Introduction

Globally, an estimated 260,000 maternal deaths occurred in 2023 [1]. Despite accounting for only 10% of the world’s population [2], Sub-Saharan Africa contributes 70% of these deaths (212,000 maternal deaths) [1]. In Tanzania, the maternal mortality ratio is estimated at 104 maternal deaths per 100,000 live births, based on the 2022 Tanzania Demographic and Health Survey and Malaria Indicator Survey. Although this represents a decline compared with previous estimates, it still translates to more than 2,000 maternal deaths annually, indicating a persistently high burden of preventable mortality [3]. Most maternal deaths in SSA occur during or within 24 hours of childbirth [4–6]. An increasing number of births in the region occur in health facilities [5]. Thus, improving the quality of delivery and post-delivery care in health facilities is a critical intervention to reduce maternal mortality [7–9] and achieve Sustainable Development Goal 3.1, which aims to reduce the global maternal mortality ratio to no more than 70 maternal deaths per 100,000 live births by 2030 [10].

Maternal death reviews (MDR) are key interventions that aim to identify and address gaps within the health care system that contributed to the death of a woman during pregnancy, delivery or the first 42 days after delivery [11]. Tanzania has been and continues to be among the top 10 countries with a high number of maternal deaths per year [12]. In 2006, Tanzania implemented maternal and perinatal death reviews for the first time, as recommended by the World Health Organisation (WHO) guidance Beyond the Numbers, Reviewing Maternal Deaths, and Complications to Make Pregnancy Safer [11]. The MDR process in Tanzania was scaled up into a surveillance system following the publication of the WHO Maternal Death Surveillance and Response (MDSR) technical guidance [13,14]. This was captured in the first national guidelines issued in 2015 and revised in 2019 [15]. As with other countries in the region, Tanzania has extended the surveillance system to include perinatal deaths [15]. The MPDSR guidelines are intended to help eliminate preventable deaths in both facilities and communities by continuously documenting and analysing the magnitude, trends, and impact of maternal mortality. The guidelines prioritise a focus on strengthening accountability in responding to gaps in health systems, investigation of community deaths through verbal autopsies, and use of the WHO applications of ICD-10 to deaths during pregnancy, childbirth, and the puerperium (ICD-MM) and deaths during the perinatal period (ICD-PM) [11–13]. The guidelines provide ground rules for conducting death reviews, which include maintaining an atmosphere of “no shame, no blame, no name,” respecting the confidentiality of the discussion and accepting criticisms from peers to improve clinical care [15].

Tanzania MPDSR guidelines are fully institutionalised at the community, health facility, district, regional, and national levels. For a meaningful critical MDR process, highly skilled clinicians organised in health facility committees must be present during the review. These professionals are generally available in hospitals (mainly regional, zonal and national), often in towns and cities. Several assessments of the national MPDSR guidelines implementation indicated delays and suboptimal quality in conducting reviews in lower-level health facilities [16,17]. Delays and suboptimal quality reviews were partly the result of a lack of commitment among health managers, time constraints, and limited training of reviewers [16–20]. Sub-optimal reviews led to incorrect assignment of causes of death, weak root-cause analysis (RCA) to identify gaps in the quality of care, and a failure to develop appropriate action plans that include specific, measurable, attainable, realistic, and time-bound (SMART) recommendations [21].

Experts recommend that MDSR be institutionalised within the Ministry of Health (MoH) to provide oversight, policy guidance, and support to ensure a functional and effective process. As a response to the COVID-19 pandemic, the MoH instituted limited physical travel for review meetings to mitigate the spread of the virus. This led to the adoption of virtual social networking applications such as Zoom, WhatsApp, Microsoft Teams, and Google Meet to facilitate technical interactions and support clinical decision-making for junior healthcare providers.

