Moving Past ‘Silver Bullet’ Solutions in the Pandemic and Beyond

Bruce Y. Lee, MD, MBA, professor of Health Policy and Management, executive director of Public Health Computational and Operations Research (PHICOR), and executive director of the Center for Advanced Technology and Communication in Health (CATCH), at CUNY SPH, has spent years advising governments how to navigate complex distribution and supply chains to carry out successful vaccination efforts, using computer simulation models. In December 2020, as the Covid-19 vaccination campaign began, he offered a warning.

“Don’t assume that current vaccine supply chains are sufficient to deliver Covid-19 vaccines,” Lee wrote in an Op-Ed for STAT. Decision makers, he advised, must anticipate factors like the size of vaccine packaging, the risk of “wastage” due to exposure to warm temperatures, and varying production schedules for the different vaccines.

“The assumption on their part is often that, once vaccines reach the market and are paid for, vaccine delivery is the easy part and they can somehow magically appear in people’s arms and mouths,” he added.

His predictions now borne out, Lee says the hope that biotechnological innovation alone could end the pandemic is emblematic of society’s misguided approach to public health crises.

“There are so many examples in health where people ask ‘What’s the one magic bullet? What’s the one diet? What’s the one ingredient?’ ” he says. “There has been a lot of focus on trying to find one cause and one effect in health and healthcare. But most of these health issues are symptoms of broken systems.”

Lee’s research is devoted to helping government, nonprofits, and the private sector address the complexity of society and its systems in responding to public health crises, using computer simulations powered by mathematical formulae.

Over its 14 years, PHICOR has advised organizations including the U.S. Centers for Disease Control and Prevention (CDC), the National Institutes of Health (NIH), the Aspen Institute, the Bill and Melinda Gates Foundation, and UNICEF. They’ve used models to estimate the cost of childhood obesity to society and predict the effectiveness of a strategy to prevent antibiotic resistant disease in hospitals. Since the start of the pandemic, their studies have addressed some of its most complicated challenges, like the level of vaccine coverage needed for herd immunity and how many respirators hospitals should keep in stock.

“Covid-19 shows how broad the effects of a public health problem can be,” Lee says. A virus that was spreading uncontrolled led to sports leagues having to shut down, businesses to change what they’re doing, many people losing jobs and places going out of business, healthcare professionals being overwhelmed. It had reverberating effects around the entire society. It showed that it’s all connected, it’s exposing and magnifying a lot of the existing problems of our society.”

The enormous impact demonstrates more than ever the importance of complex systems science in public health, notes Lee. “The pandemic exposed the weakness of single cause, single effect thinking,” he says. “Through the entire pandemic, we’ve seen situations where there’s a focus on one factor without understanding we need a combination of different strategies, a systems approach to tackle Covid-19.”

The case for modeling in public health

Meteorology, air traffic control, and finance have long used computer models to predict weather patterns, airplane travel, and financial markets. Yet systems modeling is relatively new to public health, where most research uses methodology that identifies associations between single factors and a disease. Data is analyzed and abstracted to yield a result, like that coffee is linked to lower cardiovascular disease risk.

A complex systems approach, by contrast, uses data to simulate the workings of interrelated systems and model different scenarios.

“We model a system’s complexity in order to address it,” says Lee. For example, a model of the obesity epidemic could include data on a neighborhood’s grocery stores and fast-food restaurants, access to bike paths and sidewalks, and the food advertising people are exposed to.

Lee compares this work to the popular video game Sims, where players design virtual lives and carry out actions like building a home, starting a family, or opening a business, seeing how it all plays out over time.

“We literally try to recreate what’s going on,” he says. “Our computer models can serve as virtual laboratories to help decision makers design and test different policies and interventions aimed to improve health and public health before trying them in real life.”

Advances in computing technology and data storage over the last couple of decades has made this approach much easier, but Lee is still part of a relatively small group of public health scientists using these methodologies.

His interest in modeling dates to when, as a kid, he developed a computer program that simulated playing tennis on various court conditions. As a student at Harvard Medical School and during his internal medicine residency, he started to think about how medicine could benefit from computational methods.

He recalls trying to convince a group of cardiologists they could use models to decide where to locate cardiac catheterization labs. “The response was, ‘We decide where to put cath labs, not some computer,’” he says.

He went on to work at the drug consulting company Quintiles, developing models to evaluate pricing and design clinical trials for clients before moving to academic public health.

Though he became convinced computer models could help answer different questions in health and healthcare, others weren’t so sure. “I was told many times, ‘don’t do this kind of work, you’ll never be able to publish, you’ll never be able to get grants,’ ” he recalls. He was advised instead to “ ‘pick a disease or pick a body part and do that.’ It’s good to have people who do that, but you also need the people who cut across things. I felt many of these problems were interrelated.”

