the impact of maternal health improvement on perinatal survival

INTERNATIONAL JOURNAL OF HEALTH PLANNING AND MANAGEMENT, VOL. 9,131-149 (1994)

THE IMPACT OF MATERNAL HEALTH IMPROVEMENT ON PERINATAL SURVIVAL:

COST-EFFECTIVE ALTERNATIVES

JULIA A. WALSH,' ANTHONY R. MEASHAM,' CHRISTINE N. FEIFER3 and PAUL J. GERTLER4

'Department of Population and International Health, Harvard School of Public Health. 665 Huntington Avenue, Boston, MAO2115, USA

'Health, Population and Nulrition Department, The World Bank, Washington DC. USA 'University of Cahfornia. Los Angeles

4~~~~ Corporation

SUMMARY Each year, an estimated half million women die from complications related to child birth either during pregnancy, delivery or within 42 days afterwards. When pregnant women have complications, their infants are at greater risk of becoming ill, permanently disabled or dying. For every maternal death, there are at least 20 infant deaths: stillbirths, neonatal or postneonatal deaths. Altogether, an estimated 7 million infants each year die perinatally (stillborn or deaths within the first week of life). Low cost, feasible, and effective intervention strategies include: a) improved family planning and abortion services; b) obstetric care at delivery; and, c) prenatal services.

Two hypothetical populations of one million (a low mortality and a high mortality country) are used to illustrate maternal and perinatal program strategies and priorities. In countries with high fertility, major reductions in maternal and infant deaths result both from reductions in the number of pregnancies through family planning and from improved obstetric care. Where fertility is already low, reductions result almost entirely from improved obstetric and prenatal care. The investments required are relatively low, while the potential gains are great. The cost to avert each death in a high mortality population is estimated between $800 and $1500 or as low as $0.50 per capita per year.

The priorities for programs targeting maternal and perinatal health depend on demographic, ecologic and economic factors, and should include the promotion of good health, not merely the avoidance of death. More operational research is required on various aspects of maternal and perinatal health; in particular, on the cost-effectiveness of different service components.

KEY WORDS: Cost-effectiveness; Neonatal; Perinatal; Maternal and child health; Fertility; Family planning

INTRODUCTION

More than 20 per cent of the population in developing countries is comprised of women in their reproductive years (United Nations, 1988). During a woman's

CCC 0749-6753/94/020131-19 0 1994 by John Wiley & Sons, Ltd.

132 J . A . WALSH ET At.

life, one of the greatest risks to health is child bearing. Pregnancy brings high risks for sickness, complications of delivery, disability, and death. Moreover, when pregnant women have complications, the infants these women bear have an increased risk for low birth weight (LBW), illness resulting from the complications of delivery, disability, and perinatal mortality. Maternal and perinatal health problems are inseparably linked: the major risk factors for disease and death among mothers and their newborns are the same. Not surprisingly, many of the interventions simultaneously improve both maternal and fetal health.

Maternal and perinatal mortality rates are indicators of the health of women of reproductive age and indirect measures of the quality of the health care system. In many places, health care during pregnancy is the initial contact with modern medicine that a woman seeks. Obstetric services, therefore, repre- sent a link to the general health care system, and thus have an impact on present and future pregnancies and on the well-being of the entire family.

The factors that affect the health of the mother and newborn can be divided into three: those that occur prior to conception; those during pregnancy; and those during delivery. Health service solutions alone will be inadequate without concurrent attention to the social and medical risk factors involved. Attention to a woman’s general health care needs, access to family planning, and readily available prenatal and obstetric care can, however, prevent the preponderance of perinatal and maternal morbidity and mortality.

In this article, we first discuss the public health significance of maternal and perinatal health problems. In the next section, we describe several specific maternal health programs, including: a) risk reduction prior to conception; b) strategies for pregnancy management to reduce risk and treat complications; and, c) ways in which programs might improve the outcome of pregnancy for both mother and infant. Program and research priorities are discussed in the final section.

PUBLIC HEALTH SIGNIFICANCE

Current levels and trends in the developing world

Maternal deaths and illnesses, LBW and perinatal deaths, are under-reported in most countries. The best figures for the developing world are merely estimates pieced together from multiple sources of varying degrees of reliability.

The World Health Organization (WHO) estimates that 500 000 women annually die of complications of pregnancy and delivery (Lopez, 1990). Ninety-nine per cent of these deaths occur in the developing world. It is estimated that 249 000 died in 1985, less than 1 000 in industrialized countries, 5 000 in non- market countries, and the rest in the other regions of the world (Bullatao and Stephens, 1991). The maternal mortality ratio (MMR) (the number of maternal deaths per 100 000 life births) measures the obstetric risk in a given pregnancy. The lifetime risk of maternal mortality is many times greater because each pregnancy adds to the total lifetime risk, The MMR ranges from 25 to 1660

MATERNAL AND PERINATAL HEALTH 133

in studies from developing countries (and averages 10 in developed ones), but an average woman in the developing world (excluding China) faces a lifetime risk of at least one chance in 33 that a pregnancy will result in her death. The highest overall MMR occurs in Africa, especially the Sub-Saharan region (640); the ratios are lower in Asia (420) and in Latin America (270). The maternal mortality rate (the number of maternal deaths in one year per 100 000 women of reproductive age, usually aged 15 to 49) combines the fertility rate (births per 1000 women of reproductive age) and the MMR. It is thus influenced both by the likelihood of becoming pregnant and by the risk of dying from that pregnancy (Edouard, 1985).

