By Greg Szrama III
A couple years (or was that months?) past, in February 2020, the United States government began planning in earnest for the arrival of COVID-19. By March, numerous media outlets ran pieces on the 1918 “Spanish” flu. Now as we find ourselves further along in the current pandemic, let’s take another look at what remains the most severe pandemic in modern history.
Reflecting on how COVID-19 is tracking against the 1918 flu can help us anticipate the next phases of this outbreak.
The ‘Spanish’ Flu
Despite the name, the 1918 Spanish flu pandemic did not originate in Spain. As a neutral power during World War I, Spain just happened to be the primary Western country willing to allow accurate reporting on the outbreak. News outlets in other nations could safely report on the “Spanish flu” without running afoul of wartime restrictions on reporting the outbreak in their own countries.
The first recorded cases arose in March in Camp Funston in Fort Riley, Kansas, affecting military personnel training for combat operations. From the initial outbreak, sickening about 100 soldiers, it spread in fits and starts for several months, finally tapering off by summer. The initial infection remained relatively mild, resembling a standard seasonal influenza outbreak.
In late August, a resurgent flu outbreak began. This second wave was caused by the same H1N1 influenza virus that circulated earlier that spring, so those who fell sick during the earlier wave were blessedly immune. Unfortunately, in the intervening months, the virus had mutated into a vastly more lethal strain. In October 1918, the deadliest month for the Spanish flu, approximately 195,000 Americans died.
Two smaller and less lethal waves followed in 1919 and 1920. The Centers for Disease Control and Prevention (CDC) estimates that 675,000 Americans died—more than the combined American casualties of World War I (116,708), World War II (419,400), the Korean War (36,754), and the Vietnam War (58,318). By the time the virus ran its course, upwards of one-third of the world’s population contracted it and between 5% and 20% of the infected died as a result.
Why Was the 1918 Flu So Lethal?
A disturbing pattern emerged during the second wave of the epidemic. The seasonal flu is typically most lethal among the very young, the very old, and the immunocompromised. By contrast, the second wave of the 1918 flu had its highest mortality rate among young adults. Nearly half the fatalities in the United States were in those between 20 and 40 years of age. Overall, 99% of Americans killed by the flu were under the age of 65.
One compelling explanation for the pattern of lethality comes from Paul Ewald in his 1994 book, Evolution of Infectious Diseases. Ewald believes the unique circumstances of World War I allowed the more potent strain to evolve from the milder initial strain. In a typical influenza outbreak, the most dangerous strains die out as those with the worst symptoms quickly seek care or self-isolate. By comparison, less dangerous strains proliferate because the infected with milder symptoms continue their daily lives. The 1918 flu inverted this pattern as those with more minor strains remained with their units or in the trenches. Soldiers with more intense sickness traveled from the front lines in crowded trains only to be placed in yet more-crowded hospitals, where the virus had far greater opportunity to infect others.
Modern analysis also turned up evidence that the 1918 flu likely caused a deadly cytokine storm condition. Caused by the immune system overreacting to the presence of infection, this condition leads to fluid buildup in the lungs and severe respiratory distress. Ironically, this condition can be most severe in the young and healthy as their stronger immune systems react more aggressively to infection. Many infected died within several hours of the onset of symptoms.
Ongoing research into what made the 1918 flu virus so deadly led scientists to begin sequencing the genome of the virus in 1999. Samples of the virus were procured from three different sources—two from military storage and one frozen in the Alaskan permafrost. Then in 2005, in a secure CDC facility in Atlanta (and under extraordinary safety precautions), Dr. Terrence Tumpey reconstructed the live virus. The work required direct approval from the CDC director and required many weeks of working in isolation until the virus finally grew in cultured human kidney cells.
Testing with the reconstructed virus has proved its extreme virulence. One experiment with live mice found that, four days after infection, lung tissue contained over 39,000 times as much of the 1918 virus as tissue containing a comparison virus. A separate test using a human lung cell line showed that the 1918 virus yielded as much as 50 times the amount of 1918 flu virus as a comparison virus.
The Great War At Home
The unique global situation in 1918 further exacerbated the pandemic. Doctors recorded the first case of the deadly second wave of the flu on Aug. 22, one day after the Battle of Albert launched, when Allied strength in the trenches peaked. This battle, followed by the Second Battle of the Somme and the advance on the Hindenberg Line, blazed a trail toward final victory of the Allied forces over the Central powers.
