RE: Diagnostic test performance27 Dec 2024 19:12
Maybe it "betrays" not " belies "my intelligence, Green.
Now, who is being obtuse and seemingly unintelligent now?
The "new" NPV data etc was widely discussed last August by all of us. It is not new data.
We went through all that test data last year. This is what was said and it doesn't seem like anything has changed. Ie Cizzle still keeping people in the dark.
This may be useful for people ...a definitive guide to tests really.
https://ebm.bmj.com/content/18/1/5
"For a test to be useful sensitivity plus specificity should be at least 1.5"
Ciz is right on this threshold for it's best results (95+ 61). Just over.
Now with PPV , it could be useful and that it will target the demographic to those more likely to have the disease.
Cizzles test certainly could be useful as guidance and better than CT scan for sure at the moment.
But, let's take the better figures (95% sensitivity, 61% specificity)..
If a cohort of 1000 people present for testing and 100 have lung cancer...
This means 95 with be correctly diagnosed, with a further 351 false positives.
So a total 443 presenting as positive, of home only 100 (minus the 8 not picked up) are truly positive.