SEARCH

SEARCH BY CITATION

Keywords:

  • maternal mortality;
  • process indicators;
  • evaluation;
  • developing countries

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Study populations
  6. Maternal mortality
  7. Indicators of access to obstetric care
  8. Statistical methods
  9. Results
  10. Discussion
  11. Acknowledgements
  12. References

Summary Objectives Process evaluation has become the mainstay of safe motherhood evaluation in developing countries, yet the extent to which indicators measuring access to obstetric services at the population level reflect levels of maternal mortality is uncertain. In this study we examine the association between population indicators of access to obstetric care and levels of maternal mortality in urban and rural West Africa.

Methods In this ecological study we used data on maternal mortality and access to obstetric services from two population-based studies conducted in 16 sites in eight West African countries: the Maternal Mortality and Obstetric Care in West Africa (MAMOCWA) study in rural Sénégal, Guinea-Bissau and The Gambia and the Morbidité Maternelle en Afrique de l'Ouest (MOMA) study in urban Burkina Faso, Côte d'Ivoire, Mali, Mauritanie, Niger and Sénégal.

Results In rural areas, maternal mortality, excluding early pregnancy deaths, was 601 per 100 000 live births, compared with 241 per 100 000 for urban areas [RR = 2.49 (CI 1.77–3.59)]. In urban areas, the vast majority of births took place in a health facility (83%) or with a skilled provider (69%), while 80% of the rural women gave birth at home without any skilled care. There was a relatively close link between levels of maternal mortality and the percentage of births with a skilled attendant (r = −0.65), in hospital (r = −0.54) or with a Caesarean section (r = −0.59), with marked clustering in urban and rural areas. Within urban or rural areas, none of the process indicators were associated with maternal mortality.

Conclusion Despite the limitations of this ecological study, there can be little doubt that the huge rural–urban differences in maternal mortality are due, at least in part, to differential access to high quality maternity care. Whether any of the indicators examined here will by themselves be good enough as a proxy for maternal mortality is doubtful however, as more than half of the variation in mortality remained unexplained by any one of them.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Study populations
  6. Maternal mortality
  7. Indicators of access to obstetric care
  8. Statistical methods
  9. Results
  10. Discussion
  11. Acknowledgements
  12. References

Monitoring progress towards improved maternal health has proved to be a challenge, in particular because the preferred health outcomes are difficult to estimate reliably. The measurement of maternal mortality is thought to be too costly and insufficiently precise to be of use in assessing the effectiveness of safe motherhood interventions (Graham et al. 1996). Maternal morbidity has been proposed as an alternative measure of health outcome, but research has shown that the reliability of estimates of obstetric morbidity may be poor, certainly when based on women's recall of their birth experience (Filippi et al. 2000). Process evaluation has thus become the mainstay of safe motherhood evaluation in developing countries (Maine et al. 1997; UNICEF, WHO & UNFPA 1997; Ronsmans et al. 2002).

Process evaluation will adequately inform the design or management of maternal health care strategies if the processes of care that are being measured are those that are critical for improving maternal health so that levels and trends in maternal mortality can be inferred from their measurement. Strategies ensuring that all pregnant women have access to a health professional for delivery, for example, are thought to be critical for maternal mortality reduction, and the proportion of births with a skilled attendant has now become a widely promoted indicator for monitoring progress towards maternal mortality reductions (World Health Organization 1997; AbouZahr & Wardlaw 2001). Similarly, ensuring access to specialized obstetric care is deemed essential for the reduction of maternal mortality, and many developing countries are now adopting the target of a minimum of 15% of births to take place in an essential obstetric care (EOC) facility or 5% of births with a Caesarean section (Maine et al. 1997; UNICEF, WHO & UNFPA 1997; Ronsmans et al. 2002).

Rigorous evidence linking levels of maternal mortality with specific programme inputs is scant however, and the extent to which population-based indicators of access to obstetric services reflect levels of maternal mortality is uncertain (AbouZahr & Wardlaw 2001; Graham et al. 2001; Graham 2002). Renewed calls have been made to measure maternal mortality and process indicators concurrently so that the link between process and outcomes can be firmly established (Graham 2002). Research in this area is hampered by the lack of good maternal mortality data. In this study we examine the association between indicators of access to obstetric care and levels of maternal mortality using data from 16 surveillance sites in eight West African countries. The study sites are unique in that special efforts were made to assess the levels and causes of maternal mortality, while data on place of birth, delivery attendant and mode of delivery are available for the same populations.

