Like many people, I sometimes give in to the temptation to anthropomorphize a virus. It can become an “opponent,” a “villain,” the “enemy” in a war. The media will often sensationalize it, call its spread “remorseless,” refer to it as “evil.” But, of course, a virus doesn’t have a mind or a morality. A virus simply does what it has evolved to do: reproduce and thus go on existing. Viruses uniquely need a host—actually, a host cell—to reproduce. Viruses lack all the machinery to self-replicate, so they need to invade and take over our cells. A virus must find new hosts to invade each and every day. The viruses that are most accomplished at this move through the air invisibly. Viruses with the highest transmission potential are those that infect their hosts but don’t immediately kill them. Instead, they produce minimal symptoms, allowing the host to live long enough for the virus to fully replicate and be shed, through secretions or droplets, or to remain suspended, invisible, in the air, silently present, awaiting a new host. Successful viruses produce a chain of transmission, spreading from person to person, minute by minute and day by day. With cases on the rise, this novel coronavirus was clearly doing its job. It was less clear, though, if we were doing ours.
As January ended, a flurry of events made me realize that officials in the United States were starting to take the novel coronavirus, and the evident disease it was causing, more seriously. Most notable was the official announcement that the White House Coronavirus Task Force had been created and would be headed by Alex Azar. Still, many of the efforts the task force appeared to be focusing on—various travel restrictions, symptomatic screening, and voluntary quarantine—belonged under the same heading as before: containment. As the month drew to a close, the primary belief that we could prevent the virus from coming ashore in the United States was still very much alive, still replicating in the minds of those who should have known that containment had already failed.
I assumed that Bob and Tony, both of whom were also task force members, were involved in these travel restriction and quarantine orders. The screening of incoming passengers would now take place at twenty major airports around the country. Also by the end of January, the CDC had issued a Level 3 travel alert, its highest, recommending that all travelers avoid nonessential trips to China. This was an important move, but it came too late.
Cases rose rapidly everywhere in early February, so much faster than with SARS. Many people like round numbers and likely paid attention to the fact that, on February 5, Wuhan’s health officials announced that they had reached more than ten thousand “confirmed” cases. Given China’s spotty record on honest reporting and the likelihood that its consideration of silent spread was minimal, the real number was, of course, much higher. My PEPFAR data team at the State Department and I talked about whether it was actually one hundred thousand or more. Looking back, I see our number was almost certainly closer to reality than theirs.
While I continued to believe the worldwide numbers showed that silent spread was fueling this outbreak, hard evidence of this was still missing. Without aggressive testing, no country would know how widespread the virus was already. As much as I believed we were headed toward a global pandemic, a silent invasion of the virus across the globe, I lacked the kind of evidence I could share to get countries, including the United States, to take aggressive proactive measures. So, I kept observing, focused on ensuring that Africa would be alerted, and waiting for data that was more defensible.
I wanted to prove that silent spread was at work in this outbreak. The cruise ship the Diamond Princess would provide the evidence. On February 7, I learned that the liner had experienced a viral outbreak and was currently at port in Japan. Some passengers were reported to have pneumonia-like symptoms. Japanese medical authorities investigated thirty-one suspected cases of the novel coronavirus and confirmed the presence of the virus in ten passengers. On February 3, Japanese authorities promptly issued orders that the ship’s passengers be held on board in isolation for fourteen days.
In medical science, we often create experiments and computer simulations to replicate real-world conditions and outcomes. Right away, I saw this tragic situation for what it was: If no one were allowed off, the cruise ship would become a kind of human petri dish, essentially a closed environment in which the virus could reproduce, spreading (or not) from person to person within the confines of the vessel. I am sure that, in this moment, many experts believed that locking down the Diamond Princess would contain the virus. Unfortunately for those aboard her, they were now subjects in an ad hoc laboratory experiment—except, tragically, the stakes were much higher.
An eighty-year-old was among the original passengers. He had been aboard briefly and then disembarked back in Hong Kong. Ten days earlier, he had been in mainland China but far from Wuhan. Later, he would tell officials that he had a mild cough for several days. He was mildly symptomatic but throughout the course of illness, which was ten days to two weeks, he was actively spreading the virus. On February 1, five days after he disembarked the cruise ship in Hong Kong, he became sicker, went to the hospital, was tested, and became a confirmed case of novel coronavirus infection. He likely thought that what was hospitalizing and killing people was something severe and wasn’t at all like what he was experiencing.
On February 1, Japanese health ministry officials had informed the ship’s owners that an infected passenger had been aboard the Diamond Princess. Either in concert with the ministry officials or on their own, management determined that with so few symptomatic individuals present the virus presented a low risk to the health and safety of their ship’s passengers and employees. They repeated the mistake of not accounting for the asymptomatic individuals already spreading the unseen virus through the air.
However, the ship was placed in quarantine and would not be allowed to dock, and her passengers were prevented from disembarking. The captain informed the passengers and crew about the one passenger who had contracted the virus. He also informed them that they would be quaranteed for the next fourteen days. He asked anyone who had gotten sick while on board to report to the medical facility. There, they would have their temperature taken and answer some symptom-based general health questions.
