NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Research Council (US) Panel on Reproductive Health; Tsui AO, Wasserheit JN, Haaga JG, editors. Reproductive Health in Developing Countries: Expanding Dimensions, Building Solutions. Washington (DC): National Academies Press (US); 1997.

Cover of Reproductive Health in Developing Countries

Reproductive Health in Developing Countries: Expanding Dimensions, Building Solutions.

Show details

5Healthy Pregnancy and Childbearing

An estimated 585,000 women die each year from pregnancy-related causes (World Health Organization and UNICEF, 1996). Such estimates have raised awareness of the fact that women still die in what is perceived by most to be a healthy process (Rosenfield and Maine, 1985). Since there are approximately 180 million pregnancies each year, the overall view of a healthy process is reasonable, but that view hides the huge disparity between women in developed and developing countries: about 1 in 48 women in developing countries dies of complications of pregnancy, delivery, puerperium, or abortion, compared with only 1 in 1,800 in developed countries; see Table 5-1. The risk of dying is highest for women in Africa, for two reasons: on average, they are pregnant and deliver more frequently than women on other continents and each pregnancy is riskier. Because of its much higher population, though, the majority of maternal deaths each year take place in Asia.

TABLE 5-1. Maternal Mortality by Major Regions, Circa 1990.


Maternal Mortality by Major Regions, Circa 1990.

The consequences of a maternal death for a woman's family are also profound: if she dies, the chance of death for her children under the age of 5 is as high as 50 percent in developing countries (World Bank, 1993).

Maternal and Infant Death and Disability

Maternal Mortality

The causes of maternal mortality are divided into direct causes, those which occur only during pregnancy and the peripartum period, and indirect causes, those that are aggravated by pregnancy but may be present even before pregnancy, such as diabetes, malaria, or hepatitis. Approximately 80 percent of all maternal deaths are estimated to be due to direct causes (World Health Organization, 1996). The World Health Organization (WHO) (1993c) estimates that hemorrhage is the most common direct cause, followed by sepsis and complications of unsafe abortion, hypertensive disorders of pregnancy (including eclampsia), and obstructed labor. The relative importance of different direct causes of mortality varies among studies, due in part to differences in reporting and definitions, and in part to real differences in the quality and accessibility of delivery care.

Although the definition of a maternal death includes a woman's death while pregnant or up to 42 days postdelivery from any cause (except accidental), most maternal deaths, excluding abortion-related deaths, occur during labor and delivery or soon thereafter. In a rural area of Bangladesh, for example, more than 70 percent of the nonabortion deaths occur in this short time span: 40 percent during labor or within 48 hours of delivery (primarily due to eclampsia and postpartum hemorrhage) and 30 percent between 3 and 42 days postdelivery, with sepsis and associated diseases leading the list of causes (Fauveau et al., 1988). Extending the definition of maternal death to 90 days postpartum, as has been proposed, would only have increased the number of maternal deaths by 6 percent. The critical period as that of labor and delivery is also supported by a nationally representative study in Egypt (Ministry of Health, 1994): 39 percent of deaths took place during delivery or within the first 24 hours, and 36 percent occurred within 42 days postpartum; only 25 percent took place during pregnancy itself.

Maternal Disability

Between 30 and 40 percent of pregnant women, or over 54 million women in developing countries, are estimated to experience a pregnancy-related complication annually (World Health Organization, 1993b; Koblinsky, Campbell, and Harlow, 1993). WHO estimates that 15 million women per year develop long-term disabilities from such complications as obstetric fistula, prolapse, severe anemia, pelvic inflammatory disease, and reproductive tract infections, as well as infertility.

At the country level, these estimates vary widely. For example, in community-based studies, Guatemalan women reported one in five pregnancies as complicated (Bailey, Szaszdi, and Schieber, 1994); in West Java, one in three women reported complications, not including those experienced in the postpartum period (Alisjahbana et al., 1995); and in Ghana, two of every three pregnant women had some complication, although the serious complications were infrequent (postpartum bleeding, 7 percent, convulsions, 2 percent, and postpartum sepsis, 5 percent (De Graft-Johnson, 1994). A population-based survey in El Salvador found one serious complication alone—intense intrapartum bleeding to the point of losing consciousness—was reported by nearly one-quarter of rural women (Danell et al., 1994). When the definition of complications is broadened to include perineal laceration, prolonged labor not leading to cesarean section, breech, multiple births, retained placenta, and other less severe complications, one-half of women in the Philippines reported a complication, with the most common being perineal lacerations (43% of respondents) (National Statistics Office and Macro International, Inc., 1994).

Reliability and validity of self-reported data on complications and their comparability across studies are likely to be poor (Task Force on Validation of Women's Reporting of Obstetric Complications, 1997). In a validation study undertaken at the Philippine General Hospital, Stewart and Festin (1995) developed algorithms based on reported symptoms to identify women most likely to have experienced hemorrhage, eclampsia, severe infection, and cesarean section due to obstructed labor. The algorithms were then used to identify women in the Demographic and Health Survey (DHS) sample likely to have suffered life-threatening maternal complications: 12 percent of all women in the sample were identified (National Statistics Office and Macro International, Inc., 1994). Hemorrhage was the most common complication (8%), followed by cesarean section due to obstructed labor (3%), severe infection (2%), and eclampsia (1%). The overall estimate is close to the 15 percent of women estimated to suffer ''serious" complications based on data from the United States and Canada (Koblinsky, Campbell, and Harlow, 1993; World Health Organization, 1994b). If 12-15 percent of women suffer life-threatening obstetric complications, then approximately 20 million women in developing countries warrant referral care each year.

Millions more women suffer associated illnesses that are aggravated by pregnancy—anemia, malaria, cardiac disease, hepatitis, tuberculosis or diabetes—that can indirectly cause death or further disability for the woman or newborn. WHO estimates conservatively that 12.5 million women each year are affected by these illnesses aggravated by their pregnancies (World Health Organization, 1993b; Koblinsky, 1995). Approximately 20 percent of all maternal deaths in all regions are attributed to these indirect causes (World Health Organization, 1993b). Furthermore, the nonfatal consequences of obstetric complications and associated diseases can severely affect women's quality of life, fertility, productivity, and can result in chronic reproductive morbidities that may become evident only long after delivery.


Given its high prevalence and impact on the lives and survival chances of women and newborns, anemia (hemoglobin counts below 11 g/dl) deserves special mention. Dietary iron deficiency is the most common cause of anemia, although malaria, other parasites (schistosomiasis and hookworm), AIDS and sickle-cell diseases may also contribute. Approximately 50 percent of pregnant women throughout the world are estimated to be anemic. A United Nations expert panel considers severe anemia (below 7 g/dl) an associated cause in up to half of maternal deaths worldwide (United Nations, 1991). Severe anemia may be directly associated with maternal morbidity and mortality, although available data do not often allow one to distinguish mild and severe anemia.

The impact of maternal anemia on the outcome of pregnancy—prematurity, stillbirths, spontaneous abortions, perinatal and neonatal mortality—is well documented (Levin et al., 1993). In a sample of 4,434 births in a northern Nigerian hospital, when maternal hematocrit levels measured 2 weeks before delivery were below 18 percent (the cutoff for severe anemia is 23%), nearly 50 percent of the births were stillbirths, and another 15 percent of newborns died in the neonatal period. When hematocrits were between 19-25 percent, stillbirths accounted for 22 percent of all births (Harrison, 1985). Two large studies, one in the United States and the other in Wales, found a U-shaped relationship between hemoglobin levels and the risk of low birth weight, prematurity, and perinatal mortality (Garn et al., 1981; Murphy et al., 1986). In both studies, hemoglobin levels below 10 g/dl and above 13 g/dl were associated with these poor birth outcomes.