Study populations

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Study populations
  6. Maternal mortality
  7. Indicators of access to obstetric care
  8. Statistical methods
  9. Results
  10. Discussion
  11. Acknowledgements
  12. References

We used data on maternal mortality and access to obstetric services from two population-based studies conducted in 16 sites in eight West African countries: the Maternal Mortality and Obstetric Care in West Africa (MAMOCWA) study in rural Sénégal, Guinea-Bissau and The Gambia and the Morbidité Maternelle en Afrique de l'Ouest (MOMA) study in urban Burkina Faso, Côte d'Ivoire, Mali, Mauritanie, Niger and Sénégal. Detailed methods have been reported elsewhere and are only briefly discussed here (Bouvier-Colle et al. 1998, 2001; Walraven et al. 2000; Høj 2002; Kodio et al. 2002).

The MAMOCWA study covered nine rural areas which were under demographic surveillance: one region in The Gambia with a population of 16 000 (Farafenni), three regions in Sénégal (Niakhar, population 30 000; Bandafassi, population 10 000; and Mlomp, population 8000) and five regions in Guinea-Bissau (Oio, Biombo, Cacheu, Gabu and Bafata)(Walraven et al. 2000; Høj 2002; Kodio et al. 2002). In Guinea-Bissau, the population in each site consisted of a random sample of 2000 women of reproductive age. In general, basic obstetric care is provided in health centres staffed by nurses or midwives, although a number of health centres in Sénégal are staffed by traditional birth attendants (TBA). Virtually all births with a skilled attendant take place in a health facility. The median distance to a hospital providing surgery is 30 km, with the exception of Bandafassi where the nearest hospital is 250 km away.

The MOMA study was conducted in five major cities [Abidjan (Côte d'Ivoire), Bamako (Mali), Niamey (Niger), Nouakchott (Mauritanie), Ouagadougou (Burkina Faso) and St Louis (Sénégal)] and two rural towns in the region of Kaolack (Sénégal) (Bouvier-Colle et al. 1998; De Bernis et al. 2000). In each site, all pregnant women were identified through a census and followed until 60 days postpartum by trained investigators. Any morbidity or death during this period was recorded, and place and mode of delivery, and skill level of birth attendant noted. In each urban site, obstetric care is provided in nearby health centres and hospitals staffed by nurses or doctors.

We report data on maternal mortality for the full populations under surveillance between 1993 and 1998 in Farafenni, 1988–97 in Niakhar, 1985–98 in Mlomp, 1984–97 in Bandafassi, 1989–96 for the five sites in Guinea-Bissau and 1994–96 for the MOMA sites.

Maternal mortality

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Study populations
  6. Maternal mortality
  7. Indicators of access to obstetric care
  8. Statistical methods
  9. Results
  10. Discussion
  11. Acknowledgements
  12. References

In the MAMOCWA study, special efforts were made to determine the levels and causes of maternal mortality. Between 1996 and 1999, the relatives of all women who died aged 15–49 were interviewed using a verbal autopsy. In Sénégal and The Gambia three doctors determined the final cause of death, while the Guinea-Bissau investigators used pre-defined diagnostic algorithms (Høj et al. 1999; Walraven et al. 2000; Kodio et al. 2002). In the MOMA study all deaths during pregnancy or up to 60 days postpartum were investigated using information from hospital records, questionnaires with providers and verbal autopsies, and causes assigned by three doctors and two public health researchers (Bouvier-Colle et al. 1998, 2001). Early pregnancy deaths were excluded because women were only recruited after openly acknowledging their pregnancy. In this study, maternal deaths were defined as the death of a woman while pregnant or within 42 days of termination of pregnancy, from any cause except from accidents and intentional injuries. As early pregnancy deaths could not be identified in the MOMA sites we report maternal mortality ratios with and without early pregnancy deaths for the MAMOCWA sites.

