This blog site has been archived

Testing for COVID-19: still many unknowns

Blog entry written on: ‘Effectiveness of tests to detect the presence of SARS-CoV-2 virus, and antibodies to SARS-CoV-2, to inform COVID-19 diagnosis: a rapid systematic review’ (bmjebm-2020-111511).

Authors: David Jarrom, Lauren Elston, Jennifer Washington, Matthew Prettyjohns, Kimberley Cann, Susan Myles, Peter Groves.

As the COVID-19 pandemic spread at the start of 2020, there was a global effort to develop interventions to identify and monitor SARS-CoV-2 infection. Diagnostics have been developed and adopted at unprecedented speed, accompanied by rapidly published research on these tests.

But are these tests accurate and how should they be put to best use?

Our systematic review searched for evidence on the accuracy and utility of:

  • Tests that detect SARS-CoV-2 virus.
  • Tests that detect antibodies against SARS-CoV-2.

Both test types have the potential to be valuable in different ways. Virus tests can diagnose whether we are currently infected with SARS-CoV-2 (with or without symptoms), thereby informing the need for treatment and/or isolation. But an inaccurate test could lead to people who are infected risking further exposure to others (a false negative) or people who are not actually infected isolating unnecessarily (false positive).

Antibody tests can be used in the same way, although arguably have more use post-infection to tell us whether we have had SARS-CoV-2 infection or not. This has the potential to inform the seroprevalence (how many people have already had the disease) within communities or the general population. However, whether previous SARS-CoV-2 infection results in immunity from further infection is uncertain.

There is much we still don’t know about the best way that SARS-CoV-2 testing should be used, but our review raised interesting possibilities for future testing strategies. The evidence mainly used laboratory-based tests to detect SARS-CoV-2 or antibodies to the virus: evidence on point-of-care or near-patient testing was lacking. Point-of-care tests claiming to accurately detect SARS-CoV-2 in minutes or hours have limited evidence on their use in people who needed testing: most had only been used on laboratory samples, and often only a dozen or so samples. This tells us little about how they will perform in the real world.

Despite the thousands of COVID-19 articles published each week, only a small proportion evaluated the diagnostic accuracy of COVID-19 tests; we found 38 studies for virus tests and 25 studies for antibody tests. Using the QUADAS-2 tool, we found that most evidence was at risk of bias, often due to the test only being performed in hospital patients with known COVID-19 infections.

Nevertheless, our results show some promising findings on how testing has been used to date. PCR-based testing, which has been by far the most widely used type of testing to date, has a high negative predictive value (NPV): in other words, it is highly likely that negative results are truly negative (we estimate between 96 and 99%, depending on the disease prevalence). Positive predictive value (PPV), the likelihood that a positive test result is truly positive and the disease is present, was more difficult to estimate. PPV is also more dependent on prevalence, something that is crucial to consider as wider-population-based screening strategies are rolled out, in order to avoid overwhelmingly high numbers of false positive tests.

The work described in this review was largely done in the first wave of the pandemic, and covers evidence published up to May 2020. Since then Health Technology Wales, have carried out further work assessing the evidence on COVID-19 diagnostics through our membership of the European Network for Health Technology Assessment (EUnetHTA). We co-authored a wider review on the role of antibody tests for novel coronavirus SARS-COV-2 in the management of the COVID-19 pandemic. HTW are also the lead authors of the ongoing rapid collaborative review on the diagnostic accuracy of molecular methods that detect the presence of SARS-CoV-2 virus in people with suspected COVID-19, due for publication in November.

Covid-19 Test


Image of Author

David Jarrom

Senior Health Services Researcher, Health Technology Wales

Image of Author

Lauren Elston

Health Services Researcher, Health Technology Wales

Image of Author

Jennifer Washington

Information Specialist, Health Technology Wales

Image of Author

Matthew Prettyjohns

Principal Researcher, Health Technology Wales

Image of Author

Kimberley Cann

Health Services Researcher, Health Technology Wales

Image of Author

Susan Myles

Director, Health Technology Wales

Image of Author

Peter Groves

HTW Chair, Health Technology Wales

COI: All authors declare that they have no conflicts of interest.


The views and opinions expressed on this site are solely those of the original authors. They do not necessarily represent the views of the BMJ and should not be used to replace medical advice. All information on this blog is for general information, is not peer-reviewed, requires checking with original sources and should not be used to make any decisions about healthcare. No responsibility for its accuracy and correctness is assumed by us, and we disclaim all liability and responsibility arising from any reliance placed on such commentary or content by any user or visitor to the Website, or by anyone who may be informed of any of its content. Any reliance you place on the material posted on this site is therefore strictly at your own risk.

(Visited 465 times, 1 visits today)