Dear Mr Harfort,
I am writing about your very interesting and solidly supported podcast —as is all the material you produce—on the US maternal-related mortality statistics,. You rightly conclude that the upward trend in the last two decades for the US results from a statistical mirage—the coverage of the more accurate new format proposed by the relevant international agency progressed piecemeal, one state at a time over two decades, accounting for the index's substantial growth. That is, the growth is only partly due to the improved accuracy of the new format and most probably due to the increasing temporal coverage of the sample.
I put together a graph showing this, together with equivalent indices for the major European countries. (Unfortunately this format does not allow attaching a picture.) The graph shows a near-constant level over time for the latter and a marked increase in the American one, as you pointed out. However, you also conclude that US information is utterly worthless, and with this, I disagree.
Although one should dismiss the growth trend, we cannot overlook the differential in the levels, which shows that even the lowest US figure (at the beginning of the period) is much higher than any other major advanced economy. Thus, in 2003, the US index, which supposedly underestimated mortality since the new format was not in place, was already about 50% higher than the highest European one. By the end of the period, when supposedly the US data becomes comparable to the European, it is 2.5 times as high. I think this finding is most revealing of the US's precarious, backward healthcare condition. And to this extent, it is not worthless.
