Naked Science Forum
Life Sciences => Physiology & Medicine => COVID-19 => Topic started by: set fair on 25/07/2020 20:10:18
-
The UK has done more tests per head than the other countries in this table and yet we come out bottom for finding cases.
Column 1 Country
Column 2 Daily new cases (worldometer 7 day average from july 14th)
Column 3 Daily deaths (worldometer 7 day average from july 24th)
Column 4 Testing success (column 2 divided by column 3)
Sweden 404 4 101
USA 64,087 918 70
France 582 9 65
Germany 345 6 58
Canada 331 6 55
Belgium 104 2 52
Russia 6,531 132 49
Italy 194 10 19
UK 598 62 10
If deaths always occurred 10 days after a positive result and if the infection fatality rate were 1% then the last column would give the percentage of actual cases a country detects. I guess the actual percentage is about half the figure in the last colimn.
This suggests that the UK is testing like mad and only identifying between 5% and 10% of actual cases.
Any explanations?
-
Column 4 Testing success
I am having trouble understanding in what sense "#Infections/(#Deaths 10 days later)" is a measure of testing success.
- It may be an indication of the quality of the health system, and whether it has become overloaded?
It's not even clear what "Testing success" means. If you are getting the infection under control:
- The percentage of people who come up positive in random community screening will be very low
- The percentage of symptomatic people who come up positive in COVID testing clinics will be fairly low (they will have other things like the common cold or flu or hay fever)
Donald Trump's claim that you can flatten the curve by reducing the level of testing is delusional (as is his earlier claim that the main reason that the USA has lots of COVID-19 cases is because it is doing lots of testing).
- His alternative plan (currently being implemented) is to stop hospitals reporting COVID statistics to the CDC (which is staffed by epidemiologists charged with monitoring disease outbreaks). Instead they are to report to the HHS, which has a Trump-appointed director, and is much more likely to make announcements friendly to the Trump election campaign (like suggesting that Trump has cured COVID-19).
See, for example: https://www.forbes.com/sites/brucelee/2020/07/14/trump-administration-to-hospitals-dont-send-covid-19-coronavirus-data-to-cdc/#a9392660f8ae
-
The actual number of new cases per day varies quite sharply from week to week according to public behavior, which is determined primarily by political decisions that vary from day to day and even between adjacent regions of the same country.
The number of positive tests per day depends on how many people were infected a week ago (tests are not infinitely sensitive) and how you select people for testing.
The number of deaths per day, in a country with adequate access to free health care, depends on how many people were infected 4 - 6 weeks ago.
"Cause of death" is an unreliable statistic. Even if you carry out 100% postmortem testing, you will only identify corpses that died with, not necessarily from, COVID infection. And nobody wants to waste valuable tests on corpses. The only reliable statistic is seasonally adjusted excess deaths.
Tests do not save lives either directly or in the first order, because there is no treatment. A test only has value if it prevents a carrier from infecting others, and that depends on public behavior, which depends on political advice.
-
Donald Trump's claim that you can flatten the curve by reducing the level of testing is delusional (as is his earlier claim that the main reason that the USA has lots of COVID-19 cases is because it is doing lots of testing).
You will certainly maximise the number of confirmed cases by doing lots of testing, but what really matters is the number of prevented cases, which cannot be measured and therefore has no political value. I think it was Professor Ferguson who said "If nothing happens, we will have won" - but that's a vote-loser.