Do Covid-19 tests get it wrong?

Testing for the presence of Covid-19 has been used to inform policy in many sectors. With regards to the off-shore industry pre-mobilisation testing has undoubtedly contributed to the enviable safety record of the industry during the current pandemic, however all tests are associated with potential inaccuracies.

The key questions to ask about any test are: how sensitive is the test? And how specific is the test? Or in other words how many cases that have covid-19 will we miss with the test (false negative rate)? And how many cases will be said to have infection when they don’t (false positive rate)? 

Most effort so far has been invested in ensuring high test sensitivity due to the devastating consequences of undetected cases in health-care and social care settings, and the propagation of the epidemic especially by asymptomatic or mildly symptomatic patients.  This is potentially even more important if individuals are being sent to a remote environment such as the North Sea.  

The ability of any test to detect the presence of the virus relates to the viral load present.  At high viral loads almost all commercially available tests will detect the presence of the virus.  Each test company will be able to show high sensitivity values however the situation on the ground is often very different -the results quoted for assays are often tested under idealised conditions with hospital samples containing higher viral loads than those from asymptomatic individuals living in the community. As such, diagnostic or operational performance of swab tests in the real world might differ substantially from the quoted analytical sensitivity and specificity i.e. lateral flow tests in ideal conditions indicated a sensitivity of 76% but this fell to just over 50% when tested in the field Even the gold standard PCR with a quoted sensitivity in excess of 98% in the lab is likely to run with a 90% sensitivity for most tests.

There is always a balance between increasing the sensitivity of the test to reduce the risk of disease transmissions and increasing the number of people forced to isolate or being excluded from the workforce.  

When discussing Covid-19 the term false positive has slipped into common parlance and has been taken to mean anyone who tests positive but does not go on to develop symptoms of the disease.  Our own data in testing of asymptomatic people indicates that only 20% of individuals with positive PCR tests subsequently develop symptoms.  Does this mean the other 80% are false positives?  The answer is clearly no.  We know that many infections are spread by asymptomatic carriers of the virus – i.e. the experience on board the cruise ship the Diamond Princess and also on board the USS Theodore Roosevelt.  Our own data show there is significant risk of spreading the disease in anyone with detected virus on PCR.

A true false positive test represents a mistake in the diagnostic process – it means that the presence of virus is wrongly attributed to someone with no virus and arises from contamination and should not happen provided proper laboratory governance and controls are in place.  It also emphasises the need to conduct testing in regulated centres with health care professionals.

There has been much discussion about whether the amount of virus detected is an indication of whether someone is infectious or not.   It is likely that higher viral loads are associated with increased infection risk.  Although PCR testing is not a quantitative analysis the Ct value does reflect the viral load.  In general, low Ct values (higher viral loads) are associated with higher risk of infection but the results of PCR testing only tell you what is happening on the day of testing.  We don’t know whether someone with low viral load has had a higher load the day before or whether it is increasing.  Perhaps the answer is to re-test but, as discussed above, the difficulty is that at low levels of virus the accuracy of the test is reduced and a second test that is negative may not mean too much.  It’s a bit like the needle in the haystack – if you find it its definitely there but if you don’t it doesn’t mean it’s not.  For that reason, public health authorities tend to take the line that if someone is a known contact of a case of Covid-19 they need to isolate irrespective of the results of any test.

Should there be any exceptions?  Every case must be examined in its own right and there are now situations of described of prolonged shedding of virus in a variety of case reports for COVID and is ubiquitous for RNA viruses detection. SARS-CoV, MERS, Influenza Ebola and Zika viral RNA can be detected long after the disappearance of the infectious virus.

The immune system works to neutralise the virus and prevent further infection. Whilst an infectious stage may last a week or so, because inactivated RNA degrades slowly over time it may still be detected many weeks after infectiousness has dissipated.  Over the course of the last 8 months of testing and in excess of 80,000 tests we have identified 12 cases that are persistently positive on PCR testing for Covid-19.  In these situations, it is important to look at whether there are other markers of viable virus infection and whether there is any evidence of immunity, but it is clear these cases are not the norm.  For individuals going to remote environments and where there is the very real risk of causing a significant outbreak of the disease i.e. North Sea oil rigs etc it is clear that the risks of labelling cases as false positive are significant.  

Leave a reply