Indicators of access to obstetric care

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Study populations
  6. Maternal mortality
  7. Indicators of access to obstetric care
  8. Statistical methods
  9. Results
  10. Discussion
  11. Acknowledgements
  12. References

We used five indicators of access to obstetric services, consisting of the percentage of births in the population with a skilled attendant, in any health facility, in a health centre, in hospital and with a Caesarean section (Maine et al. 1997; UNICEF, WHO & UNFPA 1997; WHO 1997). Skilled health personnel refers to doctors, midwives and nurses, excluding trained and untrained TBA (WHO 1997). Health centre births include births in any facility that is not a hospital, while hospitals refer to facilities that can perform surgery. We chose not to include the UNICEF indicator of the proportion of births in basic and comprehensive EOC facilities because this would have required a detailed inventory of each facility, which was not possible in all the sites at the time of the studies (UNICEF, WHO & UNFPA 1997). Moreover, many hospitals in the study areas did not have access to blood for 24 h, and would thus not have qualified as comprehensive EOC facilities, while still being able to perform life-saving surgery. In rural Guinea-Bissau, none of the hospitals had access to blood transfusion.

We used different sources of data to obtain the place of birth and the type of birth attendant. In the MOMA study, all Guinea-Bissau sites and in Niakhar, such information was recorded prospectively for all the women in the cohort. In the MOMA study, the name and qualification of the birth attendant was recorded in case notes for facility births or in the birth certificate for home births. In Guinea-Bissau and Niakhar, the surveillance noted the place of birth and the birth attendant, relying on the woman's report (Delaunay et al. 2002). In case of uncertainty, we allocated a skill level according to the main providers known to attend births in the health facility named by the woman. In Farafenni, all women who delivered between October 1997 and September 1998 were interviewed after birth about the place of delivery and type of birth attendant (Telfer et al. 2002). In Mlomp and Bandafassi, the proportion of deliveries in a health facility was obtained from the demographic surveillance, but data on the percentage of births in hospital or with a skilled attendant were not available (Enel et al. 1993; Pison et al. 2002a; Pison et al. 2002b). As a proxy for the proportion of deliveries with a skilled attendant, we used data for the district of Oussouye and Kédougou from a national survey conducted in 1999 (Ministère de la Santé, SERDHA et Macro International Inc. 1999).

Population-based Caesarean section rates were obtained directly from the cohorts in the MOMA study, from 1 year of surveillance data in Niakhar (1996–97), and from a random sample of the population in Farafenni (1997–98) (Bouvier-Colle et al. 1998; Dumont et al. 2001; Delaunay et al. 2002; Telfer et al. 2002). In Mlomp and Bandafassi, data were drawn from a study in 1992 in which population-based estimates of Caesarean section rates were computed by region (Bouillin et al. 1994). We used data from the region of Ziguinchor and Tambacounda for Mlomp and Bandafassi, respectively. In Guinea-Bissau, routine records from the regional hospitals in Gabu, Bafata and Cacheu were reviewed between 1990 and 1996 and the number of Caesarean sections noted. Population-based Caesarean section rates were computed by dividing the number of Caesarean sections by the expected number of live births in the community for the same period. Caesarean section rates were not available for Biombo and Oio because cases referred to the capital could not be traced.

Statistical methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Study populations
  6. Maternal mortality
  7. Indicators of access to obstetric care
  8. Statistical methods
  9. Results
  10. Discussion
  11. Acknowledgements
  12. References

Maternal mortality ratios (MMRatios) were expressed as the number of maternal deaths per 100 000 live births and were compared using relative risks (RR) with 95% confidence intervals (CI), assuming Poisson rates. Associations between process indicators and maternal mortality levels (excluding early pregnancy deaths) were assessed using the correlation coefficient, weighted by the number of live births in the population. We also examined the fit of linear and non-linear regression lines (e.g. exponential or quadratic transformations) and present those which yielded the best fit.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Study populations
  6. Maternal mortality
  7. Indicators of access to obstetric care
  8. Statistical methods
  9. Results
  10. Discussion
  11. Acknowledgements
  12. References

The MMRatios by study site, including and excluding early pregnancy deaths, are shown in Table 1. After excluding early pregnancy deaths, the MMRatio ranged from 129 per 100 000 live births in Nouakchott, Mauritanie, to 1267 in Gabu, Guinea-Bissau. Confidence intervals are wide, but maternal mortality was significantly higher in rural than in urban areas. In rural areas, the MMRatio was 628 and 601 per 100 000 live births including and excluding early pregnancy deaths respectively, compared with 241 per 100 000 (excluding early pregnancy deaths) for all urban areas combined [RR = 2.49 (CI 1.77–3.59)].

