Five years ago, SARS CoV-2 swept through Northwestern Europe and took us by surprise. Every country responded with lockdowns, closed school, and a series of other measures – but each acted in slightly different ways. And in every country – definitely in mine – there were always those convinced that our neighbouring countries had handled the crises more effectively.

How can we scientifically evaluate which countries responded well, and who did better or worse? You can’t run a randomised trial, and surveillance data – such as mortality rates – may be influenced by many other factors beyond response measures alone. While handwaving can be used, the method that may bring us closest to the truth is probably mathematical modelling.

Mathematical Modelling: A More Objective Lens
In a just published study, Dutch modellers used an existing model and observed mortality data from the Netherlands (NL), Belgium (BE), Germany GE), Denmark (DK), Sweden (SW) and the UK from the first COVID-19 pandemic wave (13 March – 1 July 2020). They calculated R-values just before measures were implemented (Rt), and the speed and depth of the decline in Rt after the implementation of measures, which reflects the effectiveness of the set of response measures in each country.

They then quantified the impact of counterfactual strategies to estimate, what would have happened to Rt and the number of deaths in the Netherlands – in the model – if we had adopted the exact same strategies as Sweden or Belgium?

Observed cumulative deaths per million inhabitants were 105, 108, 361, 546, 599 and 840 for DK, GE, NL, SW, UK and BE. Calculated Rt values ranged from 3.7 (NL and BE) to 3.2 (DK). After the implementation of measures in the Netherlands, the Rt dropped to <1 in 2 weeks, and eventually to 0.6, reflecting the fastest and deepest decline among the 6 countries.

Now what if measures from other countries (with slightly different Rt values and different declines in Rt) had been implemented instead? The answer comes as a surprise: The Dutch approach – according to the model – would have reduced the cumulative number of deaths in all other countries. Vice versa, implementing other countries’ measures in the Netherlands would have (dramatically) increased our number of deaths.

Other relevant findings: “In a fast-growing epidemic, small differences in the timing and effectiveness of measures can result in large variations in deaths… Due to differences in country characteristics and initial epidemiological situations, the outcome of the response in a particular country does not necessarily result in the same mortality as in another country; a response must always be tailor-made.”

So, this information is extremely useful for Dutch politicians, who are right now launching a parliamentary inquiry into the Netherlands’s pandemic response. Some of these politicians pursue tribunals for our policymakers (and scientists, I guess). I trust that they will meticulously analyse the methods of this study, that was co-authored by our national COVID-19 modeller, Jakko Wallinga.