The most interesting person in the (O.R.) world

Interview with Georgia Tech professor Eva Lee, whose applied work on biomedicine, healthcare delivery, pandemics, disaster response and other projects has global implications.

By Peter Horner

Eva Lee supported the U.S. Navy’s mission on the ground in West Africa for Ebola containment.

Eva Lee supported the U.S. Navy’s mission on the ground in West Africa for Ebola containment.

Eva K. Lee

Eva K. Lee

If the distinguished-looking gentleman in the Dos Equis beer commercials is the “most interesting man in the world” for his global exploits, then Eva K. Lee must be the “most interesting person in the O.R. world.” Her wide-ranging research and work on important big issues such as healthcare, biodefense, pandemics and emergency response are compelling since they impact so many people on a worldwide scale. Given the many disasters – natural and otherwise – that have afflicted the planet in recent years, Lee is certainly one of the busiest individuals in the O.R. universe. Whether serving in Fukushima, Japan, in the wake of the devastating 2011 earthquake, tsunami and near-nuclear disaster or advising the CDC as it responds to Ebola outbreaks in West Africa, Lee is either on call or onsite.

A professor in the School of Industrial and Systems Engineering at the Georgia Institute of Technology and an active member of INFORMS, Lee’s work on biodefense and emergency medicine began in 2003 with the Centers for Disease Control and Prevention (CDC) in which she focused on population protection against bioterrorism and pandemics. Her working relationship with the CDC continues to this day. In addition, she has served on several committees at the Institute of Medicine and the National Academy of Engineering and collaborated with the White House Biodefense Policy Directors since 2007.

Within the INFORMS community, Lee’s public health and biodefense work received the 2006 Pierskalla Award for research excellence in the field of healthcare management science, she was a finalist for the Wagner Prize in 2009 for excellence in operations research practice, and her work on behalf of a CDC project was a 2012 finalist for the Franz Edelman Award for Achievement in Operations Research and the Management Sciences.

In January, Lee was sworn in to serve a three-year term as one of the 13 members of the National Preparedness & Response Science Board (NPRSB). The high-impact board provides advice to the Office of the Assistant Secretary of Preparedness and Response in the U.S. Department of Health and Human Services (HHS). While NPRSB members have traditionally been bio/medical experts or local/state emergency response leaders, Lee is the first with an O.R./math background.

It’s not easy catching up or keeping up with Professor Lee, but we did, at least for 30 minutes, at the INFORMS Annual Meeting in San Francisco in November. In the following excerpts from the sit-down interview, along with several follow-ups, Lee offers an inside look at her most interesting life as an operations researcher.

You always seem to have an interesting, topical and relevant new project in many fields and many places, either domestically or around the world. Besides college professor and O.R. researcher, how do you define yourself?

I emphasize my training in mathematics. Much of my work and its impact come from being as creative as I can with mathematics. So I always define myself as a mathematician. The mathematics of pandemics fascinates me. We have been working on advancing U.S. medical and emergency response capability for the past 11 or 12 years with the CDC, the Department of Defense, and state and local public health departments. My pandemic work involves more than curbing the spread of disease; it embraces homeland security, because we seek to protect populations. My key to improving efficiencies and effectiveness in pandemic response is understanding when and how to apply mathematics. This includes improving vaccine design, discovering the biology of how humans respond to vaccines in the genomic level, modeling the disease spread across heterogeneous groups and communities, optimizing the dispensing capabilities, and mitigating overall infection and causality.

Another major focus of my research is advancing and designing systems models for biomedicine and healthcare delivery. Beyond operations efficiency, quality and safety, I devote a good amount of time on personalized treatment design and decision support within clinical settings.

You’ve worked closely with the CDC for many years. Describe that relationship.

The work with CDC began in 2003, focusing on biological defense and response preparedness. I have received CDC funding since 2007 for providing policy support as well as systems support for on-the-ground, optimized systems used to respond to many different scenarios. My relationship with the CDC touches the policy, strategic and operations levels.

I have also been appointed as a special government employee for the FDA, providing advice on the safety and quality of new medical devices. For example, how do we know that adequate systems tests have been done to assure there’s enough evidence to push a particular device out to the public?

Are there more pandemics, floods and assorted natural disasters around the world in recent years, or is it just more media coverage? What’s going on?

Some natural disasters such as floods, the consequences do seem more severe now. That’s because with more people building and living on the coasts, flooding causes more casualties and damage. Floods today are more severe because of changes that people have made to the landscape. Forested land is extremely good at absorbing and holding water. The leaves on the ground slow the water down and let the ground absorb it. A heavy rainfall can be retained by the land and absorbed slowly over several days. If you look at the statistics on disasters, I don’t know if we have more flooding going on now, but it does seem to affect more people with more severe consequences.

