Appendix 2 - Country examples of safe motherhood programs

· Bangladesh - Posting Trained Midwives in Rural Bangladesh: The Matlab Maternity Cam Project

· Bolivia - Improving Maternal and Perinatal Health in Urban and Periurban Cochabamba: Project Summary and Projected Costs

· Brazil - Training and Learning from Traditional Birth Attendants in Northeast Brazil: The PROAIS Project

· Cameroon - Risk Screening and Referral

· Ethiopia - Community Obstetrics and Maternity Waiting Homes

· The Gambia - Training Traditional Birth Attendants: The Farafenni Project

· Grenada - Public Sector Maternity Care: Summary and Associated Expenditures

· Guatemala - A Rural Case Management Strategy: The Quetzaltenango Project

· Indonesia - Regionalization of Care in West Java

· Tunisia - Strengthening and Expanding Family Planning and Maternal and Child Health Services within the Basic Health Care System

· Zaire - The Karawa Health Zone Project

· Zambia - Integrating Prenatal Care with Syphilis Control

· Zimbabwe - Expanding Access to Maternity Care

Bangladesh - Posting Trained Midwives in Rural Bangladesh: The Matlab Maternity Care Project

The Matlab district of Bangladesh is located 40 miles southeast of Dhaka, the country's capital, in the rural, flood-prone Ganges-Meghna delta region. Eighty five percent of the area's 200,000 people are Moslem, and the active practice of purdah is the norm. As such, women's mobility is limited to the household compound and they have minimal exposure to adult men other than relatives. Female literacy is only 17 percent. Women are virtually invisible in labor statistics, though their multiple responsibilities include childcare, maintenance of the physical household structure, and food processing and preparation. Female children receive less food and health care within the family than their brothers, and are often chronically undernourished.

Nearly 80 percent of women are married by age 20. Suicide and homicide are common outcomes of illegitimate pregnancies. Talking about reproductive issues is considered inappropriate, limiting access to care when problems arise. Matters related to the female genital tract are associated with shame, though women may exchange experiences freely among themselves, particularly within their compound. Walking and rickshaws, when available, are the main modes of transport to health facilities, even in emergencies. Transport by boat is also common, but boats are often not available, particularly after dark. Night travel is further limited by lack of electricity and taboos associated with women leaving home at night, especially when pregnant. Twenty five percent of all adult female mortality is due to the direct complications of pregnancy and childbirth.

Since 1966, the International Center for Diarrhoeal Disease Research/Bangladesh (ICDDR/B) has operated a rigorous demographic surveillance system in Matlab. In the late 1970s, the area was divided into an intervention area, where a maternal and child health and family planning (MCH-FP) project was implemented, and a control area, which receives government health services only. Female Village Health Workers (FVHWs) provide a full range of contraceptive methods in the intervention area through home-based delivery, monitor and manage adverse effects, provide a range of child health services, and refer women and children to MCH-FP outposts staffed by paramedics or to the Central Matlab clinic, when necessary. Contraceptive prevalence rates increased from eight percent in 1977 to 56 percent in 1989. The provision of family planning services in the area was responsible for a reduction in the number of pregnancies and, consequently, a 57 percent reduction in the number of pregnancy-related deaths. The risk women faced when pregnant, however, remained high, as reflected in a maternal mortality ratio of 550 per 100,000 live births. In addition, unsafe, induced abortion remained a primary cause of death. As part of the MCHPP project, the FVHWs also provide women with safe delivery kits and iron tablets, and traditional birth attendants are trained in hygienic delivery practices. While these interventions have helped to reduce neonatal mortality, which was their primary objective, their impact on maternal mortality has been minimal.

A retrospective study of maternal mortality in Matlab found that the main causes of death, in order of importance, are unsafe induced abortion, postpartum hemorrhage (PPH), toxemia, obstructed labor, and post-partum sepsis. Ninety-five percent of all deliveries, and 8.0 percent of all deaths, occur at home. Most deaths occur during or within 48 hours of labor and delivery. The results of this study, and the failure of the MCH-FP project to improve maternal outcomes, prompted the development of a home-based maternity care project.

Professional nurse-midwives were posted at the community level in an effort to ensure timely intervention in complicated pregnancies and deliveries and a functioning referral system. Access to emergency obstetric surgery and blood transfusion in the area is limited to the government district hospital, one hour away. Prior to the introduction of the project, the Matlab clinic had no emergency obstetric capacity. The nearby Chandpur Red Crescent Hospital remains better equipped with an obstetrical specialist on staff. The Matlab MCH-FP area was divided into a control and intervention district, comparable in terms of socioeconomic, demographic, and specific health indicators. One nurse-midwife was posted in each health outpost, serving a population of approximately 20,000. The midwives, who had received three years of standard government nursing training and one year of midwifery training, were given a brief orientation but no projectspecific training. All were from a rural background and had experience working in traditional communities. They were provided with a standardized midwifery kit, as well as with antibiotics, heavy sedatives, infusions and plasma expanders, and pitocin, to be administered intra-nasally. The focus was on supplementing, rather than supplanting, the work of traditional birth attendants, who would remain responsible for managing deliveries as far as possible. The midwives were responsible for making prenatal visits to establish rapport with pregnant women, provide information, detect and manage prenatal problems, and screen for potential future complications; encouraging labor calls and attending as many deliveries as possible; providing immediate treatment for complications in labor and delivery, when possible; organizing referral and accompanying the patient to the Central Matlab Clinic, when necessary; and visiting new mothers as soon as possible after delivery. There were about 1600 pregnancies a year in the intervention area, or an average of about 33 per month per midwife. The aim was to ensure that the complications that arose received appropriate intervention as early as possible, to prevent progression to a severe stage, given the limitations of existing services to cope with severe complications, and community resistance to referral. Examples of feasible timely interventions the midwives could carry out include early administration of anti-eclamptic drugs; complete evacuation of the uterus in case of an early retained or torn placenta; vaginal packing in cases of PPH; stitching of vaginal tears; infusion of plasma-expanders in cases of hemorrhagic shock; and administration of antibiotics to prevent severe infections. Each midwife was supported by:

· A locally recruited village man who accompanied her, especially at night; transmitted messages; carried equipment and a lantern; assisted in transporting the patient by stretcher or boat; and motivated male members of the community.

· The installation of a maternity clinic at Matlab with limited emergency obstetric capacity, at which female physicians were always available. Clinic physicians supervised the midwives, evaluated and managed referrals for which they were equipped to cope (ea. Dilation and Curettage, management of pre-eclampsia/eclampsia) and ensured timely referral of cases in need of transfusion and surgery to the district hospital. The project was not able to ensure the quality of services at the district hospital level.

· A communications strategy, which aimed to orient the traditional birth attendants, familiarize them with the referral system and care facilities (including visits to these facilities), and introduce the midwives. Efforts were also made to inform and motivate women and their families.

