They Told Us Women Were Dying From Pregnancy-Related Complications By The Droves. They Were Wrong.
Honest mistakes in data collection? Probably. Was an agenda pushed anyway? Absolutely.
A 2016 study reported that Texas’ maternal mortality numbers were alarmingly high. According to the study, from 2010 to 2012 the number of maternal deaths occurring due to pregnancy-related issues skyrocketed from 72 to 148. The study made national headlines.
There was just one problem: those statistics are false.
The real number, for Texas in 2012, turns out to be 56—which, far from registering as a public health panic, actually represents a decline.
Why did the study misreport the numbers so egregiously?
According to a write-up in the Chicago Tribune, it was “human error” that likely caused the data collection slip-ups. Ars Technica, for their part, blame “a crummy drop-down menu,” a reference to Texas’ digital process for entering a death certificate, which apparently had juxtaposed the “not pregnant within the past year” and “pregnant at the time of death” options.
But what’s interesting is that misreporting maternal mortality appears to be a national problem.
Elliott Main, the medical director of the California Maternal Quality Care Collaborative, told The Washington Post that the data errors found in this single year of data in Texas are indicative of a much more widespread maternal mortality data collection problem on a national scale, which the Texas researchers noted in their article, as well.
At the same time, only Texas seemed to get the numbers so outlandishly wrong. If we accept the above argument, is the takeaway simply that Texas employs the worst, most error-prone data entry team in the United States? Is that what we should conclude?
Setting Texas aside for the moment, the U.S. does appear to have higher levels of maternal mortality than other advanced countries. Is Main right that this is a nationwide “data collection problem”?
Researchers suggest part of the reason maternal mortality rates have gone up is we have increased our capacity to identify qualifying cases. (Yet as the Texas misreporting scandal shows, at times we’re able to “identify” these cases a bit too easily.) While it’s true that we’ve gotten more precise in our data collection practices, it’s also likely that as agencies switch to new reporting or measurement systems, temporary spikes due to over-reporting account for some of the increase.
Last summer, I researched those scary stories coming out about our maternal mortality rate and found that the narrative is wrong.
Some of it has to do with issues surrounding the reporting of late maternal deaths, defined as one that occurs more than 42 days, but less than a year, after the end of a pregnancy.
Are we over-reporting late maternal deaths?
There is a case to be made that we are.
The massively influential study put out by the Bill & Melinda Gates Foundation on this very topic — called “Global, Regional, and National Levels of Maternal Mortality, 1990–2015: A Systematic Analysis for the Global Burden of Disease Study” — alludes to possible over-reporting.
In a “Research in Context” summary box on page 1776 of the study and on a detail-rich section on page 1800, we get the following (emphasis mine):
Late maternal death statistics need to be improved. Maternal mortality surveillance studies such as confidential enquiry have showed that late maternal death is non-trivial in even low-resource settings and can account for up to 40% of maternal deaths in high-income settings. A contemporary linkage study in Mexico found that 18% of maternal deaths are missed when the definition is truncated at 42 days’ post partum. As immediate mortality continues to decrease as a result of improved antenatal, bobstetric, and post-partum care, it is therefore increasingly likely that the proportion of late maternal deaths will continue to increase. Despite knowledge of its importance, only a few countries using ICD-10 reliably code late maternal deaths. This is especially egregious because many of the same countries who have completed multiple confidential enquiries also have not recorded a single late maternal death in their official statistics. Denmark, Ireland, Finland, and the UK all fall into this category. Australia, France, and South Africa likewise completed multiple confidential enquiries and have recorded a total of eight maternal deaths combined in the entirety of their official statistics. This is the exact inverse of the USA where no nationally comprehensive confidential enquiries have been completed (although some states have established maternal mortality review boards). The USA has high MMR for a high-SDI country — and is one of the few where it is increasing — but following the lead of Mexico and much of Latin America, it is also one of the only countries that has proactively improved its civil registration system with addition of a pregnancy checkbox on the standard death certificate, so it is possible that at least a portion of the increase is related to enhanced case ascertainment.
How much of that increase is related to enhanced reporting? According to this interview with researcher Marion MacDorman of the Univeristy of Maryland, perhaps as much as 80 percent. In other words, only 20 percent of the reported increase represented a real rise in the maternal mortality rate.
That is still a hefty rise, and something we want to tackle. I have not written any of these articles to dismiss real problems in maternal mortality but to focus on the problems we actually have in the United States. Pregnancy involves all sorts of complications, many of which call for significantly different preventions and solutions. This is precisely why the late maternal death category needs better tracking.
Here’s a start: We should acknowledge that late maternal death — pregnancy-related death between 42 and 365 days postpartum — is associated with higher maternal age, which is rising in wealthy western countries. Here are figures for the U.S.:
And here’s a similar graph, showing that it’s a trend across racial and ethnic groups. (“Similar” because it looks at “first-time” mothers rather than pregnancies en toto. The extrapolation is safe in this case.)