That same night, Japanese health officials went door-to-door on the ship to identify those they believed needed to be tested for the virus, basing their decisions on the passengers’ either being symptomatic or sharing a room with someone who was. (Even this early in the crisis, Japanese research scientists had developed a test for the disease based on finding traces of its RNA in the nose.) They took 253 swabs from individuals based on the incomplete criteria they’d established.
Surprisingly, normal social functions and interactions among passengers and the crew were allowed to continue throughout the next day. The younger crew members also circulated among the passengers.
Unfortunately, if there was silent asymptomatic and presymptomatic spread, many of them would already have been exposed. Later reports indicated that the captain’s announcement had the opposite of its intended effect: passengers congregated, talking, no doubt, about the situation they had found themselves in. Some may have worn masks; others did not.
The Japanese health officials believed they had moved very quickly. The cruise ship’s management team believed they were enacting all the right protocols. Their initial, symptom-based screening had discovered only 31 people, among more than 2,600 passengers, who were believed to be infected with the novel coronavirus. The health officials viewed this as a low-risk-of-further-transmission scenario and believed the virus was now contained. They were, in their estimation, being properly judicious—protecting those on board with proactive isolation and those on shore by restricting who could leave the ship. Still, it was dismaying for me to see: If the Diamond Princess case was any indication, public health institutions around the globe, likely taking the lead from the WHO and the CDC, remained firmly focused on detecting the presence of the virus through symptoms, not testing.
Simultaneous to the news reports about the Diamond Princess, her passengers took to social media to describe what was going on aboard the cruise ship. Passengers were told to monitor their own temperature. If it increased to above normal, they were to contact the medical services personnel on board the ship. Some passengers were allowed on deck and were required to wear masks and to keep six feet away from one another. These protocols were in line with best practices, but to what degree other passengers stayed in their own rooms and stringently followed the guidelines was difficult to assess.
Meanwhile, with all the focus on the passengers, I wondered what was happening with the crew. On cruise ships, the crew is usually younger than the passengers, and they usually live in small, shared spaces while on board. The crew needed to eat and drink, and they would have to remove their masks to do so. I also believed the younger crew was more likely to have mild or asymptomatic disease and were likely silently passing the virus to one another and, potentially, to the passengers. In my mind, the Diamond Princess was a microcosm of what could happen in the wider world: the aged and vulnerable being protected and taking measures, while the young were asked to monitor themselves for symptoms. This created the false impression not only that if you didn’t have symptoms, you didn’t have the virus, but that younger people seemed unaffected by the viral spread—a fallacy that would help the virus spread widely as time went on. Additionally, with no emphasis on testing everyone on board every day, the virus would likely live out its biological imperative largely unimpeded.
I believed the Diamond Princess could still chart a successful path forward. Going back to the 2002–2003 SARS observations and my general understanding of the willingness of Asian populations to act upon government health regulations, including wearing masks, I felt that the Diamond Princess could fare better if those on board complied with the precautionary measures. If the Japanese officials on board achieved a high degree of compliance for masking and social distancing, if they made the silent spread visible by testing all crew and passengers daily and isolating the newly infected, they just might successfully mitigate against ongoing spread.
And so, the Diamond Princess became my evidence base. Data from its passengers could prove two things: that this was a highly transmissible virus due to its silent spread and that preventative measures, if enacted and complied with, could slow the spread. I started to track the ship’s case numbers as closely as those from around the globe. If I was right about Wuhan and its enormous silent spread transmission, I surmised that we’d witness the same scenario on the Diamond Princess. Unless mask and testing mitigations were followed rigidly, cases would dramatically increase.
Amid the unfolding drama aboard the Diamond Princess, one development made talking and writing about this novel coronavirus a bit easier. On February 11, the WHO stated that scientists had more properly identified the characteristics of the virus. It was now called SARS-CoV-2, and the disease it produced was known as Covid-19. I once heard that to name a thing is to exert some power over it. I hoped that would be true in this case, but unfortunately, naming it merely gave our tormentor a new identity.
As I made final preparations for my PEPFAR conference in South Africa, which would commence on February 17, the confirmed positive cases on the Diamond Princess rose. Even during my twenty-four hours en route to Johannesburg, February 14–15, the cases of Covid-19 rose: 10 . . . 61 . . . 135 . . . 174 . . . 218 . . . 285 . . . Once I arrived in Johannesburg, I continued to monitor the situation. The numbers continued to ascend from the first day of my scheduled three weeks there: . . . 355 . . . 454 . . . 542 . . . 621 . . . 634 . . . 691 . . .
The numbers were emphatic and clear. And these were just the visible cases: the Japanese continued to restrict testing primarily to symptomatic passengers and those in close contact with them, while the crew most likely had significant asymptomatic infection. Seeing spread like this in spite of the containment measures in place meant the virus must have been circulating widely, and spreading asymptomatically, before the quarantine was put in place on February 4. To those who would listen, silent spread was announcing its presence.