Because of the ongoing Hundred Days Offensive, as the plan was called, the United States (and other countries) faced a severe shortage in trained medical staff even as the virus caused peak casualties at home. The U.S. Army mobilized upward of half of all American physicians under the age of 45 (focusing on the best, of course) along with tens of thousands of nurses through the relatively young Army Nursing Corps. In addition to a dearth of personnel, hospitals and clinics had to treat patients with drastically reduced stockpiles of medicines and supplies as the war effort demanded their diversion from civilian use. Medical technology of the day also lacked the advanced antibiotics and understanding of viral replication we have since discovered.
On top of various shortages, news outlets feared reporting openly about the epidemic raging across the United States. The 1918 Sedition Act raised the specter of hard consequences should the outbreak be openly discussed. This law made it criminal to undermine the war effort; honest reporting on the severity of the pandemic would inevitably lead to accusations of just that. National health leaders repeatedly assured the public that the 1918 flu was a normal flu outbreak and that normal precautions would halt the spread. Policymakers hid the true extent of the virus, encouraging normal activity until it was far too late to mitigate the spread.
Progress and Aftermath of the 1918 Flu
Smithsonian Magazine, in a 2017 article, quoted several survivors recalling the 1918 Flu:
In Goldsboro, North Carolina, Dan Tonkel recalled, “We were actually almost afraid to breathe. … You were afraid even to go out. … The fear was so great people were actually afraid to leave their homes … afraid to talk to one another.” In Washington, D.C., William Sardo said, “It kept people apart. … You had no school life, you had no church life, you had nothing. … It completely destroyed all family and community life. … The terrifying aspect was when each day dawned you didn’t know whether you would be there when the sun set that day.”
The deadly second wave of the 1918 flu ended late that year, after the end of World War I in November. The third wave, starting in January 1919, spread from Australia to the rest of the world before tapering off in June. Less deadly than the second wave, the third surge still caused significantly more fatalities than a typical influenza outbreak. A minor fourth wave in 1920 marked the end of the “Spanish” flu.
|Why Was It Forgotten?
The most baffling aspect of the 1918 flu pandemic is its lack of lasting impact in our cultural memory. Diseases like the plague, malaria, and even typhoid are recognized and feared. Smallpox is reportedly feared as a potential bioweapon. But the flu, even the deadliest flu, is largely ignored or downplayed. Early in the COVID-19 pandemic it was even a byword for something mundane—COVID-19 was said to be no more dangerous than the seasonal flu. It bears remembering that the initial strain of the 1918 flu resembled standard seasonal flu outbreaks as well.
So why is the 1918 flu seemingly forgotten in history?
One reason goes back to wartime censorship and the more newsworthy items of the day—the advance of Allied troops toward ending a disastrous war that already stretched years longer than originally anticipated. In November 1918, for example, the largest news stories involved the peace talks in Europe, even as the country faced two consecutive weeks in which over 9,000 Americans succumbed to the flu.
Another leading theory is that the nature of the disease itself contributed to our forgetfulness. Contrast influenza to rabies, which—despite the wide available of vaccines dating back to the late 1800s—maintains an outsized place in the public imagination. Though comparatively few people contracted rabies, even in the early 20th century, most of the infected died horribly. Fear of rabies is prevalent in cultures worldwide and dates to antiquity.
In comparison, the 1918 flu appeared quickly and disappeared just as quickly. Despite its extreme virulence, the flu still has a much lower fatality rate, and deaths occurred rather quickly. Transmission is also relatively mundane, occurring largely via personal contact rather than by terrifying encounters with infected (rabid) animals.
The nature of a nation at war also gives some explanation. Those facing influenza at home viewed their fight against the disease as just another front in the ongoing global war. The second wave infected primarily young adults—the same cadre killed in combat. Lists of those killed by the flu inevitably joined the same blur as the lists of combat losses. One writer suggests those at home gained dignity in viewing the fight with influenza in warlike terms. He laments, “It is hard to make people actively aware of the fact that every nation under the canopy of heaven is at war with disease and death, and the human toll is vastly greater than that of any conflict in arms.”