Three quarters of all maternal deaths can be attributed to one of three causes: hemorrhage, sepsis, or eclampsia (Herz and Measham, 1987). More than 25 per cent of these are from abortion complications. The remaining one quarter of maternal deaths includes complications of pre-existing illness (such as hypertension, diabetes, and heart disease), or such conditions as hepatitis, anemia or latent infections (e.g. tuberculosis, malaria, sexually transmitted and other genital infections) which can become active during pregnancy. The same complications which cause death can lead to chronic disability when they are less severe. For every death, there are up to 16 complications (Datta et al., 1973).

In conclusion, maternal deaths cause about 25 per cent of deaths of women between the ages of 20 and 30 years, while complications of pregnancy and delivery cause an enormous burden of illness. Most of this can be averted by better prenatal and obstetric care.

Perinatal health problems

At no other age is the risk of death so great as in the perinatal period. Seven million perinatal deaths occur annually-almost all in developing countries. Perinatal deaths, particularly stillbirths delivered at home, are frequently under- reported. In most developing countries, perinatal mortality rates (PMR) range between 40 to 60 per 1000; whereas, in developed countries, rates range from 6 to 10 (Belsey and Royston, 1987, quoting from WHO data banks).

Perinatal mortality is largely determined by delivery care and the maturity of the fetus, as reflected by birth weight and gestational age. Studies of the relative effects of birth weight and gestational age on mortality suggest that birth weight is the predominant factor (McCormick, 1985). Many infant deaths are potentially preventable by providing prenatal and obstetric care to women. A clinico-pathological study of 702 pei-inatal deaths in Nairobi revealed that one third were potentially preventable with better obstetric care, and more than 40 per cent of deaths resulted from problems avoidable by early caesarean section or assisted delivery (Lucas et al., 1983).

Reproductive tract infections from both sexually transmitted diseases and other genital infections result in a substantial amount of perinatal as well as maternal disease (Kundsin and Hipp, 1988). In parts of Africa, more than 10 per cent of pregnant women are seropositive for syphilis, but few are treated prior to delivery (Brunham et al., 1984). The consequences of these infections

134 J . A . WALSH ET AL.

are severe and include stillbirths and infant deaths, preterm delivery, intrauterine growth retardation (IUGR), low birth weight, and neonatal sepsis (Brunham et al., 1984; Berman et al., 1987). Most of these conditions can easily be identified and treated before and during pregnancy.

Low birth weight

The WHO defines LBW as a birth weight less than 2500 g (Kramer, 1987). As is well known, birth-weight-specific infant mortality begins to rise rapidly below this value. In 1982, of the 127 million infants born, 16 per cent (20 million) weighed less than 2500 g and over 90 per cent of these infants were born in developing countries. Table 1 shows how the mean birth weight varies between developed and developing countries. The cause of the majority of low birth-weight births occurring in both developing and developed countries remains unexplained, but the major risk factors are widely recognized. Two major processes determine birth weight: duration of gestation and intrauterine growth rate. Unfortunately, most studies from developing countries do not distinguish intrauterine growth retardation from preterm infants. However, analysis of a small number of studies suggests that more than half of LBW in developing countries probably results from IUGR, while in developed countries most results from preterm delivery.

In industrialized countries, neurodevelopmental handicaps among LBW infants occur three times more commonly than among normal birth weight infants (Kramer, 1987). The risk among infants weighing less than 1500 g is ten times that for normal weight infants. Comparable statistics do not exist for developing countries. LBW infants have twice the risk of congenital anoma- lies than normal weight infants, and very low birth weights have three times the risk. LBW infants also suffer more pulmonary disease, use health services more extensively, increase family health costs, and disrupt the normal function- ing of the family.

Indirect economic costs

Information is lacking both on the economic costs and the effects of maternal and infant morbidity on the women and infants, their families, the community, and health services. Although women’s efforts are often undervalued, they sup- port families through their productive labor; poorer women are more likely to be solely responsible for their families. The loss of the mother through death or disability thus means the loss of the nurturer, provider, and de facto household head. Chronic disability may incur more treatment costs than a quick death; however, the death of the mother is usually accompanied by the death of the infant from that pregnancy.

Treatment of LBW infants and neonatal illness can drain family and health service resources, particularly when neonatal intensive care is required. Indeed, when health resources are limited, health planners should weigh the cost of

MATERNAL AND PERINATAL HEALTH 135 Table 1. Mean birth weights and prevalence of low birth weight in selected countries.

Country Mean birth weight (g) Yn Low birth weight

Africa Egypt 3200-3240 7.0 Kenya 3143 12.8

Tunisia 3210-3376 1.3 Tanzania 2900-3 15 1 14.4 Zaire 3163 15.9

China 3215-3285 6.0 India 2493-2910 30.0 Indonesia 2760-3021 14.0 Iran 3012-3250 14.0 Iraq 3540 6.1 Japan 3200-3208 5.2 Malaysia 3027-3065 10.6 Pakistan 2170 27.0

Brazil 3 170-3296 9.0 Chile 3340 9.0 Colombia 3912-3115 10.0 Guatemala 3050 17.9 Mexico 30 19-3025 11.7

Canada 3327 6.0 USA 3299 6.9

Czechoslovakia 3327 6.2 France 3240-3 33 5 5.6 Hung a r y 3 1 4 4 3 162 11.8 Norway 3 500 3.8 Sweden 3490 4.0 United Kingdom 3310 7.0

Nigeria 2880-31 17 18.0

Asia

Lutin America

North America

Europe

Source: Kramer (1987).

neonatal intensive care, which may benefit only few infants, against the wider benefits from improved preventive care for maternal and perinatal health.

Summary

Maternal and perinatal health are closely linked. Efforts to improve the health of either pregnant women or the newborn will have synergistic effects on the health of the other. Death and sickness among these population groups occur

136 J . A . WALSH ET A L .

commonly, and the effects of this ill health are pervasive and costly to society. Low-cost, feasible, and effective intervention strategies are available and need to be addressed.

