Conversations with an epidemiologist: Blind spots

Over most of April and May, as the Covid-19 pandemic unfolded in India, growing from just a handful of cases to take region after region in its grasp, Dr T Jacob John, India’s foremost virologist, epidemiologist and researcher, spoke daily with Jagdish Rattanani, journalist and editor of Foundation of The Billion Press, to build what we call the COVID-19 diary. In this interview, extracted from the April conversations of the dairy, Dr John talks about why India’s COVID numbers are unreliable, how the country has no system to validate numbers, and indeed never had, for any disease.  A golden opporunity was lost as our infection numbers were so very small in April as to be containable, but the leaky lockdown and the ignorance of an improperly briefed population – who bought into the common myths that prevailed in April, of super Indian resistance and India’s possibly COVID-unfriendly summer – rendered the whole exercise a mere ritual.   
 

Jagdish Rattanani: There is a kind of swashbuckling attitude in India that Indians are going to get away lightly from the Covid epidemic.

Dr Jacob John: When I talk to people like vegetable vendors, they pull out their mask only when I ask them about it. The mask is for the police, like the helmet which is somewhere in the back seat, to show ‘I am a law-abiding citizen’.

I also get the feeling from friends from the middle and upper classes that they are gleefully accepting the fact that the predictions of the disastrous disease have passed. We are not seeing dead bodies lying on the street. So, there is no big problem. So, there is a certain complacency at play.

JR: Possibly because they are not getting a true picture of the numbers?                 

Dr John: So, what are the numbers? What are the numbers of typhoid fever and cholera in India? Of leptospirosis and meningitis, encephalitis, pneumonia? What are the numbers of deaths due to each one of them? We do not practise what Sri Lanka, Thailand, Korea, Vietnam, Japan and even China and all the European countries, all the Latin American countries practise . . .  that is, real-time information on locally important diseases. Every doctor is mandated to report soon after he diagnosis a patient.                          

Actually, all he has to do is when he suspects a patient of a disease is to treat the case as if it was that disease and inform the public health agencies. So, the agencies have a daily tally, weekly tally, a monthly tally, of all diseases that are of public-health importance. Those diseases that are epidemic prone, those that need vaccines . . . We don't even have pneumonia numbers. COVID, of course is (often) a silent phenomenon, it is relatively mild for the majority. If a severe case goes into hospital and that patient dies and if nobody diagnosis it as COVID, it is not tested and it will not be known. If in our 700-plus districts, 100 died (of COVID) in a week’s time, it will never make headlines.

JR- That is right.

Dr John- And nobody is diagnosing and recording by diagnosis. So, COVID will never be known. So, we have this problem of a built-in blind spot or built-in under-estimation. No number in India is validatable, because there is no system to validate it. We're actually thinking that the numbers reported by the testing strategy are the total numbers. It may be off by a factor of 1000 or 10,000 or even 100,000. So, we are absolutely stuck without information, and the government will say that those are real numbers. And (people will believe) India has some hidden protection, which would be fallacious thinking.

JR- So people may be dying of COVID and we will not know anything about it.

Dr John- If an 80-year-old dies of COVID, the cause might be thought of as old age or fever or something and he will be cremated or buried, and some people may get infected within the family. They will be asymptomatic, most of them, and if a case is tested (among them) as COVID, nobody worries about it. If it is another severe case then somehow the patient may end up in a COVID hospital, but that all depends upon the alertness of the system to find cases and admit them. Families may not have the time even to admit them, and they can deteriorate quite fast also.  So, we are blind in all aspects. Numbers, severity, cases, dead . . .  none of these is counted. 

JR: So, we are in mid-April now, what do you think the real numbers must be?

Dr John: We keep to fatality here. If we take one per cent as the fatality, 400 (roughly the deaths at the time of the interview in early-mid April) multiplied by 100 is 40,000. Okay, be more conservative, and say it is 2 per cent fatality. So, multiplied by 50, we have 20,000 cases representing the 400 deaths, roughly. Now, every COVID case (where the person is not merely infected but has the disease, with its symptoms) is one out of five infections (those who will test positive for Covid but may or may not have the disease, with its symptoms). So, 20,000 multiplied by 5 is 100,000. So, we had 100,000 infections of COVID-19 on a conservative estimate, to result in 400 deaths. And we have just detected 10,000. For every COVID disease, there are four asymptomatic infected people, so one case equals 5 infected people.