We might spend some time asking why pregnancy is being delayed, but that’s a piece for a different moment. The point I’m making now is that we need to take great care to identify causes of increased maternal mortality so that we may address them with specificity and skill. Accuracy is essential. Panic causes mistakes, and last summer’s headlines were panic — unnecessary panic.
Old Narrative: The U.S. Was Bad, But Texas Was The Worst
As the narrative about Texas’ apocalyptic maternal mortality numbers was forming, many took the opportunity to push an agenda. For these opportunists, the underlying cause(s) did not seem to matter.
A Texas task force was refreshingly straightforward about this last May. A maternal mortality report released in anticipation of the Special Session mentioned potential causes for this spike, including what researchers thought (cardiac problems and drug use), what health advocates thought (cuts to Planned Parenthood funding and no Medicaid expansion), and what experts thought (obesity, age, and access to care).
The report then dismissed these potentialities with “no matter the cause…” language and went on to discuss human rights, demographics and education disparities, HIV, and political power. It’s not that these factors are irrelevant, but the statistic that a third of the deaths resulted from cardiac or blood-pressure complications appeared to get lost in the shuffle.
Yet, as I pointed out above, the maternal mortality rate spike in Texas that made national headlines was found to be overstated by about 50 percent, due to “new system error.” From the Houston Chronicle:
The new data, published in the journal Obstetrics & Gynecology Monday, found 56 pregnancy-related deaths among Texas residents in 2012, compared to 147 the health department previously supplied to national databases. As a result, the 2012 rate becomes 14.6 deaths per 100,000 live births, significantly down from the the previous rate of 38.4. …
Sonia Baeva, a Texas health department maternal mortality and morbidity epidemiologist and the lead author of the study, attributed the erroneous death certificate information to a more than 40 percent increase in the the proportion of death certificates that were submitted electronically from 2010 to 2012. On a drop-down list in the electronic system, the “pregnant at the time of death” option is directly below the “not pregnant within the past year” option.
The day after this news broke, more worry about what this might mean for Planned Parenthood and Medicaid funds emerged. Again from the Houston Chronicle, this time from Lisa Falkenberg:
[Dr. Lisa] Hollier [OBGYN in the Texas Medical Center and head of the state task force] seemed confident that the lawmakers, physicians and representatives from medical societies she had spoken with since the report came out remain supportive of improving health access and outcomes for women.
Forgive me if I’m not so sure. In Texas, a state whose only motivation to reform bad policy seems to be high-profile crises drawing national media coverage and scathing federal court rulings, I’m worried that when the glare of the national spotlight fades, so will the attention on women’s health and maternal deaths.
The next quote concerns Medicare funding for postpartum depression, followed by one that starts, “Regardless what this report says…” from a state representative from Houston. Then there are two paragraphs on Planned Parenthood funding, because supposedly they offer prenatal or postpartum care? Eventually, Falkenberg hits on a relevant problem: Medicare funding used to cut off at 60 days postpartum. State agencies have changed that but, according to Dr. Hollier, there are some paperwork problems — typical bureaucratic paperwork delays. A woman needs to know she can get coverage and needs to fill out the papers early so there is no gap in coverage. Dr. Hollier says the agencies are working on making that process smoother.
So, it turns out that Texas outcomes were not at all as dire as previously stated; that the mistake occurred because Texas was innovating data gathering to address the existing problems; that the legislature responded with a task force which quickly figured out what happened and wants to continue to address the problems; and that one of the main complaints, Medicare coverage, is currently being addressed.
Forgive me if I’m not a pessimist about the direction of maternal care in Texas.
The New, Sticky Narrative
As I argued last year, maternal mortality is rising. It is the “relative to similar countries” part that is misleading. Neither the U.S. nor the Texas rate is the worst in the developed world — probably not even close — and the actual rise is mostly due to advanced maternal age, which we generally insist on denying is a threat.
I don’t hold out hope, however, that Nicholas Kristof of the New York Times will issue any corrections to his emotionally manipulative story about Texas from last summer, “If American’s Love Moms, Why Do We Let Them Die?”, even though the stats do not support his premise.
We love mothers, or at least we say we do, and we claim that motherhood is as American as apple pie.
We’re lying. In fact, we’ve structured health care so that motherhood is far more deadly in the United States than in other advanced countries. An American woman is about five times as likely to die in pregnancy or childbirth as a British woman — partly because Britain makes a determined effort to save mothers’ lives, and we don’t.
Actually, if one of those countries is lying to its women, it is Britain. It appears that our women are five times more likely to die because the U.K. doesn’t accurately record and report their maternal deaths, while we are trying to improve our accounting for those deaths and are making errors of inclusion in our data collection. And while we are figuring our problem out, I’m not sure the British have begun to acknowledge that their problem even exists.
Alas, none of this fits the narrative of a cold and heartless U.S. healthcare system. This truth won’t get a fraction of the coverage that the terrible U.S. and Texas data have.
Unfortunately, some our worst narratives have a mortality rate of zero—these are the ones too ideologically satisfying to ever die.