Maternal mortality and near-miss review

• A maternal death is not a single number: it is a woman who died because she was pregnant or had recently been pregnant, from a cause related to that pregnancy or its care. The global goal is fewer than 70 such deaths per 100,000 live births by 2030, yet in 2023 over 700 women still died from preventable causes every day.^[1]
• The same avoidable conditions kill again and again: bleeding (about 27% of maternal deaths worldwide), hypertensive disorders (about 16%), and infections and chronic disease that pregnancy makes worse (about 23%), in the 2020 global analysis.^[9]
• Wherever care is delayed, women die in three places: deciding to seek help, reaching help, and receiving help once in a facility. These three delays frame almost every intervention.^[4]
• A confidential enquiry into every death, run nationally and without blame, is what turned maternal mortality into a measurable, improvable problem in the United Kingdom and beyond.^[5]
• Most women who nearly die do not die. Tracking near misses and severe maternal morbidity exposes the same dangerous processes earlier, with far larger numbers, and lets systems learn before a death occurs.^[6][8]
• Early warning on the ward still rests on simple repeated observations: a rising heart rate, a falling blood pressure, a rising respiratory rate. MEOWS is the obstetric version of this bedside "track and trigger" logic.^[7]

By the end of this page you should be able to:

1. Define maternal death, direct and indirect maternal death, late maternal death and near miss, and state the maternal mortality ratio.
2. Quote the scale of maternal mortality globally and nationally and name its leading causes.
3. Explain the three delays model and map a real case onto its phases.
4. Describe how a confidential enquiry works and what it produces.
5. Apply the WHO maternal near-miss criteria to identify a severe maternal outcome.
6. Explain how a maternal early warning score is calculated and why thresholds differ between hospitals.
7. Outline a two-step screen-and-review process for severe maternal morbidity.
8. Plan a maternal death or near-miss review, including low-resource adaptations.

A maternal death is the death of a woman while pregnant or within 42 days of the end of pregnancy, from any cause related to or aggravated by the pregnancy or its management, excluding accidental or incidental causes.^[3] The 42-day window is the international standard; it is used so that deaths from pregnancy complications are counted even when they occur after birth, miscarriage or termination, but deaths from unrelated causes are not.

The death may be direct or indirect. Direct (obstetric) deaths result from complications of pregnancy itself, from its management, or from a chain of events starting from either of these. Indirect (obstetric) deaths result from a disease that existed before pregnancy or developed during it, and which is not itself an obstetric complication, but was made worse by the physiology of pregnancy. A third group, coincidental deaths, are unrelated to pregnancy; they happen to occur in a pregnant or recently pregnant woman but are not caused by pregnancy.^[3]

Some deaths happen later. A late maternal death occurs more than 42 days but less than one year after the end of pregnancy, from direct or indirect causes. Because of modern intensive care more women survive the initial event only to die weeks later, so late deaths are increasingly important and in some countries now approach a third of all maternal deaths within a year of delivery.^[2]

The maternal mortality ratio (MMR) expresses maternal deaths per 100,000 live births in a given period; the maternal mortality rate uses the number of women of reproductive age as the denominator in some systems. The mortality index (MI) is the number of maternal deaths divided by the number of women with life-threatening conditions, expressed as a percentage, and it tells you bluntly how well a facility rescues women once they become very sick: MI = deaths / (near misses + deaths). The maternal near-miss ratio (MNMR) is the number of near-miss cases per 1,000 live births.

Maternal death is rare in any single hospital, but it is a sentinel event: almost every maternal death is preceded by moments at which someone could have recognised a problem, started treatment earlier or removed a barrier to care. Because the death itself is rare, counting deaths alone is a slow and noisy way to measure quality. That is why maternal mortality review is paired with severe maternal morbidity and near-miss review, which uses much larger numbers and reveals the same system failures while the woman is still alive to benefit from the lesson.

