Towards improving maternal and perinatal death surveillance and response in the African region: an analysis of 25 countries from 2015 until 2022

Abstract

Africa renewed its efforts to document maternal and perinatal deaths in 2014 following the release of the WHO’s maternal and perinatal death surveillance and response (MPDSR) guidelines. Successful implementation of MPDSR requires timely notification and a thorough documentation of maternal and perinatal deaths, the development of causal pathways, and the enactment of targeted improvement (change) actions to prevent future avoidable deaths. Similar to the Plan-Do-Study-Act (PDSA) iterative process used in the Model for Improvement (MFI), MPDSR rests on robust data reporting systems without which quality improvement initiatives are ineffective. Unfortunately, many African health systems have significant challenges with data collection and reporting, often compounded by a disconnect between public and private sectors, which erode efforts to improve MPDSR. Over the past decades, countries across Africa have produced MPDSR reports that, despite often appearing comprehensive, mask underlying operational deficiencies. These reports consistently highlight substantial barriers to implementing effective death reviews. Findings show that of 47 countries, 25 (53%) provided MPDSR reports, with East and Southern Africa contributing more than half. Notably, under 30% and 12% of maternal and neonatal deaths in the District Health Information Software (DHIS) were notified to MPDSR, and about 63% of maternal deaths were reviewed. Our analysis of MPDSR reports from 25 African countries, covering 2015 to 2022, supplemented by data from DHIS, reveals critical issues: a widespread shortage of skilled maternity healthcare workers trained in MPDSR, inadequate data harmonisation and lack of standardised maternal and newborn health metrics, insufficient funding, the absence of functioning MPDSR committees at facility level and weak leadership committed to maternal and newborn health goals. Addressing these bottlenecks is essential for strengthening MPDSR efforts and should guide WHO and other development partners to scale up maternal and perinatal death surveillance across Africa.

Summary box.

• Africa continues to record the highest maternal and neonatal mortality rates globally, underscoring the urgent need to strengthen MPDSR systems to effectively notify, document and implement improvement strategies to prevent future avoidable deaths.

• Beyond country reports, which are often superficial, masking deeper issues, this analysis exposes widespread shortages of skilled maternity personnel trained in MPDSR, inadequate data harmonisation and standardised health metrics for maternal and newborn care, the absence of functioning MPDSR committees at health facility levels and insufficient leadership to translate maternal and newborn health pledges into action.

• Implementation of MPDSR varies across WHO African subregions, with the Eastern and Southern Africa subregion, which is primarily Anglophone, taking the lead, while the Central African subregion, mainly Francophone, lags behind, despite bearing the highest regional burden of maternal and newborn deaths.

• This analysis aims to guide policymakers and development partners in the WHO African region by highlighting critical areas requiring strategic investment, including workforce training and addressing retention challenges, such as the blame culture, to advance the maternal and newborn health agenda across the continent.

Introduction

Africa continues to bear the highest global burden of maternal and neonatal mortality, a trend that has persisted for decades.^{1 2} In 2020, approximately 287 000 women died during pregnancy and childbirth worldwide, with the WHO estimating that 95% of these preventable deaths occurred in low- and middle-income countries (LMICs).³ The sub-Saharan Africa (SSA) region recorded the highest maternal mortality ratio (MMR) in 2020, estimated at 545 (from 477 to 654) maternal deaths per 100 000 live births, more than 150 times higher than the 3 to 4 deaths per 100 000 live births observed in Australia and New Zealand.³ Further, the lifetime risk of a woman in an LMIC dying from maternal causes is about 33 times higher than that of a woman in a high-income country,⁴ and resource-constrained settings account for roughly 99% of maternal deaths due to obstetric complications⁵ and 98% of stillbirths globally.⁴

To end preventable maternal deaths, stillbirths and newborn deaths, WHO recommends comprehensive maternal and perinatal death reviews as part of quality-of-care improvement under the Maternal and Perinatal Death Surveillance and Response (MPDSR) initiative.⁶ The MPDSR concept gained momentum in the early 2000s with the Centres for Disease Control and Prevention’s (CDC) 2001 Strategy to reduce pregnancy-related deaths⁷ and the WHO’s 2004 Beyond the Numbers policy document which advocated reviewing maternal deaths and complications to enhance safety.⁸ These efforts culminated in the MPDSR framework, which drew worldwide attention around 2013 following the WHO’s Technical guidance for action to prevent maternal death,⁹ which was later complemented in 2016 by the Making every baby count: audit and review of stillbirths and neonatal deaths.¹⁰ The MPDSR process begins with identifying maternal and perinatal death cases at facility level, collecting detailed information and documenting causal factors. It then involves identifying and implementing improvement actions to avert future deaths. A dedicated MPDSR team at the facility conducts death reviews, generates actionable recommendations and ensures timely follow-up.⁶

