I recall backing out the SA data and running ROW next to US across all the endpoints, as I recall when considering the CIs the data were similar, ROW a bit better at the point estimates
In general, IMO, they dropped 33wk, dropped the regional vs 33wk play, and seeked to speak towards the totality of the data. (South Africa was not mentioned).
Simply say it works, the data is bumpy, and that a P4 will tell us more. Hopefully.
But up until this point in this presentation they had not indicated that the data was regionally highly variable, they need as a matter of scientific responsibility to acknowledge that the data shown in this presentation was by no means uniform ... (implied is that there are going to be places that did worse, to the point of deliberate mentioning it, so much worse).
There is nothing gained in the regional breakout emphasis or in the 33wk emphasis. The lack of uniformity is a weakness of the trial. So they ignored it (mostly) in this presentaiton, but were obligated to acknowledge it at some level, this was an effort to tip their hat towards that without making it a focal point. This slide let us know that under the hood some of this is a bit messy, and did some hand waving for an explanation of that.
That is how I see it.