The latest UN inter-agency estimates (WHO, UNICEF, UNFPA, the World Bank Group and the UN Department of Economic and Social Affairs/Population Division) cover 2000 to 2023. In 2023, over 700 women died from preventable causes related to pregnancy and childbirth every day, approximately one death every two minutes.^[1] The Sustainable Development Goal target for 2030 is fewer than 70 maternal deaths per 100,000 live births.^[1] Reaching that target from the present position requires roughly a 15% annual reduction in the global ratio, a pace no region has sustained in recent years.

Nearly all of these deaths are in low- and middle-income countries, and within countries they are concentrated in the poorest women, in rural areas and in humanitarian settings. Conflict or institutional fragility is associated with maternal mortality ratios more than five times those of stable settings in the same estimation series.

A 2025 WHO analysis of causes of maternal death for 2020, the most current published at the time of writing, gives the direct causes for 287,000 maternal deaths worldwide.^[9]

Haemorrhage, hypertensive disorders and sepsis together have accounted for more than half of maternal deaths for decades, and over a quarter of deaths are attributable to indirect causes. Bleeding alone caused about 80,000 deaths and hypertensive disorders about 50,000 in 2020.^[9] The relative importance of indirect causes such as cardiac disease, HIV and tuberculosis rises as direct causes are controlled, which is exactly what has happened in high-income countries.

The fact that the same five clinical problems recur across countries and decades is the strongest argument for review rather than for invention. Postpartum haemorrhage, sepsis, pre-eclampsia and eclampsia, obstructed labour and complications of unsafe abortion together account for at least 60% of all maternal mortality in classic reviews of the causes.^[4] Effective interventions exist for all of them; the failure is almost always in recognition, timing, access or quality. That is why almost the whole of maternal mortality work is system work.

High-income countries show what happens when direct causes are progressively controlled: the MMR falls to low levels, indirect deaths become a larger share, and disparities between women become the visible problem.

MBRRACE-UK (Mothers and Babies: Reducing Risk through Audits and Confidential Enquiries across the UK) reported that in 2022-24, 276 women died during pregnancy or within 42 days of its end; 24 of these deaths were coincidental, leaving 252 deaths from direct and indirect causes among 1,969,321 maternities - a maternal death rate of 12.80 per 100,000 maternities (95% CI 11.27-14.48).^[2] That rate has not improved: it is 20% higher than in 2009-11, the baseline year of the previous ambition to halve maternal mortality in England, and the rate of direct maternal deaths rose 52% over the same period.^[2]

The leading causes tell the story of the modern epidemic. Thrombosis and thromboembolism remained the leading cause of maternal death overall and of direct maternal death, at more than twice the rate of any other direct cause; the next most common direct causes were suicide and sepsis due to pregnancy-related infection.^[2] After cardiac disease, the leading indirect causes were neurological conditions, other indirect causes and psychiatric causes related to substance use.^[2] Maternal suicides remained the leading cause of death between six weeks and one year after pregnancy, with psychiatric causes accounting for 33% of these late maternal deaths.^[2]

These numbers matter for two reasons. First, they show that once overall mortality is low, the challenge shifts from "what kills" to "who dies and why", which is work for audits, equity and social determinants as much as for clinical skill. Second, they show that in a high-income setting the leading cause of direct maternal death is no longer bleeding or infection but thrombosis: prevention means careful risk assessment for venous thromboembolism, not only excellent management of postpartum haemorrhage.