Table 1.  Maternal mortality in 16 sites in West Africa
Study siteLive births (N)Maternal mortality
CountryRegion/cityPeriodNumber of deathsMaternal deaths per 100 000 live births (95% CI)
AllExcluding early pregnancyAllExcluding early pregnancy
Urban areas
Burkina FasoOuagadougou94–9628318283 (122–575)
Ivory CoastAbidjan94–96303911362 (181–648)
MaliBamako94–9632789275 (125–521)
MauritaniaNouakchott94–9631054129 (35–330)
NigerNiamey94–9618884212 (58–542)
SenegalSt Louis94–9620273148 (30–432)
All urban  16 168 39 241 (172–330)
Rural areas
GambiaFarafenni93–9842451816424 (251–670)377 (215–612)
Guinee BissauCacheu89–9620811414673 (368–1129)673 (368–1129)
Bafata89–9622951715741 (432–1186)654 (366–1078)
Biombo89–9635101815513 (304–810)427 (239–705)
Gabu89–96284137361302 (917–1795)1267 (888–1754)
Oio89–9628982322794 (503–1191)759 (476–1149)
SenegalTambacounda/Bandafassi88–9738733333852 (587–1197)852 (587–1197)
Kaolack/Kaffrine, Fatick94–96152613852 (454–1457)
Ziguinchor/Mlomp85–982292108436 (209–802)349 (151–688)
Fatick/Niakhar84–9716 8668783516 (413–636)492 (392–610)
All rural  42 427257255628 (554–710)601 (529–679)

Indicators of access to obstetric care are shown in Table 2. Births in health facilities generally exceeded those with a skilled attendant in almost all sites, reflecting the fact that not all facilities are staffed by skilled persons and nearly all home births are conducted without a skilled attendant. In urban areas, the vast majority of births took place in a health facility (83%) or with a skilled provider (69%), and more than half of births (55%) took place in a hospital. Very few births (3% or less in all sites) took place with a doctor. This sharply contrasts with the rural areas where 80% of the women give birth at home without any skilled care and only 11% give birth in a hospital. The area of Mlomp in Sénégal stands out as an exception, as nearly all women deliver in a health centre staffed by TBAs who are supervised by a trained nurse-midwife. Population-based Caesarean section rates were extremely low, ranging from 0.2% in Bandafassi to 2.7% in Ougadougou, but are higher in the urban than the rural areas.

Table 2.  Urban and rural differences in maternal mortality and indicators of access to obstetric care in West Africa
Study siteDistance (km) to nearest hospitalPercentage births in population
CountryRegion/cityWith skilled attendantIn any health facilityIn health centreIn hospitalWith Caesarean section
  1. Data from: * Kédougou district (region of Tambacounde); ** Oussouye district (region of Ziguinchor); † region of Tambacounde; ‡ region of Ziguinchor. Values for the percentage hospital and health centre births do not add up to the median values of the percentage health facilities because each value represents the median for each column.

Urban areas
Burkina FasoOuagadougou0–10809426682.7
Ivory CoastAbidjan0–10757419551.5
MaliBamako0–10549161301.6
MauritaniaNouakchott0–10619175160.7
NigerNiamey0–1074732722.1
SenegalSt Louis563571562.5
All urban (median) 698323551.9
Rural areas
GambiaFarafenni10–3219187110.6
Guinee BissauCacheu0–63212110110.5
Bafata0–6321216150.4
Biombo5–6132321220
Gabu0–7118187110.7
Oio0–75111147
SenegalBandafassi25018*1830.2†
Kaolack6029696630.8
Mlomp5067**991.2‡
Niakhar7035171610.4
All rural (median) 20207110.6

