Janie Brackett, who suffers from a chronic condition that can make breathing difficult, worried recently when her weight dropped suddenly.
The 71-year-old LaGrange resident said she normally might not have noticed the drop because she hadn’t been checking her weight daily. But she is one of 20 patients at Penobscot Community Health Center enrolled in a pilot artificial intelligence program that requires her to check her weight, blood pressure and oxygen levels every day, all from home.
Her weight loss did not trigger an abnormal reading by the AI system, but Brackett was able to allay her fears by meeting with her doctor.
“It’s just peace of mind knowing the information goes to a doctor,” said Brackett, who lives with her husband in the rural town a half-hour drive north of Bangor. “I think a lot of seniors don’t take care of themselves and this can monitor their health and get concerns addressed quickly.”
Bangor-based Penobscot Community Health Center is at the forefront of a national trend in using AI to bolster telehealth services that became more important during the COVID-19 pandemic, allowing more granular analysis of symptoms from home. It also helps in rural settings where it might be a challenge to get to a health care provider.
“Most health systems are experimenting in one fashion or another with remote patient monitoring,” said Lori Dwyer, CEO of PCHC. “We can multiply our ability to care for people with the technology and we get a lot more information in real time that helps us better manage their condition.”
PCHC, which has 16 locations in Maine, started the pilot program last October and began enrolling patients in February. The participants, whose ages range from 50 to 90, have congestive heart failure, hypertension or, like Brackett, chronic obstructive pulmonary disease.
The pilot uses components from three companies that work together. The remote patient monitoring system from HealthSaas provides participants with a blood pressure cuff, oxygen-level reader and scale that can automatically send measurements to the tablet that they also get. Software from that company gets the information to an AI-driven database from AkeLex.
That database can assemble the measurements into an easy-to-read format and show whether or not they are in a normal range. If not, the third part of the pilot, PocketRN, matches an available registered nurse anywhere in the country with the patient to get them help quickly.
“PocketRN takes advantage of the existing RN workforce to find an available nurse that is licensed in your state to respond to a situation the patient’s usual care team is too busy to answer at the time,” Dwyer said.
In the event of an abnormal measurement, the pilot program also will use AI software to generate a series of questions for the patient to provide additional information to the caregiver. A patient with elevated blood pressure may have taken that reading just after they smoked a cigarette, ruling out other causes for the spike, Dwyer said.
“That gives patients a connection between what they do and what happens when they do it,” said Jill Bouchard, an adult gerontology nurse practitioner at PCHC.
Bouchard said the AI system also helps her see whether a patient’s medicine is working. One of her patients was getting high blood pressure readings in the morning. After responding to questions, she discovered he was not taking his blood pressure pills until bedtime rather than in the morning, when they were supposed to be taken.
“We increased the amount of medication and told him to take it in the morning,” Bouchard said. “The AI system allows patients to have symptoms addressed without having to come into the office and have a long waiting time.”
PCHC is considering adding a device that can help patients take medications more easily, she said. Like an electronic pillbox, it can determine if the patient has opened the daily dose and relay that to the tablet along with the blood pressure and other measurements.
The current pilot, which was funded by PCHC, runs through the end of June. The University of Maine kicked in $50,000 through a grant to pay for the remote patient monitoring equipment. PCHC wants to continue the pilot and expand its scope, including with the medication application and more participants. Dwyer estimates PCHC will have to raise up to $250,000 to expand the program.
In early June, PCHC added a small pilot with two patients diagnosed with congestive heart failure. The AI is asking them questions such as whether they have to sleep in a chair and whether their legs are more swollen than usual. Those questions can help detect whether their condition has suddenly worsened without having to go to the doctor’s office, Bouchard said.
In the 20-person trial, participants take measurements five days a week. That means health care professionals like Bouchard can learn about a patient’s health quickly and more often, as opposed to getting information from a doctor’s visit every six months to a year.
Patients, including Brackett, like knowing what their readings are more frequently and feel they are participating more actively in their own care.
“It has given me the push to do this every day,” Brackett said. “I’m not tech savvy, and this is easy for seniors or anyone to do.”