The healthcare industry faces a worldwide staffing crisis that leaves many patients underserved. Munjal Shah, CEO of Hippocratic AI, believes artificial intelligence can help bridge the gap. At the recent 2023 HLTH conference, Shah shared his vision for using generative AI for “super-staffing” healthcare roles beyond diagnosis.
The annual HLTH event in Las Vegas convenes leaders across healthcare to discuss innovation and technology. This year, much of the focus was on the potential applications of large language models (LLMs) like ChatGPT. Shah took part in a panel titled “There’s No ‘AI’ in Team” to offer his perspective.
Hippocratic AI was founded to leverage LLMs to alleviate shortages in nursing, care coordination, patient education, and other non-clinical functions. Shah said, “Right now, healthcare faces a massive worldwide staffing crisis. The World Health Organization is projecting a 10 million health workers shortfall by 2030.” He sees AI as the solution to meet demand and expand access.
On the panel, Shah acknowledged concerns about utilizing LLMs for diagnostics. Misinformation could put patients at risk. However, he argued tasks like chronic care nursing, scheduling, and explaining treatment plans are ripe for AI support. This is where “super-staffing” through generative AI can make an impact.
The premise is to combine human expertise with machine learning, termed a “centaur” approach. Hippocratic AI has hired thousands of medical professionals to train its AI. Through reinforcement learning with human feedback, the LLM mirrors evidence-based responses. The goal is to instill human-like conversational ability.
According to Shah, a trained LLM could provide services at a fraction of human costs. For example, chronic care nursing may cost $100 per hour with a human. The AI equivalent could be $1 per hour. This exponential scale allows for covering more patients.
Shah explained, “You can’t call every patient two days after they start every new medication. But at this cost structure, maybe you can.” The concept is using AI to augment human capabilities, not replace them. Together, humans and AI can reach underserved populations.
Some use cases Hippocratic AI explores include explaining billing, genetic counseling, pre- and post-op instructions, and delivering test results. A recent study found participants preferred AI responses over human doctors for empathy.
Shah sees conversational ability as a critical strength of generative AI for healthcare: “It means we finally have the technology for patient-facing conversations.” The AI can also synthesize information from diverse sources in a human-relatable way.
In Shah’s view, non-clinical functions represent low-hanging fruit for generative AI in healthcare. He aims to evolve Hippocratic AI into an integrated platform supporting the full spectrum of virtual care. Though still early, Shah is optimistic about the potential to increase healthcare access through human-AI collaboration.