In terms of disease spread, with so many more chemicals now in the environment, many viruses tend to mutate and become stronger and thus harder to contain as a result. Again that’s part of the story, but not all of it. And of course, we now hear about these things in the news so much faster and much more often than in the past, so we’re more aware of them when they happen.

Pandemics are largely facilitated by globalization and air travel – many more people today are traveling and mobilized. For example, there have been dozens of Ebola outbreaks in Africa from 1976 to the early 2000s. It didn’t spread as fast since most occurred in rural regions. There was not as much travel then, hence the disease was more contained. The current Ebola outbreak spread across densely populated regions, across the country boundaries and outside of West Africa. The border is more porous now. It is more difficult to contain.

Can you share some insight about your recent work on Ebola? What agencies do you collaborate with and in what areas?

We focus on two areas. The first is how to best utilize scarce resources. We were working to support the U.S. Navy’s mission on the ground in West Africa for Ebola containment. The United States sent 3,000 troops to West Africa to build 17 hospitals. We designed strategic models on how to best use the resources and understand how the disease spreads. That way, we would know where to both locate the hospitals and how many beds were needed in order to effectively contain the disease. Containment is key.

The second focus, with the CDC, involves looking at scenarios of the arrival of infected patients onto U.S. soil. We explore how the disease can spread under different control effectiveness. For this work, we set up a website in late August that allows us to track the confirmed cases around the world, and to project from there, through stochastic disease spread and travel patterns, how disease could arrive at different countries, and how different control measures might affect the spread. We also explored how the spread might take hold in the United States, and move from city to city.

Do you think a 21-day quarantine recommendation is a good policy for Ebola containment?

Yes, it is for precautionary measure. Given that there is no effective prophylactic vaccine nor treatment for Ebola, self-quarantine and regular reporting status to public healthcare experts are the best way to monitor individuals who may have been exposed to the virus. That way, the moment a symptom occurs, caretakers have the best chance to mount a successful intervention. This is the best we can do for susceptible individuals. If we can do a safe quarantine overseas, that will be more effective. The U.S. Navy used this strategy, since they have doctors and experts who can provide the needed intensive care.

However, for non-military U.S. citizens (serving in West Africa), overseas support for an exposed individual may be lacking. Since we have the best possible environment for intervention here in the United States, bringing such individuals back to the U.S. gives them the highest possible chance of survival.

While a similar outbreak in the United States is highly unlikely, is the U.S. prepared for an Ebola event on the scale we saw in West Africa?

With about 100 million doses of seasonal flu vaccine available, the question is, how many citizens will get vaccinated?

With about 100 million doses of seasonal flu vaccine available, the question is, how many citizens will get vaccinated?

We are not. Ebola is different than other pandemics because there are no effective vaccines widely available at the moment. A patient’s survival largely relies on supportive ICU care, which is very labor-intensive. The ICU and all personnel have to understand the disease and know how to take care of the patient. The area has to be isolated. It requires a dedicated team of care providers with special training working in a sealed-off environment in one part of the hospital. How many hospitals and ICUs have that capability? Not many. Very few. Perhaps too few.

It is not that hospitals cannot prepare to set up such an environment or personnel cannot be trained. Yes, it can be done and it does take time. But the setup would drain the already very tight hospital resources. They will lose some of their regular and emergency service capabilities. In a resource-constrained setting, the challenge amplifies.

Another question is, do all the doctors and nurses know what it takes to protect themselves? The answer is no, not yet. You really need to follow protocol, and that is not trivial. They have to be trained and tested. In some ways, the military is better at such training and responding to this kind of pandemic outbreak of a disease such as Ebola than the civilian healthcare community. It’s all about protocol adherence and compliance.

Far more people in the U.S. are going to be impacted and die from the flu than from Ebola. From your viewpoint, how well is the U.S. prepared for a flu pandemic?

The difference between Ebola and flu pandemic is that the U.S. healthcare sector is always aware of the flu. The CDC provides year-round training. When it comes to flu pandemic, the key is healthcare resources. People do not necessarily die of flu; they die of secondary infections. They will go to the hospital, require care and be susceptible to infection. That’s why protection and prevention are so important. We have about 100 million doses of seasonal flu vaccine available, so it’s a matter of how many citizens choose to get vaccinated. For a flu pandemic, there may not be sufficient vaccines to protect the entire population. Or the vaccines may not be available when we need them. A challenge, which we are working on with the CDC closely, is how to strategically vaccinate among the population so as to achieve the minimum infection and mortality.

We’ve seen that one strain of flu can mutate into another, thus reducing the effectiveness of a given vaccine. Are the flu vaccines keeping up with the mutations?

