For most of the past century, the trend in maternal mortality in the US was a good one. Fewer women were dying in connection with their pregnancies. The trend was so good in fact, that many states and cities retired their maternal mortality review committees. They didn’t seem necessary anymore.
Then something changed. Beginning in the mid-2000s, the numbers started going in the wrong direction. More mothers were dying.
Over the past decade, a new infrastructure has emerged to address that critical and tragic trend. At the center of it is the Alliance for Innovation in Maternal Health Care (AIM), which Jeanne Mahoney directs through her work for the American College of Obstetricians and Gynecologists (ACOG). 28 states (including Texas) are now implementing AIM safety bundles, with the remaining 22 states either intending to get involved or exploring involvement.
We spoke to Mahoney, who was an invited speaker and panelist at the 2019 TCHMB Summit, about the origins of AIM, the reasons for the rise in maternal mortality, and the development and dissemination of AIM’s bundles.
Mahoney came to ACOG in 2002 from the Massachusetts Department of Public Health, where she was involved in coordinating risk reduction programs for women of reproductive age.
TCHMB: What’s the story behind AIM? I was looking at the maps you showed at the conference, of the recent and rapid spread of AIM states, and it’s all very recent, isn’t it?
Jeanne Mahoney: Yes and no. In a way it starts in 1992, when a group of people interested in maternal mortality began getting together at the annual ACOG meeting. It was OB/GYNs, public health people, anesthesiologists, all sorts of people who were engaged with women’s health. At the beginning the group wasn’t meeting with a sense of alarm. The numbers had been going down for decades, and the assumption was that they would continue to do so. Then around 2008, we started seeing these numbers rising, and that changed the tenor of the discussion.
Elliott Main had been tracking the numbers in California and seeing the same kind of rise. It became very clear that it was a real problem, not just a statistical fluctuation.
We pulled together a group in 2012 to do a real deep dive into the data and to begin to formulate a national response. What we saw was that there wasn’t just one cause of the rise in maternal mortality. There were many overlapping causes, as well as a great deal that we didn’t yet understand. What we saw clearly, though, were two major cause that were remediable, maternal hemorrhage and hypertension. These were issues about care that we could do something about there and then.
At the national ACOG meeting that year we had a big meeting of about 150-180 people. We came to a consensus that we would begin to develop and deploy maternal safety bundles. We called the group the National Partnership for Maternal Safety. It was not really a card-carrying organization, but it helped formalize the structure. Out of that came The Alliance for Innovation on Maternal Health (AIM), which is staffed by us here at ACOG but is a true alliance. There are 30 different partners across multiple sectors that work with us on every aspect of AIM.
Why did the maternal mortality numbers start going up after so many years of progress? I realize it’s complex, and there are really long answers to that question, but what’s the short version?
For the two issues that we started out with, maternal hemorrhage and hypertension, I can give you some relatively straightforward answers that explain a least a substantial part of the cause.
With maternal hemorrhage we call it “too much too soon too little too late.” We are not waiting for women to have their babies. As a culture we started doing more inductions, more c-sections. We kept pushing the envelope. We have been inducing labor in too many women who don’t have medical reasons for induction, and if the babies fail to arrive on schedule, during an induction, we give the mother more and more oxytocin, and that can be dangerous.
Oxytocin causes the uterus to contract, and the uterus is a muscle like any other muscle. If you force it to contract over and over, it gets tired. After birth it is supposed to contract hard, but too much oxytocin can increase the risk of it failing to contract while the uterine blood vessels continue to pump blood—up to 1/8th of a woman’s total blood volume per minute. The risk increases the longer it takes us to recognize the bleeding.
There is a different hypertension story. Blood pressure has been going up overall in pregnant and postpartum women. We’re not entirely sure why, but that increase has intersected with a slowness, on the part of providers, to implement best practices in how to manage the blood pressure of pregnant and postpartum women.
The walls of the blood vessels of women who are pregnant and early postpartum are very thin, and so the vessels can leak at lower blood pressures than in other people. For a long time that wasn’t the training we were getting. We were taught that the danger zone for pregnant and postpartum women was the same as in other people. We know now that the danger threshold is lower than for the general public. The systolic number, the top number, should never go above 160. Normally we don’t start to treat blood pressures until that systolic number is over 180. By that point, a pregnant or early postpartum woman is much more susceptible to having a stroke and dying.
We have to retrain our providers on every level to be able to understand the hemodynamics of pregnancy and postpartum, so we aren’t sending women home from the hospital early, and we are responsive when they are exhibiting warning signs.
Another big part of the story is women dying from drugs and particularly opioids. Those are mostly killing women post-partum. In Texas right now, for instance, that is accounting for about 50 percent of maternal mortality. One of the ways this is happening is that during pregnancy, many women get some kind of medication assisted treatment for opioids, but then Medicaid runs out 42 days postpartum. Soon after, they go straight to street drugs to treat their disease of addiction, and they overdose and die far more frequently than women using opioids who have not been pregnant.
So how did you get from that big meeting in 2012 to the actual development and deployment of safety bundles? To the official launch of AIM?
We decided our bundles were going to be evidence-based best practices, and that we would pull together a team of experts to develop them. It was multi-disciplinary work-groups of 10-15 people who had to come to consensus. The groups included nurses, doctors, the head of the American Blood Bank Association, someone from the national obstetric anesthesiologists association, and others.
