It is 5:30 in the morning. My house is quiet. I am the only one awake.
I stretch. I breathe. I center myself.
Maternal mortality review days are, in a word, difficult.
For years, I sat on a committee tasked with understanding why a mother died. It was an honor. I never took it lightly.
It was service to a community I love deeply, and it rested on a conviction I have not been able to put down. That we can do better. That part of doing better is understanding, precisely and without flinching, what went wrong.
Our facilitator would open each meeting with a breathing exercise. It was never quite enough for me. I carry these women. I am these women.
I sit down at my desk and click the link for the zoom meeting. The faces arrive one square at a time. Each square is a form of expertise. Maternal-fetal medicine. Nursing. Pathology. Social work. Public health. Obstetrics. A doula. A community member.
Then the case.
She is thirty-four. Or nineteen. She died at thirty-two weeks, or eleven days after delivery, or four months later.
The maternal-fetal medicine physician walks us through the clinical picture. What was known. What was missed. The hours in which the trajectory might still have bent. The social worker lays out the conditions of her life: food insecurity, hourly shift work, prenatal appointments that cost a shift she could not afford to lose. The doula tells us what she said, and how it was heard.
Every form of expertise is present in that room except hers. She is the only person who cannot be there.
Then the form.
There are six decisions a committee must make about each death. The third asks whether the death was preventable.
Item three.
There is almost always someone who declares the mother’s death could not have been prevented. I want to be fair to that person. These are not careless people. They have given their working lives to keeping women alive, and they are sitting with a file that describes a failure they may have felt personally.
And still, the committee checks the box: Preventable.
Not occasionally. Nearly always. In the most recent federal data, 85.7 percent of pregnancy-related deaths reviewed by these committees were determined to be preventable. Room after room. File after file. The same answer.
Give the committee a death that appears to have nothing to do with medicine at all — a car accident — and it will still find its way to yes. We pulled the road. The signage. The speed limit. Whether anyone had ever asked the state to lower it. Whether a woman four months postpartum was driving to a job she should have been able to take leave from. The answer came back the same.
Here is what I have come to understand about that word.
The federal definition holds that a death is preventable if there was at least some chance it could have been averted by one or more reasonable changes to patient, community, provider, facility, or systems factors.
Reasonable.
Preventability is not a measurement. It is a verdict — a judgment about what we consider reasonable to expect of ourselves. Which means 85.7 percent is not the outer limit of what was possible. It is the limit of what we were willing to call reasonable.
Something even better was possible. Something even better is possible.
So the problem was never that we do not know how to keep women alive. We know. We write it down. We have been writing it down for years, in a federal data system built for exactly that purpose.
The problem is where we have agreed to stop.
And we should be honest about how long we have been stopping there.
Enslaved people were documented with extraordinary precision. Plantation record books tracked births and deaths across decades. One Louisiana ledger carries the title “List of Mothers, Births, and Deceased.” Another is a memorandum of the births of Negro children, kept without interruption from 1817 to 1852. There were bills of sale, probate inventories, insurance policies underwriting the lives of the people being counted. Federal census takers enumerated enslaved people in 1850 and 1860 without recording their names — age, sex, color.
The record was meticulous because the record was financial. A woman who died in childbirth was a loss of capital, and capital gets counted.
The population schedules did not name them. The mortality schedules often did. The federal government was more willing to record an enslaved woman’s name in the year she died than in any year she lived.
Care operated on the same logic. Enslaved midwives, the women known as grannies, attended most births. White physicians were summoned mainly for complications and for investigations into why an infant had not survived. Some planters reduced the field labor of pregnant women, having calculated that fewer deaths produced greater profit. Where care improved, it improved as asset protection. And when outcomes were bad, the fault was assigned to the mothers and the midwives.
We have been counting Black women’s reproductive outcomes in this country for four hundred years. We have never once built a record of whether anyone cared for them.
The ledger is gone. The arrangement that produced it — care as margin, a woman’s access to it contingent on her labor — is not. And a system organized that way will always find it easier to count her than to tend her.
A field is defined by what it measures itself against, and ours made that choice a long time ago. Black women in the United States are three to four times more likely to die from pregnancy-related causes than white women. That number is true. It has done real work. I have used it myself, in more grant proposals than I can count.
But somewhere along the way, the comparison stopped being a description and became the ambition. Close the gap. Eliminate the disparity. Bring our numbers in line with theirs.
Consider what that actually asks for. Parity with white women means matching outcomes that are themselves among the worst in the wealthy world. It is a subtraction problem, and subtraction problems can be solved from either direction — the gap narrows if white women start dying more often, and no one would call that justice.
This is not an argument against measuring disparity. It is an argument against mistaking the measurement for the goal. When we name a comparison group, we are naming a horizon. We chose one that was never adequate for anyone.
The obvious correction is to look abroad. Norway, Italy, and their peers report two to three maternal deaths per 100,000 births, sometimes fewer. What they have is not mysterious: skilled attendance at nearly every birth, continuous preventive monitoring, comprehensive postpartum support, and coverage that does not depend on employment.
Someone always objects that those are different countries — different systems, different politics, different histories. Yes. That is the argument. You cannot purchase Norway’s outcomes while preserving the arrangement described above.
But notice something else about two per 100,000. It is a better number. It is still a death rate.
We have changed the number. We have not changed what we are counting.
Even the strongest version of our current ambition — zero preventable deaths, zero preventable morbidity — is defined entirely by the absence of catastrophe. It tells us what must not happen. It says nothing about what should. Survival as the ceiling.
So let me say this plainly. Even zero preventable death is not the finish line. It is the starting line — the condition that has to be true before the real work can begin.
So what would a finish line measure?
A field measures what it has agreed to be held accountable for. Maternal health measures death because death is the only thing it has consented to answer for. Everything else — whether she was believed, whether she was included in the decisions made about her body, whether she was treated as a person — is treated as atmosphere. Nice when present. Not an outcome.
I would start here.
The interval between the moment she said something was wrong and the moment someone acted. It is in every chart already. We have simply never called it an outcome. Every Black woman I know can tell you the length of her own.
Whether decisions were made with her or to her. Consent as an event that occurred, not a signature that was collected.
Whether she left more willing to return than when she arrived. Trust is not a feeling patients owe us. It is a result we either produce or fail to produce, and it is the difference between a woman who calls at the first symptom and one who waits.
Whether her pregnancy was permitted to be unremarkable. The right to an ordinary birth — ordinary joy, ordinary fear, ordinary imperfection — as a clinical standard rather than a private stroke of luck.
Patient-reported outcome measures are established methodology. Oncology moved past mortality decades ago to survivorship and quality of life. Palliative care built rigorous instruments for dignity. The tools exist. Maternal health has simply never pointed them here.
This is not a departure from science. It is a maturation that other fields underwent while ours was still counting.
We have codes for nearly every way a body can fail. We take vital signs for heart rate, blood pressure, oxygen, temperature — the readings that tell a clinician whether a body is in distress. There is no field on that form for whether she was believed. There is no field for whether anyone in the room cared for her.
What if there were?
That is the question I want us to sit with. We have spent a generation getting better at counting the dead. The finish line is not counting fewer dead mothers, not counting any dead mothers. The finish line is past that, in territory past survival we have not agreed to measure.
Until we agree, we will keep arriving at 5:30 in the morning, checking a box that says the death was preventable, and calling that knowledge.
Something better was possible. Something better is possible.
It is time to say what it would look like.
