How AI could identify prescription mistakes before they happen
Add Axios as your preferred source to
see more of our stories on Google.
Illustration: Aïda Amer/Axios
A startup has developed a way to use AI to detect when doctors may be prescribing the wrong drug — or overprescribing opioids.
Why it matters: A system that could identify prescription mistakes before they happen could help save the thousands of Americans who die each year because of preventable medication errors, and it could contribute to controlling the opioid epidemic.
How it works: In much the same way that financial institutions use automated systems to catch outlier transactions that may signal fraud, MedAware's platform analyzes the prescription patterns of thousands of physicians to flag when a medication may be in conflict with the profile of the doctor, the patient or the medical institution.
- Gidi Stein, MedAware's CEO and a practicing physician in Israel, was inspired to start the company after hearing the story of a 9-year-old boy who died because his doctor accidentally selected the wrong drug on an electronic prescribing pull-down menu.
- "It wasn't bad judgment on the part of the physician — it was a typo," says Stein. "And you would have thought there was some kind of spellchecker to prevent this from happening."
By the numbers: Each year in the U.S. 7,000–9,000 people die due to a medication error of some kind, and the total cost of medication mistakes is more than $40 billion a year.
What to watch: Stein is particularly worried that the rapid adoption of telemedicine during the pandemic could open the door to more medication errors, as online doctors deal with patients they may know little about.
- "This is going to be the next phase of health care and that's perfectly OK," he says. "But we need to provide doctors the right tools so patients can be protected."
- With the opioid epidemic accelerating during the pandemic, Stein says platforms like MedAware can help doctors quickly identify patients who might be at a higher risk of abuse before they prescribe painkillers.
The bottom line: As in other fields, the sheer amount of health data is growing beyond the ability of humans to grasp alone, creating a need for automated systems that can save us from ourselves.