Undeterred, Lee started PHICOR in 2007 at Pittsburgh University and later moved to the Johns Hopkins Bloomberg School of Public Health.

In the fall of 2019, PHICOR came to CUNY, where Lee also teaches systems science and computer methods to MPH and PhD students.

“When you talk about CUNY, it’s the city,” he says. “This university is so integrated with the population in New York. It’s hard to think of a better place to do this work, because not only do you have access to all kinds of perspectives and people and resources, but whatever you do in New York has impact, because New York is almost like a microcosm of the world.”

Infographics created by PHICOR to accompany findings published in PLoS Computational Biology and Infection Control and Hospital Epidemiology.

Modeling Covid-19

Modeling has been critical to efforts to control the Covid-19 pandemic. Last year, PHICOR developed a simulation of a vaccination campaign to figure out when social distancing could be relaxed. They found that in areas where Covid-19 coronavirus is actively spreading, 70 percent of the U.S. population would need to get a vaccine, with an 80 percent efficacy rate, publishing the study in the American Journal of Preventive Medicine in October. The New York Times adapted the AJPM model to produce an interactive tool on their website to demonstrate the factors that could lead to herd immunity.

In an Op-Ed in The New York Times written earlier this year, Lee encouraged everyone to get whatever vaccine is available to them. He referenced a study published in the American Journal of Preventive Medicine in 2020 where his team developed a computational model that simulated the entire U.S. population, the spread of Covid-19 coronavirus, subsequent outcomes of infection (e.g., symptoms, hospitalizations), vaccines with different efficacies and vaccination timings, and the associated costs along the way. The results showed that waiting for a vaccine with a higher efficacy would result in additional hospitalizations and costs over the course of the pandemic.

The team also found that reducing the virus’s contagious period by just half a day could significantly curtail transmission, preventing up to 1.4 million cases and over 99,000 hospitalizations—even if only a quarter of people with symptoms are treated. The findings were published January 2021 in PLoS Computational Biology. An even greater number of cases would be averted if the contagious period were reduced further.

“What this study showed is that even if a treatment were to have seemingly not a huge effect, it can have a reverberating impact in terms of saving lives and preventing hospitalizations, because it’s affecting the transmission of the virus,” said Lee. “The vaccine alone may not end the pandemic, so we have to continue to look for strategies that can be paired with the vaccine to help stem the spread of the virus.”

At a time when healthcare systems are strained like never before, PHICOR has used modeling to provide guidance on the number of N95 respirators to stockpile, based on admission rates, which they published in January 2021 in Infection Control and Hospital Epidemiology. Such work will prove useful going forward, according to Michael Lin, MD, MPH, an infectious diseases physician and epidemiologist at Rush University Medical Center, who collaborated with Lee on the research. “Modeling the demand for N95 respirators can help policy makers estimate the regional and national stockpiles of respirators needed for future pandemics,” he said.

Earlier in the pandemic, when some politicians advocated a quick reopening of the economy, PHICOR modeled medical expenditures and healthcare resource needs if the country took a “herd immunity approach” and 80% of the population were infected. The costs could amount to as much as $1.25 trillion, they reported in Health Affairs in April, receiving coverage in PBS and U.S. News & World Reports.

“There was a lot of talk about the economic impact of social distancing and closing measures and not nearly as much talk of the impact of Covid-19 itself,” says Lee. “We felt it was important for policy making to better understand what these extra costs are.”

Reframing the costs of children’s physical inactivity to spur action

While modeling is most commonly used for infectious disease, Lee’s work shows it has broad application.

“His is one of the few groups that have used these techniques, particularly around children and physical activity and obesity,” says Janet Fulton, PhD, Chief of the Physical Activity and Health Branch at the U.S. Centers for Disease Control and Prevention, who is on an advisory panel with Lee. “I think he brings, with his modeling expertise, a really unique perspective.”

PHICOR’s research team drew attention in 2017 with a study in Health Affairs that estimated kids’ low levels of physical activity would cost the United States $2.8 trillion total in direct medical expenditures and lost productivity.

The team made this estimate by simulating America’s 31.7 million children between the ages of eight and 11 as they grew up, changing their body weight based on estimates of activity level and calorie intake. Currently, over two-thirds of kids in the U.S. seldom exercise. At that rate, 8.1 million children would be obese by 2020, PHICOR found. However, if half of children exercised for 25 minutes, three times per week, society would save $21.9 billion in lifetime wages and medical bills. If all children exercised that much, the savings would be $62.3 billion.

