Thread by @EdmundFordham

@lucyfrazermp is rattled and writes a round-robin email to her constituents. Keep it up, good people of SE Cambs ! (thread)

But @lucyfrazermp is misleading her constituents with things that can't be true. Asking about "false positives" is apparently a "common point".
There is "a robust system in place to adjust for potential false positives" Really ? How do they do that ? (thread)

To "adjust for false positives" you need the operational False Positive Rate (oFPR) for the testing programme. But the Govt doesn't know what this is. "The UK oFPR is unknown", says Lord Bethell, only a month ago (thread)
questions-statements.parliament.uk/written-questi…

To "adjust for false positives" we'd like to know the proportion of all those Positives that are False. This is the False Discovery Rate, or FDR. @MattHancock and @lucyfrazermp don't understand that this is different from the FPR. Look it up here. (thread) en.wikipedia.org/wiki/Sensitivi…

@lucyfrazermp: "daily case(s) [all +ves] is not .. skewed .. by false results as the vast majority of tests [all tested] return accurate results". No, Lucy, the +ves are far fewer than the tests, we're doing so many. (thread)

So, @lucyfrazermp, even if "the majority of tests" were accurate, the "majority of positives" can still be False. Especially if the disease is actually rare. (thread)

So we'd like to know the FDR for the testing programme. For that you need (i) the oFPR (which Lord Bethell says the Govt doesn't know) and (ii) an idea of the disease Prevalence. And we can't know that without "adjusting for false positives" .... (thread)

So I have written to @lucyfrazermp asking four easy questions.
1. How does the "robust system" adjusting for FPs work ?
2. How's it done if the oFPR is "unknown" ?
3. If we now know the oFPR, what is it ?
4. What is the FDR of our testing program?
(thread)

@lucyfrazermp says there's a "robust system to adjust for false positives".
In which case, @lucyfrazermp, @MattHancock et al. should be able to answer those questions easily.
Go on, Lucy, prove me wrong.
The questions are easy if you've got a "robust system". (thread)

If @lucyfrazermp can't answer the four easy questions, then where is that "robust system" ?
And @lucyfrazermp needs to write to her constituents again to explain the errors. (thread)

All good folk of SECambs should write to @lucyfrazermp too. Here are the questions to ask:
1. How does the Govt "adjust for false positives" ?
2. How's it done if Lord Bethell says the oFPR is unknown ?
3. What is the oFPR of the testing programme ?
4. What is the FDR ?
(thread)

Remember, if there is a "robust system" for "adjusting for false positives", they should know the answers to them all. Feel free to let me know how you get on.

unroll @threadreaderapp

unroll @threadreaderapp