Demystifying AI’s role in healthcare

A year ago, ChatGPT took the world by storm and thrust artificial intelligence (AI) into boardrooms and dinner parties alike. But with familiarity came foreboding. In a recent Pew Research Center survey, 52% of Americans said they were more concerned than excited about the use of AI in daily life. That wariness is even stronger in healthcare, with a separate Pew survey showing that 60% of U.S. adults are not comfortable with their providers relying on AI for medical care.

While some concerns about AI may be valid, much of the fear about its use in healthcare is due to misconceptions about the technology and how it would be applied. At the recent Cain Brothers Private Company Healthcare Conference in New York, Jill Frew, managing director at Cain Brothers, pulled back the veil on AI’s role in healthcare, with the help of panelists Allon Bloch, Co-founder and CEO at K Health; Mark Michalski, CEO at Ascertain; and Jane Sarasohn-Kahn, Founder of THINK-Health.

The panel kicked off with a quick overview of AI and how its application in healthcare differs from other, more familiar uses.
 

AI and the importance of knowing when you don’t know

AI refers to software that can solve problems and perform tasks by processing and analyzing data. Types of AI generally fall into two categories: traditional and generative. Traditional AI follows rules and patterns defined by a human, while generative AI can train itself and create new content. With both types, an AI application works by parsing data to formulate what it thinks is the best response to an input or prompt. K Health’s Bloch said that capability lies at the heart of common concerns about AI in healthcare.

“ChatGPT looks like it knows what it’s talking about, even when it’s wrong,” said Bloch. “In providing healthcare, we can’t afford to rely on technology that makes up answers. The wrong diagnosis or treatment could harm someone. ChatGPT doesn’t know how to say, ‘I don’t know,’ and sometimes that’s the best answer.”

That, Bloch added, is the big difference between clinical AI and applications like ChatGPT. Educating providers and consumers about this crucial distinction could go a long way toward making both groups more accepting of AI in medical care.
 

Administrative applications of AI

Clinical AI is just one form of artificial intelligence under the umbrella of healthcare AI. Other applications target more mundane tasks, such as information gathering and claims processing. While these solutions don’t have a direct impact on patient care, they improve healthcare delivery indirectly by reducing costs and administrative burden.

“More than a quarter of the $4.7 trillion that we spend on healthcare is on administrative cost,” said Cain Brothers’ Frew. “AI can really have an impact on not only the dollars spent but the efficiencies created for the staff and physicians.”

Some of the use cases the panel highlighted include scheduling, chronic care management, and revenue cycle management. “This current iteration of artificial intelligence doesn’t just work as a chatbot — it can actually start to navigate some of those management challenges in an automated fashion,” said Ascertain’s Michalski. “Bring in the bots for the paperwork, so our providers can be there for the patients.”

The benefits of increased efficiency go well beyond operations. As Michalski explained, being able to hand off administrative tasks to AI will help address a growing problem that threatens the entire U.S. healthcare system: burnout. He noted that, on average, a family practice doctor spends four to five hours charting — tracking and updating patient status and records — each week.

“That’s time that they’re not spending with their patients or their families,” said Michalski. “I’ve been out of medical school for 15 years, and there’s a sizeable proportion of my classmates who no longer practice medicine. The professionals who we’re all going to rely on as we age are leaving the profession. That is not just a tragedy, it’s an emergency. Those are the stakes when we’re talking about the interface between clinical medicine and operational inefficiencies.”

With the streamlining AI applications can offer, one would think physicians, nurses, and physician assistants (PAs) would be clamoring for these solutions. Unfortunately, the panelists agreed, that’s not yet the case. “There’s a lot of skepticism on the part of the providers,” said Frew.
 

Support for AI needs to come from the frontline

For the public to embrace AI in medicine, doctors, nurses, and PAs will have to embrace it first. Yet, frontline workers are hesitant to jump on the bandwagon after earlier technologies that promised to make their lives easier ended up doing just the opposite. “When technology gets introduced, doctors are quite jaded because it usually adds work rather than simplifying their life,” explained Bloch.

In the past, technological solutions have often failed to deliver greater efficiency because they required providers to gather and manage more information. As Michalski pointed out, this actually leaves the healthcare sector primed for AI, which can collect enormous amounts of data and quickly process it into actionable insights, while simultaneously alleviating the administrative burden on frontline workers.

“Current electronic health records systems are really good at increasing the scope and volume of data, so providers hate it because they feel like, in some ways, paper charting was more efficient,” said Michalski. “Now, with AI, we finally have a technology that actually reduces the scope again, so that it becomes manageable in the tight time frames that we work in.”
 

Education is the key to realizing AI’s full potential

Thanks to ChatGPT, more people than ever are aware of AI, but too few of them understand that ChatGPT is just one form of AI in a broad spectrum of solutions that can be tailored to bring improvements to almost every industry, including healthcare.

“There’s a lot of education to be done,” said THINK-Health’s Sarasohn-Kahn. “We need trusted AI ambassadors to start talking about this in plain language to real people very soon. And there’s nobody the public trusts more than their doctors, nurses, and pharmacists.”

The panelists expressed optimism that AI’s role in healthcare will eventually be understood and embraced by consumers, and that the technology will usher in a revolution in how providers operate.

“There are precious few times in your career when a technology comes along that allows you to break open some of these ossified workflows within healthcare, and really start to reengineer them,” said Michalski. “There are some things that we'll be able to automate that were historically very, very, very challenging to automate.”

Added Bloch, “I think you’ll see bigger changes in medicine over the next 10 years than in the last 50.”

To learn more about our industry expertise, visit key.com/cainbrothers, or to discuss this topic and more as it relates to your business, connect with our healthcare investment banking experts.