Click for next page ( 26


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 25
Reproductive Patterns and Women's Health In comparison with the levels in industrialized countries, reproductive mortal- ity remains high in most developing countries, particularly in rural areas. In developing countries, maternal mortality is estimated to range Tom approxi- mately 100 to 700 or more maternal deaths per 100,000 live births, with the highest levels in rural areas of sub-Saharan Africa and South Asia. These rates imply an estimated 500,000 maternal deaths annually, with over 98 percent occulting in developing countries (World Health Organization, l985b). In indus- trialized countries, maternal mortality was at about that same level at the turn of the century but has declined to a current level of less than 10 maternal deaths per 100,000 live births, with under 10,000 deaths per year. Thus, although a majority of pregnancies proceed normally and are not associated with significant health problems, there remains considerable potential for reducing the risks associated with pregnancy and childbearing. The most important causes of reproductive injury, morbidity, and mortality in developing countries are obstructed labor (and ruptured uterus), postpartum hem- orrhage, pregnancy-induced hypertension, postpartum infection, and the compli- cations of unsafe abortion. The relative importance of each of these causes varies among populations and within the same country, depending on living conditions and the availability of medical care. In this chapter, we summarize the evidence concerning the relation between reproductive patterns and women's health. As described in Chapter 2, there are at least two ways in which changes in reproductive patterns may improve women's health. First, each time a woman becomes pregnant or gives birth, she is susceptible to an increased risk of illness, injury, or mortality that she does not 25

OCR for page 25
26 co~rRAcEprioN AND REPRODUCTION face when she is not pregnant. Women who have more pregnancies or give birth to more children encounter this basic risk more frequently than women with lower fertility. Thus, a reduction in the number of pregnancies or births that women have will improve their reproductive health simply by reducing the frequency of exposure to the basic risk of illness, injury, and mortality associated with pregnancy and childbirth. Second, women experiencing certain types of pregnancies undergo an addi- tional risk. As described in Chapter 2, it has been hypothesized that higher-risk pregnancies include: first pregnancies or births, pregnancies of young women, or the combina- tion; pregnancies in women at high panty, pregnancies of older women, or the combination; pregnancies following soon after previous pregnancies; and pregnancies that are terminated by unsafe induced abortion. The effects of these higher-risk pregnancies are the focus of much of this chapter. However, before reviewing this research, we discuss the types of evidence on which our conclusions about the associations between women's health and reproductive patterns are based. SOURCES OF EVIDENCE Measurement and Data There are two standard measures of the frequency of maternal death in a population. The fast is the maternal mortality ratio, which is the ratio of the number of deaths due to pregnancy or childbirth to the number of pregnancies. However, in practice, even in an industrialized country, it is impossible to count the total number of pregnancies. Thus, by convention, the number of live births is used as the denominator. The second measure is the maternal mortality rate or the maternal cause- specific mortality rate, which is calculated by dividing the number of deaths due to pregnancy and childbirth by the number of women of reproductive ages. Unfortunately, these two measures are often used interchangeably in the public health literature, and reports of rates often are actually referring to ratios. For clarity, in this report the denominator (either women or live births) is always stated explicitly. Information on levels of maternal mortality and other indices of reproductive health is difficult to obtain, especially in developing countries. It is even more difficult to obtain data adequate to investigate the possible associations between reproductive patterns and maternal health. There are several reasons for the paucity of data on levels and trends in maternal mortality. First, vital registration

