Health industry leaders said health data collection and the use of artificial intelligence (AI) could improve health outcomes at the 2024 Hawaii State of Reform Health Policy Conference.
Peter Lewis, vice president and chief information security officer at Hawaii Pacific Health, said AI technologies enable computers to receive information that can solve issues in a way similar to the way people form solutions.
“Many healthcare organizations have been using AI technologies for many years now,” Lewis said. “A lot of the leaders are challenged with [figuring out] how [to] adopt these very innovative technologies while at the same time leaning on the right side of compliance.”
Healthcare privacy and security data, including that covered by the Health Insurance Portability and Accountability Act, also applies to AI technologies. Lewis said healthcare organizations are creating governance structures around AI, which control how these technologies are adopted, as well as steps to ensure safeguards are in place for information technology systems. He also cited steering committees that review cases on the use of AI, and establish policies for what can and cannot be done with it.
Milton Cortez, director of AHARO Hawaii, is also using AI, and reviewing protocols for processing data. He noted that data could be useful in transferring patients to Molokai Health Center, which is 38 miles long and 16 miles wide—and is close to the shoreline.
“The administration in Molokai is already looking for a place to move because of the rising sea level,” Cortez said. “The population they care for is 2,386 patients.”
— Cortez
Cortez said the use of AI for collecting data and processing information could be useful for the organization’s more than 52,000 statewide patients in establishing actual data points and transferring patients to different facilities.
Trey Sutton, co-founder and CEO of Siftwell, said AI can be used to reduce administrative burden by allowing AI models to produce patient note summaries, or by allowing AI to draft messages to patients.
Sutton said that if a healthcare organization is launching a campaign to increase breast cancer screenings, for example, a closed-loop AI effort can show curves of information related to the campaign and monitor actions of patients.
“Generative AI algorithms … basically create new data based off data it’s been exposed to already,” Lewis said.
Generative AI algorithms include ChatGPT, and one of its concepts is called “artificial hallucination,” which refers to a chatbot generating seemingly realistic sensory experiences that do not correspond to any real-world input.’ Generative AI algorithms can receive enormous amounts of data, pool it together, and start producing information known as a hallucination.
“One of the advantages of using AI applications within EMRs (electronic medical records), for example, is that the EMR is exposed to literally hundreds of millions of patient data. Companies … can statistically validate the outcomes that are coming out of those models using vast amounts of data, and that’s hopefully, one of the ways where we sort of prevent that sort of creep before it actually occurs,” Lewis said.
Lewis said there’s no shortage of ideas when it comes to applying AI within the healthcare realm, and that the focus needs to be on how facilities incorporate AI technologies to understand what will provide the greatest impact, and what is likely to be successful.