JR: Says who? Where does the data come from?

Dr John: Everybody knows that 80 per cent of the infections are asymptomatic.

JR: So, this is been understood, tested, accepted as . . . a medically proven kind of a number?

Dr John: Exactly

JR: That is not in doubt?

Dr John: That’s not in doubt. We are under-diagnosing by using a test which is in short supply and therefore very judiciously used. Public health testing is given importance to catch the infected, to quarantine and prevent further transmission, based on foreign travel, their contacts, their contact’s contact. Wherever you know there is a cluster, for example, the Delhi cluster, if someone comes in (from overseas), they will start a new outbreak wherever they settle down. So, the testing is done predominantly for these kinds of people and that will yield a certain proportion of positives. If you don’t do diagnostic testing, (which is different from testing only for the sake of public health and safety) or testing anybody with COVID syndrome, you will not know how many COVID cases are getting confirmed. The priority given to diagnostic testing is lower than that given to public health testing.

And that is what the BBC report that I saw  (https://www.youtube.com/watch?v=JIhNKZOHJ74, April 16, see above) highlighted. Two doctors in India… one saying six dead bodies arrived, all of them sounded like acute respiratory death, COVID not tested. The lady doctor saying that we are seeing symptomatic COVID-like patients but no testing. Now, no testing in a healthcare situation underestimates the total number of infections and puts healthcare staff at some risk.

JR- We had very low numbers in the beginning and could have done tremendously well in containing the spread:

Dr John: In the very early stage, the speed of expansion is proportional to the introductions into the country. If 1,000 fires were started in a forest, that will blaze much faster and much bigger than if only 10 were burning.           On February 2, the following countries had less than five cases; India, Italy, Spain, UK, Canada.

On March 1, Italy 1,694. Italy was welcoming tourists from everywhere. Spain had 84, UK 36, Canada 24, India 3, which were essentially imports from China.  During March, importation was not from China because everybody stopped people coming from China . . . all importation was from Europe, Dubai, Middle-East.

On April 1, India had 1,998 cases, less than 2000. Italy had 1,10,574, Spain had 1,04,119, UK had 29,474, Canada had 9,731. Then in five days, on April 06, India has 4,288 – up from 2,000. It doubled. Italy has 1,28,948, Spain 1,31,646, UK 47,806, Canada 15,512.

So, India’s importations were far fewer than in every other country, and India first concentrated on eastern side with the Asian countries and then concentrated on the western countries…not in terms of travel restrictions that much but in travel-and-contact tracing and testing, fairly aggressively. They tested people when they came from Italy or from Dubai. So, India had this testing net because of the low numbers we could do a better job than if we were overwhelmed with high numbers.

So, the speed of spread has been very low. Now on April 05, 4,288 is our number, and today (in mid-April) it is 11,000, so more than double in ten days’ time.

Give time, it will grow, the slope of the curve is not flattening. It is still going fairly steeply, numbers are low, but testing is low, the lowest testing rate in the world. So, slow beginning under-corrected by slow or low testing. And the worrisome part is in in spite of the lockdown, the numbers are increasing, even beyond the two weeks of lockdown. Up to two weeks, you could say what happened is a continuation from what had happened two weeks ago (importations, etc). After fourteen days of the lockdown, if we still see numbers going up, that means there were lots of leaks in the lockdown.  

JR: I know you will laugh at this, but I will still ask you. What about inherent Indian immunity that people talk about?  

Dr John:  It is age-old arrogance or the Indian superiority complex. That Indians are somehow different. Genetically we are no different. The other theory people sometimes tend to say is all weak links have been killed off in the past, because we didn’t look after them, they died. Only the strongest ones survived, so we are the children of the strongest among the species. Looking at the effect other diseases have had on us, there is absolutely no evidence to support the theory that the Indian gene is different. We had small pox, we had influenza, we had all kinds of diseases that came in, we are as susceptible as NIPAH (to anyone else). No disease has shown that we have an inherent superior gene.