RISK REDUCTION PRIOR TO CONCEPTION

The goal should be to avoid unwanted pregnancies and to lower the risks when pregnancy occurs. This can be done by improving women’s health, through better health care and nutrition, education, and family planning.

The impact of family planning on maternal mortality depends on a country’s stage in the demographic transition (Fortney, 1987). A general rule of thumb is that where fertility is already low, increased use of contraception will have relatively little effect on fertility and maternal mortality. Conversely, where the maternal mortality rate is high and the MMR low, family planning is likely to have more impact on maternal deaths, especially if deaths due to abortion are high (Fortney, 1987). Since approximately 25 per cent of maternal deaths result from unsafe abortion, the provision of safe abortion services to back up other contraceptive services is a highly effective and low-cost way to reduce maternal mortality (Blacker, 1987).

Wider birth spacing can halt the general decline in health and nutritional status that some women experience with frequent, closely-spaced births. Tar- geted family planning can help avoid pregnancies among high-risk women, thus substantially reducing maternal mortality rates. Estimates for this reduction range from 25 to 50 per cent of all maternal deaths (Blacker, 1987; Rinehart et af., 1984).

Good health for all women requires an integrated set of actions including health services, community development, and education for female children. Despite recent gains in a few countries or among women of certain classes, most women in developing countries face social constraints in terms of status and decision-making power which limit their ability to safeguard their own health (Population Crisis Committee, 1988). They still lag behind men in edu- cational attainment and literacy. Africa scores lowest for educating its women, age of marriage, and high fertility; maternal mortality is highest in Africa. The Indian subcontinent (Middle South Asia) is the next lowest on these para- meters and has maternal mortality rates nearly as high as Africa.

Poor nutrition is still a widespread problem among women, despite recent efforts to provide food supplements and nutrition education. A survey of 80 developing countries in the early 1980s showed that between 20 to 45 per cent of the women aged 1 5 4 4 consumed insufficient calories daily (Hamilton et af., 1984). Anemia incidence is even higher; up to 80 per cent of Third World women of reproductive age may be anemic (Royston, 1982).

Priorities among preconception services that improve maternal and perinatal health include; 1) improved access to the full range of family planning; and 2) health education. Wherever possible, these priority services should be integrated into pregnancy care. Another alternative is to provide family planning


and treatment for reproductive tract infections and sexually transmitted diseases together and pregnancy care separately.

PREGNANCY, DELIVERY AND THE NEONATAL PERIOD

The ability of the health system to decrease the impact of pregnancy complications depends upon the capacity to identify high-risk women, prevent complications or treat them, and refer for skilled obstetric care. When a complication presents itself in spite of prenatal care, easily accessible obstetric facilities are required. A maternity care program in rural Bangladesh, which posted trained midwives at village health posts, with referral of complicated deliveries to a central maternity clinic, reduced maternal mortality by 60 per cent (Fauveau et al., 1991). This study suggests that midwives can improve maternal survival when given proper facilities, supervision and referral.

Prenatal care is cost-effective because it can reduce the numbers of women requiring skilled obstetric care, the proportion of LBW infants, and the incidence of perinatal disease (Harrison, 1985; Trivedi and Mavalankar, 1986; Brown, 1985; Chi et al., 1981; Donaldson and Billy, 1984; and IOM, 1985). However, these effects depend on identifying high-risk women early in pregnancy and providing them with special care.

It has not been possible to be prescriptive as to the specific components of prenatal care which reduce LBW and perinatal problems (IOM, 1985). To some degree, the content of prenatal care must be adjusted according to local technologies, economics, and population needs. Prenatal programs in poor countries must select carefully the components they will include for maximum efficiency and benefit. Table 2 prioritizes the important components of prenatal care for developing countries and the lowest levels of the health care system at which this care can occur.

Carefully considering which health care personnel will deal with what level of complications is critical. The generally accepted strategy involves a regional network of community risk assessment through prenatal care, and use of first- referral level facilities, usually a district hospital, for the management of high- risk cases and treatment of obstetric emergencies. Better care during pregnancy, labor, and delivery potentially could reduce maternal mortality rates anywhere from 50 to 80 per cent (Lettenmaier et al., 1988; Kwast et al., 1984; Walker et al., 1986) and PMR by 30 to 40 per cent (Lucas et al., 1983; Kaumitz et al., 1984).

Attention to the quality of postpartum and neonatal care are further concerns. This attention during postpartum visits is especially important in those areas where women only use clinic services for maternal health needs. About 70 per cent of all babies in developing countries are delivered by traditional birth attendants (TBAs) or relatives (Lettenmaier et al., 1988). With appropriate training and supportive supervision, TBAs can provide the basic care required by women with normal deliveries. Emergency referral is necessary for those with complications.

138 J . A. WALSH ET A L .

Table 2. Prenatal interventions for preventing maternal morbidity and mortality, low birth-weight, and perinatal mortality in developing countries.

