Services on Demand
Print version ISSN 0042-9686
Bull World Health Organ vol.84 n.11 Genebra Nov. 2006
PUBLIC HEALTH REVIEWS
Lacunes de la littérature scientifique sur la mortalité maternelle mises en évidence par une revue systématique de cette littérature
Lagunas en las publicaciones científicas sobre la mortalidad materna: revisión sistemática
Diana Gil-González1; Mercedes Carrasco-Portiño; Maria Teresa Ruiz
Preventive Medicine and Public Health, University of Alicante, Campus San Vicente del Raspeig, Ap. 99 C.P 03080, San Vicente del Raspeig (Alicante), Spain
Issues related to maternal mortality have generated a lot of empirical and theoretical information. However, despite the amount of work published on the topic, maternal mortality continues to occur at high rates and solutions to the problem are still not clear. Scientific research on maternal mortality is focused mainly on clinical factors. However, this approach may not be the most useful if we are to understand the problem of maternal mortality as a whole and appreciate the importance of economical, political and social macrostructural factors. In this paper, we report the number of scientific studies published between 2000 and 2004 about the main causes of maternal death, as identified by WHO, and compare the proportion of papers on each cause with the corresponding burden of each cause. Secondly, we systematically review the characteristics and quality of the papers on the macrostructural determinants of maternal mortality. In view of their burden, obstructed labour, unsafe abortion and haemorrhage are proportionally underrepresented in the scientific literature. In our review, most studies analysed were cross-sectional, and were carried out by developed countries without the participation of researchers in the developing countries where maternal mortality was studied. The main macrostructural factors mentioned were socioeconomic variables. Overall, there is a lack of published information about the cultural and political determinants of maternal mortality. We believe that a high-quality scientific approach must be taken in studies of maternal mortality in order to obtain robust comparative data and that study design should be improved to allow causality between macrostructural determinants and maternal mortality to be shown.
Une grande quantité de données empiriques et théoriques ont été rassemblées sur les problèmes liés à la mortalité maternelle. Cependant; malgré le volume de travaux publiés à ce sujet, les taux de mortalité maternelle se maintiennent à des niveaux élevés et il n'existe guère de solutions claires à ce problème. La recherche scientifique dans ce domaine est axée principalement sur les facteurs cliniques. Cependant, cette approche n'est peut être pas la plus utile si le problème de la mortalité maternelle n'est pas appréhendé dans son ensemble et si l'importance des facteurs macrostructuraux économiques, politiques et sociaux n'est pas dûment prise en compte. Le présent article recense le nombre d'études scientifiques publiées entre 2000 et 2004 sur les causes principales de décès maternel selon l'OMS et compare la proportion d'articles consacrés aux différentes causes à la charge de mortalité associée à ces causes. Par ailleurs, il analyse de manière systématique les caractéristiques et la qualité des articles portant sur les déterminants macrostructuraux de la mortalité maternelle. Par comparaison avec la charge de mortalité qu'elles entraînent, les causes de mortalité maternelle suivantes sont proportionnellement sous-représentées dans la littérature scientifique : dystocies, avortements non médicalisés et hémorragies. La plupart des études que nous avons analysées sont transversales et ont été réalisées par des pays développés sans la participation de chercheurs des pays en développement servant de cadres à ces études. Les principaux facteurs macrostructuraux mentionnés étaient des paramètres socioéconomiques. Globalement, la littérature publiée ne traite pas suffisamment des déterminants culturels et politiques de la mortalité maternelle. Nous pensons qu'il est nécessaire, pour étudier la mortalité maternelle, d'appliquer une démarche scientifique rigoureuse afin d'obtenir des données solides et comparables, et d'améliorer la conception des études pour mettre en évidence les relations de causalité entre les déterminants macrostructuraux et la mortalité maternelle.
