Almost 65% of adults in the United States rely on prescription medication, and they are at the mercy of a surprisingly convoluted system to obtain them as needed. These 131 million patients must consult with a doctor at intervals ranging from monthly to annually, after which the physician must submit an order to a pharmacy, which dispenses the medication to the patient if its stock is adequate.
Depending on location and medication, this often occurs smoothly. However, this is not the case for numerous patients, and 27% of pharmacy customers surveyed by health tech company DrFirst have run out of medication waiting for a refill.
In the quest for a solution, Utah is using its artificial intelligence (AI) regulatory sandbox to pilot AI-facilitated refills using Doctronic technology. Dedicated regulators will be tasked with identifying and correcting safety pitfalls in the proposed AI program, then determining if the improved product should be enacted on a larger scale.
Current Dangers
Before assessing potential pitfalls in any novel technology, it’s important to look at the hazards in the system it is designed to replace, so a fair comparison will be possible.
Lack of timely access to medication is an obvious danger. As noted, more than a fourth of patients have experienced temporary lack of access, contributing to a medication non-compliance crisis that costs more than $100 billion annually in avoidable medical expenses.
In each of the steps to get the product to the customer, delays often occur. The average wait for a doctor’s appointment in a large city is 31 days, but it frequently stretches to many times that length.
After the appointment, many doctors bundle prescription call-ins, waiting to contact pharmacies until they have a significant number to request. Physicians also commonly make errors in prescribing, with 36% of patient prescription charts showing evidence of oversights. This causes further delays if the fault is caught; it can cause far worse consequences if it is not. An estimated 9,000 patients die from prescription errors every year.
Finally, a pharmacy needs the item in stock, and it is worth noting that automated inventory tracking and reordering technology is why this part of the prescribing chain is already quite efficient.
Risk Mitigation In New Project
Despite the worrisome current state of affairs, the pilot method must obviously be held to rigorous safety standards. The Utah initiative combines two complementary safety nets: a limit to a short list of 191 relatively benign meds, and a restriction to refills only. No controlled substance or similarly controversial medicine is eligible, and no patient will receive any drug for the first time through AI.
Even with these restrictions, a physician will oversee the first 250 refills, and Doctronic will test samples to ensure accuracy.
Advantages
Under these strict parameters, AI-generated refills will allow patients to skip the potential months-long deferral to see a physician.
Prescription refill appointments are typically brief, involve surface-level question lists (which are themselves often automated and not immediately read by the physician), and are frequently done online without seeing or speaking to a physician. Delaying a refill to fulfill the legal technicality for such an “appointment” is almost all risk, no gain.
Restructuring this process benefits patients, doctors, and pharmacists.
A Logical Step
Automated refills, particularly of the low-risk type Doctronic is testing in Utah, offer immense potential to save not just time and hassle, but also lives. Human error and inefficiency have long caused bottlenecks and worse in the prescription pipeline, and the system is begging for technological support.
As sponsoring Sen. Kirk Cullimore (R-UT) emphasized, the principle of “doctor, not device” still needs to reign in medicine. However, we refuse assistance from appropriate technology at our peril.