Level Activity Purpose

Community

Primary

Primary

Primary

Primary

Referral

Referral

Referral

Educate for:

Monitor and treat or refer for skilled care: Screen and treat or refer for skilled care:

Treat intercurrent illness:

Provide appropriately:

Detect and treat:

Skilled delivery for women with:

Treat:

Nutrition, tobacco and alcohol use, recognize signs of premature labor and other serious complications and self- referral for care Uterine growth, weight gain, bleeding, presentation, hypertension, edema Reproductive tract infections, sexually transmitted diseases, diabetes, urinary tract infections, cardiac disease Such as diarrhea, respiratory infections, malaria Malaria prophylaxis in endemic areas, tetanus immunization, iron and folate supplements for anemia, nutritional supplements for malnourished women Premature labor and rupture of membranes Small pelvic size, poor obstetric history, open cervix, other risk factors (e.g. age, parity) Complications of spontaneous and induced abortions, ectopic pregnancies, hemorrhage

But, knowing what is good practice does not make it available. Existing reproductive care for women falls far short of needs and lags far behind what we know is feasible (Ratnam and Prasad, 1984). Access to quality family planning and abortion services remains limited. Few women in developing countries receive prenatal care. Indeed, ‘many women see no need for prenatal care unless they have complications’ (Lettenmaier et al., 1988; Bamisaiye et al., 1986; Leslie and Gupta, 1989). Worse still, the women who receive late or no prenatal care are most often young, primiparous, poor, of a racial or ethnic minority, undernourished, and more likely to smoke or drink (Kramer, 1987), all of which makes them higher risk and more in need of care. Of the women who do receive prenatal care, most make only one visit relatively late in pregnancy. Rural women are much less likely to receive prenatal care than urban women.

Systems to identify high-risk pregnant women even in developed countries have had poor sensitivity and specificity (IOM, 1985). Even though most of the adverse pregnancy outcomes occur in women who have several of the known risk factors, many maternal and fetal diseases occur in women with no easily identifiable risk factor. Because of this phenomenon, the sensitivity and specifi-


city of systems based upon history and physical examination average around 60 per cent (IOM, 1985). These systems identify only about half of the women who eventually have LBW infants, for example.

A reasonably reliable system for classifying women who will have poor pregnancy outcome is needed to contain the cost of services, so that those at high risk can receive special, and more expensive, care. An excess of women incor- rectly identified as high risk, means that scarce resources are used for the care of those who will not benefit from the service, and may experience complications from unnecessary diagnostic and therapeutic procedures. However, a misclassi- fication as low risk denies at-risk women access to beneficial services. Some of the systems evaluated, classified more than half the women as high risk when the incidence of poor pregnancy outcome was less than 10 per cent. These systems for classifying obstetric risk were evaluated in populations where the average percentage of LBW infants is 6 per cent (IOM, 1985).

The proportion of false positives and false negatives should be less in developing countries where the incidence of the adverse outcome is higher. For example, in developing countries where the incidence of LBW may be two to five times greater than in industrialized countries, the predictive value of the screening systems should increase. Conversely, since maternal mortality occurs much less frequently than LBW and perinatal mortality, a risk assessment system set up solely to prevent maternal complications will misclassify (false positives and false negatives) a large proportion of women. In order to have a successful maternal and perinatal health system based on risk assessment and referral care, improved screening criteria are needed.

Delivery and postpartum care suffer from non-use for reasons similar to those causing women to ignore prenatal care services. Delivery costs are a special barrier to formal services, which often are also competing against cultural norms favoring home delivery. Emergency care suffers in many places from the lack of roads, reliable vehicles, petrol restrictions, and lack of public transportation, which would integrate a referral system. Postnatal care is by far the least used service; in 1982 only 5 per cent of mothers in Costa Rica made one postpartum visit, although nearly 97 per cent delivered in hospitals (Lettenmaier et al., 1988).

In conclusion, a number of prenatal care and delivery interventions have been identified which can be applied depending upon local problems and resources. However, the actual health impact of an intervention depends on several factors: 1) efficacy rates, or the ability of an intervention to prevent or treat the problem; 2) accuracy of the diagnostic tests to identify those who would benefit from the intervention; 3) quality of care; 4) patient compliance; 5) coverage, in this case, the proportion of women using the service; and, 6) frequency of the health problem. All these plus cost and feasibility will need to be considered in planning a system for improving maternal and perinatal health.

140 J. A. WALSH ET A L .

APPLICATIONS: STRATEGIES FOR TWO STANDARDIZED POPULATIONS

In this section we use two hypothetical populations, both of one million persons, to illustrate maternal and perinatal program strategies and priorities. The populations include a low mortality country and a high mortality country.

Low mortality example

The first population, referred to here as ‘Lowmort’ (for low mortality), has already passed through much of the demographic transition. The fertility rate is 2.8. We assume the maternal mortality ratio is about 200 deaths per 100 000 live births, and that about 65 per cent of women of reproductive age use contraceptives. The infant mortality rate is 51 and the PMR 35.7 per 1000 live births; life expectancy at birth is 64.4 years.

‘Lomort’ has a reasonably well developed health system, a large proportion of women receive prenatal and delivery care from trained birth attendants and the average standard of living is moderate. The impact of increasing contraceptive prevalence on maternal mortality in ‘Lomort’ would be low. Therefore, we consider other avenues in ‘Lomort’.

‘Lomort’ could make substantial gains in maternal health with relative cost- efficiency by improving the quality of existing services, focusing on utilization, and coordinating with other sectors to implement social reforms which promote women’s health. Such social reforms include legalizing abortion, raising age-at- marriage, requiring universal education, and providing welfare to the worst-off, highest-risk groups. Changing cultural perceptions of women’s roles which block their ability to obtain care, and discouraging dangerous traditional prac- tices would also be important.

’Lomort’s’ health sector might implement a 10-year, limited effort, safe motherhood program emphasizing health and nutrition education, hiring additional staff and extending community outreach and prenatal screening, and improving existing services and their utilization. The program would offer additional training for personnel; provide health and nutrition education campaigns; and exam- ine facilities, emergency systems, and referral networks, and renovate these as necessary. The cost estimates for this kind of effort have already been pub- lished (Herz and Measham, 1987).

