Recent estimates (including our analysis led by @timwrussell: eurosurveillance.org/content/10.280…) have suggested 0.5–1% of COVID-19 infections may potentially be fatal overall. Some have interpreted this to mean COVID could be similar to flu, but they're making a crucial mistake... 1/
To make a comparison with influenza, we need to look at all influenza *infections* (including asymptomatic ones), then compare with deaths. Remember, in epidemiology, always make sure you're comparing what you think you're comparing: 2/
So how often do people get infected with influenza (with or without symptoms)? A few years ago, I looked at this with @SRileyIDD & others – using serological data we estimated 40-50% of younger groups were infected each year, and 15-20% of older groups. journals.plos.org/plosbiology/ar… 3/ 
Early data from a very detailed study in South Africa found similarly high values - 369/1118 (33%) of participants had a PCR confirmed flu infection during the year isirv.org/site/images/co… 4/
So if we're calculating risk infection is fatal, we need to compare total flu infections - potentially 1/3 of population - with annual deaths. As rough estimate, this suggests 0.02-0.05% of flu infections are fatal, i.e. 10x lower than COVID (Data: cdc.gov/flu/about/burd…) 5/
More on interpreting fatality risk for the 2009 flu pandemic based on estimates of symptomatic cases and infected individuals: ncbi.nlm.nih.gov/pmc/articles/P… 6/6/
