November 16, 2023

expert reaction to study looking at an epidemic-economic model of Covid based on data from New York city

A study published in Nature Human Behaviour looks at an epidemic-economic model of COVID. 

Prof Katharina Hauck, Professor in Health Economics; and Deputy Director of the Abdul Latif Jameel Institute for Disease and Emergency Analytics, MRC Centre for Global Infectious Disease Analysis, Imperial College London, said:

“The press release and the findings presented overall correctly reflects the modelling presented in the paper, and they confirm findings from other similar papers on the economic and epidemiological.  There are a few caveats to the analysis:

“1. The model is parameterized with and calibrated to data from one specific metropolitan area, the New York metro area.  While it is plausible that it can be applied to some other urban centres in high-income countries, it is perhaps less generalizable to urban centres in South-East Asia (e.g. Singapore) because of different economic structures, pandemic mitigation, and behavioural effects.  It is also less generalizable to low- and middle-income countries, because of a much higher percentage of informal employment, younger age demographics, and other factors.

“2. Behaviour is modelled in a simplified fashion, as a function of deaths, and it is not informed by actual data (such as a survey).  Modelling of behavioural effects is actually quite often done in integrated economic-epidemiological models, I am aware of 31 models which modelled behaviour in some form.  Behaviour during subsequent waves may be very different than during the first wave modelled here.

“3. The authors did not use actual economic data for New York, but generated those from national data with the help of an additional analysis.

“4. It is great to see a strong contribution to the literature on integrated economic-epidemiological modelling. Several teams around the world are working on balancing health, economic, and social objectives in pandemic mitigation, a recent scoping review found 80 papers on integrated modelling.

“5. The paper also claims that it is the only agent based integrated economic-epidemiological model.  I am aware of two others: 1. arXiv:2011.06289 (Mellacher. Covid-town: An integared economic-epidemiological agent-based model), and 2. https://www.sciencedirect.com/science/article/pii/S0960077920304859

“6. The limitation of this model (and those of similar integrated models) is that we essentially have very little data about contacts at the workplace – this applies to contacts amongst workers, between workers and consumers, and between consumers themselves, at places where goods are purchased and services consumed.  Contacts are estimated from mobility data from users who opted in to be traced, which may results in a bias to the sample, and some other problems related to mobility data.

“7. In the press release it is suggested that lockdowns and behaviour change lead to similar outcomes and trade-offs between health and the economy.  I am not convinced this is supported by the findings of the paper, which states on behaviour change: ‘This calibrated value, which merges NPI effects with behaviour change, cannot be causally interpreted’.

“The strengths of the paper (compared to other integrated economic-epidemiological models) are that:

“1. It is an agent based model (there are only two others we are aware of).

“2. It is fitted to actual data from the first wave in NY metro area.

“3. It considers the input-output network of intermediates that industries use to produce final goods and services (I am aware of only 8 other models that do that).

“4. It models that less productive industries are more likely to reduce output during the pandemic.

“5. It allows for the differential impact of unemployment for low and high-income workers, and it assess economic loss across the income distribution.  Most other papers just investigate impact on overall GDP.  This is a real strength. 

“Overall, it is great to see a further contribution to advance the science on integrated economic-epidemiological modelling of pandemics.  We have seen how important interdisciplinary approaches are to understanding the impact of COVID-19, and to the design of mitigation policies that balance health, economic and social objectives.”

Prof Mark Woolhouse, Professor of Infectious Disease Epidemiology, University of Edinburgh, said:

“The absence of a framework for exploring the relationship between public health and economic impacts was a weakness during the Covid pandemic.  This new study by Pangallo et al. is an interesting retrospective attempt to fill this important gap.

“The authors use an agent-based model to simulate the behaviour of over 400,000 individuals and the health and economic consequences of those behaviours.  The model is calibrated to be representative of the first wave in New York City and uses real world data on economic activity and health outcomes.  It was then used to explore counterfactuals scenarios.

“The model is one of the first to allow that people may change their behaviour spontaneously, from fear of infection and its consequences.  There is good evidence that this happened during the pandemic but it was not incorporated into the models being used at the time.

“As with all models, this study simplifies a complex set of epidemiological, behavioural and economic processes.  Nevertheless, it is an advance on previous work of this kind, particularly since it uses (anonymised) data collected from New York residents at the time.

“A key finding is that there is a tradeoff between lives and livelihoods.  Harsher restrictions have a public health benefit but come at a cost.  This seemingly obvious conclusion was hotly disputed during the pandemic, particularly by those arguing for lockdowns.

“Another key finding is that the burden was felt far more by low than high income workers, mainly because remote working was not possible, so low income workers faced higher risks of infection or economic hardship.  This adds to a growing evidence base that the pandemic, and our response to it, exacerbated existing inequalities.

“The authors conclude that untargeted, blanket lockdowns were a suboptimal response to Covid.  They also show that both health and economic harms are minimised by acting quickly, in keeping with the maxim that early intervention can be less drastic intervention.

“During the pandemic in the US, UK and many other countries, the debate about whether to lock down or not became highly polarised.  This left little room for anyone to argue for the middle ground, but the potential for harms caused by Covid and by lockdown were both so great that getting the balance right was crucial.  This new study not only adds weight to that viewpoint but goes further by offering a framework for making better decisions next time.”

‘The unequal effects of the health–economy trade-off during the COVID-19 pandemic’ by Marco Pangallo et al. was published in Nature Human Behaviour at 16:00 UK time on Thursday 16 November 2023.

DOI: 10.1038/s41562-023-01747-x

Declared interests

Prof Katharina Hauck: “CoI over the past year:

I have undertaken paid consultancies for WHO related to the topic of the paper (via my consultancy Tyra Consulting Ltd where I am sole shareholder and director);

I have undertaken paid advisory roles for Pfizer and GSK for topics unrelated to the paper.”

Prof Mark Woolhouse: “I was an advisor to the UK and Scottish Governments during the pandemic I am author of a book The Year The World Went Mad.”