RHO archives : Topics : Safe Motherhood

Annotated Bibliography

This is page 2 of the Safe Motherhood Annotated Bibliography. This page contains:

• Human rights approach to safe motherhood
• Community mobilization
• Measuring maternal mortality and program impact

To access more bibliographic entries, visit page 1, page 3, page 4, or page 5, or return to the complete list of topics covered in the Safe Motherhood Annotated Bibliography. Be sure to use the Glossary if you are unfamiliar with any of the terms on this page.

Please note that PDF files require Adobe Acrobat Reader software, which can be downloaded for free at www.adobe.com/products/acrobat/readstep.html.

 

Human rights approach to safe motherhood

Amowitz, L. et al. Maternal mortality in Herat Province, Afghanistan, in 2002. Journal of the American Medical Association 288(10):1284–1291 (September 11, 2002).
A cross-sectional survey of 4,886 Afghan women in seven districts in the Herat Province of Afghanistan assessed maternal mortality and human rights issues in the area. Using the sisterhood method of measuring maternal mortality, there were 276 maternal deaths among 14,085 sisters of respondents. This indicates a maternal mortality ratio of between 557 and 630 maternal deaths per 100,000 live births per year in the previous 10 to 15 years, or a point estimate of 593 deaths. Ninety-two percent of these deaths were in rural areas. Women reported problems with lack of food (41%), shelter (18%), and clean water (14%). The majority of women (87%) had to obtain permission from their husband or male relative prior to seeking health care. While 12 percent used birth control, 23 percent wanted to use it. Only 11 percent of women received prenatal care, and only 40 women (0.9%) delivered with the assistance of a trained health care worker. Surveys of health facilities in the province found that 17 of 27 were functional; only 5 provided essential obstetric care. This study suggests that high rates of maternal mortality may be an indicator of violations of womens human rights. Prevention of these deaths requires the protection and promotion of a range of womens rights over a long period of time.

Bloom, S. et al. Dimensions of womens autonomy and the influence on maternal health care utilization in a north Indian city. Demography 38(1): 67–78 (February 2001).
Interviews were held with 300 women in Varanasi, India, to assess aspects of womens autonomy and their relationship to use of maternal health care services. Autonomy was determined in three areas: (1) control over finances, (2) decision-making power, and (3) freedom of movement. After controlling for education, age, household structure, and other factors, women with closer ties to their natal kin were more likely to have greater autonomy in each of these areas. Women with greater freedom of movement obtained higher levels of antenatal care, and were more likely to use safe delivery care. Women with close ties to their natal kin could turn to their mothers when they wanted to go somewhere, such as the clinic. This study emphasizes the need to examine different aspects of autonomy when determining what factors affect health outcomes.

Center for Reproductive Rights and Association des Juriestes Maliennes. Claiming Our Rights: Surviving Pregnancy and Childbirth in Mali. New York: Center for Reproductive Rights (2003). Available in English at: www.crlp.org/pub_bo_mali.html, and in French at: www.crlp.org/fr_pub_bo_mali.html.
This report is the first in a series of four reports focusing on womens right to safe pregnancy and childbirth. This report analyzes pregnancy and childbirth in Mali and the laws and policies that contribute to the risks of pregnancy and childbirth in that country. It evaluates the Malian governments compliance with international legal obligations to uphold womens rights to survive pregnancy and childbirth. While the government of Mali has put in place policies to improve womens empowerment and health, more needs to be done to ensure that laws and policies are implemented and benefit Malian women. This requires concerted effort by the government and international, regional, and donor agencies. The report includes several recommendations for action at each level.

Cook, R.J. et al. Advancing Safe Motherhood Through Human Rights. Geneva: World Health Organization (2001). WHO/RHR/01.5. Available at: www.who.int/reproductive-health/publications/RHR_01_5_advancing_safe_motherhood/RHR01-5-Text.pdf.
This report analyzes how human rights laws can be used to promote safe motherhood. It describes the problem of maternal mortality and morbidity, and proposes ways in which the principles of human rights can be applied to safe motherhood. These human rights include rights to life, survival, and security; rights to maternity and health; rights to non-discrimination and due respect for difference; and rights to information and education. The report also provides strategies for implementing a human-rights approach to safe motherhood. The appendices provide data on maternal mortality, a listing of human rights relating to safe motherhood, international and regional human rights treaties, a bibliography of relevant websites, and excerpts from several documents on human rights.

Cook, R.J. and Dickens, B.M. Human rights to safe motherhood. International Journal of Gynecology & Obstetrics 76:225–231 (2002).
Many of the deaths and health complications related to pregnancy and childbirth could be prevented by addressing womens inadequate education, low social status, and lack of income and employment opportunities. These inequities, and others that condition unsafe motherhood, could be addressed by observing five basic human rights for women. These are womens right to life; to liberty and security of the person; to the highest attainable standard of health; to maternity protection; and to non-discrimination. These critical human rights offer pathways for women, their families, and their communities to make motherhood safe.

Freedman, L.P. Shifting visions: "delegation" policies and the building of a "rights-based" approach to maternal mortality. Journal of American Medical Womens Association 57(3):154–158 (2002).
In this commentary, the author proposes that a rights-based approach to maternal mortality in high-mortality countries requires decision makers to shift their focus. Using an example of the delegation of anesthesia provision in emergency obstetric care, the author shows that shifting from an individual, ethics-based, clinical perspective to a structural, rights-based, public health perspective can save more lives. In this hypothetical example, an anesthesiologist has the best survival rate (99%), but the lowest coverage (10% of population). A general practitioner with anesthesia training has a 90 percent survival rate, and a 40 percent coverage rate. A nurse-anesthetist has a survival rate of 90 percent, but covers 70 percent of the population. From a clinical perspective, survival rate matters most, while from a public health policy perspective, coverage is most important. In a hypothetical population of 10,000 women needing cesarean sections, the anesthesiologist could save 990 women, the general practitioner 3,600 women, and the nurse-anesthetist 6,300 women. The author argues that decision makers should work to address the power dynamics within the health care system to favor a rights-based public health perspective.

Freedman, L.P. Using human rights in maternal mortality programs: from analysis to strategy. International Journal of Gynecology & Obstetrics 75:51–60 (2001).
Human rights can be used to guide the design and implementation of maternal mortality policies and programs. First, universal access to high-quality emergency obstetric services should be made a priority. Human rights principles can then be integrated at the clinical, facility-management, and national policy levels. At the clinical level, efforts can be made to encourage respectful, non-discriminatory treatment of patients and staff. Focusing on entitlement and accountability can promote community participation and improve functioning of health facilities. Human rights principles can also be applied to health-sector reform and its impact on access to emergency obstetric services.

Liljestrand, J. and Gryboski, K. "Maternal Mortality as a Human Rights Issue." In: Murphy, E. and Ringheim, K., eds. Reproductive Health and Rights: Reaching the Hardly Reached. Washington, DC : PATH (2001).
The authors maintain that maternal mortality is related to women's low status in society. In many countries, women have little education, limited decision-making power, few resources, and are faced with health services that are insensitive to women's needs. As a result, many women do not receive the basic maternal care they need. Approaching safe motherhood as a human rights issue places emphasis on women's rights to adequate reproductive health care, including family planning, education, nutrition, and basic health services. Motherhood can be made safer by improving women's educational, economic, and political opportunities, in addition to implementing specific strategies to reduce maternal mortality. The publication also provides a list of core resources, including websites, related to maternal mortality.

Murphy, E. and Ringheim, K. eds. Reproductive Health, Gender and Human Rights: A Dialogue. Washington, DC : PATH (2001). Available at: www.path.org/files/RH-GHR-Dialogue.pdf.
This collection of articles is a result of a meeting, "Dialogue on Reproductive Health, Gender, and Human Rights," that took place in December 1999 at the World Bank. The articles, by human rights and public health experts, focus attention on human rights and gender issues in reproductive health programs. The dialogue is an effort to increase the communication between the human rights and public health sectors, and improve the integration of human rights perspectives in reproductive health.