Despite the nationwide institutionalisation of MPDSR in Tanzania, challenges persisted in achieving the quality, depth and consistency required for effective maternal death reviews, especially in facilities with limited specialist involvement. To address these gaps, the Ministry of Health introduced the Virtual Maternal Death Review (VMDR) model in late 2021 as a health system innovation that strengthened the review process by using real-time digital platforms to bring together multidisciplinary expertise from across the country. VMDR enabled joint clinical analysis, facilitated collaboration between specialists regardless of location, improved efficiency through streamlined scheduling and digital record-sharing, promoted shared learning, established a robust platform for providing unbiased reviews and increased the accuracy of identifying modifiable factors. Since implementation in 2022–2023, the model has expanded access to specialist input, enhanced accountability among regional and district teams and accelerated the translation of review findings into targeted mentorship, supportive supervision and timely system-level actions. Through these improvements, VMDR represents an innovative advancement of the existing MPDSR process. This paper describes the implementation of VMDR in Tanzania, the patterns and modifiable factors identified, and the actions taken to improve the quality of care. As more countries seek to optimize their maternal mortality surveillance systems, learning from the experience of Tanzania may support an excellent opportunity to introduce the widespread usage of online review meetings in similar settings.

Methods

Results

A total of 369 VMDRs were conducted, and 687 maternal deaths that occurred from January 2022 to December 2023 were virtually reviewed. Reviewed cases were selected from maternal deaths that occurred in zonal referral hospitals (n = 151; 22%), regional referral hospitals (n = 159; 23%), district hospitals (n = 179; 26%), health centres (n = 136; 20%), dispensaries (n = 42; 6%), and communities (n = 20; 3%). Almost one-half of the 3,041 deaths notified to the MoH were from 7 regions (Dar es Salaam, Mwanza, Morogoro, Kigoma, Mtwara, Mbeya, and Tabora) (Table 1). Almost one-fourth (22.6%) of notified maternal deaths in 2022–2023 have been reviewed through VMDRs. The percentage of VMDR-reviewed deaths ranged from 11.3% in Dar es Salaam to 59.7% in the Pwani region. Because VMDR sessions follow a structured rotation in which each region is allocated scheduled opportunities to present cases regardless of the absolute number of maternal deaths, regions with fewer notified deaths were able to have a higher proportion of their cases reviewed virtually. This reflects the design of the review schedule rather than the preferential selection of regions.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Number of notified maternal deaths and proportion of virtually reviewed maternal deaths by region in 2022–2023, Tanzania Mainland.

https://doi.org/10.1371/journal.pone.0344858.t001

The mean age of the 687 women who died of maternal causes and were VMDR-reviewed was 29.5 years, with a range of 15–50 years; 9% of women were less than 20 years old at the time of death.

Of the 687 maternal deaths VMDR-reviewed, 90% were due to direct obstetric causes, and 9% were due to indirect causes. The leading direct causes were obstetric haemorrhage (49%), hypertensive disorders of pregnancy (16%), anaesthesia-related complications (10%), puerperal sepsis (6%), obstetric embolism (4%), and abortion-related complications (3%). The leading indirect obstetric causes were anaemia (5%), malaria (2.6%), and preexisting cardiac conditions (1.5%). The cause of death was not established in 1% of cases due to inadequate information during the review (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Distribution of causes of deaths in 687 virtually reviewed maternal deaths in 2022–2023, Tanzania Mainland.

https://doi.org/10.1371/journal.pone.0344858.g001

The leading non-medical factors contributing to the maternal deaths reviewed through VMDRs were related to insufficient clinical skills (82%), healthcare providers’ attitudes (69%), suboptimal leadership and accountability (42%), inadequate surgical care (37%), inadequate antenatal care (34%), lack of blood (32%), poor anaesthesia practices (28%), client and community factors(27%) and lack of equipment and supplies (24%) (Fig 2). These causes often overlapped, each case having multiple contributing factors.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Main clinical and non-clinical modifiable factors in 687 virtually reviewed maternal deaths in 2022–2023, Tanzania Mainland.

https://doi.org/10.1371/journal.pone.0344858.g002

The VMDRs recommended specific actions (responses) to address avoidable factors related to each maternal death. Most agreed actions were related to 1) increased capacity of staff to manage obstetric complications that included improving surgical skills, improving consultations and mentorship (e.g., tertiary hospitals to improve multi-disciplinary consultations), and improving labour and postpartum monitoring; 2) adherence to the use of guidelines and standard operating procedures; and 3) increasing blood collection, processing, distribution and storage systems. This went hand in hand with the responsibilities of Regional Health Management Teams (RHMTS) and Council Health Management Teams (CHMTS) to ensure the functionality of the health system at primary healthcare levels, including the community.

