The next prescribing model must close the gap between a script and sustained care
Doctronic, Amazon Pharmacy and Omada leaders described prescription renewals as an entry point to care that continues between visits.
Doctronic co-founder and CEO Dr. Adam Oskowitz said its Utah prescription-renewal program is beginning to show patients returning for refills earlier and remaining on medicines longer. But the larger ambition described alongside Amazon Pharmacy and Omada Health was to make the prescription an opening to continuous care, rather than the point at which the healthcare system considers its work complete.
Oskowitz characterized the adherence observations as internal data. He did not present a controlled outcomes study or quantify the improvement during the discussion. The distinction leaves an important question for the next phase of AI-enabled care: whether easier renewals translate into sustained treatment and better health, beyond a faster transaction.
The November 17 New Prescribers panel brought Oskowitz together with Tanvi Patel, vice president and general manager of Amazon Pharmacy, and Wei-Li Shao, president of Omada Health. Lydia Ramsey Pflanzer, deputy editor at Endpoints News, moderated the conversation, pressing the executives on access, payment and what actually changes when an encounter moves from an office to a digital platform.
A renewal program has boundaries
Oskowitz described prescription renewals as a practical place to start because lapses in access can interrupt an existing treatment. He said users were coming from rural areas and Salt Lake City, and included older adults rather than only the younger population the company initially expected. His account challenged the assumption that demand for digital access necessarily follows a narrow demographic profile.
“There are care deserts everywhere,” Oskowitz said. He argued that the common characteristic is an unmet need, whether the person lives far from a clinician or cannot obtain timely care in an urban area. That explanation makes continuity, not geography alone, the relevant problem for a renewal service.
He described the system’s escalation design as a safeguard when information is ambiguous or incomplete. “The AI doesn't say yes or no to a prescription,” he said, explaining that the alternative to proceeding is referral to a human doctor. He framed broader follow-up and additional clinical uses as future opportunities, rather than outcomes already established by the renewal program.
The state’s authorized-pilot description sets a narrower regulatory boundary than the phrase AI prescribing might suggest. It covers renewals of previously prescribed medications from a fixed formulary, excludes controlled or addictive substances and treatment-plan changes, and says the pilot remains in a phase requiring licensed-practitioner authorization of every request. Oskowitz’s interest in new prescriptions or dose titration therefore describes possible expansion, not existing general permission.
The pharmacist still checks the prescription
Patel said the source of a prescription does not replace Amazon Pharmacy’s dispensing checks. A prescription enters the pharmacy’s verification process and is reviewed by a pharmacist. More routes into care can expand access, but they still have to connect to a reliable process for getting the right medicine to the patient.
“That's not where the care journey ends,” Patel said of the prescribing encounter. She emphasized that the medicine must actually arrive, be taken and remain available over time. A convenient consultation can accomplish little if the next step becomes a search for stock, an unexpected price or a pharmacy the patient cannot reach.
Her most pointed questions concerned information patients often receive too late. “Why do I not know how much it's going to cost me out of pocket on my benefit before the prescription is even sent by my doctor?” Patel asked. She argued for giving patients and prescribers that information during the treatment discussion, when they can still make an informed choice together.
Patel also described people checking availability before directing a prescription to Amazon Pharmacy, as well as caregiver access that allows someone to manage another person’s medicines. Those features address different sources of friction: uncertainty about supply, limited mobility and the time required of family members. She did not argue that a digital pharmacy should replace every existing way of obtaining care.
Moving the visit online does not complete the redesign
Shao challenged the idea that reproducing a traditional encounter through telehealth is enough to change chronic-disease outcomes. The treatment plan may be created during a short visit, but the work of following it happens in everyday life. His definition of a new prescriber therefore extended responsibility beyond diagnosis and the initial script.
“There’s a big difference between the treatment plan and then the follow through with the treatment plan,” Shao said. He described care between visits as an opportunity to help a person interpret what is happening, act on a recommendation and return to the next clinical encounter with a more durable result.
That does not require every interaction to be handled by the same kind of worker. Shao described AI support for understanding meals, tracking nutrition and providing immediate information, alongside human care teams. The design question, in his account, is when a person needs more human attention and when software can remove work that should not require a clinician’s time.
“If there's a dialogue that needs to be there, that requires empathy, care, and understanding to determine the right next step, that's probably an area for the human,” Shao said. His argument tied staffing decisions to the needs of the interaction, rather than treating human and automated care as competing packages from which an organization must choose only one.
The payment question follows the patient home
Ramsey Pflanzer pushed the panel beyond the familiar promise of more touchpoints by asking who pays for them. Oskowitz argued that AI could make some interactions less expensive, while acknowledging that reimbursement and adoption still require a viable way to finance the technology. His expectation of lower-cost care was a business proposition, not evidence that the payment system has already changed.
Patel connected the same issue to delivery logistics and around-the-clock pharmacist access. Technology can remove administrative burdens elsewhere in the business, she argued, helping support services patients need. Shao added that information has value only when patients can understand why a recommendation matters and connect it to their own behavior.
The three accounts converge on a more demanding test than the number of prescriptions produced. A renewal service must prevent avoidable interruptions; a pharmacy must make price, supply and delivery usable; and a care platform must help people act between encounters. The panel’s strongest proposal was to connect those responsibilities. Demonstrating the resulting health outcomes remains the work ahead.