medical consultation with female doctor
The promise and peril of artificial intelligence made it to the world stage this week when President Joe Biden and other global leaders included AI in their addresses to the United Nations General Assembly.
Today, on a slightly smaller stage, one company is formally announcing its bet on AI’s potential to improve healthcare.
At its annual summit in San Diego, Hint Health, a digital health company that provides practice management software and other tools to Direct Primary Care (DPC) providers, is launching a new AI product in collaboration with OpenAI.
Under the DPC model, consumers pay a membership fee for sometimes unlimited access to their doctors. Typically operating outside the health insurance system, DPC providers are paid through these membership fees rather than billing an insurance company for each service provided.
The new product, Hint AI, will allow DPC providers to record patient consultations, transcribe those encounters, and automatically generate clinical documentation in the patient’s medical record.
Automating this task may sound mundane, but the potential time savings is significant. A 2022 analysis showed that doctors spend an average of 4.5 hours per day working in electronic health records, including documenting patient visits.
With Hint’s new tool, doctors must still review their notes, make any necessary edits, and then sign off. But according to Hint Health CEO Zak Holdsworth, doctors piloting the tool have generally been able to sign the AI-generated notes with minimal, if any, editing.
“The early results are shockingly good,” Holdsworth said in an interview.
Improving clinical documentation is just the beginning of how Hint will use AI, according to Holdsworth. He said the company plans to infuse AI capabilities throughout its platform. For example, he said that the company plans to release a HIPAA-compliant tool like ChatGPT, embedded into the medical records. Doctors will be able to query a patient’s medical history to prepare for upcoming visits, develop treatment plans, or identify patterns or trends in the patient’s condition. The company also plans to use AI to enable doctors to automatically generate referrals to specialists or other providers, Holdsworth said.
Even these expanded applications of AI are relatively limited compared to what will be possible in the future.
“AI has the potential to revolutionize healthcare in a way that enhances, rather than replaces, human relationships,” Aliisa Rosenthal, Head of Sales at OpenAI, said in a statement. “We are excited that Hint Health is using OpenAI’s technology to bring advanced tools to doctors who share our vision.”
Gayle Brekke, PhD, FSA, founder and chief influencer at Primary Care Mindset and host of the Nurturing the Heart of Family Practice podcast, said she believes that AI tools have the potential to improve some aspects of DPC practice. Examples include identifying dangerous medication interactions, streamlining medication dispensing, and helping diagnose skin conditions through image analysis.
“Using DPC practices to test AI-powered technologies would identify technologies that sound good on paper but are problematic in practice, preventing a repeat of the electronic health records blunder,” Brekke said.
Whereas the application of AI in traditional clinical settings has the potential to enhance efficiency and streamline administrative processes such as billing and coding, Holdsworth said he doesn’t think that AI alone will truly transform healthcare.
“If you think about advances in technology over the last three or four or five decades, none of the advances we’ve seen have actually helped transform the system. They’ve all essentially reinforced the existing incentives in the system,” Holdsworth said.
He predicts that AI, too, could preserve the status quo, spurring further disintermediation of the doctor-patient relationship and accelerating an arms race between insurance companies and healthcare providers.
“Insurance companies are going to leverage AI to make it easier to deny claims [and] on the provider side, it’ll be used as a way of making it easier to submit claims,” Holdsworth said. “It’s going to make the existing system more efficient at what it’s doing well—which is not working right.”
On the other hand, in a DPC context where incentives are not in opposition between insurers and providers, Holdsworth sees the opportunity for AI to optimize patient-doctor connections and communications.
“We can really focus on what enhances the therapeutic relationship, what is really a patient- or doctor-centric capability as opposed to a healthcare system capability,” he said.
For AI tools to take off within the DPC community, they’re likely going to have to fulfill that promise, according to Brekke.
“One advantage of DPC practices is that they are independent of third-party payers; as such, they will only adopt technologies that improve patient care or business operations,” she said.
Brekke warns that AI tools may not be as valuable in DPC practices as in other practices because some of the hassles and barriers that exist in the traditional healthcare delivery system are not factors in DPC.
“DPC affords time for patients and avoids documentation and administrative tasks that don’t benefit patients,” she said. “Why would they give over relationship- and trust-building aspects of care to AI?”
Holdsworth would agree.
“AI can actually reduce administrative work for doctors and free up their time to focus more on the patient,” he said. “This strengthens the doctor-patient relationship, ensuring a more personalized approach to care.”
AI tools can also strengthen communication between patients and doctors by helping translate medical jargon or complex information into plain English that patients can easily understand.
“This idea that I need to use AI to supplement the brokenness of the system because patients can’t get served in the way they want to get served—in DPC, that’s not really a problem,” Holdsworth said.
The other group that will need to be won over: patients. That may take some doing. A July poll showed that 70% of American adults reported feeling concerned about the use of AI in healthcare.