Letters to the Editor — June 10, 2020

India’s IFR is not Higher than Global IFR

A letter to the editor by Prof T. Jacob John (TJJ henceforth) titled ‘India’s CFR’ on 8th June correctly points out that Case Fatality Rate (CFR) for COVID-19 in the strict sense must be calculated by the number of deaths caused by COVID-19 during a specific time divided by those proven to be infected during a specific time. The point about the need for denominator and numerator sharing a comparable time period is well taken.

However, TJJ seems to confuse CFR with Infection Fatality Rate(IFR) and vice versa, when he states that “The total number, 226,713, are not all “cases” (COVID-19) but “coronavirus infections” (lab-confirmed); hence the derived rate will be Infection-fatality rate (IFR), not CFR”.

CFR is normally taken to be deaths divided by known cases, and IFR, deaths divided by known plus unknown cases. Known cases will only be a subset of the overall cases and hence IFR will always be less than or equal to CFR. Here, what is implied is the opposite.

Surprisingly, in the very next sentence TJJ contradicts himself and states, “IFR is a fraction of CFR, since cases are a fraction of infections”, which is correct. TJJ then goes on to calculate what he calls IFR using current death numbers and past case numbers, accounting for the time lag. However, the IFR he calculates is actually CFR, because it uses the number of known cases alone.

After erroneously establishing ‘IFR’ to be 6.43%, TJJ claims that it is much higher than the global IFR. This is primarily wrong because he is comparing apples and oranges; if his methodology was followed on the easily available global data, he would have reached a global ‘IFR’ much higher than India’s. However, a much lower, seroprevalence survey based IFR was conveniently used.

Calculating CFR in the middle of a pandemic is a challenge. One way of sidestepping the time-lag problem is to focus on known outcomes – deaths and recoveries. Number of known deaths divided by number of known outcomes should give us a more robust proxy to CFR, if needed. Calculated using current numbers, the global CFR is 10.4% and India’s CFR is 5.5%.

Midway through a pandemic, what is commonly called CFR, warts and all, is a matrix that is useful across the globe, and comparisons are indeed made by researchers because flaws are mostly across the board. What Prof John does is to confuse IFR for CFR, and reach wrong conclusions, as he has also done at least once earlier in an interview with Mr Karan Thapar.

Reportedly, results from ICMR’s seroprevalence survey suggest that hotspots in India show upto 30% population infected. This, if true, will pull down the Indian IFR to very low levels. It is known that due to systemic constraints, India may be severely under-reporting deaths as well as cases, even more than many other countries. However, debates should not be based on arbitrary definitions that lead to erroneous conclusions.

Oommen C Kurian,

Observer Research Foundation.

Related Topics
Recommended for you
This article is closed for comments.
Please Email the Editor

Printable version | Jul 13, 2020 7:16:39 AM | https://www.thehindu.com/opinion/letters/letters-to-the-editor-june-10-2020/article31790403.ece

Next Story