"It was totally clear she knew her products could not do x, y and z. She facilitated and approved and enabled sales of these products that could not do x, y and z, to providers who were going to use them to perform x, y and z tests on patients, tests that they could not do."
How do you see that she didn't defraud patients?
> Though they initially were divided, jurors soon agreed to acquit Ms. Holmes on the one count of conspiracy to commit wire fraud and three counts of wire fraud tied to patients, Ms. Stefanek said. She said the jury concluded that prosecutors didn’t present enough evidence to show that Ms. Holmes knowingly pitched a faulty product to induce patients to pay for tests.
> The jury was in agreement, Ms. Stefanek said, that the Theranos lab was run shoddily and not in the way to be expected from a company setting out to achieve the highest quality in medical care. But jurors thought the burden of proof on defrauding patients was higher.
> “If all we’d had to prove was that she knew there might be problems in the lab and that might end up harming patients, that would be one thing,” Ms. Stefanek said.
https://www.wsj.com/articles/jury-in-elizabeth-holmes-trial-...
To me it's direct correlation.
"If she knew there might be problems in the lab" - the only problem in the lab was the machines themselves! The blood work is known science - the issue was getting to a point where accurate blood work could be done. Which involved the machines, the machines that they separately demonstrated she knew did not and could not do 'x, y and z'. And to be clear, in your example, x, y and z are "extract sufficient blood for an accurate test", "extract blood in a way that allows for an accurate test (i.e. absent over-dilution, contamination)".
To also be clear, this can be entirely a failing of the prosecution too - but it seems there's this sense of "there's this missing -glue- that we couldn't prove", but the lab IS the machine, and if she knows the machine can't do x, y, z, then it follows that the lab can't do x, y and z. There are only two outcomes from the machine, and lab, an accurate result, or an inaccurate result (or inability to provide any result). And then if you say "My machine cannot do x, y and z", then it cannot provide accurate results (at least not by design, stopped clocks and all that). Then you get to a conclusion that her argument would be that "an inaccurate result doesn't inherently harm a patient", and convoluted logical leaps of "because the result would lead to further testing and interventions", but there was presented at trial concrete evidence that unnecessary, and harmful, interventions were ordered/performed, on the basis of the lab results from those machines.