OCR for page 25
REPRODUCTIVE PATTERNS AND WOMEN'S HEALTH 27 in most developing countries is seriously incomplete: relatively few deaths, especially in rural areas, are registered. In addition, deaths occurring very early in pregnancy (such as those caused by ectopic pregnancies), those caused by com- plications of induced abortion, and those attributed to other causes (such as malaria and hepatitis) are often not classified as due to reproductive causes. Even in industrialized countries where the numbers of births and deaths are known, causes of maternal death are not well reported. The majority of studies discussed in this chapter define maternal mortality as a pregnancy-related death, due to either direct causes (pregnancy complications) or indirect causes (other diseases such as heart disease exacerbated by pregnancy3. Poor vital registration is a problem that affects the availability of data on infant and child mortality as well as maternal mortality. However, in the case of infant and child mortality, vital registration data can be supplemented by levels and associated risk factors from national or local fertility surveys or from field studies in small areas in which data are collected over time. Data on maternal mortality, by contrast, are considerably more difficult to obtain from these two sources (Zimicki, 1989) . Reproductive histories and histories of child death are usually collected from women, but maternal mortality data are difficult to collect in this way, since women who died cannot be interviewed, and collecting data on the reproductive histories of decedents from other sources is difficult. The Demo- graphic and Health Survey Project plans to test an indirect method of obtaining maternal mortality estimates asking adults in households if they had a sister who died shortly before or after childbirth. ~Measunng maternal mortality of entire populations in developing countries is also difficult because death from reproductive causes remains a relatively rare event. For example, maternal mortality ratios of 700 deaths per ioo,ooo live births have been reported in some parts of Africa. But even with these high levels of death, a sample of 100,000 women would be necessary to yield 700 maternal deaths. By contrast, in these same populations, current levels of infant mortality rates imply that there will be between 10,000 and 20,000 infant deaths per 100,000 live births. Since maternal death is a relatively rare event, accurate measurement of maternal mortality rates requires collecting data from a much larger population than would be required for accurate measurement of infant and child mortality. The problem of sample size is somewhat smaller in the case of illnesses associated with pregnancy, since maternal morbidity is more common than mortality. However, since morbidity is a less well-defined event than mortality, it tends to be even more poorly reported. Thus, information with which to evaluate the association between reproduction patterns and maternal illness or mortality generally is not available, because data on both maternal illness or mortality and on reproductive history are required for a very large population. This report draws on studies of the association between maternal health and reproductive patterns that are based on three sources of data: general population- based studies, hospital-population studies, and hospital case series.

OCR for page 25
28 CO=RACE"iON AND REPRODUCTION Population-based studies allow nearly complete counts of live births and maternal deaths, although they suffer from the same types of underestimation as vital registration systems. Very few population-based studies concerning mater- nal mortality in developing countries have been published: four from Bangladesh (Chen et al., 1975; Khan et al., 1986; Alauddin, 1987; Koenig et al., 1988a), one from Ethiopia (Kwast et al., 1986), one from Egypt (Fortney et al., 1985), one from Gambia (Greenwood et al., 1987), and one from Jamaica (Walker et al., 1985~. Preliminary results from population-based studies in India (Bhatia, 1985) and Bangladesh (Lindpaintner et al., 1982) are available. In addition, one hospi- tal-based study from Lusaka, Zambia (Mhango et al., 1986), arguably covers a sufficient proportion of the population (85 percent of births, 90 percent of deaths) to be considered a population-based study. However, the number of deaths in each study is relatively small, so estimates for subgroups of women with particu- lar characteristics are unstable. Characteristics of the data and types of analysis used in these and the studies cited are presented at the end of this chapter in Appendix Table 3.A. Studies of hospital patients, particularly those carried out in referral hospitals, provide less accurate estimates of the incidence of maternal mortality or morbid- ity than population-based studies. There are several reasons to expect that maternal mortality ratios from hospital studies are unrepresentative. First, hospi- tal patient studies reflect the experience of only a proportion of the population during only part of the risk periodusually at the time of or immediately after the pregnancy outcome. Thus, deaths occurring outside the hospital and those occurring some time before and after delivery and discharge are usually missed. Second, rural women are likely to be underrepresented in such studies. For example, a woman living in a rural area with an ectopic pregnancy or one who has postpartum hemorrhage is likely to die before she can reach a hospital. In addition, cause-specific mortality measures may be affected by the types of complications that allow time for a woman to be moved to a hospital. Third, the population delivering in hospitals on a nonemergency basis represents a more socioeconomically advantaged portion of the population and is likely to have had greater-than-average exposure to prenatal care. Finally, since most women in developing countries do not deliver in hospitals but may be brought there on an emergency basis if the delivery is complicated, hospital populations tend to have more abnormal or complicated cases. Thus, hospital data are likely to overesti- mate maternal mortality ratios and provide a misleading picture of maternal morbidity. Even if national maternal mortality ratios cannot be accurately estimated with hospital data, if the assumption can be made that deaths in hospitals are represen- tative of all maternal deaths, or if the way in which they are unrepresentative can be identified, then cause-specific mortality rates from these sources may be useful. Because of the very small numbers of deaths identified in population- based studies, hospital-based studies provide valuable sources of data for examin- ing the causes of high maternal mortality.