A confidential enquiry into maternal deaths is a specific thing, not a synonym for "audit". The British model was begun in England and Wales in 1952 and extended UK-wide in the 1980s; Ireland joined in 2009. Its stated aim is to investigate why some women die during or shortly after pregnancy and to learn how such tragedies can be avoided, so that all pregnant and recently delivered women receive safe, high-quality care in settings suited to their needs.^[5]

Three features distinguish it from routine review:

• Confidentiality. Cases are reported and examined in confidence; the published reports aggregate cases so that no woman, family or individual clinician is identifiable. That is what makes complete case disclosure possible.
• Expert, independent review. Each case is examined by multidisciplinary assessors (obstetricians, anaesthetists, midwives, pathologists, psychiatric and other specialists) against national standards, with no involvement of the treating team.
• Thematic inference at scale. Reporting is annual rather than triennial, grouped around rolling themes (sepsis, haemorrhage, anaesthesia, amniotic fluid embolism, neurological and indirect causes), so that lessons can be acted on while they are still current.^[5]

The outputs are clinical: recommendations on recognition, escalation, protocol design and service organisation. The UK model has repeatedly been credited with improving standards of care and clinical governance, both in Britain and internationally. MBRRACE-UK reports are commissioned by the Healthcare Quality Improvement Partnership and are led from the University of Oxford.

A confidential enquiry is not the same as the "blame-free" mortality and morbidity (M&M) meeting, which is local and educational, nor the same as a disciplinary investigation. It is also not a substitute for a legal inquest where one is required. In practice all four may run in parallel for the same death.

Every maternal death, in every setting, travels through three stages. The model was described in 1994 by Thaddeus and Maine and has been used ever since to organise why women die.^[4]

1. The first delay is the delay in deciding to seek care, by the woman, her family or both. It is shaped by whether anyone recognises danger, how decisions are made in the family, the cost of care, the perceived quality of services, and sociocultural factors including the status of women.
2. The second delay is delay in reaching a facility: distance, transport, roads, fuel, communication between facilities and the ability to pay.
3. The third delay is delay in receiving adequate care once the facility is reached: shortages of blood, drugs and equipment, staff shortages and skill gaps, poor supervision, absent protocols, power and water failures, and failures of referral.

The three phases rarely operate in isolation. Barriers at the second and third phases feed back into the decision to seek care the next time, so a frightening experience in hospital makes women slower to present next pregnancy.^[4] In one hospital-based case-control study the most striking difference between women who died and controls was the third delay, and in a facility audit in Tigray, Ethiopia, 88% of maternal deaths could be attributed to medical failures.^[4]

The practical value of the model is that it forces a review to ask three separate questions rather than answering only "what was the clinical cause".

A maternal near miss is a woman who nearly died but survived a complication during pregnancy, childbirth or within 42 days of its end.^[8] The WHO near-miss approach (2011) exists so that facilities can identify these cases with explicit criteria rather than with hindsight. It groups women with life-threatening conditions and counts them alongside deaths.

The WHO framework rests on three kinds of criteria: clinical (potentially life-threatening conditions), critical interventions (such as intensive care admission and life-saving procedures) and organ dysfunction (seven dysfunction categories). The clinical layer includes five potentially life-threatening conditions: severe postpartum haemorrhage, severe pre-eclampsia, eclampsia, sepsis and ruptured uterus.^[8]

Working this out into a table:

The organ-dysfunction layer above names the seven categories used in the published WHO near-miss handbook. The source used for this page reports organ dysfunction only as one of the three near-miss layers and does not name each category, so the labels should be checked against the handbook itself before local use.

The WHO Maternal Morbidity Working Group defines maternal morbidity as any chronic or acute health condition due to or aggravated by pregnancy or childbirth that has a negative impact on the woman's well-being.^[8] "Severe maternal morbidity (SMM)" is a subset: conditions that put the woman at clear risk of dying but from which she survives.

Maternal death is the tip. Severe maternal morbidity is vastly more common and is the material a quality system can work with. In one 2022 review of high-income country studies, rates ranged from 5.0 per 1,000 deliveries in England to 16.1 per 1,000 in Canada and 13.5 per 1,000 in California; the largest international comparison reported 15.6 per 1,000 in the USA, 8.2 in Australia and 5.0 in England.^[8] The same review noted that severe haemorrhage, hypertensive disorders of pregnancy and pre-eclampsia or eclampsia were the commonest types across countries.