From thousands of miles away, the Diamond Princess was the wake-up call I was sure would spur the CDC and other public health agencies to further action. We were all now able to see the results of this floating laboratory experiment writ large. The documented spread was intense, going from 1 to 691 confirmed positives in only three weeks—and those were just the people with symptoms. If they had been testing more widely, among asymptomatic people, the real number could be two to three times greater: 1,200 to 1,800 infections. Despite the measures the Japanese health ministry had put in place, this explosive growth was clear evidence of silent spread.
Containment was not working—not aboard the Diamond Princess and not in other parts of the world.
In the United States, this wasn’t the story the Trump administration was telling. The Diamond Princess cases failed to shift the rhetorical focus on containment, and neither the president’s nor Health and Human Services’ actions were proportionate to the threat the cruise ship’s living experiment had exposed. While the president remained focused on the United States’ low case count, HHS had been preoccupied with limiting travel. They only screened by symptoms the thousands of Americans and permanent residents, and the one hundred thousand American cruise ship passengers returning to the United States from around the globe. Instead of tamping down fear, as the president’s remarks seemed designed to do, the administration and federal public health officials should have been warning us about what the Diamond Princess had shown. Instead, their language stayed much as it had been.
Containment. Containment. Containment.
Unfortunately, rhetoric would harden into inaction.
I hadn’t spoken to anyone on the White House Coronavirus Task Force since my January 31 African ambassador meeting with Bob and Tony. I tried to silence the alarm bells going off in my head by reminding myself that the confidence I had in their abilities wasn’t misplaced. I felt sure that they almost certainly had access to data I didn’t, data that had perhaps lessened their level of concern. But I struggled to reconcile the concern I felt with the lack of public action, especially in light of the demonstrable threat of silent spread. Where was the testing? Where was the warning that this virus could be spread by those without symptoms, so Americans could be on guard? Where was the full-throated preparation we at PEPFAR were doing to ensure the safety of those in Africa?
In the coming weeks, it would become even clearer that the Diamond Princess was a warning to us all, a show of force by a virus demonstrating just what it was capable of. Unfortunately, warnings work only if people are willing to act on them.
AT LEAST THE JAPANESE had a reliable diagnostic test. The same could not be said for us.
When Bob Redfield spoke at my African Diplomatic Corps meeting on the last day of January, all seemed to be moving along nicely toward the development of a diagnostic tool that could be rolled out quickly and in large numbers. Over the first half of February, though, it became clear that neither was true.
On February 8, we learned that contaminated testing materials had caused the tests the federal agencies had created to result in false positives. This production issue was resolved, but the delay was costly. We should have been on the brink of having millions of test kits manufactured and shipped to the states. Now, rather than ensuring that those kits were deployed at the first sign of an outbreak, the CDC had to validate a new test. This general disregard for and mismanagement of testing, as compared with the rapid focus on vaccines at the NIH, would continue to haunt us throughout the spring and into the summer.
We were no longer days behind; we were weeks behind. Early in a pandemic, days and sometimes even hours can be crucial. You are attempting to slow the spread before the curve becomes exponential. Basically, losing one week early on feels more like losing four weeks. Being four to six weeks behind at this initial stage meant that, by the time the federal agencies had caught up to where they were before the testing debacle, they were actually sixteen to twenty-four weeks behind the progress of the viral spread.
Though the overall number of confirmed cases in the United States was still relatively small when news of the faulty tests first broke in early February, the failure significantly delayed the states in their ability to test. While even under ideal circumstances it would have been difficult to get the state public health agencies to test widely at this early juncture, their ability to use tests merely in so-called surveillance mode (a common practice to see what viruses are circulating in a population) was severely compromised. At precisely the moment when we needed to be testing widely to see where and how the virus was spreading, we were flying almost totally blind. Tests should have been broadly available; instead, they became a precious commodity.
While the failure of the tests was deeply disturbing, I was also concerned about some of the decision making at the CDC. When the tests were designed back in January, they were built mainly for use and analysis by our country’s Public Health Laboratories and their often unique equipment. This made the kind of rapid, widespread testing of symptomatic and asymptomatic people we needed quite hard, if not impossible. Each state has at least one Public Health Laboratory. As a part of the global health security focus, many of these critical laboratories were expanded significantly with CDC pandemic preparedness funding prior to Covid-19. These labs are more like research facilities than diagnostic labs that patients would access. In fact, the general public is probably unaware of the function or location of these sites. The CDC uses the tests performed and analyzed at these labs to answer specific, and often isolated public health questions. One of their most common basic functions is to track the strains of seasonal flu circulating in a state to see how they compare to that year’s flu vaccine. From the thousands of symptomatic flu cases in a state, a Public Health Laboratory may test a few hundred collected viral samples over the course of flu season. Often, the results of the test aren’t returned to the hospital, clinic, or patient.