Shen, C. and Williamson, J.B. Maternal mortality, women's status, and economic dependency in less developed countries: a cross-national analysis. Social Science & Medicine 49:197–214 (1999).
This study assesses how well three economic theories (modernization, economic dependency, and gender stratification) explain cross-national variations in maternal mortality. Using data from a sample of 79 developing countries, the analysis found that women's status (as measured by women's level of education relative to men, the presence of health attendants at delivery, contraceptive prevalence, age at first marriage, and total fertility) is a strong predictor of maternal mortality. In societies with similar levels of economic development, if women's status is higher, maternal mortality is lower. Consistent with other studies, this analysis also found that economic dependency has negative indirect effects on maternal mortality. While more data are needed to do a more thorough analysis, the authors conclude that action must be taken to improve the social, economic, and health status of women in order to reduce maternal mortality rates.

WHO. Safe Motherhood: A Matter of Human Rights and Social Justice. WHD 98.3, Geneva: WHO (1998). Available at: www.who.int/archives/whday/en/pages1998/whd98_03.html.
Improving safe motherhood worldwide and achieving the goals of the Safe Motherhood Initiative will not happen until women are empowered and their human rights are realized. This document lists the barriers women face in attaining good health, and the steps necessary to address these constraints. Preventing maternal deaths and illness is a social justice and human rights issue. Protecting and promoting the human rights of women will enable them to make decisions about their own health, and to access quality services and information before, during, and after pregnancy and childbirth. Existing international treaties offer under-utilized opportunities to monitor and advance safe motherhood.

Yamin, A.E. and Maine, D. Maternal mortality as a human rights issue: measuring compliance with international treaty obligations. Human Rights Quarterly 21(3):563–607 (1999).
This article provides an overview of maternal morbidity and mortality, and the public health approaches being used to address maternal health problems. It then explains how the UN guidelines for monitoring essential obstetric care services can be used as human rights tools to monitor compliance with womens rights to health under international treaties. By adopting measurable and verifiable indicators, it is possible to measure the progress of states in meeting their obligations. The indicators will also show if states are taking effective steps to improve maternal care, and whether they are doing so on a non-discriminatory basis (for example, benefiting women in both urban and rural areas).

Yanda K et al. Reproductive health and human rights. International Journal of Gynecology and Obstetrics. 2003;82:275–283.
The authors call for adopting a human rights-based approach to reproductive health care at the clinical as well as national levels, especially those programs offering abortion and postabortion care. Poor women face greater maternal mortality and morbidity than affluent women because of lack of access to adequate reproductive health services. They are also more likely to resort to unsafe, less accessible and less affordable abortion services. The public health sector has the ability and responsibility to reduce these inequities. Efforts must continue to prevent unwanted pregnancies, unsafe abortions, and abortion-related deaths; to treat abortion complications; to broaden the types of providers who are allowed to perform abortions; and to enhance training for abortion providers.

Top of page

Community mobilization

Ahluwalia IB et al. An evaluation of a community-based approach to safe motherhood in northwestern Tanzania. International Journal of Gynecology and Obstetrics. 2003;82:231–240.
This evaluation is of the Community-Based Reproductive Health Project (CBRHP) in two districts of northwestern Tanzania. CARE and the Ministry of Health worked together to strengthen community-level services by training and assisting village health workers, developing community-based plans for transportation to health facilities, and increasing local participation in planning and decision making. The project activities increased knowledge of danger signs, birth planning, timely referrals, and transport of pregnant women to hospitals. At least 36 women with obstetrical emergencies from 10 villages used the community-based transport system to reach a hospital. Village health workers received more community support. These interventions have lead to better care for pregnant women and have improved links between the communities and health facilities. With more support from the Ministry of Health these improvements can be sustained and replicated in other villages.

Fofana, P. et al. Promoting the use of obstetric services through community loan funds, Bo, Sierra Leone. International Journal of Gynecology & Obstetrics 59 (Suppl. 2):S225–S230 (1997).
In the Bo District of Sierra Leone, focus group discussions with community members, traditional birth attendants, and health staff revealed that a lack of funds contributed to delays in seeking and obtaining emergency obstetrical care. To address this need, communities were mobilized to establish systems to pay for emergency care. With assistance from the Prevention of Maternal Mortality Network, community-generated loan funds were established in 1992. These were funded by levies charged to adult members of the communities, and managed by village development committees. As a result of these loan funds, utilization of the referral hospital by women with obstetric complications increased from 5 in 1992 to 12 in 1993. Utilization by women from villages without loan funds stayed about the same (12 in 1992 and 13 in 1993). Nearly half of the women from loan fund areas paid their hospital bills in full (46%). Approximately 72 percent of the cost of establishing the loan funds was covered by the project. Although six communities mobilized to establish loan funds, only two were successful. This success is attributed to strong leadership in these two communities. Without strong leadership, this type of intervention may not be successful.

Howard-Grabman, L. et al. The Warmi Project: A Participatory Approach to Improve Maternal and Neonatal Health, An Implementors Manual. Arlington, Virginia: John Snow, Inc./MotherCare (1993).
This manual details a participatory approach to improving maternal and neonatal health in rural communities. It describes the steps taken by the three-year "Warmi" demonstration project in rural Inquisivi, Bolivia, implemented by Save the Children/Bolivia with funding from USAID's MotherCare Project. The goal of the project was to organize women's groups to increase women's knowledge and awareness of specific maternal and neonatal health problems and of the resources available locally to address these problems. The project developed a model for maternal and neonatal health interventions that can be replicated by other groups and other communities. According to the Community Action Cycle model, the first step is identification and prioritization of problems, followed by planning together, implementation, and evaluation. The manual provides detailed information for following this cycle. The project's success related to the fact that it responded to needs expressed by the community members. After three years, there was a noticeable increase in women's participation in community decision making, and in women communicating among themselves about maternal and neonatal health. Perinatal and neonatal deaths declined, while use of contraception, tetanus-toxoid coverage of pregnant women, immediate breastfeeding after delivery, and the use of trained birth attendants all increased. Although the methodology illustrated by the Warmi project is time-consuming and labor-intensive, it offers a low-cost, sustainable way to provide community members with the skills to identify and address local problems.

Kandeh, H.B.S., et al. Community motivators promote use of emergency obstetric services in rural Sierra Leone. International Journal of Gynecology & Obstetrics 59, Supple. 2: S209–S218 (1997).
A preliminary study in Bombali District, Sierra Leone, indicated that the three community public health units (PHUs) were in poor condition, had inadequately trained staff, experienced shortages of drugs and supplies, and had poor recordkeeping. A series of interventions were developed to remedy these problems, and increase use of emergency obstetric services at the PHUs. These included: community meetings, distribution of drug kits, staff training, renovation of the PHUs, and use of community motivators. The motivators undertook community education about recognition of obstetric complications, formation of action groups for transport of patients (teams of young men to carry patients to the PHU or to transport), and facilitation of emergency referrals. The improved services at the PHU increased their utilization by women with obstetric complications. However, the contribution of community motivators to the improved utilization appears small. It is a costly ($423 per motivator), labor-intensive intervention, especially in terms of monitoring and supervision. The output of the community motivators (number of women with complications referred) diminished over time. Other programs interested in replicating this intervention should consider these factors.

Levitt, M. J., et al. "Getting Messages Out: Partnerships and Innovative Community Mobilization in Nepal." Paper presented at the 1997 NCIH Conference, Washington, DC, June 12–14, 1997.
Communicating with Nepal's widespread, rural population poses great challenges for the many organizations committed to reducing maternal and neonatal deaths in the country. This paper describes a unique approach taken by a coalition of groups to communicate reproductive health messages to rural audiences. Several international and national nongovernmental organizations (NGOs) worked together to plan a National Condom Day to both educate and entertain. The success of this event, planned around the Dashain holiday, showed everyone the power of working together, and sparked the creation of National Clean Delivery Awareness Day (on March 8, International Women's Day). Twenty-six NGOs and the Ministry of Health collaborated to distribute Clean Home Delivery Kits, IEC materials, and to communicate safe motherhood messages. This stimulated the official launching of the Ministry of Health's National Safe Motherhood Program, and the creation of the Safe Motherhood at the Community and Family Level Support Network. The network held several planning workshops, and by the next International Women's Day event (1997), had grown to include more than 50 NGOs. IEC materials were distributed in 70 of Nepal's 75 districts and reached 500,000 people. The network has continued its planning and collaboration, organizing events around public and religious holidays. While initially competitive, the members of the network have learned to work as partners, along with the government. Although planned to raise awareness of reproductive health, the events have also served as catalysts for action from the local to the national levels.

