To guide the scoping review, we developed a protocol, following the Arksey and O’Malley approach, as further refined by the Joanna Briggs Institute Manual and others (Arksey and O’Malley, 2005; Levac et al., 2010; Aromataris, 2020; Peters et al., 2022). The protocol, which is annexed as Supplementary File 1, describes the method and justifies our choice of research design. It outlines how we intend to carry out the six stages of the review recommended by the Arksey and O’Malley approach, which are (1) identifying the research question, (2) identifying relevant studies, (3) study selection, (4) data collection, (5) data summary and synthesis of results and (6) optional consultations.

We included all literature published or accessible online in English between 2010 and November 2022 that reports on legal issues arising during the implementation of maternal or MPDSR systems The year 2010 marks the starting point of policy initiatives that ultimately established the MDSR (Danel et al., 2011; WHO, 2016a). We limited the selection to literature discussing MDSR implementation to gauge the empirical experience of countries.

We excluded publications that exclusively deal with perinatal or neonatal review systems, as well as those solely discussing confidential enquiry into maternal death. Confidential enquiries were excluded due to their fundamental differences from the MDSR. Among others, unlike in MDSRs, reviewers in confidential enquiries are typically external bodies not affiliated with the facilities where deaths occur. Besides, such enquiries focus on facility deaths, and review is often conducted on selected cases. Furthermore, they are adopted in settings with functioning statistical infrastructure (WHO, 2004; Hussein et al., 2009; Moodley et al., 2014).

Assisted by university librarians, we initially tested our search strategy on PubMed and JSTOR and determined the search terms. While Table 1 summarizes the search keys, the full search history is provided in Supplementary File 2. In December 2022, we conducted a comprehensive search on Medline, Embase, Web of Science, PubMed Central, JSTOR, HeinOnline, and Westlaw UK databases. In addition, we conducted an open search using Oria, the search engine of the University of Bergen (UiB), Norway, to identify further studies, including grey literature. We selected the first 100 hits, based on the assumption that subsequent screening would yield few additional relevant works (Stevinson and Lawlor, 2004). Furthermore, we conducted a hand search on the WHO online library. Through June 2023, we also identified more publications from an alert system linked to Oria, which we had set up to track new publications. Finally, we reviewed the reference lists of selected papers for any further records.

To test the feasibility and ensure consistency in approach, M.B.K., A.M. and I.M. initially screened 15 titles independently. Subsequently, M.B.K. screened titles and/or abstracts of the records, and A.M. verified the results. Similarly, five records were randomly selected, and the three reviewers independently conducted a full-text review against the set criteria. During this process, we observed that several papers contained minimal data material to address our research question, and we decided to further limit our selection to papers that discuss at least one substantive legal issue, as described earlier. Following discussions and consensus, M.B.K. performed a full-text review of the records and A.M. verified the results. Discrepancies were resolved through discussion, involving I.M. A list of the studies selected is provided in the record index, Supplementary File 3.

The research team (M.B.K., A.M., I.M. and K.M.M.) developed a data charting form based on the Bain and Kongnyuy framework, categorizing the legal issues around the MDSR into three sets: data, people and the use of findings. After pilot testing with five records and gaining increased familiarity with the literature, we found that the three categories leave out some aspects and issues of a legal nature relevant to the functioning of the MDSR. We, therefore, added the fourth category, ‘legal regulation,’ as described earlier.

Finally, M.B.K. charted the data and A.M. verified the results. All authors then discussed the findings. The charting form was continuously revised to reflect the review team’s consensus regarding the form and the results included. Data items charted included key reference characteristics and background, including authors’ names, year of publication, country of focus, the scale and level of study, research design and key relevant findings. The substantive findings were recorded into the four categories by relevance. A description of key characteristics for each included study is provided in Supplementary File 4, and key results are summarized in Table 2. For further details of results, see Supplementary File 5.

