Letter to editor

摘要

Interpreting COVID-19 antibody tests. The implications of pre-test probabilities of 5% and 50% on the interpretation of COVID-19 antibody test if 1000 people were tested. The Cochrane review identified a mean sensitivity of 91.4% and mean specificity of 98.7% and these are here used for illustrative purposes. For the 5% pre-test probability, e.g. a patient with no symptoms but answers yes to a triage question, should the test have a lower sensitivity, particularly if the peak incidence and therefore likely time of infection is > 35 days ago, this would proportionally increase the false-negative and false-positive rate substantially. For example, for 57% specificity, the false-negative rate is 2.1% and the false-positive rate would be 21%. In contrast, if the pre-test probability is 50%, e.g. patient has atypical symptoms for several weeks, a positive test is compelling due to the low false-positive result, but the false-negative rate is still high at 8%. Antibody tests have high specificity, but sensitivity is variable and depends on the time since symptom onset