Severe maternal morbidity statistics: the near-misses that outnumber maternal death roughly 80 to 1

The hemorrhage started quietly. Her uterus had contracted after the cesarean, the recovery room looked ordinary, and then the pad count changed and the blood pressure began to fall. Two units of packed red cells, a second operating room, a hysterectomy she had not planned for. She left the hospital alive, with a living baby, and with a scar and a diagnosis that almost never makes the evening news: severe maternal morbidity.

I see versions of that story more often than the maternal-mortality numbers would suggest. Death is the tip. The near-misses form a much larger ice sheet underneath. The CDC defines severe maternal morbidity (SMM) as unexpected outcomes of labor and delivery that result in significant short- or long-term consequences to a woman's health. It is tracked with 21 diagnosis and procedure indicators on the delivery hospitalization record, and it is the measure hospitals and perinatal collaboratives actually use to improve care in real time.

The public underweights the scale of this problem. According to a national analysis of 2016-2017 delivery hospitalizations by Brown and colleagues in Obstetrics & Gynecology, the rate of SMM was 139.7 per 10,000 deliveries, about 1.4% of births. The Commonwealth Fund, applying that order of magnitude to recent U.S. birth totals, puts the annual count at roughly 50,000 to 60,000 women. For every maternal death, that same primer estimates there are about 70 to 80 cases of severe maternal morbidity identified at delivery alone.

What follows is the rate, the indicators that dominate it, the racial and payer gaps, and the hospital-to-hospital spread. Quality is not evenly distributed.

Year	Rate per 10,000
2012	69.5
2013	71.9
2014	72.4
2015	73.1
2016	72.0
2017	71.5
2018	76.7
2019	79.7

Source: Hirai et al., JAMA Network Open, 2022; HCUP National Inpatient Sample. Rates exclude blood-transfusion-only cases.

Race/ethnicity	Rate per 10,000 deliveries
Black	240.7
Hispanic	161.3
Asian/Pacific Islander	138.7
White	113.6

Source: Fingar et al., AHRQ HCUP Statistical Brief No. 243, 2018. Any of 21 CDC SMM indicators during the delivery hospitalization.

What counts as severe maternal morbidity

Start with the definition. Two different rates float around the literature, and they answer different questions.

The CDC's population measure uses 21 indicators drawn from hospital discharge codes: 16 diagnoses (eclampsia, acute renal failure, sepsis, amniotic fluid embolism, and others) and five procedures (blood transfusion, hysterectomy, ventilation, temporary tracheostomy, and conversion of cardiac rhythm). The current ICD-10 set applies from October 2015 forward. CDC materials note that SMM has been steadily increasing. Consequences include longer stays and higher costs as well as immediate clinical harm.

A second, stricter measure excludes deliveries in which blood transfusion is the only SMM indicator. Transfusion coding is imperfect, and validation work has found that transfusion alone often does not mark true life-threatening morbidity. Quality-improvement collaboratives and Healthy People 2030 therefore track the excluding-transfusion rate. When a journalist quotes "the SMM rate," ask which of the two is being used. The including-transfusion figure is roughly twice as high. ACOG and the CDC also endorse a facility-level screen for case review: ICU admission and/or four or more units of blood products, from conception through 42 days postpartum. That screen is for real-time multidisciplinary review, not for the administrative discharge rate.

How common is SMM

The cleanest single number for the full 21-indicator definition is still the one Brown and colleagues published using the HCUP National Inpatient Sample for 2016-2017: 139.7 SMM events per 10,000 delivery hospitalizations, about one in every 72 deliveries. The Commonwealth Fund's 2021 primer cites approximately 140 per 10,000 for the same period and puts the annual count near 50,000 to 60,000 women (about 50,500 if that rate is applied to 3.6 million U.S. births in 2020).

Longer-run AHRQ HCUP data show the same direction. From 2006 through the third quarter of 2015, any SMM rose 45%, from 101.3 to 146.6 per 10,000 delivery hospitalizations. Blood transfusion drove most of that climb (54% increase, 78.9 to 121.1 per 10,000); the other 20 indicators rose 24% (33.8 to 41.9).

Strip out transfusion-only cases and the climb continues from a lower base. Hirai and colleagues, in JAMA Network Open (2022), estimated the national excluding-transfusion rate rose from 69.5 per 10,000 in 2012 to 79.7 in 2019, without a significant discontinuity at the ICD-10 code switch. Acute kidney failure more than doubled, from 6.4 to 15.3 per 10,000. Healthy People 2030 objective MICH-05 tracks this measure and is not on track: baseline 71.5 per 10,000 in 2017, most recent 93.1 in 2022, target 64.4. Official status: getting worse.