Moore, K.M. Safer Motherhood 2000: Toward a Framework for Behavior Change to Reduce Maternal Deaths. The Communication Initiative (April 2000). Available online at www.comminit.com/misc/safer_motherhood.html.
While many aspects of safe motherhood programs have been reorganized over the last several years and new efforts have focused on allocating resources to reflect lessons learned, the communications components have not received the same attention. This paper reviews recent literature and emphasizes the need to utilize focused, innovative, participatory research and community development methodologies to design communication efforts to reduce maternal deaths. An updated, evidence-based behavior-change communication framework is needed to reprioritize key behaviors and break down each behavior for each target audience into sub-behaviors. The author offers six priority themes, ranked in order of their potential contribution to maternal mortality reduction, to help guide communication efforts. Safe motherhood communication efforts can learn from the successful strategies and models used in other health campaigns, such as HIV/AIDS-risk reduction. Collaborative community partnerships have great potential for mobilizing communities to reduce maternal deaths. Instead of using a top-down approach to educate about medical risks, models need to include comprehensive community participation that considers social risks as well.

Nwakoby, B. et al. Community contact persons promote utilization of obstetric services, Anambra State, Nigeria. International Journal of Gynecology & Obstetrics 59, Supple. 2: S219–S224 (1997).
Focus group discussions revealed that delays in seeking obstetric care contributed to maternal mortality in Njikoka Local Government Area, Anambra State, Nigeria. To remedy this, obstetric services at the local hospitals were upgraded and 48 community contact persons were recruited. Their role was to increase community awareness of obstetric complications, and to facilitate referral of women with complications. Of the 48 contact persons recruited and trained, 25 were still active after 18 months. They assisted 129 women, 14 percent of whom presented with a problem. While these community contact persons contributed to improved utilization of health services, they considered themselves as the health resource person in the community. The cost of the community motivator intervention was $635 per person, and the supervision demands were significant. If this intervention is to be replicated, the issues of compensation and supervision should be studied.

Russell R, Levitt-Dayal M. Igniting Change. Washington, DC: CEDPA; 2003. Available at: www.cedpa.org/publications/pdf/ignitingchange.pdf.
Two U.S. Agency for International Development-funded projects, Enabling Change for Women’s Reproductive Health (ENABLE) and Maternal and Neonatal health (MNH), have used social mobilization to promote change in reproductive health. Both projects have promoted communication and partnership with multiple stakeholders to effect enduring change. This report summarizes the theoretical basis for social change, and includes examples of its application and replication of successful strategies. The report concludes with challenges and lessons learned.

Schmid, T. et al. Transportation for maternal emergencies in Tanzania: empowering communities through participatory problem solving. American Journal of Public Health 91(10):1589–1590 (October 2001).
In the Mwanza region of Tanzania, poor roads, long distances, a lack of transportation vehicles, and lack of community or family plans for emergency transportation to health care services all contribute to high maternal mortality. A random sample baseline survey of community members in 50 villages and in-depth interviews with 110 community leaders, health care providers, teachers, mothers, and others revealed that no community had a transportation plan for obtaining urgent obstetric care. Most people thought such transportation was the sole responsibility of the mother. In 1996, U.S.-based CARE and CDC began activities to build community capacity for problem solving. Master trainers were trained in community empowerment to assist community leaders with the development of emergency transportation plans. An assessment done in April 2001 found that 19 villages had begun collecting funds for transportation systems; 13 villages had transportation systems available; and 10 had used them in the last three months. The project required significant technical and financial assistance over more than four years, and sustainability of community empowerment skills and activities depends on how well these practices are institutionalized within the communities.

Sibley L, Buffington ST, Haileyesus D. The American College of Nurse-Midwives’ Home-Based Lifesaving Skills Program: a review of the Ethiopia field test. Journal of Midwifery and Women’s Health. 2004;49(4):320–328.
The Home-Based Lifesaving Skills program (HBLSS) is a family- and community-focused program that transfers skills to family and community members to reduce maternal and newborn mortality. This study reports on the field test of the HBLSS in rural southern Ethiopia. There was improved performance in management of postpartum hemorrhage among those trained in HBLSS. This included skills transfer and retention and appropriate case management. Appropriate management of newborn infection was less evident. None of the communities had established reliable emergency transportation. About 38 percent of the community was exposed to HBLSS training, with strong community support. There is need for more programs to incorporate HBLSS in order to evaluate this promising approach.

Sibley, L. et al. Home-based life-saving skills: promoting safe motherhood through innovative community-based interventions. Journal of Midwifery & Womens Health 46(4):258–267 (July–August 2001).
The American College of Nurse Midwives (ACNM) has expanded its Life Saving Skills (LSS) training series to include an innovative community and competency-based program. The Home Based Life Saving Skills (HBLSS) program aims to reduce maternal and neonatal mortality by increasing access to basic life-saving measures within the home and the community. Field testing of the new program began in 2001 in India and Ethiopia. Baseline studies included a referral facility assessment, a community self-assessment, and a morbidity and performance assessment. Results were used to develop materials (such as Pictorial Learning Cards and Take Action Cards) to be used in flexible participatory trainings. The innovative program allows the use of different modules to educate community groups or individuals in topics such as "Women and Baby Problems" and "Too Much Bleeding." The HBLSS process is designed to empower communities and families to identify and solve problems related to maternal and newborn health. The HBLSS is part of a larger community partnership model that includes emergency transportation, community orientation to HBLSS, and monitoring of pregnancy outcomes.

The White Ribbon Alliance. Awareness, Mobilization, and Action for Safe Motherhood: A Field Guide. NGO Networks for Health, Washington, DC (2000). Available at: www.whiteribbonalliance.com/pdf/wrahow2guide.pdf.
This guide is designed to help developing country organizations become actively involved in the Safe Motherhood Initiative and the global White Ribbon Alliance. The guide offers suggestions for bringing a wide range of people and organizations together—at the local, national and international levels—to raise awareness and promote action toward safe motherhood. It suggests ways for a variety of organizations, from schools and religious groups to businesses and media, to work together to ensure safe motherhood. The emphasis is on communication to increase awareness of safe motherhood; build alliances among organizations; and stimulate action at the community, health service delivery and policy levels. The guide includes specific information for activities and events, examples of White Ribbon Alliances in several countries, and a guide to using the media for outreach.

World Health Organization (WHO). Home-based Maternal Records: Guidelines for Development, Adaptation and Evaluation. Geneva: WHO (1994).
These guidelines explain the development, adaptation and evaluation of home-based maternal records. Home-based maternal records are a low-cost, simple tool that offer families and communities a way to increase self-reliance and participation in health care. The record is designed to be kept by the mother, and carried with her to health visits. It includes information on her pregnancies, births of her children, postpartum, and interpregnancy periods. This WHO prototype was tested at 20 centers in 14 countries during 1984–1988. Results showed that home-based maternal records are useful to community health workers, traditional birth attendants and mothers. They increased the referral rate, the use of antenatal care, attendance at postpartum health checks, and childhood immunization rates. They serve as a reminder to women, their husbands, and other family members to take preventive steps against risk conditions, including making the necessary arrangements for funds and transportation to referral centers. Women also found the cards to be useful records of their own and their babies' health. The guide gives information on adapting the home-based record to local health needs and priorities, includes instructions on completing the records, and provides samples of the records used in various countries. Home-based records are a good source of data for program monitoring and evaluation, but their introduction and use require careful planning and training.

Top of page

Measuring maternal mortality and program impact

AbouZahr, C. and Wardlaw, T. Maternal mortality at the end of a decade: signs of progress? Bulletin of the World Health Organization 79(6):561–573 (2001). Available at: www.who.int/bulletin/pdf/2001/issue6/vol.79.no.6.561-573.pdf.
Maternal mortality is difficult to measure, especially in developing countries where few births are registered. This lack of data has forced researchers to pursue other indicators of maternal health. In this study, trends in skilled attendance at delivery and the rate of cesarean deliveries are used as proxies for obstetric care received by women in developing countries. From 1989 to 1999 there was a modest (1.7%) increase in births covered by trained assistance. The Middle East and North Africa showed the greatest improvement, while sub-Saharan Africa showed no overall change. The rate of cesarean deliveries is used as a proxy for the extent to which health care facilities offer, and women have access to, this form of essential obstetric care. Cesarean delivery can be used for the wrong reasons, so careful interpretation of the data is warranted. Very high maternal mortality rates are associated with very low cesarean rates. However, low levels of maternal mortality are associated with varying levels of cesarean deliveries. Trends in cesarean delivery rates appeared to remain steady during the 1990s. Most notably, in the countries with the lowest levels of cesarean deliveries—and the presumed greatest need for essential obstetric care services—there was the least change. These indicators appear to show slow progress in increasing use of medical care at delivery, with the exception of sub-Saharan Africa.

