Even as Kerala has become the single largest contributor to Covid-19 cases in the country, accounting for more the half of the new cases daily, most left-liberals have gone silent on the pandemic. However, a few braves among them have come forward to defend the Communist-ruled state to argue that higher cases in the state are the result of efficient management of the pandemic, not the other way around. With that goal, The Print published an article on July 30 using a dubious analysis to claim that other states under-reporting new cases by a huge margin.
The Print journalist Abantika Ghosh wrote the article based on tweets by one so-called ‘health economist’ Rijo John claiming that the results of the latest serosurvey conducted by ICMR prove this point. According to John, Bihar detected only 1 out of 134 cases, UP detected 1 out of 100, but Kerala detected 1 out of 6 covid-19 cases. Maharashtra, another worst-affected state, is reporting 1 out of 12 cases, according to the analysis based on which the Print article has been written. The figure for all India is 1 out of 33.
Thus, the article claimed that the entire country is under-reporting cases, but its extent is much more in states like Bihar and UP, where the cases have come down to less than 100 per day. The Print has claimed that cases in Kerala are more because under-reporting there is much less. The article further multiplies the alleged under-reporting factor with the state-wise official number of Covid-19 infections to arrive at ‘actual’ numbers.
According to this calculation, the actual infection in the country was 92.65 crore by May 2021, not the official number of 2.82 crore at that time. The respective numbers for the states are also spectacularly high with that calculation, which is 16.89 crore for UP, 9.47 crore for Bihar, 6.74 crore for MP, 6.17 crore for Rajasthan, 4.81 crore for Gujarat, 6.07 crore for West Bengal etc. The estimated cases for Kerala and Maharashtra using this calculation are 1.59 crore and 7.14 crore respectively. John claims that fingers are pointed towards these two states because they were detecting cases better.
While the serosurvey numbers are actually from the fourth round of the National Sero-Prevalence Survey done by ICMR, there are huge problems in using these numbers to calculate the ‘actual’ number of cases. The sero prevalence survey is a sample survey that measures the estimated percentage of people with Covid-19 antibodies, while the reported Covid-19 numbers are based on the total number of tests done on people. Comparison of estimated numbers from surveys with actual results is always problematic, as the surveys suffer from various limitations.
The current 3.16 crore total confirmed cases to date in the country is the result of over 46.64 crore tests done in the entire country. On the other hand, the sero prevalence survey was done among around 28,000 people in 70 districts in the country. Even though statistical models say that such a small sample base is good enough if the sampling is done correctly, a survey estimate can never negate the results of actual tests. The sero survey only gives an indication of how the disease might have spread, it is not designed to calculate the ‘actual’ number of cases, as Rio John and The Print did.
The sero survey measures how many people have Covid-19 antibodies. The antibodies are produced by the immune system of the body when it encounters the infection. Therefore, it provides an estimate of how many people might have contracted the infection. It is notable that a large number of people can develop antibodies after contacting the virus without developing any symptoms. Therefore, the number of people with antibodies will always be higher than the number of people who are tested positive, as generally only people with symptoms, their contacts, and high-risk category people are tested for the virus.
Covid-19 vaccines also produce the same antibodies in the body, and this is the big factor that was missed by The Print while calculating ‘actual infections’. In the total sample of 28,975 people, 2,631 people were fully vaccinated with two doses, and 5,038 people had received one dose of the vaccine. This means, there were 7,669 people in the sample who got the antibodies from the vaccine and not the infection. As more than one-fourth of the people included in the survey were vaccinated, it is completely wrong to use the sero survey numbers to estimate ‘actual’ numbers of Covid-19 infections.
The ICMR study says that antibodies were found in 81% of people with one dose and 89.8% of people with both doses, while the same was 62.3% for unvaccinated people. It means that vaccination is a big factor in the development of the antibodies in the body. If one has to estimate the actual infections based on the sero survey results, the number of vaccinated people will have to be excluded to get a truer picture. But this was not done by The Print, which resulted in unnaturally high numbers.
The sample of 28,975 people also included 7,252 healthcare workers, a high-risk group that is most likely to have antibodies either from infections or vaccines. 25% of the sample were healthcare workers, which is far more than the actual share of healthcare workers in the country. Therefore, it is another reason why the survey returned such a high proportion of people with the antibodies, and why it can’t be used to estimate the ‘actual’ number of infections in the country.
The argument that Kerala and Maharashtra are detecting cases better than states with lower cases is also not supported by another actual data, the test positive rate. The rate is 12.3% for Kerala and 3.4% for Maharashtra, but the same is only 0.02% for UP and 0.05% for Bihar. This means, for every 100 tests done, Kerala is reporting 12.3 positive cases. But UP is reporting 1 positive case in every 5000 tests done.
If actual cases in UP was much more and it was under-reporting cases, the scenario should have reversed. UP reported 37 new Covid-19 cases on July 20 after doing 2.45 lakh tests, while Kerala reported 20,772 cases from 1.53 lakh cases on the same day. Therefore, the logic of Kerala detecting cases better does not hold.
Due to these reasons, it is completely illogical to use the sero survey results to claim that India is under-reporting cases and the actual number of Covid infections is 92.65 crore. The survey only says that many people are ‘estimated’ to have antibodies against the virus, which includes vaccinated people. It does not mean the cases are being under-reported as claimed by The Print.
