SEARCH

SEARCH BY CITATION

Keywords:

  • maternal mortality;
  • risk factors;
  • maternal mortality ratio;
  • verbal autopsy

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Summary

objective  To determine the maternal mortality rate in a rural district of Tanzania and to measure the incidence of causes of maternal mortality, the presence of risk factors and the relationship with social and demographic factors.

method  From January to December 1993 a retrospective recording of maternal deaths was completed using verbal autopsy and networking.

results  A total of 76 deaths were found which is equivalent to a maternal mortality ratio of 961 per 100,000 live births for this 12-month period of time. The leading causes of death were postpartum haemorrhage with retained placenta, anaemia, postpartum haemorrhage without retained placenta, AIDS complex and obstructed labour (in descending order of frequency). Maternal deaths were seen irrespective of group factors such as access to a main road, presence of antenatal risk factors and contact with health care personnel or a nearby facility before death. Mortality was also present both in home and hospital deliveries (excluding hospital referrals). Antenatal care had been received by 97.2% of the mothers who died after the second trimester. The referral rate even in the presence of a known antenatal risk factor was 34.6%. Patient compliance to the referral was only 44.4%. Mothers and their families followed strong cultural beliefs even when they were detrimental to the mother's health. Maternal deaths were proportionately higher among women >40 who were also gravid geqslant R: gt-or-equal, slanted 5, but there was no significant increase in deaths in women < 19 years of age.

conclusion Effective antenatal care, appropriate emergency treatment of complications, access to transportation and competent referral level care with adequate equipment encompass the most effective answers to reduction of maternal deaths at a district level.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Maternal mortality is a major problem in East Africa. Its impact is most commonly quoted in terms of the maternal mortality ratio. On a multinational level in 1987, East Africa had a MMR (maternal mortality ratio) of 660 per 100,00 live births (Herz & Mersham 1987). In the same year, Tanzania reported a national MMR of 190 per 100,000 live births (MCH Unit, Ministry of Health, Tanzania). In 1990 estimates of maternal mortality revised by WHO and UNICEF (1996) from East Africa were found to be increasing. For Tanzania, the report estimated a MMR of 750 per 100,000 live births. Typically the maternal mortality ratio has been hospital/institution-acquired. More recent research has been attempting to overcome that particular limitation of the MMR data. In a Tanzanian village in 1990, Moller and colleagues prospectively followed 719 pregnancies and reported four maternal deaths consistent with a MMR of 600 per 100, 00 live births. In contrast, Walraven et al. (1994) in rural communities in Northwestern Tanzania, derived a MMR of 241 per 100,000 live births from a prospective community-based survey of a cohort of pregnant women.

The community-based surveys should be more reflective of the true magnitude of the maternal mortality in developing countries, where the vast majority of women deliver outside formal health care settings (Ngallaba et al. (1993). This study, conducted between January 1 and December 31, 1993, was intended to evaluate the level of maternal mortality in a coastal district of Tanzania, to measure the incidence of the causes of maternal mortality and to identify risk factors. The distribution of deaths according to various socio-demographic factors and the pattern of usage of the existing health care system (i.e. dispensaries/health centers/referral centers and their respective staff) were also studied.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The study was conducted in the district of Bagamoyo in the Coast region of Tanzania. For 1993, the projected population was 192,963 with 30,874 women estimated to be of childbearing age (projected fro the 1988 National Census). The major source of income is subsistence farming and fishing. The major religion is Islam. For females the illiteracy rate is reported as 35%; only 49.5% have completed primary education (grade 7). The district has 31 dispensaries, 4 health centers and one district hospital. The distribution of the population is 95% rural.

