COVID Surge: When Will Cases Fall & What Do Math Models Predict?

Prof Gautam Menon, who models diseases, explains the recent COVID trends in India and what we should watch out for.

Image of Indian map used for representational purposes.

(Photo: FIT)

(As told to Indira Basu, Asst Editor, Op-Ed, The Quint)

The notion that India's experience with the COVID-19 pandemic would be very different from that of other countries has been demolished by the second wave. India's current numbers of daily cases have far exceeded peaks in the US and Brazil, and it currently accounts for more than 60 percent of new cases worldwide.

To understand what could happen in the future, we need to understand our present circumstances better, so that we may be prepared for future waves and their consequences.

When Will Cases Begin to Go Down?

The numbers of COVID-19 cases rose very sharply after about 15 February 2021 and have been continuing to rise. At some point, when enough people have been infected,  standard epidemiological theory would tell you that it becomes increasingly more difficult for someone infected to come into contact with someone uninfected. Multiple restrictions and lockdowns have also been implemented across the country. So, what can we say about when that might happen? Is it going to happen in two months, could it happen in one month? That's where models come in.

A bunch of models, including models by Dr Bhramar Mukherjee of the University of Michigan in the USA, have suggested that this — a decrease in India’s number of active COVID cases — should happen sometime by mid May.

Professor Mukherjee has been modelling this pandemic for a while, and so has our group, a modeling group associated with the Indian Scientists Response to COVID-19 (ISRC). Combining the projections of these and other models, there is a general belief that reported numbers should turn down somewhere around the middle of May. This is just from an overall synthesis of what different models are saying currently.

Quality of Data Is Inadequate

But there is a very broad error on this. It could happen somewhat later, perhaps two weeks later. And this is mainly because the quality of the data that goes into this model has definitely been decreasing. There are many cases of people with symptoms who are not being tested because they just prefer to stay at home rather than be tested. Or they may refuse to be tested in the first place because they worry that they will be taken away, against their will, to COVID treatment centres. There are currently huge delays in testing, so any test results announced now may have come from a test that was taken 5 days later, or 3-5 days earlier.

There is also serious concern that some states are not reporting cases accurately, or even deaths, especially from rural areas.

For a mathematical model, all these inputs are required. Thus, these are the uncertainties, but  even taking  the uncertainties involved into account, I think it's reasonable that somewhere between the 15th and maybe the end of May, is a reasonable estimate at this point.

The number of those testing positive has actually shown a downturn in recent days, down by about 10 percent from the highs of more than 400,000  daily from a few days ago. This is encouraging, but we need to see consistent trends over a week or more before we can be sure that any change we might see is genuine.

Estimates for the numbers of cases at the peak are all over the place currently, but this is because models use different estimates for the fraction of those infected who experience symptoms serious enough to warrant getting tested.

What Makes the Indian Experience of the COVID Pandemic Different?

What is unusual about the Indian epidemic is that it had a peak around the 12 September 2020 and then daily case numbers came down more-or-less steadily for almost five months. Because usually the sequence of waves of the pandemic in other countries, and also in other parts of India, had seen relatively short intervals between successive peaks, it was believed that maybe India had many people who were already infected — and that this would protect us against a further rise in cases provided we were careful enough.

A model that the Indian government funded specifically for this purpose, called the SUTRA model by its originators, but more generally known as the ‘supermodel’, was excessively optimistic about the end of the COVID-19 pandemic. They predicted that India would see no cases after February 2021 since it had reached, what is called, ‘herd immunity’.

What Has Changed Since the Tapering Down of the Previous COVID Wave in India?

But several new things have happened since the tapering down of the last wave of the COVID-19 pandemic. For one, a number of variants have come to India from other parts of the world. These include the variant B.1.1.7 (UK) which is now well-entrenched in Delhi and Punjab, as well as the variant called B.1.351 (South Africa). But there are variants that are indigenous, such as the B.1.617 variant and various sub-lineages of it.

This was first seen in the state of Maharashtra and is believed to be responsible for the rapid rise of cases there. Of all the variants in India, this may potentially be the one best adapted to spread. There are other variants in the east of India and the south of India too, but it looks as though the B.1.617 variant will displace those in the future.

How Viruses Mutate

When viruses mutate, most of the time, nothing unusual happens. But occasionally they mutate to transmit more easily between people. If that happens, that lineage of the virus will take over the previous lineage and become the dominant lineage. I believe that this is what has happened with the rise in cases in Maharashtra which signalled the second wave, mainly because of the B.1.617 variant.

But a relaxation of what is called COVID-appropriate behaviour, the opening up of schools in between, election rallies and and much more mixing between people overall, certainly piggy-backed on top of this and made the situation worse than it should have been.

So, while what specific variant could drive the second wave and when  it would happen couldn’t have been anticipated in advance, we should have looked to the examples of what was happening in other countries to understand what could change the dynamics of COVID-19 inside our country.

What is the Extent of Undercounting of Deaths & Cases?