India is the world capital of TB. We were never been able to get rid of malaria. So, I wouldn’t count on any Indian advantage, except that the BCG vaccination given in India may have some dampening effect on a macro-scale, so therefore we may have slightly less burden than a non-BCG country like Italy or Spain etc. But that is not worth counting on very much. However, that may be one factor. That will be seen only in retrospective analysis; after seeing the magnitude of the height of the epidemic and the duration of the epidemic.

JR: What about the weather?        

As an optimist, I would like to believe weather might make a difference; as a pessimist, I would say, ‘no’, weather will not make a difference. As a public health person, I would say don’t count on weather; it is an unknown. So, assume it will have no effect. but if it has an effect, then thank your lucky stars.

JR- The low death rate in India has drawn deep attention from everywhere. You say someone called you from Peru about it…

Dr John: Yes. That was the first question. . . . everyone is assuming that that death rate of COVID is low in India. That may be true or may not be true. I've not seen any reliable or validated evidence for that assumption. In our practice of public health . . . in surveillance, real-time counting of cases, we cannot even compare pneumonia deaths of January April 2019 and April 2020. Data just do not exist…So, some numbers will eventually emerge in the government website. The But true numbers will never be known. We have a problem in data collection.

JR: But the counter-view is that while the death toll that has been reported may not be precise, it unlikely to be way off the mark because concealing a huge number of deaths also is not possible in a country like India. So, maybe, instead of x, it could be x plus delta, two delta three Delta . . . into 10 delta even. It may not be 10x or 100x. What would you say to that?

Dr John: That is why I brought up the pneumonia issue. If somebody dies of pneumonia and it's not diagnosed as COVID, it will not be recorded as COVID. If something like this happens in 3,000 different spots in India, it will not be picked up. It is not plane-crash-like numbers, all in one place. So, if the deaths are widely distributed, you will never hear about it. In a country of 1.3 billion, 826 deaths (late April) from lab-confirmed Covid cases is very low. But they are not obtained by the usual method of counting all cases and all deaths. The 826 is evidence-bsed, it could be 100 per cent of the cases or 50 per cent the cases, or 1 per cent of cases. Now, the ICMR has just published their data on testing of what is called a severe acute respiratory infection (SARI) and they have come up with some numbers/

JR- What do they indicate?

Dr John: I don’t remember the numbers, but these numbers were distributed in 52 districts of the country. That means it is widely distributed. Small, small numbers, one here, one there. You can’t have one COVID case in one place. You can't have just one crocodile in the water, if the water is infested, there will be lots of crocodiles, but you have not seen them.

JR: Surely, the lockdown must have helped in some way.

Dr John:  Lockdown was declared from the Monday morning of 24 March for three weeks, supposed to end on 15 April, but is now extended to May 3. So, when locked down was declared, only 519 persons had been found infected and 20 plus dead in the whole country of 1.3 billion. That shouldn't worry anybody. So, this lockdown should have been equal to New Zealand's lockdown having no further cases, because we called the lockdown just as the virus was spreading in the country. Now, one month later from the lockdown date, that is, April 27, 26,917 are positive, 826 deaths.  So I asked the Peru journalists who had called me, you'll be the judge to say if lockdown was successful. Those who argue that without lockdown the numbers could have been many times more, that's pure guess. 

JR- One could also say the lockdown is very leaky in a country like India

Dr John: Yes, the Peruvian doesn’t know that, for the Peruvian lockdown is a lockdown. Only we know that the lockdown was a fake lockdown Now, some scientists have calculated that by the end of May the number infected will be 244,000.  Given that deaths are 5 per cent of COVID cases, and COVID cases 20 per cent of COVID infections, we must have passed that in end of March itself. As we don’t have data we will never know the truth. The whole world will think we don't have a problem. And we will pretend as though we don't have a problem. That people are not dying, we have some hidden protection . . .

JR: I got it…anyway it will come out. If it is getting worse it will come up at some point.

