Coronavirus **support** thread
162 watchers
Aug 2020
6:36pm, 13 Aug 2020
7,515 posts
|
Too Much Water
JR - The same smartcard you use on buses is also used for taxis in Seoul (they don't have Uber) - it also links with online banking, food delivery, satnav and loads of other things so if you fail the test you're more or less locked out of society.
|
Aug 2020
6:37pm, 13 Aug 2020
11,145 posts
|
rf_fozzy
Last post. WA - "some" false positives. As I've tried to illustrate above, if you had a false positive rate of 10%, then for every 5 genuine cases you find, you'd find 95 false positives. If you had a false positive rate of 1%, then you'd still only find 38 genuine cases and 62 false positives. Unless the false positive rate is ~0.1% then you'll find more false positives than genuine cases. |
Aug 2020
6:57pm, 13 Aug 2020
7,300 posts
|
WA
Thank you. Maths was never my strong point
|
Aug 2020
6:59pm, 13 Aug 2020
17,202 posts
|
Bazoaxe
fozzy, I get that there are some false positives. I also see that you made some assumptions in your calculations that may or may not be right. |
Aug 2020
7:09pm, 13 Aug 2020
1,388 posts
|
Grast_girl
An interesting graph on the effects of different interventions on R. Shows only lockdown really had an effect, although most other interventions weren't implemented for long enough to really see the effect. nature.com It comes from this paper: nature.com There is also this one in the same issue. Although I found it harder to follow it covers different countries and splits the lockdown interventions: nature.com |
|
Aug 2020
7:23pm, 13 Aug 2020
11,147 posts
|
rf_fozzy
Baz- - yes 3 assumptions. 1. Prevalence of CV19 in population - this is too high, it's actually far lower than this in the UK. But I chose a higher number to illustrate the point more clearly. If the prevalence is lower, the false positive problem becomes *much greater* (as pointed out by JB above). 2. False negative rate (40%) - taken from the paper ptb posted earlier. But you can make it whatever you like. Even 99% and the problem still exists. 3. False positive rate (used both 10 and 1%). This is the biggest assumption I've made and the one that makes the biggest difference. I've tried to show that unless the false positive rate is <1%, then this is what causes the problems. So which assumption do you disagree with? And why? |
|
Aug 2020
7:43pm, 13 Aug 2020
681 posts
|
JR
TMW - That is very scary. The idea of it all being linked to your bank account is really dystopian.
|
|
Aug 2020
7:43pm, 13 Aug 2020
17,203 posts
|
Bazoaxe
Thanks for explaining those and how/why you arrived at those. So for all its faults I think you said you still correctly find that people who don't have CV19 ->99% x 90% = 89.1%. |
|
Aug 2020
9:51pm, 13 Aug 2020
11,148 posts
|
rf_fozzy
Yes, that's correct Baz, given the assumptions I've made in the calculations above (with say a 10% false +ve rate), but is irrelevant as these people face no sanction. The point is that the test will find that more people that actually don't have the disease, but the test says that they do, than people who actually have the disease and the test says they do. It's the false positives that are important. Not the correct negatives. As I said above, this may be irrelevant *if* the sanction is that you end up missing a running club session, but if it's more serious (e.g. missing school), then it becomes important. Take the analogy I used earlier. Let's say we had a test for murderers. The test finds that for every 5 murderers correctly identified and locked up, it also finds 95 innocent people who are locked up. Is this acceptable? Obviously not. So the full picture is important. As are unstanding that there may be unintended consequences to what appear to be simple actions. And that is my final post on the matter! |
Aug 2020
10:05pm, 13 Aug 2020
4,368 posts
|
run free
fozzy - this is what the large marathon events had an issue with (especially the US ones - court cases etc). So finding a test that would be more reliable is important to them.
|
Related Threads
-
Coronavirus discussion thread
Feb 2024
-
Long Covid
Apr 2022
-
Children off school, check in here for support.
Mar 2021
-
Working From Home
Jan 2021
-
How to occupy yourself during lockdown
Jun 2020
-
What are you missing at the moment?
Jun 2020
-
Coronavirus. Antibody testing. Priority post box.
May 2020
-
Hey Fetchies how are you?
May 2020
-
Silver Lining!
Apr 2020
-
Coronavirus Regulations UK: What is permitted and prohibited
Apr 2020