Los temas relacionados con la mortalidad materna han generado abundante información empírica y teórica. Sin embargo, pese a los muchos trabajos publicados sobre el tema, las tasas de mortalidad materna siguen siendo altas, y las soluciones a este problema siguen también sin estar claras. Las investigaciones científicas sobre la mortalidad materna se centran principalmente en los factores clínicos. Sin embargo, puede que este enfoque no sea el más conveniente para lograr comprender el problema de la mortalidad materna en su globalidad y apreciar la importancia de los factores macroestructurales de índole económica, política y social. En este artículo analizamos el número de estudios científicos publicados entre 2000 y 2004 sobre las causas principales de mortalidad materna según la OMS, comparando la proporción de artículos sobre cada causa con la carga correspondiente a la misma. En segundo lugar, examinamos sistemáticamente las características y la calidad de los artículos sobre los determinantes macroestructurales de la mortalidad materna. A la vista de la carga que representan, el parto obstruido, el aborto peligroso y la hemorragia están subrepresentados proporcionalmente en la bibliografía científica. En nuestra revisión, la mayoría de los estudios analizados eran transversales, y habían sido realizados por países desarrollados sin la participación de investigadores de los países en desarrollo donde se estudió la mortalidad materna. Los principales factores macroestructurales mencionados eran variables socioeconómicas. En términos generales, se observa una escasez de publicaciones sobre los determinantes culturales y políticos de la mortalidad materna. Consideramos que los estudios sobre ésta deben adoptar un enfoque científico de calidad que permita obtener datos robustos y comparables, y que el diseño de esos estudios debe mejorarse para poner de manifiesto las relaciones de causalidad entre los determinantes macroestructurales y la mortalidad materna.
Much empirical and theoretical information exists on the various causes of maternal death. However, despite the amount of published work on this problem, maternal mortality remains high in many parts of the world and the solutions remain elusive.1
Although almost all cases of maternal death are avoidable,1 standard programmes to prevent deaths have not been effective in controlling the causes of maternal mortality. In 1998, the WHO Division of Reproductive Health stated that there had been little evidence of progress in the effort to reduce rates of maternal death.2 Some authors disagree with current safe-motherhood strategies and have raised concerns that interventions are poorly implemented or lack an evidence base.3,4 In Africa, for example, there have been calls for further operational research to identify and test promising safe-motherhood interventions, in particular interventions designed to reduce the high rate of maternal mortality associated with unsafe termination of pregnancy.5 Strategies to reduce maternal mortality have also been criticized on the grounds that the component programmes are based mostly on low-grade evidence derived from retrospective and observational studies, as well as on empirical knowledge or conclusions reached via deductive reasoning.6 A further problem is that most research is focused on providing evidence for clinical decision-making.7 This tactic needs to be reconsidered to fit with strategies at the population level since safe motherhood strategies require complex public health and multidisciplinary policy approaches.4,6
In many countries, maternal mortality is the main cause of death in women of reproductive age.7,8 The severity of the problem has caused the United Nations General Assembly to focus its attention on improving maternal health. The fifth Millennium Development Goal (MDG) is to reduce maternal mortality by three-quarters by 2015.9 Clearly, interventions in reproductive health are necessary if the burden of maternal mortality is to be alleviated.
To increase the likelihood of achieving a reduction in maternal deaths, WHO launched the Partnership for Maternal, Newborn and Child Health in September 2005. This collaboration between many academic and research institutions, governments, nongovernmental organizations, professional groups and health agencies aims to take immediate action to help women and children to survive.10 The importance of addressing maternal health with the same determination as that shown for child health and the influence of Goal 5 on the other MDGs are the key points of this partnership. WHO intends to prioritize maternal health10 and to raise awareness of the fact that maternal health interventions have not been implemented in countries where pregnancy and childbirth pose the greatest threats to women's health.1
There is agreement that access to health services is essential if high rates of maternal death are to be reduced.11 Nevertheless, a focus on only the clinical causes of maternal death may not be the most comprehensive perspective from which to understand the problem of maternal mortality since it largely ignores the macrostructural i.e. the social, cultural, economic and political determinants of health.12-15
Some studies have noted the relation between maternal mortality and socioeconomic factors, such as income per capita, gross domestic product and educational level.16,17 However, less effort has been spent researching the influence of culture and political conditions on health although it has been shown that political factors, such as democracy, have a positive effect on health indicators including maternal mortality.17 The United Nations Population Fund (UNPFA) has promoted the "three delays model," a framework useful for identifying the stages at which delays can occur in the management of obstetric complications.18
There is concern that little research is done on diseases that affect people in developing countries.19 Health professionals and policy-makers from those countries may have difficulty in setting priorities20 and also there is an increasing demand for research to be prioritized according to the prevalence of a health issue.21 Therefore, although all health problems that can cause a maternal death are important, not all the underlying causes of maternal mortality will attract the same amount of scientific interest.