One study from Nigeria implicates even mild anemia in 45 percent of maternal deaths of women who also suffered a major obstetric complication (Harrison, 1985). Iron-deficient women have an increased risk of complications during pregnancy, including urinary tract infections, pyelonephritis, and pre-eclampsia (Kitay and Harbort, 1975). Laboratory studies and studies with children have linked iron deficiency with increased morbidity from infectious diseases due to impaired immune function, although this would not explain the association with pre-eclampsia (Srikantia et al., 1976; Stinnett, 1983; Bhaskaram and Reddy, 1975; Enwonu, 1990). Anemia is also associated with reduced capacity to work, affecting both women's productivity and their quality of life (Bothwell and Charlton, 1981).

Besides anemia caused by iron deficiency, folate deficiency, which causes megaloblastic anemia, is also common in developing countries. This deficiency may have independent effects on birth weight, preterm birth, and possibly, neural tube defects in newborns (Hughes, 1991).

Other common micronutrient deficiencies—particularly iodine and vitamin A—are also known to have similar negative pregnancy outcomes.

Obstetric Fistula and Genital Prolapse

Two chronic conditions warrant attention because of their high prevalence and devastating consequences—obstetric fistula and genital prolapse. An obstetric fistula is a passage in the vaginal wall leading into the bladder (vesico-vaginal fistula), the rectum (recto-vaginal fistula), or both. In developing countries where it has been reported, fistula is associated with prolonged and obstructed labor. Those at high risk of fistulas include very young women of low parity; women whose growth has been stunted by nutritional deficiency and recurrent infection in childhood that led to growth stunting; women in rural areas where health care is not accessible because of distance, lack of transport, and poverty; and women who prefer and have used traditional care and home delivery many times (Lawson, 1992; Murphy, 1981; Mustapha and Rushwan, 1971; Tahzib, 1983). A traditional practice in northern Nigeria, the gishri cut (the cutting of the anterior of the vagina as a way of treating a number of conditions, including obstructed labor, infertility, dyspareunia, and amenorrhea) further contributes to the risk of fistulas (Tahzib, 1983, 1985).

Good estimates of the prevalence of obstetric fistulas are impossible because the affected women do not seek or get care at hospitals, and health interview surveys have not elicited data on the condition. The available reports of fistula in developing countries come from most countries of sub-Saharan Africa and South and Southeast Asia (World Health Organization, 1991; Lawson, 1992).

The consequences of fistulas are devastating, especially for primiparas (women in their first pregnancies). In most cases, the baby is stillborn. The woman becomes incontinent of urine and sometimes of feces, uncomfortable conditions producing a foul odor and leading to feelings of shame and disgrace. The woman may not know about treatment, or she may be unable to seek professional help. Once the condition is seen as chronic, she is often deserted by her husband and has to seek support within her own family. Even there, she is usually segregated and not allowed to participate in certain activities, such as food preparation. In Islamic cultures, the woman may be forbidden to pray because of the requirement of cleansing oneself for prayer, which her condition of incontinence does not allow (Women's Global Network for Reproductive Rights, 1991; Murphy, 1981).

Also borne silently in most countries is genital prolapse, often considered a normal consequence of childbearing. Occurring when the vagina and uterus descend below their normal positions, this condition is associated with high parity and is often the result of damage during childbirth to the muscles and ligaments that support those organs. Genital prolapse is extremely uncomfortable, particularly for women who undertake chores in a squatting position, as is common in low-income and rural settings in developing countries. Prolapse may also be accompanied by backache (Younis et al., 1994) or urinary problems (Otubu and Ezem, 1982). Sexual intercourse may be painful, and pregnancy can lead to fetal loss and further maternal morbidity (Younis et al., 1994; Das, 1971).

As with fistula, there are no reliable large-scale population-based estimates of the prevalence of genital prolapse. A Brazilian hospital study revealed that 40 of every 1,000 women coming to a university hospital for other reasons suffered severe genital prolapse (Pinotti, Brenelli, and Moragues, 1993). Community-based studies have found between one-third and two-thirds of women with prolapse in Lahore, Pakistan, Istanbul, Turkey, and in two rural villages in Giza, Egypt (Omran and Standley, 1981; Younis et al., 1993). Clinically confirmed severe prolapse was found in almost one-third of the women in a study in Giza, Egypt; many were found to be using an intrauterine device (IUD) for contraception despite the fact that IUDs increase the discomfort of the prolapse (Zurayk, Younis, and Khattab, 1995). The Giza study also revealed an association between prolapse and reproductive tract infections (Younis et al., 1993).

Consequences for Infants

The impact of women's reproductive health on the fetus or newborn is immediate and dramatic. An estimated 7.6 million perinatal deaths (stillbirths and first-week deaths) occur each year in developing countries (World Health Organization, 1996). These perinatal deaths are associated to a large extent with their mothers' health and nutritional status prior to and during pregnancy, the management of labor and delivery, and the same maternal complications that can cause women's deaths. About one-half of all deaths of children under age 5 occur in their first month of life. Population-based studies in South Asia and the Philippines suggest that approximately 20-25 percent of perinatal deaths are associated with causes known to threaten the survival of women during pregnancy or delivery, another 20 percent are due to management practices at delivery, and more than one-third are due to women's health and nutritional status (Shah, Pratinidhi, and Bhatlawande, 1984a, 1984b; Fauveau et al., 1990; National Statistics Office and Macro International, Inc., 1994). Programs aimed at improving women's health and nutritional status and at managing obstetric complications and providing appropriate care of the newborn could reduce this enormous death toll (World Health Organization, 1994a).

The consequence of a mother's health and nutritional status, management of delivery, and early newborn care may not be death for the infant, but long-term disability. Obstructed or prolonged labor leads to asphyxia for an estimated 3 percent of newborns, resulting not only in death for nearly a one-quarter of these infants, but also in brain damage leading to cerebral palsy, seizures, and severe learning disorders for another quarter (World Health Organization, 1993b). Women with poor nutritional status (short stature, poor prepregnancy weight, inadequate weight gain during pregnancy, and anemia) or reproductive tract infections or other infections during pregnancy are more likely to have low birth-weight infants. The perinatal mortality rate of a low birth-weight baby is 20 to 30 times higher than that of a fetus or infant of normal weight. Many low birth-weight infants who do not die may suffer serious neurological problems, hearing and visual defects, and may be subject to slow development throughout their lives.

Interventions To Reduce Maternal Deaths

The most immediate pathway to maternal death begins with conception. The major complications that cause maternal death may be present initially in a mild form, progressing to severe complications in some 12-15 percent of all pregnancies. Interventions can stop this progression at several points:

  • prevent unintended pregnancies (see Chapter 4),
  • prevent diseases that will complicate pregnancy or detect and treat early signs of complications, and
  • treat complications at the mild or severe level with "essential care of obstetric complications."

Prevention and Early Treatment of Diseases and Complications

Prenatal care is primarily preventive, enabling health care staff to identify problems and illnesses that threaten a pregnancy and its outcome, to monitor and treat some conditions, and to give the pregnant woman and her family information about appropriate diet, behaviors, and delivery care.

Most studies of the effectiveness of prenatal care have focused on infant outcomes—perinatal mortality, preterm delivery, and low birth weight. But an association between use of prenatal care and maternal mortality has also been found in hospital case series, case-control studies, and informal surveys, although it is argued that this association is due to selection bias—that is, the women who use prenatal services are also more likely to take better care of themselves and use delivery services (Rooney, 1992). The major direct causes of maternal deaths in developing countries can neither be predicted nor prevented (Maine, 1991; Thaddeus and Maine, 1994).

