searchSearch data by region...
Expert Insight

Q&A: Why Aren’t We Talking to Each Other About Data?

Experts across fields have been frustrated trying to figure out how best to build and estimate models of behavior and disease spread throughout the pandemic. Increased communication and collaboration on data collection and analysis is perhaps the best way to address this issue and produce better forecasts and policy.

Share
Authors:
Joshua E. Porterfield, PhD
December 8, 2021

It has become clear through interviewing faculty from diverse departments for the Pandemic Data Initiative that there are many brilliant minds working with pandemic data, and they have similar experiences and frustrations. However, as pointed out by economics professor, Dr. Nicholas Papageorge, experts from different fields are not often working together despite their common goal. Papageorge, also the associate director of the Poverty and Inequality Research Lab, recently co-authored a paper titled “Modeling to Inform Economy-Wide Pandemic Policy: Bringing Epidemiologists and Economists Together,” which serves as the basis of our discussion. The end to this pandemic and effective preparation for future health crises will require an unprecedented level of multidisciplinary collaboration within academia.

Can you describe the interplay between public health and economic data?

It's an interesting question because there is a lack of communication and a lot of miscommunications between epidemiologists (or people focusing on an illness and health outcomes) and economists. This crisis thrust people who care about public health and those who care about the economy into a very similar policy space. It turns out that we can understand each other because we use a lot of the same mathematical tools. However, we don't communicate, so we end up both building models about pandemics that say different things, to the point where it's incredibly frustrating because we talk past each other.

“This was an economic crisis and a health crisis from the get-go.”

What made you think to start a conversation between economists and epidemiologists?

It started so organically. I began talking to one epidemiologist, and we wrote a blog post together. Then we reached out to other epidemiologists and formed a group because we thought forming a working group could be interesting. We were all frustrated. I've worked on disease spread and individual behavior, but I was seeing economists put papers together that epidemiologists didn't take seriously and epidemiologists do work that claimed to take economic factors into account, but that I didn't see as such. Then there were pieces written about bringing these disparate components together that I personally felt did not get at the heart of the issue, which is to understand what it is we don't like about each other's work. I applied for and got a grant to then officially convene a group of five epidemiologists and five economists to write a consensus paper, which is now out as a National Bureau of Economic Research working paper. We published it as a working paper because the publication process takes so long. We wanted the message out quickly.

What did you learn from this working group?

It ended up being incredibly fruitful because we forced ourselves to clarify our differences. I do not believe that glossing over differences is a useful way forward. Understanding differences in perspectives is healthy in any relationship, including this one! One of the things I finally came to understand is that we're both modeling a pandemic, using a lot of the same mathematical tools to do it, informing it with data, claiming to care about how a virus is moving throughout the population, and recognizing that people's behavior matters — which sounds like we're doing the same thing, but the devil is in the details. The differences come out in the way we model things. Any mathematical model is going to vastly simplify lots of different dimensions, and we choose very different simplifications.

“If you get behavior wrong, you're doing bad epidemiology, and if you get the disease wrong, you're doing bad economics.”

Our models have very different focuses. Epidemiologists tend to have very detailed models of the virus itself. They care a lot about transmission styles, like ventilation or how people move about spaces. Differences in how people behave given their circumstances and how that might affect the virus — we call it endogenous response — is where economists put all of our effort. It's important to understand precisely how differences in physical space and virus biology affect spread, but it's also important to understand impacts on behavior.

What are the next steps to implementing the changes you discussed?

This convening was a call to action — we can do better. It's not just restructuring how we model. That isn't going to be something that leads to some grand consensus model to inform national policy by, like, tomorrow. That's a massive undertaking. But, it can start the process. It can also start to influence the kind of data we collect and how we collect it. Data collection from this pandemic has been far from ideal.

The epidemiology data give us a bird's eye view, but cannot be used to relate individual characteristics to their behavior and outcomes. Meanwhile, our economic microlevel data that tell us about individuals either don't ask about the pandemic since they were collected earlier — or were collected during the pandemic and have very limited numbers of variables on a non-representative sample. It's so unfortunate. People from different disciplines need to come together to talk about what kinds of data we would like to have prior to the next pandemic — and then put the infrastructure in place to collect them. My suggestion would be to approach one of the national household surveys with hundreds or thousands of people and ask if we could incorporate appropriate modules at the start of the next crisis. That would perhaps provide both flavors of scientists with useful data at enough of a granular level.

How do we redesign our systems to get better data?

Communication is key. What I hope was engendered is a conversation that begins with mutual respect and understanding that we have different priorities rooted in our goals and training. The first thing that I would do is think about the questions we want to answer together. What would we have wanted to answer if we could go back to the start of the pandemic? What is it that epidemiologists need for their models to understand how the virus will move around? I can talk about my models and what I need. Then we should get together and start thinking about which metrics we believe are essential to models and which ones we're willing to give up.

The kinds of models that we like to do are already sophisticated, so we can’t just add a lot of new data or machinery. We must think about our priorities and where we can compromise, then start to think about how we communicate our models with each other. There could be an integration of models, but there’s also the prospect that we maintain our separate models and improve them with each other’s input. Then we have to harmonize messaging.

“There has been a great deal of frustration in the public and that’s our fault.”

We did not help public discourse by saying opposing things as economists and public health officials, both saying we're experts, and both giving policy advice. That left the public and policymakers confused. When you leave a vacuum like that, it becomes easy for extremists to come in and fill it with conspiracy theories and other nonsense. We are rightfully appalled by some of the extremism, yet we're partially responsible for it because we couldn't agree as scientists about society's goals, health-wealth trade offs, the data needed, and the models to help understand the pandemic. We have to come together if we are going to handle this better in the future.

Joshua E. Porterfield, PhD

Dr. Joshua E. Porterfield, Pandemic Data Initiative content lead, is a writer with the Centers for Civic Impact. He is using his PhD in Chemical and Biomolecular Engineering to give an informed perspective on public health data issues.