Positive-negative-positive: How reliable are COVID-19 tests ?
Test, test, test is what WHO told countries fighting pandemic. But even as countries are torn between conducting more or fewer tests, doubts are being raised about the worth of the tests themselves
These are testing times, literally. In the war against the pandemic that grips the world, the strategic war room planning hinges on the key operativetesting. One of the most tangible parameters of comparative gains in this war against the virus remains the number of tests per million.
So much so that President Donald Trump boasted of the US having the highest number of cases globally almost as a ‘badge of honour’ on account of the number of tests carried out. Only days later, alarmed over the rising graph he was to suggest cutting the number of tests down.
Closer home, a state government in an affidavit before a high court admitted to restricting number of tests in order to stem the panic!
A disease becomes a cause of alarm on two counts- morbidity and mortality. With a data base of millions of cases, we now know for sure that COVID-19 causes no long-term morbidity. Mortality therefore remains our only concern. Calculation of mortality is a simple exercise of arithmetic- number of deaths as numerator and the number of those affected as denominator.
First the denominator- the number of people affected. COVID-19 is an abbreviation of Coronavirus induced ‘disease’.
Disease word is rightly described as Dis-ease, meaning thereby a feeling of being unwell. Since an overwhelming majority of those who are counted as cases of COVID do not develop any dis-ease, it is therefore all about the result of a test called RT- PCR, a positive report implies infection by the virus.
With stakes high, the test ought to be fool proof. But what have caused doubts are the high number of false positives and false negatives thrown up by the tests. The sceptical layman is also asking how the test is deemed to be fool-proof when the coronavirus is so new, barely six months old, and when we know so little about it.
But such is our faith in the reliability of this test that no one barring a few seems to have subjected the test to a scientific scrutiny.
Torsten Engelbrecht and Konstantin Demeter in an article published last month in ‘offGuardian’have raised some valid concerns. OffGuardian as a journal has been mired in controversies, but the issues flagged are pertinent.
A test needs to be tested against a gold standard. In case of an infectious disease, the gold standard remains the demonstration of the microorganism (bacteria, virus or parasite as the case may be).
In case of RT PCR for COVID-19 the virus has so far not been demonstrated. How have we then relied so heavily on a test that has no validation of its reliability?
Over a century ago Robert Koch postulated that in a microbial infection the microbe must be present in every case of the disease. Microbe must be isolated from the patient with the disease and grown in pure culture. The specific disease must be reproduced when a pure culture of the microbe is inoculated into a healthy susceptible host.
With certain modifications these postulates have been the basis of defining a microbial disease and have so far withstood the test of time. In the present pandemic, with half a million deaths over six months, the disease is yet to pass muster Koch’s postulate.
The issue of a lack of a gold standard to validate the specificity (the test being specific to the disease and giving no false positives) and sensitivity (the test picking up all cases with virus and giving no negatives) has been raised by an Australian infectious diseases specialist, Sanjaya Senanayake.
Torsten Engelbrecht in the controversial paper has raised what appears to be a valid issue. “The PCR is extremely sensitive, which means it can detect even the smallest pieces of DNA or RNA — but it cannot determine where these particles came from. That has to be determined beforehand.”
But if the virus has not been isolated so far, how on earth can we be sure about it. So far no one has claimed to have isolated the virus in purified form.
For any test, including RT PCR, there are certain numbers of false positives. The predictive value of a test in diagnosis of a disease is called a positive predictive value (PPV). It depends on false positives and the prevalence of the disease. Even if we set aside all arguments put forth in the preceding discussion and go by an accepted false positivity rate (of all PCR tests) of 5%, with the presumption of a prevalence rate as high as 20 percent, the calculated PPV comes to 70% only.
In a situation as in India where prevalence rate is very low, The PPV falls down to just 30 percent. In other words, the probability of having the disease in spite of a positive rate is abysmally low. To put it more simply, millions of people in the world seem to have been wrongly labelled as being Coronavirus cases.
A person testing positive and then negative and again positive is nearly impossible. Haven’t we come across reports of such happenings? This can only be possible if the positivity is false.
Yuval Noah Harari in his book 21 lessons for the 21st century presciently writes, “Diseases may be diagnosed and treated much before they lead to pain or disability. As a result, you will always find yourself suffering from some medical condition and following this or that algorithmic recommendation. If you refuse, your insurance may become invalid, or your boss will fire you-why should they pay the price of your obstinacy?”
From individual agony of a quarantine and ostracism to a global economic disaster- all based on report of a test with a pathetically low predictive value!
Jharkhand is a small state (where I live) with a low disease load and an even lower mortality rate. Ever since the first case was detected in the state, the local newspapers publish the new cases and every single death in great details.
So far, there have been approximately two thousand and five hundred positive cases as I write. Of the positive cases that were detected on surveillance of healthy contacts, not one developed even moderate symptoms during their quarantine period. Similar has been the experience in neighbouring Bihar and Odissa. So much for the denominator in our arithmetic.
Now the numerator- the number of deaths. In Jharkhand, there have been fifteen deaths so far. The sole criterion for the label of Covid death has been a positive RT PCR test. No patient has so far died of Covid pneumonia. Each single death has been due to a condition that would have caused the fatality irrespective of the test report- Brain haemorrhage, Renal failure, Myocardial infarction, advanced tuberculosis and as farfetched conditions as head injury due to fall from roof, abdominal injury due to road accident and complications after a child birth.
All were counted as COVID death based on a report the positivity of which has not caused a single illness out of over two thousand healthy persons. Similar is the story in the neighbouring states.
No one, yes no one knows for sure how many out of the half a million deaths reported in the world have actually been due to the coronavirus, to the exclusion of those with an incidental positive (true or false) report.
Isn’t it a piquant arithmetical function? Neither the numerator nor the denominator is anywhere near correct and yet the entire world is paralysed with mortal fear.
I always wondered why they called this arithmetic function as ‘vulgar fraction’ in arithmetic. Now I know why.
Is some form of Zohnerism (the use of a fact to lead a scientifically and mathematically ignorant public to a false conclusion) at work at the global level with a definite design? My gut feeling is a resounding YES.