As political decisions and best practice should be based on evidence, the lack of information and the quality of research on the macrostructural determinants of health could be a handicap. Consequently, the objectives of our study are: to quantify the amount of scientific interest in the main causes of maternal mortality and compare the proportion of papers on each cause with the corresponding burden of the cause; and to assess the characteristics and quality of papers on the macrostructural determinants of maternal mortality.
To measure the amount of scientific interest in the main causes of maternal death, we identified and counted all papers published between 2000 and 2004 that mentioned WHO-defined causes of maternal mortality.1 The decision to start from 2000 was somewhat arbitrary and based on the idea of reviewing only recent data.
We searched several databases that covered the social and behavioural sciences, clinical medicine and life sciences: Medline, Science Citation Index, Institute for Scientific Information (ISI) Current Contents, the Cumulative Index to Nursing and Allied Health (CINAHL), PsycINFO EconLit, FRANCIS, Sociological Abstracts, and the Education Resources Information Center (ERIC). To identify the keywords, a thesaurus (such as MeSH) was consulted in all the databases. In the cases of the Science Citation Index and ISI Current Contents, which do not have a thesaurus, we used the search equation "maternal mortality AND [search term]" with the following search terms: haemorrhage, infection, eclampsia, preeclampsia, obstructed labour, unsafe abortion, AIDS, anaemia, malaria and cardiac disease. Duplicate papers in the databases were deleted.
To assess the characteristics of published work on macrostructural determinants of maternal mortality, we undertook a systematic review of original research papers identified from the previously mentioned databases using the search term "maternal mortality." We included all papers in any language published between 2000 and 2004 that mentioned any of the macrostructural determinants (social, cultural, economic and political factors) of maternal mortality in the title or abstract.
We excluded papers that dealt mainly with clinical issues such as obstetric interventions during and after delivery, the application of medical technologies, clinical trials of medicines, clinical practice guidelines, medical protocols from hospitals and primary health care centres, maternal health not associated with maternal mortality, pregnancy-related diseases (e.g. diabetes, high blood pressure, pre-eclampsia) and infant mortality. In addition, non-original papers such as theoretical reviews, book reviews, letters, editorials, summaries of conferences, historical papers and papers without abstract were excluded. After screening 2250 abstracts, we analm lysed the full text of all papers included in the final systematic review using a checklist to capture information about the papers (Box 1). To assess the level of agreement between the researchers who reviewed the papers, the codifiers and classifiers (MC and DG) independently analysed a sample of 10 papers: a high level of concordance was noted (kappa index = 88%).
We calculated frequencies and percentages of the variables included in the review using SPSS 11 and MS Excel 2000 commercial software.
We identified 1232 papers published between 2000 and 2004 about the main causes of maternal mortality. Table 1 shows the frequency distribution of published scientific articles by cause and compares this with the burden of each cause.1 Articles on eclampsia or pre-eclampsia (28.25% of all scientific papers about causes of maternal death) and all indirect causes (26.71%) were over-represented when compared with the burden of these causes reported by WHO (12% and 20%, respectively). There is a lack of information about obstructed labour (2.51%) and unsafe abortions (3.73%) in light of their prevalence (8% and 13%, respectively). Information about haemorrhage is also slightly under-represented in the scientific literature (19.72% of papers but a burden of 25%). The proportion of papers on infection (18.26%) was similar to its prevalence as a cause of death (15%).
From 2000 to 2004, we identified 2250 papers on maternal mortality. However, after applying the exclusion criteria, only 27 papers were selected.22-48 These were studies focused on the possible relation between maternal mortality and socioeconomic, political and cultural factors.
Of the 2223 papers excluded, 842 (37.87%) were theoretical or were book reviews, letters, editorials, summaries of conferences, historical papers or papers without an abstract. Another 835 (37.56%) were excluded because they focused on evaluations of obstetric interventions during and after delivery, the application of medical technologies and clinical trials. Also excluded were papers on clinical practice guidelines and medical protocols from hospitals and primary health-care centres (9.71%), maternal health not associated with maternal mortality (9.35%), pregnancy-related disease (3.95%) and infant mortality (1.52%).
