That is a very low quality study because it a) employed self-selection instead of a proper RCT design and b) didn't control for anything. At least tracking pre-existing conditions would have been necessary, when not randomizing, for the results to have any meaning.

When studying the effects of invermectin, at least checking for worms in the patients' feces would have been needed.

Post in the thread [1] gives a very plausibly-sounding hypothesis for seeing different results in non-western countries like your link.

[1] https://news.ycombinator.com/item?id=30720766