The project's communications element was helped enormously by the results of an anthropological study, which shed light on family level decision-making dynamics: while women are involved in decision-making regarding health and health care, they do not make decisions independently. When complications occur, decision-making roles shift, and the mother-in-law, elder sister-in-law, or husband take charge. Another study found that there was often conflict between young mothers and older family women with regard to appropriate health behavior. As is often found, the mother-in-law tends to be the gatekeeper.

Outcomes were measured by comparing maternal mortality ratios in the control and intervention districts during three years prior to (1984-86) and three years following (1987-89) the implementation of the project. The mortality difference between the two areas prior to implementation was not statistically significant. During the three years after the project was implemented, the difference between the two areas was statistically significant, and the ratio in the intervention area had fallen by 68 percent. In short, the introduction of the maternity care project had a substantial impact on maternal mortality.

The causes of death that were reduced by the project were, in order of importance, the complications of unsafe abortion, PPH, post-partum sepsis, and eclampsia. Other causes of adult female mortality were constant over the project period. Although abortion was not a specific focus of the project, the decrease in abortion-related mortality may have been related to earlier intervention in complications by the midwives. In addition, the midwives may have succeeded in discouraging some women from resorting to dangerous traditional abortion procedures.

Of the 4,884 registered pregnancies, 44 percent of the women were visited at home at least once during pregnancy. Fifteen percent of the pregnant women requested attendance during labor. In 9 percent of cases the midwife herself delivered the baby, and in 4 percent she attended the delivery but allowed the traditional birth attendant or a female paramedic to perform the delivery. In 2 percent of cases, the midwife was on her way. 19 percent of the women attended by midwives were referred to the Matlab clinic. Of the 1,712 women visited postpartum, 3 percent were referred to the clinic.

Some reviewers have questioned the relationship between the small proportion of deliveries attended by the midwives and the reduction in mortality. Although it is not possible to determine exactly which of the home deliveries they performed and which of the patients they referred would have died in their absence, it is legitimate to accept that the averted deaths were drawn from the patients attended or referred by the midwives.

Research was conducted to identify the factors that differentiated women who requested midwife attendance from those who delivered alone or with a traditional birth attendant. Attended women were more often of lower parity, and were more often primigravidae. Women who lived closer to the midwife's residence, women who had received prenatal care from the midwife, women with poor obstetric histories, women with pathologic signs during pregnancy (ea. vaginal bleeding), and women who experienced complications during labor were also more likely to be attended. Contrary to what was expected, users and non-users did not differ significantly in terms of socioeconomic status.

The low proportion of requests for attendance may be related to distance or to the rarity of complications. In addition, it is possible that the reluctance of family decision-makers (husbands, mothers-in-law) to call for external assistance was greater than expected. Many would not call until complications had already arisen. Others would hesitate to call for fear of referral to the district hospital, based on negative past experience and a common perception that quality of care at the facility is poor.

An attempt was made to estimate the cost of integrating the maternity care project into the MCH-FP project, and the relationship of costs to project outcomes. Between 1987 and 1989, detailed monthly cost reports were collected by resource category. The total direct cost of the three-year project was US$85,862. A substantial increase in costs occurred between 1988 and 1989 due to a major revision of salaries and services within ICDDR/B in 1989. Two caveats must be made. Firstly, the costs represent only the additional amount required to build a maternity care project onto a well-established MCH-FP project; and secondly, the costs represent those of a large international organization, with higher salary and other expenses relative to the rest of Bangladesh. In addition, comparative data on alternative health interventions or service delivery mechanisms are not provided, though work in this area is underway.

Replication of the project as currently designed will not be possible in the Bangladeshi context due to insufficient nurse-midwives and lack of national commitment. A modified version of the project, using female paramedics with 18 months training, is being tested. The project could be further improved through a strengthened communications element to improve utilization of midwifery care at delivery and reduce resistance to referral, and by improving the capacity and quality of referral-level facilities. It is clear, however, that ensuring the availability of communitybased maternity care involving professional level midwives and an adequate referral system has the potential to bring about substantial reductions in maternal mortality in Bangladesh.

(Fauveau 1991a; Fauveau and others 1991b; Fauveau and Chakraborty, 1988; Stewart 1991)

Bolivia - Improving Maternal and Perinatal Health in Urban and Periurban Cochabamba: Project Summary and Projected Costs

The goal of the USAID-supported Cochabamba project, which was launched in 1991, is to reduce maternal and perinatal mortality rates in urban and periurban areas of Cochabamba, Bolivia. An estimated 91,000 women of reproductive age, of whom approximately 19,400 become pregnant each year, live in the Cochabamba area, which has an estimated total population of 1,000,000. In 1989, maternal mortality was estimated at 480/100,000 live births and perinatal mortality at 110/1,000.

The Cochabamba project takes a comprehensive approach with a focus on preventive prenatal and postnatal care, institutional delivery care, use of trained personnel for home deliveries, and the provision and promotion of low-cost contraceptive services. The majority of the project's interventions are aimed at raising awareness of maternal health problems and building demand for available services.

Specific project components include: (1) baseline and evaluative research; (2) information, social marketing, education and communication activities and materials development; and (3) inservice training for both public sector and NGO health professionals in the provision of family planning information and services, prenatal assessment, and management of high risk obstetrical cases.

The project is coordinated by and implemented through NGOs already working in maternal and infant health in and around Cochabamba. In addition, a long-term resident advisor provides technical assistance to the overall effort.

The following chart summarizes estimated costs, as indicated in the proposed project budget, for the three year project. The estimates include costs of locally-recruited staff, local contractors, expatriate consultants, travel, per diem, equipment, commodities, and other direct costs.

(MotherCare Project. Taylor and others 1990)

Brazil - Training and Learning from Traditional Birth Attendants in Northeast Brazil: The PROAIS Project

More than 30 million of Brazil's people live in the country's eight dry, desert-like northeast states. Most are very poor; average per capita income is less than US$200. Sixty-five percent of the population and 95 percent of physicians live on a narrow greenbelt bordering the Atlantic ocean. As a result, most of the rural population has little access to formal health services.

In the late 1970s, the late Professor Galba Araujo, then Medical director of the Ceara Federal University (CFU) in Fortaleza, developed a community health program based on the following observations: many of the rural women who came to the city for hospital care did so for problems that did not require such care. Worse, many women suffering severe complications requiring hospital care arrived too late and too ill to be saved.

Birth attendants selected by rural community leaders have been trained to manage rural maternity units, provide prenatal care, and screen and refer women at high risk. They were trained in the use of a risk detection form, and to measure women's weight, blood pressure, and other indicators.

A referral system was developed to ensure the transfer of high risk patients to the CFU maternity hospital, which has been the key to the program's success. Health professionals from the maternity hospital visit weekly, to ensure that the prenatal care provided is adequate and that women at high risk are identified and provided with appropriate care. Most risk factors are identified during these visits.