Although countries in the WHO African region have shown political commitment to documenting all maternal and neonatal deaths, MPDSR roll-out in this region remains slow and uneven across and within countries and subregions.11,15 Indeed, country reports from Uganda,¹⁵ Kenya,¹⁴ Rwanda¹³ and Nigeria¹⁶ reveal wide disparities in MPDSR update and highlight substantial capacity and health-system challenges. Furthermore, country surveys indicate that MPDSR efforts in this region are often hampered by limited political and financial support, weak integration of surveillance data with civil registration and vital statistics systems and a disconnect between death reviews and broader quality improvement processes at health facilities.17,19

As of 2018, an estimated 30 (67%) countries in the WHO African region had begun integrating MPDSR into their health systems and were producing national MPDSR reports.²⁰ This regional analysis aims to evaluate progress in MPDSR implementation, identify key challenges and inform strategic actions. It specifically assesses the quality of national MPDSR reports against established recommendations and examines how these reports have influenced political actions and funding to support MPDSR efforts at country and regional levels.

Obtaining MPDSR country reports

We requested national MPDSR reports for 2015 to 2022 from all 47 WHO African region countries by engaging WHO country offices, reviewing Ministry of Health (MoH) and national statistics bureau websites, and liaising directly with maternal and newborn health (MNH) programmes. We included both standalone MPDSR reports and those embedded within maternal, newborn and child health (MNCH) reports accepting documentation in English, French and Portuguese. Those in French and Portuguese were translated into English prior to review. Of the 47 countries, 25 shared MPDSR documentation; the list of countries by WHO subregion is provided in online supplemental material S1. Maternal and perinatal death notification rates in our analysis were those reported directly in the country MPDSR reports. If a report did not provide rates but did provide the underlying counts and denominators, we calculated the rates ourselves. We did not independently retrieve or use additional District Health Information Software (DHIS2) data beyond what was presented in the reviewed country reports.

Data and information extraction

Two independent reviewers extracted data from each MPDSR report using a pre-established questionnaire (see data aspects in table 1). We first assessed compliance with WHO’s MPDSR guidelines and frameworks,^{13 21 22} and then calculated maternal and perinatal death notification and death review rates for countries lacking official rates and presented results using bar charts to highlight low performance. Next, we summarised country-specific leading causes of maternal and perinatal mortality (with bar charts for clarity) and evaluated how MPDSR response plans were implemented after each death review. The analysis concluded with a synthesis of common challenges undermining MPDSR across the WHO African Region. For countries with multiple 2015–2022 reports, we reviewed all submissions and cross-checked data consistency. Any data deemed inaccurate or discrepant MPDSR reports were discussed with the relevant MoH, which identified the most accurate and recent data source.

Table 1. Data type and information extracted from MPDSR reports and the DHIS2.

Data and information aspect	Description
Quality of MPDSR report	We assessed whether the overall MPDSR report structure included the core sections recommended by standard WHO guidelines with accurate definitions of key terms and correctly calculated metrics. Particularly, we conducted a brief review of the overall contents to assess the completeness and sufficiency of MPDSR data and information.
Death notification rate	Using DHIS2 data, we extracted maternal and perinatal death counts and the total notifications, then converted these into proportions to assess notification and reporting levels. For each country, we calculated the maternal death notification rate as dRm and the perinatal death notification rate as dRn, where dNm and dNn are the numbers of maternal and perinatal deaths notified through the MPDSR system, and dTm and dTn are the total maternal and perinatal deaths that occurred and reported in the DHIS2 system. These rates were used to gauge the completeness of death reporting across settings.
Maternal and perinatal death review rates	WHO guidelines stipulate that maternal and perinatal deaths should be comprehensively reviewed and disaggregated by sociodemographic characteristics. The rate of death review was calculated as the number of deaths that underwent review divided by the total number of deaths notified to the MPDRS system, with disaggregation by relevant sociodemographic strata.
Causes of death	Similarly, we explored data on causes of maternal and perinatal deaths and highlighted the most prevalent causes of death including deaths due to maternal complications. Completeness rate was derived by dividing the number of deaths with explicitly reported causes by the total deaths.
Delays of care	We explicitly extracted data on the three maternal delays of care to examine that carrying the most significant impact on mortality in the region.
Response actions	For each report, we documented the existence and functioning of MPDSR committees and what and how responses have been enacted to prevent future avoidable maternal and perinatal deaths.