A retrospective area-based recording of maternal deaths for the 12-month period from January 1 to December 31, 1993 was completed by one of the authors who worked in the district of Bagamoyo as a clinical medical officer. Recording was done in two steps: first, contact was made with villages, dispensaries or health centers through visits and informal meetings. During these initial contacts, interested individuals were given the task of collecting the most exhaustive and complete lists of maternal deaths for their area. These individuals were a heterogenous group of people from village government leaders to village health workers from local ?respected women? to any cadre of health professional. By the second visit, the informants (those who had been designated on the first visit) had compiled a list of maternal deaths for their area. The author and the informant went to the home/village of the deceased woman or her family. To encourage reporting by the local people, all maternal deaths reported on the second visit were followed up, although only those falling in the above period of time were included in the final data analyses. The author, a known doctor in the district, interviewed a family member or an informed individual with a detailed questionnaire which elicited sociodemographic information and a verbal autopsy. The verbal autopsy elicited the signs and symptoms of the illness and/or difficulties which occurred leading up to the mother's death and directly extended to an obvious medical reason for her death. Any attending health personnel were interviewed for collaborating data to clarify and validate the medical diagnosis.

All deaths followed the definition for maternal mortality set forward by the World Health Organization and the International Federation of Gynaecology and Obstetrics (WHO 1993). The definition is the death of a woman while pregnant or within 42 days of termination of pregnancy, irrespective of the duration and the site of pregnancy, from any cause related to or aggravated by the pregnancy or its management. All data was analyzed by the statistical software program EpiInfo 5.01.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

There were 76 reported maternal deaths for the specified 12-month period of time. The maternal mortality ratio was 961 per 100,000 live births using the projected number of live births from the 1988 national census.

All causes of death are listed in Table 1. Three deaths were due to induced abortion and one death to a spontaneous abortion, accounting for 5.3% of all deaths. The underlying direct cause was sepsis in all of these cases. In the ten maternal deaths caused by anaemia, the symptoms prior to death were a constellation of severe weakness, pallor, shortness of breath and perpheral oedema without a history of haemorrhage, fever or other possible attributable causes. For any woman whose history included haemorrhage (antenatal or postpartum), even if she exhibited the above noted symptoms of severe anaemia, death was attributed to the haemorrhage or its underlying cause (i.e. placenta praevia). There were symptoms of anaemia in 45.3% (34/76) of the deaths regardless of cause.

Table 1.  The causes of maternal deaths in Bagamoyo district: January to December 1993 Thumbnail image of

As is shown in Figure 1, 18.4% (14/76) of the women were 15 to 19 years of age and another 15.8% (12/76) 35 years or older. However, the majority, 65.8% (50/76), of the women who died were 20 to 34 years of age. Looking at the number of deaths according to gravidity, as shown in Figure 2, 34% (33/76) of the gravid ?5 women died. The next highest group were primigravida with 22.4% (17/76) of the deaths. However, 43.4% (33/76) of women of gravid 2-4 died.

image

Figure 1. Number of deaths and maternal mortality ratio (MMR) for a 12-month period calculated for each 5-year-age-group. ▪ Total number of deaths, □ MMR (× 100) per 100,000 live births.

Download figure to PowerPoint

image

Figure 2. Number of deaths for a 12-month period according to gravidity of the pregnant woman.

Download figure to PowerPoint

Overall, 77.6% (59/76) of the deaths occurred at home. Table 2 shows the relationship between place of childbirth/abortion versus place of death. Seven of the 76 women neither aborted nor gave birth before their death and are therefore excluded from this table. Only 58.0% (40/69) of the women gave birth/aborted at home and then died at home. Of these women, 45.0% (18/40) had contact with a health worker or health facility before their death. Of the deaths occurring at home, 23.0% (12/52) had given birth/aborted at a dispensary, health centre or district hospital before later dying at home. Of the 48 women who had a home birth/abortion, 8 (16.7%) later died in a health facility. Excluding the deaths caused by HIV positive/ARC syndrome, 64.2% (43/67) of all the women had contact with a health worker before their death. Another 46.3% (31/67) had an outpatient visit or a hospital admission and discharge for a related problem before their death.

