Over the past few days, jurors in Theranos founder Elizabeth Holmes’ criminal trial have been hearing about the now-infamous demos given to Theranos investors. The devices would make noise but not produce results, they ran apps that blocked error messages from appearing, and in some cases, instead of tests happening in front of visitors, blood samples were taken elsewhere for analysis.
Yesterday, the picture resolved itself further as testimony from Daniel Edlin, a product manager who was friends with Holmes’ brother Christian, continued. Edlin revealed that Holmes herself and Daniel Young, a Theranos scientist and vice president, played key roles in what ultimately transpired during demos for visitors and investors.
But first, Edlin was cross-examined by Holmes’ defense attorney Kevin Downey, who showed the court a series of emails from a US military officer who was part of the team evaluating Theranos devices for use on the battlefield in Africa. In a June 2012 email, Young wrote that he felt “very good about reliability for this deployment” after testing 100 protocols over the course of 48 hours in 110° F conditions. Yesterday, the court heard from Edlin that the device would only run properly between 72–82° F. Either Edlin or Young is mistaken or there’s some nuance here that we’ll hopefully hear about later.
Emails also revealed that some Theranos devices did make it to Africa and that they were tested in Uganda, Cameroon, and South Sudan. “The machine traveled well, seemed to functioned well in the environment. My only complaint is the touch screen. Very frustrating,” Melissa Givens, an officer with the US Africa Command, said in an email. She said she was hoping to get additional funding for a larger proposal. In his testimony, Edlin clarified that, despite the military testing the devices, no actual blood samples had been run through the machines while they were in Africa.
Yet in pitches to investors, Holmes touted the military’s use of Theranos devices in Afghanistan on medical evacuation helicopters. Edlin said that the company had developed protocols for testing at Bagram Air Field but that no devices had been sent to the country. He added that he did travel to Florida once with a machine, which was put through a battery of security tests but not any blood testing.
Marketing material inaccuracies
In further cross-examination, Downey made the case that Holmes and Young attempted to correct marketing materials to more accurately reflect the state of Theranos’ technology. In one instance, Holmes said that a disclosure at the bottom of the company’s website should be made more prominent after being alerted to call-center complaints about the matter.
The website’s statement was to read, “*Occasionally, a venipuncture may be required, based on the lab order. This is uncommon, and we aim to eliminate this scenario entirely.” Holmes wrote in an email, “I did not see this concept of the asterisk sent to me or mentioned to me at all. If we are going to say something like this, we need to own it and not contradict ourselves.”
“Uncommon” was a claim that hardly jibed with reality. Venous draws accounted for around 40 percent of all patient samples taken at Walgreens, assistant US attorney John Bostic pointed out later.
In another case, Theranos marketing materials were updated to remove a claim about results being ready in 30 minutes, instead saying, “Results in Hours. Not Days.” In one exchange over a similar claim on the website, Holmes pushed for a passage to say that results would be returned in four hours, not two. Yet here again, the marketing didn’t necessarily match reality. In trials done by Safeway, the grocery store chain, it often took three to five days to get test results back.
Who was behind the demos
Earlier in Downey’s cross-examination of Edlin, the former product manager said he was not out to deceive anyone in the demos for visitors and investors. In redirect, Bostic peppered Edlin with questions, seeking to nail down who was responsible for the shady investor demos.
What was the purpose of blocking errors on the machines, Bostic asked. “I don’t know,” Edlin replied. How does withholding results help show how well the technology works? “I don’t know,” Edlin said. Might visitors and investors want to know about inconsistencies? “I personally don’t know. I can’t speak for them,” Edlin said.
Who had final say over what results were in the demos? “Daniel did and Elizabeth did,” Edlin said.
Who did the final review of investor presentations? “Elizabeth,” he said.