Lauren Ancel Meyers was at home with her family in Austin one evening in April 2009 when she saw a news report about a deadly new strain of influenza circulating in Mexico City. She and her husband called a friend living there to find out what he’d heard about the virus. He told them one of his coworkers had died of the disease that very morning.
“It was really alarming,” Meyers, a mathematical epidemiologist with the University of Texas at Austin, told an audience at a lecture last fall. “It made my worst fears come to mind. By the next morning, everyone was afraid.”
No one yet knew that the outbreak in early 2009 was the beginning of the first pandemic of the twenty-first century—the result of a new strain of H1N1 influenza A virus, commonly known as swine flu because those who initially caught it were in direct contact with pigs. The U.S. Centers for Disease Control quickly gathered some of the world’s leading epidemic modelers—including those at Meyers’s lab at the University of Texas, which uses mathematical formulas to study and project how diseases spread. These experts began to discuss how they might prepare for a wider outbreak in the United States. Ultimately, H1N1 was estimated to have infected about a quarter of the world’s population and to have killed nearly 300,000, including more than 12,000 Americans and 240 Texans.
Even those death counts, though, didn’t represent anywhere near the devastation that the world might have suffered. To epidemiologists, H1N1 was a “near miss.” They knew they needed to get ready for a potentially far more virulent outbreak. So why, when COVID-19 emerged more than a decade later, did even they feel ill prepared?
Meyers’s explanation sounds a bit like the aphorism about generals always fighting the last war. When her lab developed a pandemic response toolkit for the Texas Department of State Health Services in the wake of H1N1, the hypothetical virus it was meant to respond to was a novel strain of the flu, rather than other threats such as today’s novel coronavirus. Because influenza readily evolves from year to year and easily spreads from person to person, it has long been the source of gravest concern for epidemiologists. “Now we talk about pandemics in general,” Meyers told me recently from her Austin home, where she has been working remotely since workplace restrictions went into effect in March. “But I think all the papers I’ve written on pandemics have really been about pandemic influenza.”
The tools Meyers developed for the state were designed to help leaders make quick, informed decisions about how most effectively to deploy stockpiles of antiviral drugs and to distribute flu vaccines—which can typically be developed in relatively short-order. The novel coronavirus required a whole new playbook. There were no proven antiviral drugs, and when the virus first arrived in the U.S. a potential vaccine seemed a long way off: in spring, many experts anticipated that even an emergency approval was at least eighteen months away (new vaccines typically take about a decade, but work on previous coronaviruses like SARS may have given researchers a head start). Even some of the social distancing methods discussed in relation to a pandemic flu seemed wholly inadequate. Meyers had discussed in her toolkit the potential need to close schools, because of the flu’s usual pattern of spreading among children first, but the idea of a widespread shutdown was uncharted territory.
As more and more officials sought guidance on the novel coronavirus pandemic’s potential trajectory in their cities early this year, Meyers’s lab added dozens to its team of scientists, engineers, and public health experts to monitor and study the spread of the disease. In March, the lab determined that COVID-19 was spreading twice as fast as had its predecessor coronavirus that was responsible for the Severe Acute Respiratory Syndrome, or SARS, outbreak of 2002–2004. Worse yet, they found that those infected by COVID-19, unlike those with SARS, appeared to be contagious even before presenting any symptoms.
The UT lab has now built models for 217 cities across the country, including 22 in Texas, in order to help guide policy decisions including when to institute restrictions on businesses, public gatherings, and travel, and when to relax such restrictions. “The pace of the science is just incredible,” Meyers said. “I’ve probably done as many analyses in the last three months as I’d done in the three years before.”
Building a detailed projection for a disease outbreak is a complex task. Meyers’s statistical model predicts likely outcomes by starting with characteristics of the virus, including how long an infected person is typically contagious, what percentage of those infected remain asymptomatic and for how long, and how many cases typically are serious or fatal. The model also factors in an array of population demographics, including the prevalence of vulnerable groups, such as the elderly in the case of COVID-19. On top of that, it layers in any countermeasures being taken, using anonymized cellphone mobility data as a proxy for how well a population is practicing social distancing. Then all of that is combined with the latest confirmed case, death, and hospitalization counts to anticipate how the disease is likely to spread and how deadly it’s likely to be. The data are fed into supercomputers at UT’s Texas Advanced Computing Center, which run thousands of simulations to determine the range of potential outcomes.
As of June 11, the Meyers lab’s statewide model, which is based on fewer inputs than local models, still anticipates a decline in the death rate over the next few weeks because of successful social distancing during the strictest period of Texas’s restrictions—though the rate will hold flat or even rise slightly in most of the major metropolitan areas. (That’s a rosier scenario than some more pessimistic models, such as the one released in late May by Imperial College London, which said Texas had the highest transmission rate for the disease among all the states.)
However, the numbers used by Meyers’ team are predicated on the assumption that most Texans will continue taking social distancing precautions. It’s an assumption that already seems to have been undercut by behavior, given that over the last week Texas has seen a significant spike in cases, hospitalizations, and deaths. Indeed, even before the model has been updated to reflect these increases, Meyers is not optimistic that Texans are doing what’s needed to control spread and avoid an infection surge throughout this month. She’s seen images of crowded Texas bars and beaches over Memorial Day weekend, as well as the massive gatherings throughout the state in protest of the killing of George Floyd. (The anonymous cellphone data that modelers use to approximate social distancing levels didn’t reflect a large uptick in movement for Memorial Day, and data from the period of the protests isn’t yet available.)
The UT projections warn that if Texans reduce, or are already reducing, their social distancing significantly, a second wave of COVID cases could come, overwhelming the state’s hospital capacity and forcing another three months of tight restrictions on businesses and public gatherings. Meyers and other public health experts, such as Peter Hotez of the Baylor College of Medicine and epidemiologist Rebecca Fischer of Texas A&M University, have repeatedly and publicly expressed their concerns that Texas is reopening too quickly. Governor Abbott, heavily influenced by certain business interests, has plowed ahead with his phased reopening, despite the fact that Texas has not met the criteria laid out by one of his own appointed medical advisers, former FDA chief Mark McClellan.
Yet even as she worries about a hasty reopening and the worst-case scenario, Meyers is careful to communicate more than just doom and gloom about our state’s prospects. “If people are taking precautions, and they’re wearing safety masks, and they are keeping their distance even as they got out in public more often, we may not see a second wave at all this summer or will maybe just have a very, very slow-growing epidemic” that will not overwhelm the capacity of our hospitals, Meyers said.
Meyers’s UT lab will continue tracking the data and refining its models to offer the best possible advice to officials looking to limit the toll of the disease. Whether those officials—and business owners and individuals—will take that advice is a whole other question. “We can tell you about what the virus can do, but we can’t tell you about what the people are going to do.”