Table 2.  Number of deaths as a function of place of death versus place of birth or abortion. Note that the total is 69 not 76 because 7 women died without either aborting or giving birth Thumbnail image of

There was a health facility near the full-time place of residence of 64.5% (49/76) of the women, but 22.4% (11/76) of these died in a village without a healthy facility because they had moved there during their pregnancy for various social and cultural reasons. The family members were asked to give their personal opinion on why their relative died or what she died from. The answers can be classified into three main categories: answers involving sorcery and witchcraft were given by 30.3% (23/76) of the relatives interviewed, an act of God or God's will was given by 36.8% (28/76), and only 32.9% (25/76) gave a medical reason for the cause of death of the pregnant mother.

There were 71 maternal deaths in the second trimester, third trimester or postpartum. 15.5% (11/71) of these mothers' babies were stillborn and another 23.9% (17/71) died in utero with their mothers. This means that 60.6% (43/71) were live births, but only 55.8% (24/43) were still alive at the time of the interview. Only 33.8% (24/71) of the children of the pregnant mothers who reached at least the second trimester before their death are still alive and well.

In this study 97.2% (69/71) of the women who died after reaching at least the second trimester had received antenatal care. All health facilities in Tanzania use a standard antenatal card put out by the ministry of health in conjunction with UNICEF. This card has a check list of various factors considered to be predictive risk factors of complications during pregnancy or delivery. All MCH aides are taught to look for these factors and if present to note them on the antenatal card and refer to the next level of care. 76.4% of the women who received antenatal care were noted to have at least one or more risk factors. But only 34.6% (18/52) of them were referred and 44.4% (8/18) followed through and went to the referral center. Looking at three of the top five causes of death, 68.4% (13/19) of the postpartum haemorrhage (PPH) deaths, 66.7% (2/3) of the ruptured uteri deaths and 100% (7/7) of the obstructed labor deaths had at least one of these recognizable risk factors (see Table 3).

Table 3.  Three treatable/preventable causes of death: postpartum haemorrhage (RPH), ruptured uterus and obstructed labour analyzed according to a selected group of risk factors Thumbnail image of

14.5% (11/76) of the deaths occurred in the intrapartum period and 71.1% (54/76) in the postpartum period. On the first postpartum day, 85% (17/20) of the deaths were due to postpartum haemorrhage with or without retained placenta. Two to four weeks postpartum, anaemia and puerperal sepsis claimed 41.1% (14/34) of the deaths. 72.7% (32/44) of the women who gave birth at home were delivered by TBAs, 25.0% (11/44) by a relative and only one woman had no assistance. All TBAs had attended the training sessions held by the MCH aid at their local health facility.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

There were few written records such as death certificates (3/76), hospital records (2/76) or antenatal cards (4/76) available for data collection. Therefore, the verbal autopsy was a useful tool to obtain the desired information in this study. There will be error due to poor recall because of the retrospective questioning. The majority of the interviewees were nonmedical family members and therefore the inability to recognize or recall pertinent medical symptoms will make answers less accurate. Also, because the author was dealing with the death of an individual, most interviewees were very concerned about issues of blame and the possible negative consequences of their answers. Usually the reassurance of the interviewer overcame this last concern. Overall the community was the best source of information on maternal deaths. In 1989, Boerma and Mati found this as well when they conducted a village survey in coastal Kenya using the system of networking to derive a MMR or 6-7 deaths per 1,000 live births. TBAs and health care workers at all levels from the dispensary up to the district hospital were not seen to be useful in this capacity likely due to the fear of blame and/or failure. Over 50% of the deaths were reported on more than one occasion from separate sources. This afforded an opportunity to fill in separate questionnaires and autopsy reports and compare the results for validity. Also, wherever possible the verbal autopsy and questionnaire was completed in its final form after interviewing more than one individual who had contact with the mother before her death.

The maternal mortality ratio (MMR) of 961 per 100,000 live births is one of the highest reported MMRs for Tanzania. Numerous papers reporting MMRs in Tanzania since 1974 show a range from a low of 180 to a high of 730 per 100,000 live births but these data are mainly institution-based (Government of Tanzania and United Nations Children's Fund 1990; Walraven et al. 1994). More recently, a large scale adult mortality study conducted in the rural districts of Morogoro and Hai showed a MMR of 748 and 230 (per 100,000 live births) respectively1.

