Underreporting of SARS-CoV-2 Infections in British Columbia

Our understanding on the epidemiology of COVID-19 is limited by the ability of health systems to ascertain SARS-CoV-2 infections. Due to asymptomatic infection and testing hesitancy in symptomatic individuals (which may vary throughout time and with age, disease severity, socioeconomic status, etc.), reported cases represent only a fraction of total infections. The gold standard to estimate the burden of disease in a population is seroprevalence surveys. However, such surveys are operationally expensive and provide only a snapshot on the ascertainment of cumulative infections. In the current poster, we introduce a method to estimate instantaneous infection ascertainment from the the instantaneous case-hospitalization fraction and the infection-hospitalization fraction (which itself is constrained from seroprevalence data). We applied these methods in an age-structured manner to estimate the instantaneous ascertainment of SARS-CoV-2 infections in the B.C. population, from the outset of the epidemic to date. This allowed us to back-calculate the true epidemic, from which we could more acutely identify epidemiological trends.