Background Since 2015, Tanzania has been implementing the Maternal Death Surveillance and Response (MDSR) system. The system employs interactions of health providers and managers to identify, notify and review maternal deaths and recommend strategies for preventing further deaths. We aimed to analyse perceptions and experiences of health providers and managers in implementing the MDSR system. Methods An exploratory qualitative study was carried out with 30 purposively selected health providers and 30 health managers in four councils from the Mtwara region between June and July 2020. Key informant interviews and focus group discussions were used to collect data. Inductive thematic analysis was used to analyse data. Results Two main themes emerged from this study: ‘Accomplishing by ambitions’ and ‘A flawed system’. The themes suggest that health providers and managers have a strong desire to make the MDSR system work by making deliberate efforts to implement it. They reported working hard to timely notify, review death and implement action plans from meetings. Health providers and managers reported that MDSR has produced changes in care provision such as behavioural changes towards maternal care, increased accountability and policy changes. The system was however flawed by lack of training, organisational problems, poor coordination with other reporting and quality improvements systems, assigning blame and lack of motivation. Conclusion The implementation of the MDSR system in Tanzania faces systemic, contextual and individual challenges. However, our results indicate that health providers and managers are willing and committed to improve service delivery to avoid maternal deaths. Empowering health providers and managers by training and addressing the flaws will improve the system and quality of care.
An exploratory qualitative study design with focus group discussions (FGDs) and key informant interviews (KIIs) at regional, district and facility levels was applied. FGDs were conducted among health providers in the wards who were members of MDSR committees while KIIs focused on health managers experiences. These methods were used to makes sure that rich data were collected from both health providers and managers. There are many MDSR committee members with different experiences on MDSR depending on their cadre, experience in maternal care and participation in the system’s activities. FGDs were used among these health providers in order to explore their perceptions and experiences with MDSR, however, little they had. Key informants interviews were used among health managers in order to explore their experiences more deeply since they have been implementing the system for a longer time than most health providers. Another reason was that logistically it would have been difficult to organise FGDs with managers. The study was conducted in four districts in the Mtwara region of Southern Tanzania, where the two districts with the highest number of maternal deaths in the year 2018 (Masasi District Council and Mtwara Town Council) and the two districts with the lowest number (Nanyamba and Tandahimba District Councils) were selected for the study. All facilities, public and private (dispensaries, health centres, district and regional hospitals) that conduct delivery services are required to implement the MDSR system in Tanzania. The dispensary is at the lowest level, which serves local residents in villages/streets and refers patients to health centres which serve local wards. The health centres refer patients to district hospitals, which in turn refer patients to regional hospitals. All levels of facilities are capable of providing antenatal care and delivery services. Caesarean sections are performed in all hospitals but only a few health centres. Mtwara has a total population of about 1.2 million people according to the 2012 census.4 The region has a well-established MDSR system in all its’ districts since 2015. These facilities have multidisciplinary MDSR committees consisting of obstetricians, medical doctors, clinicians, nurses, midwives, laboratory personnel, pharmacy staff, drivers and other supporting staff, which reviews all maternal deaths. The size and structure of the committees depends on the level of the facility, and the number and availability of staff. The medical officer in-charge is the chairperson of the hospital MDSR committee. When a death occurs at the facility, notification is sent to the district and regional Reproductive Child Health Coordinators. The facility MDSR committee reviews the death within 7 days and fills the maternal death reporting form which is sent to the district and regional health office together with a narrative summary and action plans put forward by the committee. The region and each district have quarterly MDSR meetings to discuss maternal deaths happening in the quarter and make regional and district recommendations. These are implemented and reports are sent to the Ministry of Health, Community Development, Gender, Elderly and Children (MoHCDGEC) (figure 1). Key informants were selected from one Regional Health Management Team (RHMT), two District Executive Directors’ (DED) offices, four Council Health Management Teams (CHMT) and four health facilities. Participants were purposively sampled on the basis of their position in the health system, as well as their participation in MDSR activities for at least 6 months. Furthermore, snowball sampling was used to include more key informants from other facilities. In total, 30 key informants from the DEDs office, RHMT, CHMTs and health facilities were included (table 1). Demographic characteristics of the participants (N=60) Others (teacher, lawyer, health secretary and social welfare). *Includes medical doctors and assistant medical officers. FGD, focus group discussion; MDSR, maternal death surveillance and response. A total of four FGDs with 6–8 participants were conducted; one from each selected health facility in the four districts. These included members of facility MDSR committees who were not health managers, such as clinicians and nurses from maternity/labour wards, pharmacy staff and laboratory staff. We specifically excluded health managers in the FGDs in order to avoid instances where health providers might not reveal some information due to fear of their managers. Participants were selected by information provided by those in charge at facilities or hospital matron. We worked closely with these managers who identified all members of the committee in the facility on the day of data collection. We used our inclusion criteria to select