The major causes of mortality found in this study are similar to findings of other mortality studies in developing countries (Martey et al. 1994; Toure et al. 1992; Moller et al. 1989; Kwast et al. 1986). This study did demonstrate a lower prevalence rate (5.3%) of abortion as a cause of maternal death. Mhango et al. (1986) and Ashworth (1990) in an urban setting reported rates from 25 to 50% of the total number of maternal deaths due to abortion. Kwast et al. (1986) concluded that this difference in prevalence was a reflection of rural versus urban populations.

The literature supports the concept of increased risk of death in younger (<19 years of age) and older (> 35 years of age) pregnant women (Mlay et al. 1994; Safe Motherhood Task Force 1992; Mhango et al. 1986; Guenther & Brandrup-Lukanow 1993; Okonofua et al. 1992; Alauddin 1986). The largest number of deaths in the district of Bagamoyo were seen in women in the age groups of 20-24 followed by 25-29 years of age (see Figure 1). It is possible that the larger number of deaths in these two age groups is a factor of more women delivering at those ages. Therefore comparing the MMR rather than the absolute number of deaths would take this factor into account as the MMR is based on the number of live births. The live birth estimations for each age group were calculated using data from a cohort of the pregnant women in one ward of the Bagamoyo district for the same period of time (see Table 4)2. The Lugoba ward in Bagamoyo is representative of the typical rural population in this district. A bias may exist as a result of inaccurate recall of age by the women interviewed. The resultant maternal mortality ratios based on each age group were recalculated and are shown in Figure 1. The MMR is similar across the different age groups except for women 40-45 years of age, who had a substantially higher MMR. Of note is that all the women in the ?40 age group were also gravid 5 or higher pregnancies. Therefore concerns about older women may in reality be a reflection of the high gravidity and not their age at all. These results suggest that age has no causal relationship with maternal death in the district of Bagamoyo and age as an independent risk factor is questionable.

Table 4.  Derivation of the maternal mortality ratios shown in Figure 1: the number of live births for a cohort of women from one ward of Bagamoyo during a 12-month period was used Thumbnail image of

Various socioeconomic factors are often cited as risk factors for maternal mortality (Guenther & Brandrup-Lukanow 1993). Common social problems seen in Tanzania such as unwed mothers and different fathers for each child were examined. Figure 3 shows that these social problems were not associated with higher levels of maternal deaths. Thonneau et al. (1992) demonstrated that only low family income is a risk factor for maternal death. In summary, sociodemographic variables, most importantly poverty, are related to the various MMRs globally, but they play no role in the analysis of the cause and effect relationships of maternal death for individual mothers. All causes of death reported in this study were treatable and/or preventable medical conditions. Therefore at district level the focus must be on medical interventions to affect the causes of maternal death.

image

Figure 3. Number of deaths in a 12-month period according to a, marital status; b, number of marriages; c, number of fathers of her children.

Download figure to PowerPoint

At the first level one must improve antenatal care (Mlay et al. 1994). We found that women attend antenatal care but that problems are missed and inappropriate advice is given if problems are noticed. For example, failure to detect transverse lie or to refer it leads to obstructed labour or ruptured uterus. Moodley et al. (1994) showed that an improved antenatal care decreased maternal mortality due to eclampsia despite an increasing incidence of eclampsia in the same time period. Ekstrom et al. (1996) showed that giving iron supplementation starting in the second trimester can improve haemoglobin concentrations. Early detection and treatment through properly trained and equipped health personnel in the antenatal clinic setting would definitely impact outcomes of women with anaemia, malpresentation (obstructed labour/ruptured uterus), malaria, eclampsia, tuberculosis and placenta praevia.