We analysed the data thematically using a deductive-inductive approach (Clarke and Braun, 2017). Accordingly, the data were initially organized into the four categories described earlier. Then, we inductively and iteratively developed subcategories from the gathered data sets, which formed the basis for reporting the results. We also provided a descriptive narration and numerical summary of the study characteristics as part of the process.

The protocol was presented on several occasions to gather feedback, including to members of a multi-country, multidisciplinary research project working on the MDSR system at the University of Bergen,¹ in November 2022. It was also discussed in a legal methodology course in March 2023 and at an international seminar in May 2023, which was attended by experts in medical and public health, as well as health ethics. Preliminary results were shared at a MATRISET seminar in early November 2023. Additionally, expert librarians from both the faculties of Medicine and Law at the UiB were consulted for advice on search terms and strategies.

The review decision process is reported using an adapted ‘Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews’ Checklist (Tricco et al., 2018), available in Supplementary File 6.

The initial search resulted in a total of 1174 records. All citations were imported into the EndNote20 citation manager, where duplications were removed. After removing duplicates, we screened 546 unique results by title and/or abstract, yielding 87 publications eligible for a full review. An additional three records were identified—two from citation references and one from an alert to new publications. This led to a total of 90 records undergoing full-text review, of which 31 met the inclusion and exclusion criteria. Twenty-eight of the records were academic journal articles, and three were study reports. All the academic articles were published in journals dedicated to medical, public health or health sciences (20 in total). Figure 1 shows the flow diagram documenting the screening process, and Table 3 summarizes the number of search results by source.

Of the 31 included records, 22 examined the implementation of the MDSR system in specific countries, covering a total of 17 countries. The rest reported the status of MDSR implementation in certain regions, a selection of countries or countries across the globe. Most individual countries represented were in Africa (13), two were in Asia, and the remaining two were in the Americas. In terms of the publication year, 18 records were published since 2019, and 11 were published between 2014 and 2018, with only two records published between 2010 and 2013. An approximately equal number of records examined the implementation of MDSR at both national and subnational levels. Most of the records (23) were qualitative studies, with the remainder employing mixed methods. While most of the records (27) discussed the full cycle of the MDSR system, a few of them (four) focused on one or more of the stages. Notably, none of the records exclusively or significantly discussed legal issues. Next, we present a summary of the results by category.

This category encompasses legal issues and standards that relate to how data are collected, stored, used or transferred during the implementation of the MDSR. Numerous studies discussed various aspects of the data used in the MDSR and the challenges faced in accessing different information sources and systems. The literature focused on the safety of the MDSR data after its collection and is often discussed under ‘anonymity’ and/or ‘confidentiality’, without providing further definitional information. Even if the studies showed that the privacy rights of individuals are threatened in the process, particular privacy laws were rarely mentioned.

While none of the studies included addressed pre-data collection legal considerations, a few studies briefly described the forms and timing of anonymization or the mandated destruction of collected data. While some countries required data anonymization after a facility review is conducted, in others, the data could remain identifiable until a report is published (Armstrong et al., 2014; Melberg et al., 2020; Kodan et al., 2021). However, none of the studies discussed legal standards of anonymity.

Several studies report weaknesses in securing the privacy of the data gathered (Armstrong et al., 2014; Melberg et al., 2020; Kodan et al., 2021; Abouchadi et al., 2022). In a few instances, it was also indicated that the identity of health workers and facilities was deliberately disclosed to the public as a form of punishment for poor attendance and engagement in review sessions (Said et al., 2021). A study from Indonesia further noted that the MDSR was treated as a forum for performance evaluation, leading to the public disclosure of findings, including the identity of the health workers and the institution involved (Cahyanti et al., 2021). Such breaches of anonymity and confidentiality led to poor dialogue during death reviews (Kodan et al., 2021). Recommendations for the adoption of appropriate laws to ensure confidentiality and anonymity of the process were provided although they often lack specificity (WHO, 2016a; Kodan et al., 2021; Willcox et al., 2023).