DHIS2, District Health Information Software; MPDSR, maternal and perinatal death surveillance and response.

Quality of MPDSR reports

We received 32 reports from 25 countries, of which 29 were standalone MPDSR reports, with the other three from Ghana, Liberia and Rwanda consisting of integrated MNCH reports. All reports covered core MPDSR elements, although with varying completeness. These elements of such a report comprise an introduction and background, MNH metrics, maternal and perinatal death notification rates, death review and analysis, risk and cause analyses, responses to avert future deaths, MPDSR integration into the Health Management Information Systems (HMIS), and strategies to strengthen and monitor MPDSR.

Maternal and perinatal deaths surveillance and notification

Maternal and perinatal deaths are typically captured via routine HMIS primarily the Integrated Disease Surveillance and Response (IDSR) and the DHIS2. In addition, South Africa and Kenya have engineered dedicated MPDSR systems (ie, MaMMAS) to report maternal and perinatal deaths, and Uganda uses M-TRAC to notify and document mortality. Figure 1 shows that 16/25 countries reported maternal death rates, with only seven countries notifying more than 50% of maternal deaths. South Africa approached a 100% notification rate, while Ethiopia and Cameroon reported below 10% of maternal and perinatal deaths.

Figure 1. Maternal deaths notification rates in 16 African countries from 2015 to 2020.

Overall, reporting of maternal and perinatal deaths remains low and unreliable with substantial discrepancies between MPDSR reports and DHIS2 records. In our analysis, only about 2.2 to 30% of deaths recorded in DHIS2 were notified to MPDSR, indicating major data quality concerns. Specifically, figure 2 shows that perinatal deaths were notified in only 11/25 countries and at rates far lower than maternal deaths. Notably, many countries reported stillbirths and neonatal deaths but did not provide details needed to support death categorisation (eg, macerated or fresh stillbirths and early neonatal deaths). From 2015 until 2020, an estimated 11.7% of neonatal deaths were notified regionally. Nigeria had the lowest neonatal notification (ie, 1.7%) while Uganda had the highest rate with 38.2%.

Figure 2. Perinatal deaths notification rates in 11 African countries from 2015 to 2020.

Proportions of maternal and perinatal deaths reviewed

Most MPDSR reports include death metrics such as death distribution by sex, age and health province or district but pay less attention to underlying mortality drivers such as medical factors and sociodemographic and obstetric risk determinants. Proportions of death reviews are presented in figure 3 and show varying death review coverage. South Africa continues to lead with 100% of maternal and perinatal deaths reviewed, supported by its Confidential Enquiry into Maternal Death (CEMD) programme, which rests on the principles of death identification and assessment with a focus on avoidable causes of mortality and improving care and service provision.²³ In contrast, Gabon reviewed fewer than 0.1% (ie, 1 in 1000 deaths) of maternal deaths. Cote d’Ivoire, Burkina Faso and Cameroon also show limited review coverage, each under 30%.

Figure 3. Proportions of maternal deaths reviewed in 20 African countries from 2015 to 2020.

Figure 4 highlights perinatal death review coverage. Unlike maternal death reviews, only 14/25 countries reviewed perinatal deaths and those reviews covered fewer than 40% of cases. The regional average perinatal death review rate is 18.5% with Senegal, Cameroon, Burundi and Burkina Faso reviewing fewer than 3% of perinatal deaths. Most importantly, reviews using standardised perinatal death definitions (eg, stillbirths vs neonatal deaths) have been rarely implemented across the region.

Figure 4. Proportions of perinatal deaths reviewed in 14 African countries from 2015 to 2020.

