The rapid pace of evolving artificial intelligence (AI) has hospices navigating complex decisions around the increasingly diverse range of uses for this type of technology.

This is according to Evan Kramer, senior vice president of innovation and operational excellence at Compassus. The Tennessee-based company leverages technology to strengthen the delivery, timing and experience of its continuum of home-based services, including end-of-life care. Compassus is a portfolio company of the private equity firm Towerbrook Capital Partners and the health system Ascension Health.

AI tools have reduced documentation errors, remotely monitor patients, streamline operational efficiency and improve clinical capacity, among other uses. However, expanding AI utilization has come with mounting concerns about the potential outlook of how health care will be accessed, Kramer indicated. Providers must carefully consider the array of reverberating unintentional impacts of AI integration alongside the potential benefits, he stated.

“People may be worried about AI replacing them in certain ways, but in health care there will always be a human-to-human connection” Kramer told Hospice News. “Technology enables us to be a lot more specific and tailored in a way that ultimately we serve each patient. Our particular space is unique in that there just aren’t enough people to do the work and only about half of eligible patients even receive hospice. So, there is an opportunity to do things differently and spend more time with patients.”

Slow, steady AI integration

Compassus’ continuum of home-based services includes hospice, palliative care, home health and home infusion. The company has more than 300 locations across 33 states, including a number of similar partnerships with health systems and other health care providers.

More than 70% of Compassus clinicians have seen an average 60-70% reduction in time spent on administrative tasks due to AI documentation tools.

“Our primary goal so far with AI documentation is really to be a satisfier for clinicians,” Kramer said. “It’s been a really big milestone for us to see positive results from our team. We have a lot of anecdotal evidence that it’s really moving the needle for people, and we plan to do engagement surveys for more quantitative measures.”

The company’s approach to AI integration is to first optimize, then automate these tools in ways that help scale access, strengthen its workforce capacity and ensure measurable value to patients and staff, according to Kramer. The level of service capacity can “go up dramatically,” when AI tools are employed, he said.

Compassus has leveraged AI documentation tools that automate some clinical tasks related to recertification and routine patient visits. The company’s goal is to expand its scope of AI documentation for patient admission and discharge processes, according to Kramer.

The company plans to pilot more AI tools by the end of the year. Among the various avenues are exploring AI utilization for intake, discharge and recertification processes. Other uses for AI include navigating referral and admission processes in a more timely manner and streamlining revenue cycle management (RCM) and billing, Kramer indicated.

AI integration has thus far improved staff satisfaction at Compassus by allowing for more time to be spent at the bedside versus documenting, he stated. Clinicians’ feedback on AI use has included better work-life balance and the company has also seen stronger patient and family satisfaction.

Hospices should carefully consider and evaluate the different ways they leverage AI based, Kramer said. Important factors include the potential for AI to improve quality, compliance, care coordination or reduce redundancies in clinical and administrative workflows, he stated. Also key is having systems in place that measure the efficacy and performance of AI tools over time.

“The smartest move isn’t rushing into AI, it’s fixing the manual workflows,” Kramer told Hospice News. “By automating bad processes, organizations may end up compounding inefficiencies. It’s been very data-driven and a massive step up in terms of the overall quality of care. That’s going to take some time for AI to measure things quantitatively.”