Of the 27 papers, 23 (85%) analysed the influence of macrostructural determinants on maternal mortality in developing countries: six papers reported findings from a group of developing countries,23,24,35,38,46,48 three reported data from Nigeria,29,31,44 two from Uganda,30,32 two from Zambia,22,25 and there was one paper from each on Afghanistan,34 Bangladesh,43 Cambodia,27 Guatemala,48 Guinea Bissau,28 India,45 Indonesia,26 Mexico,39 Mozambique37 and the United Republic of Tanzania.33 However, controversially, 70% of the studies19 were done by scientists from developed countries: 10 from the USA,24,27,29,34-37,40,41,43 four from the United Kingdom,38,44-46 two from Japan,23,26 and one each from Denmark,28 the Netherlands25 and Sweden.48
With respect to the study design of papers on macrostructural determinants of maternal mortality, 17 of the 27 papers were cross-sectional (63%);22,24,25,29-35,37,41-46 there were three case-control studies (11%),26,36,40 one cohort study (4%),28 three ecological studies (11%)23,38,47 and three qualitative studies (11%).22,39,48
Table 2 shows where in the 27 papers macrostructural determinants and accessibility to health care and services were mentioned. Socioeconomic status was mentioned in 18 introduction sections (67%),23-26,31,34-37,39,40,42-48 18 methods sections (67%)22-24,26,27,31-35,39,40,43-48 and 16 results sections (59%).24-26,31-35,37,40,42-48 Educational level of the woman was considered in six introduction sections (22%), 12 methods sections (44%) and 16 results sections (59%). One study47 considered political stability as a factor that might influence rates of maternal death. Research on the effect of culture on maternal mortality was rare (nine from 27 papers). Furthermore, messages from these nine papers were contradictory some supported the local culture and proposed integration with the official health system while other papers criticized the local culture as a risk factor in maternal mortality.
Mention of attendance by a traditional birth attendant appears in 12 results sections (44%) but in only seven (26%) methods sections. Accessibility of health-care centres shows similar results: although the problem of distance between a woman's home and the health-care centre is mentioned in 11 results sections (41%), this variable is only included in only three methods sections (11%).
Few studies have been found about the relationship between macrostructural determinants of maternal mortality and health care and services. The clinical variables that are most often mentioned in the results section of papers about macrostructural determinants of maternal mortality are care during delivery (n = 12 [44%])24,26,32,34,35,37,40,42-45,47 and perinatal care (n = 11 [41%]).24,27,29-32,34,35,42,43,45
In our review of published articles on maternal mortality, we found that reports on obstructed labour, unsafe abortion and haemorrhage are under-represented in the scientific literature. This finding suggests that there may be an inverse scientific interest pattern understood as a knowledge-base gap between the burden of the causes of maternal mortality and scientific interest in these determinants. Furthermore, we noted that research concentrates on clinical issues rather than on the social, economic and political factors that may affect rates of maternal death, as shown by the number of papers excluded. Most studies analysed were carried out by researchers from developed countries without the participation of the developing countries where maternal mortality was studied. There is also a lack of information about the cultural and political determinants of maternal mortality. Moreover, research on maternal mortality lacked robust methodological design and we believe that the quality of such research needs to be improved to produce knowledge about macrostructural causes of maternal mortality.
Since the mention of macrostructural variables was the main criterion for inclusion in our review, some important papers may have been overlooked given that clinical papers were excluded. For example, pregnancy-related studies in which economic, social and political aspects are not mentioned in the abstract but are considered later in the article would not have been included in our study. Another factor which may have affected the completeness of our review is that not all research papers on maternal mortality are published in scientific journals; such work also appears in institutional and nongovernmental organization reports.49 Furthermore, the difficulties encountered in obtaining data from developing countries with high rates of maternal deaths limit the quantity and quality of scientific results emerging from these countries. Overall however, the small number of papers finally selected for the review reflects the near-absence of scientific studies on the relation between maternal mortality and macrostructural determinants; it also restricts our ability as researchers to draw meaningful conclusions about the nature of work on the topic. One strength of the review, however, is that inter-reviewer agreement was strong, as shown by the kappa index.
