January 2012

There’s an App for That … And the FDA Is Figuring Out How to Regulate It
By David Yeager
Radiology Today
Vol. 13 No. 1 P. 6

This past September, the FDA held a public workshop to discuss its proposed approach to the oversight of mobile medical applications. The workshop allowed attendees to provide input regarding the agency’s draft guidance statement before it issues a final statement. Representatives from several physician specialties took part.

As mobile apps have filtered into the public consciousness, they have become increasingly popular among busy healthcare providers. However, it has become apparent that there is a gray area between how an app is marketed and how it’s used. Depending on its intended purpose, an app may be classified as a Class 1 device, which requires only general controls, according to the FDA; Class 2, which warrants some special attention from the FDA; or Class 3, which requires full premarket approval as a medical device. Class 1 and 2 devices require 510(k) clearance unless they are exempted by the FDA. Determining intent is the overriding challenge for the FDA and users alike.

“The big question, as far as the FDA is concerned, is where do you draw the line between where FDA’s regulatory oversight should come into play and where the guidance and procedure guidelines and technical standards of specialty societies like ACR come into play,” says Michael Peters, director of legislative and regulatory affairs for the government relations and economic policy departments at the ACR.

In the radiology realm, this schism is best illustrated by the development of mobile image viewers. For example, if an image viewer is being used for primary diagnosis, it requires 510(k) clearance. But an app that allows a clinician to view images through an EMR is considered a tool for secondary review, not primary diagnosis, and does not require 510(k) clearance. In the case of apps that are designed strictly as image viewers, some companies have sought 510(k) clearance to allow for primary diagnosis, while others have taken the approach of marketing their viewer as a secondary review tool. The potential problem is once the device becomes available, there is no way to control how it is used.

Image Quality Concerns
The FDA’s main concern is image quality. Aside from the significantly reduced screen size of smartphones and tablets, ambient light can vary dramatically depending on a user’s location. To deal with lighting issues, app developers invented a tap test that presents the user with a series of small targets of varying shades of gray that pop up in different locations on the screen. The app can determine if there is adequate visibility based on whether a user can find the targets and tap on them.

David S. Hirschorn, MD, director of radiology informatics in the department of radiology at Staten Island University Hospital who represented the ACR’s IT and Informatics committee at the workshop, says this method improves on the usual method of attaching photometers to a reading screen because a photometer measures only the screen’s capabilities, while the tap test accounts for the user’s visual acuity as well as ambient light. However, even if a smartphone or tablet can reliably present medical images, there are other potential distractions that could hinder a radiologist’s ability to render a diagnosis on a mobile device. For this reason, the FDA’s stance is that mobile apps should be used only when a radiology workstation is not available. Hirschorn says that’s how mobile devices are most likely to be used anyway.

“There’s no radiologist I know who wants to read 100 chest X-rays off of an iPad,” says Hirschorn. “It’s far less screen real estate than what we’re used to working with.”

Where the mobile app trail ultimately leads is anybody’s guess, but Hirschorn sees great potential in apps allowing broader access to subspecialty expertise. Rather than keeping a subspecialist on call, he says it’s much more cost-effective to “phone a friend.” He says mobile image viewers may also facilitate communication between radiologists and referrers by giving radiologists an idea of how the other half lives.

“If you start making images more available this way, through a mobile app that [radiologists are] going to use on a consultative basis, then they begin to see the world of radiology through the referring clinician’s eyes a bit more,” says Hirschorn. “And I think that’s a good thing.”

Decision Support
Another area that is likely to affect radiology is stand-alone clinical decision support (CDS), which was discussed on the second day of the FDA’s workshop. Stand-alone CDS is generally considered to be a tool that uses specific patient information to help a clinician make a diagnosis, but there isn’t a regulatory definition. The gray area here is that any app that constitutes a reference manual is not considered a stand-alone CDS tool, even if it contains formulas for helping clinicians figure out specific problems—as long as the app doesn’t make use of specific patient information. Peters says developers would prefer a comprehensive definition for stand-alone CDS. There is some debate about whether any current apps can be considered true stand-alone CDS tools.

Because the FDA tends to accept public comments after the deadline (which was October 19) when issuing guidance, it’s difficult to determine when the final version of the guidance statement will be issued. This approach gives the agency more latitude in deciding which public comments to address and fewer potential headaches if they decide to alter the guidance at a later date. Peters says the FDA’s goal is to let mobile app developers know what the agency’s thinking is on regulatory matters but also allow developers to find their own way to the marketplace. However, he expects the FDA to keep a close eye on developments in the medical mobile app market.

“I think that the draft guidance does a pretty good job of not stepping outside the boundaries of FDA’s normal oversight, but that’s one of those questions that we’re going to have to look at long and hard when we start looking at radiology mobile apps, specifically,” says Peters.

— David Yeager is a freelance writer and editor based in Royersford, Pennsylvania.