A Perfect Rapid Test Would Fail the FDA. Michael Mina Explains Why That Is the Point

Lex Fridman opens by calling rapid at-home testing the most powerful, doable, and obvious COVID solution from the start of the pandemic. Michael Mina, a Harvard professor of infectious disease and immunology, has been writing and talking about it since then. The tests are accurate for detecting contagiousness, can be made by the hundreds of millions cheaply, and leave privacy and individual choice intact. Fridman believes that if you give people information about whether they are infectious, they will mostly do the right thing. The case was obvious in May 2020, obvious a year earlier on this show, and obvious still. The conversation is about why it has not been done, and how it still could be.

Mina says the country has now lived almost twenty months in anxiety, shutdowns, and economic instability because of a virus, while a little paper test is finally reaching some households. People who buy one at CVS or Walgreens feel, for the first time, that they can control their slice of the pandemic. A negative is not perfect, but it is close enough to let you see your mom without infecting her by mistake. Politically, the ground is shifting: Biden’s COVID-19 action plan had just named testing, and rapid tests in particular, as a main pillar. Mina has been beating the drum so long that he half-jokes he hates rapid tests now. Most Americans, he confirms, have still never taken one. Testing, for them, still means a deep PCR swab at a clinic.

The at-home tests are paper strips. You swab the front of the nose, not a brain-tickling PCR swab, dunk it in a small tube of liquid, and either drip that liquid onto the strip or drop the strip into the tube. Two lines is positive, one line is negative, the same grammar as a pregnancy test. Once you know the motion it is about 30 seconds of hands-on time and roughly 10 minutes of waiting. Mina’s claim is that they are extraordinarily good at one question: am I infectious? That, he says, is the public-health question a pandemic has to keep asking.

PCR is different. Analytically it is more sensitive: it can detect one molecule where an antigen strip needs on the order of a hundred thousand. For public health, Mina does not want a one-molecule test. That extra sensitivity has been a net negative, because people stay PCR-positive for 30 to 70 days while they are only infectious for something like three to seven. A tenfold mismatch means isolating people who stopped spreading weeks ago, canceling flights, and tracing last night’s contacts of someone who was not contagious. It is a huge disincentive to get tested at all.

The virus itself grows from zero to a billion copies in a day, so you do not need extra sensitivity on the front of an infection if the trade-off is waiting one to three days for a result. You want to know, right now, that you are transmitting. PCR also fails on access: order online, wait a day for a kit, swab, ship, wait two more days, and you have burned four days, often past the infectious window. Speed, home use, and distribution are what make the cheap strip powerful.

Mina walks through a graphic he tweeted that overlays seven tests against viral load. You cannot ask “what is the sensitivity of this test” and stop there, which is what FDA currently does. Sensitivity to what? PCR will call you positive at a viral load of one and at a trillion. Against that gold standard, a rapid test that ignores people who are no longer infectious will look mediocre. Ask instead how sensitive it is when you are infectious, and he puts the number around 97 percent. Ask how likely it is to catch a superspreader, and he puts it at about 100 percent.

As viral load falls to non-infectious levels, antigen sensitivity versus PCR can drop toward zero. Mina calls that a feature. FDA and others average the whole range, declare the product 40 percent sensitive, and slow or refuse authorization. He wants at least a weighted average by transmissibility: how many samples were not transmissible, how many moderate, how many extreme. The country does not do that, which is why the tests are rare. The bar that is actually needed, he says, is 80 or 90 percent when you are infectious, where these strips already sit in the 95-to-100 range.

That mismatch is not a vibes argument. If a perfect public-health test is 100 percent sensitive and specific for the infectious stage, and people are infectious for maybe 20 percent of the PCR-positive window, then that perfect test can only score about 20 percent sensitivity against PCR at population level. FDA asks for 80 percent. “It is a mathematical fact,” Mina says, that a perfect contagiousness test cannot clear a medical-device bar built to count molecules. The test is not trying to inventory virions. It is trying to tell you whether you can infect someone today.

As a physician, Mina has one patient in front of him. That person is not asking whether they are infectious; they are asking whether they have recently been infected. For that question he wants maximum molecular sensitivity, can wait a couple of days, and does not care how many resources one answer consumes because insurance will pay. That is medicine. Public-health testing has to cover everyone you are not seeing, including tomorrow’s infection after today’s negative PCR. Accessibility, frequency, cost, and scale are then central variables. In a clinic they barely matter.

