Fig. 9

Estimating fraction of infections detected. a Naive CFR among adults aged 61–70 years by week of specimen collection alongside expected IFR among this age distribution, as estimated by Levin et al. [13] or O’Driscoll et al. [14]. Shaded regions show 95% confidence intervals. b HFR among admitted 61–70 year-olds by week of specimen collection (solid red) with standard error of proportions (shaded red), fitted sigmoid curve (solid black) and 95% prediction interval (dashed black). c Fraction of all deaths reported by date of death, based upon comparison of COVID-19 mortality and excess all-cause mortality. d Estimated fraction of SARS-CoV-2 infections among adults aged 61–70-years that were detected by surveillance: assuming that all deaths are reported and IFR is stationary; e assuming all COVID-19 deaths are reported and IFR is non-stationary due to improving clinical outcomes; f and assuming that excess deaths are unreported COVID-19 deaths and IFR is non-stationary. Dark blue estimates use IFR estimates from Levin et al. [13] and light blue from O’Driscoll et al. [14]. Shaded regions in D-F are 95% credible intervals with 1000 bootstrapped samples