An estimate of Covid-19 under-detection and under testing in India

You may have heard that India is not testing enough for Covid-19. Due to our limited resources and lack of availability of reliable test kits, it seems that it is true. So, how bad is it?

We can get an estimate by making some reasonable assumptions. I tried to reach the number by making the following assumptions

1. Covid-19 has mortality rate of 2%. Perhaps it is lower in India. But, lets use this as start point

2. Covid-19 case outcome takes 15 days.

3. In India 4% is positivity rate - For every 25 samples taken, 1 case is positive

We can do the following calculations:

Death count on 21st May 2020 = 3584

By taking 2% mortality rate, on May 6th total infections should have been 3584 * 50 = 179200

Cases actually reported on May 6th = 53007

So under-detection rate on May 6th = 179200/53007 = 3.38

I have done this calculations over the pandemic period to see if this number is improving or not. And the good news is that - yes, we are improving. On March 25th this number was close to 17! See the graph below




If you extrapolate  the graph the number comes to 2.23 for today May 22nd.

May 21st cases reported were 118235

With this estimate, total number of infections today should be 118235 * 2.23 = 263664

Based on 4% rate of positivity, total tests that India should have conducted = 263664 * 25 = 6591600 (close to 6.6 million)

India currently reported total numbers of tests done = 2,615,920 (2.6 million)

There is gap of close to 4 million or we are only testing 39.7% of what it should be


Disclaimer: I am just an amateur. I am not claiming any expertise on subject matter. Please give your feedback on how to improve this model