In response to the VMDRs, RHMTs and CHMTs set up expert teams to build capacity based on identified gaps. Examples of the actions taken by the expert teams included safeguarding correct anaesthesia practices at one of the district hospitals, ensuring that specialist doctors in one of the zonal referral hospitals started a full-time night shift, connecting regional doctors to mentors in a zonal referral hospital to brush competencies; relocation of an anaesthesia machine to another health facility; and commencement of onsite mentorship to address key identified dysfunctions. The VMDRs managed to reinforce adherence to standards of care in the management of intrauterine foetal deaths, use of misoprostol for induction of labour, removal of retained products of conception, and use of low molecular weight heparin to prevent and treat obstetric thromboembolism.

At the national level, the MoH, professional associations, and partners have intensified resource mobilisation and partnerships, creating a system for healthcare capacity building. This is shown by developing a national MNCH mentorship guide and supporting clinical attachments of healthcare providers from lower-level facilities to high-level facilities. The Government has established mentorship in health facilities and commenced a virtual perinatal death review process and a weekly virtual Continuing Medical Education Forum (Table 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. National Response using the VMDR Data.

https://doi.org/10.1371/journal.pone.0344858.t002

Discussion

Studies and reports from more than a decade of MPDSR implementation in Tanzania have consistently shown limitations in the quality of reviews and weak accountability in follow-up actions, largely driven by insufficient reviewer capacity and inconsistent clinical oversight [16,18,20]. Against this background, this study set out to document the process and lessons learned from introducing VMDR as a strategy to strengthen the review system and address these persistent gaps.

Through the review of 687 maternal deaths, VMDR revealed recurrent gaps in clinical care, including limited clinical skills, suboptimal provider practices, and provider attitudes that affected the management of obstetric complications [28,29]. The process also demonstrated how VMDR supported greater accountability, as both managers and frontline healthcare providers were more responsive to the dysfunctions identified during the virtual reviews and more engaged in implementing corrective actions [30,31].

The introduction of VMDR strengthened the quality and consistency of maternal death reviews in Tanzania by widening access to specialist expertise, improving the identification of modifiable factors, and supporting more timely follow-up actions across levels of the health system. Similar benefits have been observed in other virtual and digital review initiatives. International virtual confidential reviews conducted across 11 low- and middle-income countries, including Ethiopia, Nigeria, Kenya, Ghana, South Africa, Zambia, Uganda, Pakistan, Bangladesh, Jordan, and Morocco, demonstrated that remote expert panels improved identification of clinical and health system gaps and strengthened learning and accountability in maternal death reviews [32]. In addition, telehealth-supported maternal health initiatives implemented in low-resource settings, including underserved areas of sub-Saharan Africa and South Asia, have shown how digital platforms can enhance access to specialist input and support timely clinical decision-making in maternal care [33]. Unlike these studies, which focused on discrete clinical domains or external review mechanisms, VMDR established a continuous, nationally coordinated platform embedded within the MPDSR system, enabling collaborative review, consistent interpretation of clinical events, and strengthened accountability.

Through this approach, VMDR created a structured environment where clinicians and managers at facility, district, regional and national levels could jointly analyse cases, clarify gaps in clinical care and system functioning, and agree on targeted actions. This aligns with recommendations that maternal death reviews should directly inform ongoing quality-improvement activities and not remain isolated discussion forums [11,24]. By linking case review with follow-up actions, VMDR helped translate MPDSR findings into concrete improvements, including reorganisation of staffing arrangements, strengthened referral pathways, targeted mentorship and supportive supervision, and mobilisation of specialist support.

