Social Dynamics of Adolescent Fertility in Sub-Saharan Africa (1993)

It should not be surprising that adolescents in sub-Saharan Africa, who have the highest rates of fertility for their age in the world, face probably the highest risks of pregnancy-related mortality, of delivery complications, and of premature births or low-birthweight babies. Teenage pregnancy in Africa also has important social and economic outcomes, the most highly publicized of which stem from lost educational opportunities when pregnancy forces young women to leave school.¹ Ideally, an investigation of the consequences of adolescent childbearing and sexuality should cover a wide range of outcomes that affect not only the young mother and her child, but also other family members and society at large. Because there has been

comparatively little systematic research on the subject, we focus on what is best documented: the health consequences for mother and child. Even here, our knowledge is extremely thin on, for example, the incidence of induced abortions in Africa. After treating several other types of consequences more cursorily, we draw some inferences about the consequences of adolescent childbearing under unsanctioned conditions.

Most reproductive health problems experienced by adolescents are also experienced by older women. But they are exacerbated among the young, whether by physical immaturity, primiparity, or social condemnation. Chapter 1 alludes to the two-part typology set forth by Zabin and Kiragu (1992), of childbearing among very young married women and childbearing among young unmarried women who still may be engaged in training. This typology sets the stage for discussing the consequences of adolescent fertility in terms of health and welfare for mothers and children. Because most information that is available concerns the second type, young unmarried women, we focus on that group.

To recapitulate, in societies in which marriage and childbearing are expected to begin early, the majority of pregnancies produce highly valued children. Problems, when they do arise, arise primarily from the mother's physiological immaturity: Babies suffer from low birthweights and birth traumas, and young women are not mature enough to safely carry a fetus to term or to bear a baby. These problems are often compounded by the lack of adequate medical care. Although this pattern of early marriage and childbearing has received considerably less attention than has the emerging problem of pregnancy among urban schoolgirls, the number of women potentially at risk for these problems argues convincingly for including this pattern in any discussion of the consequences of adolescent fertility.

The second high-risk group, young women who become pregnant while still in school or training, or with no supportive men in sight, has emerged more recently; its members have experienced disruptions of standard sequences of education/training, marriage, and childbearing. Most notable examples are unmarried urban schoolgirls who become pregnant. Because this problem attracts so much attention both in the research community and in the popular press, it tends to be identified as the major locus of troublesome adolescent fertility in Africa. Yet both the category of urban schoolgirls and its associated fertility risks are more complex than we might assume. Inasmuch as urban schoolgirls are very young and physically immature they face risks of childbearing similar to those of women in the first group. However, because their risks extend through their education or training, women in this group are probably slightly older, on average, than their counterparts in the early-marriage group. Yet in the case of urban schoolgirls, social condemnation makes them reluctant to seek health services; hence, they suffer additional risks from lack of prenatal care and from

unsafe abortions. They may also suffer an increase in risk of sexually transmitted diseases, although the evidence on this point is thin.

Assessing the consequences of adolescent fertility is fraught with analytical difficulties. How, for example, do we separate the true effects of maternal age from other ''compositional" effects? Recent studies (for example, Senderowitz and Paxman, 1985; Strobino, 1987) have stressed that not all pregnancy-related complications experienced by adolescents are attributable directly to age or immaturity. That is, age itself may cause fewer health risks for either the mother or her offspring than the perceived legitimacy of the pregnancy, low socioeconomic status, poor nutrition, inadequate prenatal care, or primiparity. First births, for example, are typically more complicated than higher-order ones are. Teenagers have higher rates of maternal death in part because they have predominantly first births. Even problems such as cephalopelvic disproportion, in which a woman's pelvis is too small to permit a child's head to pass (a condition most frequently observed in young women), can also stem from poor nutrition and stunting in childhood.

Our second difficulty entails distinguishing what seem to be clear-cut causes from consequences. We have already explained why it is unclear whether fertility among adolescents causes or stems from other factors. Even less easily solved is the difficulty in assessing whether two events are related because people responded in a certain way to an exogenous event, or whether individuals, acting in anticipation, actually brought about what appears to be the prior event. To avoid these chicken-and-egg conundrums, we use the term "consequence" in this chapter simply to refer to events that appear to follow temporally the other events.

Another general caution is that the physiological, educational, or economic consequences of adolescent fertility do not occur in a social vacuum. Social contexts shape the consequences of physiological and demographic events in subtle yet profound ways. One much-discussed example of how biology and society interact is the "biosocial gap" between menarche and socially sanctioned childbearing. The wider this gap, the greater the likelihood of conceiving an unsanctioned child.

