Mumbai: India has had over 2.2 million cases of COVID-19 and close to 45,000 deaths, and the country now has the third most cases after the United States and Brazil. A big source of fear and stigma in the spread of this disease so far has been the probability of those close to an infected person contracting it. However, a global literature review conducted at the Indian Institute of Public Health (IIPH), Gandhinagar, says that 80-90% of family members do not develop COVID-19 after a primary case is diagnosed in the family. The secondary attack ratio (SAR) for a household stands at about 6% in India.

We speak with Dileep Mavalankar, author of this study and head and director of the IIPH. Mavalankar is an MBBS and MD in preventive and social medicine from the University of Gujarat, Ahmedabad. He also holds a master’s and a doctorate in public health from the John Hopkins School of Hygiene and Public Health.

Edited excerpts:

Could you take us through the broad findings of the study?

Let me give you some background of why we started looking into this. If you remember, everybody had been saying that this is a new virus, [and that] it will cause millions of cases and hundreds of thousands of deaths, etc. But five months into the pandemic, we do not see that in India. Some time in early May, we started looking at the spot maps of cases and deaths in Ahmedabad. We saw that there were very few houses where there were two dots--showing two cases in one house. [A similar analysis] was done 150 years ago for cholera cases in London by John Snow, and that is how public health started. That indicated there are not many cases in one household. And that is when we started looking for data and reading international papers. There were many papers showing [low a low secondary attack ratio]--a rate of 10%, 5%, and one paper shows less than even 5%.

When we finished the analysis in July, we were pleasantly surprised that a household that already has a case of COVID has a secondary infection in only 10-15% of cases. This is a good sign. Of course, the disease is infectious and spreads rapidly. But within the household--where there is no distancing and no masks--in spite of close contact, the cases do not seem to be happening. Before a person is diagnosed, they have symptoms for at least three or four days because normally, once some symptoms start, they take home medicines, and from local shops or local doctors, and if they do not get better, then they go for COVID testing. In spite of four to five days of close contact, family members are in some way resistant.

Is this a feature that has been seen across the world?

Yes, there are many papers--largely from China and Korea. But of course, there are papers from the UK, Italy, the US. The variation is quite a bit. Some papers have shown 5% SAR, and one paper has shown 50% SAR from China. But most papers show between 10-20%. The ICMR study in India showed about 6-8%. Some states have up to 11%.

What is it telling you, in terms of the way this virus is behaving?

My interpretation of this is [that] the virus is highly infectious, it spreads very fast through even minor or small contact; but that everybody is not susceptible to this virus, that a large number of people within our country somehow have resistance--I will not use the word immunity, because immunity is equated with herd immunity, vaccine-based immunity, etc.--but some kind of resistance, either innate or acquired, to this virus.

In some sense, this is not new. This should have been seen. If you remember 2006, chikungunya came, which was a new virus that came back by mutation after 40 years. It came back from Africa to India and many parts. Even then, about 1.4 million Indians were infected and several thousand died. But not everybody was infected. So, there are many such [examples of] innate resistance in the body, which prevent everybody from being infected.

Could we extend that to tuberculosis (TB), dengue and many other diseases as well?

Sure. If you consider tuberculosis, almost 40% of Indians are infected. But the number of people who develop disease is the interplay of people, the bacteria of TB and one’s own resistance. In this disease also, many people may be infected. For example, if you take the sero-surveillance data from Dharavi slums near Mumbai, 57% are positive. But Dharavi has only 2,500 cases with a population of about 1 million. Less than 0.25% of people are showing [up as] cases.

If you were to look at it from the other side, is it likely that all those who have been in the studies are those showing less viral load? In contrast, you have super spreaders, who have completely tilted the numbers in some cases. So what could we be looking at?

Some studies do mention that there is a lot of variation. There are a lot of families where there are zero secondary cases. And there are some families that have many secondary cases in the same family. But still, if you consider the transmission from husband to wife or wife to husband, it is only about 50%. So, 50% of the spouses do not get infected in spite of very close interaction with the patient.

Basically, my interpretation of it is of course [the] viral load of the infected case could be one [factor]. But [a] second [factor] is also the immunity or resistance of the person who is susceptible but not infected.

Could you dwell on that a little more? Let us say I have a flu, but my spouse or child does not get it. Is it the same analogy, or should we be looking at it differently?

It is the same. If you read the literature on exposure, when you cough at very close distance, lots of virions come out through droplets. So there is no reason why everybody should not be getting infected. [But] it is also the nature of the disease as well as the nature of the immunity [that impacts transmission]. This may be the reason why, even though it [SARS-CoV-2] is highly infectious, [its SAR is low].

There are various forms of immunity or bodily resistance or barriers to infection. This coronavirus is a family of viruses, and there are flu viruses, which are similar to coronavirus and there may be some cross-immunity--[which could be] a possible reason why people do not get infected in spite of being in close contact. Secondly, there is cellular immunity--it is not always the serological or antibody immunity. Third is, sometimes, what is called as surface immunity--which is in the respiratory passages; the virus is not allowed to sit and infect the cells, which is very difficult to measure.

Have we understood the concept of super spreaders better? Super spreaders have clearly tilted the scales in many cases. One person may infect 50 people in a very short time in a closed space. Is that something that we should be studying more or trying to understand better?