Based on experience in countries which have implemented similar efforts, the program’s impact on adverse events could be as much as a 65 per cent reduction in maternal mortality and morbidity, a 20 to 25 per cent reduction in the number of LBW babies, and a 35 per cent reduction in perinatal mortality over the 10-year period. Table 3 provides two sets of calculations using more conservative estimates of the progress. The lower estimate assumes the maternal mortality rate will be reduced by 25 per cent, LBW will experience a drop in incidence from 8 to 7.5 per cent and the PMR will be reduced by 12.5 per cent. The more optimistic scenario assumes maternal mortality will be reduced by 50 per cent, LBW incidence will drop to 7 per cent, and the PMR

MATERNAL AND PERINATAL HEALTH 141 Table 3. Current indicators of maternal and child health in ‘Lomort.’ A standardized

population of 1 million; and effects of a limited-effort safe motherhood program.

Indicator Before Limited-effort Limited-effort (conservative (moderate

estimate) estimate)

Population 1 000 000 1 000 000 1 000 000 Contraceptive prevalence (%)

(female acceptors 1544) 65 65 65 Fertility rate 2.8 2.8 2.8 Crude birth rate 24.5 24.5 24.5 Number of births 24 500 24 500 24 500

Maternal mortality ratio

Perinatal mortality rate (deathsi1000 live births) 35.7 31.2 26.8

Number of maternal deaths 49 37 25 Number of maternal morbidities’ 784 592 400 Number of perinatal infant deaths 875 76@ 656’ Number of low birth weight

Births averted nla 0 0 Maternal deaths averted n/a 12 24 Maternal morbidity averted nla I92 384 Perinatal infant deaths averted nla 109 219 Low birth weight averted nla 122 245 Program cost nla $480000 $480000

Cost per capita nla $0.48 $0.48 Cost/death averted n/a $3967 $1975 Costlevent averted n/a $1103 $550

(deaths1100000 live births) 200 1 503 100)

babies 1960 (8%)’ 1838 (7.5’”~)~ 17 15 (7Yq4

‘Number of maternal deaths x 16. Estimated ratio from Datta et al., (1973). *Predicted from regression. ’Assuming 25 and SOYO reductions in the maternal mortality ratio, conservative to moderate estimates for the effects of improved obstetric services from the literature. 4Assuming 0.50/0 and 1% reductions in low birth-weight incidence; these result in 6.2Y0 and 12.5% reductions in the number of low birth weight babies. 5Assuming 12.5% and 25% reductions in the perinatal death rate, conservative to moderate estimates for the effects of improved obstetric services from the literature. Source: Herz and Measham (1987).

will be reduced by 25 per cent. The total cost of the program for a population of one million would be about $480 000 annually or about $0.48 per capita. The estimated cost per death averted (maternal and infant) ranges from $3967 to $1975, and the estimated cost per adverse event averted ranges from $1 103 to $550. Adverse events include deaths, episodes of maternal morbidity, and LBW babies.

The fertility rate and numbers of births have been held constant in Table 3, in order to isolate the effect of improvements in prenatal and obstetric care alone. In reality, one would expect a further decline in fertility if services and community outreach were extended. The numbers of infant deaths and LBW babies avoided are much greater than those of adverse maternal events in spite

142 J . A. WALSH ET AL.

of the greater per cent reduction in rates for women. This highlights the fact that safe motherhood programs benefit both women and their infants.

High mortality example

Our second standardized population, ‘Himort’, is at an early stage of development. Health services are scarce, with a few exceptions, constrained by inadequate staffing, supervision, and supplies. Family planning is virtually unknown so that there are very few modern contraceptive users, except in the capital. Unsafe abortions are relatively common. Most women, about 90 per cent, deliver at home; about 70 per cent of births are attended by untrained traditional midwives or family members. The few better-off places are served by health centers and a district hospital but maternal health is not given priority.

Life expectancy at birth is 51 years; the infant mortality rate is 129 and perinatal mortality is an extremely conservative estimate of 51.6 deaths per 1000 live births. Maternal mortality is believed to range from 800 to 1400 in the various regions. Health officials estimate a national average MMR of 1000 deaths per 100 000 live births. The total fertility rate is 6.9 and the crude birth rate is 49.5.

The problem facing ‘Himort’ is how to improve maternal and perinatal health with severe limitations on resources, inadequate facilities, and inadequate transportation. Each region must individually assess its restrictions and needs and may need to plan an appropriate program which differs from those of other regions, Among the options available are preventive activities. Family planning and safe abortion services are the most cost-effective preventive activities. A second option would strengthen prenatal care and train birth attendants for pregnancy management. We begin by illustrating the likely costs and benefits of family planning alone, and then discuss the implications of a broader maternal health program.

Family planning alone

If ‘Himort’ only offered family planning services, substantial improvements in maternal and infant health outcomes could result. Social marketing, community-based distribution, postpartum projects, and voluntary sterilization services appear to be among the most cost-effective models for providing family planning services. A community-based delivery system, using TBAs and local health personnel, might be a good model for ‘Himort’.

Table 4 presents demographic data and health statistics for three scenarios: increases in the prevalence rate for contraception to 20,40 and 6O%, respectively. Through the increase in contraceptive users, fertility rates decline and the number of births is reduced, thereby reducing exposure to pregnancy risk. The MMR is held constant throughout Table 4, however, emphasizing that unlike the maternal mortality rate, it is unaffected by overall changes in fertility. In spite of the constant MMR, the number of maternal deaths and morbidity drops remarkably. Using the new fertility rates, calculated on the basis of


Table 4. Family planning: Effects on maternal and child health of three scenarios based on contraceptive prevalence rates'.