I will not comment on the several model estimates of the fatalities because there seems to be considerable undercounting in this wave, and there's virtually no good data to compare with. The extent of this is not fully known. In many places there are large queues outside crematoriums and burial grounds, and the variance between official numbers and counting on the ground seems large.

The other thing is that many people are dying who should not be dying — they are dying because of lack of access to medical facilities, ICUs and oxygen supplies.

As for the question, are the actual numbers twice as more or thrice as more as the real figures on ground? For cases, one can estimate somewhere between 20 and 30 as a reasonable number; there are about 20 undetected people for every detected case. This is what we have estimated for the last wave using our model.

It's true for this disease that many, many people will be infected by it, but not know that they have had the disease, simply because they don't show any symptoms.

When it comes to deaths, the factor is probably somewhere between 2 and 3, between actual deaths and reported deaths in the previous wave. It also depends very much on which part of the country we are discussing. Right now, in the second wave, that factor may be considerably larger, but that is all we can say.

Mathematically Can We Say That Fatalities Have Down Among Vaccinated Folks?

Can we estimate mathematically whether among the vaccinated population fatalities have gone down? Well, if we had good data, we could have estimated that. Currently, a group of scientists have written a petition to the Government of India to make data from ICMR and other government agencies available for analysis.

I hope there will be some progress on that front, since there is definitely more expertise in this area outside government than within it and a proper analysis would be very useful in our understanding.

From looking at mortality in the population above 60 before and after vaccination, there’s a good case to be made that vaccinations are already making a difference.

The applied mathematician Murad Banaji has studied mortality in Mumbai city carefully and come to a related conclusion. Reports from both Bangalore and Chennai, with the Chennai analysis coming from the ISRC group, supports this. This is heartening.

What About Disease in Vaccinated & Younger Populations?

There have been reports of ‘breakthrough’ events, where someone who has been fully vaccinated, has ended up getting COVID-19 within two weeks after their second vaccine dose.

The general experience states that it is very rare for people to develop a severe infection after two shots of the vaccine, and that, I think is reasonably clear from the data that I have been seeing.

There have been questions about whether fatalities and serious disease among the younger populations are being seen more in the second wave because the 18-45 age group largely constitute the unvaccinated lot. I would say no — it could happen for many reasons. Usually, epidemiological experience has been that the first wave of the disease is when the older people are affected, often those in care homes or other situations where they are potentially more vulnerable as a group.

In the second wave it's the younger people. So, the nature of contact at the level of age groups — younger people tend to go out more — may also have a role to play in this.

Right now, it may appear that there are more younger people being affected by this second wave, but it may also just be anecdotal. There’s an overall rise in cases, so you will see a larger number of younger people in any case.

But whether they represent a larger fraction of reported cases vis-a-vis older people is as yet unclear, and can only be assessed with large-scale data.

I should also point out that some  trends can only be seen retrospectively when you look back at the data and then try and understand where these trends came from. It's so hard to say anything definitive in the middle of an ongoing pandemic.

Why Has Delhi Been So Badly Hit by the Second Wave of COVID?

I would assign that largely to the fact that the dominant variant of the virus in Delhi is what is called the B1.1.7 variant. This is known to be more transmissible. It may also have the characteristic of immune evasion — you have antibodies from a prior infection, but when you're infected again, the virus is able to circumvent those antibodies to some extent, reducing their ability to counteract disease. (It is important to remember that vaccines prevent severe disease or death, almost surely, but don't prevent infection from COVID-19.)

Apart from that, migration into the city from nearby locations may also have played a role in increasing the pool of susceptible people. The numbers suggest that the situation in Delhi has improved, although only slightly. The test positivity still remains high, and is a source of concern.

What Is Most Worrying About the Spread of COVID-19 in India Right Now?

At this moment, I would say I'm worried about the state of West Bengal, mainly because there's a whole mix of variants there with some of them known to be more infectious than others. The general lack of attention to COVID protocol during the long election cycle has led to this, and the situation as seen from reports on the ground is worrying. Karnataka may be seeing the effects of its lockdown, recently extended, soon.

Kerala continues to be hard hit, and is an enduring puzzle. It is at least encouraging that deaths there remain low, but reports of cities such as Ernakulam running out of ICU beds may change this.

Cases are rising at a fast rate in the northeast of India as well as in Jharkhand and Goa, but it is possible that this just reflects better reporting. These are also low population states, so their impact on the Indian numbers as a whole will be small.

Bihar and Gujarat seem to be seeing substantial undercounting, at least from newspaper reports and other reports from the ground, so, the true scale of the epidemic there is hard to gauge. It is likely, at least from anecdotal reports, that the situation is far worse there than the numbers would indicate.

(Gautam I Menon is a Professor at the Departments of Physics and Biology, Ashoka University. He tweets @MenonBioPhysics. This is an opinion piece. The views expressed above are the author’s own. The Quint neither endorses nor is responsible for them.)