Evaluation of the number of undiagnosed infected in an outbreak using source of infection measurements

Multiple studies have been conducted to predict the impact and duration of the current COVID-19 epidemics. Most of those studies rely on parameter calibration using the published number of confirmed cases. Unfortunately, this number is usually incomplete and biased due to the lack of testing capacities, and varying testing protocols. An essential requirement for better monitoring is the evaluation of the number of undiagnosed infected individuals. This number is crucial for the determination of transmission prevention strategies and it provides statistics on the epidemic dynamics. To estimate the number of undiagnosed infected individuals, we studied the relation between the fraction of diagnosed infected out of all infected, and the fraction of infected with known contaminator out of all diagnosed infected. We simulated multiple models currently used to study the COVID-19 pandemic and computed the relation between these two fractions in all those models. Across most models currently used and for most realistic model parameters, the relation between the two fractions is consistently linear and model independent. This relation can be used to estimate the number of undiagnosed infected, with no explicit epidemiological model. We apply this method to measure the number of undiagnosed infected in Israel. Since the fraction of confirmed cases with a known source can be obtained from epidemiological investigations in any country, one can estimate the total number of infected individuals in the same country.