The traditional birth attendants are trained to refer all women suffering complications. Harmless or beneficial traditional practices are encouraged, and are replicated at Rural Delivery Houses. These include the vertical delivery position, with the delivering woman seated on a traditional birthing stool, and the participation of a family member, usually the husband. Women usually hold a rope hung from the rafters, or wrap their arms around their husband's neck, for leverage. Several types of maternity huts are in operation. The most simple consists of a small mud room attached to the traditional birth attendant's home and equipped with a small bed or hammock. Such units handle two to six deliveries per month and cost US$50 per month to operate. Type B units, which have more than two beds and additional equipment, cost US$100 per month. The most sophisticated unit, with 10 beds, costs about US$200 per month to operate.

Some traditional birth attendants are trained to take PAP smears, perform breast examinations, and detect gynecological cancer by distinguishing between a normal and abnormal cervix. They are also trained to provide information on all family planning methods, distribute barrier methods, and refer women who desire other methods to a nurse or, in the case of the IUD, a physician.

The traditional birth attendants deal specifically with women and their needs. Other traditional health care providers have been trained to undertake child survival activities. Faith Healers, for example, are trained to train mothers in oral rehydration therapy. Their traditional prayers and practices are respected. Community coffin makers assist the rural health program by collecting mortality data. Adolescents also play an active role. The PROAIS project offers them entertainment and sporting activities, as well as providing them with sex education and family planning and encouraging them to contribute to improving village life. Those who are willing and able serve as health agents, providing home follow-up visits for children, and participating in child growth monitoring.

Teams of professors and health science students from the university provide medical care to the villages on an interim basis and supervise high-risk, prenatal, postpartum and perinatal clinics held at the Rural Delivery Units. All health science students are required to work in a rural community prior to graduation, and many select PROAIS. Most have benefited substantially from exposure to community needs and preferences and from the opportunity to combine their formal training with the wisdom of traditional care systems. Indeed, the PROAIS program has not only succeeded in providing remote, rural areas with access to modern medical care, but has also resulted in a transfer of traditional delivery practices, many of which are beneficial, to the formal care system. A modern version of the traditional birthing stool has been developed, and the vertical delivery position, as well as minimal interference with delivery, have been absorbed into hospital obstetric routine. Hospital deliveries in Brazil are often characterized by excessive intervention, as illustrated most dramatically by the fact that 60 to 90 percent of women are delivered by Cesarean Section. The PROAIS project has reduced the incidence of this procedure to 2.4 percent, though in the private wing of the hospital, rates remain high. Immediate breastfeeding, which speeds the delivery of the placenta, is also now encouraged, and the umbilical cord is not cut until the placenta has been delivered, unless it is too short to allow the baby to reach the breast.

A World Bank project in the region includes the development of Rural Delivery Units and is based in large part on the PROAIS model. Continuing dialogue with the Ministry of Health and medical professionals has been essential to overcome continuing unfavorable opinions of traditional birth attendants. While it is clear that enormous gains have been made from training and delegating responsibility to traditional birth attendants in this project, its success has relied on effective referral-level facilities, a functioning referral system, and extensive supervision by health professionals.

(Bomfim 1991; Araujo and others 1993; Janowitz and others 1985)

Cameroon - Risk Screening and Referral

The maternal mortality ratio in the Cameroon, a country of 11.9 million people, is 420 per 100,000 live births. Access to health care is limited for most of the predominantly rural population, with one physician for every 17,466 people. Until three years ago, the government did not have a family planning program. Contraceptive prevalence is very low, at 5 percent. More than 21 percent of the female population is between the ages of 10 and 19, and early marriage and pregnancy are the norm. Teenagers account for more than 22 percent of all births each year, and nearly 18 percent of teenage pregnancies end in induced abortion, which is legally restricted and usually unsafe. The total fertility rate is six, and the average age for grand multiparity is 27.

Since the Nairobi and Niamey Safe Motherhood conferences, many African countries have begun working on the development of strategies to reduce maternal mortality. At the Central Maternity (CM), University Hospital Center (UHC), and 18 private maternities in Yaounde, the capital of the Cameroon, this work began many years earlier. The strategy adopted is based on the risk approach, which requires that all pregnant women be screened for risk factors during the prenatal, intrapartum and postpartum periods, and that those identified as being at high risk receive special surveillance and care from scarce experts, with the objective of concentrating expert attention on those most likely to need such care.

Approximately 85 percent of the 22,575 deliveries in Yaounde each year are estimated to take place in these facilities. Complicated cases are referred primarily to the CM. Between 1973 and 1978, studies were undertaken to determine the characteristics of the women who died in these facilities. The results of these studies were used to develop training courses for all health service staff on the screening and management of high risk pregnancy, including the following elements: detection of risk factors; use of the partogram; identification of contraceptive need; IUD insertion; and the establishment and maintenance of service records. Health service staff were also provided with screening and referral guidelines, and they rotate regularly through different maternity services to ensure that their clinical skills are kept up to date and to promote effective ad hoc deployment when necessary.

Midwives are responsible for screening patients, receiving referrals, and turning them over for specialist management when necessary. Physicians are involved in the selection of women at high risk, and provide back-up support. Sixty percent of all risk factors can be identified in the prenatal period. Teenage pregnancy and grand multiparity combined account for 63 percent of all high risk pregnancies, which in turn account for 27 percent of all pregnancies, and 67 percent of complications in labor and the puerperium.

Studies had found that inadequate care of hospitalized mothers contributed to 54 percent of all deaths. In addition to ensuring that all women at high risk receive appropriate surveillance, all women in labor in the teaching hospitals are now monitored using the partogram, which can be used to detect intrapartum risk. The sensitivity and specificity of the instrument has not been determined, however, and its use, which would be unnecessarily complicated, has not been taught to the midwives. Instead, their clinical judgement is relied upon. The partogram has been found to provide reliable guidance to health personnel in detecting problems in labor. It has been introduced in some maternities in rural areas, and will soon be introduced to others.

Family planning service provision began in 1975 at the CM, and in 1982 at the UHC. Government permission to start these services was granted at a time when official policy was unsupportive of family planning. Midwives are responsible for the provision of family planning services, with specialist backup. Contraceptive prevalence among service users has increased 40 percent, and there has been a decrease in the number of high parity women. The program has succeeded in bringing about significant reductions in maternal mortality. At the UHC, the maternal mortality ratio (MMR) has been maintained at 0 - 0.84 per 100,000 live births. At the CM, maternal mortality fell 40 percent between 1979 and 1989, from 200 to 120 per 100,000 live births. The populations delivering in the CM and UHC are comparable, except in terms of socioeconomic status. Nutritional status is satisfactory in both groups, however. The difference between the two institutions, given the similarities of their patient groups, is thought to be related to organizational factors. The CM has a much higher workload and more limited space, equipment, personnel, etc.

A recent pilot project in six rural communities indicates that the approach can be successfully replicated in rural areas. The program will now be implemented country-wide, with full government support. Research in progress will inform this effort and includes standardizing instruments and equipment; determining the sociocultural factors that affect the acceptability of contraception among women at high risk; and determining the knowledge, attitudes and practices of different groups to be served by the program.