In the United States the Centers for Disease Control and Prevention report that the prevalence of SMM more than doubled between 1998 and 2014.^[8] Whether that rise is real or an artefact of widening definitions and better coding, it is the reason the American College of Obstetricians and Gynecologists now recommends a systematic approach to SMM rather than a reliance on the rarity of death.^[6]

The ACOG/SMFM Obstetric Care Consensus on severe maternal morbidity (number 5, September 2016, reaffirmed 2025) starts from the observation that SMM has a high rate of preventability, similar to that of maternal mortality, and can be considered a near miss for maternal mortality because without identification and treatment some of these women would have died.^[6] It therefore argues that severe morbidity cases merit quality review on the same basis as deaths.

The screening-and-review process it proposes has two steps:

1. Screen. Identify women with severe morbidity using defined criteria - either an existing published screening list (disease or procedure based, or ICD-based) or a locally agreed list of conditions and interventions that always trigger review. This is the "case finding" step; it replaces informal memory.
2. Review. Each identified case is reviewed by a multidisciplinary team, anonymously, to determine whether there were opportunities for improvement in care, what the modifiable factors were, and what action follows. The team decides whether care was appropriate and, if not, which phase of the three delays it fell into.

In the absence of universal agreement on which conditions count as SMM, institutions and systems are advised to adopt an existing screening criterion rather than inventing one.^[6] That advice matters: the most common error in SMM work is to collect a local list that no one else recognises, so that nothing can be compared between hospitals or over time.

The commonest technical response to the third delay is an early warning system: a simple, repeated observation set at the bedside that is converted into a score, with defined triggers for escalation. NICE recommended in 2007 that physiological track-and-trigger scoring systems be used for monitoring all patients in acute care settings, including women in obstetrics, and listed six parameters: respiratory rate, oxygen saturation, heart rate, systolic blood pressure, temperature and consciousness level.^[7]

The obstetric version, the Modified Early Obstetric Warning Score (MEOWS), is based on the Stirling Royal Infirmary chart. It records six vital signs - respiratory rate, heart rate, systolic and diastolic blood pressure, oxygen saturation and temperature - together with secondary observations such as proteinuria, lochia, pain scores and whether the woman looks subjectively unwell. Values fall into white (within an arbitrarily agreed normal range), yellow (mildly abnormal) or red (abnormal) squares. A red observation, or two yellows, suggests deterioration and should trigger escalation.^[7]

Whether the event is a death or a near miss, the review should answer the same questions in the same order. A practical flow:

For a near miss this process is faster and less formal, but the same questions apply, and the aggregate numbers are what justify change. A hospital that reviews only deaths is reviewing failures it can no longer fix; one that reviews near misses is reviewing failures it can still prevent.

Where the health system is thin, the same facts take on a different shape. The three delays model was literally built in this context, and the first two delays are usually the binding constraint. The WHO near-miss approach, the MDSR (maternal death surveillance and response) framework and district-level death reviews are designed to be used with paper records and small committees.

Practical priorities in these settings:

• Know the five conditions and their treatments. Postpartum haemorrhage (uterotonics, tranexamic acid, fluid resuscitation, bimanual compression, balloon or tamponade, surgery), pre-eclampsia and eclampsia (magnesium sulphate, blood pressure control, delivery), sepsis (antibiotics, fluids, source control), obstructed labour (partograph, timely delivery) and unsafe abortion complications (uterotonics, antibiotic, uterine evacuation, resuscitation). Five conditions and the skills to treat them would remove most maternal mortality even where everything else is scarce.^[4][9]
• Track the delay, not only the disease. A district review that records which phase of delay killed the woman produces actionable recommendations: a drug stockout is a third delay, an unavailable ambulance is a second delay and a late presentation is a first delay.
• Use the simplest effective warning. A single parameter that is extreme, or any two abnormal observations, is often enough to trigger senior review where full MEOWS scoring is unrealistic; the principle matters more than the paper.
• Blood and drugs. Availability of safe blood, magnesium sulphate, oxytocin, antihypertensives and antibiotics saves more lives than additional clinical complexity. Where a district hospital cannot transfuse, that is the intervention to fix first.
• Consume evidence, do not write protocols from scratch. National guidelines (Uganda Clinical Guidelines and equivalents) and WHO's guidelines on postpartum haemorrhage and eclampsia give tested drug, dose and route information.