The flu affected communities around the globe. Even where mortality remained low, so many adults became sick that normal life all but ceased. During the pandemic, schools, restaurants, cinemas, and other businesses all shut down. Civic society frayed as places of worship closed and public services drastically reduced their offerings. Neighbors who once took meals and necessities to sick families began keeping their distance for fear of contagion. Seemingly healthy young adults developed symptoms and died within a matter of hours—a terrifyingly fast progression that was beyond anyone’s comprehension.
After the pandemic ended, the young demonstrated multiple long-term consequences. One 2006 study found “cohorts in utero during the pandemic displayed reduced educational attainment, increased rates of physical disability, lower income, lower socioeconomic status, and higher transfer payments compared with other birth cohorts.” A separate study in 2018 found that, even after controlling for other factors, “in the absence of the pandemic, the 1919 birth cohort would have been more likely to graduate from high school.” Yet another study from 2007 described links between the 1918 flu and the encephalitis lethargica epidemic in the 1920s.
The pandemic led to further indirect effects in the provision of medical care. With no cure, and with doctors stymied by the disease, nurses took the forefront in treating the sick. Nursing even then remained a predominantly female profession and the ranks of nurses included many women from the first large classes of graduates in the 1890s. The heroic care provided by nurses encouraged the celebration of women in the workplace and helped allow for greater numbers of women both in health care and in other professions.
The Forgotten Pandemic
Despite the lethality of the 1918 flu, it quickly faded from public consciousness, so much so that historians sometimes refer to it as the “forgotten pandemic.” Carla R. Morrisey, writing for U.S. Navy Medicine in 1986, describes how formal publications all but ignored the pandemic and how it left little impression in the cultural consciousness. Personal remembrances, letters, and autobiographies shows how frightening and life-changing the disease became. Morrisey describes this as both a mystery and a paradox—that Americans as a people displayed complacency toward the disease and yet individually acknowledged it as one of the most influential events of their lives.
The fact remains that the death toll of the 1918 flu is surpassed only by the Bubonic plague and smallpox. Those other diseases, however, required centuries to achieve their death tolls. The 1918 flu came, killed, and left—in a matter of months. Morrisey writes, “No infection, no war, no famine has ever killed so many people in so short a time. Its sheer devastation is incomprehensible. The flu killed millions of people in 1 year or less. In the United States alone, 550,000 people died within the 16-week period from October 1918 to February 1919.”
Are We Better Prepared Now?
Our understanding of diseases has advanced drastically in the last century, as have the technological advances in treating them. By 1918 science generally accepted the germ theory of disease but many people still believed that “miasma”—bad air—was the main cause of influenza. Not until 1933 would the influenza virus itself be isolated, and another 12 years passed before the first flu vaccine became widely available. In 1918, scientists used specialized filters to prove the existence of infectious particles too small to be optically imaged; the scanning electron microscope used to directly image virus particles would not be invented for another 13 years.
Even with greatly advanced knowledge and tools, however, prevention still relies on the exact same techniques: wash your hands with soap and water, maintain social distance, avoid touching your face and eyes, sneeze or cough into a tissue or your elbow. At least public messaging campaigns are no longer needed to discourage people from spitting in public. Follow-on infections from influenza continue to be a major contributor to its lethality and new drugs like Tamiflu, while effective, face the potential of increased incidence of drug-resistant strains of the virus.
We face many of the same political headwinds that directly contributed to the lethality of the 1918 flu pandemic. Historian Kenneth C. Davis, author of More Deadly Than War: The Hidden History of the Spanish Flu and the First World War, discussed the potential for another outbreak in an interview with PBS in 2018. Asked whether this could happen again, he states that our precautions are weakened “when we deny science, when we ignore sound medical advice for short-term political considerations. Those things all factored into the spread of the Spanish flu 100 years ago, and those are things that could happen again today, if we weaken our defenses at the CDC, if we weaken our defenses in terms of cooperating with foreign governments about sharing information about viruses.”
The aforementioned Smithsonian Magazine article from 2017 echoed this concern. John M. Barry, author of The Great Influenza: The Epic Story of the Greatest Plague in History, writes, “In my view, the most important lesson from 1918 is to tell the truth. Though that idea is incorporated into every preparedness plan I know of, its actual implementation will depend on the character and leadership of the people in charge when a crisis erupts.” He described a pandemic “war game” in Los Angeles where the scenario involved managing an emerging pandemic with the first case in California identified symptomatically at a local hospital. Prior to the exercise, Barry briefed the participants on the 1918 flu, emphasizing the breakdown in social services and the necessary steps to retain public trust. He writes:
The participant with the first move was a top-ranking public health official. What did he do? He declined to hold a press conference, and instead just released a statement: More tests are required. The patient might not have pandemic influenza. There is no reason for concern.