There was a relatively close link between levels of service use and maternal mortality (excluding early pregnancy deaths), although the variation in mortality explained did not exceed 45% for any of the indicators (Figures 1–4). Logarithmic transformations of the maternal mortality ratio provided the best fit for all indicators, and findings are presented for the latter only. The correlation coefficient for the association between the MMRatio and the percentage of births with a skilled attendant or in hospital was −0.65 (R2 = 0.43, P = 0.006) and −0.54 (R2 = 0.29, P = 0.045) respectively. All areas with MMRatios in excess of 450 per 100 000 live births had fewer than 30% of births with a skilled attendant or less than 15% of births in hospital. There was also a relatively close link between the population-based Caesarean section rate and the MMRatio (R = −0.59, R2 = 0.34, P = 0.027). All sites with MMRatios of 450 per 100 000 or more had population-based Caesarean section rates of less than 1%, whereas lower maternal mortality levels were achieved with widely varying Caesarean section rates (from 0.6% in Farafenni to 2.7% in Ouagadougou). There was no association between the percentage of births in a health centre and the MMRatio (R = −0.42, R2 = 0.18, P = 0.132).

image

Figure 1. Scatter diagram of maternal mortality and the percentage of births in a health centre.

Download figure to PowerPoint

image

Figure 2. Scatter diagram of maternal mortality and the percentage of births in hospital.

Download figure to PowerPoint

image

Figure 3. Scatter diagram of maternal mortality and the percentage of births with a Caesarean section.

Download figure to PowerPoint

image

Figure 4. Scatter diagram of maternal mortality and the percentage of births with a skilled attendant.

Download figure to PowerPoint

There was marked clustering of levels of services use in urban and rural areas. Stratified analyses within urban and rural areas did not reveal any clear pattern, however, and none of the associations within strata were statistically significant (data not shown).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Study populations
  6. Maternal mortality
  7. Indicators of access to obstetric care
  8. Statistical methods
  9. Results
  10. Discussion
  11. Acknowledgements
  12. References

This study discloses huge inequalities in access to obstetric care and levels of maternal mortality in urban and rural West Africa. Most rural women give birth at home in the absence of skilled care, while urban women tend to give birth in a hospital with a skilled attendant. Concomitantly, maternal mortality is extremely high in rural areas, and substantially lower in urban areas. However, within urban or rural areas, there was no obvious association between maternal mortality and skilled obstetric care, hospital births or Caesarean sections.

In this ecological study, associations have to be interpreted with caution because the patterns of maternal mortality observed may reflect urban–rural differentials in factors other than use of obstetric care. Without doubt, a whole array of factors will determine why some women use services more or less often, ranging from financial barriers to issues of women's status and autonomy, but whether such factors affect maternal mortality other than through facilitating access to obstetric care is uncertain (Loudon 1992). Nutritional deficiencies have not been unequivocally linked to the incidence or severity of obstetric complications (Ronsmans 2000), except possibly for vitamin A deficiency (West et al. 1999). Physical differences between West African women from different ethnic origin, on the other hand, have been hypothesized as possibly contributing to varying levels of maternal mortality, and this may be worth exploring further (Høj 2002). While there is no doubt that contraceptive use may affect the maternal mortality ratio by preventing abortions and high-risk births, its absolute effect on the maternal mortality ratio is believed to be rather limited (Trussell & Pebley 1984; Fortney 1987).

This study has a number of other weaknesses that warrant caution in the interpretation of results. First, bias may have been introduced because the methods to ascertain maternal deaths and service use differed between sites. Secondly, some of the estimates of maternal mortality and service use covered different time periods and population groups, and changes over time or between populations in coverage of maternal health services may have prevented relevant comparisons. And thirdly, we could not adequately measure mortality in early pregnancy, and the levels of mortality are likely to have been underestimated.

The urban cohorts were followed at regular time intervals, maternal mortality assessed prospectively using multiple sources of data and service use data mostly obtained from the health providers themselves. In contrast, while mortality data in rural areas were collected prospectively, verbal autopsies went back in time and service use data mostly relied on women's recall. Recall is not thought to be a major problem for events as traumatic as a maternal death (WHO 1995), and given the special efforts made to ascertain maternal mortality in the rural areas, maternal deaths after 3 or 4 months of pregnancy are unlikely to have been missed. Moreover, while women may not have known the qualification of the person attending their labour, they are unlikely to have misreported their place of birth or whether or not they had undergone a Caesarean section.