Prenatal care may be effective in detecting and treating two conditions that underlie the direct medical causes of maternal and perinatal death: anemia and hypertension. For the mother, one of the most important benefits of prenatal care can be the provision of information about obstetric emergencies and linkage to sources of care. For this reason, it is best to consider prenatal care and delivery and early postpartum care as parts of a whole, rather than as separate interventions.

The benefits of adequate prenatal care for the health of the infants are somewhat clearer, but difficult to evaluate, because women who get inadequate prenatal care are more likely to have other risk factors for poor pregnancy outcomes (Kramer, 1987). Some observational studies with reasonable controls for confounding variables have found that better prenatal care is associated with such improved outcomes as lower rates of intrauterine growth retardation and premature delivery (e.g., Coria-Soto, Bobadilla, and Notzon, 1996). Tetanus toxoid immunization during pregnancy has been crucial in reducing neonatal and maternal tetanus (Fauveau et al., 1990). Screening for, and treating, syphilis during pregnancy is beneficial to both mothers and infants.

Risk Assessment

Predicting those pregnancies that will go on to suffer mild or severe complications on the basis of risk factors (as opposed to medical signs or symptoms of complications) has not proved useful. Demographic factors, such as age, parity, or a combination thereof, have not proved to be sensitive and specific enough. In a Philippines' population-based survey, for example, background characteristics of respondents reporting symptoms of major obstetric complications for births in the past 3 years revealed minimal variation among the age or parity groups, suggesting that these demographic indicators are inadequate predictors of risk of an obstetric complication (National Statistics Office and Macro International, Inc., 1994). In a program context, the use of such risk factors as the basis for referral could either overwhelm the maternity care system, or, if only specific ones are used, focus attention on those women who contribute minimally to maternal deaths. In Guatemala, for example, 70 percent of the women would be considered at risk if age and parity were used as risk factors, while the system can only manage 20 percent (Schieber, 1993). In Egypt, only 16.4 percent of deaths occurred among women aged less than 20 or more than 40 years, two common demographic factors. Although these women have a higher risk of death, the numbers of such women are low in comparison with women who die between 20-40 years of age simply because most pregnancies occur to women aged 20-40 (Ministry of Health, 1994).

Educational background did produce some variation in the proportion of women with any symptom of obstetric complication in the Philippine survey. However, most of the variation could be explained by differences in cesarean section due to obstruction, the one complication that requires hospital treatment. This result has been interpreted to indicate that educated respondents had better access to medical services (National Statistics Office and Macro International, Inc., 1994).

Any woman, regardless of age, parity, socioeconomic status, or education, can develop a complication at any stage of pregnancy, delivery, or the postpartum period. If a complication develops, it may be an emergency (e.g., hemorrhage) or may be taken seriously only when it reaches the stage of being an emergency (e.g., sepsis, eclampsia, obstructed labor).


Anemia is targeted for detection (if screening is carried out) or prevention (if iron tablets are given to all pregnant women) during prenatal care. However, three decades of prenatal distribution of iron and folate tablets to pregnant women has had little impact on the levels of anemia. This outcome does not reflect lack of efficacy of iron supplementation: in supervised trials in both developed and developing countries, iron tablets are associated with improved maternal hematologic status (depending on the initial anemia status and dose and duration of supplementation) (Mahomed and Hytten, 1989; Sloan, Jordan, and Winikoff, 1992). The side effects that sometimes accompany iron consumption or women's dislike of the pills themselves (due to smell or taste) often have been blamed for the failure of programs to reduce anemia, although a recent literature review found that side effects accounted for only 10 percent of the noncompliance. Rather, in most cases, women did not take their pills because they never received them or received them in inadequate numbers (Galloway and McGuire, 1994). Not enough pills are purchased by health facilities or governments because of lack of funds or a perception that anemia is not a serious health problem.

Yet even in areas where supplementation trials have taken place, overall prevalence of anemia often remains high. In India, for example, the prevalence of anemia decreased only from 88 percent to 56 percent in the highest iron dose group (240 mg) after a well-conducted trial by Sood et al. (1975). The authors concluded that it is difficult to treat a severely iron deficient woman and provide for increased fetal needs through oral iron supplementation alone during the relatively short period of pregnancy. Thus, in developing countries, prolonged supplementation beginning before women become pregnant may be a more effective strategy to benefit the majority of the population (Sloan, Jordan, and Winikoff, 1992; Galloway and McGuire, 1994). Community-based distribution schemes along with counseling on why, how, and when to take pills and where to obtain refills could increase access and compliance.

Long-term strategies to improve the overall iron status of the general population include increased production and consumption of iron-rich foods, increased family income, fortification of commonly consumed food with iron, and reduction of work for women. Food-based solutions for increasing iron consumption are not promising because most diets are plant-based and contain only small amounts of absorbable iron, except when fermented and germinated foods are added (which reduce the inhibitors of iron in the diet). Even when such additives are provided, food-based solutions seem to require increasing family income or the control of it by women.

Fortification schemes have proved successful in developed countries and have virtually eliminated iron deficiency anemia in small children (Dallman, 1989, 1993). The key to successful fortification is finding a food that all vulnerable groups consume and regulating the private sector to ensure compliance with norms for fortification. These are not easy tasks. Even if an appropriate food is identified, women would still need to take supplements to meet their iron requirements during pregnancy. Reducing maternal workloads may decrease women's overall requirement for iron, and all programs addressing anemia should include counseling of family members so they can help reduce a pregnant woman's activities, particularly during her third trimester.


Eclampsia and pre-eclampsia, the life-threatening complications of hypertensive disorders in pregnancy, have proved difficult to predict or prevent. Although eclampsia is considered the terminal stage of hypertensive disorders of pregnancy, a high proportion of cases occur in women in developed countries who did not have previous hypertension and proteinuria. This may be due to detection and management of patients with classic signs of pre-eclampsia, leaving a higher percentage of atypical eclampsia cases, or it may be that eclampsia can occur so rapidly that prior signs are not noted. At any rate, detection of the classic form of the disease is made difficult because its natural progression is not well understood, and the relative importance of the degree and timing during pregnancy of the symptoms (hypertension, proteinuria, edema, or other biochemical abnormalities) is unclear (Rooney, 1992). Predicting pre-eclampsia is also elusive: only severe obesity and a history of pre-eclampsia have been found to be independently associated with severe pre-eclampsia (Stone et al., 1994). Primary prevention with low-dose aspirin or with calcium supplementation is being studied, although evidence is accumulating against the use of aspirin among low-risk women for prevention of pre-eclampsia (Villar and Bergsjo, 1996).

Blood pressure measurements during pregnancy continue to be recommended, although the minimum number and timing of the measurements to detect cases of severe pre-eclampsia or eclampsia is unknown (Rooney, 1992). Neither edema nor proteinuria detect pre-eclampsia as well as prenatal measurements of blood pressure (Golding, Shenton, and MacGillivray, 1988; Hall, Chng, and MacGillivray, 1980). However, urinalysis for proteinuria is recommended for all women on the first prenatal visit (along with blood pressure measurement) because of the severity of the disease if found in the early part to pregnancy. After the first visit, proteinuria should be detected during all visits only for nulliparous women or those with previous pre-eclampsia or hypertension (Villar and Bergsjo, 1996). Women with moderate or high hypertension with proteinuria require referral and treatment.

The wide variation in case fatality rates for women from eclampsia or pre-eclampsia among countries suggests that differences in care can improve outcomes. Women with pre-eclampsia and eclampsia experience better outcomes when they have access to and use professional care (Rooney, 1992). Rest, antihypertensives, and anticonvulsants are presently considered possible treatments. Evidence is insufficient to determine the effect of bedrest, which in any case is often an unrealistic prescription. Antihypertensive drugs have been shown to prevent further increase in hypertension, but their effect on preventing pre-eclampsia is still uncertain (Rooney, 1992). For women who have already progressed to convulsions (eclampsia), a recent multicenter trial showed that the anticonvulsant drug of choice should be magnesium sulfate rather than diazepam or phenytoin, as it decreases the risk of recurrent convulsions (The Eclampsia Trial Collaborative Group, 1995).

