Artificial Intelligence
Elizabeth Briand
The concept of artificial intelligence has been around for nearly 75 years, since mathematician Alan Turing asked the question, “Can computers think?” In the ensuing decades, as computers have become faster, stronger, and capable of once unimaginable feats, artificial intelligence also has evolved—and in recent years, that evolution feels like it has accelerated to breakneck speed.
In health care, the idea of using computers to assess and solve medical problems has been in place since the early 1970s, when researchers at the University of Pittsburgh created INTERNIST-1, a computer-assisted decision tree designed as an educational experiment to one day help patients in remote environments.
Putting AI into practice
Today, the use of AI has become more pervasive, especially as health care systems look to manage costs while improving patient care. The most common applications of AI in health care currently focus on cancer screening prediction, medical imaging and diagnosis, and clinical decision support systems, said Khoa Nguyen, PharmD, clinical assistant professor at the University of Florida College of Pharmacy.
“In cancer screening, AI helps with early detection and analysis of cancer patterns, while in medical imaging and diagnosis, it assists in analyzing various medical scans like X-rays, MRIs, and CT scans to identify abnormalities,” Nguyen said. “The third major application involves clinical decision support systems, which provide tools for risk assessment and prediction, helping health care providers make informed decisions by analyzing patient data and suggesting appropriate treatment options.”
In pharmacy, AI can be applied to two main areas of need: prevention and treatment. On the prevention side, it can predict medication effectiveness as well as potential adverse effects. “AI also addresses one of the pharmacy field’s biggest challenges—medical adherence—by helping to identify nonadherence risks and suggesting ways to improve compliance,” said Nguyen.
Nguyen noted that AI can be valuable in predicting treatment efficacy—citing psychiatric care as an example—where determining medication effectiveness can take up to a month. “AI can streamline this process by better predicting which medications will work best for specific patients, potentially reducing the time and uncertainty involved in finding the right treatment regimen,” he said.
As good as advertised?
The potential of AI has health care providers across the country and around the world leaping at the chance to incorporate it into their care models. A November 2024 article published in Health Management Policy & Innovation titled “Why AI Is Good for Our Health but May Hurt Our Wallets” suggested that while AI may have a multitude of clinical benefits, it also may end up costing health systems more than it is saving. The article, written by a team of Stanford University School of Medicine researchers, noted, “Current approaches to evaluating, regulating, and paying for AI in health care incentivizes use of AI in a manner that is likely to increase total cost of care.”
The problem, the researchers suggest, stems from “disconnected regulation and reimbursement approval decisions, anemic health information technology budgets, and complex revenue structures across stakeholders,” all resulting in the potential need to get creative with business models to subsidize AI tools.
“The impact of AI on time and resource savings in health care is still difficult to measure as we are in the very early stages of its adoption,” said Nguyen. “Unlike the rapid integration of AI in everyday consumer applications, its implementation in health care settings has been considerably slower.”
It will take more time to determine whether AI is living up to its lofty expectations. “The cautious and gradual integration reflects the complex nature of health care systems and the need for thorough validation before widespread implementation,” said Nguyen.
Moving forward with innovation
There is no doubt that AI is poised to become a ubiquitous component of health care. To maximize its effectiveness, though, it may be best to look before leaping, ensuring that resources are in place to support the program, that it has been well vetted, and that it is the right fit for the role it needs to fill.
“Just as current technology has revolutionized access to information and improved operational efficiency and safety, AI is expected to further enhance these aspects,” Nguyen said. “The technology will continue to evolve and advance, ultimately improving quality of life for both health care providers and patients by streamlining processes, enhancing decision-making capabilities, and providing more sophisticated tools for health care delivery.” ■