Researchers Adjust Antibody Tests to Improve Study Data Quality
July 28, 2020
An important part of the research process is making the best decisions possible at each stage and continually making changes as more and better information becomes available. Learning about the novel coronavirus is real-time discovery.
Because COVID-19 antibody tests are relatively new, they must undergo a review process to confirm that they work and that their results are statistically “valid.” Tests are measured by sensitivity and specificity. Sensitivity is a measure of how well a test finds antibodies in someone who has been infected. Specificity is a measure of how accurate the test is at confirming negative results.
The antibody tests we have been using as part of the COVID-19 Community Research Partnership have undergone 3 separate reviews: 1 by Wake Forest Baptist Health, 1 by LabCorp, and 1 by the National Cancer Institute (NCI), at the request of the Food and Drug Administration (FDA).
In the first 2 reviews, the test sensitivity was 90% and test specificity was 100%. This meant that the test identified antibodies in 9 out of 10 samples from patients known to have COVID-19 and did not identify any antibodies in samples that had been collected and saved from before COVID-19 existed (no false positives).
In the third review, test sensitivity showed 93% for one type of antibodies (IgM) and 77% for the other type (IgG), meaning there were a larger number of false negative tests. Test specificity was 97.5% for IgM and 100% for IgG.
Although the test we have been using performed well on the first 2 reviews and the average of the 3 reviews showed overall good performance, the third review raised concerns that the test may not be sensitive enough (too many false negatives). As a result and because we continually want to improve the quality of our data, we have changed our antibody tests.
The at-home antibody tests in this study are used for epidemiologic research purposes only. We want to use the most accurate results we can, but the purpose of the testing is to get a good estimate of the number of people who may have been infected. We know there will be some false positives and false negatives. Your individual results should not be thought of as clinical testing and should not cause you to change your behavior.
Antibody test results in this study give us critical insights into the characteristics and spread of COVID-19 in our communities. We are still learning what it means to have COVID-19 antibodies. The virus is too new for anyone to have a full understanding of the level of protection the antibodies may offer you – and how long that may last.
At our current prevalence estimates (greater than 10%), a positive test is likely to be true. The sensitivity of the test that we have been using is less than 90%, so it has likely missed at least 10% of positives.
If you have questions about your results, please contact your primary care physician, or call one of our representatives at 336-70-COVID.
We encourage you to listen to your local officials and take appropriate precautions, such as wearing a mask, social distancing and frequent handwashing, to prevent the spread of this disease.
The COVID-19 Community Research Partnership collects antibody data with either an “at home collection device” made by Neoteryx, in which participants obtain blood samples via a finger prick and mail the sample back to us, or an “at home test kit” from Scanwell Health, in which participants obtain blood samples by a finger prick, perform the test themselves, and take a photo of the results to send via a smartphone app.
Test samples using the Neoteryx collection device will now go to LabCorp instead of Wake Forest Baptist Health for testing. The collection method and device remains the same.
Scanwell was using a test kit made by Syntron Bioresearch, but they now use test kits made by other companies. The collection method and use of the Scanwell smartphone app remains the same.
Using two different types of tests allows us to compare the results and ensure that we are getting consistent data. This is part of our ongoing assessment of the tests.
Scanwell Health varies the tests that they use based on availability, cost and quality.
We have been working with Scanwell Health for several months, helped them evaluate several tests on the market, and placed a strong emphasis on choosing a high-specificity test to avoid overestimating the spread of infection. Together, we selected a test made by Syntron Bioresearch that we believed to be the best available test at the time.
Since the test made by Syntron underperformed for sensitivity in the recent NCI review, Syntron will improve the test and resubmit it for another review. In the meantime, Scanwell will use other tests that have performed well in all the categories on the NCI review.