Other Diseases

Approximately 30,000 maternal deaths yearly may be caused by tetanus (Fauveau et al., 1993). Where tetanus is a cause of maternal mortality, adequate protection for mothers may be provided during prenatal care by the tetanus toxoid immunization typically aimed at protecting newborns.

In the Philippines, tetanus toxoid injections have been given to all adolescent girls to increase the probability that they will be adequately protected the first time they give birth. Reproductive tract infections, including sexually transmitted diseases, should also be screened for and treated during prenatal care. As we note in Chapter 3, at a minimum, syphilis should be screened and treated during pregnancy, and newborns should be given prophylaxis for gonococcal and chlamydial eye infections (ophthalmia neonatorum). These eyedrops can be given by traditional birth attendants: they are simple, inexpensive interventions that are highly cost-effective in most parts of the developing world.

Obstructed and Prolonged Labor

Obstructed and prolonged labors are estimated to cause 40,000 maternal deaths each year, with many more survivors suffering obstetric fistulae while their newborns suffer death or brain damage from asphyxia. Prompt detection and management of obstructed and prolonged labor can have a beneficial effect on the outcome of pregnancy for both mothers and infants. Risk factors for this complication during pregnancy (e.g., height, foot size, or history of poor previous outcome) have not proven to be specific or sensitive enough (Fortney, 1995; Maine, 1991). But monitoring during labor allows for early intervention or referral with the consequent reduction in a number of sequelae of severe obstructed labor.

A recent WHO multicenter trial of more than 35,000 women who delivered in eight hospitals in Indonesia, Thailand, and Malaysia found the partograph a beneficial tool for detecting labors that progress too slowly and led to guidelines for labor management (World Health Organization, 1994a). The partograph is a chart to record the progress of labor and other essential fetal and maternal observations. When the partograph shows an abnormal progress of labor, drugs may be used to improve the pattern of contractions for a normal delivery, or a cesarean section may be performed. Through comparison of patients before and after introduction of the partograph, this multicenter hospital study showed that the number of prolonged labors (> 18 hours) was halved, the rate of postpartum infection (sepsis) was cut by over one-half, and the number of still-births fell from 5 to 3 per 1,000 babies. Fewer drugs were also needed, and caesarean sections for women without complications were avoided, with no adverse impact on the condition of the fetuses.

If labor is followed with the partograph by a medical professional in a home or health center, it is assumed that time would allow for referral to a medical center where prolonged or obstructed labor could be managed. However, there has been no trial on use of the partograph in the home or health center by medical staff.

Essential Care for Obstetric Complications

Since the major causes of maternal mortality cannot be predicted or prevented well enough during pregnancy to allow reliance on primary prevention and screening, improvements in maternal death rates will require that women have access to facilities with trained providers and equipment that can carry out essential care of obstetric complications. Defined by WHO first in 1985 with refinements made in 1995 (World Health Organization, 1995), essential care of obstetric complications consists of:

  • the ability to carry out surgery (i.e., caesarean section, treatment of sepsis, removal of an ectopic pregnancy),
  • the ability to provide intravenous oxytocin,
  • the provision of anesthesia,
  • medical treatment (for shock, sepsis, anemia, and hypertensive disorders of pregnancy),
  • replacement of blood,
  • manual procedures (e.g., removal of placenta, repair of episiotomies and perineal tears and vacuum extractions),
  • monitoring of labor (including use of the partograph),
  • management of problem pregnancies (severe anemia, diabetes, twins, malpresentation),
  • manual vacuum aspiration for treatment of incomplete abortions, and
  • provision of special care for neonates (e.g., resuscitation).

Historical experience provides evidence for the effectiveness of this approach. It was not until the mid-1930s, with the introduction of medical technologies to treat obstetric complications, that maternal mortality began to decline in several European countries and in the United States. After antibiotics, blood transfusions, and improved surgical techniques in caesarean sections and safe abortions became routinely available in the industrialized world, maternal mortality all but disappeared (Loudon, 1991). Prior to that time, infant mortality had declined dramatically in the United States and Europe, but maternal mortality had remained constant. While the causes of infant death are extremely sensitive to environmental factors and respond quickly to improved sanitation and nutrition, the majority of obstetric complications that cause maternal deaths cannot be averted simply by improving women's overall health or nutritional status (Loudon, 1991).

In Sweden, where mortality statistics have been kept since 1750, the maternal mortality ratio declined from 900 to 6 per 100,000 live births between 1750 and 1980, with two-thirds of the decrease occurring in the eighteenth and nineteenth centuries. This decline, not reported in other European countries, has been attributed to home-assisted births by trained midwives and the use of aseptic techniques. The decline in the maternal mortality ratio in Sweden in the twentieth century was attributable to the same factors as in the rest of Europe and United States (Hogberg and Wall, 1986; Hogberg, Wall, and Brostroin, 1986).

Implementation of some elements of essential care of obstetric complications in a few developing countries have also resulted in substantial declines in maternal mortality (World Health Organization, 1995). Sri Lanka's maternal mortality ratio dropped dramatically: from 555 per 100,000 live births in 1950-1955 to 239 in the 1960s and to 95 in 1980. A nationwide extension of the health center system and expansion of midwifery skills are credited with this rapid decline. A major shift toward birth with trained personnel occurred over this 30-year period, with a major impact on the proportion of deaths attributed to sepsis.

Maternity Care And Survival

Use of Maternity Care

The use of medical services for delivery lags far behind use of prenatal care in most developing countries. Home birth, either alone or with someone from the community, remains a strong preference. WHO estimates that only 37 percent of births in developing countries take place in a health facility; more than 60 percent of births—or 55-60 million infants annually—take place with only the help of traditional birth attendants, family members, or no assistance (World Health Organization, 1993a).

The reasons for the widespread acceptance of prenatal care throughout developing countries were captured nicely by Bolivian women who stated, ''because you're in a delicate condition," "to see if the baby's okay" (The Center for Health Research, Consultation and Education and MotherCare/John Snow, Inc., 1991). In 39 of 43 countries covered by DHS surveys between 1985 and 1994, coverage for prenatal care was found to be higher than for delivery care from a trained health provider (doctor, nurse, or nurse-midwife) (Macro International, Inc., 1994). In sub-Saharan Africa, for example, 15 of 22 countries surveyed had achieved over 75 percent prenatal care coverage (women's own definition of prenatal care was used), and only 2 had below 50 percent coverage (Macro International Inc., 1994). But only one country, Botswana, achieved over 75 percent of deliveries with professional health care providers. Between 50-75 percent of pregnant women in one-half of the other countries used medical services for delivery, while in the other one-half of the countries less than 50 percent did so.

Latin American and Caribbean countries have somewhat higher coverage rates than sub-Saharan African countries. In 3 of 11 Latin American countries (Trinidad and Tobago, the Dominican Republic, and Colombia), more than 75 percent of women both used prenatal care and gave birth with the assistance of a professional health care provider. In five countries, prenatal coverage rates ranged between 50-75 percent, and four of these had delivery assistance in the same range. Bolivia had prenatal care coverage in that range, but delivery assistance was slightly lower at 47 percent. In only one country, Guatemala, was coverage for both types of services less than 50 percent (Macro International, Inc., 1994).

In the Asian countries surveyed, three-quarters of pregnant women used prenatal care, covering women in Indonesia, Philippines, Sri Lanka, and Thailand; in Bangladesh and Pakistan, only one-quarter of women used prenatal care. Three of every four Sri Lankan women, one-half of Thai women, and two-thirds of Philippino women had a professional attending their deliveries. In the remaining three countries, Bangladesh, Pakistan, and Indonesia, between one and three out of every ten women had a professional in attendance.