The reason the tools are missing is that FDA evaluates antigen tests as medical devices against PCR, without having to weigh time, cost, or access. Companies only know how to file medical claims, because there is nothing else to file. The United States has defunded public health for so long, Mina says, that it does not have a language, a law, or a pathway for a public-health test. He is trying to coin the term. FDA’s in vitro diagnostics division is doing the job it was given: reject applications that miss a medical bar. He has been hard on the agency, but he now thinks the failure sits higher. FDA itself says it does not review public-health tools. That sentence is the lever.

When he took a plan titled, in an Atlantic piece, as a way to get lives back to the White House, the reply was that it sounded interesting and checked out, except there were no tests at scale. He calls that defeatist. Everything happening now, variants included, was predictable a year earlier. Nobody treated the paper strip as urgent because it felt esoteric, and because the pandemic was always about to be over.

Mina’s ask of the president is one sentence: any tools used for public-health testing during this public-health emergency will be designated as public-health tools. FDA can then stop forcing a square peg through a round hole. CDC could take over, define metrics around high-viral-load detection, speed, and scale, and immediately certificate tests that have already performed well for months in Europe. That, he says, could put hundreds of millions of additional tests into the country almost immediately, instead of six-to-twelve-month trials that fail because they cannot keep up with PCR. He judged an XPRIZE rapid-test contest; a few days in a lab produced full performance metrics, and a systematized protocol could be an hour.

You cannot simply slap “public health tool” on a box and start shipping. FDA will call it a crime. The precedent for skipping EUA already exists, though, in pooled testing. Companies such as Ginkgo, working with around a hundred labs, have run pooled PCR on millions of schoolchildren, five to twenty-five kids per tube, with no FDA authorization because collection is on-site and the sample ships to a lab. A rapid strip used for the same purpose, Mina notes, is treated as a medical device. CMS, adopted by FDA, says a single result that alters one person’s behavior is medicine. A pool that alters twenty-five people’s behavior at once, quarantining a whole class on one positive tube, is somehow public health. He finds no logic in the line.

Masks sailed through as a CDC tool on evidence Mina considers trickier than the antigen data. Doctors, he argues, do not pretend to understand fluid dynamics, but they do believe they understand tests. The loudest “hell no” came from physicians who saw PCR-positive, antigen-negative pairs and called the antigen a false negative. From a public-health view that pair can be a PCR false positive: the person is not infectious. That category does not exist in clinic thinking. Early decision-makers were clinical lab people, microbiologists, and physicians, none of whom are trained in medical school to optimize for county-scale throughput.

He uses a TSA analogy. An agent’s job is this bag, this lane, the best scanner. How many other lanes exist, and whether people are walking around security, is not that agent’s training. A physician’s duty is do-no-harm to this patient, not how many tests a country can run in a day. The two goals are sometimes at odds. The morning of the taping was Mina’s first day back at Harvard since February 2020. He scanned a badge and was handed a PCR tube to return by noon. If he were a superspreader right then, the school was giving him a free day to infect people and a result two days later. The design did not match the goal. The president, he says, has authority to task CDC with certifying public-health tests; free-for-all shipping would create pandemonium and bad product, but the U.K. already ranked antigen tests outside its medical-device agencies and started deploying in weeks.

Once vaccines arrived, many people treated COVID as solved. Mina wrote an op-ed on why rapid tests were still needed. Trials were designed to stop disease; most did not even measure transmission, let alone make it an authorization metric. Expecting a virus like this to be shut down at the nose was, he thinks, naive. The first trial readouts produced a short-sighted celebration that missed three features: the shots might not stop spread, new variants might break through vaccine immunity, and efficacy was measured at peak protection in the first months after the dose, which inflates the long-term picture.

CDC and the administration went all-in on the claim that breakthrough cases were rare and that vaccinated people did not really transmit, without great data, especially once Delta arrived. When people then watched whole vaccinated parties blow up, they read it as vaccine failure. The shots, Mina says, are still keeping most people out of the hospital. They are not stopping transmission. Until society decides it is willing to let illness and spread run, it needs another tool. Herd immunity, on this path, is not arriving. Letting the virus train on vaccinated hosts is a boot camp for mutations around vaccine antibodies, and that makes him nervous.

Fridman asks whether the plan is everyone, every day. Mina does not want that. He wants a dynamic program: every household keeps a box in the cupboard and stops testing when the community has been quiet. Wastewater monitoring of viral RNA paints a map of prevalence. When the signal rises, send a text, put it on the news, tell that county to use one test per person per week. Weekly testing will not stop transmission, but it will show where outbreaks are. In those zip codes, test every two days or every day for seven days, squash the cluster, then stop again after a week or two with no cases. He does not want people in testing purgatory. He wants to get through the pandemic.

