In this blog for anyone interested in rapid point-of-care tests for diagnosing COVID-19 infection, Jac Dinnes from University of Birmingham discusses her recently published Cochrane Reviews are systematic reviews. In systematic reviews we search for and summarize studies that answer a specific research question (e.g. is paracetamol effective and safe for treating back pain?). The studies are identified, assessed, and summarized by using a systematic and predefined approach. They inform recommendations for healthcare and research. More to help healthcare professionals and members of the public interpret the evidence for their accuracy.
This page was updated on 13 April 2021
Tests for diagnosing COVID-19 infection are an incredibly important tool to help reduce the spread of infection in our communities, schools and workplaces, and have received a huge amount of attention in the press and on social media over the last year. Measuring the accuracy of tests is not always straightforward or easy to understand, especially when the same test can be used for different purposes and in different groups of people.
What are rapid point-of-care tests?
Our Cochrane Review of rapid point-of-care tests includes two types of test, antigen tests and molecular tests. Both types of test use swab samples taken from the nose or throat, can be used outside of a specialist laboratory, and provide results in less than two hours. This makes them suitable to use at the ‘point-of-care’ or while the person tested waits for the result, unlike the PCR test which often takes up to 24 hours to provide a result.
Rapid molecular tests detect the virus’s genetic material using methods similar to the PCR test. They require more equipment and are more expensive to use than rapid antigen tests.
Rapid antigen tests (commonly known as ‘lateral flow tests’ or LFTs) are the ones that lots of people will have heard of – they come in small disposable plastic cassettes that look a bit like pregnancy tests. The swab sample is applied to the cassette and a line appears when antigen from the virus is detected – usually within 20 to 30 minutes.
The use of LFTs is rapidly expanding across the world so for the rest of this blog I will summarise what our review tells us about the accuracy of LFTs, look at some of the most common uses for the tests, and how we might expect tests to perform in different groups of people.
How do we measure whether a diagnostic test is ‘accurate’ or not?
The accuracy of a test is not just about how well the test identifies people who have a disease or condition but also accounts for test results in people who do not have the disease. Two measures are used. A measure of a screening or diagnostic test’s ability to correctly detect people who have the disease. More is the proportion of people with the disease who get positive results on the test. A measure of a screening or diagnostic test’s ability to correctly identify people who do not have the disease. More is the proportion of people who don’t have the disease who get negative results on the test. The closer sensitivity and specificity are to 100%, the more accurate the test.
Sensitivity and specificity are not fixed values but can vary because of differences in the populations tested, the way the tests are done, or the way that the presence of the disease is confirmed. All else being equal, it is important to establish how well a test works in different populations of people, for example in people with or without symptoms, or in children compared to adults.
What is the evidence for LFTs?
In our Cochrane Review Rapid, point‐of‐care antigen and molecular‐based tests for diagnosis of SARS‐CoV‐2 infection, which we updated for the first time in March 2021, we included 48 studies evaluating 16 different LFTs published up to 16 November 2020. They tested nose or throat samples from a total of 20,168 people including 5,666 who had COVID-19 confirmed by a PCR test.
Although more than half of studies included people at COVID-19 test centres or who had been in contact with someone with a positive test (so meeting national criteria for covid tests), others included anonymised samples left over after PCR tests, or did not provide enough information for us to judge who the participants were. LFTs were often interpreted without knowing the result of the PCR test (so the PCR result could not have influenced whether the test was judged to be positive or negative). However, the tests were not always done in the way that the test manufacturers recommend. This means that different results might be seen when the tests are used in practice.
In people with COVID-19, LFTs correctly identified the infection in an average of 72% of people with symptoms, compared to 58% of people without symptoms (test sensitivity). For people with symptoms, tests had higher sensitivity when used in the first week after symptoms first developed (an average of 78% of confirmed cases had positive LFTs). This is likely to be because people have the most virus in their system in the first days after they are infected.
In people who did not have COVID-19 (had negative results on PCR tests), LFTs correctly ruled out infection in 99.5% of people with symptoms and 98.9% of people without symptoms (test specificity).
There was a lot of variation in the sensitivity of different brands of tests. We combined results from different studies that evaluated the same tests. When we only included studies that followed the manufacturer instructions, only one test (SD Biosensor STANDARD Q) met international standards of ‘acceptable’ sensitivity and specificity.
Possible benefits and harms from using LFTs
The possible benefits from using LFTs are mainly related to the speed of diagnosis. Faster diagnosis of cases of infection should allow people to self-isolate or quarantine more quickly, and could trigger contact tracing. Modelling studies have shown the biggest impact on transmission of infection happens if contact tracing is started within 24 hours of a positive test.
