Adam Rosendorff, a former director of Theranos' lab, was the second employee to testify in Elizabeth Holmes trial. He kept his work email accounts. Rosendorff signed a non-disclosure contract when he joined Theranos. Forwarding work emails can be a violation of that agreement. He was concerned that he would be held responsible for the company's problems, just like Surekha Gangakhedkar.
He was right to be concerned: he is one of the people Elizabeth Holmes Defense is trying to blame.
The company was more concerned with PR and fundraising than it was about patient care.
The defense admitted that there were problems at Theranos' lab in opening arguments in US v Elizabeth Holmes. Homes' attorney Lance Wade stated that the laboratory director was ultimately responsible for problems in Theranos lab. Sunny Balwani is also being tried separately as a co-defendant. The defendants are facing 10 counts of wire fraud as well as two counts each of conspiracy to commit wirefraud.
Today's testimony focused on bad Theranos testing. Numerous emails were added, proving Holmes was well aware of Theranos' problems and actively trying to resolve them. Rosendorff attempted to convince Theranos that FDA-approved tests be used instead of those developed by Theranos. Perhaps even more telling was the email that Rosendorff was exempted from.
Rosendorff was not shy in his statements when he testified. Rosendorff stated that he quit because of several reasons. He said that one reason was the inability of management to conduct proficiency testing as required by law. I felt pressured into vouching for tests I didn't believe in. He said that he came to believe that the company was more concerned with PR and fundraising than patient care.
Rosendorff believed that Theranos would be the next Apple when he joined Theranos in March 2013. The company's sheen quickly wore Rosendorff out and he started looking for other jobs that summer. He said that August was the time when Theranos was launched in patients. The clinical lab was nervous about testing real people. The launch pace was extremely fast and hurried.
Raising alarm bells
Elizabeth Holmes sent an email at 1AM on August 31 asking for information about the status of validation tests. One employee replied that none had. Rosendorff stated that the launch was set for September 9, 2013, and that it had been scheduled by Holmes and Balwani.
He tried to stop it, growing more worried. Rosendorff was becoming increasingly worried and tried to stall the launch. He sent Holmes and Balwani an email about a week before the launch. He felt that the lab needed more people with better training.
In an email, he stated that he needed to take a few weeks to resolve these logistical and medical issues. This would delay Theranos testing for patients. Holmes even met him to discuss his concerns. He felt that Balwani was not taking him seriously.
Rosendorff described Holmes' nervousness during that meeting as a sign she was not her usual calm self. He said that she was slightly trembling.
The launch went as planned, despite Rosendorff's concerns.
Theranos devices failed so often that I began to doubt their accuracy.
His testimony corroborated some of the statements made by previous Theranos employees. Erika Cheung's concerns about the elimination of outliers in Theranos data were echoed by him. Both Cheung and Gangakhedkar affirmed that the Edison machines often failed quality control.
While a quarter of Theranos devices routinely failed quality control, failures of commercially-available tech was far less frequent. Rosendorff stated that Theranos devices were so often failing quality control tests, it raised doubts in mine about their accuracy. Balwani, who replied to him, wrote back that he had emailed Balwani regarding the failure rates. However, this is not true.
Lab directors normally deal with doctors who are concerned about laboratory results. This was Rosendorff's previous job at the University of Pittsburgh. Theranos appointed Christian Holmes, the brother of Holmes, to be a liaison for any complaints. Rosendorff claimed that he was forced to give explanations for bad results, even though they didn't question the tests.
The tests turned out to be bad. Rosendorff sent an email in May 2014 to his staff stating that any future tests would be done on FDA-approved devices.
Complete chaos
This did not happen.
Christian Holmes actually wrote to his sister, CCing Balwani in June 2014, complaining that the lab was creating problems.
Holmes replied that she had shed connect with her brother and then added Sunny / This is already taken care of.
