Hacker News Reader

Comment by seehafer | Hacker News Reader

What's notable about the meat of the complaint sent to the FDA is that it's less "this technology actually doesn't work" and more "this company has no culture of experimental integrity".

From the article:

"hadn’t fully assembled the proprietary machines used for the herpes study when the experiments began" ... "company underreported the rate at which the machines broke down during the study" ... "employee also alleged that some crucial parts of the devices ... were modified to improve their accuracy"

FDA is generally skeptical of validation data not generated from "production equivalent" technology, but they get even more skeptical if you tell them something is X and it turns out to be Y down the road. That (rightly) sends them on a fishing expedition.

View on HN

You only fudge the results if the technology doesn't work. The FDA does not look kindly on such shenanigans.

Not necessarily to the former. Definitely to the latter.

To elaborate point 1:

- If the crucial parts of the devices that "were modified to improve accuracy" were modified early on in the study the data could be representative of the final product.

- The same applies for the devices not being fully assembled. If they weren't fully assembled, and the data were still good enough to defend a 510k, then the data should represent worst case usage, unless full assembly would somehow make the technology less performant.

Yes it is always possible their one test works, but reading between the lines it appears that Edison is flaky and hard to get to work consistently. Things like changing the pipet tips halfway through are amateur hour - issues like this should have been picked up in R not D.

really sounds if they went to certification's stage too early, regulatory bodies do not tend to support rapid iteration of products

> regulatory bodies do not tend to support rapid iteration of products

I disagree. 21 CFR 820 is deliberately written such that it does not prescribe any one specific development process; AAMI TIR-45 for example aligns agile methodologies with medical device software development and is an FDA recognized consensus standard.

What regulatory bodies do demand is that one document what one will do, and then, actually do what one has documented. Sounds like Theranos had trouble with that.

Thank you for pointing this out. You'd be surprised how many people even within medical device think design control necessarily means some slow-as-hell waterfall development method.

The company is already 12 years old.

I fully understand that part, perhaps the FDA needs to adapt as well. Testing doesn't need to be static, there is room for cooperative certification and corrective actions to insure the test completes. With the rapid advances in some areas of technology you could run into a situation where you obsolete the technology you are working on as certification issues reveal improvements.

While the medical field certainly has additional concerns I do see a time where, with proper processes in place, that the final delivery may not exactly be what entered the certification phase.

But they seem to have pivoted several times already. The founder's original patent was for a wearable device that transmits readings to your doctor.

Reply on news.ycombinator.com