This picture shows that older people are more likely to need hospitalisation. Once the authors adjust for age, the effect of number of children is not significant. Quite inappropriate that the main figure is the _unadjusted_ analysis. https://twitter.com/dr_michaelmarks/status/1310567170091474951

Like, the one thing we absolutely know about this virus is that bad outcomes are very strongly correlated with age. Association tests of anything to outcome that are not age adjusted should trigger massive alarm bells.

Also, showing survival curves without confidence intervals is misleading. Keen eye can immediately sense something is up by the coarseness of the yellow and green lines. See, e.g. below by @alehins .