Biostatistics: using data and models to fight Covid-19

We are all familiar with these images from some countries: Completely full intensive care units, doctors working frantically despite being ill, being forced to decide who lives and who dies due to critical shortages of respirators. How can you forecast Covid-19 impact on healthcare systems to avoid overload? Biostatistician Professor Frank Klawonn studies this issue.

At virtual.MEDICA 2020, he presented a lecture titled "How to predict the potential overload of health systems by COVID-19" in the virtual. MEDICA LABMED FORUM. In this MEDICA-tradefair.com interview, he talks about the role of biostatistics in the Covid-19 pandemic and reveals successful measures to manage the crisis.

Professor Klawonn, your research focuses on biostatistics. What is meant by that?

Prof. Frank Klawonn: Biostatistics are the statistical processes and methods applied to the collection, analysis, and interpretation of biological data and especially data relating to medicine. Statistical methods principally have general validity, but medicine presents unique challenges. Unlike technical systems, there are limited options to collect data and conduct experiments based on ethical issues alone. You can study identical technical systems. Apart from identical twins, this is not possible when it comes to human beings. We also frequently deal with a small number of cases but many variables, such as genes for example. Think of "personalized medicine" in this setting, which requires the analysis of ever smaller groups with more and more variables. My research group explores this set of problems.

How important is this approach in the current situation?

Klawonn: Statistical methods allow you to quantify the uncertainty associated with simulation models, which are used to forecast the spread of infection. However, the quality of forecasting depends largely on the data as there are many parameters to consider. For example, the reported daily new infection rates also depend on the number and types of tests. When you primarily test symptomatic individuals, the number of reported cases is obviously lower but yields a higher percentage of those who test positive compared to routine testing of specific occupational groups. Unfortunately, there is a lack of data on how many tests are attributed to preventive screening, how many positive tests that have been carried out in asymptomatic contacts and those tests that were performed due to symptoms. Data on infection clusters is also not directly available.

How can this information be used to predict the potential overload of health systems?

Klawonn: All things considered, I think the available data only permits limited effective forecasting. Having said that, statistical methods and modeling are still great approaches to track trends, which is especially important given that there is a one-to-two-week delay before we can see the initial impact of any taken measures on infection rates. The time horizon for the intensive care situation is even longer. People affected by COVID-19 who require intensive care or ventilators often need long-term treatment before they can be transferred to a regular ward or – in the worst-case scenario – pass away. This means that even after a hard lockdown, the intensive care units will continue to fill up and exhibit high hospitalization and utilization rates for a long time to come.

How can we use the results of your research? What need for action do you see?

Klawonn: I deal with the immediate impact of the Covid-19 pandemic, which includes the number of infected people, intensive care bed requirements, fatality rates and seroprevalence. Policymakers should incorporate the research findings pertaining to these factors – which they actually do. However, all aspects that others study intensively must ultimately also be factored into the political decision-making process. This entails the effects on the economy, schools and childcare providers and the impact on children and parents, postponed screening tests and elective surgeries due to critical bed shortages, psychological and social aspects and so much more. Politicians must put all these individual components together, assess and prioritize them, and finally make the right decisions. I am glad that this is not my personal responsibility. I have little experience in these non-medical fields, which prevents me from identifying and recommending specific action steps.

How can we successfully tackle the coronavirus pandemic in your opinion?

Klawonn: I am not really equipped to give any advice on this issue. If you were to look at just the number of infections or severe cases, you would have to declare nationwide hard lockdowns, announce temporary border closures, and have stringent contact tracing. Fortunately, this is not a feasible option for us in Germany. However, you should definitely avoid overloaded healthcare systems to prevent triage protocols and images like the ones we saw in Bergamo, Italy. Unfortunately, this also means that we will likely have to rely on short-term strategies for the time being and avoid a premature relaxation of social distancing measures and – in some cases – be willing to temporarily tighten restrictions again, at least on a regional level.

We know what we can accomplish from a medical viewpoint and understand the impact of certain measures on the pandemic. It is crucial to consider the effects this has on all other areas of our lives and mitigate them with corresponding countermeasures. It is relatively easy – albeit expensive – to achieve this from an economic perspective. However, things get more complicated when it comes to education as it affects daycare centers, schools, and colleges. The arrival of vaccines offers hope, but we still need a lot of patience and a willingness to accept restrictions and trade-offs until enough people have been vaccinated against the coronavirus.