help workers who have been displaced due to the fallout from COVID-19 find new employment opportunities in senior living. In October, researchers at the University of California, San Diego, School of Medicine published a research paper about their use of natural language patterns-understanding software and other machine-learning tools to identify degrees of loneliness in independent living residents.
Juniper’s effort, Katzmann said, is “a very tentative first step into using AI to drive results in senior living. That’s what’s so exciting to me about it. I see that as the future for a lot of the work that we do. …Rather than sickness treatment, it’s about health and prevention.”