Similarly, it is unlikely that variation in time periods or denominator populations has affected the results. Access to obstetric services has been persistently low for at least a decade in sub-Saharan Africa (Cisse et al. 1998; AbouZahr & Wardlaw 2001) and there is little evidence of declining trends in maternal mortality over time, except in Farafenni, The Gambia, where substantial declines have been observed (Walraven et al. 2000; Kodio et al. 2002). Our reliance on district-based estimates of service use for a few rural sites, – as opposed to data from the cohorts themselves – may also have affected the results. The rural populations studied here live in very remote areas, however, and use of referral services is low even among the more urbanized in the district (De Bernis et al. 2000). In Niakhar, Sénégal, the cohort-based estimates of the percentage of births with a Caesarean section were very similar to those reported for the district of Fatick (0.40% and 0.35% respectively), suggesting that bias might have been minimal. Excluding the sites with district-based estimates of service use from the correlational analysis did not substantially alter the findings (data not shown).

Our estimates of maternal mortality are much lower than those reported by WHO and UNICEF for 1995 (Hill et al. 2001). For the eight countries discussed here, Hill et al. estimate that maternal mortality ranges between 923 and 1379 per 100 000 live births, compared with our estimates of 241 and 628 per 100 000 live births for urban and rural areas respectively. Illegal abortion may be common in urban areas (Faye et al. 1996; Desgrées du Lou et al. 1999), and mortality from unsafe abortion has been estimated to account for 12% of maternal deaths in West Africa (WHO 1998). Abortion mortality alone, however, cannot explain the huge differences between our estimates of maternal mortality and those reported by Hill et al. Sole reliance on nationally representative data may limit the interpretation of the WHO/UNICEF estimates and the inclusion of local information from demographic surveillance sites with good estimates of maternal mortality should be considered.

In this study, nearly all births with a skilled attendant took place in a health facility (MOMA), and no inference can be drawn about the role of the skilled attendant per se. Similarly, the association between maternal mortality and hospital births should not be interpreted as meaning that a strategy of hospital births should be encouraged. Graham et al. (2001) highlighted the fact that a professional label alone may not be a good proxy for competence or skill, and that efforts should be made not only to assess providers’ skills, but also the enabling environment, including equipment, supplies, drugs and transport for referral. The lack of a link between maternal mortality and health centre births perhaps suggests that the quality of care offered in these health centres was not sufficient to prevent maternal death. Many so-called comprehensive EOC facilities do not offer 24-h emergency services and life-saving equipment and drugs may be missing (Cisse et al. 1998). High costs of services, poor attitudes of staff towards women and poor management also contribute to poor quality of care (The Prevention of Maternal Mortality Network 1995). Further research into feasible and reliable indicators of the quality of obstetric care and of skilled attendance is necessary.

We did not adequately measure indicators of need for obstetric care, which aim at identifying the sub-group of women thought to require specialized obstetric care (Maine et al. 1997; UNICEF, WHO & UNFPA 1997; De Brouwere & Van Lerberghe 1998; Ronsmans et al. 2002). In settings with poor access to life-saving obstetric care, however, Caesarean section rates may reflect the extent to which the need for life-saving obstetric care is met, although a substantial proportion may not be performed for maternal indications (Ronsmans et al. 2002). While the debate on the minimum or optimum Caesarean section rate continues, the findings of this study suggest that small differences in Caesarean section rates may hide huge differences in maternal mortality, and that substantial gains in maternal health can be achieved with Caesarean section rates far below 5%.

