By: Hannah Lin (CC'23)
Could you talk a bit about your background?
I am an epidemiologist with a background in emergency medical services. I’ve worked in various places—hospitals, communities, morgues. That’s where I got a start thinking about epidemiology. Those are what sparked my interest in population data and epidemiology and the power of these to make positive change for large groups of people.
Could you describe your COVID-19 research at the moment?
Most of my interests are in trauma and injury— people getting in car crashes, people getting shot with firearms. Some of my work via emergency medical systems has dealt with mass events—responding to disasters, whether they be from epidemics, natural disasters such as floods or earthquakes, or things like mass shootings or bombings. So I have a history doing that.
For COVID, a pandemic is yet another form of excessive demand, a form of illness on a large scale. So I started applying what I did with hospitals and healthcare systems to the pandemic by looking at it from a number of different perspectives. One of those perspectives is just the sheer volume and whether we have the right healthcare resources to respond to the volume of cases. The other is the geography of it all—where the greatest risks are, not just simply calculating the risks regardless of place-based disparities.
How has your progress been?
We have the pre-print publication, and you’re talking to me on a day when I’m working on the peer review revision of that same publication for other scientists to look at. Peer review is when scientists send things out to their fellow scientists—epidemiologists, in this case, and we also have public health scientists looking at this.
This is a publication looking at counties across the US and their COVID-19 case forecasts over 6-week windows going into the future to look at which counties are at greatest risk of experiencing high levels of COVID-19. It also looks at matching the demand by patients who are seriously ill with the supply of medical resources in counties across the US. As you can imagine, there are different levels of access to hospitals by county in the US. It varies greatly. Basically, the work is trying to predict which counties are at greatest risk of having their demand of COVID-19 patients exceed the supply of hospital resources, especially critical care.
We have the mapping on our websites—it’s always dynamic, it gets updated every Monday morning with the newest figures: https://tinyurl.com/beh-columbia-maps. We also have another website that our programmers made, predicting for every county in the US, looking at the curves for the COVID-19 cases in those counties: https://cuepi.shinyapps.io/COVID-19/.
Have you faced any setbacks in gathering these data and making these projections?
I’ve been doing this for quite some time, thinking about the population-level surge of illnesses and injuries. The earliest work I did for this was the Oklahoma City bombing and then 9/11. For the Oklahoma City bombing, I was working at the federal government and we had a response to analyze the situation, the people who were injured, and how they were distributed to the hospitals in the surrounding Oklahoma City area. That was a relatively small event, actually, in the grand scheme. The federal government invested a little bit in solutions, but not overwhelmingly.
And then 9/11 happened, and there was vast investment in all manner of different things related to disaster response, not just mass casualty trauma, which is a high volume of trauma—like what happened on September 11th—but there was also concern about bioterrorism. So that spilled over into things like what we were going to do if someone weaponized smallpox or released anthrax over a city. There was a lot of resource investment in response, data gathering, modeling, and so forth that then suddenly went away because there were no events in the interim.
I think that was a bad decision, and what’s happened is that now, the challenge for us in building these models is that the pandemic caught us by surprise and we at Columbia University and various other universities have had to re-energize past work and pick up the slack. Now, I will say there are people who have been carrying this and doing this all the way through, so kudos to them; I think it’s fantastic that they’ve done that, but they’ve done it with limited resources. Mostly, all this response modeling—whether it’s the health system, testing and tracing, or the demand of the cases themselves—we’ve had to redo it, in some cases from scratch. The federal government and state governments have been supportive, but it’s taken them a little bit of time to get their response systems up and running again, to be honest with you. Some people would argue it’s taken too much time—there are lives that have been lost in the time it’s taken for policymakers in general to get on board with all this.
So the data we’ve been gathering have been available to us very freely. We did have some challenges early on in finding the data especially for every hospital in the US—these data are not in a good, publicly available, centralized place. They were in a lot of different databases around the nation. We had a large team of people who took the date and pieced them together to make something serviceable so that we could advise the coronavirus task force, advise the CDC, advise the US Army Corps of Engineers, advise states and counties, in terms of where they should create the next hospital beds. Many different agencies and action-oriented individuals used the data that we created.
So you’ve reached every hospital in the US?
We have data on every hospital in the US, yes. Almost 6000 of them.
Wow, that’s very impressive. So you talked about forecasting in 6 week windows--what is the general timeline of your research, and where is it going in the future?
