|
Aug 2020
5:14pm, 13 Aug 2020
4,363 posts
|
run free
Think some of the schools in the UK have already purchased temperature sensors to monitor students entering the premises....
|
|
Aug 2020
5:16pm, 13 Aug 2020
17,201 posts
|
Bazoaxe
For me it’s another tool in our armoury. Albeit it has its flaws.
We can accept this along side other things in our effort to manage Covid or we can rail against these measures as we don’t like them and their implications or accuracy.
I still believe we should all be accepting some changes In our lives. The consequences of going back to February is just not acceptable.
|
|
Aug 2020
5:27pm, 13 Aug 2020
673 posts
|
BexleyKev
Having had an exceptionally high body temperature it isn’t the same as feeling hot due to being out in the sun or exercise. Conversely, having put ice bags on my forehead made me feel a little better but the temperature was still above normal. We have quite an amazing biological answer to over heating and self regulating temperature.
|
|
Aug 2020
5:33pm, 13 Aug 2020
12,980 posts
|
Ultracat
When I get very hot and sweaty I am more likely to check my BG levels rather than my temperature, I have a thermometer so may check my temperature if I can.
|
|
Aug 2020
5:41pm, 13 Aug 2020
11,143 posts
|
rf_fozzy
Baz. That's not the point. Understanding the uncertainty is crucial.
Apologies to all those who get this already, but I'm going to spell it out, as it's important.
Let's say 1 person in 100 has CV19 (i.e. 1%).
We do a thermal imaging test that detects 60% of people with CV19 and misses 40%.
The thermal imaging test also wrongly detects 10% of people who *don't* have CV19 (I've guestimated this number - if someone can find a real number for this, then can update the below) - this is the false positive rate.
So this method will
correctly find that people have CV19 -------->1% x 60% = 0.6%
miss that people have CV19 -------------------> 1% x 40% = 0.4%
Incorrectly find that people have CV19 ------>99% x 10% = 9.9%
Correctly find that people don't have CV19 ->99% x 90% = 89.1%
So if this test says you have CV19, the actual chances of you *really* have CV19 is (0.6)/(9.9+0.6) = 5.7% chance.
So the false positive rate is absolutely crucial!!
If the consequences are trivial (miss out on a running session with a club), no worries.
If the consequences are not trivial (missing school for 2 weeks), then it matters.
So I agree with larkim's point above.
But let's not miss the point that we must not lose sight of our test limitations when we decide to use these things.
|
|
Aug 2020
5:45pm, 13 Aug 2020
11,144 posts
|
rf_fozzy
Note that if you don't like a false positive rate of 10%, let's reduce it to 1%
The chance of this test correctly identifying you have CV19 is then is still only 38%!
|
|
Aug 2020
5:46pm, 13 Aug 2020
3,114 posts
|
Little Miss Happy
But the consequences shouldn't last anything like as long as that fozzy because it should prompt the person to get a test - worst case scenario should be about three days of self isolation, best case conformed not Covid within 24 hours. I don't believe I have Covid, I don't go anywhere to contract it, but despite being asthmatic and able to claim a medical exemption and finding it uncomfortable I wear a mask if asked to do so - it's a minor inconvenience that may save me, someone I love or even a random stranger so worth it.
If I recall correctly a skin temperature reading is often lower than that which you'd get taking a tympanic temperature.
|
|
Aug 2020
5:56pm, 13 Aug 2020
680 posts
|
JR
In the case of South Korea bus stop (which started this discussion) it could essentially mean you were stranded at work unable to get home - how would that work? Or in the case of kids - sent home - what if no parents home or couldn’t get home? Or denied a holiday you could have spent £1000s on?
|
|
Aug 2020
6:02pm, 13 Aug 2020
23,613 posts
|
Johnny Blaze
TBFTF I thought most scientists generally subscribed to the view that a bad test can be worse than no test. As pointed out, when prevalence in the community is low, a high false positive rate can do more harm than good.
I suppose that this can be moderated as a viewpoint by considering "what the test is for": is it to stop a person getting on a bus or is it to consign them to (say) 2 weeks in quarantine (wrongly) or to convincethem (wrongly) they have an invulnerability because they are a false positive for antibodies.
Generally speaking I would aim for the highest specificity possible. Crude testing will yield crude results and potentially endanger people's livelihoods and health. Can't end well in the long run. Just my two pennorth.
|
|
Aug 2020
6:26pm, 13 Aug 2020
7,299 posts
|
WA
At the risk of giving fozzy concussion from all the head banging, surely some false positives are better than no positive positives?
|
Coronavirus discussion thread
Jan 2025