Participation of senior clinical leaders and health system managers further reinforced implementation, as their presence strengthened accountability, heightened awareness of identified challenges and facilitated timely responses. These patterns mirror broader lessons from MPDSR implementation research across sub-Saharan Africa, where leadership engagement and structured review processes are consistently associated with improved responsiveness and more effective action [18,19]. Together, this evidence highlights the value of VMDR as a coordinated national mechanism that advances both learning and accountability, complementing global efforts to enhance the quality of maternal healthcare.

From the VMDR review process in Tanzania, maternal deaths predominantly occurred in health facilities and were largely associated with preventable or modifiable factors, including gaps in clinical knowledge, delayed recognition of complications, inconsistent adherence to standards of care, and inadequate supervision. These patterns closely mirror findings from MPDSR studies in Tanzania and other sub-Saharan African settings, where substandard clinical care, health system delays and poor decision making remain major contributors to maternal mortality [18,19,21,28,34]. In addition, this study identified provider attitude as an important contributor to suboptimal care, a finding that aligns with evidence from Kenya, Ghana and India showing that disrespect, poor communication and limited autonomy undermine women’s experiences of facility-based maternity care and compromise quality [35]. Together, these findings highlight that improving maternal outcomes requires strengthening both technical competence and the behavioural dimensions of care, reinforcing the dual clinical and person-centred focus emphasised in the VMDR process.

The relatively high proportion of anaesthesia-related maternal deaths identified through VMDR may partly reflect stronger documentation for these cases, but it also suggests persistent gaps in anaesthesia skills and facility readiness for emergency obstetric care. Similar concerns have been raised in evidence from low and middle-income countries, where anaesthesia-related mortality, though often underreported, represents an important and preventable contributor to maternal deaths [36]. These findings highlight the need for targeted skills building, strengthened supervision and continued clinical mentorship for anaesthesia providers at all levels of the health system.

This study had some limitations. Case selection depended on the completeness of clinical documentation and the quality of prior facility or district reviews, which may have resulted in underrepresentation of cases with inadequate records, a challenge noted in other MPDSR assessments in Tanzania and comparable settings [21,37]. The reliance on facility-based deaths limited the assessment of community-level delays, including delays in seeking care or reaching a facility, which remain important contributors to maternal mortality in many sub-Saharan African contexts [27]. These limitations reflect long-standing structural challenges within MPDSR implementation rather than weaknesses of the VMDR model itself. They underscore the need for continued investment in accurate documentation, strengthened surveillance systems, and improved integration of community level information to ensure a more comprehensive understanding of preventable maternal deaths.

The distribution of VMDRs by cause is not representative. Still, it generally aligns with global, regional and national causes of MDS, where obstetric haemorrhage and pre-eclampsia are the leading causes [5,6,28]. One exception is the high proportion of maternal deaths related to obstetric anaesthesia documented in VMDRs, which is several times higher than in the published literature. A recent systematic review found that anaesthesia accounted for an estimated 3.5% of direct maternal deaths in low- and middle-income countries [36]. Whereas the selection of cases submitted for virtual reviews may have influenced the percentage distribution of causes of maternal deaths, the true magnitude of anaesthesia-related deaths in Tanzania remains unclear.

Conducting VMDR consistently and regularly may help key players in maternal health care and related disciplines overcome their comfort zones and provide quality care through improved competence, practice, and attitudes. The success of VMDRs relies on regional, subnational, and national champions, including professional associations, who are dedicated to the entire review process and following up on recommendations.

The quality of documents used in reviews may directly impact the recommendations from maternal death audits. Missing or incomplete patient information and data on the care received may lead to unsatisfactory conclusions on the’ causes and circumstances of maternal deaths and failure to identify substandard care and corrective actions correctly [36–38]. In VMDR, data abstraction is thorough, with clear event narration and documentation on the cause of death and underlying factors. If records are incomplete, an additional step is triggered via WhatsApp to reach out to all healthcare providers involved in the case management and complete missing information. This approach has enabled the generation of clear documentation on the cases that were reviewed and, in turn, enabled accurate descriptions of medical and non-medical causes that contributed to the death and improved the quality of recommendations. The same has influenced the government and stakeholders to adopt the best approaches to improve the quality of care, focusing on improving healthcare providers’ competency, practice and attitude through pre-service courses and in-service mentorship and supported supervision.

