[from the duuuhhh files~cr]
LUND UNIVERSITY
Research news
In a recent study, researchers at Imperial College London developed a model to assess the effects of various measures to contain the spread of the coronavirus. However, the model had fundamental flaws and cannot be used to draw the published conclusions, Swedish researchers from Lund University and other institutions claim in the journal Nature.
WATCH: Three reasons why mathematical models could not predict the spread of the coronavirus – https://www.youtube.com/watch?v=nwT8_CyIcSI
Imperial results showed that it was almost entirely the complete social lockdown that suppressed the wave of infections in Europe in the spring.
The study estimated the impact of various measures such as social distancing, self-isolation, closing schools, banning public events, and locking yourself.
“Since the measures were implemented over a few weeks at roughly the same time in March, the mortality data used simply does not contain enough information to distinguish their individual effects. We demonstrated this through a mathematical analysis. On this basis, we then ran simulations with the original code from Imperial College to illustrate how the sensitivity of the model leads to unreliable results, ”explains Kristian Soltesz, Associate Professor of Automatic Control at Lund University and first author of the article.
The group’s interest in the Imperial College model was piqued by the fact that it explained almost all of the reduction in transmission during the spring from lockdowns in ten of the eleven countries modeled. The exception was Sweden, which never imposed a lockdown.
“In Sweden the model offered a completely different measure to explain the reduction – a measure that seemed almost ineffective in the other countries. It seemed almost too good to be true that an effective lockdown was in place in every country but one, while another measure appeared unusually effective in that country, ”notes Soltesz.
Soltesz carefully points out that it is entirely plausible that individual measures had an effect, but the model could not be used to determine their effectiveness.
“The various interventions do not seem to work in isolation from one another, but are often interdependent. A change in behavior as a result of one intervention influences the effect of other interventions. How much and how is more difficult to know and requires different skills and collaboration, ”says Anna Jöud, Associate Professor of Epidemiology at Lund University and co-author of the study.
Analysis of models from Imperial College and others shows the importance of reviewing epidemiological models, the authors say.
“There is a focus in the debate about data sources and their reliability, but there is almost no systematic review of the sensitivity of different models in relation to parameters and data. This is just as important, especially when governments around the world use dynamic models as a basis for decision-making, ”emphasize Soltesz and Jöud.
The first step is to perform a correct analysis of the model’s sensitivities. If they are too big of a problem, more reliable data is needed, often combined with a less complex model structure.
“When much is at stake, it is wise to be humble in the face of fundamental restrictions. Dynamic models can be used as long as they take into account the uncertainty of the assumptions on which they are based and the data on which they are guided. If this is not the case, the results correspond to the assumptions or assumptions, ”concludes Soltesz.
###
From EurekAlert!
4.4
5
be right
Item rating
Like this:
Loading…
