Hospices Traverse AI Danger Zones

Hospices are carefully navigating a rapidly changing landscape of artificial intelligence (AI) innovation. Among the significant challenges topping providers’ minds are ensuring goal-concordant care delivery, data accuracy and safeguarding patient privacy.

Due diligence is a crucial step for hospices when exploring different uses for AI technologies, according to Paola Pouponneau-Nisbett, senior vice president of home- and community-based services at MorseLife Health System.

Ensuring interoperability between AI tools and a hospice’s existing electronic health record (EHR) systems is key to patient data safety and compliance, Pouponneau-Nisbett said. Significant privacy issues can arise when these technologies are misaligned, she said during a recent Hospice News webinar.

“Whatever AI platform you use, make sure it is private,” Pouponneau-Nisbett told Hospice News during the webinar. “Make sure it has all of those safeguards. Make sure that it is HIPAA compliant. In order to ensure that quality of data, you have to make sure as you’re starting to use the program that has interoperability [with] already existing legacy systems. Sometimes there are [data] breaches that occur because of that integration.”

Florida-based MorseLife Health System provides home health, hospice, palliative care and Programs of All-Inclusive Care for the Elderly (PACE). The nonprofit also offers rehabilitation, long-term, independent and assisted living services, as well as Parkinson’s and cannabis-based therapies.

Careful consideration is needed when using AI in the hospice space, according to Pouponneau-Nisbett. Providers need AI platforms that are well-versed not only in data protection, but also in hospice-specific regulatory requirements, particularly those related to quality and documentation.

Having internal audit and review processes that oversee AI integration and utilization is significant to safeguarding both patients and providers, she indicated.

Hospices have used AI in various ways to improve workflows, operational efficiency, clinical capacity and documentation and revenue cycle management. Other uses have included predictive data analytics and patient monitoring systems that are aimed at improving timely access and utilization.

AI has increasingly become an avenue for public outreach, with more hospices seeing patients and their families using the technology to guide end-of-life decisions. While this could move the needle on improved awareness, some hospices have voiced concerns about the ethical considerations involved and the accuracy of information that AI tools provide.

Patient data privacy and security are key focuses of hospices’ approaches to AI innovation, said Bill Riemer, senior vice president of information systems and technology of Empath Health. Equally important is considering the regulatory, legal and clinical aspects of hospice care delivery, he stated.

Empath Health is the parent company of 17 affiliates and four philanthropic foundations. The Florida-based nonprofit’s Full Life Care model includes hospice, home health, palliative and primary care, bereavement, PACE and adult day programs, advance care planning, HIV and sexual health services, as well as pharmacy and medical equipment assistance.

AI programs can help streamline clinical documentation, but need a hands-on approach to governance, according to Reimer. Hospices need ways to ensure that these systems have built-in mechanisms that align care models with patients’ goals. This, in part, involves establishing strategic clinical and compliance oversight around technology processes, he said.

“[It’s] keeping in touch with the tech aspect of it on what is safe and what is not safe,” Reimer told Hospice News during the webinar. “The risks include not only patient data safety, but also making sure that there’s a clinical oversight and governance around [AI] to avoid any kind of misaligned care goals. The challenge is ensuring information is both effective and accurate across the board. Make sure that [AI] is hardwired in your governance process and you have ethical considerations, compliance, regulatory and legal, as well as technical and clinical aspects in that governance discussion and review.”

AI technologies are in the early stages of evolution, with hospices only beginning to tap into the various uses, according to Matt Tanzer, president of BAYADA Hospice.

New Jersey-headquartered BAYADA provides hospice, adult and pediatric home health, personal care and private duty nursing. Established in 1975, the nonprofit provides services across the United States, Germany, India, Ireland, New Zealand and South Korea.

BAYADA has taken a cautious approach to AI integration, carefully vetting technology companies and clearly communicating hospice-specific needs, Tanzer indicated during the Hospice News webinar. Hospices need AI tools that incorporate regulatory requirements and clinical and operational needs, he said.

A main challenge is that new AI tools and technology companies are increasingly emerging, with some of lower quality and weaker capabilities compared to others. This can make it difficult for patients and providers to discern and navigate safe, effective and accurate AI utilization, according to Tanzer.

“A huge danger of AI in some of these applications and important to get this right from a safety consideration is that this cannot replace clinical judgment,” Tanzer told Hospice News. “It’s important to get this right from a safety consideration. This is to support, it’s to help and enable. But there is a risk. It’s being able to reinforce that concept … with the governance, and the training and the tools to make sure that whatever we’re providing has those safeguards around it. It is critically important to ensure safety and efficacy of these new tools.”

AI can be a “powerful tool” to help strengthen access and improve care coordination, the American Telemedicine Association (ATA) indicated in a recent report. The association’s recently published report, Policy Principles on Artificial Intelligence (AI), examines different ways to ensure that AI innovation helps to build trust and a more connected health care system.

A key is to ensure that AI technologies are designed for accountability, transparency and stronger engagement. Also important is having clear regulatory guardrails and safeguards to mitigate bias and protect data privacy and security, ATA reported.

“Policies should encourage collaboration between technology developers, providers, and educators to ensure that AI complements clinical expertise and supports the long-term sustainability of the healthcare workforce,” the association stated in the report shared with Hospice News. “Policies should balance individual privacy and confidentiality with innovation, recognize that privacy-preserving approaches may necessitate appropriate limits on disclosure, tailor safeguards to system function and risk, and align with evolving federal and state standards.”