Maternal Death Review in a Hill District
A hill district with 210,000 people and about 4,800 births a year reports 12 maternal deaths in the past year, giving an institutional maternal mortality ratio of 250 per 100,000 live births. A confidential review finds that 7 women died at home or on the way to the district hospital, 3 died within 24 hours of arrival, and 2 died postnatally of sepsis. The district hospital, 55 km from the furthest village, has no blood bank, no functioning ambulance at night, and one obstetrician who is often away for court or leave. Only 35% of births are attended by a skilled provider; traditional birth attendants are trusted and charge less. Families report needing permission and cash before seeking care. The district has funding for one major investment this year: a maternity waiting home near the hospital, a blood bank, two ambulances, or training and equipping traditional birth attendants and midwives.
1. Assess how much confidence you can place in the reported maternal mortality ratio of 250 per 100,000 and what you would verify.
Read the worked answer
The arithmetic (12/4,800 = 250 per 100,000) is correct, but reliability is limited. Numerator risks: home deaths may be unregistered or misclassified, with haemorrhage recorded as anaemia or sepsis as fever; deaths of residents occurring outside the district are missed; early-pregnancy deaths may be coded as abortion or omitted; there is no verbal autopsy. Denominator risks: live births may be projections rather than registrations; births to non-residents distort the figure; an institutional ratio is not a population ratio.
With only 12 events, random variation is large: the exact Poisson 95% confidence interval is roughly 129–437 per 100,000, so a single year cannot support fine comparisons or trends. Before acting, verify by reviewing all facility maternal deaths and theatre records, conducting verbal autopsy or a reproductive-age mortality survey for deaths of women aged 15–49, and cross-checking village and traditional birth attendant registers, police and cremation records. Confirm live births from civil registration and disaggregate deaths by cause, place and time from arrival.
If under-reporting is found, the true ratio is likely higher, which strengthens rather than weakens the case for action. The corrected pattern, not the headline ratio, should drive the choice of intervention.
2. Relate the 12 deaths to the three-delays model and recommend which single investment to prioritise.
Read the worked answer
Delay 1, deciding to seek care: permission, cash and trust in traditional birth attendants fit the 7 home and on-the-way deaths. Delay 2, reaching care: 55 km and no night ambulance also fit those 7. Delay 3, receiving quality care: no blood bank, an often-absent obstetrician and sepsis fit the 3 deaths within 24 hours of arrival and the 2 postnatal sepsis deaths.
No single investment addresses all three delays. A maternity waiting home targets delay 2 for women identified as high-risk and can reduce delay 1 by giving families a concrete plan, but it will fail without night transport and a blood bank, and risks becoming a low-quality holding area. A blood bank addresses obstetric haemorrhage and the early facility deaths but does nothing for the 7 who never arrived. Ambulances address delay 2 broadly.
Recommendation: choose the maternity waiting home only if it is paired with a contracted 24-hour ambulance and a written referral protocol; otherwise choose the ambulances. Whichever is selected, implement no-cost fixes now: a roster guaranteeing an obstetrician or trained midwife on site, a blood donor network, haemorrhage and sepsis protocols, partograph training, and community birth-preparedness with transport funds. Justify the choice against the distribution of deaths, not against what is easiest to procure, and track referral time, waiting-home occupancy, blood use and deaths within 24 hours of arrival.
3. The review names a midwife and a doctor, and a newspaper demands the report. How should confidentiality and accountability be handled?
Read the worked answer
Separate system failure from individual misconduct. Most maternal deaths arise from absent blood, transport, staffing or protocols, not from one clinician's negligence. The review should be confidential and legally protected, with a no-blame culture for honest error, but it must not become a cover-up.
Share a de-identified summary and action plan with the public and the newspaper, acknowledging the deaths and stating what will change. Keep names and case details confidential. Inform the families and seek their consent before publishing any details. Give the named staff the right to respond and offer them support. If there is evidence of gross negligence or wilful misconduct, refer it through the professional council and disciplinary process, not through the press. Protect whistle-blowers, and ensure the review has a legal mandate and terms of reference with tracked recommendations.
Explain to the journalist why naming individuals would deter future reporting and damage the review system, while offering the systemic findings. The ethical test is whether the response improves safety and respects the bereaved family, or merely allocates blame. Accountability is demonstrated by visible system change, not by sacrificing two staff members to a headline.