The exchange has a few noteworthy aspects. It discredits the notion that Holmes might have been misled regarding the state of the clinical laboratory and its tests. It also shows that she clearly took control of the situation and even told Balwani to step down. It's remarkable, however, for the third person missing from this chain: Rosendorff. He may not have known that he only instructed FDA-cleared devices.
He said that he first saw the emails in court. Rosendorff was not included in the other emails containing patient complaints.
This test no longer has any diagnostic value.
Rosendorff tried to persuade Theranos to discontinue its HDL (good cholesterol) test in February 2014. Balwani was not listening to his email, so he sent him an email and included Holmes. Rosendorff stated that he received a lot of criticism for suggesting that the lab use approved devices.
He testified that Holmes, Balwani and VP Daniel Young were the ones who pushed back. Rosendorff stated that Young did not have a medical degree but offered many suggestions on what could be wrong with the tests. Rosendorff's testimony generally spoke about Young in a tone suggesting that he hated the VP.
Rosendorff stated that there were many problems at the lab. He cited tests of bicarbonate and an email in which he said that two-thirds were below the normal range. He emailed September 2014 to say that the test had lost its diagnostic value.
These lab results are confusing. What should I do with these lab results?
Theranos was not honest with providers or physicians about this. Instead, customer service representatives were told to tell providers and physicians that results weren't reported because of temporary inaccessibility.
Another email was sent by a healthcare provider claiming that it did not know Theranos' methods were not FDA-approved. It also complained about blood sugar and testosterone. These lab results are not what I want. The email was read.
Rosendorff stated that part of the problem was the fact that all validation studies were done with venous blood. To get enough fingerpricks blood to run all the studies you would need to prick every patient's finger. Venous blood was then drawn as an option. However, if you draw blood from fingerpricks it is much more likely that red blood cells are destroyed, which can lead to inaccurate readings.
One of the poor tests was called ISE, which is for sodium, potassium and chloride. Young, a statistician and not a doctor, said that the test's faulty results were due to red blood cells breaking down. He recommended that the blood be examined visually to confirm.
Since September 2013, this bad assay has been used to test patients. The tests were still causing problems in June 2014.
This was problematic for many reasons. This was first because FDA-approved machines could detect it immediately and not after strange readings. Rosendorff also testified that evidence of red blood cells being damaged wasn't always apparent to the eye.
Since September 2013, this bad assay was being used on patients. Youngs recommended fixes for the problems in June 2014. Theranos canceled tests that were below and above accepted values.
This was also a problem. We received an email from a doctor on October 27, 2014, which was addressed to a customer support rep. A test was ordered because the patient had low sodium levels. Rosendorff was contacted by customer service, who asked Rosendorff if that was the value. Will it be voided if she returns and the value is not reduced?
Rosendorff sent this email to Balwani, Holmes, and wrote, "Im not sure of the clinical value of a salt assay in which it is not critically important and the exact situations that require reporting abnormal sodium results are voided."
It is obvious that we have problems with calcium, potassium and sodium.
Too bad was the calcium testing. Christian Holmes wrote to Young & Balwani asking for information about the high rates of calcium reporting. This seems to be the most common question we receive about accuracy.
This was then sent to his sister. He wrote that it was obvious that we had issues with sodium, potassium, calcium. Although I cannot tell them that we are wrong, they continue to send patients on to Quest [another company for blood-testing] when we report high levels. The results still come back as normal.
Rosendorff was not included in this chain and raised concerns about it.
Rosendorff testified that he didn't believe Theranos could solve these issues at this stage. He stopped speaking with doctors and told a few physicians that he did not trust the results.
Holmes is not only aware of the issues, but she is also the one who is responsible for them. Her defense did not mention what Rosendorff's cross-examination might bring. Federal authorities also investigated one of his other employers, and another company was under investigation by the justice department. This was before the jury could enter the courtroom.
The jury might not hear anything about Rosendorff's employment after Theranos. The defense will have a lot of work ahead. Holmes was probably more affected by Rosendorff's testimony than James Mattis, who was technical in nature. She will have to provide ample evidence to support her claim that fraud was mostly Balwanis' fault.