Following are some of the most significant areas thought to be key to the successful implementation of this program on a national scale: ensuring sufficient political commitment, adequate coordination, uniform data collection and the use of standardized instruments; providing adequate monitoring, supervision, and continuing training to service providers; and undertaking information, education and communication activities to sensitize the community, taking into account existing sociocultural and religious taboos.

The advantages of this approach relate primarily to social justice: it helps ensure that women most in need of care have priority access. It also permits the development of an alert system of referral and feedback between the levels of the health system; it entails reorganization and inservice training at all levels; it extends the reach of the health system to the most peripheral rural areas; it is based on prevention rather than cure, which may make it more cost-effective; and it relies on community responsibility and involvement.

(Leke 1991; Nasah and others 1991)

Ethiopia - Community Obstetrics and Maternity Waiting Homes

Attat Hospital, a 55-bed rural community-based establishment in Central Ethiopia, covers a population of 1.5 million. The hospital's catchment area contains 300,000 people who are one to two days's walk from access to transport in case of an emergency.

The hospital's immediate target area for implementation of a primary health care program includes 15 villages with a population of about 15,000 located within a radius of 12 kilometers (equivalent to some two hours walking). Each village has a development committee of five people which meets monthly with a representative of the hospital's public health team. A women's group concerned with social and community development exists in every village as well. These groups have initiated small income-generating projects for women, and constructed wells and toilets with a subsequent decrease in diarrheal disease. Natural family planning methods are gradually being accepted and practiced in the area.

A maternity care system with efficient access to the Attat Hospital was organized through the area's functioning primary health care program. Thirteen traditional birth attendants and 13 village health workers were trained and worked in village health posts. Prenatal clinics are conducted in three villages by the traditional birth attendants and nurse-midwives from the hospital. Pregnant women at high risk are identified and referred to the "tukul", a maternity waiting home located near Attat Hospital, approximately two weeks before their expected date of delivery. High risk criteria include poor obstetric history - preterm labor, operative delivery, fistula, referral for retained placenta, and referral for hemorrhage, puerperal fever; and present pregnancy - hemorrhage, raised blood pressure, malpresentation (transverse/twins/priapara breech), severe anemia, and very young, very short primipara.

The maternity waiting home was built in the style of a local house and provides pregnant women with a temporary residence where they can be observed prior to delivery. All the labor and the majority of construction materials for the home were provided by the community. All women who use the facility are accompanied by a relative, bring their own food and buy firewood locally. While in residence, the women attend the hospital's prenatal clinic and are visited by a hospital nurse once a day. The home contains 15 beds and the average length of stay is 15 days.

During 1987, 151 women were admitted to the maternity waiting home. All the women had received prenatal care either in the hospital or the outreach clinics. Seventy-two women admitted to the home had unfavorable obstetric histories; 39 others had complications during the index pregnancy. The remaining 40 were not identified as being especially high risk (17 grandmultipara, 10 primigravida, 1 lived very far away, 12 unknown) but they wished to be near the hospital when they went into labor.

Of the 72 women with unfavorable obstetric history, one went home prior to delivery and 34 had cesarean sections. Of 15 women with a previously ruptured uterus, seven were delivered abdominally; the other eight were delivered vaginally under close supervision. Of the women with previous stillbirths, all had livebirths. Some of these women had two, three or four prior stillbirths.

There were 13 maternal deaths among women admitted directly from their homes to the hospital (MM ratio: 2,120/100,000 live births) but none among those who entered the maternity home first. Causes of maternal death included ruptured uterus, eclampsia, hepatic coma, severe sepsis and placenta previa. The stillbirth rate for direct hospital admissions was ten times higher than for maternity home admissions (253.5 vs. 28.2 per 1,000 births). Of the four stillbirths which did occur in the maternity home, intrauterine death had occurred in two of the cases prior to their arrival. There were no ruptured uteri nor craniotomies among the women admitted via the home.

(Poovan and others 1990)

The Gambia - Training Traditional Birth Attendants: The Farafenni Project

Estimates of the maternal mortality ratio in The Gambia are among the highest ever documented: 1000-2000 per 100,000 live births. Most of the country's 800,000 people live in small, scattered villages, with limited access to health care. The country's climate is typical of Africa's sub-sahel, with a long dry and short wet season. Women in this predominantly

Moslem country have little decision-making power. They are responsible to the men around them, first their fathers, and then their husbands. Marriage is arranged through contract and bride price. Contraceptive use is rare and total fertility is high, at 6.5 births per woman. Life expectancy for women is 45 years.

Following the Alma Ata Conference that launched the global "Health for All" strategy in 1978, the Government of The Gambia decided to reorganize its health system based on the primary health care (PHC) model. Village elders were asked to select one male and one female to be trained as a village health worker (VHW) and birth attendant. Most villages with over 400 people are now served by a government-trained birth attendant.

The Government of the Gambia asked the United Kingdom Medical Research Council (UK MRC), which has had laboratories in The Gambia for fifty years, to evaluate the impact of its PHC scheme, including the effect of traditional birth attendants on pregnancy outcomes. The UK MRC's evaluation was undertaken in a rural area on the north bank of the River Gambia, near the town of Farafenni, and began with a year's collection of baseline data in 1982-83. This was followed by three years of data collection after the PHC scheme had been implemented in the area.

After an initial period of consultation with village leaders, VHWs and birth attendants were selected for training. The birth attendants, most of whom were elderly, illiterate, and already served as untrained birth attendants, received the standard, 10-week government training course. They were supplied with birth kits, which include clean dressings, scissors, string, oral ergometrine and disinfectant. Traditional birth attendants conduct deliveries in their homes, and advise women on prenatal and postnatal care. They also refer and accompany women with complications to the health center, where they can be delivered by a trained midwife. Women in both PHC intervention and non-PHC control villages have access to health centers and trained midwives, but the health centers have no surgical or blood transfusion capacity. The nearest hospital equipped with such facilities is 200 miles away, and can only be reached after crossing the river on an unreliable ferry.

The population of the area at the midpoint of the survey was 13,780, including 2,738 women of reproductive age. Data were collected using three methods: (1) monthly morbidity surveys during pregnancy; (2) three cross-sectional surveys following confirmation of pregnancy (using urine tests), performed at the end of the dry season and at the end of the rains; and (3) mortality studies of all women who died during their reproductive years using the verbal autopsy technique, to determine a likely cause of death and its relationship to pregnancy. The maternal mortality ratio was found to be 2,360 per 100,000 live births during the pre-intervention year with primigravidae, multigravidae, and women under 20 or over 40 years of age, most at risk of dying.

Following the introduction of the scheme, the proportion of women who attended prenatal care increased, but not significantly. The change was similar to increases registered in non-PHC control villages over the same period. Sixty-five percent of deliveries in the intervention area were attended by traditional birth attendants. A small proportion (4 percent) of deliveries in nonPHC villages were also attended by traditional birth attendants from neighboring PHC villages. There was a significant increase in the number of deliveries attended by trained midwives in PHC villages, but not in non-PHC villages. Mortality fell in both PHC (2,716 to 1,051) and nonPHC (1,498 to 963) villages, but the difference in the change in ratio is not statistically significant.