You probably read in the news that the flu vaccine for this recent season was about 20 percent effective. It is tricky since the World Health Organization recommended on which strain to anticipate in February 2014. By the time the vaccines became available in the fall, due to drift variants in the influenza viruses, the vaccine produced is a mismatch against the circulating strains. This truly underscores the challenges and importance in rapid vaccine development.

Scientists are designing a flu vaccine right now that is supposed to be a universal vaccine, which means that no matter how the flu virus mutates, the vaccine will deliver. It’s supposed to be ready in three or four years. Along these same lines, I have been working with the Emory vaccine group, looking at how to effectively design a better vaccine. We just finished a study on vaccine immunogenicity with five years of data. We were able to identify how individuals actually respond to flu vaccines, so that will be helpful in the design as well as in target delivery.

I understand you were sent to Japan following the devastating tsunami in 2011.

I was one of the U.S. scientists on the ground helping with the response effort. I was there as a principal investigator supported by the National Science Foundation, which provided the funding. I went there with my CDC emergency response colleague. We reported some of our findings to federal leaders. I had access to the patients who had radiation exposure, and they provided lots of information.

Many of the other scientists were experts in earthquakes and flooding. My expertise was on the operations and healthcare side of the emergency response, as well as on the radiological emergency response, which I have also been working on with the CDC for many years.

We designed systems to help with decontamination and rapid screening, and to help with health monitoring large populations that have been exposed to radiation. In Fukushima, I met with nuclear plant workers and displaced individuals who lived within five or 10 kilometers of the nuclear power plant. Amazingly, many of these individuals had no idea how to respond to such an event. That was not good. It was a heartbreaking yet very humbling experience for me. These individuals lost so much and were in crisis, yet they offered many details to me about their experiences. They truly helped me focus on advancing the research and the future training design.

Given its history, how could Japan not have been prepared for such a disaster?

The Japanese were really good dealing with the earthquake, and they were really good dealing with the tsunami. Even little kids, when they hear the emergency warning sirens, they know exactly how to react, what to do. But they had no idea about the radiological situation. This was one of the largest radiation leakages ever, and it’s still not clear whether it was a failure of the nuclear plant following the earthquake, or whether it was because of the plant fire and explosion following the tsunami and flooding. Before the disaster, everyone, including government officials, trusted the nuclear plant officials regarding the complete safety of the nuclear plants. Hence, they were unprepared.

Following the Fukushima disaster, many of the nuclear plant workers and displaced individuals had no idea how to respond to the event.

Following the Fukushima disaster, many of the nuclear plant workers and displaced individuals had no idea how to respond to the event.

Can you provide details on your decontamination work?

For Fukushima, our task was to provide support on decontamination based on our system to optimize available resources to make sure all individuals were effectively screened. This includes both the responders and the affected population. We were prepared for this exact scenario – an emergency response for both workers and citizens, and helping set up centers for radiological decontamination and monitoring. The system is designed for both actual operations as well as for planning and training.

So what’s your latest, interesting, topical and relevant new project?

We have many different projects, particularly in the healthcare sector. I love the implementation and getting the O.R. methodology down to the hospitals and providers. One of our biggest projects right now involves analyzing a large amount of data on patients that go to the doctor on a regular basis, some of whom have chronic diseases.

What are you looking for in all that data?

We’re looking at many different things. We’re investigating practice variance; for example, you and I have the same illness, but different providers treat us differently. So what evidence and what outcomes demonstrate which approach is the best for which group of patients so that every doctor can adopt it strategically to different patients? This allows personalized evidence-based treatment design, potentially offering the best possible outcome. That’s the first thing.

Second is predictive health. Some of these individuals are healthy and disease free. They come in for check-ups or they have some minor thing so they see a doctor. We have all their vitals, so that means we can track them and their lifestyle and how they progress. What type of risk factors do they have? Should they show signs of symptoms, we’ll be able to detect early and intervene early, such as with pre-diabetic patients and patients with a high risk of developing certain diseases.

The third thing we’re looking at is how do you manage quality healthcare over a large population? How can you positively influence individual behavior? How do you conduct patient intervention on a day-to-day basis in a cost-effective way? How do you close the gap in health disparity? Many challenges ahead. It’s really an exciting time in the healthcare industry.

We want to uncover useful knowledge from the data, and translate this knowledge into system models that facilitate decision-making – one that allows for early risk/disease detection  and target intervention and personalized treatment design. We want to accomplish this based on evidence; understand what works for what demographics and what type of disease characteristics. There are lots of avenues to explore, and I have a lot of passion in all these areas.

Peter Horner (horner@lionhrtpub.com) is the editor of OR/MS Today and Analytics magazines.