Thanks so much for your thoughtful feedback.
To your first point - thank you for catching this inaccuracy. I have updated the text in that section.
To your second point - the models I created in this article probably aren't valid for several reasons (e.g. other assumption violations, confounding) and were for concept review / coding demonstration purposes only. You definitely bring up an important point here, though, especially since sample size is relevant conceptually. I therefore added notes both describing this specific problem with the WHO data and highlighting the fact that these models likely have other issues.
Thanks for reading,