- Research article
- Open Access
- Open Peer Review
Maternal death reviews at Bugando hospital north-western Tanzania: a 2008–2012 retrospective analysis
BMC Pregnancy and Childbirthvolume 15, Article number: 333 (2015)
Unacceptably high levels of maternal deaths still occur in many sub-Saharan countries and the health systems may not favour effective use of lessons from maternal death reviews to improve maternal survival. We report results from the analysis of data from maternal death reviews at Bugando Medical Centre north-western Tanzania in the period 2008–2012 and highlight the process, challenges and how the analysis provided a better understanding of maternal deaths.
Retrospective analysis using maternal death review data and extraction of missing information from patients’ files. Analysis was done in STATA statistical package into frequencies and means ± SD and median with 95 % CI for categorical and numerical data respectively.
There were 80 deaths; mean age of the deceased 27.1 ± 6.2 years and a median hospital stay of 11.0 days [95 % CI 11.0–15.3]. Most deaths were from direct obstetric causes (90); 60 % from eclampsia, severe pre-eclampsia, sepsis, abortion and anaesthetic complications. Information on ANC attendance was recorded in 36.2 % of the forms and gestation age of the pregnancy resulting into the death in 23.8 %. Sixty one deaths (76.3 %) occurred after delivery. The mode of delivery, place of delivery and delivery assistant were recorded in 44 (72.1), 38 (62.3) and 23 (37.7 %) respectively.
Routine maternal death reviews in this setting do not involve comprehensive documentation of all relevant information, including actions taken to address some identified systemic weaknesses. Periodic analysis of available data may allow better understanding of vital information to improve the quality of maternity care.
Maternal mortality remains high in most sub-Saharan African countries, including Tanzania.[1, 2] Even among women who seek care in health facilities for pregnancy-related complications early enough, case fatality rates may be high , indicating sub-optimal care and the imperative of improving the quality of care for better maternal survival.
The maternal mortality ratio in Tanzania is 454 deaths per 100,000  and levels have not declined significantly in the past two decades to meet the 5th millennium development goal by 2015.  Direct obstetric causes of maternal deaths account for at least three quarters of these deaths [1, 3], and although life-threatening pregnancy-related complications may start at home for some women, most come into contact with the health system before they die . Sub-optimal care is also common in health facilities.  Utilization of essential services for maternal and newborn survival such as skilled delivery attendance and emergency obstetric care is not always possible even among women living within 5 km of health facilities. [3, 6].
Of the at least 7900 maternal and 39,000 newborn deaths expected to occur in Tanzania in 2013 [2, 7], a significant number could have been saved with better quality of care in health facilities as critical skills, equipment, drugs and other essential commodities for saving maternal and newborn deaths often lack. [8–10] Nevertheless, lessons learnt from managing individual cases may be equally important for improving the quality of care women receive in health facilities. For example, gaps identified in a near miss or maternal deaths reviews can inform providers on how to avoid similar future deaths. One strategy advocated to address the latter is health facility maternal death review [11, 12]-defined as a qualitative, in-depth investigation of the causes of and circumstances surrounding maternal deaths occurring at health facilities .
In Tanzania, guidelines for maternal and perinatal death reviews (MPDRs) were produced and rolled out in 2006. The guidelines aim to improve the process of maternal death reviews with an ultimate goal of improving case management for better maternal and newborn survival.  They also require a review team in each health facility (membership eligibility stipulated) which is supposed to meet within 24 h of the death occurring. Typically, a review session for a maternal death at a health facility takes approximately one to two hours and the process involves presentation of a case summary by someone who was not involved in the case management, identification of gaps and strengths in case management and drafting of an action plan with timelines and specific roles to the identified levels of care to act on the recommendation (s) in order to prevent similar deaths in future. Review committees are also required at all other levels of health administration in the country (district, region and national) to meet at stipulated times. But available evidence suggests that except for health facility and district meetings, regional level meeting are not always convened and no national meeting has ever been convened.  Review reports from health facilities are subsequently forwarded to the district where reports are also sent to the region and to the responsible section within the Ministry of Health and Social Welfare.
