Crises that aren't

From At least five interesting things to start your week (#25) by Noah Smith.  The subheading is Maternal mortality, "living paycheck to paycheck", media negativity, diffusion of the tech industry, and the effect of corporate tax cuts

In management consulting, you always have to be careful whether there is a real problem or only a signal problem.  Time and again you encounter an issue where there is a perceived problem.  You drill down into the data and then discover that a definition change occurred, the boundaries between territories moved, some contextual element changed without notation, etc.  Sometimes numbers are a signal, sometimes they are noise.  

Smith has the following example.

Maternal mortality in the U.S. is not as bad as we thought

For a number of years, Americans have been freaking out about high maternal mortality rates. In my New Year’s post I used maternal mortality as a major indicator of how much better the present is relative to the past, but the U.S. has been doing a lot worse on this metric relative to other countries. Here’s a graph from the Commonwealth Fund in 2022:

Source: Commonwealth Fund

Click to enlarge.

And the numbers have only gotten worse in recent years.

This just added to the overall sense of the U.S. as a nation in decline. It seemed to be of a piece with our low and falling life expectancy, our overpriced health care, and so on. It was one more fact that a lot of people knew that contributed to the general sense of crisis in the late 2010s.

Only there’s one slight problem: The recent increase may not have happened at all. Starting in 2003, U.S. states started rolling out a change to death certificates — a checkbox for pregnancy at the time of death. This resulted in a whole lot more deaths getting labeled as pregnancy-related. Joseph et al. (2021) tell the story:

Rigorous studies carried out by the National Center for Health Statistics show that previously reported increases in maternal mortality rates in the United States were an artifact of changes in surveillance. The pregnancy checkbox, introduced in the revised 2003 death certificate and implemented by the states in a staggered manner, resulted in increased identification of maternal deaths and in reported maternal mortality rates. This [paper] summarizes the findings of the National Center for Health Statistics reports…[C]rude maternal mortality rates did not change significantly between 2002 and 2018, [and] age-adjusted analyses show a [21%] temporal reduction in the maternal mortality rate…Specific causes of maternal death, which were not affected by the pregnancy checkbox, such as preeclampsia, showed substantial temporal declines.

And here is a graph:

Click to enlarge. 

They go on to show that causes of maternal death specifically associated with pregnancy (e.g. eclampsia and amniotic fluid embolisms) have declined, even as causes of death less related to pregnancy (e.g. hypertension and diabetes) have risen. That suggests that we’re looking at a change in how maternal deaths are reported, rather than a worsening of obstetric care in the U.S.

The mainstream medi is always in search of a calamity to tout.  Babies and mothers - what a great target.  Except, they are not.  Just a measure issue.  

Possibly.  Seems like some more effort is needed to be confident but at this point, it feels like one more fake story.