The fact that more than two-thirds of the studies were carried out by researchers from the developed world without the participation of the developing countries could be problematic. Work by researchers from developing countries is underrepresented in health research: data from the developing world are mostly obtained by researchers with resources and values from developed countries.50 Enabling developing countries to play a leading role in the study of their maternal diseases and health contexts seems crucial if changes to health status are to be achieved. The involvement of workers from the countries themselves may also result in more work being produced on high prevalence factors in maternal mortality, such as obstructed labour, unsafe abortion and haemorrhage. Policy-makers and health professionals including traditional health-care staff should be able to access information and be involved in the process of generating and disseminating information that contributes to best practice and political decisions related to women's health.
The failure of some health strategies in developing countries could be related to a lack of knowledge about the cultural and political structures of such countries: this problem might have been the reason behind the lack of success resulting from some initiatives implemented in the 1990s.11 Although maternal mortality ratios have decreased in some countries such as El Salvador, Ethiopia, Honduras, Mozambique and Uganda, they have increased in other places, including Guatemala, Nicaragua and Pakistan, despite the measures taken.11 In particular, the lack of involvement of the local culture in the implementation of health strategies may not have had a positive effect on maternal mortality.
An integrated approach
The clinical model of reinforcing health services and trained staff may not be enough to solve the problem of maternal mortality.16 Our review of the literature shows that the effects of cultural and political issues on maternal mortality are largely ignored. However, political variables, such as democracy, have been shown to have an important influence on health indicators such as maternal mortality ratio.17 This means that in some contexts characterized by chronic poverty, researchers are not considering the complete set of determinants that may affect rates of maternal deaths. Gender equity and traditional culture23-26,31-34,37,40,42-47 must be taken into account to improve maternal health. For instance, it could be interesting to use the three delays model as a research tool to allow macrostructural factors and health factors to be integrated into one analysis.18
Obtaining information about the relation between maternal mortality and the coordination of the different institutional levels is also an important factor in measuring the real magnitude of maternal mortality. In the case of death directly related to the distance between women's homes and health-care centres, it is not only health-care institutions that should be examined, but also other factors such as infrastructure, transport and social services. All relevant governmental departments, not only health, should also be involved in tackling the problem.9
Recommendations for the future
Most studies of maternal mortality focus on clinical factors. However, the lack of new knowledge and the poor quality of macrostructural determinants of these maternal mortality studies affect policy-makers and health professionals.4,12 As it is recommended that political decisions be based on scientific evidence, scientific literature on maternal mortality should be re-examined carefully before putting reproductive health policies into practice and the number of studies on such macrostructural variables must be increased. Furthermore, the problem of maternal mortality must be understood as a whole. We must look beyond the difficulties that pregnant women may face in accessing health services to develop a broader framework that includes sociocultural and political contexts as well as other structural boundaries such as access to health services. Furthermore, a high-quality scientific perspective must be achieved to obtain robust comparative data and study designs should be improved to allow causality to be proven. In addition to these changes, papers should be written so that the information is presented with coherence between the different sections of the paper.
Finally, to tackle the issue of maternal mortality, a new perspective based on identification of new information sources in different environments is required. This could be achieved by not only taking into account information from public and private health services, but also by looking at the community's own experiences.
Competing interests: none declared.