I was stunned. This official had not actually told a lie, but he had deliberately minimized the danger; whether or not this particular patient had the disease, a pandemic was coming. The official’s unwillingness to answer questions from the press or even acknowledge the pandemic’s inevitability meant that citizens would look elsewhere for answers, and probably find a lot of bad ones. Instead of taking the lead in providing credible information he instantly fell behind the pace of events. He would find it almost impossible to get ahead of them again. He had, in short, shirked his duty to the public, risking countless lives.
And that was only a game.
How Does the 2019 Novel Coronavirus Compare?
The 1918 flu, as an H1N1 influenza strain, comes from an entirely different class of virus than the coronavirus strain that causes COVID-19 infections. Even so, with both diseases being respiratory infections, they bear striking similarities. Both are spread through close contact and through aerosolized droplets containing virus particles. Both cause mild respiratory symptoms in most of the infected but progress in some patients to acute respiratory distress syndrome (ARDS). In both diseases the most critical cases may include patients suffering from cytokine storm, causing or compounding the onset of ARDS. Both pose high risk of superinfection, further complicating treatment of the respective diseases.
The diseases bear their own distinct infection patterns. The COVID-19 disease is most lethal among the elderly and the immunocompromised, rather than the young adults at highest risk of 1918 flu complications. Patients infected with the coronavirus may take as long as two weeks to begin showing symptoms, whereas the 1918 flu followed typical influenza patterns with patients developing symptoms within two to seven days of exposure. Those suffering COVID-19 infections also present symptoms not usually associated with influenza, including severe shortness of breath and new loss of taste or smell.
Unfortunately, we see additional similarities in the public health response to the two diseases. In the early stages of the current pandemic, the Chinese government actively stifled the release of information related to SARS-CoV-2, echoing the self-censorship the Sedition Act caused. In a report in The Lancet, the first recorded case of human-to-human spread of the disease occurred in early December. Chinese authorities continued to deny human-to-human spread through most of January.
In one notable case, Dr. Li Wenliang, then practicing medicine in Wuhan, China, raised warning of a SARS-like illness on Dec. 30, 2019, in a private message to other physicians. The Public Security Bureau summoned him and forced him to sign a statement recanting his messages. On Feb. 7, 2020, he died of COVID-19, presumably contracting the disease from a patient he treated.
Worldwide, the response to the pandemic has remained hampered by lack of medical facilities and medical supplies. In a May report, the U.S. Department of Homeland Security laid out strong evidence that China delayed revealing the extent of the coronavirus pandemic through most of January while simultaneously buying large quantities of personal protective equipment. This exacerbated the supply shortages experienced worldwide in February as the disease quickly spread throughout Europe.
Hospitals found their response further degraded when supplies purchased from China—including respirator masks, tests, and other gear—proved unsafe to use. This scarcity of supplies directly mirrors the shortages in the 1918 flu pandemic where the ongoing war effort diverted supplies from civilian use, compounding the damage. Many hospitals, especially in Italy, found themselves with more patients than beds and faced severe shortages in critical equipment like mechanical ventilators. As in 1918, physicians had to focus on palliative care and make heart-wrenching decisions on provision of care.
In the United States, public health officials remained cautious in their response, mirroring the failed strategy John Barry noted in Los Angeles years ago. As late as March, officials in New York refused to issue a shelter-in-place order, even encouraging continued usage of mass transit. Public messaging at the federal, state, and local levels included confused and often contradictory directives—to wear masks or not, to isolate or not. When the scale of the pandemic became undeniably clear, health departments found themselves scrambling to line up surge capacity to handle cases. By late March, New York became the global epicenter of the pandemic as hospitals struggled to provide a response to the disease.
The Coming Months
Thankfully, most areas in the United States have not seen the same volume of cases that New York and other international hot spots saw. We have heretofore successfully “flattened the curve” while simultaneously increasing the capacity of our health care systems to cope with disease. As many others have repeated, this does not mean our problems are behind us. We did not lessen the overall number of anticipated infections, we merely stretched out the timeline to ensure that our health care systems are not overwhelmed.