Health indicator 20% 40% 60% Prevalence Prevalence Prevalence

rate rate rate

Fertility rate' 5.3 4.3 2.6 Number of births 41 170 34 183 28381

Maternal mortality ratio

Perinatal mortality rate

Number of maternal deaths 412 342 284 Number of maternal morbidities3 6592 5412 4544 Number of perinatal infant deaths 2017 1586 1246 Number oflow birth weight babies' 5764 (14%)' 4444 (13Yo)j 3406 (12%)'

(deathdl 00 000 live births) 1000 1000 1000

(deaths/ 1000 live births) 494 46.44 43.94

Births averted Maternal deaths averted Maternal morbidity averted Perinatal infant deaths averted Low birth weight averted

8330 15317 21 119 83 153 21 1

1328 2448 3376 531 968 1308

1661 2961 4019

Program cost' $500 000 $1 500 000 $4 500 000 Cost per capita $0.50 $1.50 $4.50 Cost/death averted $806 $1338 $2962 Codevent averted $139 $229 $505

'Prevalence rate is per cent of women 1544 using contraceptives. 'Predicted from regression. 'Maternal deaths x 16. Estimated ratio from Datta et al., (1973). T h e decreases in perinatal mortality from the original high of 51.6 through contraceptive prevalence rates of 20,40 and 60%, are 5 , 10 and 15% each, respectively. 'Decreases in incidence of low birth weight of 1% for each 20% increase in contraceptive prevalence, together with decreases in numbers of births, result in total decreases in number of low birth weight babies of 22,40 and 54%. 'Longer time periods are obviously needed for achieving higher rates of contraceptive prevalence. These longer time periods are likely to involve adjustment costs not included here.

the contraceptive prevalence rates, revised rates of LBW are also derived. Peri- natal infant death rates are assumed to drop 5 per cent for every 20 percentage point increase in contraceptive prevalence.

The demographic changes presented in Table 4 may take several decades. Costs presented assume that expenses increase exponentially as population coverage goals increase, but these estimates ignore adjustment costs which could occur over the long periods and which may be necessary for major increases in contraceptive use to take place. The data do show that the cost-efficiency of family planning for improving maternal and perinatal health decreases as contraceptive prevalence increases. The cost per death averted ranged from $806 to $1338 to $2962 as contraceptive prevalence reaches 20, 40 and 60% of couples at risk. Costs per adverse event averted range from $139 to $229 to $505.


Table 5. Estimated impact of comprehensive maternal health plan and cost information for family planning only and family planning with limited and moderate obstetric

program in ‘Himort’ (US$).

Family Family planning and planning and

Family limited moderate planning obstetric obstetric

Before only program progrm

Population Contraceptive prevalence (“h)

(female acceptors 1544) Fertility rate Crude birth rate Number of births

Maternal mortality ratio (deathdl00 000 live births)

Perinatal mortality rate (deaths 1000 live births)

Number of maternal deaths Number of maternal

Number of perinatal infant

Number of low birth weight

Births averted Maternal deaths averted Maternal morbidity averted Perinatal infant deaths averted Low birth weight averted

Program cost

morbidities

deaths

babies

Cost per capita Coddeath averted Codevent averted

1 000 000 1 000 000

0 20 6.9 5.3

49.5 41.17 49 500 41 170

1000 1000

51.6 49 495 412

7920 6592

5643 2017

7425 (15%) 5764 (14%)

nla 8330 nla 83 nla 1328 nla 537 nla 1661

$500 000 $0.50

nla $806 nla $139

1 000 000 1 000 000

20 20 5.3 5.3

41.17 41.17 41 170 41 170

800’ 600’

47.8’ 36.8’ 329 247

5264 3952

1968 1515

5352 (1 3%) 4940 (12%)

8330 8330 166 248

2656 3968 586 1039

2073 2485

$0.96 $2.00 $1303 $1554 $179 $258

$980 000) $2 ooo 0004

‘Assumes additional decreases in the maternal mortality ratio of 20% and 40% for limited and moderate efforts, respectively, over the decreases achieved by family planning alone. ’Assumes additional decreases in the perinatal mortality rate of 12.5% and 25% for limited and moderate efforts, respectively, over the decreases achieved by family planning alone. ’Combines $500000 cost of family planning with $480000 cost of limited effort. 4Combines $500000 cost of family planning with $1 500000 cost of moderate effort.

In the next section, it is assumed that ‘Himort’ has adopted the goal of a family planning prevalence rate of 20 per cent, and that it also undertakes additional reforms in the area of maternal health care.

Comprehensive maternal health program

The goals for a comprehensive maternal health program in ‘Himort’ would differ according to the existing infrastructure, socioeconomy, and ecology of


particular sub-areas. The first effort, a limited one, is most appropriate for the very poor, isolated desert and rain forest areas that make up the bulk of the country. The plan includes resources necessary for a limited prenatal care and birth attendant training program. The second, moderate, effort would best be applied in areas that are not as poor, are densely populated, have existing health facilities including a hospital, and where women utilize health services more. The moderate plan includes the resources necessary for a moderate prenatal care and birth attendant training program, again in addition to the family planning effort. (We have adapted the program design and cost estimates developed by Herz and Measham ( I 987), who provide a more detailed discussion of these items.)

Limited effort. The goals of the limited effort are to reduce the MMR by 20 per cent through prenatal and delivery care, and trained birth attendants. In addition, we expect the plan to reduce the incidence of LBW babies to 13 per cent (dropping from 14 per cent with family planning alone) and perinatal infant mortality by an additional 12.5 per cent. These decreases would be the result of reducing the number of high-risk pregnancies and of providing better care to those who become pregnant. The components of the limited plan would be as follows:

- upgrading existing facilities to ensure the availability of maternal health and establishing four more centers (two with caesarean section and surgical family planning capacity);

- investing in an emergency transport system, one four-wheel drive vehicle for each new center, so that more women can reach the existing service areas;

- introducing risk-screening and developing plans for at-risk women to deliver in health facilities; three mobile units equipped with radios and staffed by three health personnel; and maternity villages for those referred to facilities for delivery;

- strengthening all community-based services so that prenatal and uncompli- cated delivery care is readily available for training all TBAs and providing them with basic medications, compensation for family planning activities, and radios;

- coordinating outreach so that facilities and trained personnel are efficiently used; and,

- conducting research activities to identify the most effective approaches.

