But again, I think it's morally fine for an uncertified person to run a laboratory as long as they are running it correctly. But the Theranos labs were not being run correctly, and they were lying about the results that their machines could deliver.
You can build a medical product that has a higher sensitivity and/or specificity than the other products on the market without regulatory oversight. But the Edison had neither of those and they were pretending it did. That was the problem.
How do you determine if they're "running it correctly"?
The whole point of certification is to require a minimum set of skills/knowledge to do so.
Insert "faster" or "cheaper" as well, but that is my point. There are very clear ways to measure if your medical device is effective.
A lot of the regulations around laboratories exist because people have died or been injured as a result of lab screwups. Theranos hid entire labs from inspectors, let uncertified people run laboratories, and a lot more that wasn't fraud, but definitely violated regulatory requirements.