Coronavirus: "A pandemic is a behavioral phenomenon"

At the virtual.MEDICA 2020 trade fair, Prof. Dirk Brockmann delivered the keynote address in the MEDICA CONNECTED HEALTHCARE FORUM on digital epidemiology, which got a big boost thanks to the coronavirus pandemic. It can help us understand how human behavior influences the course of the pandemic.

In this MEDICA-tradefair.com interview, Prof. Brockmann talks about natural experiments and explains how wearable technology and collected data can give us valuable insights into the spread and the dynamics of infectious diseases.

Prof. Brockmann, what is digital epidemiology?

Prof. Dirk Brockmann: The field primarily builds on a combination of new methods and data sources. This refers to data that is being collected via smartwatches, wearable devices, or smartphones, for example. This data is often not generated under controlled conditions but stems from so-called natural experiments, meaning activities of daily living.

From a methodological perspective, digital epidemiology lies at the intersection of computer science, health research or medicine, making it similar to bioinformatics. This science field emerged because data processing was needed to make gene sequencing data analysis possible. Today it is a well-established branch of science.

Can you give us an example of data from natural experiments?

Brockmann: Mobility data is one example. Many people own a smartphone today, and it traces and tracks their movement and location. We can learn a lot from this and obtain relevant epidemiological information.

Another example is contact networks in hospitals. You equip staff and patients with RFID chips, allowing detailed contact tracing and analysis. The assessment of these networks is critical to understand the spread of nosocomial infections in hospitals.  

You also spearheaded the development of the German Data Donation App of the Robert Koch Institute. What real-world activities can you analyze by harnessing this data?

Brockmann: A study from early 2020 indicated that fever prompts an increase in your resting heart rate, which can be measured by wearable devices. If a person’s data deviates from the statistical norm, we can deduce that they run a fever.

Many people voluntarily shared their data with us via the Data Donation App. Fever is a common symptom of Covid-19 – though fever does not necessarily mean you are infected with Covid-19. If many people in a large group run a fever, it could be an indicator of the spread of Covid-19.

Which new insights are especially important to you?

Brockmann: Contact networks teach us many new things, albeit several aspects are still unclear. We can only understand infectious disease dynamics if we understand the contact networks. Traditional epidemiology looks at the number of cases. However, that is just one result of the dynamics pertaining to contacts between infected and non-infected individuals. This is hard to measure even though the German Corona Warning App is an important approach that helps map and trace contacts.

We also learn a lot about the relationship between infectious disease dynamics and human behavior. There is a strong correlation. Epidemic models describe the disease as a dynamic process with rising or falling infection rates and constant human behavior. Meanwhile, a pandemic lasts longer and is more dangerous. Policymakers respond, people wear masks. Then the number of cases is decreasing, restrictions are eased, making some people more careless. It is an interaction we must learn to fully understand. Digital tools play an important role in this situation because they can map the behavioral changes as it pertains to mobility, for example.

I think the most important insight is this: a pandemic is a behavioral phenomenon, even more so than it is a virological or medical challenge.

What is in store for the future of digital epidemiology?

Brockmann: It is a growing field of research that will become an important part of science just like bioinformatics. An infectious disease is a great way to explain this concept: You have the human being as the host and the virus as the intruder. We need to fully understand this system. Bioinformatics already plays a key role in understanding the biochemistry of the virus.

This is not yet the case from the host’s perspective. This setting is about psychology, behavioral research, and social networking. The interface between computer sciences and social sciences is only now emerging. It is also an area of epidemiology we must delve into: we must understand and analyze the behavioral patterns of populations and individuals.

If you are currently studying social sciences or epidemiology, you do not automatically learn computer programming. But that day will come, and it may become a standard like it is in the field of bioinformatics.