By Eric Taub
There’s no other topic that has generated as much attention in the general and business press in the past year as artificial intelligence (AI). Based on the breathless coverage, one could imagine that the technology has just been invented. In actuality, it’s been around for years.
AI fuels such voice recognition products as Amazon’s Alexa and Apple’s Siri, as well as chatbots, customer service scripting and Netflix’s viewing suggestions, among others.
However, with the release of ChatGPT, awareness of AI technology has come to the masses. Today, anyone can use AI to garner information from the web, compose documents in a variety of writing styles and conjure up graphics with just a few words of instruction.
Ethical Considerations
As with all new significant inventions, pundits are claiming that AI, with its potential power to take over mundane jobs, will either destroy the world or save it. We’ll all be slaves to a technology, or conversely use it to harness our abilities far beyond today’s imagination.
While the government develops guidelines to protect the public from AI’s worst potential excesses, AI tools are being created that promise to increase customer satisfaction and improve business efficiencies by freeing us from the most boring, repetitive tasks and having machines do them instead.
Nowhere is that truer than in healthcare and related fields such as skilled nursing care and senior living residences. All three business segments are faced with rising costs, mountains of paperwork, growing documentation requirements, requests for insurance pre-authorizations and constant staff turnover. New AI tools on the cusp of market introduction promise to help senior-focused organizations by giving them important new solutions to remain profitable and grow.
While exciting, the industry needs to be cautious in its introduction of AI, assuaging fears among many that it will decrease the quality of care.
That’s important because, according to a recent survey by The Harris Poll and sponsored by Athenahealth, a revenue cycle management and clinical records company, there are serious concerns in healthcare about the use of AI.
For those in the field who were wary of the technology, 60 percent were mostly worried about the loss of human touch. Another study conducted by Dynata found very similar concerns among consumers.
In a study published in June by the Journal of Hospitality Marketing & Management, consumers were significantly less likely to buy a product if it was advertised as using AI.
Workflow Applications
For the healthcare industry, the best use of AI will be through creating efficiencies in administrative tasks and other back-office functions, The Harris Poll study found. And that’s especially important in senior living, with residences commonly experiencing high staff turnover rates. If AI can make an operation more efficient, it can safely reduce its employee count.
AI tools are always meant to supplement, but never replace, human decision making.
“These are predictive models based on past behavior and data,” says Scott Code, vice president for LeadingAge’s Center for Aging Services Technologies. “They’ll never be 100 percent accurate. And they always need to support the staff and the facility’s residents.”
High staff turnover in senior living, which some say could at times be as high as 80 percent annually, presents a serious problem for maximizing resident well-being. “Now, too much information stays in staff persons’ heads. With 80 percent employee turnover, you’re losing 80 percent of residents’ information, unless that information was written down,” notes Dylan Conley, chief technology officer of LifeLoop, based in Englewood, Colorado.
The company provides workflow software to senior living residences, allowing them to keep track of such tasks as meals, messaging, repair work orders and resident transportation requests, and also provides engagement and curated content to residents.
The company is now working on integrating AI into resident activity calendars. Using aggregate patient information as to their likes and dislikes and event attendance history, staff can more efficiently generate novel weekly activity schedules for their clientele that cover all seven dimensions of wellness: physical, emotional, intellectual, spiritual, environmental, social and occupational.
The staff thinks of its weekly activities goals, taking into account weather conditions, age range, activity level and other factors, and AI will generate a unique activities calendar each week.
LifeLoop’s goal is to cut in half the amount of time required to produce a weekly activity calendar, which currently takes about four to five hours to complete.
To ensure that resident information is available to all staff members, the company is also working on automating the interview process, using AI to automatically extract salient information such as a resident’s interests and goals to add to its database for future calendar creation.
Eventually, that database will also store a resident’s weekly habits and then automatically ask the individual if he or she wishes to reserve transportation to a hair salon or other location based on the person’s traditional schedule.
Connecting People with Computers
Yardi, a provider of real estate management software for senior living and other industries, is developing an AI product dubbed Virtuoso. It will be able to respond to voice requests for access to data files. Staff will also be able to add to a resident file by simply stating the information verbally.
The company already uses AI to help maintenance staff repair items in a facility. A tech can send a photo of a leaky toilet to the Yardi app, and AI will return standard information on the model and a list of parts that typically wear out.
If the toilet cannot be fixed, the tech can order the part through the Yardi software, and the process will follow the facility’s budget approval and requisition system signoffs before it allows the purchase.
In the future, Yardi expects to integrate service requests into its consumer-facing app; when a resident reports a beeping sound in his or her unit, for example, an AI-generated request will suggest possible causes and solutions.
If it’s a simple fix that can be done by the resident, the operator will save the cost of sending out a tech to the unit.
Patient-Centered Approach to EHRs
While senior and assisted living residences are not healthcare facilities, they benefit when their residents enjoy better health, enabling them to live longer with an improved quality of life.