Harms from using LFTs result from missing true cases of infection (In diagnostic tests, a conclusion that a person does not have the disease or condition being tested, when they actually do. More results) or from positive test results in people who are not infected (false positives). Negative LFT results in people who have COVID-19 could lead to increased transmission of infection, especially if the negative result encourages them to be less strict about social distancing and other infection control measures. Positive LFT results in people who do not have COVID-19 may lead to unnecessary self-isolation (including time off school or work) for the individual and their close contacts and additional confirmatory testing.
The balance of benefit to harm, and relative impact from false negative or In diagnostic tests, a conclusion that a person does have the disease or condition being tested, when they actually do not. More results, changes depending on the underlying percentage of people with confirmed COVID-19 (The proportion of a population who have a particular condition or characteristic. For example, the percentage of people in a city with a particular disease, or who smoke. More) amongst those being tested. We show this changing balance of benefits and harms in the examples below using summary results for the SD Biosensor STANDARD Q test in studies that followed the test manufacturer’s instructions.
Using lateral flow tests in people with symptoms of COVID-19
Used in people with symptoms, the STANDARD Q test correctly identified 88.1% of people with COVID-19 (sensitivity) and 99.1% of people with no COVID-19 infection (specificity).
The two examples here show that when used in people with symptoms, the relatively small number of false positives from LFTs could be an acceptable trade-off for faster diagnosis and contact tracing of the people correctly diagnosed with infection. The number of cases missed (false negatives) could be of concern, especially when the underlying number of cases of infection (prevalence) is high. This reinforces the need to follow preventive measures and self-isolation following a negative LFT result.
Using lateral flow tests in people with no symptoms of COVID-19
LFT manufacturers usually do not evaluate their tests in people who do not have symptoms of COVID-19 (asymptomatic people). Although people without symptoms can have high levels of virus (high viral load) for a similar length of time as those with symptoms, it is difficult to pinpoint when that will be, especially if they have had no known contact with someone who has tested positive.
When it was used in people with no symptoms (one An investigation of a healthcare problem. There are different types of studies used to answer research questions, for example randomised controlled trials or observational studies. More of 127 people), the STANDARD Q test correctly identified 69.2% of people with COVID-19 (sensitivity) and 99.1% of people with no COVID-19 infection (specificity). Because of the very small numbers, we can’t be sure that these percentages are correct.
This example illustrates the considerable change in the balance of harms and benefits from LFTs when applied to a much bigger group of people with very low prevalence of infection, and using a test with relatively ‘good’ sensitivity and high specificity for detecting asymptomatic infection. Although some cases are correctly identified, these can easily be outweighed by the number of people with false positive results who, along with their close contacts, have to take time off work or school. The impact from false positive results could be lessened by following up with a PCR test.
LFTs will continue to be an attractive and potentially important tool for managing the spread of COVID-19 infection. Most of the evidence available to date is based on people with symptoms and we are much less certain as to how well LFTs perform in people with no symptoms. We know that more studies are now available and will produce an updated version of this review in the near future.
Please note, we cannot give specific medical advice and do not publish comments that link to individual pages requesting donations or to commercial sites, or appear to endorse commercial products. We welcome diverse views and encourage discussion but we ask that comments are respectful and reserve the right to not publish any we consider offensive. Cochrane UK does not fact check – or endorse – readers’ comments, including any treatments mentioned.
People who comment on our blogs often have further questions, including about the effects of treatments. If you have a question about a health claim you have heard or read (e.g. does the use of petrol pumps spread COVID-19 (coronavirus disease)?) you can submit it to iHealthFacts, a fact-checking resource.
Dinnes J, Deeks JJ, Berhane S, Taylor M, Adriano A, Davenport C, Dittrich S, Emperador D, Takwoingi Y, Cunningham J, Beese S, Domen J, Dretzke J, Ferrante di Ruffano L, Harris IM, Price MJ, Taylor-Phillips S, Hooft L, Leeflang MMG, McInnes MDF, Spijker R, Van den Bruel A. Rapid, point‐of‐care antigen and molecular‐based tests for diagnosis of SARS‐CoV‐2 infection. Cochrane Database of In systematic reviews we search for and summarize studies that answer a specific research question (e.g. is paracetamol effective and safe for treating back pain?). The studies are identified, assessed, and summarized by using a systematic and predefined approach. They inform recommendations for healthcare and research. More 2021, Issue 3. Art. No.: CD013705. DOI: 10.1002/14651858.CD013705.pub2. Accessed 01 April 2021.
Declaration of interests:
Dr. Dinnes reports grants from NIHR Birmingham Biomedical Research Centre at the University Hospitals Birmingham NHS Foundation Trust and the University of Birmingham (Grant Reference Number BRC-1215-20009), during the conduct of the study.