Doctronic's AI for prescription renewals gets OK from Utah, stirs some debate

When most states try to improve access to healthcare, they approach it as a workforce issue. Take Iowa Gov. Kim Reynolds, for example. Her state ranks 44^_th in patient-to-physician ratio. She responded last year by agreeing to fund hundreds of new medical residencies. The state’s medical society called the move a “game-changer.”

But the by-leaps-and-bounds advances in artificial intelligence (AI) have prompted complicated questions: What if the care-access problem can be addressed without recruiting more providers? Indeed, what if sometimes it’s precisely the humans in the healthcare loop that act as barriers to care?

Matt Pavelle, co-founder and co-CEO of Doctronic, notes that sometimes it can be difficult for patients to get — or attend — in-person medical appointments for routine care such as prescription renewals. According to the Centers for Disease Control and Prevention, medication nonadherence creates more than $100 billion in avoidable medical expenses each year.

“A person’s inability to get their prescription renewed is thought to account for about 30% of that,” explains Pavelle, “and it disproportionately affects poor, rural, and older Americans.”

Pavelle and his co-founder, Adam Oskowitz, M.D., Ph.D., a vascular surgeon at the University of California, San Francisco, wondered whether AI could help close that gap. Most prescriptions filled by pharmacies are renewals, and most renewals are low-risk transactions in which the patient is by definition experienced with the medication. In the age of breathtaking AI advancement, why not allow a well-trained AI algorithm to dispense such routine prescriptions?

Doctronic developed an AI-based product to do just that. In July 2025, the company released a study showing that its large language model-based AI system was able to come to the same diagnostic conclusions as human clinicians in four out of five urgent-care telehealth cases. In 99% of cases, the AI chose a treatment plan that aligned with the clinician’s plan. Those data convinced Pavelle and Oskowitz that their system could be a safe and reliable provider of prescription renewals. The data also impressed Utah’s Department of Commerce, which in January 2026 announced that it had partnered with Doctronic on a first-of-its-kind automated prescription renewal pilot program.

Pavelle tells Managed Healthcare Executive that the new program has a number of safeguards in place to ensure accuracy and a positive patient experience. Any close calls or edge cases are referred to licensed providers in Utah. The company also runs prescriptions through a database to ensure there are no potential drug interaction concerns. In addition to safety and accuracy metrics, the state is also keeping a close eye on the AI system’s impact on refill timeliness and costs.

Margaret Woolley Busse, M.A., MBA, the executive director of the Utah Department of Commerce, said in a news release that the arrangement represents a “vital balance between fostering innovation and ensuring consumer safety.”

It also positions Utah as a leader in AI at a time when many states see the technology as a potential driver of economic development. “By providing temporary regulatory relief that enables innovation and crafting agreements that facilitate the deployment of AI in healthcare, the state aims to foster innovation, enhance patient care, strengthen provider capacity, and reduce costs,” the news release said. Doctronic’s program went live in mid-December 2025. Pavelle says it’s proving to be popular with patients.

“We see incredibly strong early interest from patients who’ve experienced running out of medication and waiting days for a refill,” Pavelle says. Providers, he says, are “broadly impressed by the system.”

Generating debate

Still, the idea has sparked controversy. The Utah Academy of Family Physicians said it had “significant concerns” about the pilot program, arguing it represents an “apparent willingness to move forward with AI without the necessary guardrails and without the understanding that human involvement is still necessary to truly guarantee appropriate healthcare.”

In a news release, Robert Steinbrook, M.D., health research group director at Public Citizen, the consumer advocacy group, characterized the Utah program as “a dangerous first step toward more autonomous medical practice.”

N. Lance Downing, M.D., a practicing physician and clinical assistant professor at Stanford Medicine, does not see this early foray into AI prescriptions as a major concern — or, for that matter, as a major technological achievement.

“I don’t think this is necessarily a risky or even impressive task,” he tells Managed Healthcare Executive. Rather, he says, “It’s an interesting front door for more valuable applications of AI.”

He says patients who use this or similar platforms should be aware of the system’s safety backstops and how their data will be used. They should also understand the company’s business model. It matters, he says, who the client is.

“When an AI refill platform is also selling to payers, we should ask whether the primary optimization target is patient well-being, adherence and safety — or pharmacy spend,” he says.

Limits and logistics

But the AI product could also be an attractive cost-saver for other links in the healthcare chain. For instance, Pavelle says the company is talking to health systems about ways the software can reduce “low-value” administrative work for physicians. Expansion to other states is also “in the works,” he says.

Expansion might also come in the form of an expanded prescription-writing portfolio for Doctronic. At present, 190 drugs are eligible for refills in Utah, Pavelle says. Some categories of drugs, such as controlled substances, attention-deficit/hyperactivity disorder therapies, and injectable drugs, are excluded. But Pavelle says there are other types of prescriptions that are low-risk and protocol-driven besides renewals. He sees medications for conditions such as strep throat, recurrent urinary tract infections and stable chronic conditions that could be prescribed using AI in the not-too-distant future.

“Two years out, we’d hope AI handles more routine work that doesn’t require physician judgment, freeing physicians for cases that do,” he says.

Finding a place for physicians

Yet, finding ways to “free up” physicians without excluding them from the healthcare “loop” can be a balancing act. For instance, a physician whose original prescription is renewed through Doctronic may not know that their patient received a renewal. The job of communicating with primary care providers (PCPs) is left to patients, Pavelle says.

“We encourage patients to maintain their relationships with their PCPs,” he says. “We’re preventing gaps in care when the existing system is slow, not replacing the relationship.”

Although PCPs do not get alerted to refills, Pavelle says, the company does send its refill data to Surescripts, so the patient’s medication history is updated and available to any electronic health record connected to Surescripts’ database. The software is also designed to promote in-person care as needed. Cases are automatically escalated to a licensed Utah physician if there are signs that the patient is not getting appropriate follow-up care. There are also limits on how many prescription renewals a patient can consecutively fill through Doctronic to ensure they see their human physicians periodically.

Utah Republican State Sen. Kirk Cullimore, who sponsored the legislation creating Utah’s “regulatory mitigation” program to help court AI companies, said the mix of safeguards and technology “reinforces the principle of ‘doctor, not device,’ ensuring that automation supports, rather than replaces, human judgment.”

Some fear, though, that the line between doctor and device may become too blurred. Doctronic’s homepage uses first-person language to announce, “I’m your private and personal AI doctor.”

Steinbrook said such language is unhelpful. “AI is a software application, not a licensed physician or other medical professional,” he said in the news release. “There is no such thing as an ‘AI doctor’ or ‘medical-grade AI.’”

Yet, advancements in artificial intelligence (AI) have prompted complicated questions: What if the care-access problem can be addressed without recruiting more providers? Indeed, what if sometimes it’s precisely the humans in the healthcare loop that act as barriers to care?