Disclaimer: I am not a statistician so be kind
I’ve been trying to work something out. There’s been so much speculation about how reliable PCR tests are and how unreliable antigen (or lateral flow) tests are.
It seems to me that the issue lies in the speed of the results. Here, in France, if we find out we’ve been in contact with someone who tested positive, we aren’t required to isolate if we have had our two vaccine doses. In addition, if we decide not to work in this scenario, we aren’t paid for our time off.
Several people I know have waited over two days for a PCR response and understandably after extended periods of furlough can’t accept lower pay anymore.
What I would like to show is a comparison of antigen tests which would rapidly but less accurately isolate positive cases, against more effective but slower PCR tests. I would like to say that I am neither for or against either form of testing, it would just be interesting to see the comparison between the following two cases:
People trusting antigen results as accurate therefore isolating immediately.
People opting for more accurate PCR results but not isolating whilst awaiting results for 48 hours.
I envisage the things we would need to consider would be:
Accuracy of both positive and negative results in both testing methods
False positives and false negatives
Incubation period
R number
Please excuse my extreme ignorance towards the wonderful work of statisticians, but I find this an intriguing and important thing to understand.
How would I start to calculate this? And does the community have suggestions for the visualisation of the curve?