If some people skip the test or lie, that is fine. Coverage of 98 percent is better, 50 percent is still a lot better than zero. The only numerical goal is to push R below one: 100 infected people going on to infect 95. Physicians hear that and call it failure. For this virus it is a massive win, because a few weeks of it turns a large outbreak into a small one. Let 100 infect 140 instead, and in four weeks you have about 600 people rather than 20. That gap is the entire value of frequent cheap testing.

Manufacturing, as a generous estimate, can be around 50 cents a test. Germany has authorized some 60 or 70 high-quality rapid tests; you can buy one there for 80 cents and the maker still profits. Competition drives the price down. America’s bottleneck is not a missing factory. The largest diagnostic companies in the world will not sell billions of tests into the U.S. because they do not want to play FDA’s clinical-trial game. Open the door, Mina says, and existing scale walks in. He has half-seriously wished Elon Musk would make the strips, and has talked with Tesla people, including Jim Keller, about machines that ask whether a thing must cost a thousand dollars or can cost ten. The real gate is still authorization, not the converter line.

If tests are used to enter work or school, honor-system results are a hard sell. Mina has pushed “test to stay” so a child who would otherwise be quarantined can attend on a negative. Employers do not want nursing stations in the lobby. At a Milken Institute conference, every morning he scanned a QR code on the box; eMed, a proctoring service, watched him on video, verified the strip, and issued a laboratory report from the couch. CDC already accepts that workflow for people flying back into the United States from places like Heathrow: antigen test on a phone, TSA honors it. The tools exist.

Americans are wary of government-run vaccine passports. Mina’s line is that a person has a right to know their status on their own terms, and that public health must never hold that knowledge hostage. The pandemic-era subtext, as he hears it, was: if I cannot have the data, you cannot know either. That is backwards. People who will not test because they will be reported produce neither data nor safety. At work or school, a name attached to a result is hard to avoid. At a restaurant or a friend’s house, a report that lives only on your phone is enough. eMed, as a CLIA-waived digital lab, is legally required to report. An over-the-counter verification service would not be, and could let users opt in. Public-health colleagues get queasy. Mina, who is one of them, still prefers a used, unreported test to no test.

Vaccine hesitancy is already a distrust of authority. Asking those same people to report status to a central agency is, Fridman says, incoherent. Mina agrees, and wants verified services that can stay private. Companies that want to remain in favor with CDC and FDA have little incentive to build the no-report version. That is why, on his last visit, he was still pushing for government production. He has given that up. If scale were truly the issue, one company should be able to make 20 million tests a day with a faster machine. EUA and market access are the issue.

Biden’s action plan has several pillars. The two Mina watches are a rule that companies with more than a hundred employees must ensure unvaccinated workers test weekly, and a $2 billion effort to cut PCR prices by 35 percent, cheapen antigen tests, and make 280 million of them. He is glad testing is finally on the page. The problem is the “or.” Vaccinated people transmit just fine, so “vaccinate or test” is incoherent if the goal is to cut spread. It should be vaccinate and test. As coercive vaccination policy he will not litigate it on air; as science communication he thinks talking down to people like children with candy is a mistake. Honesty about uncertainty is what earns trust.

Fortune 500 leaders call him asking where they will find tests for 8,000 employees. Language about a “pandemic of the unvaccinated” is, in his view, divisive and false. It is a pandemic of a virus. The unvaccinated are often scared and have not been given much reason to trust public health. Tell people plainly: vaccines are not stopping transmission well, Delta and future mutants and waning are biology, and the shots are still what keep you out of the hospital. If the population goal is still to stop spread, use the tools that actually do that. Saying breakthroughs are rare, then watching whole vaccinated parties blow up, makes people wonder whether the entire program failed. That is shooting yourself in the foot.

Two hundred and eighty million tests is less than one per person per year in the United States. It sounds huge; it will not keep CVS and Walgreens stocked. That is crisis one: expectation without supply, which turns into anger. Crisis two: millions of newly mandated testers will swamp PCR labs into five-day delays. The only scalable answer is the paper strip, and the way to get it is to let companies that can already manufacture, including the global majors sitting out the FDA game, sell here. Mina supports the plan in principle. He also thinks a timer has started. If shelves go empty, the emotional attachment people now have to the little test becomes blame, and there will be political cost. Testing had been the rare COVID tool without a red-blue split. He has briefed the far right of Congress and the super-liberal side, plus governors of both parties, because a home test can stay unreported and asks only whether you can see your mother today. “Vaccinate or test” starts to turn “I have to get tested” into a loyalty test.

