Coronavirus

Covid-19 Is All About the Curve

Weizmann Institute molecular geneticist Prof. Doron Lancet thinks the nature of the virus, more than the lockdown, is causing infections to fall

Globes • TAGS: Virus, Biology, Culture, Mathematics

 

Prof. Doron Lancet of the Department of Molecular Genetics at the Weizmann Institute of Science says that the Covid-19 curve has developed in the same trajectory in many countries regardless of their lockdown policies. By his reckoning, the significant statistic is not the number of cases but the pattern of development, which is similar in every country including Israel and is currently falling. According to Lancet, it was not the lockdown that led to the change in direction in Israel, or at least not just the lockdown, but something natural in the development of the disease and he insists that a comparative analysis proves this. By this logic, even the complete relaxation of the lockdown (although this is not necessarily what he recommends) would not result in the kind of figures being seen in Italy or New York.

Lancet says, In the eyes of many people the coronavirus is a mysterious monster making people very ill and even killing them. But an intelligent mathematic analysis puts things into perspective and makes it easier to cope with the virus.

All of us have heard the concept 'flattening the curve.' The origin of that curve is the SIR model (susceptible, infected, recovered and immune). In this model 'flattening the curve' means activities that will slow reaching the peak by dispersing the number of infections over a longer period of time, so that the number of patients at any given will never exceed the limit possible for hospitalization and treatment. Lets look a little differently at these statistics and check the curve of the accumulated number of patients, where the final line is ideally defined as the moment that there are no new cases (percentage growth equals zero), and the number of patients reaches the base line of the graph.

Let's say that the early and late parts of the virus behave exponentially. The meaning of exponential is that the number of patients who test positive rises every day by a certain percentage compared with the previous day. So for example in the two weeks between March 11th and 25th the number of infections rose 25% every day (a rise of 1.25). As a result, by the same accounting as compound interest, the number of patients during this period rose from 82 to 2,369, almost 30 times! But when we heard the alarming news each evening about 'a new peak in the rise in the number of infections,' we ignored the fact that this was precisely what was expected from an exponential curve! Unfortunately, most of the time, they didn't report whether the rate of increase continued to be exponential with the same steepness. If we understand the percentage change, we would see it go down over time almost like at the beginning of the virus. It's important to point out that had the multiple of the coronavirus in Israel stayed at the level of increase of 1.25 per day, then there would have been 84,000 infected today and 600-800 dead but that didn't happen. Something is causing patients to infect less people than in the past.

The Ministry of Health and the government are convinced that this happened because of isolation, social distancing and the lockdown, which they imposed in Israel relatively early in the development of the disease?

Not necessarily and I will explain which data show this. First of all it's important to stress that the graphs show a gradual fall in the steepness of the slope. Although as the weeks have gone by the number of infected patients continues to rise, but along every section growth represents a fixed fraction of the overall infected patients. Most important of all, the fraction continues to shrink, from 25% at the start to 15% by March 31 and 10% by April 5 and 4% more recently.

And wasn't that because of the lockdown?

We have here a combination of human acts such as hygiene, isolation and lockdown, with natural developments - probably the attributes of the virus and the immune systems of people. I'm strongly inclined to think that the natural reasons were dominant, and that's because of the similarity of the curve in different countries despite the extreme differences in restriction policies. The sharp fall in the steepness of the slope was not from the responses to the ‘Purim catastrophe’ and also doesn't show substantial changes in response to events such as the tightening of the lockdown.

But in fact all the countries whose curves behaved the same way are countries that used lockdown policies, except for Sweden?

Sweden is actually a case that supports my conclusions. In Sweden they took minimal isolation measures. Despite that, their infected patient growth curve fell by 35% per day in the first week to 7% after four weeks. Moreover, the number of infected patients per million residents in Sweden was slightly smaller than in much stricter countries like Israel. It's hard to see how such facts as these go together with the claim that the improvement in the growth slope for the number of cases stemmed mainly from the strict lockdown policy.

In countries where the damage done by the coronavirus was especially severe and in our country, where the damage was much smaller, the current daily rise is at a rate of a few percentage points, in other words the curves of most nations is nearing the desired level (in China and South Korea they are already there). In addition, a certain universality can be found in the comparison of the number of infected patients per million residents in various countries. The typical number is about 1,000 cases per million and that's a value that is common to very different countries in terms of size and government policies. By the way, the significance of that is that more than 99% of all people in the world have remained healthy!

