Read on Twitter
1/n. Some data on how test performance interacts with local epidemiology
2/n Test performance assumptions
3/n. Epidemiology
4/n PPV and NPV now
5/n PPV and NPV at peak in December
6/n PCR is good for ruling disease in and out
Although in asymptomatic patients, especially when prevalence is low, it’s 50:50 whether it’s a true positive
Although in asymptomatic patients, especially when prevalence is low, it’s 50:50 whether it’s a true positive
7/n LFDs can rule out disease in low risk settings.
LFDs look good for helping manage close contacts
LFDs may a role in risk assessing people with symptoms in particular low risk contexts Eg. Differentiating covid / non-covid in people with common low grade symptoms
LFDs look good for helping manage close contacts
LFDs may a role in risk assessing people with symptoms in particular low risk contexts Eg. Differentiating covid / non-covid in people with common low grade symptoms
8/n LFDs look pretty awful at ruling in disease in low prevalence settings
9/n History of symptoms and covid contact is nearly everything. Getting the approach to assessing this right is more important that any test.
10/n In terms of elective screening, LFDs have an added advantage of telling you about the situation on the day of the procedure.