Akalin, M.Z., et al. Why perinatal mortality cannot be a proxy for maternal mortality. Studies in Family Planning 28(4):330–335 (December 1997).
Perinatal mortality (fetal (>28 weeks gestation) death plus newborn (<1 week) death) has been proposed as a proxy for maternal mortality in monitoring safe motherhood programs, because it occurs more frequently but is still related to obstetric factors. This analysis of data from Matlab, Bangladesh found that perinatal and maternal mortality rates were not correlated statistically and did not exhibit the same trends, probably because different types of obstetric complications caused maternal versus perinatal deaths. The authors suggest using process and output indicators to monitor safe motherhood programs.

AMDD Working Group on Indicators. Program note: Using UN process indicators to assess needs in emergency obstetric services: Morocco, Nicaragua and Sri Lanka. International Journal of Gynecology and Obstetrics 80(2):222–230 (February 2003).
Needs assessments conducted in Morocco (2000), Nicaragua (2001), and Sri Lanka (2002) are the first step in the Averting Maternal Death and Disability (AMDD) Program in these countries. The studies found that the three countries are making progress in ensuring that emergency obstetric care is available to all pregnant women. A high proportion of births take place in facilities in each country, and the met need for women with obstetric complications is elevated in Sri Lanka (75%) and Nicaragua (53%), and lower in Morocco (37%). The proportion of all births by cesarean section is within the recommended range in Nicaragua and Sri Lanka, and a bit low in Morocco. The case fatality rates are low in all three countries. Despite these positive overall indicators, regional disparities indicate not all women are benefiting from improved services.

AMDD Working Group on Indicators. Program note: using UN process indicators to assess needs in emergency obstetric services: Pakistan, Peru and Vietnam. International Journal of Gynecology and Obstetrics 78(3):275–282 (September 2002).
Needs assessments in Pakistan, Peru, and Vietnam were conducted in 2000 to determine the availability, quality, and use of emergency obstetric care facilities according to the UN process indicators. The results show that each country appears to have an adequate number of comprehensive facilities, but insufficient basic emergency obstetric care facilities. In Pakistan, the assessment shows disparities in access across districts, with Karachi West having fewer facilities. In Peru, the data indicate that obstetric facilities are often used to treat abortion-related complications, and thus contribute to the met need indicator. In both Pakistan and Vietnam, it is difficult to generalize at the provincial or national levels because data have only been collected at a limited number of facilities.

AMDD Working Group on Indicators. Program note: Using UN process indicators to assess needs in emergency obstetric services: Bhutan, Cameroon and Rajasthan, India. International Journal of Gynecology & Obstetrics 77(3):277–284 (June 2002).
This second program note reports on the findings of needs assessments in Bhutan, Cameroon and Rajasthan, India, as part of the Averting Maternal Death and Disability (AMDD) program. The results indicate the need for specific interventions in each country to improve the availability, use, and quality of emergency obstetric care.

Bailey, P.E. and Paxton, A. Program note: Using UN process indicators to assess needs in emergency obstetric services. International Journal of Gynecology & Obstetrics 76:299–305 (2002).
The Averting Maternal Death and Disability (AMDD) program is helping governments and international organizations improve access to emergency obstetric care. The first step in each project is a needs assessment, during which data on the U.N. process indicators is collected. This note reports on the results from Mozambique, Nepal, and Senegal. While one indicator alone cannot provide sufficient information for assessing a program, taken together the indicators are informative for the designing, monitoring, and evaluation of maternal health programs. Comparing the results with recommended or expected levels for each indicator allows program managers to identify problem areas. The indicators will be assessed periodically during the AMDD program to monitor progress toward improving access to, utilization of, and quality of emergency obstetric services.

Danel, I. et al. Applying the sisterhood method for estimating maternal mortality to a health facility-based sample: a comparison with results from a household-based sample. International Journal of Epidemiology 25(5):1017–1022 (1996).
The goal of this study was to compare maternal mortality estimates using the sisterhood method in a population-based household sample and a facility-based sample of adults attending outpatient health facilities. A total of 9,232 interviews were conducted in the 91 target health facilities in a rural region of Nicaragua. Results from the facility-based survey estimated a lifetime risk of maternal death of 0.0144 (1 in every 69 women). This estimate was identical to that from the household survey in the same region 8 months earlier, and corresponds to a maternal mortality ratio of 241 deaths per 100,000 live births. The authors concluded that the sisterhood method provides a robust estimate of the magnitude of maternal mortality and suggested that guidelines for utilizing this approach in similar settings be developed.

Filippi, V. et al. Women's Reports of Severe (Near-miss) Obstetric Complications in Benin. Studies in Family Planning 31(4):309–324 (December 2000).
This study in Benin, West Africa, tested whether women accurately report the signs and symptoms of severe and less severe obstetric complications for a retrospective interview-based diagnosis of maternal morbidity. Three groups of women (with severe complications, with mild complications, and with a normal delivery) were identified through hospital records and interviewed at home. The women's responses were compared with their hospital records to assess the validity of their recall. Women with a severe, life-threatening complication were identified as "near miss" cases. The questionnaire was fairly accurate in detecting eclamptic seizures, abnormal bleeding in the third trimester for a recall period of three to four years, and all episodes of hemorrhage for a recall period of two years. Questions about dsytocia and genital tract infections produced less accurate responses, except when information about treatment was included. Better results were achieved for antepartum and acute events. Samples from different groups of women would produce different information, especially given the differences evident between lay and clinical perceptions of health. However, the authors conclude that community survey interviews could be useful to assess eclampsia and bleeding in developing countries. This information could indicate the relative importance of complications in order to establish program priorities and interventions.

Font, F. et al. Maternal mortality in a rural district of southeastern Tanzania: an application of the sisterhood method. International Journal of Epidemiology 29:107–112 (2000).
In this study, 4,734 women in the Morogoro Region of Southeastern Tanzania were interviewed using the sisterhood method to estimate maternal mortality. The resulting maternal mortality ratio of 448 deaths per 100,000 live births is much higher than the Tanzanian government's estimate for the region, but much lower than the levels estimated by WHO and UNICEF. Maternal deaths accounted for 19 percent of all deaths among women age 15 to 49. The sisterhood method offers a low-cost way to estimate maternal mortality in this small, rural area where specific health information is lacking.

Gichangi, P. et al. Rate of cesarean section as a process indicator of safe-motherhood programmes: the case of Kenya. Journal of Health, Population and Nutrition 19(2):52–58 (June 2001).
Using data available through the routine data-collection system of the Ministry of Health in Kenya, this study assessed the rate of cesarean section as an indicator of obstetric care in the country. Data were gathered from census reports, WHO/UNICEF, and World Bank publications, and from the Kenya Demographic and Health surveys. In Nairobi, data were collected from the teaching and referral hospital, from the largest maternity hospital in Kenya, and from a private hospital. Overall, there were no significant changes in the rate of cesarean section between 1994 and 1998 in the district hospitals. The hospital-based cesarean rate was 6.3 percent in 1997, varying from 37.7 percent in urban Nairobi, to 0.3 percent in remote Isiolo. The overall population-based cesarean rate was 0.95 percent. In two districts, data from the health-information system were compared with data from hospital operating theatre logbooks, and showed the two to be closely correlated. This study indicates that the data routinely collected by the Ministry of Health in Kenya can be used to evaluate the delivery of maternal health services. Based on the national health system, the cesarean section rate does not exceed 2 percent for the country, reflecting limited availability of obstetric care. Cesarean section offers an easy-to-measure indicator, but overall data collection needs to be strengthened at the local level.

Goodburn, E.A. Using process indicators to monitor and evaluate obstetric services in developing countries. Journal of the American Medical Womens Association 57(3):145–148 (2002).
Use of the U.N. process indicators for monitoring emergency obstetric care helps identify inadequacies in maternal health care using data available from health registers and other health records. The indicators have been put to use in Malawi, Morocco, Indonesia, and Nepal. They are most useful when used as a set or in progression, and can help determine if the main problem relates to access or to quality of care. However, they do not provide information on where facilities are needed, why facilities are not used, or if women are being managed correctly when they come for care. Other types of research are needed to answer these questions, and to better understand quality of care. The indicators are useful in improving maternal record-keeping, and will help integrate obstetric care information into routine health information systems.

