Tests to inform COVID-19 diagnosis
Topic Status Complete
This is a report on the effectiveness of tests that detect SARS-CoV-2 virus or antibodies to SARS-CoV-2, to inform COVID-19 diagnosis.
This report has been produced to assist the Welsh Government, and Health and Social Care in Wales, in their response to the coronavirus disease 2019 (COVID-19) pandemic. It provides a rapid review of published evidence on the effectiveness of virus and antibody tests to inform COVID-19 diagnosis. It is based on the most recent available evidence at the time of publication but will be updated frequently.
Plain language summary
Heath Technology Wales (HTW) looked for evidence to assess the accuracy, efficiency and cost effectiveness of testing for the presence of coronavirus 2 (SARS-CoV-2) for COVID-19 diagnosis. Tests for COVID-19 fall into two basic categories; those that test for the presence of the virus itself (‘virus’ tests) and those that test for the presence of antibodies produced in response to the virus (‘antibody’ tests). There are lots of different types of tests from both categories, most of which are done in a laboratory, but some can be done nearer the patient (these are known as ‘point of care’ tests), such as at hospital or GP clinics. In the studies identified, most of the tests were performed in laboratories on hospitalised patients who were either known to have COVID-19 or presumed to have it.
For both virus tests and antibody tests it is difficult to be sure of the accuracy of testing, as there is not yet a recognised way to conclusively prove a test result is correct, i.e. does a positive result really mean the virus/antibody is actually present and does a negative result really mean the virus/antibody is not present?
For virus tests, the best estimate is that, for every 100 people who are actually infected with the virus, only 89 would test positive. The remaining 11 would test negative despite actually being infected with the virus (essentially an incorrect or ‘false’ negative result). However, there are a number of different factors that appear to influence the results and accuracy of these tests, including the type of sample used, the body part sampled, and the length of time between when symptoms start and the test is given. The evidence for antibody tests can’t be summarised in the same way as virus tests because there are so many different antibody types and targets, and test accuracy varies widely.
Overall, for both virus tests and antibody tests there are important gaps in the evidence. There is limited evidence of how well testing works in the community, rather than in hospital, or in people who are infected with coronavirus but only have mild symptoms of illness. HTW will continue to review the evidence as more sources become available and will update our reports accordingly.