Data on the use of postpartum services are relatively scarce, but the rates are typically lower than rates of institutional delivery. In the African countries for which data are available, coverage for postpartum care has for the most part not exceeded 40 percent, even in urban areas. Latin American rates are slightly higher than those for Africa, and Asian women seem the most likely to attend a postpartum clinic. However, data are sparse, and information is not available on the types of providers who are serving women in this period nor women's reasons for seeking such care (World Health Organization, 1993a). Even rates of coverage provide little information about the quality of care or the reasons for its use.1 What these coverage rates do show is that a substantial portion of women remain outside delivery and postpartum services for reasons that could range from taboos on mobility to preference to remain at home to the unavailability or inaccessibility of appropriate care. Women living in rural areas or having no education use all these services less than other women, while differences in age, parity, and birth interval are less consistent in determining coverage rates (Govindasamy et al., 1993).

Pathway to Maternal Survival

Since most women in developing countries experience labor and delivery outside the formal health care system, we discuss obstetric care under a four-step pathway to maternal survival, assuming that labor begins for a woman in her home (see also Thaddeus and Maine, 1994). From the point that an obstetric or newborn complication occurs, four steps are required to promote survival of the woman or the baby:

Step 1:

recognizing a life-threatening complication by the woman, her family, traditional birth attendant, or others in attendance;

Step 2:

deciding to seek care, typically by family members if the woman is in poor condition;

Step 3:

reaching quality services, which often involves overcoming impediments such as distance, cost of or lack of transport, cost of the services, geographical or weather constraints, and perceived poor quality or attitude of the providers; and

Step 4:

obtaining appropriate care for obstetric complications.

Death of a woman, fetus, or newborn is likely to ensure if one of these steps is not taken.2 This same sequence is relevant when a woman suffering complications has made contact with a health care provider. If the provider diagnoses the problem, decides on the appropriate care, and refers to quality care on a timely basis or provides adequate care, death is more likely to be averted for the woman and her baby.

The family side of this chain of events was followed in a retrospective study in the altiplano in a remote area of Bolivia (Bartlett, 1991). In only one-quarter of the cases resulting in maternal and neonatal deaths were the symptoms of a life-threatening condition recognized by the family in time, and in another one-quarter of cases the symptoms were not recognized at all. The most common response among the families who tried to do something about the problem was to administer home-made remedies to the sick mother or infant; fewer than one-third of families who took any action sought care from a source with at least some appropriate knowledge and resources (9% of the fatal cases).

Step 1: Recognition of Complications

Pregnancy is generally considered a time of well-being. Evidence for this view comes from focus groups with pregnant and recently delivered women in projects in rural areas of West Java, Indonesia, Bangladesh, and Nigeria; in rural and urban settings in Jamaica; and a periurban/urban area of Bolivia (see Appendix B). Labor and delivery, however, are anticipated with some trepidation: it is rightly considered, in the words of an Indonesian woman, the time when she is "caught between life and death." Jamaican women were even more expressive: "Having the baby is life and death … is not cutting a slice of cake" (Wedderburn and Moore, 1990:26). When the delivery is over, however, fears for the woman's life generally pass, and attention shifts to the newborn, for whom the first month of life is seen as challenging.

If a complication for either the woman or newborn arises, it is often considered fated, with little that can be done. In Indonesia, most pregnant women are even reluctant to talk about possible complications because they believe talking may be prophetic. Hessler-Radelet (1993), in exploring these issues in West Java, determined that a complication may not be considered serious if it disappears after birth (e.g., as does swelling), does not cause pain (malposition), is treatable with medicine (e.g., fevers and headaches), is common in pregnancy (swelling), is common outside of pregnancy (e.g., fevers), does not restrict daily activities, or is inevitable (fate or the will of God). Bolivian women, for example, consider swelling beneficial, believing it to result from the accumulation of blood that is necessary for the birth to take place. Even fever following delivery is considered a normal part of the birthing process for most women. Bolivian women are the exception, however, greatly fearing complications in the postpartum period; several of the conditions they describe as dangerous include chills and fever. Although Bolivian women with fever are treated first at home, they may also go to the hospital (if referred) because of the perceived seriousness of this complication (The Center for Health Research, Consultation and Education and Mother-Care/John Snow, Inc., 1991).

In Indonesia as well as in Bolivia, malposition is viewed by women and traditional birth attendants as the domain of the attendants. Following the seventh month, Indonesian women visit the attendants regularly for checks on the baby's position. Soothing massage is given to help pregnant women maintain the inner calm deemed necessary during pregnancy, labor, and delivery and to "correct" the baby's position (Ambaretnani, Hessler-Radelet, and Carlin, 1993). Referral may not take place until prolonged or obstructed labor is already under way. Bolivian women also seek attendants to correct the position of the baby through massage as well as manteo, a gentle rocking of the woman in a blanket (The Center for Health Research, Consultation and Education, and MotherCare/John Snow, Inc., 1991).

Bleeding stands out as the one maternal complication that commands attention almost everywhere. During pregnancy, or in the intrapartum or postpartum periods, the extent of bleeding is watched apprehensively—"bad" blood must leave the body, but "good" blood must remain. But there are exceptions even to this rule: Yoruba women of Nigeria stated that bleeding causes a pregnancy only to be interrupted, and the woman will take more than nine months to deliver (Public Opinion Polls, 1993).

The Quechua and Aymara women of Bolivia believe they must watch what enters their bodies to control the necessary shedding of blood. A balance of hot and cold must be achieved after delivery, as "dirty blood" of childbirth must be lost, but not to the point of excessive bleeding. Cold air and being unclothed, or underclothed, as is often the case in a hospital delivery, are feared, as coldness suspends the expulsion of the "dirty blood" and can cause swelling of the body and a series of very dangerous puerperal illnesses according to women (The Center for Health Research, Consultation and Education and MotherCare/John Snow, Inc., 1991).

Since traditional birth attendants oversee many deliveries, their recognition of symptoms of complications and knowledge of what to do about them can be critical for maternal and infant survival. In focus groups with attendants in Bolivia, Indonesia, and Nigeria, discussion of complications during any period of the reproductive process did not elicit much reaction. Their major role, as perceived by themselves, is the appropriate positioning of the baby. In West Java, for example, malpresentation was the major discussion topic. Malpresentations are managed by the attendants unless labor is prolonged, which one considered to be more than 8 hours. No attendant admitted a "difficult labor." Traditional birth attendants occasionally referred a patient to a health center or hospital, but most often the referral is declined, they stated, for reasons of expense, distance, and fear of being away from family and friends. In Indonesia and among the Fulani and Hausa of Nigeria, attendants said they will not even advise referral if they believe that the family will not comply.

Bolivian attendants mentioned pain in the back, vagina, and belly and vaginal bleeding as pregnancy-related complications, and almost all of them attributed these symptoms to the poor position of the baby. Management includes massage of the abdomen or advice on rest; a few attendants do refer women to formal health services. Delivery complications they mentioned include a bad position of the baby, emergence of the umbilical cord, arm, or leg prior to the head, and the umbilical cord wrapped around the baby. These complications mean that women have not taken care of themselves or pushed too early, stated the attendants, and referral to a doctor is in order, although some mentioned they try to accelerate delivery by cutting the amniotic sack with a razor blade or massaging the abdomen. Most said a normal delivery is 4 to 14 hours, but others said 1 to 3 days. If labor lasts longer than normal, the woman is given hot teas and massage.

From these qualitative studies it can be concluded that neither women nor traditional birth attendants perceive obstetric complications as necessarily requiring medical assistance. Apart from bleeding, problems are generally explained by and thought to require other types of more spiritual interventions, described below. The attendants' role is primarily involved with the baby's position, and to her most other problems are explained by this central factor.