The study shows that framing public health crises in terms of financial costs helps draw attention to them, which complex systems modeling is well-positioned to do. “We saw that there was this gap in understanding what the cost savings for physical activity is,” says Marie Ferguson, MSPH, a project director who has worked with Lee for the last five years. “We know we need to get kids more active, that’s great. But policymakers really want to see how that affects their bottom line.”

The Aspen Institute’s Project Play, an initiative that works to promote youth physical activity through sports, and the Detroit-based Ralph C. Wilson Foundation, have successfully used PHICOR’s modeling to encourage investment in kids’ sports in the Great Lakes region.

“Bruce’s work is essential in our space because people need to see the benefits well beyond sports of more kids getting physically active,” says Tom Farrey, executive director of the Sports and Society Program at the Aspen Institute. “The CEO of Wilson Foundation puts those numbers front and center when he speaks to the business community about why they need to care about this work.”

Reducing drug-resistant infection rates in hospitals

In 2019, health centers in the Chicago-area asked PHICOR to help evaluate a strategy to prevent the transmission of antibiotic-resistant infections called carbapenem-resistant Enterobacteriaceae (CRE) during patient stays, a rising challenge for healthcare systems. PHICOR’s model tested what level of healthcare system participation was needed to deploy a registry that tracks and alerts when patients with “superbug” infections are admitted.

The findings were promising: even if there was “modest” involvement of only 25 percent of the largest health facilities, CRE cases would decline by nearly three percent over three years. If all facilities participated, there would be a meaningful 12 percent reduction in new carriers and a nearly eight percent decrease in the prevalence of infection.

The results have helped support the Illinois XDRO (extensive drug-resistant organism) Registry, which has been accessed by a growing number of hospitals and nursing homes in the state, and a potential model for other state or national efforts.

Navigating vaccine delivery, including by drone

Lee learned the importance of supply and distribution chains many years ago, when he was asked by a funder to study how to develop thermostable vaccines, which can survive temperature fluctuations during shipping and storage. After building models, Lee identified the more immediate problem.

“We started realizing a lot of these distribution systems are inadequate,” he says. What good was developing new vaccines, he wondered, if they couldn’t get where they needed to go?

PHICOR worked in 2014 with the West African nation of Benin on a vaccination campaign for rotavirus, which is responsible for the dehydration deaths of 215,000 children annually, mostly in low-income nations. The team created a model that simulated the journey of thousands of vaccines across the country, representing each refrigerator, vehicle, and vial involved in the distribution process and obstacles like transportation bottlenecks and waste from expired vials. With certain changes to the supply chain, PHICOR found Benin could save $50,000 to $90,000 annually and improve vaccination rates, reaching virtually all children.

The model led the government to modify its supply chain so vaccines wouldn’t travel through as many storage locations and vehicles to arrive at clinics. By 2019, after employing these strategies and a mass communications campaign, Benin has rolled out the vaccine nationwide.

PHICOR has even tested the viability of using drones to transport vaccines over rough terrain in the nation of Mozambique. Distribution faces significant hurdles during the last leg of that country’s supply chain, from the district storage office to local clinics, as they are transported on motorbikes, rafts, bicycles, or even animals.

The group modeled how drones would fare in various, realistic scenarios—encountering bad rainfall and cyclones or wild animals, and even getting shot down out of fear they were part of a military operation—and compared the costs to the existing system. Used frequently enough, the model showed, drones could save anywhere between 20 to 50 percent of costs over land transport.

“It shows that new technology doesn’t have to be expensive,” Lee wrote in the MIT Technology Review.

Communicating in a pandemic and beyond

Lee has written about his research and other public health news for outlets like The New York Times, Time, The Guardian, and a regular Forbes column.

He believes communication is an essential part of his work. “Science is like that tree in the forest. If it falls and no one hears it, then it doesn’t really make an impact. If scientists only circulate information amongst themselves, it doesn’t change society,” he says.

“He’s very good at taking complex, complicated ideas and distilling them down in a way that’s easy for everyone to understand,” says Sarah Bartsch, MPH, a project director who started working with Lee 10 years ago.

Covid-19 has magnified the challenge of evidence-based communication about health.

“We’ve seen many situations where people were actively trying to spread misinformation,” Lee says. “At the same time, we have a lot of evidence and knowledge behind how to handle pandemics, including the coronavirus. The big challenge is, how do you communicate scientific facts and scientific knowledge to all decision makers, policymakers and the general public, so they understand why things are being done, or what needs to be done?”

While Covid-19 has highlighted the challenge of making the case for complex responses to large-scale problems, it’s also shown there is no other good option, Lee says. “Once the pandemic ends and is no longer on the front pages every day, we can’t go back to how we were before, where we weren’t using enough systems approaches.”