This category explores the forms, exacerbating factors and effects of concerns of a legal nature that actors involved in the implementation of the MDSR manifest in relation to their participation in the MDSR process. The studies highlighted a pronounced sense of vulnerability among participants, with a predominant theme of experiencing or fearing blame for maternal deaths. Health workers, who are the backbone of the MDSR process, emerged as the principal target of these concerns.

Most studies reported a widespread association of the MDSR system with blame (Agaro et al., 2016; WHO, 2016a; Abebe et al., 2017; Smith et al., 2017a; Melberg et al., 2019; Naik et al., 2020; Kinney et al., 2021; Willcox et al., 2023). Broadly stated, it was believed that the way the MDSR was conducted and its outcomes were widely used, or were amenable, to placing blame on individuals and institutions involved in the process. Some studies mentioned practical experiences that justified the perception or fear of the respondents; however, the reasons were not clarified in many others.

The experience of blame took various forms. For example, the MDSR resulted in or was associated with court cases, including instances of criminal and civil litigations (WHO, 2014; Melberg et al., 2019; Mukinda et al., 2021), cases of work suspension (Kinney et al., 2020; Mukinda et al., 2021) and referral to human resource offices for a likely disciplinary action (Shittu and Kinney, 2017; Mukinda et al., 2021). Other forms of blame included oral or written warnings (Kinney et al., 2020; Said et al., 2021); other unspecified punishments or punitive actions (Abouchadi et al., 2013; Kouanda et al., 2022; Russell et al., 2022); an external investigation (Melberg et al., 2019); having to explain to higher officials, often politicians (Abouchadi et al., 2013; Kouanda et al., 2022); being called out or reprimanded during death review or the risk of being considered unskilled by one’s colleagues (Abouchadi et al., 2013; Armstrong et al., 2014; Shittu and Kinney, 2017); and generally feeling threatened (Kinney et al., 2021).

Health workers were reported as the actors most often blamed in relation to the MDSR (WHO, 2016a; Abebe et al., 2017; Lardi et al., 2019; Kinney et al., 2020; Abouchadi et al., 2022). Among the category of health workers, low-level or front-line health workers and junior doctors seemed to be subjected to blame more often than others (Scott and Dairo, 2015; Singh et al., 2015; Melberg et al., 2019; Naik et al., 2020; Cahyanti et al., 2021). A study from Ethiopia reported that while doctors were concerned about formal legal accusations, lower-level health workers were worried more about informal sanctions (Melberg et al., 2020). Apart from health workers, health institutions (Cahyanti et al., 2021; Boyi et al., 2022), traditional birth attendants (WHO, 2014; Russell et al., 2022), family members of the deceased, community members and the deceased women themselves were also subjected to blame of some form (Scott and Dairo, 2015; Melberg et al., 2019; Willcox et al., 2023).

The studies also revealed factors reinforcing the sense of blame and insecurity associated with the MDSR process. Several studies reported, for instance, that the increase in malpractice cases had such an effect (Shellhaas and Conrey, 2018; Melberg et al., 2019; 2020). Community members in Ethiopia also hesitated to report community maternal deaths owing to the heightened emphasis placed on institutional delivery (Abebe et al., 2017). Other contributing factors included a strong political will to reduce maternal deaths (Abebe et al., 2017; Melberg et al., 2019; Kouanda et al., 2022), hostile media coverage of such deaths (Abouchadi et al., 2022; Kouanda et al., 2022) and lack of understanding about the objectives of the MDSR (Armstrong et al., 2014; Kinney et al., 2021; Russell et al., 2022).

