A prescription refill program launched in Utah earlier this year is igniting a significant debate about the readiness of artificial intelligence to handle tasks traditionally performed by doctors. The program, which allows Utah residents to obtain prescription refills online via an AI chatbot named Doctronic, aims to increase convenience for both patients and prescribers.
However, the initiative has also raised alarms among medical professionals, legal experts, and public health officials. It highlights critical questions regarding the regulation of AI in medicine, the necessity of physician oversight, and the implementation of patient safety measures. A central point of contention is the existing state and federal laws that restrict prescription authority to licensed medical professionals. Proponents argue that these laws, foundational to American medicine for over a century, should be updated to accommodate AI technologies like chatbots.
"We have crossed a threshold in terms of giving something that is not human a medical license, whether or not we want to call it that," said Dr. Eric Bressman of the University of Pennsylvania. Experts like Bressman are not inherently opposed to AI in prescribing but emphasize that AI systems should meet rigorous standards comparable to those of human doctors, who undergo extensive training and testing before licensure.
Doctronic's pilot program in Utah operates within a "regulatory sandbox," a framework that permits state officials to grant waivers for AI companies developing promising technologies. The program is currently supervised by a five-member board of AI specialists, none of whom are physicians. The board asserts that numerous safeguards are in place, including a period where human doctors review all Doctronic refill orders. The company anticipates transitioning to fully automated refills in the near future.
Concerns have also been voiced by medical professionals within Utah. The head of the state’s medical licensing board, Dr. Alan Smith, stated that he and his colleagues were informed of the program's launch through news reports. In a letter to the state in March, 11 board members requested the program's suspension, citing risks associated with automatically renewing medications that could have side effects or dangerous drug interactions. Smith, speaking as an individual physician, noted that patient conditions can change, making previously prescribed medication inappropriate.
The regulation of medical technology is often fragmented, with federal agencies like the Food and Drug Administration (FDA) overseeing technology and states regulating medical professionals. Doctronic executives view their AI as part of the state-regulated medical practice, but some experts believe the AI's direct impact on medical care may fall under FDA purview. When asked, Doctronic executives did not confirm if they had sought FDA approval, stating their focus is on patient access to healthcare.
In the Utah program, residents can use the Doctronic website to confirm their identity. The AI chatbot then collects information on prescriptions and medical history, verifying valid prescriptions through a national pharmacy database. Approved renewals are sent to local pharmacies. More complex requests are transferred to a doctor working for Doctronic's telehealth service.
Doctronic's co-founder, Dr. Adam Oskowitz, envisions the AI handling routine medical tasks, thereby enabling physicians to manage a larger patient volume. Several other states, including Texas and Wyoming, are also exploring waivers for AI technologies. Meanwhile, legislative proposals in states like Iowa and Idaho aim to establish formal licensing for AI medical services, often drawing from template legislation provided by the nonprofit Cicero Institute.
According to Adam Meier, director for health policy at the Cicero Institute, resistance to medical AI is partly driven by economic concerns within the healthcare workforce. The American Medical Association has echoed concerns about prescription renewals, emphasizing they are not simple administrative tasks.
Utah's AI office head, Zach Boyd, indicated that Doctronic has been conservative in its initial rollout, frequently escalating routine decisions to human doctors. In response to safety considerations, certain medications, including one for irregular heartbeats, have been removed from the list of eligible refills.
While Utah has released preliminary data and Doctronic plans to publish peer-reviewed studies, its current published research on the technology is a non-independently reviewed paper authored by company scientists. This study examined Doctronic's ability to diagnose medical conditions from telehealth consultations, finding its diagnoses matched human doctors 80% of the time.
Critics like Bressman argue that Utah should have required comprehensive data on prescription refills before the program's implementation, suggesting the state is currently trusting the company's assurances. Bressman likens the current approach to AI in medicine to the unregulated standards of the early 20th century before national benchmarks for training and licensing were established.
The FDA, while committed to encouraging medical innovation, has stated it has not authorized any AI chatbots for prescription refills and is focused on safety. Industry observers anticipate that companies like Doctronic may expand their services across states with varied regulatory landscapes. However, Daniel Aaron from the University of Utah's law school cautioned that such expansion risks eroding public trust and prompting backlash in the long term.