OCR for page 25
REPRODUCTIVE PATTERNS AND WOMEN'S HEALTH 29 Case series studies are based on reports about a series of deaths, which could be deaths from all maternal causes or from a specific cause, such as ruptured uterus. These series can provide information about case-fatality rates. Unfortu- nately, most case studies contain no description of the population from which the cases are drawn. Thus, they do not provide evidence on the level of maternal mortality or the influence of reproductive patterns on women's health. Methodological Issues Aside from problems associated with the available data, studies of the associa- tion between reproductive patterns and maternal health are more limited than those available on child health for two methodological reasons. First, few of the available studies use multivariate statistical techniques to control for potentially confounding variables (see Appendix Table 3.A), as do most of the studies of child health on which we draw. Thus, it is more difficult to conclude that specific high-risk characteristics are causally related to maternal mortality. For example, poor women both may have more births (and thus achieve higher parity) and may be less likely to obtain adequate interpartum care. This could result in an upward bias in the estimated effect of high parity on maternal morbidity and mortality. Second, many studies of maternal health fail to hold constant the effects of other reproductive variables. For example, the association of poor pregnancy outcome and births to teenagers may be pardy due to the fact that many of these births are first births. EFFECTS OF YOUNG MATERNAL AGE AND PRIMIPARITY Most studies of pregnancy complications or maternal mortality have investi- gated their association with either young maternal age or first panty, but not with both. Because age and parity are strongly associated, it is often unclear whether the age-specific and parity-specific patterns reflect the same basic risks based on parity, whether they have independent effects, or whether they act in combina- tion. An answer to this question would require more studies of maternal mortality that control for age and parity simultaneously, but few such studies are available (see Koenig et al., 1988a; Walker et al., 1985~. Both population-based and hospital studies indicate that the first pregnancy is strongly associated with a higher risk of maternal mortality. For example, population-based studies from Bangladesh, Ethiopia, and Gambia indicate that the risk of maternal mortality is up to three times higher for the fast birth than for subsequent births. Also, in developed countries and in some developing coun- tries, such as Jamaica, where the possibility of death associated with any birth is much lower than in ADica or South Asia, there is also a higher risk associated with the first birth. There is conflicting evidence about whether pregnancies at maternal ages below 20 are inherently riskier than pregnancies at ages 20 through 24. The