In summary, the introduction of the PHC scheme was associated with some improvements in maternal outcome. The traditional birth attendants may have had an effect by encouraging prenatal care attendance and by referring and accompanying women with complications to the health center for delivery by a trained midwife. These improvements cannot be attributed solely to the PHC scheme; however, the Farafenni health center was upgraded during the study period and transport options improved, which may also have had an impact. It is also likely that the system of surveillance played a role, though the researchers tried to interfere as little as possible. Further improvements could be made:

· The traditional birth attendants could play an important role in identifying women who need to deliver in the health center in advance and persuading them to do so. This will require the development of a system whereby women can stay near the center prior to delivery, and a major effort to make it socially acceptable for women at risk, and especially primiparas, to leave the village for delivery.

· Education is needed to reduce the proportion of women who do not use the services of the traditional birth attendants.

· The traditional birth attendants could also play a greater role in providing family planning to women at high risk.

· The health center could be equipped to provide surgery and blood transfusions. The Government has plans to ensure more peripheral availability of such services.

In conclusion, traditional birth attendants have an important role to play, but cannot bring about major reductions in maternal mortality unless they are supported by accessible, well-equipped referral centers.

(Greenwood 1991; Greenwood and others 1990)

Grenada - Public Sector Matemity Care: Summary and Associated Expenditures

Maternal and perinatal mortality is relatively low in Grenada despite limited use of advanced medical technologies. There were six maternal deaths in 1987-1988, among 5737 live births, or a ratio of 104 per 1,000,000 live births. A network of seven primary health care centers and 29 smaller visiting stations provide prenatal care. An estimated 75-100 percent of women use these services, with an average of 5-6 prenatal visits per woman. Ambulance services are available at two health centers and at the island's three small hospitals. Almost 90 percent of all births are performed by nurse-midwives, with physicians attending the more complicated cases. About 90 percent of all births occur in medical facilities. The following table shows estimated maternity care expenditures in the proposed 1991 Ministry of Health operating budget. Estimates are shown for two of the hospitals - the main General Hospital, and one of the smaller hospitals, Princess Alice and for all the health centers and visiting stations. For the purposes of these estimates, cost allocation was based on an analysis of utilization records. Selected unit costs are also shown.

(MotherCare Project, Laukaran 1990)

Guatemala - A Rural Case Management Strategy: The Quetzaltenango Project

Maternal mortality estimates in Guatemala range from 100-144 (government estimates) to 1,0001,700 (World Bank estimates) per 100,000 live births. Three of the five leading causes of all hospital discharges are pregnancy-related. Forty percent of the country's nine million people are illiterate, and more than two-thirds live in extreme poverty in both rural and urban areas. Access to health care is limited, with hospital capacity for a maximum of 25 percent of all births, and 4.4 physicians per 10,000 population.

The health system, which is highly centralized, is divided into health regions, areas, and districts. The district chief is responsible for a network of health posts, which are staffed by rural health technicians and auxiliary nurses, who in turn are responsible for supervising traditional birth attendants on an informal basis. The supervising auxiliary nurses, whose training emphasizes the health needs of children under age five, have little training and practical experience in obstetrics. This reflects the heavy emphasis on child survival in recent decades and limited programmatic attention to maternal mortality. Guatemala's 20,000 traditional birth attendants attend 60-70 percent of all births, but have little functional interaction with the referral system. Community organization and emergency transport are extremely limited.

The Government recently initiated a decentralization policy, and the Instituto Nutricional de Centroamérica y Panama (INCAP) was requested to develop a "local health system" in the high priority highland districts of Quetzaltenango as a test case, prior to country-wide implementation. As part of this effort, INCAP conducted an operational study to determine how high risk pregnancies were perceived, detected, and managed, focusing on all levels of the health system. A maternal and neonatal health project has been developed on the basis of the study findings, in collaboration with the MotherCare Project of John Snow, Inc.

Maternal mortality in the area was found to be 234 per 100,000 live births. Most deaths occurred in the area's one hospital (57 percent), followed by deaths at home (37 percent) and deaths en route (6 percent). Most were due to hemorrhage (41 percent), sepsis (35 percent) and eclampsia (16 percent). In the case of hemorrhage, 52 percent died within 2-6 hours, 74 percent within 24 hours and 98 percent within 48 hours. In the case of sepsis, days elapsed between onset and death. Seventy one percent of the deaths that took place at home were attended by traditional birth attendants, who recognized problems or sought help too late, and had no knowledge of simple management techniques (ea. external uterine massage in the case of post-partum hemorrhage).

Traditional birth attendants, who attended over 90 percent of all births in Quetzaltenango, were found to have limited understanding of the concept of risk. They recognize certain situations as dangerous, but often attribute them to luck or divine will, and do not know how to prevent or manage problems. Their opinion of the formal health system, from which they receive little support, is low. Harmful practices are common, including the widespread and often inappropriate use of oxytocin to "give force to the labor." Community members were found to have some knowledge of high risk situations and to perceive hospitals and doctors as the most appropriate source of care. Nonetheless, their opinion of the health system, due to poor treatment, fear, lack of confidence, high cost, and long waiting times, is low, and they are reluctant to use available services. The formal health services do not use a risk screening and management approach, and lack institutional norms and basic screening equipment. The referral system is non-functional, as are information and registration systems for patient management. Health staff have little knowledge of the conditions in which traditional birth attendants work, and tend to view their practices as dangerous, even when they are not.

The study was followed by a long program development process, which began with the presentation of findings to health personnel and the development of a collaborative plan of action. The program, which is now in the early implementation phase, aims to reduce mortality by accelerating the detection and referral of cases, and by ensuring appropriate management at all levels of the health system. The traditional risk approach is inappropriate in this context, given the low absorptive capacity of the formal health system. Referring all first births alone would far exceed this capacity. Therefore, an approach based on a small number of actual, high risk events, those which are associated with the greatest risk of mortality, has been adopted.

Strategies include traditional birth attendant training and supervision, through a modular, participatory approach that builds on their own experience and is appropriate to the local culture; the establishment of new relationships between all levels of the health system, focusing on the traditional birth attendants as the critical link, and promoting mutual respect; increasing the assessment and problem-solving skills of medical and nursing staff) increasing the registration of births through a simple technique of traditional birth attendant reporting; and ensuring that the information collected is used for improved decision-making. Community education, using interpersonal media, will be undertaken to improve recognition of danger signs, health care seeking behavior, and compliance with referral. Outcome and process indicators in the four intervention districts will be compared with those in four comparable control districts.

Traditional birth attendant training materials are being revised based on specific, priority obstetric/perinatal complications and the specific tasks the traditional birth attendants will need to perform to prevent death. A pictorial maternity card depicting these high risk situations will be developed, and will be managed by the traditional birth attendants themselves. The traditional birth attendant will be able to send patients to health facilities and alert health staff to risk situations using the card. Health staff can also use the card to alert the traditional birth attendants to additional problems encountered, facilitating joint patient management.