These CDC Public Health Labs are an important but niche research asset. Because they were designed for research or limited surveillance, they are not high-throughput (i.e., fast-processing, high-volume) clinical or commercial labs. Worse, most have been significantly understaffed for decades. As a result of all this, the Public Health Labs were quite unsuited for testing on the scale needed for this current crisis. And yet, the CDC had created a test that could be processed only at these low-throughput facilities, with their somewhat slower, older, more labor-intensive, lower-tech platforms and equipment. In comparison to higher-throughput, commercial labs, the CDC versions could process fifty to one hundred samples every four hours, far less than the five hundred to one thousand samples every four hours processed at the clinical and private diagnostic labs. If you think of them like software, the operating system at a Public Health Laboratory would be version 1.0 when the vast majority of clinical and hospital laboratories would be version 10.5.
Almost no hospital, clinic, emergency room, urgent care clinic, or commercial lab uses the unique platforms the Public Health Laboratories do. The vast majority of the laboratory capability in the United States lies with private hospitals, clinics, and commercial labs. So, designing a test that relied primarily on the Public Health Labs meant that the bulk of U.S. processing equipment was sitting idle. With the CDC limited to about 0.1 percent of the country’s laboratory capacity, all those high-throughput systems, the 99.9 percent of our national laboratory capacity, were sidelined when they were needed most.
The only reasonable explanation for this oversight was that the CDC had never envisioned using its test on the scale necessary to test widely and identify silent spread from the asymptomatic and presymptomatic cases. If you test only the symptomatic, you don’t have to account for the three- to tenfold increase that silent spread accounted for. In the case of SARS-CoV-2, you might be able to do significant and impactful research somewhere down the line using these facilities, but you wouldn’t be able to test proactively to identify where the virus was in the present moment. In other words, you could look into the past, but you couldn’t effectively view what was happening in the present enough to shape the immediate future.
A more tactical approach to testing development and production in an emergent crisis would have been to work with all the assets at Health and Human Services and the commercial manufacturers to create a test format that could be processed by university research, public, and private labs. This would have ensured the rapid development of tests for all our preexisting platforms across the United States. From where I was sitting, on the outside, I didn’t know whether this was because of some failure on the part of either the CDC or another of the departments within HHS, like the Office of the Assistant Secretary for Preparedness and Response (ASPR). The bottom line was that when it came to testing, cooperation with the private sector wasn’t happening.
The testing bottleneck that hampered our earliest response to the pandemic and continued for far too long had its roots in this early decision not to engage the private sector. That choice reflected the CDC’s belief that the virus it was then facing was like one it had encountered in the past. The United States had never needed to test on this scale before, and the CDC believed that it wouldn’t be necessary in this crisis, either. Unfortunately, all the test kits in the world are useless if you don’t have enough capacity to process them.
The problems didn’t end there. In declaring a public health emergency on January 31, the United States had initiated what are called “emergency use authorizations.” These allow the Food and Drug Administration (FDA) to take emergency measures to protect the nation’s public health in the face of a chemical, biological, radiological, or nuclear threat. Infectious diseases fall under that umbrella of possible dangers. The point of the emergency use authorization is to help make medical countermeasures (vaccines, drugs, biological therapeutic products, decontamination systems, and a host of other devices and equipment) readily available more quickly to serve the public’s needs. The system is also designed to speed the usual labyrinthine, but necessary, FDA approval process for vaccines and medicines. Still, just how much more streamlined the other approvals would be under emergency use was difficult to gauge. Having any one agency managing the approval process for as many measures as might be needed in a pandemic would put tremendous pressure on a small number of staff. This would be particularly true when it came to reviewing the new testing kits.
An FDA team worked diligently to evaluate and approve the CDC’s initial test. This was the good news. However, central to its approval was a statement that declared the tests approved for use with symptomatic people. This one statement created an immediate and terrible restriction. So, not only had the initial CDC test experienced false positives and been developed in a format that would limit its processing primarily to the Public Health Labs, but it was approved solely for use on those showing symptoms. The test was expressly not approved to do the most important thing we needed it to do: diagnose and isolate the asymptomatic, presymptomatic, and very mildly symptomatic people who were spreading the virus unknowingly.
Screening and testing only those with demonstrable symptoms would account for, at best, between only 25 and 50 percent of infected people. You can’t stop outbreaks by detecting fewer than half the cases. If you have incomplete data, then you might draw incorrect conclusions.
As it turned out, that’s just what happened.
FROM SOUTH AFRICA, I tried to sift through the implications of the testing debacle, but something else, aside from the immediate logistical setbacks of the failed tests, troubled me, something larger and more ominous: bias.
The failure to course-correct on the silent spread following the Diamond Princess disaster, combined with the design of tests solely for the Public Health Laboratories, demonstrated that those in charge weren’t looking at this situation through the correct lens. The CDC obviously expected the new coronavirus to behave like seasonal or pandemic flu, and they assumed they’d be able to spot it circulating in the public in the same way they spot the flu every year: by relying almost exclusively on symptoms. This was a huge miscalculation.
