Two questions.
One, what is India’s death rate from Covid-19 infection?
This is different from the case fatality rate, which is simply the number of deaths represented as a proportion of the number of cases. The case fatality rate, or IFR, is the number of deaths represented as a proportion of the number of infected. Warning: the numbers of deaths around the world, directly due to the coronavirus disease, or indirectly caused by it, are being underestimated, but in the absence of accurate data and records of deaths, there is no point trying to guess how much.
The IFR is an important metric because it accurately measures possible deaths from disease, and also the chances of dying from it. However, measuring it requires knowing the denominator: the number of infected people in a population.
That is why serological surveys, blood tests looking for Sars-CoV-2 antibodies (in this case), are important. These surveys measure the prevalence of a disease in a population, or the rate of infection. Since deaths from Covid-19 are recorded anyway, it is possible to calculate the IFR.
Around 20 serosurveys have been conducted in various parts of India covering about 110,000 people. I decided to choose four of these, excluding the ones with smaller samples. These four surveys, two in Delhi and one in Chennai and Ahmedabad, collectively covered around 78,000 people. The Ahmedabad survey showed that 17.6% of those studied had been exposed to the virus; the Chennai one 21.5% and the two Delhi surveys 23% and 29%. The Ahmedabad and Chennai surveys were conducted in July; and both from Delhi in late June / early July and early August. Delhi’s third serosurvey is nearing completion and results are expected next week. Based on these numbers, it can be assumed that at least 20% of the population in large cities has been exposed to the disease.
What does that do to the Delhi IFR? There have been 4,567 deaths from Covid-19 in Delhi as of Sunday night. An infection rate of 20% in the population would mean four million infections. That translates into a death rate from infection of 0.11%. This compares with a fatality rate of 2.4% in Delhi. In Ahmedabad, the IFR is again 0.11. And in Chennai it is 0.13. All numbers are based on the current population and the total number of deaths in these cities as of Sunday night. Three cities in different parts of the countries, all with similar IFRs, is interesting but also expected. Pandemics respect the math (it is also why I have repeatedly denounced states that pretend all is well on the basis of positivity rates that have not tracked such rates around the world).
The second question, how common is a Covid-19 reinfection?
This is due to media reports claiming that the first case of coronavirus reinfection in India has been identified in a Bengaluru hospital. There have been 27 million coronavirus infections worldwide, but only two known and established cases of reinfection (both at the end of August). There are probably more, but establishing reinfection requires sequencing the viral genome from the first infection in the patient and then sequencing the viral genome from the second infection in the patient and showing that the two are different. Doctors and scientists established this in the case of a Hong Kong man whose second infection was asymptomatic (detected because he was traveling) and an American woman whose second infection resulted in severe symptoms.
This does not appear to have been done in the case of the Bengaluru patient, which does not rule out reinfection, but also does not rule out residual infection or false negative tests.
Scientists are very interested in reinfections because they influence vaccine development. If there are major reinfections (two out of every 27 million are not) and if the immune system of these patients worked the way it was supposed to, the vaccine development process becomes a little more complex; it is always easier to develop vaccines. for diseases in which an infection results in lifelong immunity (or at least immunity for a significant period of time).
.