To improve relationships between levels of the health system and the functioning of the referral system, regular meetings will be held between traditional birth attendants and district health staff to identify training needs and discuss the problems traditional birth attendants confront. Traditional birth attendants have been taken to visit the hospital, to familiarize themselves with the surroundings in which their patients will be attended, and to exchange points of view on patient management with hospital staff. The chief of obstetrics and gynecology is considering making it possible for the traditional birth attendant to remain with the mother in hospital, as well as implementing other changes to make the hospital environment more comfortable for women. Physicians were also taken to communities to gain a greater understanding of prevailing conditions, including transport and resource constraints.

The project aims to develop a replicable model, including norms, information and referral forms and training materials. It will be promoted to facilitate replication throughout Guatemala, and possibly in other parts of Central America.

(MotherCare Project: Schieber 1991)

Indonesia - Regionalization of Care in West Java

Estimates of the maternal mortality ratio in Indonesia range from 150-720 per 100,000 live births, with hemorrhage, infection and toxemia responsible for 75-80 percent of all deaths. Less than 50 percent of Indonesian women receive prenatal care, and more than 70 percent are anemic. Eighty percent of all deliveries are attended by traditional birth attendants. Though traditional birth attendant training has been undertaken on a wide scale since the 1970s, it has failed to improve pregnancy outcomes. The training and supervision of traditional birth attendants is inadequate, and their functional interaction with the formal care system is minimal. Since the national Safe Motherhood meetings held in Indonesia in 1988, the Government of Indonesia has made a commitment to improving pregnancy and delivery care services, through a strategy based on increasing the number of traditional professional midwives and posting them at the community level. Ongoing work in the Tanjungsari sub-district of West Java, Indonesia's most populated province (32 million), aims to make recommendations for continuing development in national maternal care policy, training curricula and program inputs.

In 1985, a study was undertaken in Tanjungsari to evaluate the impact of a revised system of traditional birth attendant training using the risk approach. Ninety percent of all deliveries in the sub-district are carried out by traditional birth attendants, some of whom are men. According to government policy, traditional birth attendants in Indonesia can only be trained in preventive and promotive care. The program used pictorial maternal and child cards, which were kept by the woman but filled in by traditional birth attendants and other attending health staff, to improve traditional birth attendant recognition of risk factors, and promote appropriate referral to the health center or hospital, depending on the specific condition identified and which facility was best equipped to handle it. They were also trained to weigh the mother and to use a specially-developed one minute hourglass to take vital signs, and were taken to visit the hospital to visit mothers and infants and familiarize themselves with the hospital environment. In the control area, traditional birth attendants received conventional government training only.

Maternal mortality in the area (500 per 100,000 live births) was not affected by the risk approach training program. Traditional birth attendants trained in the risk approach referred 22 percent of women, compared to only 8 percent in the control area, but misreported 29 percent of risk factors, and missed 30 percent. Perinatal mortality decreased by a total of 23 percent, but greater reductions were recorded in the control area, which was probably due to a spillover in training between the intervention and control areas. Of the 20 'women who died between 1988 and 1989, 18 were referred, but six refused referral, many due to distance from the health facility. The terrain in the area is difficult, and adequate transport virtually nonexistent. Most women are currently transferred by hand-carried bench. Distance from the main road, however, was not found to influence mortality levels. In addition to difficulties associated with poor communications and transport, the quality of the referral system is compromised by poorly-equipped facilities, which are poorly linked to one another as well as to the community.

As a result, the development of a coordinated system for the delivery of maternal and child health (MCH) services, including prenatal, intrapartum and postnatal care, is proposed, beginning with an augmented package of services at the village level. Comprehensive health service posts (known as "birthing huts," though they will provide the full range of maternal and child health services) will be established. They will serve as the meeting point between formal and informal care, and will be managed by traditional birth attendants and the village midwife and coordinated by the village head and village committee. The huts will be equipped to provide prenatal care (equipment will include scales, blood pressure gauges, and dipsticks to measure protein in urine), a clean place for delivery, family planning, and child health care. They will also serve as a transfer point for referral: they will be located both near the main road and near the villages, will be equipped with a stretcher, and will have access to four-wheel drive vehicles. The huts will also serve as the village drug store.

Specific sites will be chosen by village heads, and the communities will be responsible for hut maintenance. Two way radios will link the huts with area health centers (and the health centers with the hospital), to ensure immediate emergency notification and to facilitate the provision of advice. The huts will also serve as information, education and communication centers, and will aim to improve community awareness of risk and the benefits of referral. Each of three subdistrict health centers will be responsible for supervising three to four huts. Only one of these facilities has an inpatient service, and will be equipped to serve 20 percent of all deliveries. An emergency van and driver will be available on a 24 hour basis. The hospital will be equipped to attend 5 percent of all deliveries, or those at highest risk. All health personnel will receive continuing training in treatment and preventive care protocols. Appropriate screening tools, norms, training modules, and information systems will also be developed at all levels. Continued interaction of all levels will be fostered through a system of meetings.

Longitudinal studies to assess the program's impact on outcomes will be undertaken in the intervention and control areas, and its results will be presented to health authorities, professional organizations, and universities. The program is expected to provide valuable information to the government's midwifery program, by providing a model of the midwives' fixed village-level facility (the MCH huts). It will also provide information on the effectiveness of a bottom-up approach to improving the accessibility of care, which will be invaluable during the transition from traditional birth attendants to trained midwives in village-level delivery care. The program is also expected to show that the training and posting of midwives must be supplemented by improvements at all levels of the referral system, as well as by efforts to ensure appropriate links between them, if improvements in maternal outcomes are to be achieved.

(Alisjahbana 1991; Alisjahbana and others 1991; Alisjahbana and Thouw 1991)

Tunisia - Strengthening and Expanding Family Planning and Maternal and Child Health Services within the Basic Health Care System

Tunisia has made significant progress on the demographic transition and is a trailblazer in the provision of family planning services on the African continent and within the Arab world. A substantial rise in the age at marriage (from 19 to 24 years for females) linked with improvements in female education and employment before marriage have all contributed to the recent decline in fertility levels from 7.1 to 3.5 children per woman. Abortion is legal in Tunisia, and the number of abortions has remained virtually constant over the last 15 years, demonstrating that the decline in fertility is due in large part to family planning services. Contraceptive prevalence is over 64 percent in some parts of the country, yet because of uneven distribution in contraceptive use based on regions and schooling, the needs for contraceptive use are not yet satisfied.

Studies suggest that maternal mortality is about 100 per 100,000 women for Tunisia as a whole but is as high as 1,000 per 100,000 in rural areas. Although almost 60 percent of all deliveries are assisted, studies indicate that a large proportion of maternal mortality can be prevented by timely referral to appropriate levels of care.

