Clinical and human oversight are important considerations in building artificial intelligence (AI) systems for safe, effective utilization in the hospice space.
Human touch is intrinsically part of hospice care, and a deeply needed component of AI implementation, according to Dr. Michael Lalor, CEO of Trellis Supportive Care. The North Carolina-based nonprofit provides hospice and palliative care, advanced care planning and caregiver support.
AI technology is moving rapidly, requiring careful navigation in end-of-life care delivery, Lalor said at the Hospice News ELEVATE conference. Hospices need to carefully weigh the opportunities alongside potential impacts on clinical documentation, operational burden and organizational culture.
“Responsible [AI] readiness is not so much about how fast you’re able to implement these products, but how intentionally you do it,” Lalor told Hospice News during the conference. “I think it will be important as you’re implementing these tools that you keep that in the back of your mind all the time. Can you improve care coordination? Is it something that builds organizational trust? Making sure that you have real clinical insight of the tools. If your AI adoption is outpacing clinical judgment, organizational governance or your organizational culture, you’re probably moving too fast.”
Hospices need to consider and establish clear AI governance, ownership and accountability, said Stephanie Adair Reed, vice president of hospice clinical services at Liberty Home Care & Hospice. Part of Liberty Health, the hospice provider operates across North and South Carolina and Virginia.
Technology standards must align with hospice requirements and regulatory standards, according to Adair Reed. Sustainable, responsible technology exploration requires building trust, and establishing safety guidelines for patients, staff and hospice leadership at both clinical and back office levels.
“Governance breaks down when no one truly owns it,” Adair Reed said at the conference. “If everyone assumes that IT, or compliance, or clinical leadership is reviewing performance, then no one actually is. So, there needs to be safeguards put in place. Some may think AI governance is just writing policies or buying software, but the hard part is deciding who is responsible when something goes wrong and making sure that the tool stays safe as workflows and data change. Human accountability is the most important guardrail.”
AI governance in end-of-life care can be challenged without strong data management policies in place, according to Andrew Olowu, chief technology officer of the health care technology platform Axxess. Insufficient oversight or clear protocols for patient data can risk privacy and satisfaction in significant ways for a hospice provider, Olowu said during the ELEVATE conference.
AI oversight involves staff education and engagement in the adoption of technology as it evolves in end-of-life care delivery, Olowu indicated. Hospices need to consider multiple aspects of the impacts of AI integration on patient care, operations, billing, compliance and quality. Systematic, adaptable approaches to technology are key.
“You want to keep an eye on making sure you protect patients’ data, making sure your staff is upskilled to be able to handle a lot of the new technology,” Olowu told Hospice News. “You have to have that oversight of what’s going on in the data, in the people and in the infrastructure. You need to have a continuous monitoring mindset. It’s an AI-risk management framework [that] allows you to have the ability to govern, map, measure and manage your risk very clearly.”
Companies featured in this article:
Axxess, Liberty Health, Liberty Home Care & Hospice, Trellis Supportive Care