Causes and risk factors of maternal and perinatal deaths

Common maternal and perinatal risk factors reportedly include sociodemographic factors such as maternal age, residence, marital status, religion, and obstetric factors like parity, gestational age, antenatal care, delivery status and mode, and the presence of health conditions such as HIV and anaemia. The leading causes of maternal death are obstetric haemorrhage (OH) and hypertensive diseases of pregnancy (HDP), which are particularly predominant in Nigeria and Madagascar (see online supplemental material S2). Other reported causes include maternal sepsis, labour dystocia, maternal anaemia and post-abortion complications, though data were insufficient for detailed analyses. Only eight of the 25 countries used ICD-10 for death classification. Thirteen countries attributed deaths to the three delays of care²⁴ with region-wide estimates indicating that 45.7% of deaths are due to delayed treatment, 43.2% attributed to delayed care-seeking and 15.8% due to delays reaching facilities. Uganda’s review linked the third delay to shortages of skilled birth attendants. Causes of perinatal mortality were documented in five countries, namely Gabon, Mali, Rwanda, Uganda and Zimbabwe, with 35 to 70% of these deaths representing stillbirths and most neonatal deaths occurring within the first hour after birth. The leading causes of perinatal death are asphyxia, prematurity and sepsis. For instance, asphyxia and prematurity accounted for 18.9% and 17.1% of neonatal deaths in Rwanda, respectively, while in Uganda birth asphyxia accounted for more than 50% of early neonatal deaths.

Responses and actions to prevent future maternal and perinatal deaths

Nearly two-thirds of MPDSR reports (59%) contained recommendations and actions aimed at preventing future maternal and perinatal deaths. These recommendations varied by health system context with some countries such as Ethiopia, Cameroon, Sierra Leone, Madagascar, Kenya and Uganda enacting MPDSR recommendations by level of death audit (eg, community, health facility, district, regional and national levels) while others like Zambia and South Africa layered these recommendations by geopolitical boundaries and, in some cases, by specific medical causes of death. For actionable responses to prevent future deaths, most countries urged community engagement, referral system strengthening and better inter-facility connectivity to address the first delay. To tackle the third delay, countries called for more skilled health workers and more equitable resource allocation based on workload. Specifically concerning actionable recommendations, we systematically checked whether each recommendation was SMART (ie, specific, measurable, achievable, realistic and time-bound) by identifying the target population (‘specific’ criterion) and by examining whether the recommendation contains a clear indicator (‘measurable’ criterion) that can be realistically achieved given the country’s context (‘achievable’ and ‘realistic’ criteria) within a specific timeframe (‘time-bound’ criterion). Our findings revealed that many of these recommendations lacked SMART attributes, limiting their implementation specifically in Benin, Senegal, Mali, Sierra Leone, Comoros and Madagascar. By contrast, countries implementing the CEMD, such as South Africa, reported more explicit plans (eg, the Safe Caesarean delivery intervention package) that addressed c-section competence gaps of South African anaesthetists and obstetricians and supported the enhanced basic-ANC-plus model to identify high-risk pregnancies early. In Uganda, the CEMD-led MPDSR informed the strengthening of comprehensive emergency obstetric and newborn care (CEmONC), prioritising 427 facilities across 25 districts and engaging communities and development partners (eg, USAID, RHITES, etc.), thus contributing to an 8% reduction in maternal and perinatal mortality from 2015 until 2020. Uganda’s CEmONC strengthening also improved process metrics, such as newborns who suffered from birth asphyxia, a proportion that declined from approximately 4.0 to 2.7% between January and September in 2020.

Monitoring and strengthening country MPDSR

Performance assessment of MPDSR was included in 91% of country reports, and all 25 countries examined had explicit policies and guidelines supporting MPDSR. Yet, most reports highlighted challenges in engaging communities and the private sector and difficulties linking and improving reporting and death notification systems. Community health programmes often confront technology constraints, such as limited access to health information software (eg, that enables linking data reporting systems between the private for-profit and the public sectors) opening to data quality issues, which we empirically detected in some MPDSR reports. The value of data quality control for quality improvement (QI) in healthcare is well-established,²⁵ and inaccurate MPDSR data or incomplete causal pathway analysis hinders evidence-based responses. Our analysis also found that many countries lack functioning facility-level MPDSR committees with substantial gaps in MPDSR skills where those committees exist, exacerbated by blame and confidentiality concerns surrounding maternal and perinatal deaths. MPDSR implementation is often limited to basic clinical case reports rather than comprehensive audits, and health workers face demotivation due to high workloads and understaffing, as seen in Rwanda and Tanzania. Two particular system-level challenges which we noted are the blame culture and the inadequate ICD-10 classification skills in Ethiopia, Madagascar and Zimbabwe, for instance. These challenges affect human resources for health (HRH) retention initiatives and worsen maternity care capacities. Limited funding and weak leadership further hamper progress. Common recommendations emphasise linking MPDSR to facility-level quality improvement, fostering teamwork and a no-blame culture and increasing funding for skills development.