In Sweden, which has a similar size population to Israel, over 2,000 have died compared with 200 in Israel. What's the difference between us other than the lockdown? Is it possible that the number of people testing positive there for Covid 19 does not reflect reality because so few tests were done?

In contrast to the surprising similarity in the behavior of the curve of infected patients and the per capita number of cases in different countries, the rate of fatalities among the overall cases is hugely different from country to country. Typical examples (as of March 26) were Germany and Israel - 0.5%, Denmark and Sweden - 2%, Belgium - 4%, Netherlands 6%, and Italy 10%. In Sweden, there has since been a steep rise in the mortality rate to 11% today, but the rate of infection and the gradual fall remains very similar to that of other countries. That is to say there is no evidence of unrealistic morbidity. It seems that the percentage of cases ending in death depends mainly on the preparedness of the health system and the dedication of staff, and it's difficult to surmise that a mechanism like the severity of the lockdown had a direct influence on this percentage.

New York City has a similar size population to Israel. But the number of cases and fatalities was far higher. Isn't that proof that if we don't reach those sort of numbers, then we won't actually see herd immunity?

New York City did indeed have ten times the morbidity compared with Israel. It's reasonable that the enormous prevalence of cases in the population in the city was connected to a failure to keep a distance and urban crowding. In extreme situations like these the high rate of mortality is a result of an absolute failure of the health system.

Going back to the natural process that according to you has restricted the development of the disease, I still don't understand what is the mechanism that explains this. Is it herd immunity?

That's an excellent question and the answer is not simple. Even with the level of infection in New York only 1.5% of the population have tested positive for the virus, and that's a long way from the numbers like 60%-70% needed for herd immunity. What is clear is that the disease restricts itself through some sort of mechanism and I leave it to the big experts to solve that riddle. One possible explanation for this mechanism is that the spread of the disease was far higher than reported with additional infected carriers being asymptomatic. There are indications that need confirmation as well as realistic testing tracks to clarify this. Alternatively, it could be that a substantial part of humanity has hereditary genetic traits that prevent the virus from attacking, which reduces the number of people vulnerable to infection and increases even more the percentage of those who in effect were infected.

In Bnei Brak too, where social distancing and the lockdown were adopted more slowly, there has been a higher level of infections and fatalities than in other Israeli cities. Doesn't that show that these measures are a factor that had a major influence?

In Bnei Brak it is true that the frequency of infections was more than four times bigger than in the rest of the population, but as with New York there is validity to the explanation that population density and difficulty with social distancing were the explanation. But it turns out from the analysis that I conducted that in that city as well and others like it, the slope of the curve is falling at a similar rate to the entire country, although somewhat later. These differences should be used as a parallel criterion in analyses of foci of where the virus hit hardest, in addition to tracking the number of infections in relation to the size of the population in the city.

What do you recommend regarding an exit strategy?

I was offered to present decision makers and the broader public with quantitative criteria to help plan and to check progress. We are talking about following the fall in the rate of growth in continual comparison with what is happening in other countries. It should be stressed that these changes are statistical fluctuations so that the absence of a fall for one or two days is not room for concern - you have to relate to trends in the change over a full week at least. This will allow predictions that can lead to a quantitative estimate of the damage that can be caused by an exit strategy and can also forecast the future regarding the risk to the health system so that it won't succeed its capability capacity for treatment. A quantitative forecast in this direction, will perhaps allow, based on the innovative that I developed for the comparison of the change in the slope in different countries. Similar efforts have been made for example by Prof. Isaac Ben-Israel at Tel Aviv University.

When the moment comes to gradually ease social distancing and check on the results, it is very important that comprehensive testing will be carried out among the population to gather data. Following the spread of the virus among asymptomatic people may provide insights into the natural factors flattening the curve as well as giving far swifter answers than following the symptoms of the disease two weeks after the infection. Checks like these should be done quickly before there are social changes so that there will be a base line for comparison. All these are examples in which the coronavirus should be judged with the proper caution as part of the exit strategy, for which we all wish will be successful.