The second step involves improving emergency obstetric services. We found that most women had contact with a health facility or health personnel before their death, and that a significant number of women died on a main road or in a village with a dispensary. The emergency services need to improve from the first contact with a rural health personnel up to trained physicians at the referral centre. At the most basic level, the ability to stabilize and transport to a nearby centre that can deal with obstetric emergencies is essential (Safe Motherhood Task Force 1992). Greenwood et al. (1990) showed a reduction in the maternal mortality rate in a rural setting in the Gambia which they concluded was partly due to improved transportation. This requires that basic equipment and training be available at the first line of the health care system and that at this level there is access to transport. Lamb et al. (1984), also in rural Gambia, reported no maternal deaths during an eight-year period in their study area. They attributed this to an on-the-spot, 24 h-fully available qualified midwife or physician who was attached to another project in the same area. Okonofua et al. (1992) in Nigeria concluded that training of health care personnel should emphasize first-aid manoeuvres and use of medication to treat postpartum haemorrhage and retained placenta. This training should be extended to TBAs and serious consideration given to even supplying a local, centrally-located, trained TBA with oral ergometrine, as 18/19 (94.7%) of the PPH deaths were home births. Maternity waiting homes may be another alternative for rural communities to overcome the lack of basic essential services in their community or the delay in treatment of complications due to lack of adequate transportation (Chandramohan et al. 1994). Figa-Talamanca (1994) advocates the use of maternity waiting homes but suggests that their success depends on good antenatal care.

Lastly, the quality of decision-making and competence of referral-level physicians, the expedience with which their care is available and the level of equipment and medicine available to them needs to be markedly improved. Lunan (1996) discusses the lack of appropriate training for physicians to deal with the obstetric emergencies of the developing world. Martey et al. (1995) discuss a community-based postgraduate programme for obstetrics and gynaecology in West Africa, which is overcoming the previous more traditional approaches to training. Sundari (1992) reports that lack of know-how and even faulty patient management at the referral hospital level play a significant role in treatable causes of maternal death. Having a viable alternative may motivate health workers at the basic levels to detect and refer high risk cases early and motivate mothers to follow through with the referral.