Depending on the community or region, other potential components could include: training nurse-midwives to enlarge the pool of blood transfusion, surgical family planning and caesarean section providers; strengthening health services management and giving higher priority to maternal health; and, encouraging community groups to become involved in women’s health and safe motherhood.

The total cost of the program for a population of one million, including


family planning program costs, would be about $980 000 annually or about $0.98 per capita. The cost per death averted (maternal and infant) would be about $1303, and the cost per adverse event averted about $179.

Moderate efSort.The goals of the moderate effort are to reduce the MMR by 40 per cent through prenatal and delivery care, facility development, and training of birth attendants. In addition, we expect the plan to reduce the incidence of LBW babies to 12 per cent and perinatal infant mortality by an additional 25 per cent. These reductions are more conservative than most cited in the literature. The components of the plan can be listed as follows:

- establishing a community outreach system for prenatal care, providing nutrition advice and pregnancy risk screening, making appropriate referrals and encouraging use of health facilities;

- increasing the number of health posts to one per 10 000 population using community aid, and training all TBAs in outreach and routine care activities;

- increasing the number of health centers by building five new ones in 5 years, and using these as referral centers for pregnancy complications;

- adding ten maternity beds to the district hospital and an operating room with capacity for high-risk deliveries, caesarean sections, and surgical contraception;

- training additional health personnel at each level so that a regional network of services exists with increasingly complex services offered at health posts, health centers, and the hospital, respectively; and,

- developing an emergency transport system and training all personnel for appropriate referrals.

The total cost of the program for a population of one million, including the family planning program costs, would be about $2 000 000 annually or about $2 per capita. The cost per death averted (maternal and infant) is about $1554, and the cost per adverse event averted is about $258. Table 5 summarizes the impact and cost information for the family planning effort alone, and the family planning effort plus limited and moderate obstetric efforts.

Summary

As the examples illustrate, efforts for maternal and perinatal health should involve appropriate combinations of family planning, prenatal care, and obstetric improvements. Health services should also give priority to screening and referral networks and emergency transport systems.

In countries with high fertility, major reductions in maternal and infant deaths result both from reductions in the number of pregnancies through family planning and from improved maternal care. Where fertility is already low, reductions result almost entirely from improved maternal and perinatal care. Altogether, the investments required for safer motherhood and healthier starts to life are relatively low, while the potential gains are great. However, the health system


must have the capacity to make services available to people in all parts of the country. Extending service coverage is especially critical for those countries whose current budget is allocated largely to the urban centers.

Finally, a word of caution is in order. The declines in adverse outcomes that we have suggested are no more than best estimates of the likely impact of the measures proposed, based on the limited evidence available from the literature, but they do offer a focus for debate and question.

PRIORITIES

Priorities for resource allocation

Maternal mortality, perinatal mortality, high fertility and LBW are all high- priority problems. Technically feasible and affordable approaches exist which can significantly reduce the incidence of perinatal and maternal disease and death, and increase the prevalence of contraceptive use. Because maternal and perinatal health problems have long been neglected, they should be at the top of the priority list in most countries-especially those in which maternal mortality, neonatal mortality, and LBW are high (i.e., South Asia and sub- Saharan Africa).

The priorities for particular programs targeting maternal and perinatal health depend on the demographic situation, particularly the fertility rate and level of contraceptive use, and on ecological and economic factors. When countries are deciding on the balance between maternal health and other health programs, it is worth considering that maternal health affects the health of infants, children, and the dependent elderly.

The focus in this article has been on mortality and morbidity; the other side of the equation, the magnitude and quality of health, has largely been ignored. When implementing safe motherhood programs, the priority should be to promote good health, not merely to avoid death. Countries should look at the quality of life for newborn children and the quality of life for women both during and after the reproductive period. These priorities will require intersectoral collaboration, strong political will, and a willingness to recognize the importance and value of women.

Family planning and access to safe abortion as a back-up are clearly a priority. Successful family planning programs require demand for spacing children and limiting family size. Health improvements must be accompanied, therefore, by concomitant efforts to improve female education, and women’s economic opportunities. Declines in fertility will then make reproductive health care services more affordable.

Access to prenatal and competent obstetric care also merits very high priority. The content of these services, their quality, and utilization certainly influence their effectiveness. Indeed, merely improving the quality of existing care and increasing use of existing services, could have a substantial impact in many countries.


Priorities, for operutional research

Maternal and perinatal health are severely understudied areas. Where programs have been implemented which exert their impact on the health of women (such as nutrition supplementation), the effects they have on infant mortality have more often been the focus. We have little evidence for factors causing mortality rate declines, no measures of the economic impact of death or disability, nor any firm idea of the prevalence or duration of maternal and perinatal illness. There is insufficient information to relate different service components directly with effectiveness and cost. More field research is needed in this area, and more programs should include an evaluation component.

REFERENCES

Bamisaiye, A., Ransome-Kuti, O., Famurewa, A. A. (1986). Waiting time and its impact on service acceptability and coverage at an MCH clinic at Lagos, Nigeria. J Tropical Pediatrics 32, 158-161.

Belsey, M. A,, Royston, E. (1987). Overview of the Health of Women and Children. Presented at the International Conference on Better Health for Women and Children through Family Planning. Nairobi, Kenya.

Berman, S. N., Harrison, R., Boyce, W. T., Haffner, W. J. J. , Lewis, M., Arthur, J. B. (1987). Low birthweight, prematurity, and postpartum endometritis. J. Am.