Goodburn, E.A. et al. Monitoring obstetric services: putting the UN guidelines into practice in Malawi. I: developing the system. International Journal of Gynecology & Obstetrics 74:105–117 (2001).
The Malawi Safe Motherhood Project is the first large project to introduce the UNICEF/WHO/UNFPA Guidelines for Monitoring the Availability and Use of Obstetric Services as part of a routine government monitoring system. This article documents the challenges and the lessons learned during the development of the new system. The implementation process, which took one year and cost US$100,000, included needs assessment, tools development, operations research, field testing, and training. The project found that the UN Guidelines required substantial adaptation in order to be applied to a routine monitoring system based on the national Health Information Systems periodic reporting system. Introduction of new monitoring indicators for maternal health was done methodically and included widespread consultation, systematic clarification of definitions, rigorous testing, and simplification and testing of established systems. Several technical problems were encountered in calculating the denominators and numerators for the utilization indicators. Systems of routine recording, reporting, and analysis of maternity services data need to receive greater focus, and continued supervision and training are required of those collecting the data. The process indicators derived from routinely collected data will be valuable for measuring the implementation of improved maternity services.

Graham, W.J. Now or never: the case for measuring maternal mortality. Lancet 359(9307):701–704 (February 23, 2002).
The author of this Viewpoint article argues that the lack of data on maternal mortality is both part of the problem and vital to the solution. Reliable data on the effectiveness and cost-effectiveness of the various strategies now being implemented in developing countries are needed. While technical obstacles make the collection of accurate data difficult, strong leadership is needed to obtain rigorous data. The lack of resources is a cause and an effect of inadequate data. With few large-scale efforts to address maternal mortality, data to demonstrate impact are limited, and few donors want to spend scarce resources on measuring the consequences of programs. Evidence-based decision making requires quality evidence (data). For example, although skilled attendance at delivery is being promoted as a way to reduce maternal mortality, there are no data showing a causal relationship. Sustainable evaluation capacities must be established at national levels to provide the data necessary to choose the most effective programs. The Initiative for Maternal Mortality Programme Assessment (IMMPACT) has begun to develop methods for achieving these goals.

Graham W et al. The familial technique for linking maternal death with poverty. Lancet. 2004;363(9402):23–27.
Using data from 11 household surveys (DHS) in ten developing countries, the authors assessed any associations between women’s poverty status (defined by educational level, source of water, type of toilet and floor) and survival. As poverty level increased, the proportion of women dying of non-maternal causes generally increased, and the proportion dying of maternal causes increased consistently. Existing data can be used to reveal the relationship between maternal death and poverty, and expose the rich-poor gap.

Graham, W.J. et al. Demonstrating programme impact on maternal mortality. Health Policy and Planning 11(1):16–20 (1996).
This article explores the problems in measuring levels and trends in maternal mortality, including the infrequency of maternal deaths in some settings, underreporting, the difficulty of determining the cause of death, and the lack of indirect demographic estimation techniques. Given the difficulty of measuring maternal mortality, the authors argue that it is not a feasible indicator for assessing safe motherhood initiatives. Instead, they conclude that safe motherhood programs should be assessed on their operational performance and on improvements in other health outcomes.

Graham, W. and Airey, P. Measuring maternal mortality: sense and sensitivity. Health Policy and Planning 2(4):323–333 (1987).
This article reviews problems in measuring maternal mortality and explores their program implications. The authors criticize maternal mortality measures for focusing exclusively on the risk of dying during pregnancy and delivery without considering the risk of becoming pregnant. After reviewing data on maternal mortality, they recommend targeting limited health resources to women in their prime childbearing years rather than to women at higher risk of maternal death because of age or parity.

Graham W, Hussein J. Measuring and estimating maternal mortality in the era of HIV/AIDS. Presented at: Workshop on HIV/AIDS and Adult Mortality in Developing Countries, August 26, 2003; New York. UN/POP/MORT/2003/.
This paper explores the difficulties presented by measuring maternal mortality in the context of HIV/AIDS. This includes both the effects of HIV/AIDS on the risk of maternal death, as well as the effects of HIV/AIDS on measuring maternal deaths. The former includes a number of biological and behavioral synergies that occur and influence women’s risk both negatively and positively. These act at the individual level (HIV/AIDS as direct cause of death) and also affect all pregnant or reproductive age women (reduced quality of care in overburdened health care system, reduced risk of conception). HIV/AIDS affects the already difficult task of measuring maternal deaths by influencing the classification of death, the quality of data collected, and the use and interpretation of the data collected on maternal mortality.

Høj L et al. Maternal mortality: only 42 days? BJOG: An International Journal of Obstetrics and Gynaecology. 2003;110(11):995–1000.
This prospective study in Guinea-Bissau followed 15,884 women of childbearing age with biannual home visits for six years. There were 14,257 pregnancies and 350 deaths. All pregnancies were self-reported, and all deaths were investigated by interviewing the woman’s relatives. Eighty-two deaths occurred in the first 42 days following miscarriage/delivery; 16 women died 43–91 days post pregnancy; another 16 died during the period 92–182 days; and 18 died between 183 and 365 days. In areas where living conditions are difficult, the health of women post-pregnancy is affected for longer than 42 days. Extending the WHO definition of maternal mortality to encompass deaths within three months of delivery might be more accurate and could increase maternal mortality rates by 10–15 percent.

Høj, L., et al. Maternal mortality in Guinea-Bissau: the use of verbal autopsy in a multi-ethnic population. International Journal of Epidemiology 28:70–76 (1999).
This study attempts to develop standard criteria for a verbal autopsy (postmortem interview) that can be used by medical personnel with basic training to determine the cause of death. This data on magnitude and cause of death are important for the development of programs to reduce maternal mortality, especially in areas with limited medical facilities and no death registry. A structured interview was applied to all deaths of women of fertile age in a cohort of 10,000 women living in 100 clusters in Guinea-Bissau, Africa, over a six-year period. Of the 350 deaths of women of reproductive age, 32 percent were judged to be maternal. Seventy percent of the deaths were given a specific diagnosis, with hemorrhage (42%), obstructed labor (19%), and puerperal infection (16%) being the most common. Analysis shows that a diagnosis was more likely if the respondent was present during the last illness, and if the death occurred less than one week after delivery. Contrary to what was expected prior to the study, husbands proved better informants than co-wives, and men in general provided more specific information than women. The study concludes that if husbands had been the only informants, only 14 percent of the deaths would have remained unclassified (versus the 30 percent found in the study). The successful application of verbal autopsy requires a good knowledge of local culture, beliefs, and disease phraseology. Verbal autopsy should be a low-cost, viable tool for use by health care personnel with minimal training.

Hussein, J. et al. Monitoring obstetric services: putting the UN Guidelines into practice in Malawi: 3 years on. International Journal of Gynecology & Obstetrics 75:63–73 (2001).
In 2000, three years after implementation of a new system of monitoring for maternal health, the Malawi Safe Motherhood Project documented improvements in the availability, utilization, and quality of obstetric care services. This is the first report from a large-scale project utilizing the process indicators recommended in the "UN Guidelines for Monitoring the Availability and Use of Obstetric Services." Participating in developing the new monitoring system created a sense of ownership and interest in analyzing and using the data, which has aided in implementation of the new system. The system has provided data that are of immediate relevance to service providers, managers, and policy makers. While there have been difficulties in ensuring that obstetric complications have been recorded correctly, the data make it possible to see short-term changes and identify immediate needs in service delivery. The authors caution that some of the improvements shown may be due to project inputs and better recording of data. While the availability of basic emergency obstetric care facilities has improved in the country, it is still inadequate.