It was hard to come to consensus, having that many different organizational folks sitting around the table. Our hemorrhage bundle, for instance, took us almost two years, because we needed to learn how to work together. At the same time, that process was immensely valuable, both on its own terms and because when we finally did achieve consensus, we already had those organizations on board, affirming this work.
The key with the bundles is that they are very specific in certain ways, and very open in others, so that the hospitals have a lot of flexibility in how they implement them. These are the practices that you need to do, and now you figure out how to do it. If we talk about a medication, for instance, we don’t put dosages in, or instruct providers how to give them. That makes it a lot easier on the implementation end, and so our bundles endure. The hemorrhage bundle, which we developed in 2013-2014, hasn’t changed at all. We review it every 18 months or so, to see if it still applies, and although we have added some references, the itself bundle remains the same.
How did you get the bundles out into the world? It is one thing to have a good set of tools. It’s another thing to get people to use them, and to find the resources to support that.
We began working on the bundles in 2013. In 2014 we applied for and received a grant from the Maternal and Child Health Bureau at HRSA for what we were now calling the Alliance for Innovation on Maternal Health, or AIM. We came out with the hemorrhage and hypertension bundles around then, and then we kept going. We now have 10 bundles, including bundles on venous thromboembolism prevention, postpartum care, reduction of racial disparities, and opioid use disorder.
In terms of getting buy-in more broadly, it helped that from the outset we had support from the broad coalition of organizations involved in the development of the bundles. They were partners in the grant application and gave us credibility as well as avenues for dissemination and outreach. We also worked out a really unique deal with the major disciplinary journals for the simultaneous publication of the commentaries that add implementation support to each bundle. These commentaries flesh out the bundles with definitions, references, discussion of issues that need more discussion, and so on. Whenever we release a bundle the commentary is published in three to five separate peer reviewed journals simultaneously.
Even with all that support, though, at first we were hard pressed to find states to participate. No one understood what we were doing, and we had to sell them on it. Over time, as we’ve established ourselves, and demonstrated the efficacy of the bundles, getting buy-in has become easier.
What is the efficacy?
We are seeing in the hospital outcome data that the states that are doing any bundle at all have reduced their severe maternal morbidity rates 8-25% across the board. A lot of that improvement, we believe, is the result of just pulling the teams together and working on an issue together. It is changing the way people are thinking. They are working together, and that produces good outcomes.
Do the bundles make the work of providers harder? Is it more work?
Yes and no. The bundles themselves are best practices we should be doing anyway. They’re substituting more effective practices for less effective practices. The data part is a little trickier, because in many cases it is adding several process and structure measures that need to be entered into a data system. Doing rapid cycle data driven quality improvement provides the basis for action and is highly valued.
You mentioned the racial disparities bundle. That feels like a very different challenge to tackle than the other bundles. It’s one thing to say that you should always have a hemorrhage cart nearby in situations x and y. It’s very different to tell providers to be less racially disparate, when we don’t even know most of the causes of racial disparities.
It is very tricky, and also very important. We see these incredible disparities in morbidity and mortality. But also variation in the degree of disparity, which suggests that there is a lot of room for improvement. New York City, for instance, has a twelve-fold difference between black and women dying. Illinois has a six-fold difference. Across the country we are seeing a 3 to 4-fold higher rate of black maternal mortality.
This is not just about black women coming in with higher risk factors. One thing we are finding is that there are communication issues. We are not listening to some women as well as we are to others. So somebody says, “I am having pain,” and instead of exploring the possible causes of that pain, the response is, “You’ve just had a baby, of course you have pain.” That happens more to black women than white women. We need to find ways not just to close the gap but to improve communication with all women.
Right now we have a grant from the Robert Wood Johnson Foundation to try to quantify those voices of women, to develop measures that we can track and give back to providers to let them know how they’re doing and how they can improve.
That sounds fascinating, but hard. How would you do that? How do you quantify that kind of subtle communication?
It’s complicated. The grant with RWJ involves studying interactions and identifying key words and phrases that can signal, for instance, that communication is failing. And not just words, nonverbal communication as well. The goal will be to train medical providers to be alert for those key words and gestures that signal that they need to listen more closely. And not just the physicians, the other providers as well, so that there are multiple people listening to and looking out for women.
It is not going to be easy work, but we have hospitals who want to be part of the pilots, which is an indication of the interest in addressing the issue.
One thing I noticed at our conference this year was how many nurses were there. That surprised me, but it probably shouldn’t have. Is that true nationally?
Yes. In every meeting we have for AIM, there are more nurses than docs. Some of that is a result of who the hospitals are choosing as their representatives. They’re more likely to send nurses. But it’s also because nurses really like this work, because it empowers them to provide the best care for their patients. They’re not having to negotiate conflicting instructions from different doctors, and in many cases, it provides them the means to intercede if the care is not being provided according to the bundle.
What’s on the horizon for AIM?
We’re working with more and more states, like Texas, on implementing bundles. We have projects that are focusing on improving care, and customizing bundles, for rural settings. We are rural-izing them, as one of our partners in North Dakota said. We are doing work internationally. Did a project in Malawi, for instance, in which were able to reduce maternal mortality due to hemorrhage in three hospitals by 83%.
We’ll keep working to improve the health of mothers.