Covid-19 solutions through data

Data as a guide to balance societal trade-offs in COVID-19 epidemic

11th of April 2020

A series on understanding and misunderstanding of the SARS-Cov-19 virus spread (the main cause of COVID-19 disease) through data, statistics and modelling.

Table of Contents

Why advanced modelling does not help (yet)?

Not having sufficient data means garbage-in, garbage-out in a data science slang. The available data is severely incomplete. There is important data that is missing and without which simple math is misleading and advanced modelling provides only guesses. Locking down countries based on guesses can be deadly too.

The available data on country level consists only of those who experienced COVID-19 or flu related symptoms severe enough for health authorities to invite an individual for testing. In some countries only those who required hospitalisation were tested, e.g., USA. In some countries testing is free for everyone with symptoms and easily accessible for those without symptoms, e.g., South Korea.

The data that is missing is on those who got infected by the virus, but did not develop any symptoms or at least not severe enough to be tested for the virus. The evidence from Cruise ship Diamond Princess (an example of a specific population where everyone was tested) shows that 46% of all infected didn’t experience any symptoms, not even a month after testing.

THE SCIENTIFIC FACT IS THAT WE DON’T KNOW YET NEITHER HOW INFECTIOUS THE VIRUS IS NOR HOW LETHAL THE DISEASE IS.

How to get the missing data?

By testing individuals for antibodies of SARS-Cov-19 virus (the cause of COVID-19 disease) on a representative sample of chosen population. Such surveying will enable to answer, scientifically, the research question: How infectious the virus is and how lethal the disease is?

A societal trade-offs and an appeal to researchers and governments

It might sound expensive to do testing for antibodies on a representative sample, but it is a small fraction of the costs created by wrong policy measures to curb the epidemic. We should not follow data if it is known that is biased and if it is known that the bias cannot be removed without doing additional studies. We should be honest with this, do additional studies, collect the necessary data and make data-based decisions after. Governmental decisions should not be based on results obtained by modelling incomplete data, because results based on incomplete data are mere guesses.

We should make all the efforts to collect as soon as possible the missing data. The developed countries should play a role in this endeavour since they have the required resources, not only financial, but also in terms of data registers that can support a development of an efficient population level representative sampling design to conduct testing for antibodies of SARS-Cov-19 virus. By doing this, we will be able to collect the required data and provide scientific answer to the question: How infections the virus is and how lethal the disease is?

We should not rely on any modelling approaches without having data enabling to accurately estimate this information. Data-wise, this is the core information to be able reliably model possible scenarios that policy makers could use for decision-making.

A country like Finland should make all the efforts to quickly implement such a study since it has the best data registers in Europe to implement it quickly and with minimum costs. Finland should play a role in answering this scientific question however all countries are encouraged to conduct such studies too and look for the specificity of the environment they are in, for example proximity to ‘hot zones’ such as Italy and Spain in Europe.

If researchers act together swiftly, a scientific design for such testing and implementation can be done in three weeks. It is mostly time to organise everything. Experienced researchers in survey methodology can prepare the design and survey protocols in a matter of days and from then on is just a matter of implementation – actual testing and data collection. From that moment on, with information of such data, modelling and simple math becomes informative.

Statistics for Understanding – Statistics for Reliable Solutions – Statistics for Helping

Tarastats Statistical Consultancy | Fredrikinkatu 61A, 00100 Helsinki, Finland | www.tarastats.com | info@tarastats.com

 

Tarastats Statistical Consultancy © 2020