John Snow, Inc. Safe motherhood indicators: lessons learned in measuring progress. MotherCare Matters 8(1):1–28 (May 1999). Available at: www.jsi.com/intl/mothercare/PUBS/mcmatters/pdf/Vol8%201.pdf.
MotherCare organized a workshop in June 1998 to assess selected maternal health process indicators; the presentations and debates are presented in this article. The most promising indicators for measuring access to and use of services include met need for essential obstetric care; unmet obstetric need; cesarean section rate; who delivers the woman, and where the delivery takes place. Indicators that are useful to assess quality of care are case fatality rate (and number of maternal deaths) and referral rates. The discussion includes information on how to calculate, use, and interpret each of these indicators, along with data sources, and examples from specific countries. The most important data source for access and quality of care indicators is the birth or delivery room register. However, these registers need to be revised to include specific information on complications, woman's home, and referring source. The cesarean section rate and information on where the woman delivers and with whom often can be gathered through existing data-collection systems or through surveys, and can give some indication of access. Met and unmet obstetric need are most usefully applied at the district or provincial level. The case fatality rate and referral rate need more field testing, and other quality-of-care indicators are needed to assess project and program impact.

John Snow, Inc. Assessing safe motherhood programs—India and Bangladesh. MotherCare Matters 7(1):1–20 (May 1998). Available at: www.jsi.com/intl/mothercare/mcmatters/mcm7_1.htm.
This issue discusses MotherCare's efforts in assessing Safe Motherhood programs in India and Bangladesh. It presents the achievements and shortcomings of both programs, identifies lessons learned, and recommends strategies to improve future Safe Motherhood and reproductive health interventions in each country as well as in a broader context.

Khosla, A.H. et al. Maternal mortality and "near-miss" in rural north India. International Journal of Gynecology & Obstetrics 68:163–164 (2000).
This review of maternal mortality and morbidity in rural north India during 1998 found 31 maternal deaths and 224 cases of life-threatening health emergencies. Lack of antenatal care, and deliveries and induced abortions by untrained village midwives (dais) were associated with both deaths and near-misses. Government policies and actions are needed to provide basic health care facilities, and stop induced abortions by unqualified people.

Kilonzo, A. et al. Improving surveillance for maternal and perinatal health in 2 districts of rural Tanzania. American Journal of Public Health 91(10):1636–1640 (October 2001).
As part of a community-based reproductive health project, village health workers in rural Tanzania were trained to collect data on maternal and perinatal health during health education visits to pregnant and postpartum women. Any deaths identified were tracked and displayed on a community monitoring board. Among the 904 pregnancies, there were 4 maternal deaths (maternal mortality ratio of 460 per 100,000 live births), and 69.4 fetoneonatal deaths per 1,000 live births. This community-based surveillance system more accurately assessed deaths than facility-based systems, and the results can help program managers understand the causes of poor pregnancy outcomes, identify areas for improvement, and monitor the effectiveness of interventions. Such systems can be used at the community and district levels to assess trends, but sustainability of the system depends on community commitment and support.

Kodio, B. et al. Levels and causes of maternal mortality in Senegal. Tropical Medicine and International Health 7(6):499–505 (June 2002).
A direct, community-based assessment of maternal deaths using verbal autopsy in three unique groups in rural Senegal found high levels of maternal mortality, but lower than those estimated by WHO and UNICEF. The study obtained records of all deaths of women and births for periods of 14, 13, and 10 years in Niakhar, Mlomp, and Bandafassi, Senegal. Relatives of the women who died were interviewed and causes of death were assigned by three physicians. The maternal mortality ratios in Mlomp (436 deaths per 100,000 live births) and Niakhar (516 deaths) were similar, while in the more remote area of Bandafassi the ratio was 852 deaths per 100,000 live births. Two-thirds of the maternal deaths were from direct obstetric causes, primarily hemorrhage. The authors conclude that demographic surveillance systems are useful for the measurement of maternal mortality, as long as special studies are done to determine causes and levels of maternal deaths. Not surprisingly, the highest levels of maternal deaths in this study are found in the very remote areas with limited health infrastructure.

Lech, M. and Zwane, A. Survey on maternal mortality in Swaziland using the sisterhood method. Paediatric and Perinatal Epidemiology 16:101–107 (2002).
Using data from the 1993–1994 Multi-Purpose Household Survey in Swaziland, this study estimated maternal mortality for the country. Using the sisterhood method, adults in the households (4,371) being interviewed were asked about their own sisters. Based on responses to questions about numbers of sisters surviving to adulthood and numbers dying of pregnancy-related causes, maternal mortality was estimated. This method indicated a maternal mortality ratio of 229 deaths per 100,000 live births for the period 7–8 years before the survey. Not surprisingly, this figure exceeds the official ratio of 125 deaths based on facility records.

Leete, R. Issues in Measuring and Monitoring Maternal Mortality: Implications for Programmes. Technical and Policy Paper No. 1, UNFPA (1998).
This technical report discusses important issues in measuring and monitoring maternal mortality. The report is written for those concerned with formulating, monitoring, and evaluating programs to reduce maternal mortality to better understand the scale and frequency of maternal mortality in populations, to recognize the problems of measuring maternal mortality, and to suggest indicators that can be used to monitor progress. Topics covered include frequency of maternal deaths, survey- and model-based estimates, and process indicators of maternal health.

Maine, D. et al. Guidelines for Monitoring the Availability and Use of Obstetric Services. New York: UNICEF/WHO/UNFPA (October 1997).
This manual offers two approaches to monitoring progress toward reducing maternal mortality. The traditional approach to measuring the impact of programs on maternal mortality is hindered by the now well-known methodological problems (infrequency of deaths, underreporting, misreporting) posed by using maternal mortality rates and ratios. The manual offers an alternative approach based on monitoring the interventions aimed at reducing maternal mortality. It outlines a series of process indicators that assess the accessibility, use, and quality of obstetric services, and provides guidance on data collection and interpretation. These indicators include the number of essential obstetric care (EOC) facilities, their geographic distribution, the percentage of women with complications delivering in EOC facilities, the cesarean section rate, and the case fatality rate (a measure of quality of care).

Milne L, Scotland G, Tagiyeva-Milne N, et al. Safe motherhood program evaluation: theory and practice. Journal of Midwifery and Women’s Health. 2004;49(4):338–344.
Although there has been much discussion of which indicators to use to evaluate safe motherhood programs, little has been mentioned about the application of different approaches to evaluation. This article reviews current theories of evaluation and provides example of their use. Most safe motherhood program evaluations use multiple methods without explaining the rationale for their choice. While several approaches are necessary to evaluate complex interventions, the needs of those using the results can help guide selection of evaluation questions to be answered. Disseminating results, along with a discussion of the approach, why it was selected, and relative strengths and weaknesses can help improve assessments of safe motherhood programs.

Mswia, R. et al. Community-based monitoring of safe motherhood in the United Republic of Tanzania. Bulletin of the World Health Organization 81(2):87–94 (2003).
Maternal mortality in the United Republic of Tanzania was monitored using sentinel demographic surveillance of 77,000 reproductive age women, and prospective monitoring of mortality at an urban site (Dar es Salaam), a poor rural district (Morogoro District), and a wealthier rural district (Hai District). From 1992–93 to 1999, maternal mortality rates (maternal deaths per 100,000 women of reproductive age) declined in the three study areas. Maternal mortality ratios were higher in each of the study areas than those estimated by official, facility-based statistics. The most significant decline was seen in Dar es Salaam, where deaths rates fell by 72 percent (from 134 to 37). Most of the deaths could have been prevented by adequate medical care. The data also showed that an additional year of education for household heads was associated with a 62 percent lower risk of maternal deaths at the community level, after controlling for other factors such as proportion of home births and occupation. This study shows that community-based sentinel surveillance offers a cost-effective and accurate way to determine estimates of mortality.

Murray, S. et al. Tools for monitoring the effectiveness of district maternity referral systems. Health Policy and Planning 16(4):353–361 (2001).
Effective referral systems are necessary to reduce maternal mortality and severe morbidity. Even where such systems exist, it is difficult to evaluate their effectiveness. This study used a variety of tools to review pregnancy-related referrals in Lusaka, Zambia. The information collected provided information about the distribution of births across different facility levels, population coverage, use of essential obstetric care (EOC) facilities by women with complications, progress toward reduction of maternal mortality by referral facility, inappropriate use of EOC level facilities, and perinatal outcomes at peripheral facilities. Most births in Lusaka are covered by public-sector services. Less than a third of the 2820 referrals to the University Teaching Hospital (UTH) during a two-month period fit the U.N. definition of a "complicated case" requiring essential obstetric care. However, the need for EOC is met in an estimated 68 percent of cases in Lusaka. Around ten percent of women in labor in urban Lusaka clinics are referred to UTH, and eight percent of deliveries at UTH are self-referred. Most of these data come from routinely collected information, and can be incorporated into routine monitoring by district health managers. The selection of tools and data need to be made locally, but should examine different aspects and levels of the referral system.