Historically, the CDC’s flu-testing effort is focused more on identifying the particular strain of flu that’s active in a community, rather than preventing its spread. Millions of people get the flu each year, millions have flu-like symptoms, but very few people are actually tested for the flu—the vast majority are diagnosed on the basis of their symptoms alone, with doctors reporting cases through “presumptive diagnosis” rather than “definitive diagnosis,” which requires actual testing. Also, by and large, the CDC’s flu surveillance system, built with ease of burden in mind, works very well, we think—for the flu. (Ease of burden means not taxing too greatly the limited public health labs we have with processing too many tests.) We expect that within a relatively small degree of variance, around twenty-five to thirty-five thousand people per year will die from the flu, and year after year, this has generally proven consistent—but this is a modeled estimate, not a definitive number.
Of course, we all prepare for the worst-case scenario of an influenza virus outbreak. In 2005 and 2006, while I was at the CDC, the White House Homeland Security Council outlined the “National Strategy for Pandemic Influenza” and the “National Strategy for Pandemic Influenza Implementation Plan.” In respiratory pandemic circles, leaders often talked of the risk of widespread avian flu or other H1N1 or H5N1 infections. We were always on high alert for these zoonotic viruses. This was evident in 2009, when the swine flu spread across the United States. We learned critical lessons from the 2009 flu execution, but those lessons were very much focused on vaccine distribution and employing the full clinical capacity of the U.S. pharmaceutical industry, including pharmacies.
Part of the CDC’s past success at combating the flu had been attributable to the establishment of the Pandemic Influenza Implementation Plan, an annual desktop exercise. It simulates a flu outbreak to see how the plan would help mobilize a response to a viral outbreak of the H1N1 or H1N5 variety. The belief was that by being fully prepared for the flu—in particular, the H1N1 and avian flu strains—the HHS/CDC preparedness plans would, according to the CDC’s website, “enable HHS to respond more effectively to other emerging infectious diseases as well.”
That was a worthwhile goal, to be sure. But my long history in the military had taught me that, while plans are great, you need to be flexible as the situation on the ground unfolds and not become locked into “the plan.” In the military, you are taught to allow your response to evolve, and evolve rapidly, based on your frontline information. In the case of SARS-CoV-2, those in charge had to observe what this unique virus was actually doing and be agile enough in their thinking and actions to change strategy early and aggressively.
You also have to learn the nature of the opposing force and how it operates. Simply put, the CDC was prepared for the type of viruses with which it had demonstrated past success: H1N1 and H1N5. But SARS-CoV-2 was a different kind of virus. The CDC believed its plan would be effective enough to deal with other emerging viruses as well as it did influenza viruses, but that expectation didn’t account for institutional inflexibility and the variability of a different strain of virus. As a result, our nation’s top public health agency would cling too long to the wrong response model, one based on a different type of virus, while not accounting fully for every type of spread, not just symptomatic. We can’t implement a response based only on past experience or on a model that worked for a different disease.
Later, many people would be led to believe that the Trump administration threw out the Obama pandemic plan, causing a delayed and ineffective response. This is not true. The pandemic plan is and has been continually updated from administration to administration through the CDC. Covid-19 merely revealed the holes in that plan. The perceptions and assumptions that created those holes were present in previous iterations of the plan throughout multiple administrations, both Democratic and Republican.
Experience can be a help or a hindrance. In the early 1980s, with AIDS, we simply didn’t know what was making people die, and we felt helpless as an unidentified virus silently spread. After being caught flat-footed by HIV’s asymptomatic spread, which took us years and thousands of lives to understand, I found my guard going up whenever a new virus seemed to exhibit even the faintest tendency toward asymptomatic behavior. Obviously, I knew that this novel coronavirus and HIV transmitted in very different ways, but when evaluating the data, I kept an open mind regarding the former’s possible silent spread.
This past experience with HIV, along with my limited evidence from other countries and from the Diamond Princess, told me that this was transmitting differently. But those two things—past experience and existing evidence—would not be enough to move opinions at the CDC. Based on what my experience working at the CDC had taught me about its intransigence, I knew that I would need more. No one, no organization, wants to be wrong, but to show that an organization as large and important as the CDC has been so wrong requires a lot more proof. And at that point, I couldn’t prove conclusively to anyone, even myself, what I fully suspected was true.
Given my position outside the domestically focused federal agencies and the White House, all I could do was share the information I had with someone I trusted. I saw the gaps in the flu model approach and communicated my concerns with Matt Pottinger early and often.
He got it. He advocated for a different, more urgent approach.
Though I had armed him with sufficient rationales he could present to others, the administration remained stuck in the flu groove. They stuck to the game plan without realizing that the rules this virus played by were dramatically different. Matt sensed that because he wasn’t a medical doctor nothing he said would carry the authoritative weight of the truth I was relaying through him.
