4 ms·

"If only 1.7m tests are being conducted per day, what can that 386,000 really mean? Is it that 0.0004% of the country has come down with the virus since yesterd

"If only 1.7m tests are being conducted per day, what can that 386,000 really mean? Is it that 0.0004% of the country has come down with the virus since yesterday, or that nearly 23% did? Obviously, the true number lies between those absurd extremes, but who knows where?" Why is there not a proper estimate on how big a part of the population is infected? And with which strains? Surely with a randomized, representative survey you could find this out, instead of all this speculation? It also would give authorities better grounds for their decisions.

ignoramous 6y ago
> It also would give authorities better grounds for their decisions. The authorities need to do those studies or commission them. Instead, the authorities are busy politicising vaccine distributions (vote us in for free vaccines) [0] and polarising on religious fault lines. The country has gone to the wolves. [0] https://archive.is/dNYo4 https://archive.is/dNYo4
throwawaysea 6y ago
In general it is quite hard to get accurate representative samples if there is an unknown or high degree of variance between locations (are you scaling up the right samples to arrive at an accurate estimate). As for knowing which strains, that would certainly be useful but it requires significant sequencing infrastructure, which is not easy to scale up. The Washington Post noted (https://www.washingtonpost.com/world/2021/04/29/india-surge-variant-genetic-sequencing/ https://www.washingtonpost.com/world/2021/04/29/india-surge-...) that India is #102 among nations in sequencing per capita. It notes: > Sequencing takes a high level of infrastructure and funding. India, which spent very little per capita on health care before the pandemic, has not been able to quickly make the investments to scale up, said Gupta.
sokoloff 6y ago
If I understand your calculation, it seems like the lower bound is around 0.03%, not 0.0004%.
blackoil 6y ago
Tests are not randomly distributed, but estimates of actual cases is 4-10x. It also varies by state. Some states have very active tracing and testing. Some are actually trying to undermine tests, death count.

ignoramous 6y ago

> It also would give authorities better grounds for their decisions. The authorities need to do those studies or commission them. Instead, the authorities are busy politicising vaccine distributions (vote us in for free vaccines) [0] and polarising on religious fault lines. The country has gone to the wolves. [0] https://archive.is/dNYo4 https://archive.is/dNYo4

throwawaysea 6y ago

In general it is quite hard to get accurate representative samples if there is an unknown or high degree of variance between locations (are you scaling up the right samples to arrive at an accurate estimate). As for knowing which strains, that would certainly be useful but it requires significant sequencing infrastructure, which is not easy to scale up. The Washington Post noted (https://www.washingtonpost.com/world/2021/04/29/india-surge-variant-genetic-sequencing/ https://www.washingtonpost.com/world/2021/04/29/india-surge-...) that India is #102 among nations in sequencing per capita. It notes: > Sequencing takes a high level of infrastructure and funding. India, which spent very little per capita on health care before the pandemic, has not been able to quickly make the investments to scale up, said Gupta.

sokoloff 6y ago

If I understand your calculation, it seems like the lower bound is around 0.03%, not 0.0004%.

blackoil 6y ago

Tests are not randomly distributed, but estimates of actual cases is 4-10x. It also varies by state. Some states have very active tracing and testing. Some are actually trying to undermine tests, death count.