Even if it’s half reliable it’s evidence of much undetected spread. Interesting to see result of Kirby study.
-
-
Replying to @Vic_Rollison @orpheuseurydice
Half reliability isn't really the way it works. A study is either reliable or it isn't.
1 reply 0 retweets 2 likes -
Replying to @JamesUVanDyke @orpheuseurydice
I haven’t looked at the paper itself, just the Age report. But wouldn’t the only way the entire study be contradicted is if the antibody test literally didn’t work?
3 replies 0 retweets 0 likes -
Replying to @Vic_Rollison @JamesUVanDyke
The Age report contains all the info needed to conclude that no firm conclusions can be drawn from the study. 41 of 2991 tested +ve to antibodies to the virus. False +ve rate of test = 1.09. Article itself says sample size is too small to make confident estimate; understatement!
1 reply 0 retweets 1 like -
Replying to @orpheuseurydice @JamesUVanDyke
They’ve probably used largest sample they had access to - convenience sample of hospital admissions so not universalisable. I have no idea what sample would be needed to make the results definitive, but I do know opinion polls infer 16 million voters by surveying only 1,200
1 reply 0 retweets 0 likes -
Replying to @Vic_Rollison @JamesUVanDyke
Yes. The fact that it is too small a sample size would be a qualification made by the authors of the study themselves. Equally problematic is the false +ve rate for antibodies. 9 in 100, ie nearly 10% are false +ves. This error is amplified if you extrapolate.
1 reply 0 retweets 1 like -
Replying to @orpheuseurydice @JamesUVanDyke
If 10% are false positives, then the study is still 90% accurate so still shows a large amount of undetected spread.
3 replies 0 retweets 1 like -
That is incorrect. With prevalence of 2 in 1,000, 10% false positives means that 998 tests will pick up 99 false positives, 2 tests will be true positive, for a rate of 2% positive predictive value (or 2% 'accurate')
2 replies 0 retweets 1 like -
Dr Coatesworth says the model estimates “somewhere between 0-185,000 cases”. I don’t think I’d base any policy decisions on that!pic.twitter.com/VjoGcttFNY
1 reply 0 retweets 1 like -
Somewhere between 0-185,000, but everyone just hopes it’s zero and moves on. They found undetected spread - double tested positives to make sure. So much wishful thinking going on. 40% asymptomatic. Easy to see how undetected spread happens.
1 reply 0 retweets 0 likes
I'm not sure what you mean by "hope" here. The point others have rightly made is that the confidence interval includes 0, which means the results are not statistically distinguishable from a situation where all of the positives in the study were false positives
-
-
So how do you explain their explanation of their confidence of 0.28% positive based on double testing? They’re the ones who did the study. I’m just reading their analysis of their results.
1 reply 0 retweets 0 likes -
I'm not going to speak for the authors, merely pointing out the fact that with a confidence interval including zero it is possible that there were no true positive results in the study. This is the uncertainty that I believe Dr Coatsworth was expressing in the quote above
0 replies 0 retweets 2 likes
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.