OCR for page 25
30 CONTRACEPTION AD REPRODUCTION largest population-based study, with 14,631 first births (Koenig et al., 1988a), shows no increased risk, though smaller studies from the same area (Chen etOal., 1975) and from Indonesia (Chi et al., 1981) and Jamaica (Walker et al., 1985) indicate a slightly elevated risk of death. A significant problem with these studies, however, is that all births to women under 20 are combined in recent studies. There is evidence that the increase in risks is most important for young women under age 17, particularly those under 15. For example, a hospital study in Nigeria showed that the risk of mortality varied inversely with age, with women 15 or younger having 10 times, l~year-olds 4 times, and 17-19-year-olds twice the risk of those ages 20-24 (Harrison and Rossiter, 1985~. In Tanzania there were 2-3 times more deaths during their first pregnancy or birth among the younger women (Arkutu, 1978~. The possibility that this pattern is influenced by selection bias cannot be ruled out, however, since younger women may be less likely to be brought to hospitals unless they have serious complications Harrison and Rossiter, 1985~. Studies in both developed and developing countries that have considered causes of morbidity or mortality in younger and primigravid women indicate that pregnancy-induced hypertension is most common among women during their first pregnancies and more common among younger women (Arkutu, 1978; Efiong and Banjoko, 1975; Faundes et al., 1974; World Health Organization, 1988~. Obstructed labor because of the pelvis's being too small to allow the child's head to pass is most common in young primigravid women (Aitken and Walls, 1986~. Malaria is more frequent and the infection appears to be heavier during first, and to a certain extent, second pregnancies, than during later preg- nancies (McGregor et al., 1983~. The increased risk for younger primigravidas may reflect not so much in- creased physiological risk as socioeconomic differences between the younger mothers and women who have their first child at ages 20-24. For example, women who have their first child earlier may be from poorer families (Efiong and Banjoko, 1975) and have less access to or make less use of prenatal care than those who have their fast child after age 20 aelley and Madeley, 1983~. EFFECTS OF OLDER MATERNAL AGE, HIGH PARITY, OR BOTH The problem of confounding between age and parity exists as well for births to older women, which are in most cases also high-order births. A pattern of generally increasing risk of maternal death with each successive birth after the second or third birth is evident in the information from three population studies. Women of parity 5 or more have about 1.5 to 3 times the risk of maternal death than women at the lowest-risk parities (2 and 3~. In general, within any parity, older women, particularly those over age 35, tend to be at higher risk of death (Koenig et al., 1988a; Chi et al., 1981; Walker et al., 1985~. Older, multiparous women are more likely to have problems with malpresen- tation, in which the fetus lies in a position other than in the usual head first

OCR for page 25
REPRODUCTIVE PA"E^S AND WOMEN'S HEALTH 31 position (a breech or transverse lie, for example). Mlpresentation may occur because the muscles of the uterine wall become flaccid with repeated stretching of successive pregnancies. The condition can result in uterine rupture, hemor- rhage associated with rupture, or with unsuccessful attempts to remedy the situation (such as the excessive use of oxytocin-containing medicine, abdominal pressure on the uterus, or manipulation of the fetus) and infection. Another cause of hemorrhage is placental abnormality. Faundes et al. (1974) found an abnormal placenta to be more common in women above age 35 and in women who had had 5 or more previous births. Antepartum hemorrhage can arise from several causes, the most important being placenta previa (a condition occurring when the placenta overlies the cervical opening of the uterus, resulting in massive, fatal hemorrhage at the time of attempted delivery unless a cesanan section is performed) and abruptia placenta (a condition occurring when the placenta separates prematurely from the uterus prior to delivery of the baby), which if untreated leads to fetal death, hemorrhage, and blood clotting. Postpar- tum hemorrhage, which is the most common complication among high-parity women, arises primarily from uterine atony (or lack of contraction of uterine muscles), sometimes secondary to a retained placenta (a condition in which the placenta is not expelled after delivery of the baby). EFFECTS OF SHORT INTERVALS BETWEEN BIRTHS Although a number of researchers have hypothesized what is called a mater- nal depletion effect of short interbirth intervals that would increase the risk of maternal mortality (Jelliffe, 1976; Omran and Standley, 1981; Rinehart and Kols, 1984; Winikoff, 1983), no study yet identified has specifically addressed the issue of the relationship between birth interval length and maternal mortality in devel- . . Opmg countnes. There is indirect evidence from Matlab, Bangladesh, suggesting that short intervals are not associated with a higher risk of maternal mortality: for each 5- year age group of women, the risk of mortality decreased with increasing parity, at least through parity 6 (Koenig et al., 1988a). Among women of the same age, those of higher parity will, on average, have had shorter birth intervals. Thus, this study seems to suggest that women with the shortest birth intervals are least likely to die. The reason, however, could be due to selection: the healthiest women may become pregnant more quickly and therefore achieve higher panties. EFFECTS OF PREGNANCY IN INCREASING MORTALITY FROM OTHER CONDITIONS Pregnancy increases the likelihood that a woman will die of certain conditions (i.e., case-fatality rates are increased). These conditions include chronic illnesses that antedate the pregnancy, such as rheumatic heart disease, diabetes, sickle-cell disease and AIDS, as well as acute infectious diseases that the woman contracts