Looking at our experiences treating the 1918 flu, we can glean a few details of what to expect. Most notably, we are almost certain to see additional waves of infection. For reasons not entirely known, respiratory virus spread is hindered by summer weather. In the coming months, the rate of new infections will likely decrease as our capacity to test for the presence of the virus continues to increase. Even so, the virus will certainly remain in circulation for the foreseeable future. As the weather turns cooler, when typical flu season begins, we can anticipate a new increase in cases of COVID-19.
Public policy between now and then will play a large factor in how severe future waves of infection become. Certain measures are already underway, such as increasing rates of testing. Others are just now starting, such as easing lockdown orders and allowing regular commerce to resume. Even though infection rates are ticking up where lockdown policies are eased, holding those policies in place too long may be just as counterproductive as ending them too early. Public officials risk burning through the goodwill necessary when then next wave does arrive.
The Anti-Mask League of San Francisco is an instructive example from 1918. The league was formed to protest the lack of scientific basis for wearing masks and the infringement on civil liberties caused by a local ordinance concerning their use. Protests today come from a similar place of frustration, anger, and fear, with people understandably worried about their livelihoods or even their continued health. Already alarms are being raised regarding increased severity of depression and anxiety, including increases in suicide rates, during our present lockdown orders. This overhang will continue to wear on the public, fraying social cohesion at the edges, and creating additional challenges for policymakers in balancing competing needs.
The long-term effects on those infected also remain unknown. Certainly, the most severely affected may face lifelong complications such as decreased lung capacity. Less obvious, however, are the effects on those with otherwise mild symptoms. Just as the 1918 flu affected those in utero in an unpredictable long-term fashion, doctors are finding strong evidence of pediatric multisystem inflammatory syndrome in children who survive infection.
Thus far, the pace of infection remains startling, with millions of known infected and potentially vastly more who are unknown. Yet there remains cause for hope. In many districts, isolation policies are effectively reducing the rates of new infections. Progress also continues toward a vaccine, with several already approaching or in human trials. Each day physicians learn better how to treat the severe cases, even finding success with ages-old techniques such as proning (laying intubated patients on their fronts to ease breathing).
We also see many positive parallels with societal response between 1918 and present day. In 1918, the American Red Cross arranged a massive and effective volunteer response. In various cities around the United States, this included organizing volunteer nurses, aiding in provision of masks and other supplies, and even ensuring adequate provision for gravediggers and caskets. They matched volunteers with sites having the greatest need, ensuring that medical professionals had support in caring for the afflicted.
In present day, volunteers are mobilizing to supplement medical stores through sewing efforts, through 3-D printing of personal protective equipment, and again through medical professional volunteerism. One Mayo Clinic intensivist wrote a diary of a week spent providing relief in a New York COVID-19 intensive care unit. Hearing of the situation, he wrote, “I served for eight years on active duty with the U.S. Army, including a deployment to Afghanistan in 2016 as a medical director for the major military hospital in Kabul. Frankly, the situation my friend described was worse than what I had seen in Afghanistan. This was a hospital at war.” Feeling convinced of his duty to help, he arranged emergency clearance with the aid of the Society of Critical Care and, within a matter of days, boarded a plane for New York.
As with the 1918 flu, we can look forward to a resilient recovery. While the economy is obviously hurting and we do not have a wartime mobilization to employ the otherwise furloughed, we can surely look forward to pent-up demand expressing itself in increased consumption once the lockdowns ease. The recovery will likely last several years, but we also have much greater knowledge of treating disease, so we are unlikely to see the massive fatalities experienced in 1918. Caution is still warranted in the coming months and years, certainly including a certain amount of social distancing for vulnerable populations until a vaccine is available, but we continue to be better positioned than any time in human history to deal with an epidemic on this scale.
One Big Question
The largest and possibly most important unanswered question is this—what will we learn from this experience? Around the globe, social activity has all but stopped in the effort to combat this disease. Will we do our best to forget the events going on around us, as seemed to happen with the 1918 flu pandemic? Or will we finally take to heart lessons about openness, about communication, and about how complex these situations are?
There are no good answers, or at least no satisfying ones, to the questions this crisis brings out. Will we be able, in the future, to extend grace to the public figures (and to our neighbors) who find themselves in lose-lose situations, trying to balance too many conflicting priorities and interests? We have an opportunity to better prepare our civilization for dealing with future pandemics—and we owe it to future generations to take every opportunity to do so.