The worst part of COVID, Fridman says, is uncertainty, which becomes fear: vaccine-hesitant people afraid of the shot, vaccinated people afraid of the hallway. Testing gives confidence back. He finds himself nicer after a negative, because he is not putting people in danger. Mina agrees. He once required a rapid test at the door of a room in the masked-indoor phase of the pandemic; everyone relaxed. The alternative is masks that might give around 20 percent protection in a shared room, versus a rapid-test program that he puts nearer 95 to 100. Restaurants are scary because nobody is masked and you cannot see who is infectious. A room of same-day negatives is a different psychological object.

COVID response, he argues, is truly an information problem. Society did not close because everyone was positive; it closed because nobody knew who was. A classroom quarantines for ten days not because the whole class is sick, but because one case leaves the rest unknown. If infectious people glowed green, nothing would have shut and outbreaks would have been easy to walk around. The same logic applies to isolation. CDC and WHO picked ten days, or fourteen for quarantine, because they did not know when each person stopped being infectious. Some rare people spread for fourteen days; many are done in four. The extra week is blindness. Wake up, see two lines, stay home; see one line, go out. People will comply more with a result they can see than with a decree they feel fine ignoring. It even becomes a small experiment: come on, immune system.

Asked what listeners can do, Mina flinches. A year earlier on This Week in Virology he invited emails and was flooded; that episode helped produce rapidtests.org, which automated letters to members of Congress. FDA, he now says, will not change and should not be asked to, because it has decided how it regulates medical devices. The ask is that people with sway request the president use executive power to designate these as public-health tools and let CDC, NIH, or designated academic centers evaluate them on the right metrics. Ten new tests could be authorized in two days. He is finally talking to people close to the White House because the vaccine-only silver bullet did not pan out. Over time, he thinks the country can reopen even without a full testing fix, as adults slowly “grow out” of first-exposure severity the way babies do, perhaps looking at the virus as not particularly dangerous for most people by late 2022 or early 2023, with the elderly still at risk. He does not want that heard as advice to get infected. The faster, cleaner path is still information: tests are not big medical things, they are how you see the cars in the next lane.

Fridman asks about Mina’s time as a Buddhist monk, barely covered on the previous episode. Mina describes meditation as an active effort, not the absence of activity: exercise in quieting the mind, with the explicit intent to detach. In Buddhist thought, attachment, even to loved ones, keeps the cycle of rebirth. He does not personally believe in rebirth; he thinks we die and vanish. The practice was still powerful. Most people close their eyes and the blotches and thoughts start. Reaching a point where you can shut the mind down into silence is addictive, and a slope. He practiced about fifteen hours a day. The forest-monk end point is an absence of attachment so complete you would not mind losing a foot. The hat parable: love it and suffer when it is lost, hate it and suffer when you see it, or be ambivalent so that shredding it raises nothing. American life, he notes, is built for emotional rises whose average is supposed to come out happy. The monk’s line is flat. He does not call that the only right life, but he does call the exercise a way to become aware of your own mind.

He stopped after the 2004 Indian Ocean tsunami. He was in Sri Lanka, lived in a refugee camp, and asked what he could do. The answer was to keep meditating, for karma. Coming from Western roots he could not sit with devastation in front of him. He has not really recovered a daily practice, but he knows there is a part of himself most people never get to meet, and he can still use it to settle physical or emotional pain. He is not a Sam Harris of free-will metaphysics; he treated the work as utilitarian, minimizing thought and noticing sensation. On possessions he is split: stuff is a burden, empty surfaces feel peaceful, and he likes the idea of being able to pack up. Reality is messier. They are about to have a baby, boxes of bottles and tiny hats keep arriving, and he sometimes has to tell himself it is okay because it is not permanent. Fridman adds that even under mortgages and credit you are still free, and that matching the richer neighbor is good for innovation and bad for a finite life. Mina agrees.

Looking at the country’s failure to deploy a cheap test, Fridman asks if Mina is optimistic about civilization, given worse pandemics and other threats ahead. Mina is not. He does not dwell on it because it is too existential, but he thinks the American experiment is showing its limits and will probably become another rise-and-collapse blip in the history books. The inability to recognize danger, in the pandemic or in climate change, is extraordinary; climate, he says, can make this pandemic look like child’s play. Fridman bets that builders outpower destroyers across history unless a negative peak wipes everyone out. Mina agrees up to a point: you cannot build a society overnight, but you can kill one, and he thinks about food crises and instability. Until recently, the people engineering a better world did not generally rule it. Now unelected operators of huge companies hold as much or more power than elected leaders, and how they use it is open. He hopes that, as with electric cars or finally holding a rapid test, feeling the better product creates a paradigm shift, because channeling people’s desire to do good is a more durable way to keep power than fear.

Fridman closes by thanking him for still telling an engineering story when the moment feels solution-less. Mina’s last accounting is simple. He cannot not do this. If it saves one life, two years of lobbying were worth it.