Effects of blame are also commonly discussed in the studies. Accordingly, the perception that the MDSR was punitive or an understanding that actors involved were not adequately protected led to intentional non-reporting or under-reporting and in some cases selective reporting of maternal deaths (Singh et al., 2015; Mutsigiri-Murewanhema et al., 2017; Melberg et al., 2019; Kouanda et al., 2022; Yameogo et al., 2022). It is also observed that the fear associated with the MDSR particularly resulted in the non- or under-reporting of abortion-related maternal deaths in settings where abortion is criminalized or socially stigmatized (WHO, 2014; Singh et al., 2015; Naik et al., 2020; Russell et al., 2022; Yameogo et al., 2022). Other effects identified included discouraging health workers from participating in review meetings or when they did, hesitance to freely speak (WHO, 2014; Said et al., 2021; Boyi et al., 2022; Russell et al., 2022) and intentionally misclassifying, falsifying or misattributing causes of death (WHO, 2016a; Abebe et al., 2017; Melberg et al., 2019; Willcox et al., 2023).

Moreover, several studies indicated that fear of legal consequences associated with the MDSR directly affected access to and the quality of healthcare. Among others, health workers and health facilities were found to unnecessarily refer severely ill women. In other cases, pregnant women were refused admission and treatment due to the fear of being accused in case of maternal death (Singh et al., 2015; Melberg et al., 2020; Abouchadi et al., 2022; Russell et al., 2022). Similarly, it was reported that increased vulnerability to accusations of death deterred junior doctors from pursuing obstetrics as a career (Melberg et al., 2020; Willcox et al., 2023).

To mitigate the manifestations and effects of blame in the context of the MDSR, most studies provided certain recommendations. Among the most frequently reiterated suggestions were providing training, assuring health workers about the ‘no name, no blame’ principle and orienting health workers and managers about the objectives of the MDSR, emphasizing the importance of adhering to the non-blame principle (WHO, 2014; Mutsigiri-Murewanhema et al., 2017; Smith et al., 2017a; Compaore et al., 2022; Yameogo et al., 2022). A considerable number of studies also called for legal protection, including the adoption of appropriate laws as a potential solution, even though these calls were frequently general in their framing (Abebe et al., 2017; Shittu and Kinney, 2017; Kodan et al., 2021; Willcox et al., 2023).

This theme describes the extent to which the MDSR files and documents are safeguarded against unintended use and particularly if and how they are insulated from legal and administrative processes, as recommended by the WHO Technical Guidance (WHO, 2013) and relevant literature. We found limited papers discussing this aspect.

From the relevant studies, there were a general understanding and consequent concern that the MDSR process could be used in legal or disciplinary proceedings (Scott and Dairo, 2015; Abebe et al., 2017; Melberg et al., 2019; Cahyanti et al., 2021; Yameogo et al., 2022). Among others, MDSR-related documents were used in legal proceedings in Tanzania (Said et al., 2021). Local authorities also reportedly used such files to attribute fault to healthcare professionals and facilities or evaluate the performance of facility managers (Melberg et al., 2019; Abouchadi et al., 2022). Moreover, MDSR files and findings were deliberately referred to the human resources department of health facilities. They were then utilized to discipline personnel suspected of failures revealed by the process (Cahyanti et al., 2021; Said et al., 2021). Apart from the deliberate use of the MDSR for punitive purposes, several studies also documented the challenge of separating the MDSR from other punitive and disciplinary processes (Armstrong et al., 2014; Abouchadi et al., 2022; Russell et al., 2022; Willcox et al., 2023). As such, the need to ensure separation, including by law, was emphasized (Shittu and Kinney, 2017; Lardi et al., 2019). While a few studies indicated that the MDSR is insulated from other processes by law, for instance, in South Africa (Smith et al., 2017a), other studies reported that in some countries, it is prohibited to use the MDSR in legal or disciplinary measures, without clarifying how this is enforced (Willcox et al., 2023).

Under this theme, we present findings related to the extent to which laws, serving as regulatory and facilitative tools for the MDSR or its components, were discussed in the reviewed studies.