2. AbouZahr CL. Lessons on safe motherhood. World Health Forum 1998;19:253-60. [ Links ]
3. Maine D, Rosenfield A. The Safe Motherhood Initiative: why has it stalled? Am J Public Health 1999;89:480-2. [ Links ]
4. Gülmezoglu AM, Say L, Betrán AP, Villar J, Piaggio G. WHO systematic review of maternal mortality and morbidity: methodological issues and challenges. BMC Med Res Methodol 2004;4:16. Available from: http://www.pubmedcentral.gov/picrender.fcgi?artid=481067&blobtype=pdf. [ Links ]
5. Luck M. Safe motherhood intervention studies in Africa: a review. East Afr Med J 2000;77:599-607. [ Links ]
6. Bullough C, Meda N, Makowiecka K, Ronsmans C, Achadi EL, Hussein J. Current strategies for the reduction of maternal mortality. BJOG 2005;112:1180-8. [ Links ]
7. Campbell O, Gipson R, Hakim Issa A, Matta N, El Deed B, El Mohandes A, et al. National maternal mortality ratio in Egypt halved between 1992-93 and 2000. Bull World Health Organ 2005;83:462-71. [ Links ]
8. McCarthy J, Maine D. A framework for analysing the determinants of maternal mortality. Stud Fam Plann 1992;23:23-33. [ Links ]
9. World Health Organization. Health in the Millenium Development Goals. Geneva: WHO; 2005 Available from http://www.who.int/mdg/publications/mdg_report/en/index.html. [ Links ]
10. World Health Organization. New global partnership will take immediate action to help women and children survive. Geneva: WHO; 2005. Available from: http://www.who.int/mediacentre/news/releases/2005/pr41/en/index.html. [ Links ]
11. Benagiano G, Thomas B. Safe motherhood: the FIGO initiative. Int J Gynaecol Obstet 2003;82:263-74. [ Links ]
12. Khan KS, Wojdyla D, Say L, Gülmezoglu AM, Van Look P. WHO analysis of causes of maternal death: a systematic review. Lancet 2006;367:1066-74. [ Links ]
13. Grimes DA. Unsafe abortion: the silent scourge. Br Med Bull 2003;67:99-113. [ Links ]
14. Ross L, Simkhada P, Smith WCS. Evaluating effectiveness of complex interventions aimed at reducing maternal mortality in developing countries. J Public Health (Oxf) 2005;27:331-7. [ Links ]
15. Thaddeus S, Nangalia R, Vivio D. Perceptions matter: barriers to treatment of postpartum haemorrhage. J Midwifery Womens Health 2004;49:293-7. [ Links ]
16. Costello A, Osrin D, Manandhar D. Reducing maternal and neonatal mortality in the poorest communities. BMJ 2004;329:1166-8. [ Links ]
17. Franco A, Álvarez-Dardet C, Ruiz MT. Effect of democracy on health: ecological study. BMJ 2004;329:1421-3. [ Links ]
18. Barnes-Josiah D, Myntti C, Augustin A. The "three delays" as a framework for examining maternal mortality in Haiti. Soc Sci Med 1998;46:981-93. [ Links ]
19. World Health Organization. World report on knowledge for better health: strengthening health systems. Geneva: WHO; 2004 Available from: http://www.who.int/rpc/meetings/world_report_on_knowledge_for_better_health.pdf. [ Links ]
20. Fraser DW. Overlooked opportunities for investing in health research and development. Bull World Health Organ 2000;78:1054-61. [ Links ]
21. Mexico, 2004: Global health needs and new research agenda. Lancet 2004;364:1555-6. [ Links ]
22. Maimbolwa MC, Yamba B, Diwan V, Ransjo-Arvidson A. Cultural childbirth practices and beliefs in Zambia. J Adv Nurs 2003;43:263-74. [ Links ]
23. Herrera LR, Kakehashi M. An international data analysis on the level of maternal and child health in relation to socioeconomic factors. Hiroshima J Med Sci 2001;50:9-16. [ Links ]
24. Shiffman J. Can poor countries surmount high maternal mortality? Stud Fam Plann 2000;31:274-89. [ Links ]
25. Stekelenburg J, Kyanamina S, Mukelabai M, Wolffers I, Van Roosmalen J. Waiting too long: low use of maternal health services in Kalabo, Zambia. Trop Med Int Health 2004;9:390-8. [ Links ]
26. Taguchi N, Kawabata M, Maekawa M, Maruo T. Aditiawarman, Dewata L. Influence of socio-economic background and antenatal care programmes on maternal mortality in Surabaya, Indonesia. Trop Med Int Health 2003;8:847-52. [ Links ]