The selection of cases for VMDR is delegated to different regions. Therefore, the review team does not have control over the documentation and presentation of each case submitted for virtual review. Maternal deaths that are identified by regions have already been reviewed by health facilities and sometimes by district and regional committees. The selection of the VMDR cases may have been biased toward cases with complete medical information and documentation. Maternal deaths with incomplete or inadequate information or those whose care spanned over several wards or health facilities or required readmission may not be selected for VMDRs, which reduces the opportunity to learn from and formulate actions toward quality improvement targeting the most difficult cases. As VMDRs usually do not extend to maternal deaths outside health facilities, few community deaths were reviewed by the VMDR. Further, some facility MDR committees may not review maternal deaths due to indirect maternal causes, which may explain the low contribution of these causes among VMDR cases.

Conclusion and recommendations

Maternal and Perinatal Death Surveillance and Response (MPDSR) remains a critical but complex undertaking in low-resource settings. Although widely implemented, many countries continue to face challenges in sustaining high-quality reviews and ensuring that findings translate into meaningful action. The WHO implementation guidance recognises that MPDSR must be adapted to the country context, with mechanisms that reflect the realities of each health system.

Tanzania’s experience with Virtual Maternal Death Reviews (VMDR) demonstrates that structured digital platforms can reinforce MPDSR implementation by widening access to expert reviewers, improving the consistency of clinical analysis, and strengthening accountability for follow-up actions. The ability of VMDR to convene clinicians and managers from multiple levels of the health system in real time has created an environment where gaps in care can be clarified quickly, lessons shared more broadly, and recommendations acted upon with greater urgency. The sustained participation of national and regional leaders has also helped uphold core MPDSR principles of confidentiality and a non-punitive review culture, which have been essential for maintaining engagement.

Overall, VMDR has shown feasibility and acceptability as a routine review approach and has strengthened the link between case review and corrective action. Continued leadership commitment, investment in capacity building, and reinforcement of a supportive review culture will be key to preserving these gains. As countries explore innovative models to overcome persistent MPDSR implementation challenges, the Tanzanian VMDR experience offers a promising example of how digital solutions can enhance learning, improve clinical oversight, and promote accountability for maternal health outcomes. Further evaluation of VMDR in comparable settings would help build the evidence base for WHO and global partners to consider broader adoption of similar virtual review models across member states.