Takeaway message and policy implications

This review shows that MPDSR implementation in the WHO African Region is still hampered by multiple challenges. MPDSR requires robust data management systems supported by skilled and committed personnel within enabling political and financial environments, conditions that are often lacking in many African countries.^{26 27} Using 25 countries representing 53% of countries in WHO AFRO which were assessed for MPDSR implementation over a period of 7 years from 2015 until 2022, we found that over half of countries examined (56%) belong to the Eastern and Southern Africa subregion, suggesting somewhat more advanced MPDSR progress in these subregions than in Western (36%) and Central Africa (8%) subregions. Although all country MPDSR reports contain core MPDSR information, coverage and completeness vary markedly, death notification and review are suboptimal, with many reports resembling simple patient summaries rather than full MPDSR implementation. Six of the 25 countries still operate the Maternal Death Surveillance and Response (MDSR) model without a dedicated perinatal component. Perinatal death reviews are particularly weak, often lacking details to distinguish stillbirths from early neonatal deaths or to apply a consistent perinatal definition^{28 29} as per the WHO guidelines for instance (ie, a perinatal death is a death of the baby that occurs at 22 completed weeks of gestation and over, during childbirth, and up to seven completed days of life).³⁰ Without harmonised data reporting and metric calculation, regional trends and MPDSR impact remain difficult to assess. Regional estimates indicate coverage reporting of only 48% of maternal deaths and 11.6% of perinatal deaths with country-level corroboration. For instance, only 45.3% of Ethiopian health facilities reported and notified maternal and perinatal deaths timely, and just 54.5% conducted death reviews.³¹ Varying but still unsatisfactory levels of maternal and perinatal death notifications and MPDSR implementation have similarly been observed in Rwanda,¹³ Kenya,¹⁴ Tanzania¹² and many African, Asian and Latin American LMICs.2232,35

Conclusion

This review highlighted that important challenges that impede the successful implementation and upscaling of MPDSR in Africa include widespread shortages of skilled maternity workforce with specific training in MPDSR, inadequate data harmonisation, including standardised metric definitions, poor or inadequate funding, the absence of functioning MPDSR committees at health facility level, and the lack of strong leadership to materialise political MPDSR pledges. These bottlenecks form a significant pillar of focus for future interventions and should guide countries, WHO and other development partners in their efforts to improve and scale MPDSR in the region. Policy discussions and research indicate that MPDSR provides an opportunity for health systems to understand the determinants of maternal and perinatal deaths in order to improve the quality of care and prevent future avoidable deaths. If successfully implemented in Africa where mortality remains high, this could nudge efforts to attain global and regional maternal and newborn health targets, especially those pledged in the Sustainable Development Goals agenda,³⁶ the Global Strategy for Women’s, Children’s and Adolescents’ Health,³⁷ the Every Newborn Action Plan (ENAP) and the strategy³⁸ towards Ending Preventable Maternal Mortality (EPMM) strategy.^{39 40} However, MPDSR is a complex approach requiring concrete actions and full commitment of health workers, health facility managers and policymakers and should be supported by a bold financial commitment. To attain the overall MPDSR objective and reduce maternal and perinatal mortality, the continent needs to boost MPDSR investments geared towards workforce enhancement including MPDSR training, improve and harmonise data quality and definitions of maternal and perinatal metrics, create and operationalise MPDSR committees in each health facility, and supplement these efforts with a strong political commitment with community and facility-level involvement. Finally, this analysis unveiled the existing subregional disparities in the implementation of MPDSR and hinted that countries implementing MPDSR are not necessarily those with higher maternal and perinatal mortality rates in the region. We thus suggest conducting further primary research to explore factors that could explain these subregional imbalances to inform remedial policy actions.

The author is a staff member of the World Health Organization. The author alone is responsible for the views expressed in this publication and they do not necessarily represent the views, decisions or policies of the World Health Organization.

Data availability statement

Data may be obtained from a third party and are not publicly available.