OCR for page 25
32 CONTRA CEPTiON ID REPRODUCTION while pregnant. Examples of these acutely infectious diseases are hepatitis and epidemic malaria, for which case-fatality rates are higher for pregnant than for nonpregnant women (Morrow et al., 1968~. More generally, women who are malnourished or in poor health may be more likely to experience problems during a pregnancy. This is particularly true for women who are severely anemic. EFFECTS OF UNSAFE INDUCED ABORTION - In countries where safe abortion is not available, many women suffer serious complications and often death as a consequence of unsafe abortion procedures. Unsafe abortion can cause uterine perforation, hemorrhage, uterine and general- ized infection, acute bleeding disorders, and embolism to the lung and brain. Kwast et al. (1986) report that postabortion complications are the most com- mon cause of maternal mortality in Addis Ababa, Ethiopia, especially among primigravid, unmarried women employed as domestics and students. Koenig et al. (1988a) attribute 18 percent of maternal mortality in Matlab Thana, Bangla- desh, to postabortion complications. Additional insights into the health consequences of unsafe abortion can be derived from the experience in Romania, where restrictive abortion laws were enacted after a history of relatively liberal laws, resulting in a sevenfold increase in maternal mortality due to abortion (Tietze, 1983~. At the same time, data from the United States demonstrates the extremely low risk of safe, first-trimester abortion procedures, making them one of the safest surgical procedures per- formed. CONCLUSION A reduction in the number of pregnancies and births that a woman experiences enables her to unambiguously reduce her risk of reproductive complications and maternal mortality merely by reducing the number of times she is exposed to that risk. While there are potential risks to using contraception, the research reviewed in the next chapter demonstrates that the risks a woman assumes in using contra- ception are very small compared with the health benefits of reducing exposure to pregnancy and birth-related health problems. Moreover, some contraceptives have noncontraceptive health benefits (see Chapter 4~. Data from Bangladesh show declines in the maternal mortality rate (per 100,000 women) associated with reduction in fertility (Koenig et al., 1988a). Other evidence presented in this chapter indicates that certain changes in repro- ductive patterns may also be beneficial to women's health, over and above the effect of reducing the absolute number of births women have. The basic pattern observed for the relationship between fertility and maternal mortality is that risk of mortality is highest for first pregnancies and for fifth and subsequent pregnan- cies. This pattern exists whatever the overall level of maternal mortality, but, as

OCR for page 25
REPRODUCTIVE PATTERNS AND WOMEN'S HEALTH 33 conditions improve, the U-shaped curve of risk with parity is not only lower, but also flatter. Extremes of age, lack of medical care, poverty, and some infectious diseases also increase the risk. However, no studies have shown a reduction in the maternal mortality ratio (per 100,000 live births) as a result of changes in age or parity distributions of births. For first births, particularly among young women, use of contraception can reduce maternal mortality by delaying first births until after age 20 or by reducing the number of unwanted pregnancies that might otherwise result in abortion. However, teenagers may not use contraception effectively, resulting in excessive use of abortion as a means of preventing unwanted births. This is not inevitable, however, as shown by the experience in Canada and Scandinavian countries, where more effective sex education efforts and easier access to contraception resulted in declines of both pregnancies and abortions among young women (Tietze, 1983; Henshaw, 1986~. Clearly, where safe abortions are not available, effective family planning use by teenagers is even more important as a means of reducing mortality associated with unwanted first births. Evidence presented in this chapter suggests that changing reproductive pat- terns can reduce the maternal mortality rate (per 100,000 women) by: reducing the total number of pregnancies each woman has in lifetime; reducing He incidence of high-risk pregnancies (high parity, very young maternal age and older maternal age, pregnancy among women with major health problems (e.g., hypertension, diabetes, heart disease, and malaria); and reducing the demand for abortion to terminate unwanted pregnancies in countries where safe abortion is unavailable. APPENDIX TABLE 3.A Summary of Studies of Maternal Health Study Measures of Location Matemal Condition Tabulation Vanables POPULATION-BASED STUDES Alaudd~n, 1987 Bangladesh Matemal modality ratios Bhatia, 1985 India Matemal deaths Chen, Gesche, Bangladesh Ahmed, Chowdhury and Moseley, 1975 Matemal mortality ratios Age, parity, landholding, economic status, education, . ,. gravity Age, rural/urban, panty Age, parity, living children, . . grave sty

