The TLDR: A massive serological survey of Delhi shows that almost one in four people in the city were infected with the coronavirus at some point. The government thinks this is a good result. Critics beg to disagree. 

First tell me about this survey

The result: The government collected 21,387 samples across 11 districts between June 27 and July 10. It found 23.48% of those tested have antibodies. This means their bodies were exposed to the virus, triggering an immune response. 

The methodology: Delhi residents were administered the IgG Enzyme Linked Immunosorbent Assay (ELISA) test. This is a highly sensitive test that only detects the presence of antibodies. It is not used to check if someone has Covid. And the level of antibodies it measures are typically produced two weeks after infection. What this means: that 23.48% likely reflects the number of infections in early June.

The value: The problem with this coronavirus is that many cases are asymptomatic or only display mild symptoms—and therefore go undetected. Serological surveys help fill in this gaping holes in our data:

“From a public health standpoint, knowing how many and who have already been exposed to SARS-CoV-2 gives a clearer picture of how widespread the virus is in local populations. This is extremely useful because public health measures depend on how far coronavirus has already penetrated into the general population.”

Ok, so what does this survey tell us?

There are three points of data that are worth your attention.

One: The official number of cases in Delhi is only 123,747 (including recoveries), which is less than 1% of the population. But the survey shows that it is highly likely that the actual number is around 4.7 million—or 23% of Delhi’s 19.8 million population. 

Two: In May, the government conducted the first such survey in 83 districts across India. Results from 60 of them showed that only 0.73% of those tested had been exposed to the virus—reaching 1.09% in urban areas. The May percentage for Delhi: 9-11%. In other words, the number of infections have more than doubled since then.

Three: On a cheerier note, if there are indeed 4.7 million Covid cases, the city’s fatality rate—percent of cases that result in death—is only a miniscule 0.07%.

Four: The survey is a snapshot from June when the city’s numbers were climbing. Delhi was conducting 9,500 tests daily, and 37% of them turned up positive. But in July, of the 25,000 daily tests, only 9% are positive—leading to hopes that Delhi may have finally hit its peak. 

Point to note: Anywhere between 40-60% of a population has to be exposed to the virus in order for ‘herd immunity’ to kick in—that’s when enough people have produced antibodies, and therefore block the spread of the virus. But 23.48% is nowhere close.

So is this good or bad news?

The government’s take: is that the results are good, and show that it has done everything right:

“In a densely populated city like Delhi, if the prevalence of the disease is 22.86%, it shows the efforts taken by the public and government in preventing the disease. The containment strategies adopted by the Ministry in coordination with NCDC and the Delhi government have helped us to prevent the further spread of the virus. This shows that if scientific principles are followed well then a pandemic can be controlled well.”

The experts' take: Independent experts are not so sure, and they make two key points. One: the big gap between the official numbers and the survey’s estimate indicate big gaps in testing:

“This means a huge number of cases went undetected either because they were asymptomatic or because they were missed under the testing strategies over the past two or three months.”

Two: Results indicate that there are more serious cases of Covid that are also missing from the official data. And here’s why. As per the medical consensus, around 80% of all cases are asymptomatic or have mild symptoms. That means 20% are moderate to serious. If we extrapolate that breakdown to Delhi’s numbers, this is one worrying conclusion:

“If only 20 per cent of the estimated uncounted cases of 4.5 million were symptomatic, there would be over 900,000 cases with symptoms… Delhi has so far done only 830,000 tests and found around 125,000 positive. The testing strategy has definitely missed symptomatic patients.”

What about the rest of India?

We won’t know for sure until ICMR releases results of the nationwide survey. But the private lab, Thyrocare, released its data of antibody tests—and it indicates 15% of the country may have already been exposed to the virus. That’s 18 crore Indians! The hottest spots: Bhiwandi, Thane (44%), followed by Peenya Dasarahalli area in Bengaluru and Anand Vihar in New Delhi.

The bottomline: It is clear that the virus has spread farther and wider than the official numbers indicate—but this is to be expected. Experts across the world agree that the real number in any population may be 10X the recorded cases. But the entire point of a serological survey is to deliver a needed reality check—not offer an excuse to double down on denial. 

Reading List

• The Telegraph has the best reporting on the survey results. 
• ScienceDirect explains why serological surveys are important. 
• Firstpost's detailed look at Spain’s serological survey flags its limitations. 
• The Wire raises questions about the government’s ELISA testing kit. 
• The Hindu explains why serological tests should not be used to issue immunity passports—because there is no evidence one cannot be reinfected.