We are now doing a second literature review of disease spread outside of the home. Basically, there are two spheres in which the disease gets spread: inside the home, where you have continuous contact with the person for 12-18 hours, and during the lockdown almost 24 hours; [or outside the home] when you go out in a shop, mall or closed space like an office, [where] the exposure may be much limited. There, the super spreader and the superspreading events [are being seen]. There have been papers describing a choir group in Washington state in the US where 55 of 60 people present for a two-hour choir practice got infected from one person.

It is also whether the viral load is very high, what kind of exposure you have, what kind of closed environment you have. There have been such things also in some Indian factories. So we are also planning to study the spread in workplaces.

Are you getting a sense from the literature and the data that you have looked at that India or Indians are in any way different or more immune, less immune, or similarly immune?

My understanding is--and this is based on reading of data appearing in the media, newspapers, and understanding Indian societies--in spite of a very dense population, the proportion of people infected is much less than in many other countries. Second is, our nature or severity of infection also seems to be less. Recently, you saw data saying that only 0.5% of infected people in India are on ventilators, whereas in other countries, it is much higher. Our mortality is 2-3%; in other countries it is 10% [or more, despite] better health services, or much better nutritional status, [and] much better treatments available.

In some sense, there is something that has kept the numbers [low]. And it is not only India--if you see Bangladesh, Pakistan, Myanmar, Sri Lanka, [they] have even lower [numbers] than us. Not only [do] our old health systems work, but there are some innate factors in spite of very congested populations and very poor populations. One of the pathologists here has been talking about [whether] anaemia [is] a protective factor. And we have lots of anaemia. Earlier, we talked about the BCG vaccine, malaria, etc. So there are many reasons we could have better immunological status to prevent the disease from getting very infectious as well as very severe in form.

Going by the varied rate of growth, and given what you know about the disease--and the fact that it does not spread with the same intensity that we thought it would--what is your sense hereon?

The peaks of this pandemic seem to be happening in a very localised way. If I look at Ahmedabad data a little more closely, the first part of Ahmedabad to be badly infected was the central zone of the city, which is the walled city. And now, since the last one and a half months, that zone has very few new cases coming up. The disease spread to outside zones of Ahmedabad, and now that has also gradually started declining. And if you look at the dynamics between Ahmedabad and Surat, Surat had very low cases initially in April and May. Now Ahmedabad has fewer cases than Surat. So my sense is, it will go from one place to another as effective interventions and actions happen, as well as immunity builds up. Both happen. Whatever the number of cases reported, international literature shows that there are 20-100 and sometimes 200 times more subclinical or asymptomatic cases, and immunity builds up. So the pandemic will go down.

Recently, we had interaction with Karl Friston, a neuroscientist from the UK who has developed a new hypothesis based on European data analysis and modelling. His view is [that] half of the population in Europe is not even susceptible to the disease. He says that the immunity and the herd immunity effect can be seen at as low as 20-30% or 35% of infection rates (not cases, but seroprevalence rates). Because 50% may not be susceptible.

My sense is, in many big cities like Mumbai and Delhi, it has started declining. Of course, it will spread to rural areas. But, the transmission in rural areas is much less, because the congestion is much less.

As you look at the transmission and its intensity, what is your sense about the life-cycle of this disease at this point of time? Will we continue to see waves till there is a cure or a vaccine? Or will it die out or slow down dramatically in some other way?

My sense is we will not see waves in the same cities, as much as waves passing on to other areas that are as of now not so much infected and do not have immunity. As I said, if you see the seroprevalence in the central zone of Ahmedabad [or] Delhi, the average is about 23-24%. Mumbai, in slums like Dharavi, [it] is 57%. If you have that high immunity, I do not see a major second wave. But those areas that were not infected, of course, will keep on having [new cases].

But the other thing that is not much discussed is we also see that the virus seems to be getting weaker; or it is hitting populations that have higher immunity. So you are seeing milder cases as compared to what we saw in March and April. So that may also happen, and it has happened in past epidemics--that as time progresses, the infecting agent becomes milder. So you may have more cases, but proportionately less death or mortality.

And is that something that evidence is already showing?

Yes. Unfortunately, the data is not available so easily to analyse and see the week-wise mortality. If you see, we do [look at] cumulative mortality. If we do week-wise mortality, even in Gujarat which has higher mortality like 4%, the weekly mortality is 1.5% now, as compared to 7% in April and early May.

Since you mentioned Gujarat and Ahmedabad, what is your sense on why the mortality rates there have been relatively higher compared to some other cities?

It is not clear. One is that the central zone of Ahmedabad is very congested. So when the pandemic had initially started, it had much higher infectivity there. There is also some evidence from the Biotechnology Mission Labs in Gujarat that the strain in Gujarat has more infectivity or virulence. Of course, Gujarat has higher levels of diabetes, hypertension and obesity. That could be another reason.

You said you are initiating a couple of new studies. Could you sum that up for us?

The second global literature review we are doing is on spreading outside the home--secondary attack rate in office, work and casual contact, etc., which will take a couple of weeks to finish and then another two weeks to publish.

Secondly, we are doing primary data collection to find out the district-level estimates of SAR in Gujarat by sampling people who had this infection.

The third study that we have planned is on the issue of superspreading events: Are there models where we have some understanding from industry or offices or some transport areas, where many cases got infected from one case? And what is the understanding and relevance of that for the disease spreading in our country?

We welcome feedback. Please write to We reserve the right to edit responses for language and grammar.