DESPITE MY GROWING DOUBTS, I held out hope that the CDC would somehow be able to right the testing ship, that it would see the errors in its approach thus far and move to correct them. This was easy to imagine because I remained on the outside at this point, with no idea what was happening behind the scenes to address my concerns.
From his perch on the inside, Matt was less optimistic. As early as February 13, the day before I left for South Africa, Yen and I exchanged texts. Matt had told her that there was a lack of leadership and direction in the CDC and the White House Coronavirus Task Force. Both had failed to assess the significant lessons from the Diamond Princess. Instead, their primary concern about the situation was that a small group of Americans was on board the ship. The State Department had a plane at the ready to fly to Yokohama to bring them back home. I agreed that this was an absolute must-do and immediately I laid out a simple plan of action: getting them onto buses masked and socially isolated, driving them to the airfield, repeating masking and distancing on the plane, testing them repeatedly. Yen wrote that Matt had said a plan similar to mine was offered up, but Bob Redfield and Alex Azar had wanted another night to think about it. Yen then wrote exactly what I was thinking: unless they were evacuated quickly, everyone on board that ship was going to get infected.
In their hesitation to immediately protect the Americans on board, I sensed a larger pattern at work, one of limited action and too-cautious overthinking. I was all about now-now-now action; they were all about “wait for irrefutable data.” I also heard faint echoes of Matt’s comments about a lack of decisive, proactive leadership. It’s understandable—no one wants to make a mistake. But in this case, waiting to be absolutely right would result in broader community spread and more fatalities, not fewer. With pandemics, acting rapidly and aggressively based on limited data, pushing the envelope and being prepared to fail spectacularly, can actually save lives.
These potential errors of omission were compounded as I looked at the messaging coming out of the White House. Throughout February, I became increasingly concerned about the Trump administration and how it was—or, in most cases, wasn’t—communicating with the public. One of the most essential components of any public health initiative is clear, effective, consistent messaging, stating over and over what you know and don’t know. Yet, any optimism I had about the consistency in their messaging had been quickly dashed against the rocks by weeks of deflections, diminishments, and reversals.
At a campaign rally on February 10, days after the ill-fated CDC tests were shipped more widely, President Trump said, “Looks like by April, you know, in theory, when it gets a little warmer, it miraculously goes away. I hope that’s true. But we’re doing great in our country.” The next day, he doubled down on his overly optimistic assessment: “In our country, we only have, basically, twelve cases, and most of those are people recovering and some cases fully recovered. So, it’s actually less.”
A government should not base its public health policies and emergency measures on miracles and hope. No one preventive measure is ever 100 percent effective, not even vaccines. You must have overlapping measures, duplications and redundancy. Whether it is containment, vaccines, or miracle cures, believing too much in any one thing—putting all your eggs in one basket—is dangerous. It gives the public a sense of invulnerability, of absolute protection, allowing them to take risks when you need them to take precautions. It’s why we layer protections in cars—crumple zones, airbags, seatbelts—each element adding some overlap and redundancy, resulting in more lives saved. This is what true mitigation looks like.
The president’s confidence in containment and control of this virus led many to underestimate its potential impact. The American people were the very ones who needed to take this virus seriously, and were the ones who would eventually need to comply with whatever mitigation measures were recommended or mandated. His overly rosy assessments only clouded the main issue: Containment alone would not work here. We couldn’t prevent the virus from breaching the invisible geographic borders. It was already on planes, ships, cars, buses, and subways, moving around the globe at speed. The virus had not been contained anywhere—not on the Diamond Princess, nor in Wuhan—so there was no reason to believe the United States would prove to be the exception. But this logic didn’t stop the administration from suggesting that we had little to worry about. On February 2, President Trump told Fox News’s Sean Hannity, “We pretty much shut it down coming in from China.” As late as February 24, the president tweeted, “The Coronavirus is very much under control in the USA.” On both those occasions and others, he fundamentally misrepresented the urgency of the situation and contradicted what we were already learning from other countries.
Another problem was that, at nearly every turn, someone in a position of authority was characterizing the risk to average Americans as “low.” I agreed that the risk for very serious illness might have been low, but the risk of getting infected was not. This virus was proving to be highly transmissible. The curves I was tracking across the globe clearly told that story. This virus was more contagious than SARS-CoV-1 and the seasonal flu. Other evidence existed to demonstrate this.
After all, the risk to the average passenger on the Diamond Princess certainly wasn’t low, nearly one-third of passengers showed symptoms in less than three weeks. And if the virus was being transmitted widely, the chances of someone with a weakened or aged immune system contracting it were greatly increased. More exposure meant more infections, which meant a greater frequency of serious illness and death.
I wanted to believe that, behind these optimistic assessments from the president, an aggressive and coordinated mitigation response was being planned and set in motion by HHS, that the White House and the federal agencies were working full-out, and not basing their approach on hope. I believed they weren’t publicly contradicting the president, perhaps believing that it was important not to create panic, but were working feverishly behind the scenes.
But then I saw Tony and Bob repeating that the risk to Americans was low.