Dr John: See the thing is, what will come out (in public knowledge) is a cluster of deaths like in Dharavi. There are people dying of pneumonia, one in one hospital, another in another hospital, and 100,000 in one hundred thousand hospitals across the country, no news. That is the problem in this kind of epidemic. The revenue department has some data on it because they have to certify before cremation is done. Otherwise, you cremate in your own backyard. Nobody will even know. So, like the Peruvian, everybody will think that India has some magic trick.

JR: Correct, but the reality is very different.

Dr John: We have been living with this kind of discrepancy all our life. We have TB under control, but we have the worst TB problem in the world. We have the world’s highest per unit population, not because of 1.3 billion people. Nobody bothers about it, nobody even talks on it. WHO doesn't talk about it. We don't have a cholera control programme or a typhoid control programme. And nobody talks about it. I mean, WHO doesn’t talks about it, WHO is fully aware, there is a country office in India. So COVID is also one of these diseases, nobody will talk about it. Nobody will know about it.

JR: It will become endemic and keep killing.

Dr John: Yes. It will become endemic and we will have to live with it. And who is at risk? Only people with diabetes, hypertension, chronic lung disease and senior citizens. And so seasonal influenza bumps of few people every year, that is not on the radar, not recorded.

JR: We are now close to May. How could it get in terms of numbers?

Dr John: I think a large number of senior citizens may die. But it will be one grandfather, one grandmother (scattered across the country), and if you don't test for COVID, you will not see this reality. So, our calculation is that we already have lakhs of infections by now. We had given a range for mid-April, somewhere 2 lakh and 7 lakh infections.

JR- Currently (end April) the government says we have over 25,148 total cases, and 9,064 discharged or cured, and 1,152 deaths. After four weeks of the lockdown.

Dr John: At the start of the lockdown there were only 257 positive. So, under lockdown it grew to 25,000. But the lockdown extension will result in zero traffic accidents. And other human-to-human transmitted infections like the flu, diphtheria and whooping cough may be under control because interaction among children is virtually nil. So, all sorts of infections will be at a standstill. But COVID is not at a standstill. That means either the lockdown is very, very leaky, or in spite of the lockdown intra-family spread is occurring or it's coming to the attention of the system. So, whether these numbers are going up because of the lack of discipline …

JR- In India one cannot get more discipline than this. Such a large country with a corrupt police force . . . 

Dr John: So, anyway, that's where we are. We have no hope of good news in the near future. The numbers are creeping up. 257 four weeks ago, four weeks after, 25,000, it's a very steep climb of the graph.

JR- Yes, in a lockdown. Every other country has stabilized under their lockdown.

Dr John:  Yeah. So, our lockdown is ritual tokenism, and it's really not a real lockdown. And, we are still flying blind.

JR- But the government’s hands are also tied . . . it cannot not do another lockdown, isn't it?

Dr John- Lockdown beyond three weeks is acceptable only if you want to aim for eliminating the virus, Then, such drastic measures are in a certain sense acceptable. You are trying to make the graph go down and that will require every chain of transmission to be identified and pursued and interrupted. That can be done in some countries or in a small state, provided each state does it by itself. Like New Zealand. New Zealand's lockdown was highly successful because they wanted it done so that the infection is virtually eliminated, so that you can come back full swing back to business. Right or wrong, I do not know. I thought the idea of a green district, orange district, red district was a workable concept. In green districts economic activity starts, orange districts try to become green; red districts to become orange . . . each state could focus its attention on the green districts with a different approach than the red district. And the district people should do the district job. 

JR: Yes, more local it goes, the more effective it will be. They will also understand the peculiarities of each region.

Dr John: Yes. There is no new importation that is happening. It’s all local fires catching up. So, my prediction – and there is nothing to show that my prediction is likely to be wrong – I still expect a high peak. The healthcare system may be overwhelmed, but healthcare people are in bad shape. But all this has not hit people between the eyes. And that is a result of people not been taken into confidence and told the technical side of the bio-medical problem. How is the virus spreading, how fast does it spread…and what exactly is the problem in India, what should everybody do, why social distancing or physical distancing are necessary. When you say not more than five people should assemble, what I see is this: four people sitting together and sharing tea at a tea shop. Sitting next to each other and no mask, chatting away. So, that scene has to change actually, people have to understand that physical distancing means actually physical distancing, mask means mask. Even now leaders are not wearing a mask when they come to the television camera, whereas you will not see a Japanese leader on television without a mask. Whereas, United States has a double line, a railway track. Have you seen that?