Step 2: Deciding to Seek Care

Use of formal health care for delivery is rare not just because of perceptions about the services, lack of transport, distance, or funds. Home is strongly preferred as the place of birth for a variety of reasons. The decision to seek professional assistance for women suffering with complications is made even more complex by what is considered "appropriate care." Appropriateness depends on the perception of the interpreter as to the nature of the complication (physical versus spiritual) and of the seriousness of the condition. For bleeding, women most often seek medical care. But for other complications, such as obstructed labor—which is believed in one country to be caused by adultery during pregnancy—it is perceived to be more "appropriate" to consult a diviner and confess one's sins than to seek care from the formal health care system. According to the available qualitative studies, the demand for delivery services can be weak, even when other obstacles to access are absent.

In Indonesia, the period of labor is seen to place women in a vulnerable position between life and death—death may come to her as life is entering this world. However, the forces that keep a woman in labor on the side of life are linked to the home and the inner calm that is found there, not to a health facility. Because Indonesian women do not want to think about negative events or plan ahead for such (which will disrupt their sense of inner calm), they instead plan for the best-chance scenario, hoping that a complication will not arise. Should a complication arise during delivery with a traditional birth attendant (who assists with 85% of deliveries in Indonesia), she may or may not be referred, depending on whether the attendant believes the woman will accept referral. For bleeding, however, most Indonesian women state they would seek help at the health center from a midwife or doctor (Ambaretnani, Hessler-Radelet, and Carlin, 1993).

In Bolivia, the home provides privacy; no strangers to look, laugh, and touch the woman. As Bolivian women said (The Center for Health Research, Consultation and Education, 1991:23):

At home it's private, you don't pay anything. … They don't understand us well in the hospital; besides, my friends tell me that they touch everything—our genitals—and also there's a lot of health staff.

Although the complications of delivery that can kill may be recognized by Bolivian women, they are not considered amenable to medical intervention (The Center for Health Research, Consultation and Education, 1991). Locally recognized causes of maternal death include arrebato, a disease caused by failing to bind the woman's head, and the rising up of the magre, an organ formed behind the navel during pregnancy that should be "lowered" during childbirth with the womb, baby, placenta, and blood. There is no organ equivalent in biomedical anatomy, but the fear of the magre rising is so great that women bind their abdomens to keep it down. To prevent these complications, Bolivian women not only bind their bellies, but also wrap their heads, drink teas and other home preparations, and massage bellies (both to apply heat and to assure the proper position of the baby). Few women believe that these problems could be prevented by going to prenatal care and to the hospital for delivery.

The "secret of pregnancy," the feeling of shame, and the uncertainty about the outcome are so strong among the Fulani and Hausa in Nigeria that no preparation is made concerning labor and delivery by women and their husbands. The home is the natural place for delivery, more familiar and less threatening than a medical facility, and does not expose one to shame. Little or no assistance is sought by these women during delivery, even in their home (Public Opinion Polls, 1993).

The postpartum period, described most often by women as non-life-threatening, is generally filled with traditions and taboos. Many are harmless or beneficial, but some can interfere with the use of valuable postpartum services. Women in Indonesia are prohibited from leaving their home for 40 days, so they do not seek postpartum care from a formal health provider unless and until the situation becomes very grave (Ambaretnani, Hessler, and Carlin, 1993). Nigerian women go through a period of hot baths or massage with a hot napkin after delivery. Most Yoruba women claimed that this will rid the body of blood that has coagulated inside the woman during delivery. Failure to observe the hot bath period (7 to 40 days) is believed to meet with dire consequences—swelling and smelly vaginal discharge (Public Opinion Polls, 1993). Medical providers are not seen as beneficial at this time.

Step 3: Reaching a Facility that Can Provide Care

There are many logistical and provider barriers to access for women trying to reach appropriate services once they and their families have determined to do so (Leslie and Gupta, 1989; Sundari, 1992; Kutzin, 1993; Thaddeus and Maine, 1994). Lack of facilities with skilled personnel and equipment and supplies to support essential obstetric care, poor attitudes of providers, unavailability and cost of transportation, and the high costs of services are referred to repeatedly as the major obstacles to using services. Specific country-level efforts aimed at detailing these barriers to access essential obstetric services have only recently begun. The lack of services (including facilities, skilled personnel, drugs, and equipment) has been underscored in India, Bangladesh, Indonesia, Guatemala, and Bolivia, as well as in the African countries of Nigeria, Ghana, and Sierra Leone.

A UNICEF survey in three districts of India in 1993 found that two districts had a deficit of emergency obstetric care beds (those assigned for patients with emergency obstetric complications), assuming one bed is needed per 1,000 births. (Emergency obstetric care in their definition is a subset of what we have termed essential care for obstetric complications, consisting of caesarean sections, blood transfusions, as well as the availability of oxytocin, sedatives, and antibiotics.) In all three districts, the emergency obstetric beds were concentrated in district hospitals, resulting in a very high bed occupancy rate (110 percent in one of the districts). Between 4 and 10 percent of births were taking place in these first referral units, although 15 percent of births are expected to have serious complications (as described above). Actually, only 3.0 to 6.5 percent of complicated births were managed in these facilities.

Using a list of 38 drugs as essential for effective management of common obstetric complications that can lead to maternal death (World Health Organization, 1995), the UNICEF survey found that not one of the 18 facilities observed had all the drugs. In fact, less than one-half of the essential drugs were available in most sites. Trained personnel to provide caesarean surgery was missing in 14 of the 18 facilities, and none of the first referral units had staff with neonatal resuscitation skills. Most sites did have anesthesia capability or a surgical specialist who might need only a short orientation; all but three of the sites needed staff training in blood transfusion services.

Similar results were found in Indonesia and Bolivia. Each of three districts in South Kalimantan, Indonesia, had one or more government hospitals, but obstetricians were absent in two of the three districts. Although one midwife was available per 100-200 pregnant women, the midwives did not have the supplies or equipment (or the regulations to support them) to manage serious complications. Resuscitation equipment and skills were notably lacking, as was the skill to remove retained placenta, a common complication. Less than 5 percent of pregnant women delivered in the hospitals of the two districts for which there were data (Achadi et al., 1994).

In five districts in Bolivia, no obstetricians or anesthesiologists were found in two of the district hospitals. In a large urban district, no blood bank was available to support the district hospitals. There was a general shortage of client education materials. Privacy for provider-client counseling was lacking in the district hospitals (Seoane and Castrillo, 1995).

In Nigeria, Ghana, and Sierra Leone, there were declines in deliveries in seven referral facilities from 1983 to 1989, paralleling increased costs to patients for drugs and services (especially in Ghana and Nigeria). These declines are believed to coincide with the introduction of fees in five of the seven sites. The effect of user fees on the number of complicated obstetric cases seen at a referral site is mixed, although the scant data suggest that the patients with complications are continuing to come in for management (Prevention of Maternal Mortality Network, 1995).

In Indonesia, however, the obstacle of cost cannot be overestimated. According to one Indonesian woman: "I have to live through today, before I can think about tomorrow. I can't put away money for a hospital birth, because if I do, we may not be able to eat tonight" (quoted in Ambaretnani, Hessler-Radelet, and Carlin, 1993) Traditional care in Indonesia for prenatal and postpartum massages and child care plus delivery costs a total of U.S. $7.50 (1994), but a complicated delivery at a health center would cost nearly three times as much. If a hospital delivery with caesarean section is required, it could cost even 100 times the rate for traditional care (Achadi et al., 1994). Yet if a recognizable complication arises during delivery, most Indonesian families say that they will spare no expense to ensure that the woman and her baby are safe. But one husband expressed anger that his wife was referred to a hospital and then had a normal delivery there. He felt that they had spent a tremendous amount of money on something that they did not really need (Ambaretnani, Hessler-Radelet, and Carlin, 1993).