On February 8, Tony said that the chances of contracting the virus were “minuscule.” On February 29, he said, “Right now, at this moment, there is no need to change anything you’re doing on a day-to-day basis.” I now believe that Bob and Tony’s words had spoken to the limited data they had access to from the CDC. This lack of comprehensive domestic data, and the reliance on projections from this very limited data, haunted the pandemic then and over the months and years to follow. If you weren’t testing enough, then you were dependent on the CDC’s interpretation of the data and on the flu model, and that could lead to vastly underestimating the extent of the silent invasion.
Maybe I just had a wider lens. I had access to more unreported, real-time global data, but maybe they had data in the United States that I did not. I trusted them, and I felt reassured every day with them on the task force. In the meantime, all I could see was how high the risk to Africa was, and I continued to concentrate my efforts there.
In hindsight, I know now that the overly optimistic language being used by the president and agency leadership was rooted in the lack of definitive data and at least partially rooted in a debate I’d experienced many times in my career at the CDC and with other medical professionals. When you have a virus that produces a spectrum of disease ranging from no acute symptoms in some people to deaths in others, you need to carefully explain in real terms what this means at the community level. When you have inadequate data to make a definitive statement, you say so, even if it’s not politically expedient to do so. You inform the public that we don’t know really how widespread the virus is because we don’t have the data systems in place to fully track it. Too often, scientists believe they need final, perfect data; or that scientific concepts are too complex for laypeople to understand. So, rather than finding the right communication strategy to convey difficult ideas, they oversimplify. But oversimplifying the message can be problematic.
Equating Covid-19 to seasonal flu (an illness people knew) implied that we knew more about it than we did. And not making Americans aware of the potential depth and breadth of the threat caused some to dismiss the virus as nothing more than a bad cold or the flu. Later, when we tried to convince them of the unique seriousness of the virus and of their need to protect vulnerable people in their families and community, they became confused, still believing the earlier oversimplified statements.
I don’t believe this level of simplification is ever needed. I believe that it’s the role of public health officials to explain complex ideas without reducing them to irrelevance. Not being completely transparent early on about what is known, what is not known but suspected, and what needs to be studied will only create later confusion. During my time on the task force, I heard repeatedly “The public will never understand the nuance. So, we are just going to say X, and not include the Y part.” But when the Y part is critical, and it later appears, people question everything you said before and, consequently, everything you tell them afterward. Talking down to people is never effective. Raising their awareness through effective communication is. Arming the public with every detail, telling them what we are studying, prepares all of us to learn together.
Months later, in September 2020, it was revealed in interviews Bob Woodward conducted with the president in March that Trump’s optimistic public comments contradicted his very real concern about the dangers the virus posed. If the president had just gone public and said these things to the American people it would have gone a long way toward conveying the seriousness of the situation then and now—no scientific nuance required.
This is a dangerous virus.
It transmits easily from person to person.
We need to protect the vulnerable, the elderly in the community and in nursing homes.
You’ll want to protect your family, your friends, and your local community.
The virus may not be in your area yet, but it likely will be at some point.
To keep it from doing great harm, please do these things: practice good hygiene, socially distance, wear a mask when indoors, and if you can’t keep safely apart, limit your interactions with others.
And most important, remember this: You may not feel ill, but you can still be infected. By getting tested regularly, you can prevent harming your family members, friends, and coworkers and still go about life fairly normally.
We will keep you informed, day by day, of everything we are learning. Even if we don’t have the final study or the final answer, we will let you know what we are doing.
We will let you know what we think will work while we get the data to ensure that it does.
We will learn together. Day by day. Together.
I was stunned when I learned in September of the president’s remarks to Woodward, and I remain unable to square the former president’s comments from that March interview with what he was saying publicly prior to that. In a White House in which people were unable to keep anything secret, I never heard anyone say that the president had expressed such sentiments to them regarding his understanding of the level of seriousness the virus presented. He never said anything like that to me, Bob, or Tony; nor to any of the agency leads of the various divisions within HHS; nor, to my knowledge, to any of his closest advisors. Word of such remarks would certainly have leaked, but it didn’t. Among the various mysteries surrounding this pandemic, President Trump’s comments to Woodward remain shrouded in what I can only think of as ego. Perhaps the president wanted Bob Woodward to know that, at a time when only a select few had the “privilege” or the depth of insight or the ability to see into the future, he had seen the truth of the situation.
Among those select few, working in deep background, was Irum Zaidi, my PEPFAR chief epidemiologist and data person. Irum assisted us in setting up databases around the globe so that we could bring accountability and transparency to the sixty-thousand-plus clinical sites PEPFAR supports in nearly fifty-five countries. Through her work, Irum knew another “data person,” who had access to figures about the novel coronavirus from around the world and very specific data from China. This individual was taking a great risk in passing it along to Irum, and his courage serves as an example for all of us.
