According to MRC's Report 9: Impact of non-pharmaceutical interventions (NPIs) to reduce COVID-19 mortality and healthcare demand, Imperial's MRC Centre is using an individual-based, simulation transmission model -- apparently a modification of the work discussed in the article, Modeling targeted layered containment of an influenza pandemic in the United States.
It's clear from Report 9 that the MRC Centre has considered quite a number of factors in their model, including heterogeneity of US and UK populations. On the other hand, the authors dramatically revised their policy recommendations in Report 9, as they updated parameters of the model. In fact, it's worth quoting a few paragraphs from their report [emphasis added]:
Perhaps our most significant conclusion is that mitigation is unlikely to be feasible without emergency surge capacity limits of the UK and US healthcare systems being exceeded many times over. In the most effective mitigation strategy examined, which leads to a single, relatively short epidemic (case isolation, household quarantine and social distancing of the elderly), the surge limits for both general ward and ICU beds would be exceeded by at least 8-fold under the more optimistic scenario for critical care requirements that we examined. In addition, even if all patients were able to be treated, we predict there would still be in the order of 250,000 deaths in GB, and 1.1-1.2 million in the US.
In the UK, this conclusion has only been reached in the last few days, with the refinement of estimates of likely ICU demand due to COVID-19 based on experience in Italy and the UK (previous planning estimates assumed half the demand now estimated) and with the NHS providing increasing certainty around the limits of hospital surge capacity.
We therefore conclude that epidemic suppression is the only viable strategy at the current time. The social and economic effects of the measures which are needed to achieve this policy goal will be profound. Many countries have adopted such measures already, but even those countries at an earlier stage of their epidemic (such as the UK) will need to do so imminently.
In switching from a mitigation strategy to a suppression strategy, the MRC Centre gives the impression that it's model isn't particularly robust to the values of its input parameters.
I'm not sufficiently familiar with the details of their model to offer a reasoned critique. On the other hand, switching from a mitigation strategy to a suppression strategy is a very significant change, and I come away from reading their research reports with the impression that their model isn't as robust as they thought.
George Box is attributed with the phrase, "All models are wrong; some are useful." With this in mind, my concern with the relatively sophisticated efforts of Imperial's MRC Centre is that their model may be too nuanced and sophisticated to be genuinely useful in practice.
In my own efforts to model the incidence of the virus, I've used the model that I understand is the simplest of the deterministic models used in epidemiology: the susceptible-infected-removed (SIR) model. In particular, it captures the characteristics of the infection curve that appear to be most important:
- an initial period of near-exponential growth
- followed by an inflection point, when the second derivative is zero
- following by a peak in infections, when the first derivative is zero
- following by a decline in the number of infections.
The SIR model has a small number of parameters, particularly the basic reproduction number, R0, of the epidemic, which can be used to evaluate various scenarios for transmission.
In the basic SIR model, individuals move from being susceptible to being infected to being removed (recovered, immunized, or deceased), according to a series of ordinary differential equations. The rate at which individuals move from being susceptible to infected is determined by the reproduction number, R0, while the rate at which individuals move from being infected to removed is determined by a second parameter. For more on this model, see Compartmental models in epidemiology, and A Contribution to the Mathematical Theory of Epidemics.
I'm not suggesting that this simple model is the best one can use to model the COVID-19 pandemic. Surely, practicing epidemiologists have models that are more useful for this purpose. But for the purposes of a financial analyst or an investment strategist trying to capture the key characteristics of viral transmission, the basic SIR model seems to provide a useful trade-off between parsimony and robustness.
For readers interested in going beyond this basic model, an online Epidemic Calculator using a Susceptible-Exposed-Infected-Removed (SEIR) model is available online.
A very well-done online COVID simulator has been produced by Martin Eichner, Stefan Brockmann, and Markus Schwehm.
While these models vary in sophistication, they all capture the basic characteristics of viral transmission discussed above. For the purpose of financial analysts, a useful check when using such models is to compare the results produced by various models, to assess the extent to which relevant inferences are robust to the choice of model.
No comments:
Post a Comment