The scoping review highlights the limited attention given to the legal dimension of the MDSR, in both coverage and the language used. For instance, even though the MDSR involves the exercise of public power, including the authorization of state organs to collect and access health information, this aspect was rarely raised. Consequently, the studies generally depict the MDSR as a mere technical tool that does not require a legal basis to exist. Similarly, while discussions about the safety of data abound in the studies, no attempt is made to articulate them in terms of the right to privacy, the closest legal expression of the interests discussed. Furthermore, legal tools such as the right to access to information or court subpoena of witnesses, which are common themes in the public health surveillance legal literature (Hodge, 2003; Neslund, 2007; Neslund et al., 2010; Gostin and Wiley, 2016), are not mentioned in the studies. A similar observation can be made regarding mandatory reporting, one of the widely reported themes in the review. The studies do not discuss the extent to which the relevant laws on mandatory reporting fit the purposes and fast-paced process of the MDSR.

Law exerts a powerful influence on health by structuring, perpetuating and mediating risk factors, underlying conditions and facilitating interventions (Burris et al., 2010; Gostin et al., 2019). This foundational role of law in shaping public health outcomes underscores the significance of legal frameworks in supporting health systems That said, the limited attention paid to laws and the legal dimension of the MDSR in the studies can be attributed to various factors. To begin with, the role of law in governing public health or framing issues in public health as legal has traditionally been nominal (Fox, 2001; Institute of Medicine, Board on Population, Health, 2011; Gostin et al., 2019). This is partly explained by a tendency among health professionals, who mostly oversee public health programmes, to prefer technical and more flexible approaches to governing public health work and resist formal and often power-restraining laws (London, 2008; Gostin and Wiley, 2016). Furthermore, laws represent just one of many regulatory tools available to health administrators. We also posit that the predominance of researchers with medical or health science backgrounds in the studies included for this review has likely played a role in the limited attention given to legal issues. It also appears that legal scholars have not been engaged in studying the MDSR. Future studies on MDSR implementation may benefit if they pay increased attention to the legal issues that affect its functioning. In the same vein, it is imperative to include legal competence in both research and MDSR supervision teams.

The review demonstrates that data used in MDSR contain information that identifies individuals. In many cases, collected data seem to be sufficient to potentially identify the deceased, relatives and health workers who provided care, among others. Even though mitigating measures such as anonymizing identifiable data were in some instances required, they appear to be insufficient. To begin with, these requirements come late in the MDSR process (Scott and Dairo, 2015; Smith et al., 2017a; Abouchadi et al., 2022). Moreover, it is unclear whether anonymization, when implemented, is mandated by law and whether specific anonymity standards are provided in the respective countries. Consequently, insufficient anonymity remains a major challenge for the MDSR, feeding and reinforcing the fear of unwarranted blame, mainly among health workers (Armstrong et al., 2014; Kinney et al., 2020; Abouchadi et al., 2022; Willcox et al., 2023).

Informational privacy is at stake here. As a right recognized under international and national laws, informational privacy guarantees rights holders to control the acquisition, use or disclosure of their identifiable data. As such, obtaining consent from the right holder is generally necessary to legitimize the collection, use, transfer or other forms of processing their data (Bygrave, 1998; Neslund, 2007; Witzleb, 2014; Ball, 2017; Johnson, 2018; Mariner et al., 2019). Like many other rights, however, the right provides exceptions under which personal data can be processed without the consent of the right holder under such conditions as legality, necessity and proportionality. Briefly stated, a law authorizing the exception is necessary. Besides, it should be demonstrated that the data are essential for the intended purpose and that the impact on the right holder is proportionate to the benefits gained. It is also part of the right to have personal data effectively de-identified once it achieves the purpose for which it was given or acquired (Bygrave, 1998; Hodge, 2003; Witzleb, 2014; Ball, 2017; Mariner et al., 2019; Schneider, 2019).

Closely related to informational privacy is the protection of confidentiality. Often considered a subset of the right to privacy, confidentiality represents the safeguarding, primarily from disclosure, of information shared within relationships of trust and confidence. While informational privacy grants individuals the right to restrict access to information that is legally recognized as private or personal, its counterpart, confidentiality, emphasises the responsibility of those entrusted with confidential information not to disclose it to others or use it in a manner that contradicts its intended purpose (Lowrance, 2012; Gostin and Wiley, 2016; Hulkower et al., 2020).