According to the data shared with us, we learned that the virus was preying specifically on the elderly and laying waste especially to people with comorbidities (diabetes, lung disease, and the like). This was in line with what the world would eventually learn about the risks of severe Covid-19, but Irum and I understood these risks in mid- to late February. We passed along the need for greater vigilance and testing to our African PEPFAR people. Dr. John Nkengasong, the head of the African CDC, spoke to all those assembled in Johannesburg, his information striking the perfect balance among preparedness, prevention, and not panicking. Africa needed to be ready.
I had to believe the same diligent preparation was occurring in the United States. The president might be presenting one scenario, but Bob and Tony, the agencies, and the scientists had to be responding to a different one, right? I stood astride a fulcrum of hope and fear, optimism and frustration.
I shared my fears and frustrations about the U.S. response privately with Matt, Yen, and others. Publicly, I stayed in my lane—Africa, where HIV/AIDS continued to rage. I wasn’t going to step outside my chain of command (the State Department and my global mandate). The wheels in government grind slowly; I understood that. Besides, domestic policy was outside my area of responsibility.
That said, I cared deeply about the United States and what Covid-19 might do to the American public and my family scattered around the country. I had no qualms about continuing to assist Matt from the shadows, out of the spotlight. He was inside; I was outside. And I’d never been one to pull rank and publicly criticize my colleagues’ actions if I wasn’t there to see what they were seeing. That wasn’t my leadership style, and I knew that encroaching and criticizing seldom worked.
When I sent my insights to Matt, he would decide if they were useful and whether to share them with the White House Coronavirus Task Force. Fearing blowback for stepping outside my area of responsibility, I asked him not to use my name when discussing the opinions and data I was providing. If it meant influencing the White House’s approach, I was fine with my ideas being seen as his.
ON FEBRUARY 23, A Sunday, I was working all day, and so was Matt. He reached out to me again, this time by phone. There was urgency in his voice. He was especially concerned that at a recent task force meeting, Alex Azar and acting White House chief of staff Mick Mulvaney had made the decision to ask Congress for, as Matt later put it, “a tragicomically tiny budget supplemental” for therapeutics, vaccines, PPE, and testing. The ask was for less than a billion dollars, a figure that didn’t even account for the money needed to develop vaccines. It was not terribly surprising: the figure was proportional to the low level of threat they believed the virus posed.
Matt wasn’t calling just to gripe. He asked, again, that I come work at the White House. Only, this time, he was no longer looking for me to be a voice, a spokesperson. The situation at the White House had become critical. He needed me to step in and take on a more comprehensive role. Without a background in public health and medicine, Matt was just another nonexpert in the room competing to be heard.
I told him again that I couldn’t leave Africa and PEPFAR with critical planning hanging in the breeze. The next day, Matt phoned again. (He was certainly persistent.) He believed that the situation had deteriorated even more, both within the White House itself and globally.
“I’ve sent some more suggestions to Yen,” I told him. “Can’t we keep doing things this way?”
But Matt, to his credit, continued to press me; the status quo was untenable. The White House was vetting other names for a task force response coordinator role. Matt wanted to know if he could include mine. I told him I needed a night to think about it.
I spent most of the night awake, considering my options. I went over in my mind all the reasons I’d given before about not wanting to work in the White House. But this time was different. When Matt made me his first offer, back in November, it seemed like the want-to-need ratio was 85/15. When he came back to me in 2020, at the end of January, I’d adjusted this to 80/20 want-to-need. Now, though, I saw it as 25/75. The balance had clearly shifted.
The virus was spreading silently. Without extensive testing capacity, the United States would be completely unprepared. The administration had already done and said things that would undermine Americans’ faith in their public health care system’s guidance. Without that faith, how effective could any ask for behavioral changes be?
But Matt’s urgency represented another degree of concern: the unknown. If he was this concerned, what else was happening? What else would happen? With one of the highest security clearances, Matt had access to all kinds of information that I did not.
I knew if I said yes and got the job, I would likely stay until the crisis was over or the Trump administration ended—whichever came first. Was I prepared to make that commitment? I texted Yen; I trusted her opinion. She knew my background, and she knew me. She also knew the difficulties in the White House. Amid all the back-and-forth, she wrote me a short, simple text: “Remember, your country needs you,” followed by a winking emoji.
Had that message come from anyone else, it might not have tipped the scales. But, somehow, that simple statement of “need” weighed just enough to confirm my decision. Yen was a mother; she knew about doing what was needed, about setting aside your own wants and desires to help others. Her message echoed what I’d been saying to myself. For so many years, my decisions had been guided by this principle: Go where and do what is most needed.
On February 26, Matt called me expressing greater worry. He told me that every moment I delayed making my decision, I could potentially be costing American lives. Having served in the military for twenty-nine years, I knew that when a senior leader asks you to take on a mission for the American people, you commit to the mission. And you stay on mission 24/7 until it is completed or you are removed. That’s what soldiers do. That’s what anyone called to serve does. As simple as Yen and Matt’s message was—that I was needed—it contained multitudes.
The next day, I phoned Matt to let him know he could add my name to the shortlist.
If they decided they needed me, I was in.