Whilst most deaths in health facilities are possibly reviewed and reported, implementation of these guidelines has not yielded much evidence to support that the maternal death reviews (MDRs), although conducted routinely, provide results that are acted upon to improve the quality of maternal and newborn health services. Maternal and perinatal death forms are filled routinely and sent to various levels as stipulated in the guidelines, but evidence of critical analysis of whether results from such reviews are communicated back, or even facility level recommendations are acted upon is scarce.  Furthermore, emerging evidence from epidemiological studies in the country suggests that despite significant improvement in utilization of health facilities among women with life-threatening pregnancy-related complication in recent years , maternal deaths have not decreased proportionately [1, 2, 4]; suggesting sub-optimal quality of care. MDRs aim to provide lessons to prevent similar subsequent maternal death events in health facilities, but if women continue to die in health facilities for causes that are identified routinely, doubts are abound for the effectiveness of the current MDR strategy to address the high level of such deaths. Furthermore, the country is currently reviewing the MPDR guidelines to align with the global demand for surveillance and response as the better way of understanding and responding to the levels, causes and trends of maternal mortality. [15–17].
We report the results from the analysis of data from maternal death reviews at Bugando Medical Centre 2008–2012, a tertiary hospital in Mwanza Tanzania and discuss how lessons from the results have helped to institute measures to improve care provided to women with maternal complications in the institution. The results from the analysis were also used to inform an on-going review of the existing MPDR guidelines in the country.
The study is a retrospective analysis of maternal death reviews at a tertiary hospital north-western Tanzania in the period 2008–2012. This is a 900 bed capacity and the main referral hospital for the population in the Lake and western zones constituting almost 37 % of the entire Tanzania mainland population, estimated at 43,625,354 people in 2012.  There are about 7000 deliveries at this hospital annually, most being referrals from lower level health facilities.
All copies of the reviewed maternal deaths covering the study period were accessed (80 deaths). Typically, maternal deaths reviews are conducted as they occur, usually within a week of occurring by a team led by senior obstetricians in a morning meeting. Case summaries of deaths are prepared by resident doctors in obstetrics and gynaecology not involved in the case management (usually one such doctor is assigned a case to summarize and present). In contrast to what is stipulated in the national guidelines on who should be involved in the reviews, the circumstance at the hospital demands teaching of both undergraduate medical and nursing students who are invited in such meeting as part of learning the process. Senior midwives in the department are also involved in the review besides clinicians and midwives involved in the case management. However, anaesthetists and pharmacists are rarely invited to participate.
Once the review has been done, copies of the review findings are sent to the hospital Director and the zonal reproductive and child health office housed at the hospital. The latter is responsible for forwarding the filled forms to various levels of reporting; usually to the district reproductive and child health coordinator who also forwards the information to the regional and national offices quarterly.
We reviewed multiple data sources at the hospital for deaths that occurred in this period: labour ward, medical records, theatre registers and emergency and casualty department. Missing information for two deaths not included in the existing maternal deaths review forms were identified. Corresponding review forms were found in the deceased patients’ files and included in the analysis.
Information on the forms were entered into an excel spreadsheet and missing information was extracted from case notes in patients’ files if available. Data were then transferred into Stata statistical package in windows version 12 (StataCorp LP College Station, Texas, USA) for cleaning, coding and analysis. Categorical data were analysed into frequencies to understand the various components of review that were completed and filled and causes of deaths and associated factors while numerical data into means ± SD, range as well as median with 95 % CI.