Nilses, C. et al. Self-reported reproductive outcome and implications in relation to use of care in women in rural Zimbabwe. Acta Obstetricia et Gynecologica Scandivica 81:508–515 (2002).
Health interviews were held with 1,213 women aged 15–44 years in villages in rural Zimbabwe to assess morbidity, use of health care services during pregnancy, and pregnancy outcomes. According to these self-reports, 94 percent of women with completed pregnancies attended antenatal care during their last pregnancy, and average time of booking was 4.6 months. The women preferred institutional deliveries, with hospitals (58%) and clinics (27%) the most common places for delivering. Twenty-seven percent of women reported a complication during pregnancy, and most complications were taken care of at health care institutions. Prolonged labor was the most common complication reported during delivery. The cesarean section rate was 6.3 percent. While there are limitations to the validity of self-reported data, these findings offer womens perspectives on the care they receive during pregnancy and delivery. Because so many women deliver outside of institutions in developing countries, self-reported data can be a valuable resource for program and needs assessments.

Olsen, B.E. et al. Estimates of maternal mortality by the sisterhood method in rural northern Tanzania: a household sample and an antenatal clinic sample. British Journal of Obstetrics and Gynaecology 107(10):1290–1297 (October 2000).
This study of maternal mortality in the northern Mbulu and Hanang districts of Tanzania surveyed households and women at an antenatal clinic to estimate the maternal mortality ratio and lifetime risk of maternal death. The study was carried out in 1995 and 1996 using the indirect sisterhood method of estimating maternal mortality. The risk of maternal death was 362 per 100,000 live births in the household survey, and 444 in the antenatal clinic survey. The lifetime risks of maternal death were 1 in 38 in the household survey, and 1 in 31 in the antenatal clinic survey. The risk of death was lower for women attending antenatal clinics closer to the hospital than for those attending clinics further away. There was no significant difference between the results based on the household and antenatal clinic data. The authors conclude that, using the indirect sisterhood method in health facilities, it may be possible to obtain a rapid assessment of local maternal mortality levels at low cost.

Pathak, L.R. et al. Process indicators for safe motherhood programmes: their application and implications as derived from hospital data in Nepal. Tropical Medicine and International Health 5(12):882–890 (December 2000).
Hospital data from five districts in Nepal (Rupandehi, Baglung, Kailali, Okhaldunga, and Surkhet) were analyzed to assess the availability, use and quality of obstetric care between 1997 and 1998. Following the guidelines issued by UNICEF, WHO, and UNFPA (Guidelines for Monitoring the Use and Availability of Obstetric Services, UNICEF et al. 1997), this study identified tremendous underuse of maternity services in Nepal. The number of comprehensive emergency obstetric care facilities (EOC) was adequate in four of the five districts, but none had the minimum acceptable level of basic EOC facilities. Rupandehi, with 21.5 percent of births in an EOC facility, met the guideline's minimum acceptable level for proportion of births in a hospital (15%), while the other four districts were below 5 percent. The "met need" for obstetric care was 14.9 percent in Rupandehi, and less than 5 percent in the other districts. The cesarean section rate was low, between .2 percent and 1.4 percent. The case fatality rate was 4.0 percent in Rupandehi Zonal Hospital, the only comprehensive EOC facility in the five districts. These indicators will help the Nepal Ministry of Health monitor and improve obstetric care, and should only be used in a series to measure progress.

Prual, A. et al. Severe maternal morbidity from direct obstetric causes in West Africa: incidence and case fatality rates. Bulletin of the World Health Organization 78(5):593–602 (2000).
Data on maternal morbidity help determine how many women are likely to need essential obstetric care, and can be used to monitor and evaluate safe motherhood programs. This multicenter, prospective, population-base study measured the incidence of maternal morbidity in a cohort of 20,326 women in six West African countries. Each pregnant woman was contacted four times: at inclusion in the study; between 32 and 36 weeks of amenorrhea; during delivery; and 60 days postpartum. Overall, there were 6.17 cases of direct severe maternal morbidity per 100 live births, although the ratio varied significantly from 3.01 percent in Bamako, Mali to 9.05 percent in Saint-Louis, Senegal. The main causes were hemorrhage (3.05 cases per 100 live births), obstructed labor (2.05 cases per), hypertensive disorders of pregnancy (.64 cases per), and sepsis (.09 cases per). At least 3 to 9 percent of pregnant women required essential obstetric care. There were about 30 severe obstetric complications for every maternal death. Study comparisons are hindered by: (1) the lack of a standardized definition of maternal morbidity; (2) the classification of severe morbidity depends on the level of health services; and (3) some life-threatening conditions depend on the level of care (e.g., uterine rupture) while others are less dependent (e.g., eclampsia). This study documents an unsatisfactory quality of maternal health care even in large urban settings with good access and high utilization.

Robinson, J. and Wharrad, H. The relationship between attendance at birth and maternal mortality rates: an exploration of United Nations data sets including the ratios of physicians and nurses to population, GNP per capita and female literacy. Journal of Advanced Nursing 34(4):445–455.
Data from the United Nations were used to perform regression analyses using the number of physicians and the number of nurses per 1000 population, the proportion of births attended by trained health personnel, GNP per capita, and female literacy as independent variables, and maternal mortality rates (MMRs) as the dependent variable. Linear regression analyses showed positive associations for MMRs and the ratio of physicians to population (73 percent, n=136), the ratio of nurses to population (56 percent, n=136), and the proportion of births attended by trained personnel (83 percent, n=118). Multiple regression analysis showed that attendance at birth, GNP, and physicians per 1000 population explained 87 percent of the variation in MMRs. This study cautions that its findings are only as reliable as the data set it uses, but the strong correlation between skilled attendance at birth and MMR indicates the importance of this variable in reducing maternal deaths.

Ronsmans, C. et al. Questioning the indicators of need for obstetric care. Bulletin of the World Health Organization 80(4):317–324 (2002).
In searching for better ways to measure maternal mortality, there has been a shift away from measures of health to indicators measuring use of health care services. Defining, measuring and interpreting indicators is not straightforward. Measuring the met need for obstetric care depends on accurate and uniform definitions. This paper reviews the indicators of service use and obstetric care currently being used, and discusses issues related to data collection. Although the U.N. process indicators and their expected levels are useful, the authors state these should be supplemented by more precise, clear-cut, and well-defined categories of severe complications. Use of "observed versus expected ratio" (OVER) measures and major obstetric interventions for absolute maternal indications (MOIs for AMIs) can help determine more accurate measures for the local area. Greater involvement of the private sector in data collection and of health professionals in determining definitions will ensure more valid and reliable measures of maternal care.

Ronsmans, C. et al. Use of hospital data for Safe Motherhood programmes in South Kalimantan, Indonesia. Tropical Medicine and International Health 4(7):514–521 (July 1999).
This study uses four indicators of use of essential obstetric care (EOC) and one indicator of quality of care to assess the use of obstetric care at facilities in three districts in South Kalimantan, Indonesia. Overall the findings show low use of services. In the most urban district, Banjar, where facilities coverage is highest, fewer than 14 percent of all deliveries take place in an EOC facility, 2 percent of expected births are admitted with a major obstetric intervention (MOI), and 1 percent of expected births have an MOI for an absolute maternal indication. The use of EOC in Barito Kuala is consistently lower than in other districts, no matter which indicator is used. This study indicates that in settings with low use rates, general rates of EOC facility use appear to be as satisfactory an indicator of access to EOC as more elaborate indicators showing reason for admission. The inequalities in access to care revealed by this study may be a more important stimulus for change than other (often inaccurate) reports of high levels of maternal mortality.

Ronsmans, C. et al. Decline in maternal mortality in Matlab, Bangladesh: a cautionary tale. Lancet 350:1810–1814 (December 20, 1997).
This article examines trends in maternal mortality over an 18-year period in program and comparison areas of Matlab, Bangladesh. Maternal mortality in program areas declined when outreach services by trained midwives and an active referral system were introduced. However, direct obstetric mortality also fell in one of the comparison areas lacking such services, presumably because women in both regions accessed emergency obstetric services. The Matlab data demonstrate the difficulty of evaluating the impact of any short-term intervention, no matter how rigorous the research.