There were a number of cultural values and beliefs which played a role in the medical attention-seeking behavior of the women and their relatives. One common situation involved women who returned to their mother's home or the maternal side of the family for delivery regardless of the type of obstetrical care they required. Only 32.9% of the relatives gave a medical reason for the cause of death. Despite beliefs of sorcery or God's will, a large number of women had contact with the health service after ‘becoming ill’ or recognizing some sort of problem. Thaddeus et al. (1994) demonstrated that the quality of care that a woman received influenced her decision to seek medical care more than the important considerations of cost or of distance away from the care. If the effectiveness of the health care system were improved at all three points, antenatal care, basic first line emergency treatment including stabilization with safe transportation, and referral level care, the compliance with referral and treatment would likely improve simultaneously.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The study was funded by Deutsche Gesellschaft für Technische Zusammenarbeit, Eschborn, Germany.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • 1
    Alauddin, M (1986) Maternal mortality in rural Bangladesh: the Tangail District. Studies in Family Planning 17, 1321.
  • 2
    Ashworth, MF (1990) Harare Hospital maternal mortality report for 1987 and a comparison with previous reports. Central African Journal of Medicine 36, 20912.
  • 3
    Boerma, JT Mati, JK (1989) Identifying maternal mortality through networking: results from coastal Kenya. Studies in Family Planning 20, 24553.
  • 4
    Ekstrom, EC Kavishe, FP Habicht, JP Frongillo, EAJr, Rasmussen, KM Hemed, L (1996) Adherence to iron supplementation during pregnancy in Tanzania: determinants and hematologic consequences. American Journal of Clinical Nutrition 64, 36874.
  • 5
    Figa-Talamanca, I (1996) Maternal mortality and the problem of accessibility to obstetric care; the strategy of maternity waiting homes. Social Science of Medicine 42, 138190.
  • 6
    Government of the United Republic of Tanzania and United Nations Children's Fund (1990). Women and Children in Tanzania: A Situation Analysis. Dar Es Salaam.
  • 7
    Greenwood, AM Bradley, AK Byass, P, et al (1990) Evaluation of a primary health care programme in The Gambia; I: The impact of trained traditional birth attendants on the outcome of pregnancy. Journal of Tropical Medicine and Hygiene 93, 5966.
  • 8
    Guenther D & Brandrup-Lukanow A (1993) Maternal Health and Family Planning. GTZ Health, Population and Nutrition Division and WHO Collaboration Centre for Health Systems Development. Macmillan Press Ltd.
  • 9
    Herz B & Measham AR (1987) The safe motherhood initiative: proposals for action. Background paper prepared for the ?Safe Motherhood Conference?. Nairobi, 1-13 February.
  • 10
    Kwast, BE Rochat, RW Kidane-Mariam, W (1986) Maternal Mortality in Addis Adaba, Ethiopia. Studies in Family Planning 17, 28899.
  • 11
    Lamb, WH Foord, FA Lamb, CM Whitehead, RG (1984) Changes in maternal and child mortality rates in three isolated Gambian villages over ten years. Lancet 20, 9124.
  • 12
    Lunan, CB (1996) Obstetrics and gynaecology in the developing world. British Journal of Obstetrics and Gynaecology 103, 4913.
  • 13
    MCH Unit, Ministry of Health 1987. Tanzania Government Document.
  • 14
    Mhango, C Rochart, R Arkutu, A (1986) Reproductive mortality in Lusaka, Zambia: 1982-1983. Studies in Family Planning 17, 24351.
  • 15
    Moller, B Lushino, O Kabukoba, J, et al (1989) A prospective area-based study of the outcome of pregnancy in rural Tanzania. Uppsala Journal of Medical Science 94, 1019.
  • 16
    Moodley, J Daya, P (1994) Eclampsia: a continuing problem in developing countries. International Journal of Gynaecology and Obstetrics 44, 914.
  • 17
    Martey, JO Dian, JO Twum, S Browne, En Opoku, SA (1994) Maternal Mortality and related factors in Ejisu District, Ghana. East African Medical Journal 71, 65660.
  • 18
    Mlay, R Massawe, S Lindmark, G Nystrom, L (1994) Recognition of risk factors in pregnancy among women attending antenatal clinic at Mbagala, Dar Es Salaam. East African Medical Journal 71, 56266.
  • 19
    Ngallaba S, Kapiga SH, Ruyobya I & Boerma JT (1993) Tanzania Demographic and Health Survey 1991/1992. Bureau of Statistics, Planning Commission, Dar Es Salaam.
  • 20
    Okonofua Fe, Abejide, A Makanjuola, RA (1992) Maternal mortality in Ife-Ife, Nigeria: a study of risk factors. Studies in Family Planning 23, 31924.
  • 21
    Safe Motherhood Task Force with the Ministry of Health and Family Care International, March 1992. Safe Motherhood Strategy for Tanzania.
  • 22
    Sundari, TK (1992) The untold story: how the health care systems in developing countries contribute to maternal mortality. International Journal of Health Services 22, 51328.
  • 23
    Tanzania Sensa (1988). 1988 Population Census: Preliminary Report. Bureau of Statistics, Ministry of Finance, Economic Affairs and Planning, Dar Es Salaam.
  • 24
    Thaddeus, S Maine, D (1994) Too far to walk: maternal mortality in context. Social Science and Medicine 38, 10911100.
  • 25
    Thonneau, P Toure, B Cantrelle, P Barry, TM Papiernik, E (1992) Risk factors for maternal mortality: results of a case-control study inconducted in Conakry (Guinea). International Journal of Gynaecology and Obstetrics 39, 8792.
  • 26
    Toure, B Thonneau, P Cantrelle, P Barry, TM Ngo-Khac, T Papiernik, E (1992) Level and cause of maternal mortality in Guinea (West Africa). International Journal of Gynaecology and Obstetrics 37, 8995.
  • 27
    Walraven, GE Mkanje, RJ Van Roosmalen, J Van Donegen, PW Dolmans, WM (1994) Assessment of maternal mortality in Tanzania. British Journal of Obstetric and Gynaecology 101, 414417.
  • 28
    WHO (1993) International Classification of Disease. 9th revision. Women's Groups, NGOs & Motherhood Maternal Health & Safe Motherhood Programme, Division of Family Health, WHO, Geneva.