The analysis is part of a larger study titled “Improving quality of maternal and newborn care in Tanzanian health facilities: Lessons from a mixed method assessment of health facility maternal deaths reviews in the Lake zone of north-western Tanzania” with research permits from the Catholic University of Health and Allied Sciences (CUHAS)/Bugando Medical Centre and National Institute for Medical Research (NIMR) ethical committees (no CREC/018/2013 & NIMR/HQ/R.8a/Vol.IX/1543 respectively). Permission for the retrospective access to the deceased patients’ hospital records was also requested and obtained from the administration of Bugando Medical Centre, including the head of the department of obstetrics and gynaecology as required by the national regulations. Data extraction was done by one of the authors who was not involved in the management of any of the patients and de-linked the deceased patients’ information to ensure data anonymity in the subsequent steps of data handling.
Table 1 shows some characteristics of the maternal deaths. Most were young with the mean age of 27.1 ± 6.2 years (range 15–41 years). The majority were aged 20–35 years (80 %), had unknown level of education (57.5 %) and resided in Mwanza region (83.5 %). The median age was 26 years (95 % CI 25–29) (information not shown in the table). Almost 74 % were housewives or peasant farmers and married (88.8 %).
Table 2 shows additional characteristics of the deceased women. Few were five times or more pregnant (11.3 %). On average, a patient stayed at Bugando hospital for almost two weeks before death (mean 13.5 ± 5.7 days) and median 11.0 [95 % CI 11.0–15.3] (not shown in the table) and just 20 % stayed for less than one day.
In most cases, the review forms were incompletely filled. Of the total 80 reviewed deaths, only 29 (36.2 %) had information on whether the woman attended ANC and 19 (23.8 %) had information on gestation age of the pregnancy resulting into the death. Sixty one deaths (76.3 %) occurred after the women had delivered and of these the mode of delivery was recorded in 44 (72.1), 38 (62.3 %) had place of delivery recorded, 29 (47.5 %) had the delivery assistant recorded and only 23 (37.7 %) had gestation age at delivery recorded. In addition, only 39 (62.9 %) had the outcome of delivery recorded; 25 (41 %) with live birth, 11 (18.0) fresh stillbirth and the remaining three (4.9 %) macerated (information not shown in the tables).
Half of the total deaths (40 out of 80) had information on whether a post-mortem was performed or not. Likewise, filling in the causes of death as required by the forms (both direct and indirect causes) were not always done. Of the total 80 deaths, information was filled for only 72 (90); 58 (72.5 %) with direct causes highlighted but 16.3 % of the deaths although identified as such, specific causes were not filled in the review forms. Eclampsia and severe pre-eclampsia accounted for 26.2 % of all deaths while sepsis and postpartum haemorrhage accounted for 17.5 and 11.3 % of all deaths respectively. Both abortion and anaesthetic complications accounted for one in ten of all deaths. The filled indirect causes of deaths did not provide an in-depth understanding of the individual causes but 5 % of the deaths were attributed to indirect causes with AIDS accounting for 3.8 % of all deaths. Of the total 80 deaths, 68 (85 %) had various factors associated with the deaths identified and recorded on the review forms. Late referral from another health facility, delay in receiving appropriate care at the referring health facility and at Bugando Medical Centre and delay in deciding to seek care were the commonest factors identified.
Furthermore, complete data to allow analysis of all deaths on whether they occurred before, during or after delivery as well as on actions that were suggested during the review meetings and possibly taken thereafter were missing, an indication of the quality of the meetings.
The study highlights some issues with maternal deaths review at a tertiary health facility in Tanzania. Notably, maternal death reviews are conducted but relevant information is not always recorded as per available guidelines. The missing information includes planned actions to prevent similar deaths in future, thus defeating the purpose of the reviews. Nevertheless, the retrospective analysis allowed for identification of such bottlenecks and institution of appropriate measures such as review of the management for eclampsia and severe pre-eclampsia patients to include admission to a high dependence unit for intensive monitoring and care and review of skills and experience of hospital anaesthetists to ensure that operating theatres coverage is always with competent and experienced people to prevent anaesthetic deaths. Maternal deaths review forms are also reviewed regularly after initial reviews to ensure that they are filled appropriately before they are submitted to appropriate levels of the health system.