There’s a research component of this in building a model that can then be used going forward into the future. And it does function in 6 week increments, but also in daily increments—every day over those 6 weeks. We’re in the process of not just looking forward and building a model that will continue to function to make these predictions, but also looking back and seeing how well our model has been doing.
That’s what I’m in the process of doing today—looking at that model and its accuracy because we want to know if there were overpredictions or underpredictions, over-responses or under-responses. Of course, we don’t want an under-response when large populations of people are at risk—we don’t want to under-allocate resources because people die in those situations. But something that folks might not understand is that over-response is also very dangerous. Allocating more resources than you need is very dangerous because those resources could be sent somewhere else instead of waiting around and not being used. So this model is really an attempt to balance appropriate resource allocation, what we call optimal resource allocation, using various math models.
Is your team still keeping prior projects going even while you focus on COVID-19 right now?
Yes, for sure. I actually run Columbia’s CDC—Centers for Disease Control—funded injury science and prevention center. We work on all manner of different things, from mass shootings to car crashes and falls. We just had a huge conference last week. There’s been a pent-up desire to have a national conference on injury prevention, especially in the context of the pandemic. A lot of discussions were had, ranging from domestic violence and child abuse on up to suicide. This was all in the purview of the CDC-funded center. We have this conference annually and we typically have it as a regional conference, where people come from New York, New Jersey, and Connecticut come (that’s our regional responsibility on behalf of the CDC), but this year, since we couldn’t hold it in person, we opened it up, and word got out. What was usually about 70 people turned out to be 350 people who came from as far as Alaska to join the conference and talk about issues of injury prevention.
Our work as epidemiologists is in many areas, not just infectious diseases and injury prevention – Columbia epidemiologists also lead global work in cancer prevention, cardiovascular disease, the opioid crisis, mental health, and other things. All of this functions and continues to move on. In the context of the pandemic, because so much has changed, but also, broadly speaking, we know that these have always been important and we need to continue to face these other challenges as well.
On a related point, and maybe this is the most important answer to your question: the social structure of the pandemic and the horrible disparities, the socioeconomic disparities that have been generated by the pandemic are something that I think our epidemiologists have appreciated from the beginning, but I don’t think the nation is appreciating enough. We have folks that have really been working overtime to think about and reduce the health disparities created by this pandemic, whether this be in our prisons or as a result of structural racism that has festered for generations; equally, I would say, to our people who are working to study the actual COVID-19 cases and the disease itself. So I can’t stress that enough. These health disparities that have been generated by the pandemic are a second disaster upon the first that we really need to take note of and do something about.
Are there any common misconceptions you’ve heard regarding COVID-19 that you would like to rectify or respond to?
Well, this morning I was reading about indecision by universities. I don’t know if that’s a misconception, but indecision about what to do. It’s tough—university leadership, and the public health people who are oftentimes suddenly placed in these positions of leadership in universities around the nation, are really doing some difficult tradeoffs with safety of incoming students and the desire to go back to classes and to teach. So I feel for them making these tradeoffs, and it’s quite a challenge at the same time: decisions need to be made, and folks need to know, at a minimum, the thinking of these impromptu leaders—a little more transparency in the thinking—and what the tradeoffs are that they’re making. Telling people about the details of those tradeoffs, even if there’s not a firm answer at any given moment, I think would be quite helpful.
Definitely agree. Last question: I know you already touched on this a bit, but what is your perspective on the future, both regarding your research and the broader impacts of COVID-19?
Our research-for-action is going to march on. We have a group, for instance, organized just prior to the pandemic, that has begun studying gun violence in the United States. It brought together 3 dozen Columbia scientists who are interested in studying gun violence who otherwise would not have met. This remains a very important issue for the United States. In some cities and rural areas, it is going to be an issue that is competing or will compete with the pandemic itself. So there are issues like this that we are not just going to abandon in the face of this pandemic. That said, there is a new context to contend with, I think. In terms of where resources are going to go, that context is going to be important to think about for us. It will be important to manage where there are real risks of COVID-19 and future outbreaks, where there aren’t, and where other health issues rise to the top. So we need to think about that and not slow our research on other very, very important health issues in the United States and the world, whether that be the issue of tobacco, which remains very high; the issue of diet and nutrition (food sciences), which also remain very high; climate change and health is another enormously important challenge that we and everyone else in the world continues to face. I would hate to see these things be given a backseat. But we do need to deal with the current issue of the pandemic and get that to a more manageable situation for everyone’s sake.