Ross, J.A. et al. The Maternal and Neonatal Programme Effort Index (MNPI). Tropical Medicine and International Health 6(10):787–798 (2001).
The Maternal and Neonatal Programme Effort Index (MNPI) is based on responses to an 81-item questionnaire completed by 10 to 25 observers in 49 countries. Ratings are given for the current year and three years ago. The index assesses maternal and neonatal health services and related areas such as training. The questionnaire is organized by stages in providing care: (1) setting maternal health policy and organizing support, (2) developing facility capacity, (3) expanding access to services for pregnant women, and (4) providing care to pregnant women and newborns. The index shows that country programs score about half of the maximum, with lowest scores for access to treatment for rural women and highest scores for neonatal care. Regionally scores do not differ greatly, but South Asia has especially low scores, while East Asias are quite high. Over the three-year period assessed, scores rose by about 10 percent. When divided into three groups according to maternal mortality levels, most of the 14 components assessed distinguish the high-mortality from the medium-mortality countries, and about half of the components distinguish the medium-mortality from the low-mortality countries. The MNPI offers useful measures for assessing components of program effort and correlates somewhat to maternal mortality levels.

Sloan, N. et al. The etiology of maternal mortality in developing countries: what do verbal autopsies tell us? Bulletin of the World Health Organization 79(9):805–810 (2001).
Verbal autopsies or histories have been used to describe causes of maternal deaths in developing countries. These often are based on interviews of family and community members who observed occurrences related to the death, and may be the only information available, especially for women delivering at home. This study reanalyzed data gathered in a verbal autopsy study in rural Mexico to assess its accuracy for identifying medical cause of death. Complete verbal autopsy information was obtained for 145 out of 164 maternal deaths reported in 1995 in three rural states. The cause of death as identified by various classification systems (death certificate, WHO verbal autopsy flowchart, primary cause verbal autopsy, sole cause verbal autopsy, and mentioned in verbal autopsy) varies widely. This suggests that verbal autopsies have little validity in attributing cause of death to a single specific medical cause. This is consistent with the WHO view that multiple causes of death should be considered in determining program priorities. At best, verbal autopsies may reconfirm that mortality among poor women with little access to medical care is higher than among wealthier women with better access to care.

Songane, F.F. and Bergstrom, S. Quality of registration of maternal deaths in Mozambique: a community-based study. Social Science & Medicine 54:23–31(2002).
This community-based study (covering 207,369 people) in central Mozambique assessed maternal deaths using existing health facility records, interviews with health personnel, and community collaborators to report deaths of women of reproductive age. During the one-year study period (August 1996 through July 1997) there were 204 deaths of women of reproductive age, and 40 of these were classified as maternal deaths. The study revealed significant underregistration of maternal deaths in publicly available sources of information. The provincial directorate of health failed by 87 percent to record maternal deaths. The Civil Register and the Funeral Services only registered 9 percent of the maternal deaths. Case notes were often missing for deaths at institutions, and the quality of the information in the notes was poor. There is need to identify a practical and reliable way to collect information about maternal deaths on a routine basis. In this area, linkages with private cemeteries and the introduction of a recording sheet might be a useful way to identify deaths of women of reproductive age.

Stanton, C. et al. Every death counts: measurement of maternal mortality via a census. Bulletin of the World Health Organization 79(7):657–664 (2001).
This study evaluated the use of a national population census as a way of measuring maternal mortality in five developing countries (Benin, Islamic Republic of Iran, Lao Peoples Democratic Republic, Madagascar, and Zimbabwe). Using standard demographic methods, the data on adult female deaths and births were evaluated and adjusted as necessary. The number of maternal deaths was adjusted relative to the adjustment of female deaths. The data suggest that a national census is a promising tool for measuring maternal mortality. However, it requires careful evaluation and adjustment of the data as necessary. Governments could learn from the example of these five countries, and enlist the support of statistical agencies in measuring maternal mortality. To justify adding maternal mortality to the list of items tracked by the census, a countrys Ministry of Health should also commit itself to using the results for program and policy decision making.

Stanton, C. et al. An Assessment of DHS Maternal Mortality Indicators. Studies in Family Planning 31(2):111–123 (June 2000).
This study assessed the quality of maternal mortality data collected in 14 Demographic and Health Surveys (DHS) in 13 countries. The data for reported deaths were found to be complete for most of the variables studied. However, sibling estimates of deaths are likely to be underestimates. The sampling errors found with maternal mortality ratios is far greater than those associated with other DHS indicators. Because of this imprecision, the authors conclude that the DHS maternal mortality data should not be used for trend analysis, and that the maternal mortality module of the surveys should not be used more often than every ten years in one country.

Stewart, M.K. et al. Issues in measuring maternal morbidity: lessons from the Philippines Safe Motherhood Survey Project. Studies in Family Planning 27(1):29–35 (January–February 1996).
This review of the 1993 Philippines Safe Motherhood Survey examines whether surveys can gather reliable population-based data on maternal morbidity. Respondents were diagnosed with specific obstetric complications based on their reports of symptoms experienced during labor and delivery. A validation study comparing hospital delivery records with interview results found that women more accurately reported some obstetric complications (including eclampsia and sepsis) than others. The survey found that 12 percent of women giving birth over the past three years experienced symptoms of at least one complication. The authors concluded that, while surveys require considerable resources, they may be the only way to measure the prevalence of maternal morbidity.

UNFPA. Making Safe Motherhood a Reality in West Africa: Using Indicators to Programme for Results. New York: UNFPA (2003).
This report summarizes results from needs assessments undertaken in Cameroon, Côte d'Ivoire, Mauritania, Niger, and Senegal using the UN emergency obstetric care process indicators. With support from UNFPA and the Averting Maternal Death and Disability program (AMDD), each country collected data for each of the process indicators. The report provides data on each indicator for each country as assessed in April 2001, and highlights interventions being developed to address the inadequacies identified. Future reports will summarize project interventions.

Wang C et al. Reproductive health indicators for China’s rural areas. Social Science & Medicine. 2003;57(2):217–225.
Reproductive health indicators were developed for use in rural China’s low-, middle-, and high-income counties using two participatory techniques: nominal group process and Delphi survey. Guided by the criteria of practicality, feasibility, and measurability, 55 grassroots reproductive health workers developed a final list of 126 indicators. These were then prioritized through a Delphi process among 103 nationally and internationally known reproductive health experts. This process resulted in a list of 21 community-based reproductive health indicators and 8 determinants for measuring achievements for rural China. These were field tested in several sites, and only one determinant was unobtainable due to lack of records. While several of these replicate the UN Indicators, the authors speculate that some of the UN Indicators were not included because it would be infeasible to obtain the data or were not appropriate to Chinese culture. Due to the unique sociocultural characteristics of areas, the indicators identified in this study may require further adaptation but provide another comprehensive way to evaluate reproductive health care.

Wardlaw, T. and Maine, D. "Process Indicators for Maternal Mortality Programmes." In: Berer, M. and Ravindran, T.K. eds. Safe Motherhood Initiatives: Critical Issues (1999).
In order to monitor progress of the Safe Motherhood Initiative, the authors argue that indicators that show change in a relatively short period of time (3 to 5 years) are needed, and that these indicators should be based on data that are relatively easy and inexpensive to collect. Guidelines published by UNICEF, WHO, and UNFPA present six process indicators that fulfill these requirements (see Maine et al. 1997). Use of these indicators will help governments, agencies and communities rally to reduce maternal deaths.

World Health Organization (WHO). Beyond the Numbers: Reviewing Maternal Deaths and Complications to Make Pregnancy Safer. Geneva: WHO; 2004. Available at: www.who.int/reproductive-health/publications/btn/btn.pdf.
Difficult as it is to get accurate data on maternal deaths and illness, it is important to also understand the underlying causes of these deaths, why they happened, and how they can be prevented. This guide includes information on how to gather this type of information and what it reveals. Different approaches (surveys, vital records, disease surveillance, verbal autopsies, facility-based death reviews, confidential inquiries, clinical audits, and severe morbidity interviews) are discussed, along with their advantages and disadvantages. This guide offers health program directors, researchers, and providers a tool to gather data, better understand underlying factors, and highlights the importance of acting on this knowledge.

Top of page