To reach populations in remote geographical areas of the country and those in the poorest social strata, the government is developing an integrated approach to delivering family planning and maternal and child health services by improving access to basic health care services. Safe motherhood is one of the key objectives of this strategy. The program provides all women with better access to family planning services and all pregnant women with access to prenatal care and safe deliveries through midwives and obstetrical nurses and a quality referral system to higher levels of care. Although all women will receive at least one visit for prenatal care, women identified with high-risk pregnancies will receive follow-up visits. Services at the first-referral level of care are being improved through the provision of equipment and the refurbishing of district hospitals, diagnostic centers, and rural maternity centers.

To increase the coverage of postpartum visits for mothers and children, the Faculty of Medicine in Sousse has developed a maternal and child health and family planning program.

The doctor or midwife who attends the birth schedules a postpartum visit for the mother and her infant forty days after the birth. (In Tunisia, this marks the end of the traditional period of seclusion of the mother and her newborn.) The mother is seen by an obstetrician and the child by a pediatrician. The proportion of women and infants who receive postpartum care as well as family planning and other essential health services has increased substantially as a result of this program.

(Voltaire and Weissman 1993)

Zaire The Karawa Health Zone Project

Maternal mortality in Zaire, a country of 38 million people, is 800 per 100,000 live births. Sixty to 80 percent of all births take place at home and are attended by traditional birth attendants. Total fertility is high, at 6.1. Per capita GNP is US$160.

In 1980, health officials instituted a policy of decentralization based on the primary health care approach. The country was divided into 306 health zones, which serve as the organizational units for health programming. Each zone covers 150-200,000 people and has at least one referral level facility. Zone health teams have full authority to deliver services in the manner they feel best serves the needs of the zone population and are responsible for ensuring effective linkages between community-based health activities, health centers, and the zone hospital, at which they are based.

The Karawa Health Zone (KHZ), which covers 19,000 square kilometers and is home to 300,000 people, is located in the northern Equateur Province. The local economy is based primarily on subsistence farming, and infrastructure is poorly developed. There are no paved roads or public transport options, and most people travel by foot, bicycle, or motorized twowheeled vehicles. Iodine deficiency is very common, resulting in a high incidence of cretinism and, as a result, cephalopelvic disproportion (CPD), a major cause of obstructed labor.

An ongoing program in KHZ aims to strengthen maternal health care within the primary care system, based on what can realistically be done to improve the accessibility and quality of maternal care given existing resource constraints. At the village level, traditional birth attendants are trained in basic delivery practices and the identification and referral of women at high risk. They are also responsible for encouraging prenatal care attendance. Traditional birth attendants are selected for training by a Village Development Committee composed of community leaders. They receive a small stipend during the training period (five days per month for six months) and a tee-shirt, cloth badge, and locally-made birth kit (soap, razor blades, oral ergot, mercurochrome, cord ties, etc.) upon completing the course. Kit resupplies are made available at a subsidized rate.

KHZ's 30 health centers, which serve 5-10,000 people, are staffed by a public health nurse and auxiliary personnel. Health center staff oversee routine deliveries, refer women with complications to the zone hospital (CEUM), and provide prenatal care and family planning. Most of these centers have minimal equipment and supplies and very limited emergency obstetric capacity. Four health centers have now been equipped to provide emergency obstetric care, and are staffed by an obstetric nurse or auxiliary midwife, a public health nurse, and auxiliaries. Most nursing staff have completed a two-year nursing program and a third year of training in either public health or midwifery. Auxiliary midwives undergo a two-year midwifery training program and receive some training in general hospital nursing.

The maternity center at the CEUM hospital is supervised by one physician and staffed by one nurse-midwife, supported by auxiliary midwives and obstetric nurses. Center staff are responsible for prenatal and family planning clinics, and selected obstetric nurses are trained to perform emergency surgery, including cesarean section, symphysiotomy, repair of ruptured uterus, and hysterectomy.

A pregnancy care monitoring study conducted in 1984-86 found that despite considerable effort to improve maternal care, maternal mortality in KHZ remained high, due primarily to deficiencies in the referral system at all levels. The women who died in the hospital were often admitted following prolonged labor, and lack of transport, distance, cost and other factors were found to cause significant referral delays. Traditional birth attendants did not refer all women in need of medically-supervised delivery, and many of the women who were referred did not comply.

Further research was undertaken to investigate the factors influencing referral and utilization in greater detail. The first component of this research focused on the effectiveness of traditional birth attendant training. Traditional birth attendants were found to make appropriate referrals in some cases, but to refer only about 20 percent of women with acute complications. They exhibited some confusion over risk factors and signs of acute problems. While traditional birth attendants are not usually called to assist before delivery, and are thus more likely to think in terms of problems in delivery than in terms of antecedent risk factors, less than half mentioned hemorrhage as a danger sign when asked, and less than a third mentioned prolonged labor. Traditional birth attendants are not reimbursed for referrals, and lose their delivery fee when they do, which may act as a disincentive. The study recommends limiting the type and number of risk factors taught to traditional birth attendants, and investigating the use of visual aids to enhance risk assessment; ensuring that traditional birth attendants have a plan of action before emergencies arise; and developing a system to ensure that traditional birth attendants receive some form of remuneration when they refer clients.

User fees are charged for services at all levels of the health system. Health centers, for example, must generate sufficient revenue to cover their operating costs. Hyperinflation in recent years has increased the numbers of people who cannot afford to pay for care, a factor which must be addressed in efforts to improve service utilization. Negative perceptions of referral facilities also limit utilization of available services. The study confirmed that traditional birth attendants are more accepted by the community as sources of care than formal health system staff, and recommends community-based health education to increase awareness of danger signs, help overcome fears of referral, and promote the development of community-initiated emergency protocols. In addition, formal health system staff need to improve collaboration with traditional birth attendants and communities in general.

A sixty-bed maternity waiting home was built in KHZ but is severely underutilized, due primarily to a lack of community involvement in service design and consequent lack of consideration of community concerns, particularly those related to food preparation. The study recommends investigating providing food, cooking facilities or fuel, and/or a staff member to assist patients with food preparation, laundry, etc.

Further analysis of the factors associated with maternal mortality found that duration of labor for 24 hours or longer was associated with the greatest risk of death among women delivering in hospital, due primarily to referral delays. Delays stemmed primarily from lack of transport, lack of cooperation of family members, and efforts to seek non-medical alternatives. Neglected obstructed labor was the cause diagnosed for four-fifths of the hospital patients with prolonged labor. The analysis confirmed the importance of reducing delays for women with obstructed labor by promoting earlier recognition and referral. The study recommends the development and improvement of management protocols, including visual charts for providers at all levels, to facilitate the identification of women at risk; training providers at all levels to recognize CPD as early as possible, including training staff at the health center level in the use of the partogram; and possibly training formal care providers in complete pelvic mensuration, a technique not currently in use in KHZ.

While screening for CPD using clinical pelvic mensuration requires specialized training, screening for such risk factors as maternal height can be conducted by traditional birth attendants. While short stature is associated with increased risk of CPD, the proportion of women at risk in different populations varies, and there is no universally accepted at-risk level for height. All women who have previously undergone a cesarean section should be referred, as should all cretinous women.

