Estimating True Prevalence Through Questionnaire Data

A method using questionnaire data for estimating the level of under reporting during an outbreak is presented. It is based on rewriting the conditional probabilities for getting tested, being infected, and having symptoms. It shows very good agreement with seroprevalence studies of blood donors. On the one hand, this shows the strength of questionnaires when testing the general population during an outbreak as a means to find the true prevalence. On the other, applying it to covid-19 demonstrates that the asymptomatic cases likely make up around 50% of the infected.