Table of Contents
Published: October 30, 2026
Read Time: 7.8 Mins
Total Views: 30
Like many Americans, I went to see You Can See Everything this week, the documentary in which Nathan Fielder and Lance Oppenheim follow Elizabeth Holmes through the 34 days before she reported to federal prison, and I left the theater fascinated and somewhat bewildered by the people it portrays. The question the film keeps circling, whether we are watching someone deceive us or whether we simply lack the knowledge to understand what she is telling us, is one I asked myself more than a decade ago, sitting in my office at the city health department while Holmes and her team explained why their technology would transform laboratory testing in New York.
A Request From City Hall Brought Theranos to the Health Department
In July 2013, I was Deputy Commissioner for Disease Control at the New York City Department of Health and Mental Hygiene. I was responsible for overseeing the city’s public health laboratory, the facility that confirms outbreaks, identifies unusual pathogens, and anchors the laboratory-based surveillance on which much of disease control depends. That summer I received a request from City Hall asking me to meet with a promising new laboratory testing company. I signed a nondisclosure agreement and joined an hour-long video conference with Elizabeth Holmes and her team, accompanied by the director of our public health laboratory.
I still have my notes from that meeting. The company told us it had developed a new way to sample blood in very small volumes, and that in addition to being more convenient for patients, the method would reduce testing costs by roughly 30 percent. They said that a wide panel of assays had already been validated, that the regulatory work at both the federal and New York State levels was complete, that the company held a contract with the Department of Defense, and that testing would be available across New York within months at Walgreens and Duane Reade pharmacies. That last claim arrived two months before Walgreens and Theranos publicly announced their partnership in September 2013.
My colleague and I pressed them repeatedly about the underlying science and the data supporting it. Each time, Holmes gave us the same answer: everything had been approved, and the company could not disclose business secrets.
Nothing We Knew About Laboratory Science Made Their Story Plausible
When the call ended, the two of us were dumbfounded, because the claims ran against everything we understood about how specimens are collected, how tests are validated, and how regulatory approval actually works. Blood drawn from a fingertip behaves differently from blood drawn from a vein; squeezing a finger to coax out a few drops can rupture red cells and dilute the sample with tissue fluid, distorting results for analytes as basic as potassium, and dividing a tiny specimen across dozens of separate assays forces dilutions that erode precision with each step. Overcoming those problems would have been a genuine scientific achievement, and genuine achievements in laboratory medicine leave a trail of evidence behind them: comparison studies against reference methods across hundreds of specimens, precision and reproducibility data, established reference ranges, and, in New York, the extensive documentation the state requires before a laboratory may test specimens from its residents.
None of that evidence was offered, and the explanation for its absence did not hold up. Proprietary engineering can remain proprietary while performance data are shared, which is how diagnostic companies routinely work with regulators and public health laboratories. When a company asks for trust and declines to provide the data that would justify it, the absence of evidence becomes a meaningful finding in its own right.
Composure and Connections Can Make Experts Doubt Themselves
Still, I doubted myself. Holmes spoke with such composure and certainty, and arrived with such powerful connections, that I kept wondering whether I was the problem, whether there was something about the science or the regulations that she understood and I did not. That reaction, I have come to believe, is part of how deception at this scale succeeds. Confidence and institutional endorsement quietly shift the burden of proof onto the skeptic, so that the person asking for data begins to feel like an obstacle to progress, while the person withholding it appears to be a visionary protecting something precious.
We closed the meeting by offering to evaluate the device in our public health laboratory before any wider use in the city, which seemed to us a reasonable step for a technology that would soon be testing New Yorkers at their neighborhood pharmacies. We never followed up, and neither did they.
Journalism and Inspection Eventually Exposed What the Data Would Have Shown
When the Wall Street Journal published its first investigation into Theranos in October 2015, reporting that the company ran only a fraction of its tests on its own device and that former employees doubted that device’s accuracy, I called my colleague, and we both said the same thing: “We knew it!” The consequences that followed were severe. Theranos voided two years of results generated by its proprietary analyzer and sent physicians tens of thousands of corrected reports, and after federal laboratory regulators rejected the company’s plan to fix deficiencies at its California laboratory, Walgreens ended the partnership and closed every Theranos testing site in its stores.
Those corrected reports represent patients who made decisions about their health, and physicians who made decisions about their care, on the basis of numbers that were wrong. The system did eventually catch the problem, through investigative reporting and through inspections conducted under federal laboratory law, but it caught it after the testing had reached the public, which is the most expensive point at which to discover that a diagnostic test does not work.
The Friction in Diagnostic Regulation Is What Protects Patients
Looking back from the current moment, when the political appetite for deregulation is strong, this episode reminds me how grateful we should be for the protections we have at the state and federal levels for diagnostic testing, and how fragile some of them are. Last year, a federal court vacated the FDA’s rule on laboratory-developed tests, the category of tests designed and run within a single laboratory, and the agency declined to appeal and formally rescinded the rule. Laboratory oversight under federal law focuses largely on whether a laboratory performs its procedures reliably, which leaves substantial room for tests whose clinical claims have never been independently reviewed. Theranos ran its testing inside its own laboratory, and the regulatory gap that allowed it to reach patients before its claims were scrutinized has, if anything, widened.
To people who work in technology or in other fast-moving sectors, the systems that govern diagnostic testing often seem slow and frustrating, a set of hurdles that delay innovation. That friction, however, is what improves the science and the technology, because requiring companies to publish their results and provide data to regulators forces weaknesses into view while they can still be fixed. The laboratories that protect our food supply and detect outbreaks earn public trust through precisely this kind of documented, verifiable performance, and the diagnostic innovations that have genuinely expanded access, such as home collection kits for sexually transmitted infections, are valuable because their accuracy has been measured against established methods before anyone relies on them.
The title of the film captures what our institutions need in order to keep people safe. Regulators, public health laboratories, and the clinicians who order tests need to see everything, including the data a company would prefer to keep hidden, and they need that access most when the person asking for trust is persuasive, well-connected, and certain. My colleague and I had the training to recognize that the Theranos story did not add up, and we still found ourselves wondering whether we were the ones who had missed something. A system that depends on individual experts finding the nerve to doubt a charismatic founder in a single meeting is a system designed to fail; a system that requires evidence from everyone, every time, is the one that protects patients.
