Optimal Regulation of Medical Devices
There’s no shortage of opinions about whether the FDA’s process of approving medical devices is too lax, too strict, or just right. After all, there’s a lot at stake, both for patients and for the $110 billion market for medical devices in the U.S. But evidence, one way or another, has been sorely lacking. A study by LDI Senior Fellows Matthew Grennan and Robert Town goes a long way toward filling that gap.
The investigators take advantage of differences between the European Union (EU) approval process and our own to quantify the tradeoffs between access and safety concerns. Basically, the US applies a “safe and effective” standard while the EU only certifies the safety of the product. Meeting the “effectiveness” standard often requires large-scale, costly clinical trials. Thus, many medical devices are approved for use in the EU before they are approved in the US (if at all).
In an analysis of the market for coronary stents, Grennan and Town conclude that the more rigorous standards in the US compare favorably with the looser standards of the European Union (EU), and that the current FDA process is close to optimal in terms of patient welfare. Considered in isolation, the EU would benefit from premarket clinical trials that last six months longer than current requirements. However, because the EU is able to “free-ride” off of the information generated in US trials, the current EU policy represents the optimal response to current US policy.
Town and Grennan used market quantity and price data from the EU and US to estimate the length of approval that would optimize patient welfare. In the absence of actual data on outcomes, they inferred quality by modeling how consumer choice behavior (as indicated by the device’s market share) changes as more becomes known about the product. Their model allows them to estimate the rate of “learning” in pre-market clinical trials and after the product enters the market. As “learning” increases, consumer uncertainty about risk is reduced, and market share is increased. As Town explains to Knowledge@Wharton (video clip follows),
[W]e compared the devices that were common in both markets, and how their market shares evolved over time. And finally, we compared that to devices that were only offered in the EU. There are typically many more devices available for any purpose in the EU, since that market is much easier to enter. We used that information to make inferences about the information that was being generated by the regulatory process, and then translate that into real consumer welfare.
Descriptively, the study finds that the EU has access to more, newer devices than does the US. Between 2004 and 2013, the EU had more than three times as many manufacturers and devices in the market. In an average month, 47% of the stents used in the EU were unavailable in the US. The average lag time for devices that eventually enter both markets is 10 months.
But that access comes along with greater risk and more uncertainty about quality, as measured by the average market share, which is lower earlier in the device’s life cycle in the EU than in the US. The inclusive market share for a device in the EU reaches a stable level after about two years on the market; in contrast, the inclusive market share is stable over the product lifetime in the US.
The study also provides information about how the FDA process could be improved, by estimating what might happen if the US embraced calls for more active “post-market surveillance.” Currently, there is little observational learning that occurs after a device enters the market. If, however, the rate of learning post-market approached the level of learning pre-market, the gains could be as high as 19% over current practices.
Grennan and Town caution that their results apply to stents only, a second-generation device, and may not apply to the greater informational needs required for the introduction of first generation devices. But their study adds to current public policy debates about how the FDA regulates drugs and devices. For more detail on these public policy issues, see the Wharton Public Policy Initiative’s Issue Brief on the study.