Indicator	Rate per 10,000
Blood transfusion	121.1
Disseminated intravascular coagulation	11.0
Hysterectomy	11.0
Acute renal failure	6.5
Sepsis	5.2
Shock	4.3

Source: Fingar et al., AHRQ HCUP Statistical Brief No. 243, 2018. First three quarters of 2015.

The most common indicators

If you open a discharge abstract flagged for SMM, blood transfusion is what you will see most often. In 2015, transfusion was performed in about 83% of SMM cases and occurred at 121.1 per 10,000 deliveries, according to the AHRQ brief. Disseminated intravascular coagulation and hysterectomy were next, each at 11.0 per 10,000. Acute renal failure and sepsis followed, and rates for those indicators (along with shock) more than doubled between 2006 and 2015.

Procedures cluster. Over half of deliveries with shock, amniotic fluid embolism, sickle cell crisis, or DIC also received a transfusion in 2015; one-third of deliveries with shock also involved a hysterectomy. A National Health Statistics Reports analysis of case composition found transfusion in 59.9% of SMM cases, DIC 11.0%, acute renal failure 9.3%, hysterectomy 9.0%, and eclampsia 6.6%. Eclampsia sits at the sharp end of the hypertensive disorders covered in our review of preeclampsia and pregnancy complications. Delivery hysterectomy is permanent fertility loss layered on a crisis; broader patterns are in our page on hysterectomy rates. Cesarean delivery is not itself an SMM indicator, but it amplifies hemorrhage and infection risk (see our cesarean section rates statistics).

SMM versus maternal death

Maternal death is rare enough that a hospital can go years without one. SMM is not. The Commonwealth Fund's primer puts annual U.S. maternal deaths around 650 to 750 in the years it reviewed, against 50,000 to 60,000 SMM cases, and estimates roughly 70 to 80 cases of severe illness at delivery for every maternal death. That ratio is why SMM is the better day-to-day quality signal.

The two curves do not always move together. Between 2006 and 2015, AHRQ found SMM rising across every racial and ethnic group while in-hospital mortality during the delivery stay fell. Among Black women, SMM climbed from 164.0 to 240.7 per 10,000 deliveries even as in-hospital death fell from 19.0 to 10.9 per 100,000; among white women, SMM rose from 76.3 to 113.6 while death fell from 5.5 to 3.7 per 100,000. Survival improved; the volume of critical illness did not. Our companion analysis of preventable maternal deaths covers mortality; SMM is the larger numerator for hospital improvement work.

Racial disparities that refuse to close

The Black-white gap in SMM is not subtle, and it has not closed.

In 2015, AHRQ reported SMM at 240.7 per 10,000 delivery hospitalizations among Black women, 161.3 among Hispanic women, 138.7 among Asian/Pacific Islander women, and 113.6 among white women. The Black rate was 112% higher than the white rate, nearly identical to the 115% gap in 2006 (164.0 vs. 76.3). A decade of rising morbidity left the relative Black-white disparity essentially unchanged. Hispanic and Asian/Pacific Islander gaps relative to white women narrowed somewhat; the Black gap did not.

Black women were overrepresented among SMM deliveries (24.2% of SMM cases vs. 14.6% of non-SMM deliveries in 2015). That is not explained by age: Black mothers delivering that year were younger on average than white mothers, and the mid-20s is the lower-risk band for SMM. Comorbidity load magnifies risk without erasing race. Brown and colleagues found SMM rates of 48.5 per 10,000 with no Elixhauser comorbidities, 238.6 with one, 379.9 with two, and 560 with three or more. Black women were less likely to have zero comorbidities, but race and chronic disease are not the same variable, and the hospital of delivery is not a neutral backdrop. The same architecture appears in our review of Black maternal health disparities.

Insurance and income

Payer status tracks risk almost as tightly as race does. The two are entangled.

In the AHRQ 2015 data, SMM rates were 175.0 per 10,000 when Medicaid was the expected payer and 176.5 when the delivery was uninsured or self-pay, compared with 120.8 under private insurance. Deliveries with SMM were more often Medicaid-paid (51.4% vs. 42.9% without SMM). Community income moved the same way: 177.7 per 10,000 in the lowest ZIP-code income quartile versus 122.2 in the highest.

Medicaid does not cause morbidity. It covers a larger share of births to people with fewer resources, more chronic disease, and less continuous preconception care. It also holds the policy levers for postpartum coverage, doula payment, and quality-collaborative requirements. Regional rates in 2015 ran 165.4 per 10,000 in the Northeast and 163.6 in the South, against 132.4 in the West and 116.3 in the Midwest. Large metro areas (154.0) and teaching, public, safety-net, and minority-serving hospitals also sat higher. Case mix explains some of that; it does not automatically explain all of it.