27. White PM. Heat, balance, humors, and ghosts: postpartum in Cambodia. Health Care Women Int 2004;25:179-94. [ Links ]
28. Hoj L, Da Silva D. Hedegaard, Sandstrom A, Aaby P. Maternal mortality: only 42 days? BJOG 2003;110:995-1000. [ Links ]
29. Salihu HM, Adamu YM, Aliyu ZY, Aliyu MH. Pregnancy-associated morbidity in Northern Nigeria. J Obstet Gynaecol 2004;24:367-71. [ Links ]
30. Mbonye AK. Risk factors associated with maternal deaths in health units in Uganda. Afr J Reprod Health 2001;5:47-53. [ Links ]
31. Orji EO, Ogunlola IO, Onwudiegwu U. Brought-in maternal deaths in south-west Nigeria. J Obstet Gynaecol 2002;22:385-8. [ Links ]
32. Kyomuhendo GB. Low use of rural maternity services in Uganda: impact of women's status, traditional beliefs and limited resources. Reprod Health Matters 2003;11:16-26. [ Links ]
33. Mswia R, Lewanga M, Moshiro C, Whiting D, Wolfson L, Hemed Y, et al. Community-based monitoring of safe motherhood in the United Republic of Tanzania. Bull World Health Organ 2003;81:87-94. [ Links ]
34. Amowitz LL, Reis C, Iacopino V. Maternal mortality in Herat Province, Afghanistan, in 2002: an indicator of women's human rights. JAMA 2002;288:1284-91. [ Links ]
35. Bulatao RA, Ross JA. Which health services reduce maternal mortality? Evidence from ratings of maternal health services. Trop Med Int Health 2003;8:710-21. [ Links ]
36. Geller SE, Rosenberg D, Cox SM, Brown ML, Simonson L, Driscoll CA, et al. The continuum of maternal morbidity and mortality: factors associated with severity. Am J Obstet Gynecol 2004;191:939-44. [ Links ]
37. Chapman RR. Endangering safe motherhood in Mozambique: prenatal care as pregnancy risk. Soc Sci Med 2003;57:355-74. [ Links ]
38. Ronsmans C, Etard JF, Walraven G, Hoj L, Dumont A, De Bernis L, et al. Maternal mortality and access to obstetric services in West Africa. Trop Med Int Health 2003;8:940-8. [ Links ]
39. Castro R, Campero L, Hernández B, Langer A. A study on maternal mortality in Mexico through a qualitative approach. J Womens Health Gend Based Med 2000;9:679-90. [ Links ]
40. Harper MA, Espeland MA, Dugan E, Mayer R, Lane K, Williams S. Racial disparity in pregnancy-related mortality following a live birth outcome. Ann Epidemiol 2004;14:274-9. [ Links ]
41. Nannini A, Weiss J, Goldstein R, Fogerty S. Pregnancy-associated mortality at the end of the twentieth century: Massachusetts, 1990-1999. J Am Med Womens Assoc 2002;57:140-3. [ Links ]
42. Okonta PI, Okali UK, Otoide O, Twomey D. Exploring the causes of and risk factors for maternal deaths in a rural Nigerian referral hospital. J Obstet Gynaecol 2002;22:626-9. [ Links ]
43. Paul BK, Rumsey DJ. Utilization of health facilities and trained birth attendants for childbirth in rural Bangladesh: an empirical study. Soc Sci Med 2002;54:1755-65. [ Links ]
44. Magadi M, Diamond I, Madise N. Analysis of factors associated with maternal mortality in Kenyan hospitals. J Biosoc Sci 2001;33:375-89. [ Links ]
45. Pallikadavath S, Foss M, William Stones R. Antenatal care: provision and inequality in rural north India. Soc Sci Med 2004;59:1147-58. [ Links ]
46. Graham WJ, Fitzmaurice AE, Bell JS, Cairns JA. The familial technique for linking maternal death with poverty. Lancet 2004;363:23-7. [ Links ]
47. Buor D, Bream K. An analysis of the determinants of maternal mortality in sub-saharan Africa. J Womens Health (Larchmt). 2004;13:926-38. [ Links ]
48. Roost M, Johnsdotter S, Liljestrand J, Essén B. A qualitative study of conceptions and attitudes regarding maternal mortality among traditional birth attendants in rural Guatemala. BJOG 2004;111:1372-7. [ Links ]
49. Dickersin K, Min YI. Publication bias: the problem that won't go away. Ann N Y Acad Sci 1993;703:135-46. [ Links ]
50. Sumathipala A, Siribaddana S, Patel V. Under-representation of developing countries in the research literature: ethical issues arising from a survey of five leading medical journals. BMC Med Ethics 2004;5:5. [ Links ]
(Submitted: 26 December 2005 - Final revised version received: 27 June 2006 - Accepted: 29 June 2006)