“I know developers who sat in my office and told me they didn’t build a life-saving feature because they didn’t know if it would be regulated or not,” said Morgan Reed, president of an app developers’ interest group called ACT.
Three documents from the FDA remove some doubts by describing various types of medical software and what sorts of claims would be regulated.
Reed said it’s now clear that if a doctor makes the decision and diagnosis, and uses software to make more informed decisions about treatment, that software would not be regulated.
It would be regulated, however, if the software makes a diagnosis. For instance, if it notifies a user that they have a specific medical condition.
Reed said this category of software is so important as it takes some of the guesswork out of medicine.
Previously, doctors made decisions based on patients they’ve seen with a similar condition and background. “If you’re lucky, that’s 500 people who look like you,” said Reed. But with new technology from “thousands of doctors on millions of patients,” he said, treatments will be more evidence-based.
Apple, Google, Fitbit and others are all working with FDA as it figures out how to fast-track the approval process for digital health companies.
Some people in the industry think the guidance does not go far enough.
Some clinical decision support software is more “risky” to a patient than others. For instance, some app makers will recommend chemotherapy to a cancer patient based on information in their database. Others will merely use data to predict a risk score for populations of people who are more likely to develop migraines.
Bradley Merrill Thompson, an FDA expert and lawyer with the firm Epstein, Becker & Green, would have liked to see FDA distinguish between these various use-cases based on risk.
“We didn’t get that,” he wrote in an email. “Worse, it appears based on the guidance that the FDA is not interested in drawing that line.”