It is not only the lack of supplies, logistics, or costs that create barriers. Provider attitudes remain a major hurdle. An example is seen from a focus group report from Nigeria. Doctors in the Yoruba community believe that good prenatal care could prevent many of the complications they see, but that "pregnancy is considered to be a natural thing …" and "ignorance" is pervasive. Hence, women obtain prenatal care only late in pregnancy. Nurses and midwives claim this late registration is due to the fact that attendance for prenatal care is very expensive, and women want to delay registering as long as possible (Public Opinion Polls, 1993).

Step 4: Provision of Appropriate Care

The few studies of the quality of maternity care point to major deficiencies in the systems of care available and to a large gap between what facilities and life-saving skills are available and what is needed.

Indicators of the quality of emergency care for obstetric complications at a facility include the time interval between admission to treatment, facility trends in case-fatality rates for all complications, the caesarean section rate, and trends in numbers of deaths (maternal and perinatal); the proportion of perinatal deaths contributed by stillbirths or early neonatal deaths and proportion contributed by full-term babies may also be useful. Such data have only recently begun to be collected in most places. Medical records are typically the source of such information, and they are notoriously lacking or incomplete. Medical or verbal autopsies of maternal deaths, in which causes of death and avoidable factors are determined by a medical team that reviews each death, have proven to be a very useful way to monitor deficiencies.

Trends in the time from admission to treatment for emergency obstetric cases were observed in three referral sites in two African countries, Ghana and Nigeria. From a review of medical records, it was found that waiting time had increased from an average of 5 hours in 1983 to 15.5 hours in 1988 in Calabar (Ghana) for women who died of hemorrhage. In Zaria (Nigeria) the waiting increased from 3.5 to 6.9 hours between 1983 and 1988 (Prevention of Maternal Mortality Network, 1995).

Although data on admission to treatment for most complications is often lacking, caesarean sections require an operating theater; typically, the dates and times of procedures are recorded. The average admission-to-treatment interval for emergency caesarean sections for six referral facilities in three districts of India was 1.5 to 5 hours. These long intervals were caused by the need to locate doctors who were on call and time for them to come to the facility. In another setting, a delay was caused by a patient's relatives having to purchase anesthetics outside the facility. Case fatalities in these sites ranged from 2.8 percent to 6.9 percent. Low case fatality rates were found in some cases because the most serious cases were referred to another facility or left the facility against medical advice.

In Quetzaltenango, Guatemala, only 8 percent of patients had no wait upon arrival at the referral hospital; 45 percent waited up to 1 hour, and 47 percent waited more than 1 hour. Women in a study indicated that they would not accept referrals from traditional birth attendants to hospitals because of long waiting periods (O'Rourke, 1995).

Most incriminating of the quality of care is a nationally representative study of 718 maternal deaths in Egypt in 1992 (Ministry of Health, 1994). Avoidable factors were assigned by a local advisory group of a panel of doctors for each governorate that met weekly to review each maternal death in that area; their decisions were reviewed by a central advisory group at the national level. The leading avoidable factor was poor management and diagnosis by obstetric teams—47 percent. Patient factors, particularly delay in seeking (or compliance with) medical care, were blamed in 42 percent of deaths studied.

Review of the maternal deaths revealed that there was no referral system in place in the Egyptian health care system, and that this gap affected when and where cases were referred. A disproportionately large number of the women had seen a private practitioner who delayed referral to a hospital. No protocols for managing complications were available, and most cases were managed by junior staff. Both traditional birth attendants (dayas, who deliver nearly 60% of all births in Egypt) and general practitioners (who are sought only infrequently) played minor roles in the avoidable factors (12% each) (Ministry of Health, 1994).

A similar investigation of 1,173 maternal deaths in China in 1990 also pointed to maternal services as the highest contributor to avoidable factors (48%), followed by individual and family delays in using health care and transport problems. All factors occurred more frequently at village and township level (68 percent) than at county or provincial level (32 percent) (World Health Organization, 1994b).

In Guatemala, four departmental hospitals appear to have adequate numbers of beds and skilled staff available to manage serious complications, but they are highly underutilized for several reasons: traditional birthing positions (squatting, kneeling) are not permitted; no provision is made for respecting women's privacy; and language barriers make communication between providers and women and their families virtually impossible. Technical skills could also be questioned, as two cases of postsurgery tetanus had been reported in a previous year (Colgate-Goldman et al., 1994).

Lessons For Saving Lives

Each of the steps in the pathway to survival may require one or more intervention(s), especially in rural settings where hom e birthing is not only preferred, but may be a necessity given the distance and adequacy of available health services infrastructure to respond to life-threatening complications. in urban and periurban sites, where women intend to deliver in a health facility, and do so to a large extent, interventions may aim at promoting more selective use of facilities. In both rural and urban areas, efforts to improve the quality of care are needed to complement community efforts to overcome the barriers to access. Appendix B presents more details of the setting, implementation, and results of selected interventions at each of the four steps discussed below.

Step 1: Recognition of a Problem

What seems like a simple and obvious lesson comes from projects in West Java, Indonesia (Alisjahbana et al., 1995), Bolivia (Howard-Grabman, Seoane, and Davenport, 1993; Bower and Perez, 1993), and Forteleza in Northeast Brazil (Janowitz et al., 1988; Bailey et al., 1991): focus on families because they are most involved in observing the progress of pregnancies and the ones who will benefit most by a good outcome of pregnancy for both the mother and newborn. Families and those who influence families, not just pregnant women, should be the focus of efforts aimed at recognition of obstetric and newborn complications. They should also be informed about where to take a woman when such complications are recognized. The woman herself often can tell if she has a problem, but she may not be in any condition to advise or influence others. Specific groups—husbands, mothers-in-law, sisters—as well as pregnant women and traditional birth attendants, may be targeted. Providing messages through different media and consistently over time may emphasize the seriousness of the message. Caution must be exercised in crafting locally appropriate messages, however. For example, pregnant women in Indonesia did not want pictures of possible complications near them as they believed they were sure then to have the complications (Hessler-Radelet, 1993).

Traditional birth attendants provide labor, delivery, postpartum, and newborn care for the majority of women in many developing countries. Hence, it is tempting to train them in recognition of the danger signs of complications, but the results of such efforts have been mixed. Where women use formal medical services for birth if they have a complication, as appeared to be the norm for one-half the Guatemalan women in a study near the town of Quetzaltenango, training traditional attendants did not seem to improve the use of services for care of complicated cases (Schieber, 1993; Bailey, Szaszdi, and Schieber, 1994). Similarly, a randomized controlled trial in four cities of Latin America showed that an intensive intervention of home visits for health education had no impact on pregnancy outcomes—or even on use of postpartum services for women with complications—where women are already using services for labor and delivery and making a high number of prenatal visits (Villar et al., 1992). A home visit program with health education is probably not the answer to changing the behaviors of women who are already using services.

However, where women do not use formal services even when they suffer a complication, training traditional attendants may increase the use of services when needed, as shown in a rural area of West Java (Alisjahbana et al., 1995). However, there can be a backlash, whereby women no longer seek assistance because it is known that many women will be referred to an expensive setting. This problem must be addressed through a communications effort directed to women and their families. It is important that traditional birth attendants—and those who use their services—believe that making a referral adds to their credibility rather than detracts from it. Connecting attendants to medical facilities, and making them welcome at the facilities with the women, has also proved beneficial.

