What do test results really mean?

Try this:

1) Imagine a disease that affects one person in a thousand.

2) Now imagine there is a test that can detect this disease before symptoms show. And imagine that test is 95% accurate. That is, it will detect a positive case 95% of the time. That’s not a bad test.

3) Now Imagine you have gone for the test, and to your surprise you have tested positive.

Given all that information, do you have the disease? How sure can you be?

If you said you were 95% sure you had the disease, then you would agree with most doctors when asked the same thing. But you would all be wrong. You almost certainly do not have the disease, in fact the chance that you have it is only one in fifty!

How can that be?

Well, remember that in a sample of 1000 people tested, one person on average has this disease. And a test that is 95% accurate is 5% inaccurate. So in that same sample of 1000 healthy people, there will be roughly 50 false positives.

So, in this fairly reasonable example, there are roughly 50 positive test results for every one actual case.

Are those numbers realistic? Obviously there is a huge range in both the reliability of tests and in the prevalence of diseases. But you can play with the numbers all you like, and you’ll find the problem doesn’t go away, it just becomes more or less severe.

This is one reason why I never go for routine screening tests. I only ever have a medical test of any kind when I have reason to think something might be wrong, or if I need a certificate of health for some purpose, eg work. And there have long been calls to abandon routine cancer screening for this exact reason: false alarms can start chains of events that could actually cause more death and suffering than the diseases they are looking for. However, in our culture of disease-phobia most people find that notion very hard to take in, either rationally or intuitively. And that makes disease a great tool for manipulating the masses.

The fact is that for many tests, especially for viruses, we have no idea how good they are.

Viral tests have particularly poor diagnostic power, since there is absolutely no correlation between a positive viral test and clinical illness: these tests are supposed to be able to detect disease agents in people who are clinically well. So, how do we know that these tests signify the causes of disease and not some random stuff, and who makes that decision? Elephants abound in that particular room, and knowing some of the science does not make this go away.

We now even have asymptomatic carriers who are not contagious! What is this super-dangerous thing they are carrying that doesn’t make them sick and can’t be passed on? I can’t even get my head around what ‘non-contagious asymptomatic carrier’ actually means. To me it sounds like ‘healthy person’ to the tune of moving goalposts.

And because of the holes in the science of virology, there is no way to test the test: there is actually no benchmark for determining the presence of the supposed virus, other than the test itself. Imagine you need to measure something but you don’t have a ruler; so you decide to make a ruler, but you have nothing else against which to check its accuracy. Everything is then guesswork. That is the same problem as designing a viral test. In short, some or all viral tests could be utterly hopeless, and there is no external way to check: they can only check their rulers against the thing they are measuring and hope their guesses about one thing or the other are correct. That is the sad state of the science at the moment, and if it sounds nonsensical to you, that’s because it is.

There are rival theories – good ones – and so maybe it is time to let the virus go. But this is all tangential, and leaving it aside for now, let’s just assume that viral PCR tests actually mean something diagnostically: false positives mean we still end up in the same place anyway.

There are four occasions I can think of when a false positive definitely could do you more harm than good. And some diseases fall into all four categories.

1) when the disease is thought to be especially serious, since the stress of the diagnosis could itself do a great deal of harm to your health.

2) when the disease could affect your life path. It’s tricky working as a surgeon, for instance, with a diagnosis of Hepatitis C, or AIDS. Even an unconfirmed or suspicious result could hang over you.

3) when the normal practice is to treat the disease before symptoms develop. There could be many people treated with dangerous drugs unnecessarily because of false positives.

and 4) has perhaps the biggest ramifications of all; when the disease is (believed to be) contagious.

For example, imagine testing positive for the current pandemic strain. You might be one of, say, ten false positives for every true positive (or one of fifty, or one of a thousand!). But you are unlikely to be told that. More likely you will be told the test is the test, the result is the result: trust it and move on.

Here is the kicker: symptoms or not, everyone you know must then be tested as well. And thus another generation of false positives emerges. By now the problem should be obvious.

So with mass screening the numbers can explode very quickly, giving the impression of an outbreak where there is none. The actual term for this is ‘pseudo-epidemic’.

It’s vital therefore to take into account actual clinical illness, and not just a colour change in a test tube or a readout on a screen.

At this moment we have around five million people enduring a radical lockdown in the Australian state of Victoria. The city of Melbourne has a plague of positive test results, not of clinical illness. And the response of the authoritarian government has been more testing and more false positives. This is a classic pseudo-epidemic.

The ramifications of shutting down an entire city and its region are serious and long-lasting, and this is happening on the belief that a positive test result equates to a ‘case’. Worse still is when ‘cases’ of serious disease must have radical treatment. The experience of SARS in 2003 was that many people, with perhaps quite survivable illness, were immediately put on cocktails of strong drugs that may have caused their deaths. Thus the ‘seriousness’ of the problem was incorrectly confirmed by their demise, and on it went.

It’s a well known cycle, and an easy trap for gung-ho politicians. Unfortunately the style of health policy-making in Victoria is very much to take expert advice, but only from the experts who support the government position. It’s an Emperor’s New Clothes situation.

And thus, as I have been saying for years, ALL viral pandemics are pseudo-epidemics, ALL are self-fulfilling prophecies caused more by fear and superstition than actual pathogenesis. After nine years in university learning about the sciences, my fear is not viruses but the knee-jerks of ignorant people in public office.

Whilst it is possible that there have been some genuinely virulent and deadly viruses in the world, the evidence does not strongly support their existence, and nor are they needed to explain any pandemic. ALL pandemics have other explanations. So at this stage scepticism is very sensible. All you need do to kill a great many people is declare that there is a deadly virus, then let panic and desperate measures do the rest.

In summary, the fastest way to stop these outbreaks, is to stop testing, announce that the world is safe once more, and let everybody go back to work.

 

~