The review in the link is not so impressive. I quote;
The Theranos numbers
have two problems: They deviate significantly from the Stanford numbers and, more important, they deviate from themselves over a 24 hour period. Glucose concentration can vary hourly, not so platelets or hematocrit.
Then he quotes the Stanford numbers;
Stanford Hematology disagrees:
— Platelets: 320, no concerns, come back in two or three months
— Hematocrit: 41.1, ditto
And then at Stanford, a day after that:
— Platelets: 297, a 7% drop
— Hematocrit: 41.7, +1.5%, up a bit
, but not alarming
So it is claimed to be suspicious when the Theranos numbers vary over 24 hours but it is OK when the Stanford numbers vary over 24 hours
I agree "deviate significantly from the Stanford numbers" does not look good but I do not know how much variation in blood test results between different labs is normal?