Hospital-level variation

Two women with similar charts can face very different odds depending on where they deliver. That is a measured finding, not rhetoric.

Howell and colleagues studied 353,773 New York City deliveries from 2011 to 2013. SMM occurred in 2.5% of deliveries overall: 4.2% among Black women and 1.5% among white women. After adjustment, Black women still had twice the odds (odds ratio 2.02). Risk-standardized rates across 40 hospitals ranged from 0.8 to 5.7 per 100 deliveries. White women were concentrated in lower-morbidity hospitals (65% of white deliveries in the lowest tertile, versus 23% of Black deliveries). Differences in delivery hospital may account for as much as 47.7% of the city's Black-white SMM disparity.

The gap is about more than who walks through the door. It is also about which door she walks through, and what that unit does when bleeding does not stop. Perinatal quality collaboratives, hemorrhage and hypertension bundles, and mandatory SMM case review are the operational response. Hirai's 2012-2019 analysis found significant excluding-transfusion SMM increases in 21 states and a decrease in only one. State lines change Medicaid policy, hospital consolidation, and workforce density, the same structural facts behind maternity care deserts and provider shortages.

Why the rate keeps rising

No single factor explains the trend line. The CDC lists rising maternal age, pre-pregnancy obesity, pre-existing chronic conditions, and cesarean delivery among population shifts that can push SMM upward. Brown's comorbidity dose-response is consistent with that story: three or more chronic conditions carried more than an elevenfold crude increase compared with none. Coding also moves numbers: transfusion documentation, the ICD-10 transition, better recognition of kidney injury and sepsis. Some of the rise is better capture. Not all of it is.

The data leave no room for complacency. Healthy People 2030's excluding-transfusion rate is moving away from its target. The Black-white SMM ratio has stayed near two-to-one for a decade. Hospital performance inside one city can differ sevenfold. Those numbers are not measurement quirks. They map where prevention is failing.

For clinicians: run the bundles, review every near-miss, staff postpartum for the patient who looks fine at hour six and is not at hour fourteen. For families: ask before labor how the hospital handles hemorrhage and hypertension, and who covers nights. Anyone planning another pregnancy after a complicated delivery can use tools such as our ovulation calculator and period calculator for timing, alongside a hard look at residual risk with an obstetrician who has the discharge summary in hand. SMM is common enough that every labor unit will see it, and common enough that systems tracking only deaths will miss most of the story.

Frequently asked questions about severe maternal morbidity

What is severe maternal morbidity?

Severe maternal morbidity is the CDC's term for unexpected outcomes of labor and delivery that cause significant short- or long-term health consequences. It is identified in administrative data with 21 diagnosis and procedure indicators, including blood transfusion, hysterectomy, eclampsia, acute renal failure, and sepsis. Ventilation is also on the indicator list. Hospitals also use ICU admission and large-volume transfusion as real-time review triggers.

How common is severe maternal morbidity in the United States?

Brown and colleagues found 139.7 SMM cases per 10,000 delivery hospitalizations in 2016-2017 using the full CDC indicator list. That is roughly 1.4% of deliveries, or about 50,000 to 60,000 women per year according to Commonwealth Fund estimates. The rate excluding transfusion-only cases was 79.7 per 10,000 in 2019 and 93.1 per 10,000 in 2022.

What are the most common SMM indicators?

Blood transfusion is by far the most common. AHRQ data for 2015 put transfusion at 121.1 per 10,000 deliveries, with disseminated intravascular coagulation and hysterectomy each at 11.0 per 10,000. Among patients who already have SMM, NCHS has reported transfusion in 59.9% of cases, followed by DIC and acute renal failure. Hysterectomy and eclampsia appear next in that case series.

How does SMM compare with maternal death?

SMM is far more common. The Commonwealth Fund estimates about 70 to 80 cases of severe maternal morbidity at delivery for every maternal death, with annual SMM counts of roughly 50,000 to 60,000 against several hundred deaths. SMM is therefore a more sensitive quality measure for hospital improvement than mortality alone.

Are there racial disparities in severe maternal morbidity?

Yes. In 2015 AHRQ data, Black women had an SMM rate of 240.7 per 10,000 deliveries compared with 113.6 for white women, more than double. That relative gap was essentially unchanged from 2006. Hispanic and Asian/Pacific Islander women also had higher rates than white women, though those gaps narrowed somewhat over the decade.

Does the hospital where you deliver matter?

Yes. In New York City, risk-standardized SMM rates across hospitals ranged from 0.8 to 5.7 per 100 deliveries. Black women were less likely than white women to deliver in low-morbidity hospitals, and researchers estimated that delivery site could explain as much as 47.7% of the city's Black-white SMM disparity. Hospital quality is a lever. It is not background noise.