Public health is a legitimate ground for which an exception is made under the conditions outlined. As such, for personal data to be used without the informed consent of the right holder, this needs to be mandated by law, and the necessity and proportionality requirements should be met. In the context of the MDSR, this implies that a national law should mandate it, specifying the manner in which identifiable data can be processed. Efforts to improve MDSR implementation need to give due attention to these requirements, as a matter of legal obligation. The availability of identifiable data concerning health workers in files utilized for death review, along with the confidential nature of the MDSR process, underscores the critical need to ensure privacy and uphold strict confidentiality, in compliance with legal obligations.

This is also justified on instrumental grounds. It is argued that people cooperate better in providing health data, which forms the backbone of public health works, if they trust that their data are collected, used and shared responsibly (Hodge, 2003; Gostin and Wiley, 2016). This can arguably be extended to health workers and their participation in the MDSR. By providing rules and procedures for weighing competing interests, laws on privacy rights and their limitations have the potential to improve trust in the system. This aligns with the growing understanding that health programmes are more effectively implemented when the rights and values they implicate are duly taken into account and carefully balanced (Gostin and Mann, 1994; Bygrave, 1998; Toebes, 2015; Annas and Mariner, 2016). From the review, it is evident that the mere inclusion of principles in technical guidance are insufficient for ensuring data privacy and process confidentiality within the MDSR framework. Given that privacy is a well-established interest and extensive jurisprudence exists to balance rights in a manner that advances public health, implementing countries are advised to consider legislative measures to overcome these inadequacies and support the effective implementation of the health programme.

The review documents health workers’ continued worries about the MDSR system (WHO, 2016a; Shittu and Kinney, 2017; Melberg et al., 2019; Boyi et al., 2022; Willcox et al., 2023). Factors that reinforce the fear of blame around the MDSR range from heightened political importance attached to mortality numbers in global health to the lack of sufficient legal protection that the objectives and confidential nature of the MDSR necessitate (Abebe et al., 2017; Melberg et al., 2020; Kouanda et al., 2022). The politicization of health data and its potential to compromise the integrity of health systems and accountability mechanisms are well documented (Adams, 2016; Østebø et al., 2018; Blouin, 2020; Davis, 2020). The review also highlighted that the fear of blame not only hinders MDSR implementation but also may lead to additional negative health outcomes. For instance, there are indications that some health workers, feeling insecure about the requirement to register and review maternal deaths, refer severely ill women to other facilities unnecessarily or decline to provide care.

To support regulatory efforts to improve the MDSR, we recommend a broad classification of the effects of blame on MDSR into two categories: legal and non-legal. Associations of the MDSR with litigations (Melberg et al., 2019; Kinney et al., 2021; Mukinda et al., 2021), suspension from work (Kinney et al., 2020; Mukinda et al., 2021), other disciplinary actions (Shittu and Kinney, 2017; Mukinda et al., 2021) and oral or written warnings (Shittu and Kinney, 2017; Said et al., 2021) seem to stem from inadequate legal frameworks protecting health workers in relation to their involvement in the MDSR. These outcomes often involve or seem to involve legal institutions and pertain to rights and interests frequently protected by laws, such as liberty, job security or employment benefits, which broadly qualifies them as legal.

Legislative initiatives seem to have the potential to mitigate these challenges. While the particular legal design adopted depends on the legal system in question, laws such as the laws of evidence, procedure and labour laws deserve particular attention, considering their strong bearing on the legal forms of blame. Examining the extent to which existing laws address the challenges that the MDSR is facing should form part of the solution. This is necessary because merely declaring the separation of the MDSR from legal or administrative measures, expressing a political commitment to ensure privacy and confidentiality or adhering to the ‘no blame’ approach is unlikely to have a legal effect when concrete legal cases arise. Public health law studies document that the quality of such ‘incidental laws’ substantially affects the efficacy of health programmes (Burris et al., 2010; Gostin and Wiley, 2016). Related to this, it is also important to strengthen dispute settlement and complaint-handling mechanisms in the health system. Without robust procedures, interested parties tend to resort to using the MDSR as a forum for litigation and blame (Melberg et al., 2020; Said et al., 2021).

