> As for the article, this is modelling, not data (both the retrospective and the predictive). There is no direct causation being shown. You simply cannot make a claim like "X million people will die because USAID was terminated" based on this sort of evidence. It's not factual. It's not logical. Just emotionally driven claims intended to manipulate public opinion.

How would one (ethically nor not) perform a randomised trial of this kind of intervention? Like, the Lancet definitely hop on far too many bandwagons for my taste, but it's almost certain that the withdrawal of USAID will cause ~millions of otherwise avoidable deaths.

And ultimately, it's the choice of the US government and people whether or not to fund these programmes. Personally, I think that they are worthwhile, but others may disagree.

But really, you're holding the evidence required here to a very high standard, one that many things you believe would probably not meet.