FDA’S TAX ON MEDICAL DEVICE INNOVATION

Yesterday, FDA published its user fee schedule for fiscal year 2023, which just began this month.?Two things caught my eye.

1. The user fee for 510(k)s is just shy of $20,000.?
2. The user fee for a de novo submission is well over $130,000.?

Both fees can be reduced if the applicant can qualify as a small business.

That’s only part of what I’m calling a tax on innovation.?When a client asks me what’s the difference between the 510(k) and the de novo processes, I tell them very roughly they should expect the de novo process to:

• Take twice as much time
• Cost twice as much, and
• Be at least 10X more uncertain in its outcome

The basis for the greatly increased uncertainty is the fact that, as compared to the 510(k) process, there are few useful roadmaps in the form of guidance documents or prior FDA decisions to tell us what FDA will expect as evidence for a particular novel, new device.?Worse yet, in the absence of those roadmaps, the de novo process will be dominated by the personalities of the FDA reviewers, and an applicant should expect the composition of the reviewers to change over the time it takes to get a de novo decision.?I’ve read many minutes from de novo presubmission meetings where within the span of six months, the FDA staff meeting with the company changed greatly, and the issues of concern shifted radically with the new people.?

For those of you who like statistics, the MDUFA IV Performance Reports indicate that across the years, pretty consistently a little better than 96% of accepted 510(k)’s succeed, where the yearly success rate for accepted de novo’s bounces between about 45% and 62%.?So I’m not making the up “10X less certain.” ?A 4% failure rate versus a more than 40% failure rate is a 10X difference.

If you’re not familiar with the de novo process, it is specifically intended for innovative but medium or low risk medical devices.?If a new device is “substantially equivalent” to existing low or moderate risk devices, the 510(k) route is available.?If the new device is innovative and likely risky, the FDA requires premarket approval.?The de novo process occupies the middle ground of medium risk but innovative.?Most FDA consultants these days will tell you that FDA is trying to push more companies into de novo if they have anything truly innovative.

How does this play out??Here’s a typical day for me.?A group of researchers and entrepreneurs will contact me with some exciting new AI-driven technology that has wonderful healthcare applications.?The technology appears to me to offer great potential to save lives.?The client's first two questions to me will be, is our product regulated, and if so, what is the path through FDA in terms of time, cost and certainty?

Last week, FDA modified its guidance documents on software to greatly expand the scope of software it plans to regulate.?Even before that, because these startups typically are trying to improve human health through better care, my response will frequently be, yes your product is regulated, and given the novelty of the claims and functionality you presented, it will likely be in the de novo category.?

Upon learning about the de novo category, they will say that they, and the investors who back them, will simply not support an initiative that includes the cost, time and uncertainty of the de novo process.?So, we will then spend a few weeks very painfully paring back everything that’s too innovative – in both function and claims – so they can make the argument that they are only incrementally better than some product already on the market.?

From the data released yesterday on FDA AI clearances, it does not appear that my experience is unique.?One take away is that many AI entrepreneurs are trying to stay in the 510(k) zone rather than go the de novo route for truly innovative stuff.

And so I go to bed sick to my stomach.?On behalf of American patients, there has to be a better way.?We can’t tax substantially the innovation we so desperately need in healthcare.