Even as the world struggles to live and make a living, faced with the dual threat of a rapidly spreading viral illness and economic slowdown, a new study published on the preprint server medRxiv* in May 2020 says that relaxing non-pharmacological interventions will allow the SARS-CoV-2 virus to become endemic.
When faced with a new infectious illness, it is of the first importance to understand how it is caused and how it spreads. The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has a complex pattern of spread, which makes mathematical modeling a useful tool to analyze it.
Undocumented Infection Contributes to Spread
The new paper by Benjamin U. Hoffman of Columbia University presents an epidemiological model that can predict the spread, the extent of sickness, and the death rate of the virus in New York State. It includes important measures such as the extent of invisible infections and their effect on the rate of spread, the development of robust immunity, weather-related changes in viral transmissibility, and how non-pharmacological interventions (NPIs) help to slow down and reduce the amplitude of the outbreak.
NEW YORK - APRIL 01, 2020: A long line outside of Whole Foods in Tribeca, New York as the store has implemented social distancing measures during the COVID-19 pandemic. Image Credit: Jennifer M. Mason / Shutterstock
This news article was a review of a preliminary scientific report that had not undergone peer-review at the time of publication. Since its initial publication, the scientific report has now been peer reviewed and accepted for publication in a Scientific Journal. Links to the preliminary and peer-reviewed reports are available in the Sources section at the bottom of this article. View Sources
One of the most important findings of the current study is that the immense extent of undocumented spread drove the rapid spread of the virus in New York State. The model provides a reproduction number R0 of 5.7 at the peak, which is far above the rate of spread reported anywhere in the world. This means that each infected person spread the virus to 5-6 other people.
The model also predicts that by July 2020, the R0 will reach a baseline level of 4.4, which is still almost double that reported for the USA, as well as for Wuhan and other regions of China (2.2-2.6). This surprisingly high R0 is due to the vast number of undocumented infections that were present in New York State when the outbreak began.
This figure was calculated on the basis of serologic testing for SARS-CoV-2 antibodies, which showed that despite a reported caseload of about 2% of the population in New York City, the prevalence of infection was about 15% to 20%. This led to the estimation of a 75% undocumented infection rate.
Using this parameter, the model estimates that New York State has experienced almost 700,000 undocumented infections by April 28, 2020, even though the Department of Health, New York State, reports only about 300,000 confirmed cases. This agrees with the findings of another recent study, which showed that the initial rapid spread of the virus across China was due to the large pool of untraced infections.
The presence of a vast and invisible reservoir of untraced infection explains why the outbreak in this state grew with such incredible speed. Individuals who did not know they were infected continued to mingle with those who were still healthy, rather than self-isolating. As a result, many more contacts were infected with the virus.
The initiation of social distancing measures in New York State brought about an effective reduction in the viral spread. These will have to continue in force for years, however, to avoid the repetition of large outbreaks of illness that will put hospitals equipped for tertiary care to the severest test.
Is It Possible to Relax NPIs?
The model also gives the assurance that social distancing can be carefully reduced by about 30% of the current levels without a marked rise in the rate of infection and death due to COVID-19. The model shows a minimal increase of 2,000 deaths by September 1, 2020, in such a scenario. On the other hand, a reduction by more than 50% will cause a drastic surge in sickness and more than doubling of deaths, with an effective instantaneous reproduction number R(t) greater than 1.
R(t) is different from the basic reproduction number R0 in that it describes the susceptible fraction of the population over time, and thus helps to find the potential for a population outbreak in a mixed group of susceptibles and resistant individuals, over time.
As a result, NPIs must include, besides continued social distancing, better testing, isolation of cases, and tracing of contacts, as many countries have demonstrated across the world already. The model also demonstrates the need to pursue phased relaxation of NPIs.
Using long-term simulations, with the social distancing parameter consistently kept above 0%, the model projected the result of a steadily maintained improvement in NPIs. The results can be used to predict the course of an outbreak in any city anywhere in the world, though it is built on the data gathered from the current New York State outbreak.
The Effect of Crowding
Since the rate of spread of infection is very probably in proportion to the population density, the model uses this relationship to make this type of extrapolation easier, simply by introducing the right population density for the city or region or even country that is being studied.
To illustrate this, the researchers contrast the rapid spread within New York State compared to the much slower and smaller outbreak in San Francisco’s Bay Area. One reason may have been that the latter went under home isolation orders a week in advance, on March 16, 2020, compared to March 22, 2020, for New York State.
However, another major reason is the higher population density in the latter, at 2.700 people per square mile in New York State counties included in the current study compared to 1.800 in the San Francisco Bay Area counties, namely, San Francisco, San Mateo, Santa Clara, Alameda, and Marin.
The model would thus correctly predict a lower chance of transmission in the latter area. Thus, the use of such a model will help to understand the changing and differential transmission patterns of the virus in both US and other international cities.
Seasonal changes in the rate of infectivity are also incorporated into the current model because of earlier reports showing the effects of such change on the endemicity of infection.
The single model incorporating all these variables is important in that it provides a means to analyze the effects they cause on both long- and short-term SARS-CoV-2 transmission.
Limitations of the Model
The current model has its limitations, one being that all the data came from the New York State Department of Health. Errors in such data will lead to over- or under-estimation of the true characteristics of the outbreak. Moreover, the poor level of testing in the first part of the New York outbreak could have led to a serious underestimation of cases, tilting the conclusions towards higher viral infectivity.
The use of fixed values for several variables, such as the mean incubation period, the time to hospitalization, and the effective rate of undocumented infections, could also introduce error. The values were chosen based on literature but may vary with time.
A third potential issue is the use of seasonal variability in other viruses to model that of the current virus, using a conservative measure since the actual effect of temperature and humidity on the spread of SARS-CoV-2 is still to be known. This may significantly affect the actual seasonal variation in the spread.
Finally, the effect of introducing an effective vaccine or drug is not depicted in the current model since these are only on the horizon and, at best, may still be a year away. While the level of confidence in estimating when they will become available is still zero, they are likely to be the most critical interventions to fight or prevent endemic infection worldwide, and in the long run.
Forming Stable and Sensible Policies
The model thus uncovers the most important characteristics of SARS-CoV-2, which will help to inform policies against the virus in the short-term until the emergence of effective pharmacological interventions. The simulations confirm that while NPIs are very useful in reducing sickness and death over the short term, it is only through the development of durable immunity in over 70% of the population that the virus can be prevented from becoming endemic.
Despite resuming NPIs over the winter of 2020/21, the endemic spread is predicted in New York State over the next five years without an effective vaccine or drug
In their absence, says the paper, “These data predict a significant second outbreak in early 2021 that can be mitigated, but not avoided entirely, through the resumption of strong social distancing measures.”
The author concludes that the spread of the virus can be controlled only by social distancing and other non-pharmacological measures that change with the situation but operate to keep people safely apart. Secondly, says the paper, the virus is likely to cause endemic infection if the infection does not cause durable immunity or if an effective vaccine is not found.
This news article was a review of a preliminary scientific report that had not undergone peer-review at the time of publication. Since its initial publication, the scientific report has now been peer reviewed and accepted for publication in a Scientific Journal. Links to the preliminary and peer-reviewed reports are available in the Sources section at the bottom of this article. View Sources
Article Revisions
- Mar 7 2023 - The preprint preliminary research paper that this article was based upon was accepted for publication in a peer-reviewed Scientific Journal. This article was edited accordingly to include a link to the final peer-reviewed paper, now shown in the sources section.
High mutation count in new COVID-19 variants does not increase immune evasion, study finds