Sensitivity and specificity: taking the measure of at-home tests
DEAR DR. ROACH: How reliable are home COVID-19 tests? I’ve read one article that said that if you tested negative you could be certain that you did not have COVID-19, but if you tested positive there was a significant chance of a false positive. Another article said exactly the opposite. What are the facts? — A.L.
ANSWER: You’re asking about how accurate the home tests are. There are two primary measures of tests: the sensitivity and the specificity. Sensitivity is the probability that a given test will detect the condition, if it’s there. For at-home COVID-19 tests, the sensitivity has been reported to be between 79% and 95%. However, how the sample is collected can affect the sensitivity, and the real-world performance may not be quite as good as what the manufacturers report. Timing is also important, and the optimum time for testing is probably three to five days after an exposure. The test is most likely to be positive just before and just after a person begins having symptoms.
The specificity is the probability that the person really has the condition when the test is positive. At-home tests generally perform well, with reported specificity ranging from 92% to 99%.
So, the tests can give both false positives and false negatives. A negative test is only somewhat reassuring: I have had several patients have negative at-home tests but positive in-hospital PCR tests. However, the specificity is high: If you test positive, the likelihood is high that you really have COVID-19.
One important point, though. Test results need to be interpreted within a person’s context. If you were exposed to a person known to be contagious with COVID-19 and now have fever and cough, a positive test result is virtually certain to mean you have COVID-19. However, if you have had no exposures, and have been sitting at home not seeing anybody, and you take a test that comes out positive, that’s almost certain to be a false positive. The calculation of the positive predictive value — when the test result is positive, do you really have COVID-19? — relates to both the likelihood you had COVID-19 before knowing the test results and the test characteristics. The same is true of the negative predictive value, i.e., if the test is negative, are you really free from COVID-19? Your doctor can help you interpret the results for your specific situation.
DEAR DR. ROACH: I used to give whole blood two to five times a year but several years ago I went on meds for an enlarged prostate and was told by the Red Cross that they could not use my blood due to the risk of birth defects if a pregnant woman received blood containing finasteride.
Aren’t platelet donations washed? Could I donate platelets? — J.T.
ANSWER: In general, the requirements for donating platelets are the same as for donating blood. That means you should wait a month after taking finasteride before donating blood or platelets.
I do understand that the risk to a pregnant recipient of the blood is lower than if she received whole blood (or packed red blood cells), but blood banks are very cautious about ensuring the safety of the blood supply.
I had another reader in the same situation ask whether his blood could just be given only to men; again, the blood banks want maximum safety and flexibility in giving blood products.
Dr. Roach regrets that he is unable to answer individual letters, but will incorporate them in the column whenever possible. Readers may email questions to ToYourGoodHealth@med.cornell.edu or send mail to 628 Virginia Dr., Orlando, FL 32803.