Blacker, J . G. C. (1987). Health impacts of family planning. Health Policy and Planning Mrd. ASSOC. 257,1189-1194.

2, 193-203. Brown, S. (1 985). Can low birthweight be prevented? Family Planning Perspectives 17,

1 1 2- 1 1 8. Brunham, R. C., Holmes, K. K., Eschenbach, D. (1984). Sexually transmitted diseases

in pregnancy. In: Holmes, K. K., Mardh, P.-A., Sparling, F., Wiesner, P. J . (Eds). Sexually Transmitted Diseases. New York: McGraw-Hill Book Co., 782-815.

Bullatao, R. A,, Stevens, P. (1991). Estimates and projections of mortality by cause: A global overview, 1970-2015. In: Mosley, H., Jamison, D. (Eds). Health Sector Priorities Review.

Chi, I. C. Agoestina, T., Harbin, J. (1981). Maternal mortality at 12 teaching hospitals in Indonesia; an epidemiological analysis. Intl. J. Gyn. & Obst. 24,259-266.

Datta, S . P., Srinivas, D. K., Kale, R. V., Rangaswamy, R. (1973). Evaluation of maternal and infant care in a rural area. Indian J. Med. Sc. 27, 120-128.

Donaldson, P. J., Billy, J. 0. G. (1984). The impact of prenatal care on birthweight. Medical Care 22, 177-188.

Edouard, L. (1985). The epidemiology of perinatal mortality. World Health Statistics Quarterly 38, 289-301.

Fauveau, V., Stewart, K., Khan, S. A., Chakraborty, J. (1991). Effect on mortality of a community-based maternity-care programme in rural Bangladesh. Luncet 338, 11 83-1 186.

Fortney, J. A. (1987). The importance of family planning in reducing maternal mortality. Studies in Family Planning 18, 109-1 15.

Hamilton, S., Popkin, B., Spicer, D. (1984). Women and Nutrition in Third World Coun- tries. New York: Praeger.

Harrison, K. A. (1985). Childbearing, health, and social priorities: a survey of 22,774 consecutive hospital births in Zaria, Northern Nigeria. Br. J. Obstet. Gynecol. Supple- ment 5,l-119.

MATERNAL AND PERINATAL HEALTH 149 Herz, B., Measham, A. R. (1987). The safe motherhood initiative: proposals for action.

World Bank Discussion Papers, No. 9. Washington DC: World Bank, May. Institute of Medicine, IOM (1985). Investigators of the Johns Hopkins Study of Cervicitis

and Adverse Pregnancy Outcome. Association of Chlamydia trachomatis and Mycop- l a m a hominis with intrauterine growth retardation and preterm delivery. Am. J. Epide- miol. 129, 1247-1257.

Kaumitz, A. M., Spence, C., Davidson, T. S., Rochat, R. W. Grimes, D. A. (1984). Perinatal and maternal mortality in a religious group avoiding obstetrics care. Am. J. Obstet. Gynecol. 150, 826-83 1 .

Kramer, M. S. (1987). Determinants of low birthweight: methodological assessment and meta-analysis. Bull. W H O 65,663-137.

Kundsin, R. B., Hipp, S. S. (1988). Impact on the Fetus of ParentalSexually Transmitted Disease. New York: New York Academy of Sciences.

Kwast, B. E., Kidane-Mariam, W., Saed, E. M., Fowkes, F. G. R. (1984). Report on maternal health in Addis Ababa, Sept, 1981-Sept 1983: a community based survey on the incidence and etiology of maternal mortality and the use of maternal services; and confidential inquiries into maternal deaths in Addis Ababa, Ethiopia. Addis Ababa, Ethiopia: Save the Children Federation.

Leslie, J., Gupta, G. (1989). Utilization of Formal Services for Maternal Nutrition and Health Care. Washington, DC.: International Center for Research on Women.

Lettenmaier, C., Liskin, L., Church, C. A,, Harris, J. A. (1988). Mothers’ lives matter: Maternal health in the community. Population Reports Series L , No. 7. Baltimore, MD: Johns Hopkins University, Population Information Program, September.

Lopez, A. D. (1990). Causes of death: an assessment of global patterns of mortality around 1985. World Health Statistics Quarterly (1983). 43,91-104.

Lucas, S. B., Mati, J. K. G., Aggarwal, V. P., Sanghvi, H. (1983). The pathology of perinatal mortality in Nairobi, Kenya. Bull. Soc. Path. Ex. 76, 579-583.

McCormick, M. (1985). The contribution of low birthweight to infant mortality and childhood morbidity. New Engl. J. Med. 312, 82-90.

Population Crisis Committee (PCC) (1988). Country rankings of the status of women: poor, powerless andpregnant. Washington DC: PCC, Jun. (Briefing Paper No. 20).

Ratnam, S. S., Prasad, R. N. V. (1984). Inadequacies of present health services and strategies to improve maternal and child health in developing countries. Int. J. Gyn. dt Obst. 22,463466.

Rinehart, W., Kols, A., Moore, S. H. (1984). Healthier mothers and children through family planning. Population Reports, Series J , No. 27. Baltimore, MD: Johns Hopkins University, Population Information Program.

Royston, E. (1982). The prevalence of nutritional anemia in women in developing countries: a critical review of available information. World Health Stutistics Quarterly 35,52-9 1 .

Trivedi, C., Mavalankar, D. (1986). Epidemiology of low birthweight in Ahmedabad. Indian J o f Peds 53,195-800.

United Nations (1988). Demographir Yearhook, 1988, UN, New York. Walker, G. J. A., Ashley, D. E. C., McCaw, A. M., Bernard, G. W. (1986). Maternal

mortality in Jamaica. Lancet 1,486488.

the impact of maternal health improvement on perinatal survival

Documents