Step 2: Deciding to Seek Care

Little is known about the decision making in families with regard to seeking care for complicated obstetric or neonatal problems. However, it appears that it is families that make the decisions. They may be influenced by traditional birth attendants, but as seen in the periurban Quetzaltenango project (Bailey, Szaszdin, and Schieber, 1994) and in the rural areas of Forteleza, Brazil (where a project to connect attendants and their clients with hospital back-up had been operating for 10 years) (Bailey et al., 1991; Janowitz et al., 1988), attendants were usually by-passed (as was the health center level staff) when a complication occurred.

Hence, interventions should focus on families not only for recognition of problems, but also because they are the primary decision makers for determining whether and where they will seek care. While we know of only one effort to date to involve men directly in such efforts, in Inquisivi, Bolivia (Howard-Grabman, Seoane, and Davenport, 1994), men are the obvious target audience in many places because they control the cash reserves of the household or their permission must be sought for movement of women.

A positive image of the referral site has proven effective in increasing use in Quetzaltenango, Guatemala (Bailey, Szaszdi, and Schieber, 1994). With training on protocols to manage complications and a sensitization of staff to cultural issues so that the hospital doors were opened to families and traditional birth attendants, more women began to use these services. There was no advertisement of these changes, but the news passed rapidly by word of mouth: a ''woman-friendly" hospital appears to generate its own demand. The reverse of this was confirmed in the Danfa project in Ghana: women preferred traditional birth attendants because of their fear of anticipated treatment, described as painful and disrespectful, in the hands of medically trained staff (Eades et al., 1993).

Step 3: Reaching a Source of Care

The barriers to reaching services are manifold even if the family is positively disposed to using them. Distance to a referral site may be an obstacle, but three experiments aimed at ensuring transport, with taxi subsidies (Poedje et al., 1993), an ambulance (Alisjahbana et al., 1995), and a revolving fund for transport (Prevention of Maternal Mortality Network, 1995), did not alone increase the use of services. Transportation was not seen as the major obstacle in any of these three settings; rather, it was the costs or the perception of poor care at the end of the journey that proved more formidable. But improvements in transportation have been found useful in some settings. A low cost system of ambulances, relying on country boatmen to get women with complications to a hospital in rural Bangladesh, was an important part of a successful maternity care intervention there (Fauveau et al., 1991; Maine et al., 1996).

Posting certified midwives in rural health centers to provide obstetric care in times of emergencies has appeared to be effective in reducing maternal mortality in rural Bangladesh, where use of formal health care for labor and delivery is exceedingly low (Fauveau et al., 1991). However, posting certified midwives in rural health centers on provincial or national level is costly and difficult, as is shown by efforts in Indonesia. Certain obstetric problems may be managed or stabilized at a peripheral level prior to referral (e.g., antibiotics for infections, sedatives for eclamptic patients), but severe cases may by-pass the midwife anyway and be taken directly to a hospital, and midwives may not see enough cases of severe complications each year to make the investment worthwhile. In Indonesia, for example, one midwife serves approximately 2,000 people. With a birth rate of 28 per 1,000 population and assuming a level of severe complications at 15 percent of all pregnancies, a village-based midwife would see 10-20 complicated cases per year—if all complicated cases in her catchment area came to her. If a village-based midwife's tasks also included such efforts as improvements in women's health (e.g., family planning, iron/folate tablet distribution) and newborn and infant survival and health (breastfeeding promotion, cord care, eye care, warming of the newborn, as well as assistance for diarrhea and respiratory infections), however, it might make the investment worthwhile. Determining the most useful role for health centers and their staff in safe motherhood interventions is an important topic for further research.

Maternity waiting homes are another intervention that has been tried to address the problem of distance. Such homes provide a site where women who live far away or who have had a prior poor birth outcome can stay within reach of the hospital before labor begins. Studies in Ethiopia and Zimbabwe have reported improvements in maternal mortality for women who use them (Poovan, Kifle, and Kwast, 1990) and in perinatal mortality (Chandramohan, Cutts, and Millard, 1995). But as Chandramohan, Cutts, and Millard (1995:266) state: "For a maternity waiting home to be effective, a high proportion of women must attend an prenatal clinic, there must be an effective screening and referral system in place, and hospital delivery must be acceptable to mothers."

Yet another alternative, birthing homes, where medically trained personnel can be called to assist in a difficult delivery at some distance from a hospital, have been tried in Forteleza, Brazil, as well as in Tanjungsari in West Java. The birthing homes with traditional birth attendants did not appear as attractive to women as either home or hospital delivery. Even 10 years after the initiation of a birthing home in Forteleza, only one in ten women sought to deliver there. Traditional birth attendants at the birthing homes in Fortaleza did prove to be effective in screening appropriate cases for referral to the main hospital, so an intermediate care facility like birthing homes may help promote a cost-effective system of referral (Bailey et al., 1991).

Step 4: Obtaining High-Quality Care

From efforts in Guatemala, Nigeria, and Uganda, the lesson is simple and repeats lessons from other areas of public health: high-quality care generates demand. Use of delivery services increased where training of medical providers in advanced obstetric care was available (O'Rourke, 1995; Payne, Hooks, and Marshall, 1993; Mantz and Okong, 1994). This occurred although there had been no formal advertising of the improvements; word of mouth seemed to spread quickly, resulting in increased numbers of clients at a hospital within a matter of months.

Training has been the major intervention to improve quality of care—training in life-saving skills (Marshall and Buffington, 1991) and interpersonal skills communication for midwives (Family Health Services et al., 1993). While highly appreciated among midwives, training of one cadre of workers is not enough to sustain the practices taught. Programs must also support the use of these skills with policies that allow the trainees to carry out their new skills, improved management and supervision, an information system that allows them to monitor their own progress, logistics and supplies, and training of those to whom the midwives will refer women (typically, specialist doctors).

Protocols for the management of obstetric and neonatal complications have proved necessary (but not sufficient) for medical care providers to guide and coordinate their actions and know their limits and next steps. They have provided the standard against which to measure appropriateness of actions, especially if a maternal or perinatal audit is in place. Following institutionalization of neonatal care protocols in a referral hospital in Quetzaltenango, Guatemala, the hospital early neonatal mortality rate decreased from 33 to 26 per 1,000 women between 1989 and 1992, with a remarkable drop in avoidable factors assigned to the staff, emphasizing the pediatric staff's improved response to newborn problems (Schieber et al., 1995).

Audits themselves can provide an internal mechanism to monitor and change provider practices. Where they have been implemented, as in India (Bhatt, 1989) and Egypt (Ministry of Health, 1994), they have proved a powerful tool, going beyond raising awareness of the problems. Factors determined to be avoidable can be used in monitoring on-going services. Using both audits and obstetric rounds for on-going services, and bringing together all levels of providers responsible for emergency obstetric clients might suffice to prevent future maternal deaths due to avoidable factors.

Governments, in consultation with private-sector providers, need to establish a regional or national framework for maternity care, including the allocation of resources for each level of care. Many previous efforts to reduce maternal mortality in developing countries have foundered because they relied on attempts to train traditional birth attendants and screen high-risk pregnancies and refer women to expensive, distant, and ineffective sources of treatment. To save lives and reduce morbidity will require expenditures to upgrade facilities and train and supervise providers, as well as efforts to raise awareness at a community level of the signs of obstetric complications and how and where to seek appropriate care.



When facilities exist, service providers may lack the basic equipment and supplies, the person with the appropriate skills may be absent, or routine monitoring during pregnancy, labor, and delivery or in the puerperium may result in no follow-up of abnormal findings.


The problems we classify under steps 3 and 4 (getting to care, poor quality of care) would also influence the family's decision whether or not to seek care. We discuss them sequentially, though, as a convenient way to classify needed interventions.

Copyright 1997 by the National Academy of Sciences. All rights reserved.
Bookshelf ID: NBK233286


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (2.6M)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...