The Power of Lockdown in Combatting COVID-19
How Public Safety Measures Are Helping Reduce the Spread of COVID-19 in the United States
This post is a shortened version of a preprint we have placed on arXiv. For more details on the analysis, the full paper is available here: https://arxiv.org/abs/2004.03200
I, along with my co-author, Tarun Amla, conducted statistical analysis, using a causal model, to predict the spread of COVID-19, in the United States, as impacted by the implementation of stay-at-home orders and lockdown measures, among most states.
Countries all around the world have been working to try to combat the COVID-19 pandemic through public safety measures, which predominantly involve promoting improved hygiene practices, like frequent and lengthy hand washing, enactment of social distancing measures, such as guidelines for keeping more than six feet apart from others and cancellation of large events and bans on gatherings of more than a few people, and, stay-at-home or lockdown orders, requiring the closing of non-essential businesses and banning outdoor movement, except as it pertains to health, sustenance necessities, or those businesses deemed essential. This is in addition to schools and universities having been closed, and most workplaces shifting exclusively to a work-from-home model.
The United States has become the worst hit in terms of the number of cases, and, as a result, has enacted many of these public safety measures. Though there is no nationwide lockdown, most states have issued stay-at-home orders and have closed all non-essential businesses, in addition to social distancing measures.
The current social distancing and lockdown measures are vital to our society’s survival and we simply wished to provide a brief, simple analysis to make it clear why that is and to implore people to obey these measures. We further argue that increased social distancing measures can serve to further prevent this disease from getting completely out of control.
We are not epidemiologists, but rather engineers. The bulk of what is being pushed by experts in the field, however, is proving inaccurate due to the reliance on analogy to previous outbreaks of similar viruses, and also ineffective in making the case for the importance of continued implementation and obedience of these public health measures.
Our analysis was conducted by using in our training set data from the COVID Tracking Project from the beginning of the outbreak until 04/03/2020, which we broke into two parts, one representing the days prior to the impact of widespread lockdown measures, and the other representing the days since. The first data set was simply modeled as exponential growth (not accounting for any damping pressures, as we do not project it out to a point where predator-prey or other population level forces become relevant), as a benchmark for expected trends in the absence of lockdown and social distancing measures. We modeled the second set of data as an exponential decay of the growth rate, to represent the impact of lockdown measures on reducing the spread of COVID-19.
In the above figure (Figure 1), we show that, according to our model the peak number of daily infections should be ~39k and this peak should be hit on April 09, 2020. Additionally, we can see that our model predicts this pandemic to end with a total of ~1.35M infections, as opposed to the no lockdown model, which would have exceeded that number and continued unimpeded by April 02, 2020.
To predict hospitalizations, we treated hospital arrivals as a fixed percentage of new infections. This was found to be 13.24%. Additionally, we treated hospital departures (which included both recovery and death) by calculating the mean cycle time, according to Little’s Law. From this, we found a mean hospital stay of 6 days.
As shown in figure 2 above, our model estimates a peak of ~31k hospitalizations that should occur on April 11, 2020.
Lastly, to predict the number of deaths, we treated a fixed percentage of hospitalizations (which was found to be 40%), along with the mean hospital stay cycle time to apply a lag.
As we can see from figure 3, we anticipate a peak of ~2.1k deaths on April 15, 2020 and a total number of ~72k deaths over the course of the pandemic.
The White House Coronavirus Taskforce was predicting far higher death tolls of between 100k and 240k, but our analysis indicates that the stay-at-home and lockdown measures are working.
For a thorough accounting of our methodology and a more in-depth review of our results, please see our preprint on arXiv.
What is very clear from this analysis is that the effect of the stay-at-home and lockdown measures is profound and the absence of them, or disobeying them, is devastating. The current measures are working and must be obeyed and maintained (and ideally expanded) to save lives.
Bootsma and Ferguson, in a paper studying the impact of the United States public health and safety measures taken in response to the outbreak of the 1918 Spanish flu in Europe, found a profound link between the speed with which such measures are implemented, how long and effectively they are maintained, and how effective they are in mitigating the spread and death rate caused by the pandemic. As a result, the United States did not suffer nearly as terrible of consequences as Europe did during that pandemic.
A 60 day nationwide lockdown would most likely see us through the COVID-19 pandemic. Maintaining of general social distancing practices and use of face-masks, and improved hygiene habits (washing hands frequently, avoiding physical contact, etc.) for a period of time thereafter, upon lifting of lockdown, would allow us to transition back to a new normal, while allowing us to limit the level of the overall outbreak, as best we can.
Ideally, this period of continued social distancing and general precautions would last until at least a vaccine has been developed and is readily available, or, similarly, an adequate anti-viral pharmaceutical treatment or cure is on the market and accessible to the masses. For more specificity on what that should look like, we would defer to the appropriate experts and authorities, namely the CDC and the United States Federal Government once we reach that juncture.