To increase efficiencies when an elder resident does need medical attention, Wellsheet, a software interface solution that sits on top of popular electronic health records (EHR) products including Athenahealth, Cerner and Epic, among others, uses AI to integrate patient data from across healthcare facilities. This gives physicians a more complete view of the person under their care.
The application uses AI to pull data from various healthcare sources and then surface the most relevant content for clinicians, contextualized and prioritized for their needs. Armed with this additional, more comprehensive data, physicians are able to spend less time with their patients.
Preliminary data from Wellsheet notes that client RWJBarnabas Health in New Jersey has experienced a 40 percent reduction in the average time a physician needs to spend using the EHR system, over a 16 percent drop in the average length of a hospital stay, and reduced levels of physician stress, according to Wellsheet CEO Craig Limoli.
Problem-Solving Capabilities
AI has also been harnessed to improve the well-being of dementia patients. TapRoot’s Ella app uses a patient’s health information to help caregivers make the best-informed decisions on how to resolve issues with dementia patients. For example, the app can suggest strategies that are most likely to tone down a tense situation with an agitated individual, rather than escalate it.
A resident suffering from dementia who refuses to bathe may need to be approached in a particular way or from a specific angle. If a resident had worked as a children’s book writer in his or her prime and has fond memories of the creative experience, the app might suggest reading from one to calm the patient down, rather than yelling at him or her to get in the bath.
Data indicates that by using TapRoot interventions, challenging, disruptive behaviors from patients are reduced by 87 percent and the use of psychotropic medications by 20 percent.
The app, which is now used in 30 memory care communities serving 1,200 patients, has just launched a consumer version, called Elbi, for at-home caregivers. The consumer app includes a care guide listing the 50 most common concerns facing those with dementia, such as falling and tripping.
Consumers pay $25 per month to use Elbi, which enables up to five different caregivers to work with an individual. The institutional version, Ella, costs licensees $350 per month, which covers the care of up to 35 patients, notes TapRoot co-founder, Dr. Linda Buscemi. The price rises by $50 to $80 for each additional set of 30 to 40 individuals.
Staff shortages, increased acuity among the elder population and a reduction in nurse training time is creating a “perfect storm” for the need to use AI, according to Mordy Eisenberg, co-founder of Tapestry Health, based in Stratford, Connecticut.
“Tech solutions in senior living are antiquated,” says Eisenberg. “I want to use AI for pattern recognition. Right now, because of staffing shortages, you can have a resident see 21 different healthcare staff members each week, and you have no longitudinal information on the patient; they have no idea what they’re looking at.”
Power of Predictive Analytics
Tapestry, whose software solution is currently used by 1,500 skilled nursing facilities serving 100,000 patients, is using AI to predict the likelihood of future hospitalizations based on past patterns.
Taken alone, shortness of breath in a congestive heart failure patient may mean nothing serious. But if the software indicates that the patient has also gained weight and has swelling in his or her extremities, that could be a strong predictor of a likely heart issue.
Armed with that information, residents can now be ranked according to their likelihood to need hospitalization, allowing operators to allocate resources more efficiently.
To increase knowledge of residents’ health, Tapestry also uses room-based radar sensors that track patient heart and respiratory rates in the bedroom and bath (the use of radar eliminates the need for cameras and preserves resident privacy). Added to the residents’ medical data, this gives staff an additional tool to track health conditions.
Tapestry is now in trial studies with the same radar to track falls. Residents who fall and then get back up on their own could indicate a propensity for increased fall likelihood, giving staff the opportunity to inform the family to stay alert to that possibility.
While concentrating on skilled nursing homes to date, Tapestry sees a “tremendous market in assisted living with a number of pilots now ongoing,” says Eisenberg.
Less than a decade ago, Wi-Fi was often a luxury in senior living. Now that it’s commonplace, the technology is being harnessed not only for internet access but also for AI-powered resident well-being.
Cognitive Systems’ Caregiver Aware system uses Wi-Fi to detect motion; when the signal that normally bounces off walls and floors is disrupted, motion can be discerned. By monitoring movement over time, the collected data uses AI to make predictions when movement anomalies are discovered.
Caregiver Aware is distributed by Electronic Caregiver of New Mexico under the CareAware name. The system can monitor a room up to 2,000 square feet. Four smart plugs transmit the signal; the set of plugs cost $60, plus a $20 per month subscription fee.
Next up: using that same signal to monitor activity levels and sleep interruptions, then partnering with third parties to add additional information from wearable devices and glucose monitoring, creating a comprehensive database of resident vitals.
While the company is currently selling its product only to those aging in place, “senior living sales are our logical next step,” says Geordie Hagerman, executive vice president for commercialization at the Waterloo, Ontario-based company. “We’re trialing with senior living residences now.”
Man, Machine Must Coexist
All of these examples use AI not to take over operations, but to serve solely as an adjunct to human oversight. In the end, it’s people, not machines, that must decide how AI will be harnessed.
“I believe in constitutional AI,” says LifeLoop’s Dylan Conley. “We must consider our responsibility to the public. AI requires supervision. That’s one of our tenets and is part of our AI mindset.”