The concept of herd immunity hinges on the numbers of truly infected persons in the population who have recovered and can now block the transmission to uninfected persons.
A recent report from Amdavad (Ahmedabad) Municipal Corporation reported a large survey of COVID-19 antibodies in the population of the city. Based on the antibody tests per million population, the survey of 30,054 persons claims to be the largest survey performed in the world. The report states that the study was conducted “ with a sample population of 4770 per million in comparison to a recent Spanish study where sample population ratio was 1302 per million and a US study with a sample population of 255 per million and a very small ICMR study with a minuscule sample population ratio of just 79 per million”.
Recommended ArticlesView All
India's cotton exports to shrink close to 30 percent in FY23
IST3 Min(s) Read
This survey was conducted from June 16 to July 11. It comes close to a report of a large population survey conducted in Delhi, sampling 21,837 persons between June 27 and July 10. The Delhi survey revealed an antibody prevalence of 23 percent. Both the Amdavad and Delhi study reports pointedly state that the antibody positivity rates are far below the levels required for herd immunity.
These reports raise several questions on: the accuracy of the antibody tests; proximity to herd immunity threshold (HIT) and duration of protection conferred by the antibodies. The use of these tests for characterising viral exposure in a population as opposed to an individual’s immune status has also been debated in the context of ‘immunity passports’ for persons returning to work.
All diagnostic tests are rated for accuracy on the basis of sensitivity (the ability to detect persons in whom the disease is truly present) and specificity (the ability to exclude persons in whom the disease is definitely absent). Generally, there is a tradeoff between sensitivity and specificity. Exceptionally good tests have high sensitivity and specificity but neither is 100 percent.
However good these two values are, the ability of tests to correctly predict which person in a group has disease depends on what the prevalence of the disease is in that population. If the prevalence of the disease is high, as in hospitalised persons, the number of false positives will be low. When the prevalence is low, as in the general population where many are uninfected, the test can yield several false positive results even if the intrinsic test characteristics of sensitivity and specificity remain the same. Even for a test with a sensitivity of 95 percent and specificity of 90 percent, the probability that a person who has tested positive actually has the disease is only 51 percent if the survey has been conducted in a 20,000 person sample of a population that has only 10 percent prevalence of the condition.
The problem of false positive results in COVID-19 comes not only because of these test characteristics but also because cross-reactive antibodies produced by other coronaviruses can give a positive test response. This is true not only of the SARS-CoV1 virus but also of other milder coronaviruses that cause the common cold.
There are two implications of this. False positive test results can inflate the estimated infection rate in the general population, much more than they do in a high prevalence hospital population. In the above example, the prevalence rate based on a ’case’ count that includes both true and false positives will be estimated at 18.5 percent rather than 10 percent. Larger the survey in a low prevalence population, the worse the distortion. Also, a positive test result cannot be used in such a setting to label a person as either having been infected or now having gained immunity. For this reason, WHO has dismissed the idea of introducing immunity passports, which proposed a positive antibody test as a permit for people to return to work.
The concept of herd immunity also hinges on the numbers of truly infected persons in the population who have recovered and can now block the transmission to uninfected persons. The threshold for this has been variably estimated by optimists to be between 20-40 percent, by middle-grounders to be between 50-60 percent and by conservatives to be 70-80 percent.
Most believe that at least 50 percent of the population must be infected. We are far from that. It must also be recognised that even if 60 percent of persons in Delhi are infected, anyone in the remaining 40 percent will remain protected only when they are shielded by the cordon of immunity only in Delhi but will become vulnerable if they travel to another city or town where the virus has by then infected only 10-20 percent.
For all of India to cross the threshold will take a long time, even if some small pockets do so in a few months. We need to observe all the precautions to protect ourselves against the virus, without meandering under the illusion that the magic cloak of herd immunity has already descended on us.
—K Srinath Reddy is President, Public Health Foundation of India (PHFI). The views expressed are personal
First Published: IST