The finding of incompletely filled review forms did not provide enough information to understand all circumstances surrounding the deaths as lessons gained would not lead to better understanding of the management provided to women and identification of dysfunctions to address in order to avoid similar deaths in future in the absence of case notes from the deceased patients’ files. Two main reasons were identified to explain this finding. In some cases, important information was not recorded at the referring health facility and the situation did not allow obtaining such details after the woman had died (an indication of poor referral details); but in some, extraction of the information from the patient records into the review forms was incomplete. Unfortunately, regular review of all filled forms lacked thus not allowing identification of such weaknesses to recommend appropriate and timely remedies. Although measures to avoid missing information on various actions decided during the reviews and how they are followed-up when they are filled has been addressed at Bugando Medical Centre, similar weaknesses likely occur in other facilities in the country. We recommend that the on-going-review of the national maternal and perinatal deaths review guidelines recognize this as a weakness and include measures to avoid it. Also quarterly, semi-annualy or annually reports may involve few cases for an in-depth analysis and planned analysis after some years may be a more pragmatic option to consider, especially is hospitals with limited caseloads.
The finding that approximately 80 % of all deaths occurred when the deceased had stayed in the hospital for at least a day and almost a third stayed for at least four days highlights the quality of care women receive at this hospital and in other facilities referring such women. Measures to improve the quality of care at health facilities in the lake zone are imperative for better maternal survival. The on-going Lake Zone Maternal Mortality Reduction Initiative by Bugando Medical Centre with over 35 district and regional hospitals aiming at improving the quality of care in these health facilities is a welcome measure. Results from a one year experience of implementing this intervention are encouraging and will be reported separately. Additionally, country level analysis of the coverage, equity and quality of key reproductive, maternal, newborn and under-five children indicators shows that the lake and western zones are the least performing in the country impacting negatively on the overall country MDG4&5 progress.  Subsequently, pertinent issues contributing to the sub-optimal progress are being addressed, including but not limited to ensuring availability of essential medicines, equipment and other supplies; recruiting, deploying and retaining of skilled providers; improving community engagement in service availability and utilization and improving basic infrastructure for effective referral through increased regional and council health budgets and the Big Result Now Initiative. The latter is a special presidential initiative to accelerate progress in key national sectors, including health.
Of equal note are the relatively low documented AIDS-related deaths in this hospital. Recent estimate suggest that such deaths may account for almost 6 % of all deaths in the country.  Effective PMTCT services in this part of the country may partly explain this although errors in ascertaining causes of deaths cannot be ruled out. However, similar to previous studies in the country, direct causes of deaths accounted for most deaths [1, 3].
The finding that late referral of patients from health facilities occurred in almost a third of all deaths indicates that improving the quality of care in the referring facilities is imperative for better maternal survival in this setting. Training of available staffs in these facilities for essential competences in management of pregnancy-related complications and quick referral where required is a short term option but good supportive supervision is vital to reinforce the acquired skills and to identify and address challenges in appropriate skills and within the health system. Ensuring that health facilities have the necessary supplies and equipment for managing maternal complications is another necessary intervention.
The main strength of this assessment is to highlight the causes of maternal deaths and related factors which informed subsequent steps to improve on the management of women with obstetric and related complications in the hospital. As a result, periodic analysis of the filled maternal and perinatal deaths review forms has been included in the hospital quality of care assessment and other measures instituted as required by the current national maternal and perinatal death review guidelines and plan for accelerated reduction of maternal, newborn and child deaths [13, 21].
Some limitations of this study are worth mentioning. Collection and analysis of retrospective data posed some challenges. Notably, the quality of data could not be controlled in retrospect but very much depended on the previous practice and mechanism of ensuring good quality data. Additionally, like what has have been reported elsewhere in Africa [22–25], we encountered incompletely filled MDR forms, errors we had no control over. The forms were also of different review cycles and possibly done at different times of staff competences and as a result, consistency in case management might have lacked. Routine facility MDRs at Bugando Hospital and indeed in the entire lake zone and countrywide do not involve collecting in-depth community information on factors associated with maternal deaths but highlights of what might have happened before the woman reached the health facility. Subsequently, available data may not allow understanding of factors that might have prevailed before contact with the health system thus unable to comprehensively address challenges women face in accessing available services.