JR: What do you mean?

Dr John: A double line as in a railway track that goes across the Atlantic. Starting in January, ending in December and continuing into next January. This is the normal path of death in numbers  . the numbers of death will be within this railway track. Because this railway track is created by past experience, you accumulate the last ten years’ data and project it for 2021. On the graph the number of deaths will be on the Y axis the season in the X axis. So, this is going up in the flu season, down in the non-flu season, and this kind of up and down is very visible. So, this railway track does up and down, gently going up and very gently coming down and very gently going up again. During the influenza years, during the winter, the dots jump out of this railway track.

JR: OK 

Dr John: That's how they know when there is an abnormal number of deaths. Within these two lines, deaths are the normal average number. Only when it exceeds (do they know something is out of the ordinary). But we don’t have anything near that. Ten years ago, the million death study came up with  with the information that malaria deaths were 200,000 per annum in India, while the official data said 2,000.

JR: That is interesting.

Dr John: They excluded other causes of fever and they concluded that the deaths are from malaria because it mostly in malarial areas. The reviewers were divided. One reviewer said this cannot be published because the official data is 2,000 per annum from malaria in India. .The other reviewer said India doesn't have a good review and numbering system. So, the methodology is correct and 200,000 is probably nearer to the truth than whatever the government gets, so publish it. The Lancet editors said said they wanted a tiebreaker, an independent opinion from outside the reviewers.

JR: And it came to you?

Dr John: They sent it to me. I read it very carefully. And I did some research of my own. And I found that in the previous year of the study, Madhya Pradesh had published a government official document that the number of women, pregnant women, who died of malaria in Madhya Pradesh in one year was 2000, equal to the all-India numbers that the malaria programme was officially projecting. One state was showing that in pregnant women alone deaths were 2,000. So, I wrote to Lancet and said we need better numbers than 2,000 because in one state (alone) pregnant women with malaria death number is 2,000 and the million study had reasonably good methodology, so I strongly recommended that the paper be published. If there is any problem we will see afterwards. Within three months, India revised its number to 20,000. Thereafter, they have numbers above 20,000. So, that’s illustrative of the Indian numbers. They pick some numbers, I don’t know from where. WHO needs some numbers, so give decent numbers.

JR: How is this million death study conducted?

Dr John: They took a million deaths, reported, from from government records, and they sent people to all the households of the million deaths. They were trying to figure out what caused the million deaths. So, they could ask a question like, did the person have fever and was a fever the main thing before death. If the person had paralysis, like stroke or was unable to speak, then was it a stroke or a brain problem, or chest pain. There are questions to categorize deaths as respiratory disease death, neurological death, fever without other things, or kidney death, or traumatic deaths. So, they had this huge classification . . . snake-bite death (even). And they came up with numbers which were so divergent, as to 2,000 official numbers and 200,000 study numbers in a million deaths. So that is the story. For a Peruvian journalist. I am a representative of India. I'm an ambassador. So, I cannot shut off the questions. I've got to be honest. And I, I have to be seen as arguing reasonably, which is of course, bad for the country, bad for the country's reputation. So, I wish nobody asked me questions.

JR: The only argument against our having high COVID numbers is by now we should have had many patients in that case.

Dr John: We are getting many more patients. Where is the surveillance? Where is the testing? OK, we test 100 severe acute respiratory illness (SARI) cases, but if there are one million cases, how good is the sample of 100 to inform us of the reality?  We don’t have data. We don’t have a surveillance system that tells you how many acute respiratory infection cases are  happening today. If the number is known, then an ICMR study can be extrapolated to the whole. If not, you will never know if the sample was large enough or not.              

If the frequency of SARI cases (for instance) was 1 per cent last week but 3 per cent this week, and if the number of SARI cases is known,  then immediately we could calculate. We need true surveillance of all cases that all hospitals see, all pooled together, through the public health system, on a daily basis. We don’t have that system, so we will never know.  That is the worry that I have today.