Supporting information

Acknowledgments

We thank our Ministers, Permanent Secretaries, Chief Medical Officer, Directors from MOH and PORALG, and professional bodies and associations for supporting VMDR. We sincerely thank the VMDR steering and coordination team for their tireless work to make this intervention run uninterrupted for three years from 2022 to date. The names these members are as follows; Sunday Alfred Dominico(MPDSR Focal, MOH), Felix Ambrose Bundala – Assistant Director (MPH), Saturin Manangwa – MOH (Nurse Midwife), Ziada Sellah – MOH Director Nurse Midwife Services, Upendo Mamchony – MOH (Nurse Midwife), Cecilia Msafiri – MOH (Nurse-Midwife specialist), Agness Mbio – MOH (Nurse Midwife), Faraja Jacob Mgeni – MOH (Nurse Midwife), Jackline Ndanshau – MOH (Nurse Midwife), Juma Daimon Nyakina – Bukoba Regional Referral Hospital (RRH) (OBGY), Paul John Sanka – Dodoma RRH (Nurse Midwife), Grace Aziza Machenje – MOH (Nurse Midwife), Naibu Samweli Mkongwa – MOH (MPH), Ulimbakisye Yoram Macdonald – MOH (OBGY), Victoria Lyimo – MOH (Nurse Midwife), Enock Mwantyala – MOH (MPH), Angela Leornard – MOH (Paediatricaian), Rachel Greyshom Yangwa – MOH (Nurse Midwife), Ephraim Kaphilimbi Shilla – MOH (Nurse Midwife), Fidea Obimbo – MOH (Nurse Midwife), Naomi Chamhene - - MOH (Nurse Midwife), Tito Amosi Chaula – Nkinga RH (OBSGY), Alfred Laison Mwakalebela – Iringa RRH (OBGY), Edwin Rwebugisa Lugazia – MUHAS (Anaesthesiologist), Donald Micah Maziku – Tosamaganga Hospital (OBSGY), George Rweyemamu Alcard – Mbeya RRH (OBSGY), Elias Asiwelo Kaminyoge – Mbeya (Meta) Zonal Hospital (OBGY), Chrisostom Clarence Lipingu – USAID-Afya Yangu/Jhpiego (OBGY), Rita Elias Lyamuya – Morogoro RRH (OBGY) Ali Said – MUHAS (OBGY), Elias Gregory Kweyamba – TTCIH (OBSGY), Berno Kasmir Mwambe – Marie Stopes Hospital (OBGY), France John Rwegoshora – Mbeya (Meta) Zonal Hospital – (OBGY), Titus Bonifas Mmasi – Monduli DC (OBGY), Shedrack Andrew Magambo – UDOM Hospital (OBGY), Patrick Joseph Kushoka – DCMC Hospital (OBSGY), John Deogratias Lawi – Iringa RRH (OBGY), Matilda Michael Ngarina – MNH (OBS/GYN), Pendo Samwel Mlay – KCMC (OBGY), Emmanuel Imani Ngadaya – Tosamaganga Hospital (OBGY), Chetan Pratapsinh Ramaiya – (OBGY), Nestory Andrew Mkenda – USAID-Afya Yangu (OBGY), Naomi Amon Mwamanenge – Aghakan Hospital (Neonatologist), Lucy Lawrence Mpayo – Neonatologist, Halima Mohamed Kassim – Dodoma RRH (Paediatrician), Joyce Stephen Gimonge – Geita RRH (Paediatrician), Emmanuel Rhubera Mafwiri – Njombe RRH (OBGY), Georgina Balyorugulu – Kamanga Hospital (Paediatrician), Yustina Adam Tizeba – Shinyanga RRH (Paediatrician), Tatu Seif Mbotoni – Iringa RRH (Paediatrician), Castory Jerome Mwanga – Simiyu RRH (Paediatrician), Beatrice Erastus Mwilike – MUHAS (Nurse midwife), Karoli Jerome Muna – Kondoa DH (OBGY), Enock Hashim Dolli – Bukoba RRH (Paediatrician), Mwanahawa Ramadhani Suba – Nkinga RH (Paediatrician), Amina Shomari Shabani – Morogoro RRH (NurseMidwife), Happyphania Mathew – Kibaha TC Hospital (Paediatrician), Jafari Bakari Lutavi – MNH (Nurse Midwife), Bahati Katembo – Benjamin Mkapa Hospital (Nurse Midwife), Beatrice Bernard Kambuga – DIHAS Milembe (Nurse Midwife), Elizabeth Mkashabani – Paediatrician, Florentina Mashuda – CUHAS BMC (Paediatrician), Dismas Matovelo – CUHAS BMC (OBGY), Monica Mumwi – Ndanda RH (Paediatrician), Mary Sapali – Iringa RRH (Nurse Midwife), Nikolas Albert Sagumo Chotta – Dodoma RRH (Paediatrician), Lucy Mabada – TAMA (Nurse Midwife), Feddy Mwanga – TAMA (Nurse Midwife), Hawa Ngasongwa – Morogoro RRH (Paediatrician) Elisamia Ngowi – Aghakan Hospital (Paediatrician). Also, we would like to thank the Regional Council and Health Management Teams for their support in case identification, preparation, presentation, and follow-up of the implementation of VMDR action plans. We thank everyone who participated and provided input to VMDR from Tanzania and abroad. Without all of you, the VMDR would not be possible.