GREG SZRAMA III is a freelance author and a software engineering manager at Accenture.
| Limitations on Pandemic Modeling
Policymakers at all levels are relying on epidemiological, computational, and statistical models in responding to the COVID-19 pandemic. While these models and studies based on them are born of the best intentions, they also have consistently failed us in predicting the spread of the COVID-19 disease. In preparing for this article I spoke about this topic with Timothy Atkinson, a Ph.D. candidate in computer science. In his research into Bayesian networks, he has developed expertise in statistical and computer modelling.
Early in the U.S. response to the pandemic, the Institute for Health Metrics and Evaluation (IHME) Model became the most cited in planning public response. This model has now received many well-documented critiques, so I instead focused my discussion with Atkinson on general limitations of modeling COVID-19 and how the public response has gone awry.
We began by discussing the primary function of such a model—its predictive ability. A model should be capable of accurate predictions for any arbitrary point in time for which prior data is known. If it proves unable to do so, then one or more of the underlying assumptions is wrong.
In Atkinson’s estimation, people quickly lost faith in public models, such as the IHME, because of their lack of predictive ability. Early predictions ranged from 50,000 dead upward to 2 million or more in America alone. In one sense, this wild variance in predictions is expected when modeling a never-before-seen disease such as COVID-19. The real issue is the confidence with which these predictions were presented.
Atkinson believes much of the issue with the accuracy of the early modeling attempts lay in a few key faulty assumptions. Most important to his mind is that we believed we could accurately identify the infected population. Of additional importance is the assumption that Americans would respond to certain policies in the same ways as citizens of other countries with vastly different local cultures.
Over time we saw our faults and adjusted models to match reality on the ground. Newer models prove much more capable of predicting infection rates, hospital capacity, and effectiveness of differing policy choices. This increased accuracy is just as expected as the early failures, and the important thing to remember is humility in presenting and using predictions from our epidemiological models. Over-reliance on wrong models will only contribute to wearing away the public trust necessary to combat an epidemic disease.
IHME Model Shortcomings– https://www.acpjournals.org/doi/10.7326/M20-1565
“Deadly immune ‘storm’ caused by emergent flu infections”; ScienceDaily; February 27, 2014.
“The Deadliest Flu: The Complete Story of the Discovery and Reconstruction of the 1918 Pandemic Virus”; Centers for Disease Control and Prevention; December 17, 2019.
“1918 Pandemic Influenza Historic Timeline”; Centers for Disease Control and Prevention; March 20, 2018.
“How the Horrific 1918 Flu Spread Across America”; Smithsonian Magazine; November 2017.
“Mortality Statistics 1919”; Department of Commerce, Bureau of the Census; January 29, 1921.
“Is the 1918 Influenza Pandemic Over? Long‐Term Effects of In Utero Influenza Exposure in the Post‐1940 U.S. Population”; Journal of Political Economy; August 2006.
“Fetal Shock or Selection? The 1918 Influenza Pandemic and Human Capital Development”; National Bureau of Economic Research; September 2018.
“Children and Encephalitis Lethargica: A Historical Review”; Pediatric Neurology; August 1, 2007.
“The Forgotten American Pandemic: Historian Dr. Nancy K. Bristow on the Influenza Epidemic of 1918 “; George Washington University, History News Network; June 11, 2012.
“Have Americans forgotten the history of this deadly flu?”; PBS; November 16, 2018.
“Readiness for Responding to a Severe Pandemic 100 Years After 1918”; American Journal of Epidemiology; August 9, 2018.
“The Influenza Epidemic of 1918”; US Navy Medicine; May-June 1986.
“Superinfections and Coinfections in COVID-19 – Separating the signal from the noise”; MedPage TODAY; April 28, 2020.
“The American Red Cross and Local Response to the 1918 Influenza Pandemic: A Four-City Case Study”; Public Health Report; April 2010.
“DHS report: China hid virus’ severity to hoard supplies”; Associated Press; May 4, 2020.
“Internal Emails Show How Chaos at the CDC Slowed the Early Response to Coronavirus”; ProPublica; March 26, 2020.
“Caution Warranted: Using the Institute for Health Metrics and Evaluation Model for Predicting the Course of the COVID-19 Pandemic”; Annals of Internal Medicine; April 14, 2020.