OCR for page 25
34 CONTRA CEPTiON ID PRODUCTION APPENDIX TABLE 3.A Continued Measures of Study Location Matemal Condition Tabulation Variables Fortney, Susanti, Egypt, Total deaths, Cause of death, age Gadalla, Saleh, Indonesia maternal deaths, ~~ Feldblum, and Potts, maternal mortality, 1985 rater and ratio' Greenwood, Greenwood, Rural Gambia Matemal mortality Bradley, Williams, Shenton, Tulloch, Byass and OldD~eld, 1987 ration Age, parity, prenatal visits, attendant, cause of death, place of delivery, place of death, time of death relative to delivery, pregnancy outcome Khan, Jahan and Begum, Bangladesh Maternal modality Age, panty, place, cause of 1986 ratio' death, attendant Koenig, Fauveau, Bangladesh Chowdhury, Chakraborty, and Khan, 1988 Kwast, Rochat, Ethiopia Kidane-Mariam, 1986 Lindpainter, Jahan, Bangladesh Satterthwaite, and Zimicki, 1982 Matemal mortality rater and ratio, matemal deaths, total deaths Maternal deaths, maternal mortality ratios Matemal deaths, maternal modality ratios Age, panty, gravidity, year Age, parity, education, marital status, income, prenatal care, occupation, wontedness of pregnancy Age, gravidity, cause of death, area Walker, Ashley, Jamaica Matemal deaths, Cause of death, time relative to McCaw and Bemard, maternal mortality delivery, age, panty 1985 ratios Aitken and Walls, Sierra Leone 1986 HOSPll~AL AND CL~IC-BASED STUDIES Cephalopelvic Maternal height dispositions and . . .. . . pnrmgravlaas She of the fetal head in relation to the maternal pelvis of women pregnant for the first time)

OCR for page 25
REPRODUCTIVE PATTERNS AND WOMEN'S HEALTH 35 APPENDIX TABLE 3.A Continued Measures of Study Location Matemal Condition Tabulaion Variables Arkutu, 1978 - Tanzania Pnnugravidas Complications, mode of (women pregnant for delivery, duration of labor, the first time) age, birthwight Chi, Agoestina, Indonesia Matemal deaths Cause of death, underlying and Harbin, 1981 maternal mortality candidon, age, parity, urban/ radon rural admission, anemic/not, hospital Efiong and Nigeria Comparison of very Income class,height,prenatal Banjoko, 1975 young and old canplicaia~, duration of primigravid women labor, mode of delivery, blood loss, birthweight Faundes, Fanjul, Chile Deliveries, Hypertension, ampresentatic~n, Henriquez, Mora, neonatal mortality placental hemorrhage, and Tognola, 1974 congenital malfommaia~ - parity, postpartum hemorrhage Harrison and Nigeria Maternal deaths, duration in hospital, ethnicity, Rossiter, 1985 pregnancies, residence, religion, education, maternal mortality parity, age ratios Jelley end Madely, Mozembigue Infommation from Age, health center, 1983 prenatal clinic forms category, attendants McGregor, Wilson, Gambia Deliveries Placentalparasitemia, area, and Billewicz, 1983 urban/rural, parity, sex, Mhango,Rochat, and Zambia Births, Age, panty, cause-specific, Ar~cutu, 1986 matemal deaths, type of prenatal care; maternal mortality Case reports: cause of death, parity, marital status, socioeconomic status, time relative to delivery Morrow, Ghana Hepatitis patients, Coma status,pregnancy status, Smetana, Sai including pregnant sex, age and Edscanb, 1968 and postpartum females 1 Deaths due to pregnancy or childbirth per 100,000 live births. 2 Deaths due to pregnancy or childbirth per 100,000 women of reproductive ages.