Imperial College London’s Model of Coronavirus Measures
Researchers at Imperial College London developed a model to assess the impact of various measures implemented to control the spread of the coronavirus. However, a study published in the journal Nature by Swedish researchers from Lund University and other institutions argues that the model has fundamental flaws and cannot accurately describe the conclusions drawn from it.
The Effectiveness of Lockdowns in Europe
According to Imperial’s findings, it was primarily the implementation of complete societal lockdowns that suppressed the wave of infections in Europe during the spring.
Assessing Different Measures
The study examined the effects of various measures such as social distancing, banning public events, closing schools, self-isolation, and the lockdown itself.
Kristian Soltesz, the first author of the article and associate professor at Lund University, explains, “The measures were implemented at about the same time over a few weeks in March. Due to this, the mortality data used does not provide enough information to distinguish their individual effects. We have demonstrated this through mathematical analysis and simulations using Imperial College’s original code, which revealed how the model’s sensitivity leads to misleading results.”
The Curiosity Surrounding the Imperial College Model
The Swedish researchers became interested in the Imperial College model because it attributed almost all of the reduction in transmission during the spring to lockdowns in ten out of eleven modeled countries. The exception was Sweden, which did not implement a lockdown.
Soltesz notes, “The model presented a different explanation for the reduction in Sweden, which appeared to be almost ineffective in other countries. It seemed too good to be true that an effective lockdown was in place in every country except one. At the same time, another measure appeared surprisingly effective in Sweden.”
The Complexity of Assessing Individual Measures
The researchers emphasize that individual measures do not seem to work in isolation but are often interdependent. Behavioral changes resulting from one intervention can influence the effectiveness of other interventions. Anna Jöud, co-author of the study and associate professor at Lund University, states, “How much and in what form these interventions influence each other is difficult to determine and requires collaboration and various skills.”
The Importance of Reviewing Epidemiological Models
According to the authors, analyzing models like the one developed by Imperial College is crucial. They highlight the need for a systematic review of the sensitivity of different models to parameters and data. This is especially important as governments worldwide rely on dynamic models for decision-making.
Considering Limitations and Uncertainty
The researchers stress the importance of acknowledging the limitations and uncertainties inherent in dynamic models. They suggest conducting thorough analyses of model sensitivities and, if necessary, using more reliable data and simpler model structures.
Soltesz concludes, “A lot is at stake, so it is wise to be humble when faced with fundamental limitations. Dynamic models are useful as long as they consider the uncertainty of assumptions and the data supporting them. Otherwise, the results are no better than mere assumptions or guesses.”