Are Britain's Maternity Services Safe?

Britain's maternal mortality rate stands at 9.7 per 100,000 maternities as of 2022, up from 8.2 in 2014 and roughly double Norway's rate of 4.2. The figure places the UK above the EU average for a metric that should, in a well-resourced health system, be falling. Sixty per cent of NHS maternity units are rated 'requires improvement' or 'inadequate' by the Care Quality Commission. The disparities within those numbers are stark: MBRRACE-UK data show Black women are 3.7 times more likely to die in childbirth than white women, and Asian women 1.8 times more likely — a gap that has persisted for two decades with minimal improvement despite repeated national commitments to close it.

Two landmark investigations exposed the scale of institutional failure. The Ockenden Report (March 2022) examined Shrewsbury and Telford NHS Trust and found 1,223 cases of harm over 20 years, including 201 baby deaths and 9 maternal deaths deemed potentially avoidable. Poor staffing, normalisation of bad outcomes and weak governance ran through every finding. A year later the East Kent maternity inquiry (March 2023) identified 45 babies who died unnecessarily, with a further 40 deaths that may have been avoidable — the same pattern of understaffing, poor CTG interpretation and a culture of blame rather than learning. These were not isolated trusts; they were systems operating under the same national pressures.

The staffing crisis underpins much of the risk. England carries 2,500 midwife vacancies, a rate of 12.3%, and the Royal College of Midwives estimates 2,500 more midwives are needed simply to deliver NICE-standard care. A 2023 RCM survey found one in three midwives considering leaving the profession. NICE recommends every woman have a named midwife throughout pregnancy — a model shown to reduce preterm birth by 24% and stillbirth by 16% — yet only 14% of women in England receive continuity of carer. The Maternity Safety Strategy (2016) and the Three-Year Delivery Plan (2023) committed to halving stillbirths and neonatal deaths by 2025. Neonatal mortality has fallen 25% since 2010, but maternal mortality remains stubbornly, unacceptably high.

The racial disparity in maternal mortality reflects compounding structural factors. Research consistently shows Black women's pain is underestimated by clinicians and their concerns are less likely to trigger escalation — a pattern of implicit bias documented across multiple NHS trusts. Higher rates of pre-eclampsia, gestational diabetes, and cardiovascular conditions in some ethnic groups are partly driven by socioeconomic deprivation and its health consequences, while language barriers reduce access to antenatal education and informed consent. Continuity of carer — seeing the same midwife throughout pregnancy, birth, and postnatal care — improves outcomes by 24% for preterm birth, but delivering it requires restructuring shift-based rotas around caseload teams, which trusts with 12% vacancy rates cannot staff. The midwifery training pipeline takes three years and attrition during training runs at 25%. The Three-Year Delivery Plan committed £165 million but focused primarily on workforce retention rather than expansion.

Several measurement limitations affect how maternity data should be interpreted. MBRRACE-UK reports maternal mortality as a three-year rolling average specifically because annual numbers are too small for statistical reliability at national level — this means genuine changes in trend take years to appear in the data. The 9.7 per 100,000 rate uses “maternities” as its denominator, which differs from the “live births” or “pregnancies” denominators used elsewhere, complicating international comparison. CQC maternity ratings reflect a point-in-time inspection and services can change substantially between visits. Midwife vacancy data counts unfilled posts against an establishment figure that trusts themselves set — and which can be understated to manage headcount budgets, masking the true shortfall. Ethnic disparity data relies on self-reported ethnicity in NHS records, where coding inconsistencies and missing data are well documented.

The racial disparity in maternal mortality reflects compounding structural factors. Research consistently shows Black women's pain is underestimated by clinicians and their concerns are less likely to trigger escalation — a pattern of implicit bias documented across multiple NHS trusts. Higher rates of pre-eclampsia, gestational diabetes, and cardiovascular conditions in some ethnic groups are partly driven by socioeconomic deprivation and its health consequences, while language barriers reduce access to antenatal education and informed consent. Continuity of carer — seeing the same midwife throughout pregnancy, birth, and postnatal care — improves outcomes by 24% for preterm birth, but delivering it requires restructuring shift-based rotas around caseload teams, which trusts with 12% vacancy rates cannot staff. The midwifery training pipeline takes three years and attrition during training runs at 25%. The Three-Year Delivery Plan committed £165 million but focused primarily on workforce retention rather than expansion.

Several measurement limitations affect how maternity data should be interpreted. MBRRACE-UK reports maternal mortality as a three-year rolling average specifically because annual numbers are too small for statistical reliability at national level — this means genuine changes in trend take years to appear in the data. The 9.7 per 100,000 rate uses “maternities” as its denominator, which differs from the “live births” or “pregnancies” denominators used elsewhere, complicating international comparison. CQC maternity ratings reflect a point-in-time inspection and services can change substantially between visits. Midwife vacancy data counts unfilled posts against an establishment figure that trusts themselves set — and which can be understated to manage headcount budgets, masking the true shortfall. Ethnic disparity data relies on self-reported ethnicity in NHS records, where coding inconsistencies and missing data are well documented.