Despite the limitations of this ecological study, there can be little doubt that the huge rural–urban differences in maternal mortality are due, at least in part, to differential access to high quality maternity care. Whether any of the indicators examined here will by themselves be good enough as a proxy for maternal mortality is doubtful however, as more than half of the variation in mortality remained unexplained by any one of them. Further research will need to examine the combination of indicators which point to the actions required for improving maternal health, including those that measure access to and quality of skilled attendance for all, and specialist care for some. When alternative safe motherhood strategies are considered and monitoring schemes designed, however, explicit attention will also have to be given to how reductions in rural–urban inequalities in access to obstetric services can best be achieved.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Study populations
  6. Maternal mortality
  7. Indicators of access to obstetric care
  8. Statistical methods
  9. Results
  10. Discussion
  11. Acknowledgements
  12. References

We thank all the collaborators of the MAMOCWA study for their support with data collection and analysis, the European Commission/INCO-DC for funding the MAMOCWA study (contract number: ERBIC18CT970248), and Marie-Hélène Bouvier-Colle and Alain Prual for providing the mortality data for the MOMA sites and commenting on an earlier draft of the paper.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Study populations
  6. Maternal mortality
  7. Indicators of access to obstetric care
  8. Statistical methods
  9. Results
  10. Discussion
  11. Acknowledgements
  12. References
  • AbouZahr C & Wardlaw T (2001) Maternal mortality at the end of a decade: signs of progress? Bulletin of the World Health Organization 79, 561568.
  • Bouillin D, Fournier G, Gueye A, Diadhiou F & Cisse CT (1994) Surveillance épidémiologique et couverture chirurgicale des dystocies obstétricales au Sénégal. Cahiers Sante 4, 399406.
  • Bouvier-Colle MH, Prual A, De Bernis L (1998) Morbidité maternelle en Afrique de l'Ouest. Résultats d'une enquête en population a Abidjan, Bamako, Niamey, Nouakchott, Ouagadougou, Saint-Louis, et Kaolack. Ministère des Affaires étrangeres, Coopération Française, Paris.
  • Bouvier-Colle MH, Ouedraogo C, Dumont A, Vangeenderhuysen C, Salanave B & Decam C, for the MOMA group (2001) Maternal mortality in West Africa. Rates, causes and substandard care from a prospective survey. Acta Obstetricia et Gynecologica Scandinavica 80, 113119.
  • Cisse CT, Faye EO, de Bernis L, Dujardin B & Diadhiou F (1998) Césariennes au Sénégal: couverture des besoins et qualité des services. Cahiers Santé 8, 369377.
  • De Bernis L, Dumont A, Bouillin D, Guye A, Dompnier JP & Bouvier-Colle MH (2000) Maternal morbidity and mortality in two different populations of Sénégal: a prospective study (MOMA survey). British Journal of Obstetrics and Gynaecology 107, 6874.
  • De Brouwere V & Van Lerberghe W (1998) Les besoins obstétricaux non couverts. L'Harmattan, Paris.
  • Delaunay V, Marra A, Levi P & Etard JF (2002) Niakhar DSS, Sénégal. In: INDEPTH Network. Populations and Health in Developing Countries, Vol. 1: Population, Health and Survival at INDEPTH Sites.IDRC, Ottawa, pp. 279285.
  • Desgrées du Lou A, Mselatti P, Viho I & Welffens-Ekra C (1999) Le recours à l'avortement provoquéà Abidjan. Une cause de la baisse de la fécondité? Population 54, 427446.
  • Dumont A, de Bernis L, Bouvier-Colle MH & Breart G, for the MOMA study group (2001) Caesarean section rate for maternal indication in sub-Saharan Africa: a systematic review. Lancet 358, 132833.
  • Enel C, Pison G & Lefebvre M (1993) De l'accouchement traditionnel à l'accouchement moderne au Sénégal. Cahiers Santé 3, 441446.
  • Faye EO, Diouf A, Moreira P, Sangare M, Bathija H & Diadhiou F (1996) Morbidité et mortalité provoqués dans quatre sites dakarois. Archives of Public Health 53 (S1), 50.
  • Filippi V, Ronsmans C, Gandaho T, Graham W, Alihonou E & Santos P (2000) Women's report of severe (near-miss) obstetric complications in Benin. Studies in Family Planning 31, 309324.