• Women after pregnancy. Late maternal deaths, including suicides, are a substantial minority of deaths and are under-reported. Postnatal follow-up, mental health support and access to contraception are preventive interventions, not optional extras. The UK data place suicides as the leading cause of late maternal death, with psychiatric causes 33% of those deaths.^[2]
• Women with severe morbidity. Those who survive a near miss often have lasting disability: secondary infertility, organ damage, psychological injury. "Maternal survival" measurements that count only deaths miss this entirely, which is another reason near-miss tracking is needed.
• Age and parity. Very young and older mothers carry higher risk; in the UK women over 40 had nearly three times the risk of maternal death of women aged 25-29.^[2] These are associations, not verdicts about individuals, and they work through the prevalence of comorbidity and social disadvantage.
• Ethnicity and deprivation. In the UK Black women had nearly three times the risk of White women and the most deprived areas twice the rate of the least deprived.^[2] Similar gradients are reported in every well-measured system. They are multi-factorial (comorbidity, access, service quality, racism) and do not reduce to personal behaviour. They are exactly what a review system exists to expose rather than ignore.

These are adjacent topics on MedGrasp; no links are available yet from this page: antenatal care and preconception care; antepartum haemorrhage; caesarean section and episiotomy; postpartum haemorrhage (within antepartum haemorrhage); pre-eclampsia and hypertensive disorders of pregnancy; sepsis in pregnancy; venous thromboembolism.

• Maternal death is death during pregnancy or within 42 days afterwards, from any cause related to or aggravated by pregnancy or its management, defined by WHO and ICD-MM and coded direct, indirect, coincidental or late.
• The maternal mortality ratio is maternal deaths per 100,000 live births; the global aspiration by 2030 is under 70, against roughly 700 preventable deaths a day in 2023.
• WHO's 2020 causal analysis attributes about 27% of maternal deaths to haemorrhage, about 16% to hypertensive disorders, and about 23% to underlying conditions such as HIV, malaria, anaemia and diabetes, with other direct causes making up the rest.
• Where systems are strong, indirect causes such as cardiac and psychiatric disease now dominate; thrombosis is the leading direct cause of maternal death in the UK.
• Five conditions - postpartum haemorrhage, sepsis, pre-eclampsia and eclampsia, obstructed labour and complications of unsafe abortion - account for at least 60% of all maternal mortality. ^[4]
• Maternal death review is a confidential, expert, system-level process, not a blame exercise and not a local M&M meeting. It began in Britain in 1952 and is the model behind MBRRACE-UK and many national systems.
• The three delays - deciding to seek care, reaching a facility, and receiving adequate care once there - organise the causes of almost every maternal death anywhere.
• A maternal near miss is a woman who nearly died but survived; tracking these cases gives systems far larger numbers than deaths do, with the same lessons.
• The WHO near-miss criteria combine potentially life-threatening conditions (severe postpartum haemorrhage, severe pre-eclampsia, eclampsia, sepsis, ruptured uterus), critical interventions (ICU, life-saving procedures) and seven categories of organ dysfunction.
• Severe maternal morbidity has a high, mortality-like rate of preventability; the ACOG approach is a two-step screen and review (identify then judge quality), because no single agreed list of conditions exists.
• Early warning in obstetrics (MEOWS) is bedside track and trigger on six vital signs plus observations; thresholds vary by hospital and the chart alone does not save anyone - training and escalation do.
• Review is only useful if the loop is closed: recommendations tracked, interventions delivered and re-audited.
• Ethnic and socioeconomic gradients persist in every well-measured system and are precisely what review systems exist to expose and reduce.