It’s Not Because She’s a Woman — We Just Chose to Believe Her
First thing’s first: What Elizabeth Holmes did was undeniably wrong. Her blood-testing company, Theranos, used third party machines for testing; lacked necessary FDA approval; mislead investors; and results were not peer reviewed. You can read more about the scandal in John Carreyrou’s new book, Bad Blood: Secrets and Lies in a Silicon Valley Startup, while listening to Taylor Swift’s hit song, Bad Blood. If you want to get really pumped up, start following the news on the upcoming film in which actress Jennifer Lawrence plays Elizabeth Holmes herself.
If you’re still curious to learn more about the Theranos drama, simply google “Elizabeth Holmes” and reflect upon the numerous photographs of a young, confident woman wearing a black turtleneck. Or watch her interview with CNBC where she proudly shows off the nanotube container. It probably won’t take long before you read descriptions of the “visionary CEO” who imitated the late Steve Jobs in both style and personality.
More troubling than the turtleneck, though, is the fact that countless media personalities, journalists and mildly sadistic spectators explain how Holmes achieved such rapid notoriety because people wanted to believe that she could be the next female Steve Jobs.
Even Vanity Fair reported that Holmes was favored because she was a woman emerging onto the tech scene, despite the fact that there are thousands of female scientists at Stanford (Holmes dropped out at the age of nineteen). To claim that Holmes got this funding because she is a woman is disingenuous — try telling that to the hundreds of struggling female founders in Silicon Valley (though if you want to help change that, email me, we’re fundraising).
We didn’t believe her because she was a woman; we believed her because the fake technology looked spectacular. Rather than chat about her personality, presentation style and ex-boyfriend, I’d like to dive into the technology itself.
Theranos incorporated as a consumer healthcare technology company in 2003 with clinical laboratories that offered comprehensive testing from blood samples collected by finger punctures (rather than the traditional vacuum container). To clarify, this technology has actually been around for years, albeit not at the micro level, and was first developed for diabetic patients requiring frequent blood glucose testing. Introduced over three decades ago, finger-stick glucose monitoring empowers diabetics by gaining better glycemic control.
Diabetes patients in the hospital require frequent blood glucose monitoring, sometimes up to every two hours. Prior to finger-stick glucose monitoring, they needed to provide 10 cc of blood for each specimen, which can add up to a lot of blood loss for laboratory studies (and that’s just for glucose measurements). This might mean having 5–10 cc of blood drawn twelve times per day, a huge volume of blood, which may lead to anemia.
When the new glucose monitoring technique correlated well with regular venus testing, hospitals began adopting the device, then clinics, and eventually, communities. Diabetics could get a prescription from their doctor to use a glucometer, which revolutionized parts of diabetes management. Diabetics requiring frequent tests could do the quick finger-stick measurement, get their blood sugar measurements instantly, and avoid an unnecessary laboratory visit.
The finger-stick technology was later used widely for people taking anticoagulants (blood thinners) such as Warfarin. Patients on blood thinners must remain within certain ranges; if you’re above the range, you’re at high risk for bleeding, and if you’re below, you’re at high risk for clotting. Typically, patients would go to the lab, have a sample taken, wait for the analysis, and follow the doctor’s orders on how much blood thinner to take. The finger-stick apparatus instantly showed results and enabled patients to take these tests at home. Finger-stick tests are commonplace; you or someone you know routinely takes finger-stick blood tests.
At the time Theranos came onto the scene in 2003, people were already going to Walgreens to get shingles and flu vaccines, blood pressure tests, and cholesterol measurements. Elizabeth founded her company on the claim that nanotainers delivered the same accurate results as standard techniques requiring larger volumes of blood. She also presented standard clinical correlation data using linear regression models.
I believe that people were rooting for her not because she was a woman on the path to becoming the next turtleneck-wearing tech monolith, but rather, because it would have been great to have a full suite of lab tests from a microscopic drop of blood. That’s why investors were excited. The premise that a tiny amount of blood provides the same quality of results as standard testing is intuitively really impressive. It would mean that more people, such as home health nurses, can draw blood; it means saving blood and money; and it means helping patients with a fear of needles (a fear Holmes possessed).
This sounded great, except that there isn’t a big difference to a patient between nano amounts of blood and two or three drops of blood. Compared with blood work done prior to the 1970's, a few drops is much better than 5–10 cc of blood in a vacuum container. Additionally, microscopic drops of blood do not offer the same array of testing as a standard clinical lab.
When patients visit the lab for a biochemical panel, generally accepted statistics (95% confidence interval) include a false positive rate of 5, 30, and 46 percent for sample sizes of 1, 12, and 20, respectively. Keep in mind, no test is 100% accurate.
We know how to predict false positives and negatives when drawing 10 cc of blood; however, it’s unclear whether using nano amounts of blood conforms to the same rigorous industry standards and were thought to be error prone. Given the aforementioned statistics, could you imagine what the false positive rate would be when ordering 100 simultaneous tests?
Without peer review and FDA approval, we cannot expect to trust any medical device, drug or clinical outcome. Even generic drugs must conform to stringent FDA definitions: absence of significant side effects compared to proprietary drugs as well as bioavailability in the range between 80–125% as indicated by several standard pharmacokinetic parameters.
Holmes wanted to shake up the ponopoly on clinical laboratory testing and global pharmacy chains offering finger-stick biochemical determinations. And it makes sense — it’s a huge opportunity. Insurance companies want patients to go to certified clinical labs that meet industry standards, and Holmes offered a potentially cheaper alternative to disrupt that entire system.
We believed Holmes because we didn’t see a reason not to. She worked with a team of supposed experts, presented convincing data, and hid work and research from public scrutiny. I hope, however, that the downfall of Theranos is a warning to all health-tech companies making big claims to investors, customers and, most importantly, patients. I also hope that investors and customers take to heart that if something seems too good to be true, it probably is.
Before investing or purchasing from companies that work with people, medicine and disease, hire an unbiased consultant for a second opinion, and confirm that the data is accurate and intellectually honest. Also, FDA approval is not optional — either something needs the approval or it doesn’t. Hire someone to do a background check and exercise caution.
Healthcare innovation is promising and exciting, but we need to remember we’re dealing with real people’s lives and livelihood. We must hold companies accountable to industry standards and prioritize consumer safety.