participants for FGDs. Participants were selected on basis of being an MDSR committee member, having participated in MDSR activities for at least 6 months and were excluded if they were health managers. In total, 30 health providers from MDSR committees participated in the FGDs (table 1). Most of the participants were clinicians, nurses and nurse midwives. Half of them were females, most (28) were aged 31–40 years and most (37) had 6–24 months experience with MDSR activities (table 1). Data collection in each district started with KIIs where health managers from management teams, facility leaders and government officials were interviewed. It was followed by FGDs with facility (hospital/health centre) MDSR committee members. Discussions focused on how the MDSR was initiated in the region/district/facility, the dissemination and role of guidelines, how participants were initiated and introduced to their roles, how notification of deaths and the review process are currently done, and how action plans are formulated and implemented. FGDs also included discussions of barriers and facilitators of the MDSR system, its initiation and current status. Furthermore, issues on community maternal deaths and reviews were explored. The interview guide included open ended questions with probes and focused on the MDSR implementation cycle. All interviews were conducted in Swahili and were audio recorded. KIIs lasted for approximately 45–60 min while FGDs were done for 60–120 min. The data collection process was inspired by Lincoln and Guba following an emergent design where data collection was done concurrently with continuous data analysis.16 17 Preliminary results were shared with the research team in the field after each day’s interviews, and all interviewers reviewed the data together in daily meetings in order to agree on areas to be further explored during upcoming interviews. This influenced the design of the research tool and allowed for purposive sampling of participants depending on the need. Assessment of saturation of data was also performed during the daily interviewers’ meetings. The data collection process was completed after 28 consecutive days. The study tool was adapted from a qualitative interview guide used in a study done in Uganda by Agaro et al18 with modifications to suit the current context and designed to ensure that all aspects of the MDSR system were explored. Complementary methods (KIIs and FGDs) and purposive sampling of study participants allowed different aspects and different angles to be explored from the health providers in the facilities, health managers and government officials at the DEDs offices. Participants selected were diverse (table 1) and had different professions and different tasks within the MDSR implementation. The interviews were conducted by three interviewers to reduce researcher-induced biases. All interviewers kept field notes that were used during data collection and analysis. The notes included the context description, dates, place and time of data collection, participants’ interaction during FGDs and response to understanding of questions. In each daily meeting the interviewers discussed issues that materialised during the interviews and went through their field notes. Some of the categories were identified and discussed during these meetings. All recorded interviews from KIIs and FGDs were transcribed verbatim for analysis. Thematic analysis as inspired by Braun and Clarke19 was used to develop themes that best described the findings in the data inductively. The first author read through the transcripts a number of times to understand the trend of the data. Themes that were identified during data collection (by AS and NS) in the field notes were also reviewed. Then open coding was done, followed by abstraction of codes to form initial subcategories. All the transcripts were in Swahili language, translation to English was done during coding where the Swahili speaking authors (AS) did the initial coding by reading through the Swahili transcripts and the codes generated were written in English. Microsoft Excel (2007) computer program was used during coding and formulation of categories and themes. The transcripts were written in one column of excel sheet then the code phrases were written in the next corresponding column to facilitate tracking of codes and original transcripts. The English generated codes were shared with other Swahili speaking authors (NS, AP and SM). Then the first (AS) and last author (MM) organised the initial subcategories and merged them to generate categories. The categories were then reviewed and collated to form subthemes, and these were in turn reviewed and abstracted into themes. Preliminary results were presented to the coauthors (NS, CH, SM and AP) for verification and consistency checking. Their comments inspired the final themes and interpretation. This study was part of a lager project, which involved review of documents and interviews with health providers in Mtwara and Lindi regions. We did not involve patients as a study population. During the planning stage of the study, before ethical clearance was sought, the first author visited all districts in the two regions and held meetings with regional, district and facility health managers to get their opinions on how the study should be done, which documents are available and who and what issues needed to be explored. Their opinions were valuable during the planning and proposal preparation of the study. The results of the study will be shared with the health providers and managers in the two regions by oral presentations and a written report. All participants received details about the study and its aims and were then asked to sign an informed consent form before the interview commenced. Confidentiality and secrecy were ensured by not using the names and positions of participants in the health system and conducting the interviews in rooms where no one else had access. The FGD participants were also assured that their findings would not be discussed with their managers in the facility. Use of identifying information was also avoided during report writing to ensure confidentiality. It was explained to participants that their participation or non-participation would not cause problems for themselves or their work. The interviews were audiorecorded with participants’ permission. The data (transcripts and audio recordings) were protected by saving them on the main researcher’s (AS) computer that only he had access to. The field notes were stored under lock and key and were only available to the researchers.
N/A