Improving access to surgical interventions is also key. As mentioned above, obstetric nurses in two health centers have been trained to perform cesarean section and symphysiotomy, the latter which requires fewer surgical materials and assisting staff, is associated with lower risk of infection, and can be performed using local anesthesia. Research was undertaken to compare symphysiotomy with cesarean section as an alternative intervention in cases of obstructed labor, and found that it was not associated with a significantly higher level of perinatal mortality. Further research comparing the long-term health outcomes of the two procedures is needed.

Additional improvements in the equipment and staff available at health centers, and efforts to address transport and communications deficiencies, will also be needed to maximize the effectiveness of the maternal care system in KHZ.

(Duale 1991; Duale and others 1991; Hermann and Duale 1990; White, Thorpe and Maine 1987; Smith and others 1986)

Zambia - Integrating Prenatal Care with Syphilis Control

High rates of maternal syphilis, ranging from 4 percent in Rwanda to 26 percent in The Gambia, have been reported in Africa. The consequences of maternal syphilis spontaneous abortion, perinatal or infant death, congenital syphilis - are dire, and are estimated to occur in 5 to 8 percent of all pregnancies in Africa that survive until the second trimester. The consequence of untreated syphilis for the mother, including death from syphilis of the cardiovascular or nervous system, can also be devastating .

In their work in Lusaka, Zambia, Dr. Subhash Hira and colleagues found that 8 percent of prenatal clients tested positive for syphilis, of whom 58 percent had adverse pregnancy outcomes, including abortion, stillbirth, prematurity, low birthweight, and congenital syphilis. Given the magnitude of this problem, the availability of a relatively inexpensive screening tool that can be used while the woman is attending the clinic (the plasma reagin test, RPR) and the existence of an effective treatment (one injection of benzathine penicillin), an intervention study was implemented to determine the impact of prenatal care-based syphilis control on these adverse outcomes. Measured against a baseline as well as a control group, the interventions (health education plus screening and treatment for maternal syphilis) were found to be very effective, bringing about a 66 percent reduction in adverse outcomes within approximately one year.

While the interventions were successful, they could have been even more so. Unfortunately, not all women who attended the prenatal clinics received the intervention, even under study conditions:

· Only about 60 percent of women received syphilis screening in the study centers.

· Only 15 percent of women received a second screening in their third trimester (third trimester screening is necessary because women can become infected or re-infected with syphilis during late pregnancy and still pass it on to the fetus).

· Only about half of those who tested positive were treated.

Why did the interventions not reach their potential? The time required to change beliefs and practices was probably underestimated, as is often the case. Early prenatal care is not valued in most developing countries, a fact that also continues to be a problem in some developed countries.

Another possible explanation pertains to the administration of prenatal care: routine prenatal test diagnoses are only useful if staff are well-supervised and -motivated to respond to positive test results. Conditions such as high blood pressure, anemia, and poor weight gain can signal potential complications during pregnancy, yet are often overlooked by clinic staff. There are multiple reasons for such oversight: clinic staff may serve multiple functions and may be occupied with more immediate problems (measles, fevers, diarrhea); medication is lacking (e.g., iron folate tablets); and health care providers are unsure of treatment protocols. In addition, the use of education as the sole means of preventing a condition such as high blood pressure is often not acceptable to the patient or to the provider. While these reasons may not be the primary explanation for the sub-optimal implementation of the Zambia demonstration project, they may well describe the environment in which care was provided in the study clinics.

The findings of the Zambian demonstration project strongly indicate the need for further trials of a prenatal care-based syphilis control intervention which focus on health education, screening, and treatment in areas where prevalence of maternal syphilis is high. In addition, the study highlights the need for building in "checks" to ensure that prenatal services function as planned, i.e. that women are screened for syphilis and other danger signs during pregnancy, and then treated accordingly; their partners must also be screened and treated

(Hire and others 1990).

Zimbabwe Expanding Access to Maternity Care

Zimbabwe has made impressive gains in providing health services to its 10 million people during the decade since independence in 1980. Basic health care services have been extended to cover the rural population, over 80 percent of the children are fully immunized, malnutrition has declined substantially, and rural water supply and sanitation projects have been undertaken throughout the country. Modern contraceptive prevalence of over 40 percent of women of reproductive age is the highest in Sub-Saharan Africa. Nevertheless, there are still underserved populations. Maternal mortality rates remain unacceptably high for rural women. Although more than 90 percent of expectant mothers attend antenatal clinics, over 30 percent of the deliveries are unattended and the leading causes of maternal deaths are infection, hemorrhage and abortion.

The government has emphasized primary health care and has organized the public and nongovernmental institutions plus some private facilities into a four-tiered system of national health service delivery. Basic health care, including family planning is extended to the community level through a network of rural hospitals, local health centers and community workers. At the second level, district hospitals serve as the first line of referral. Next, at the third level, there are eight provincial and four general hospitals which provide specialist services as well as obstetrics, gynecology and pediatric care. Five central hospitals serve as national referral facilities.

The Government of Zimbabwe, facing severe fiscal constraints and recognizing the importance of strengthening the country's primary health care infrastructure and services, developed a program to improve service delivery in the health, population and nutrition sectors. The government's program, which is supported by a consortium of donors led by the World Bank and including UNFPA and donor agencies from Norway, Denmark, Sweden, the United Kingdom, and the European Community, aims to strengthen maternal health by improving coverage of antenatal, delivery and postnatal services, ensuring that facilities are properly equipped and staff adequately trained, improving communication and transport links for patients being referred to higher levels of care, improving and expanding family planning and nutrition programs, and providing better health education to parents and children. In order to better understand the magnitude and consequences of the problem of maternal deaths in Zimbabwe, the Zimbabwe Medical School is conducting a community-based study on maternal mortality.

In a special effort to reach underserved populations living in rural areas, the government has targeted sixteen districts to receive a package of rural health facility upgrading, transport, communications, and health manpower development. The aim of this is to ensure that each family has access to a rural health center within eight kilometers, offering a full range of preventive and outpatient curative care, equipped to perform uncomplicated deliveries, and supported by a district hospital able to provide the full range of first-line referral services.

The program is providing pregnant women with improved access to safe deliveries through trained nurse midwives, a quality referral system, and emergency obstetric care. The project is strengthening basic and postbasic training for nurses, midwives, and community midwives in midwifery skills. The aim is to increase the percent of nurses trained in midwifery to 50 percent overall and 60 percent in rural health centers and hospitals. The skills of practicing midwives are being improved through better supervision and on-the-job training that ties skill development to the demands of the job.

Standard equipment for emergency obstetric care, including essential routine delivery equipment, is being provided to facilities in the 16 project districts. In addition, an emergency referral system, including four-wheel drive vehicles, ambulances, motorcycles, and a point-topoint radio communication system is being installed in those areas without access to telephone lines to support the transfer of patients, routine supervisory systems, and pharmaceutical distribution.

(World Bank 1992)