We conclude that maternal deaths reviews and documentation of information from such reviews at Bugando Medical Centre in the period 2008–2012 were not always comprehensive enough to provide all the necessary information required in the current national guidelines and to understand the quality of care women received before demise. This is currently addressed through improving providers understanding of the essence of the reviews, the review process as well as documentation of information from the reviews, including proposed actions to prevent similar deaths in future. The roll-out of the new national guidelines should also aim to build strong capacity for tertiary institutions like Bugando, so that they function as resource centres for training skilled professionals in maternal and perinatal death reviews.
Countdown to 2015: Fulfilling the Health Agenda for Women and Children: The 2014 report. Available at http://www.countdown2015mnch.org/documents/2014Report/Countdown_to_2015-Fulfilling%20the%20Health_Agenda_for_Women_and_Children-The_2014_Report-Conference_Draft.pdf; 2014.
World Health Organization: Trends in maternal mortality: 1990 to 2010. WHO, UNICEF, UNFPA and The World Bank estimates; World Health Organization, Geneva Switzerland; 2012.
Hanson C. The epidemiology of maternal mortality in Southern Tanzania. PhD thesis. London School of Hygiene and Tropical Medicine. London UK: University of London; 2012.
National Bureau of Statistics, ICF Macro: Tanzania Demographic and Health Survey 2010. NBS & ICF Macro. Dar es salaam, Tanzania; 2011.
Mbaruku G, Van Roosmalen J, Kimondo I, Bilango F, Bergström S. Perinatal audit using the 3-delays model in western Tanzania. Int J Gynaecol Obstet. 2006;106(1):85–8.
Magoma MTN: The effectiveness of antenatal birth plans in increasing skilled care at delivery and after delivery in rural Tanzania: A cluster randomized trial. PhD thesis. Department of Epidemiology and Population Health, Infectious Diseases Epidemiology Unit. London School Of Hygiene and Tropical Medicine, University of London. London: UK; 2010.
United Nations Children’s Fund: Levels and Trends in Child Mortality. Report 2013. Estimates Developed by the UN Inter-agency Group for Child Mortality Estimation. Unicef, World Health Organization, The World Bank & United Nations. United Nations Children’s Fund, New York, USA; 2013.
Jhpiego: Quality of Care for Prevention and Management of Common Maternal and Newborn Complications: A survey of the quality of maternal and newborn care in 12 regions in Tanzania. Report 1: Findings on Antenatal Care. Maternal and Child Health Integrated Program [MCHIP].Plotkin M; Tibaijuka G; Makene CL; Currie S; Lacoste M. Jhpiego, Washington, D.C. USA; 2011.
Jhpiego: Quality of Care for Prevention and Management of Common Maternal and Newborn Complications: A study of 12 regions in Tanzania. Report 2: Findings on Labour, Delivery and Newborn Care. Maternal and Child Health Integrated Program [MCHIP]. Plotkin M; Tibaijuka G; Makene CL; Currie S; Lacoste M. Jhpiego, Washington, D.C. USA; 2011.
Ministry of Health and Social Welfare. Service availability and readiness assessment (SARA) 2012. Dar es salaam: Ifakara Health Institute; 2013.
World Health Organization: Beyond the numbers. Reviewing maternal deaths and complications to make pregnacy safer. WHO Geneva. Available at http://www.who.int/maternal_child_adolescent/documents/9241591838/en/ 2004.
World Health Organization: Maternal death surveillance and response: technical guidance information for action to prevent maternal death, Available at http://www.who.int/maternal_child_adolescent/documents/maternal_death_surveillance/en/. accessed 18/07/2014; 2013.