The studies also report other forms of blame that are less directly related to legal processes or institutions, falling broadly under the ‘non-legal’ category. For instance, health workers are reprimanded or are made to feel that they are judged as unskilled during death reviews (Armstrong et al., 2014; Shittu and Kinney, 2017; Melberg et al., 2019) or that after the MDSR, they need to explain results to higher officials, often politicians (Abouchadi et al., 2022; Kouanda et al., 2022). Clarifying the objectives and importance of separating the MDSR from punitive and individual evaluation systems through training and appropriate reminders appears to be an important way to deal with these challenges. Persuasion and public awareness are considered important tools for implementing public health policies (Gostin and Wiley, 2016). Integrating legal solutions with non-legal interventions, based on the nature of the blame and the legal system involved, can potentially enhance MDSR implementation.

In connection with the fear of blame, the language of accountability recurs in the studies reviewed. In this regard, we observed diverse and at times conflicting usages of the notion. Generally, accountability in the context of the MDSR is portrayed as both an ally and a foe and is discussed at two main layers: the health system level and the individual level.

The studies typically portray a positive correlation between MDSR and accountability when the focus is on the health system. As a value or culture in the health system within which the MDSR functions, it is reported that accountability supports the MDSR or is produced or enhanced by it. Here, accountability relates to transparency and better coordination in the health system, its willingness and ability to assess actions and learn, explain the circumstance of death to bereaved families (Willcox et al., 2023) and most importantly act based on recommendations and ensure effective follow-up (Compaore et al., 2022; Kouanda et al., 2022; Willcox et al., 2023). Overall, the MDSR serves as an accountability mechanism—a tool designed to enhance maternal health through monitoring, reviewing and responding.

On the other hand, when used in relation to individual actors in the MDSR, accountability mostly represented blameworthiness, punishment or other negative consequences, resulting from some form of failure. Accountability thus embodies something workers want to avoid and as such a challenge to MDSR implementation (Shittu and Kinney, 2017; Melberg et al., 2019; Said et al., 2021). The MDSR is thus seen in this context as a forum and vehicle to identify those responsible for maternal death and punish their actions.

The dual use of the term ‘accountability’ is understandable, given its complexity. Accountability is a layered concept involving transparency, taking responsibility, identifying those responsible and sanctioning inappropriate action (Schedler et al., 1999; Brinkerhoff, 2004).

This conflicting understanding may be exemplified by the challenging roles of directors or facility managers in the MDSR. While many of the studies report that chief medical directors chair MDSR reviews (WHO, 2014; Shittu and Kinney, 2017; Mukinda et al., 2021), this position is not always appreciated. While the involvement of such officers in death reviews is considered crucial to conducting orderly meetings and translating recommendations into action, their responsibility to ensure quality health services delivery and their ability to take administrative measures against employees cause tensions in meetings (Said et al., 2021). Both roles are often described as accountability. A similar dilemma is observed concerning the involvement of lay members of the community in reviews—while their participation in the process helps with feedback, it also heightens suspicion among health workers to prevent them from speaking freely in the presence of ‘outsiders’ (WHO, 2014; Ayele et al., 2019).

Some observations can be drawn from the discussion on accountability. Notably, the distinction underscores that the primary purpose of the MDSR is to identify and address failures within the healthcare system. Moreover, utilizing the MDSR for punitive actions or disciplining individual implementers is counterproductive. It is evident that maintaining a completely blame-free process is challenging, and the argument is not against individual responsibility for illegal or unprofessional conduct. Rather, it emphasizes that the MDSR is not a suitable platform for regulating healthcare workers. This necessitates a careful delineation of the process from punitive measures, without implying that professional misconduct should be tolerated.

Additional data underlying this review can be shared upon reasonable request to the corresponding author.