How many #COVID19 cases are asymptomatic? The *incorrect* 40% figure keeps being repeated (such as claim below in 21 Oct NEJM Perspective), but best estimates are now closer to 17% based on systematic reviews of valid studies.

2/ There are 13 reasonably good studies of asymptomatic rates (checked everyone in a well-defined at-risk group & > 7 days follow up) when combined suggest a rate of around 17% - https://jammi.utpjournals.press/doi/pdf/10.3138/jammi-2020-0030

3/ Note that we need to separate (a) the pre-symptomatic (= every infected person before symptoms!) from (b) the asymptomatic throughout infection (so need follow up to check if symptoms develop - many studies don't) - which some commentary & reviews don't do.

4/ Transmission rates are also lower by about 42% for asymptomatic cf symptomatic cases (p=0.047); 5 of 13 studies assessed this). Why? Could be less time infectious or that not coughing – less droplets, less environment contamination.

5/ Combined this means < 10% of transmissions likely to be from asymptomatic cases - so uncommon and chains of asymptomatic transmission are unlikely - but small enough that good #TestTraceIsolate can pick up most, and cut the chain links.

6/ IMPLICATIONS: Good #TestTraceIsolate finds many of both pre-symptomatic & asymptomatic folk - Iceland find 50% folk diagnosed in quarantine - but asymptomatic can only be isolated by contact tracing; pre-symptomatic can be isolated by contact tracing or becoming symptomatic.