Ministry of Health and Social Welfare. Maternal and perinatal death reviews guide. Ministry of Health and Social Welfare. Reproductive Health Section. Dar es salaam, Tanzania: United Republic of Tanzania; 2006.
Armstrong CE, Lange IL, Magoma M, Ferla C, Filippi V, Ronsmans C. Strengths and weaknesses in the implementation of maternal and perinatal death reviews in Tanzania:perceptions, processes and practices. Tropical Medicine and International Health. 2014;19(9):1087–95.
Hounton S, De Bernis L, Hussein J, Graham WJ, Danel I, Byass P, et al. Towards elimination of maternal deaths: Maternal Deaths Surveillance and Response. Reproductive Health. 2013;10:1.
Danel I, Graham WJ, Boerma T. Maternal death surveillance and response. Bull World Health Organ. 2011;89:779–779A.
PMNCH: Implementing Maternal Death Surveillance and Response (MDSR). Asia-Pacific Leadership and Policy Dialogue for Women and Children. Available at http://www.who.int/pmnch/media/news/2012/accountability_implement_mdsr.pdf; 2012.
National Bureau of Statistics, Office of the Government Chief Statistician: 2012 Population and Housing Census: Population Distribution by administrative Areas. National Bureau of Statistics. Ministry of Finance Dar es salaam and the Office of the Chief Government Statistician. President’s Office, Finance, Economy and Development Planning Zanzibar. Dar es Salaam, Tanzania; 2013.
Ministry of Health and Social Welfare. The National Road Map Strategic Plan To Accelerate Reduction of Maternal, Newborn and Child Deaths in Tanzania 2008–2015. Dar es salaam: Sharpened One Plan; 2014.
World Health Organization: Trends in maternal mortality: 1990 to 2013 Estimates by WHO, UNICEF, UNFPA, The World Bank and the United Nations Population Division. World Health Organization, Geneva Switzerland; 2014.
Ministry of Health and Social Welfare: The National Road Map Strategic Plan To Accelerate Reduction of Maternal, Newborn and Child Deaths in Tanzania 2008–2015. Ministry of Health and Social Welfare, Dar es Salaam. United Republic of Tanzania; 2008.
Kongnyuy EJ, Mlava G, Van Den Broek N. Facility-based maternal death review in three districts in the central region of Malawi: an analysis of causes and characteristics of maternal deaths. Womens Health Issues. 2009;19(1):14–20.
Kongnyuy EJ, Van Den Broek N. The difficulties of conducting maternal death reviews in Malawi. BMC Pregnancy Childbirth. 2008;8:42.
Vink NM, De Jonge HC, Ter Haar R, Chizimba EM, Stekelenburg J. Maternal death reviews at a rural hospital in Malawi. Int J Gynaecol Obstet. 2013;120(1):74–7.
Okong P, Byamugisha J, Mirembe F, Byaruhanga R, Bergstrom S. Audit of severe maternal morbidity in Uganda--implications for quality of obstetric care. Acta Obstet Gynecol Scand. 2006;85(7):797–804.
Data analysis and report writing was supported by Evidence for Action (E4A) project Tanzania, a maternal and newborn health project funded by UKaid through OPTIONS Consultancy LTD UK. The analysis is part of the E4A efforts to understand maternal death reviews in Tanzania and contribute evidence to inform the on-going review of the national maternal and perinatal deaths review guidelines. We offer special thanks to the administration of Bugando Medical Centre for permission to use the hospital data and colleagues in the department of Obstetrics and Gynaecology for providing clarifications on some issues. Our sincere appreciation to all hospital staffs for assisting variously: labour ward, operating theatres, mortuary and medical records.
We declare no competing interest.
MM developed the research concept, analysed the data and developed the initial manuscript. AM collected the data and participated in the analysis and writing of the manuscript. CM, RM, AK, BG reviewed the extracted data, contributed to the analysis and writing of the manuscript. All authors approved the final manuscript.