  • Fortney J (1987) The importance of family planning in reducing maternal mortality. Commentary. Studies in Family Planning 18, 109114.
  • Graham W (2002) Now or never: the case for measuring maternal mortality. Lancet 359, 701704.
  • Graham W, Filippi V & Ronsmans C (1996) Demonstrating impact using maternal mortality: an impossible dream? Health Policy and Planning 11, 1620.
  • Graham WJ, Bell JS & Bullough CHW (2001) Can skilled attendance at delivery reduce maternal mortality in developing countries? Studies in Health Services Organisation & Policy 17, 97130.
  • Hill K, AbouZahr C & Wardlaw T (2001) Estimates of maternal mortality for 1995. Bulletin of the World Health Organization 79, 182193.
  • Høj L (2002) Maternal mortality in Guinea-Bissau. Levels, causes and risk factors. PhD Thesis. Faculty of Health Sciences, University of Aarhus.
  • Høj L, Stensballe J & Aaby P (1999) Maternal mortality in Guinea-Bissau: the use of verbal autopsy in a multi-ethnic population. International Journal of Epidemiology 28, 7076.
  • Kodio B, de Bernis L, Ba M, Ronsmans C, Pison G & Etard JF (2002) Levels and causes of maternal mortality in rural Sénégal. Tropical Medicine and International Health 7, 499505.
  • Loudon I (1992) Death in childbirth. An International Study of Maternal care and Maternal Mortality 1800–1950. Clarendon Press, Oxford.
  • Maine D, Akalin MZ, Ward VM & Kamara A (1997) The Design and Evaluation of Maternal Mortality Programs. Centre for Population and Family Health, School of Public Health, Columbia University, New York.
  • Ministère de la Santé, SERDHA et Macro International Inc. (1999) Enquête Sénégalaise sur les Indicateurs de Santé, 1999. SERDHA et Macro International Inc., Calverton, MD, USA.
  • Pison G, Wade A, Gabadinho & Enel C (2002a) Mlomp DSS, Sénégal. In: INDEPTH Network. Populations and Health in Developing Countries, Vol. 1: Population, Health and Survival at INDEPTH Sites.IDRC, Ottawa, pp. 271278.
  • Pison G, Guyavarch E & Sokhna C (2002b) Bandafassi DSS, Sénégal. In: INDEPTH Network. Populations and Health in Developing Countries, Vol. 1: Population, Health and Survival at INDEPTH Sites.IDRC, Ottawa, pp. 263269.
  • Ronsmans C (2000) Maternal mortality. In: Nutrition and Health in Developing Countries (eds RSemba & M, Bloem) The Humana Press, Totowa, New Jersey, pp. 3156.
  • Ronsmans C, Campbell OMR, McDermott J & Koblinsky M (2002) Questioning the indicators of need for obstetric care. Bulletin of the World Health Organization 80, 317324.
  • Telfer MI, Rowley JT & Walraven G (2002) Experiences of mothers with antenatal, delivery, and postpartum care in rural Gambia. African Journal of Reproductive Health, 6, 7483.
  • The Prevention of Maternal Mortality Network (1995) Situation analyses of emergency obstetric care: examples from eleven operations research projects in West Africa. Social Science and Medicine 40, 657667.
  • Trussell J & Pebley AR (1984) The potential impact of changes in fertility on infant, child and maternal mortality. Studies in Family Planning 15, 267280.
  • UNICEF, WHO & UNFPA (1997) Guidelines for Monitoring the Availability and Use of Obstetric Services. UNFPA, New York.
  • Walraven G, Telfer M, Rowley J & Ronsmans C (2000) Levels of maternal mortality, its causes and contributing factors in rural Gambia. Bulletin of the World Health Organization 78, 603613.
  • West KP, Katz J & Khatry SK et al. (1999) Double blind cluster randomised trial of low dose supplementation with vitamin A and β carotene on mortality related to pregnancy in Nepal. British Medical Journal 318, 570575.
  • World Health Organization (1995)Verbal Autopsies for Maternal Deaths. Report of a WHO workshop. (WHO/FHE/MSM/95.15) WHO, Geneva.
  • World Health Organization (1997) Monitoring Reproductive Health: Selecting a Short List of National and Global Indicators. (WHO/RHT/HRP/97.26). WHO, Geneva.
  • World Health Organization (1998) Unsafe Abortion. Global and Regional Estimates of Incidence of and Mortality due to Unsafe Abortion, with a Listing of Available Country Data. (WHO/RHT/MSM/97.16) WHO, Geneva.