A.I. Helps Shape the Future of Hospice Care

Artificial intelligence technology is in its infancy, but hospice organizations are already applying the technology to provide patient support, reduce costs, and improve efficiency.

“AI systems tend to make fewer mistakes than humans and can help in providing extra layer of assurance and inspection,” Jye Su principal analyst for ABI Research told Hospice News. “This by no means replaces fully trained and certified hospice professionals. This is meant to assist these professionals to make better decisions.”

Arizona-based Infinity Hospice Care used an application called Applied Insights, developed by health care technology developer WellSky, to achieve a 97% increase in patient visits during the last seven days of life. The program monitors electronic health records to identify signs that the patient’s final days are nearing. This kind of predictive modeling helps ensure that patients receive essential care and helps hospices ensure they comply with associated Medicare requirements.

Another example is Carrollton, Tex.-based Hospice Source, which provides durable medical equipment and respiratory therapy services exclusively to hospice patients. The company recently implemented an AI-based application called ServicePower to to improve responsiveness to patient and clinician calls. Among other functions, the system identifies the necessary services, recommends the fastest travel routes to the patient’s location, and indicates which staff members should respond based on their documented skills.

Referring organizations are also exploring AI. Stanford Hospital in California is piloting a deep learning application to predict the life expectancy of their patients, enabling physicians to refer those patients to hospice or palliative care sooner rather than later, according to Stanford Health. Early referrals benefit patients and their families, as well as increase the revenue streams of the hospice organizations providing that care.

“Nearly 30 percent of Medicare beneficiaries have hospice care for less than seven days— far too short a period to benefit from the skills of the hospice interdisciplinary team,” said National Hospice and Palliative Care Organization President Edo Banach. “People need earlier access to hospice and they need access to pre-hospice palliative care.”

In addition to anticipating patient needs and outcomes, hospice providers are learning to use AI systems for monitoring patients in real time.

“We have observed that AI can already use cardiac, respiratory, and motion sensors to predict worsening heart conditions, apnea, sepsis, fall risk, pressure ulcer risk, arrhythmia, or asthma attacks seven hours in advance,” Lian said. “This can reduce patient deterioration costs by 38 percent, patient falls by 43 percent, and pressure ulcers by 64 percent,” Lian told Hospice News.

Massachusetts-based NVNA and Hospice last month announced a partnership with technology company Health Recovery Solutions to bring remote physiologic monitoring systems to home-based hospice patients. The hospice uses bluetooth biometric devices capture and automatically transmit information about the patient’s status to dedicated nurses who monitor the telehealth system.

While hospice providers explore the potential of existing AI systems, researchers are hard at work developing the next generation of AI products to boost hospice efficiency and improve bedside care.

Timothy Bickmore, Ph.D., professor of computer and information science at Northeastern University in Boston, is leading a clinical trial that brings virtual conversation assistants into the homes of more than 100 hospice patients to help them stay informed and socially engaged. The system also clinicians to changes in the patient’s condition, enabling a faster response.

Through a computer tablet, patients can interact with an animated character that can answer questions as well as provide referrals to palliative care or hospice specialists. The system is designed to help patients stay socially engaged through interaction, telling stories, jokes, and a spiritual care module that can help the patient manage emotions associated with dying.

“The system provides support activities in line with the patient’s spiritual or cultural beliefs,” Bickmore told Hospice News. “This includes reminding them of religious holidays, recommending prayers, guided meditations for stress reduction, and a range of other activities.”

According to developers adapting AI systems for hospice and palliative care, costs of these systems can be minimal depending on the AI service providers. Most existing solutions are cloud-based with cost structures based on usage. Savings from increased efficiency, automation, and resource management can help mitigate expenses. A 2018 report from ABI research estimated that AI applications would yield $21 billion in savings across all health care settings by 2021, Home Health Care News reported.

“Service automation is one way to maximize staff resources and improve return of investment [for AI systems]. Automation is also good to resolve labor crunch by making existing workers more efficient,” Lian said

Hospice leaders that are considering these systems need to carefully assess the technology. A rising number of startup companies are developing health care AI systems using popular conversation assistants akin to Apple’s Siri and Amazon’s Alexa, but programs like these are not yet precise enough for medical application.

“Studies have found that when patients ask popular conversation assistants medical questions, they sometimes give inaccurate information that can influence the patient to make the wrong decision in seeking care,” Bickmore told Hospice News.