More generally, society shapes key values that dictate when young women begin bearing children and how their giving birth is regarded by their families as well as by health and welfare services. For very young married women living in rural areas, where society may define early childbearing as normal and even desirable, the social and economic risks of not bearing children probably outweigh the physical risks of bearing children. By contrast with remote rural areas, urban areas have better health care facilities, not to mention long-term educational and training opportunities—factors that should, in theory, improve health outcomes for young women and their children. Governments have made marked efforts to create and staff low-

cost clinics for mothers and young children in the last 15 years. Yet although many of the risks of adolescent pregnancy can be offset by affordable prenatal care and the use of contraceptives, adolescents often avoid public medical facilities where their presence would expose the fact that they were engaging in illicit sexual activities. We can appreciate their problems by describing a typical small clinic and what transpires within it.

Most government-sponsored clinics in Africa have what strikes outsiders as an appalling lack of facilities. Many clinics stock only a few medicines; drugs are perpetually ''on order." Equipment is minimal. A rural clinic may have an examining bed with a thin covering blanket, a table and chair, a scale, a few syringes, a large roll of cotton bandaging, and some suture material. In the waiting area are a few benches rubbed shiny from years of sitters who slide down to keep their place in line. Yet despite the poverty of their physical resources, most clinics try to maintain cleanliness. Syringes are disinfected; walls with ancient flakes of paint are scrubbed regularly; even dirt floors are swept.

The biggest surprise to outsiders is that despite their lack of facilities, most clinics that keep regular hours are patronized by long rows of patient women who sit or stand in line, sometimes for hours at a time. During days when under-fives are seen, small fussy children—some coughing and feverish, some well and feisty—sit or squirm with their mothers, who clutch tattered yet carefully retained "Road to Health" cards, on which each child's new weight and immunizations will be recorded. On prenatal clinic days, long lines of pregnant women appear, many with small toddlers. As each woman's turn comes, she swings her child onto her hip and enters the consulting area.

Yet the popularity of clinics is also a drawback to adolescents. The consulting area offers little privacy, especially, as is often the case, if it is separated from the waiting area only by a thin cloth. The crowded waiting room is constantly buzzing with gossip: whose child has been sent to secondary school, who was seen alone with whom, who may be pregnant, who is leaving her husband, and so on. The sheer lack of privacy surfaces in other ways as well. Clients know that records are written down and kept and that clinic staff are frequently their neighbors or relatives.

How, then, does a pregnant adolescent with no sanctioned attachment to a man fare in such a setting? Many pairs of eyes shrewdly appraise her condition and exchange contemptuous looks. Some women actually make comments, and all ears are trained to the consulting area as she enters. There, she is likely to be chastised by the attending staff for her condition, and her morals are likely to be loudly questioned. Describing a study of a contraceptives clinic in Ghana, Huntington et al. (1990) provide confirmatory observations:

Unlike older clients, who reported attentive professional services, unmarried adolescents reported open hostility (Huntington et al., 1990). "[N]urses continued to give what were perceived by the younger clients to be lectures on morals, and then summarily dismissed them." The following quotes were collected from younger clients (p. 175):

And if gossip among patients is not intimidating enough, nurses themselves gossip about who recently appeared at the clinic and what they came for. Mojisola Olaneyan's short story "It's Wonderful Being a Girl" (1992, manuscript) depicts such a conversation between two nurses:

Not surprisingly, such social pressures make young women extremely reluctant to utilize medical facilities of which they may have considerable need. But although they are reluctant to come to public clinics to obtain contraceptives, it is clear from numerous surveys, including the Demographic and Health Survey (DHS), that unmarried adolescents manage to obtain

contraceptives. Evidence from The Gambia shows that most unmarried young women patronize sources such as pharmacies or shops to obtain their supplies. (In some countries, market traders undoubtedly stock supplies of certain contraceptives.) And young unmarried women who use condoms rely largely on their male partners to obtain them (Bledsoe et al., 1993). Even coming to a prenatal clinic is itself a shameful act—almost an admission of guilt. Not only do few pregnant adolescents attend prenatal clinics; a neighborhood clinic is a place of last resort to an unmarried adolescent who wants contraceptives or to have her baby checked and immunized. (Grandmothers are usually dispatched for this purpose.) Fears of stigma or of losing career opportunities make many young women risk abortion or forgo medical services altogether rather than risk public ridicule. These factors, coupled with their relative immaturity and their first birth experience, put adolescents at high risk for pregnancy problems.

Considerable research has shown the debilitating, sometimes fatal, results of childbearing by young African women. The following sections document the health complications that adolescent fertility can entail for the mother and for her children.

Pregnancy-related complications are alleged to cause up to half of the deaths among women of reproductive age in developing countries (Lettenmaier et al., 1988). In many areas, for every woman who dies, between 10 and 15 may suffer long-term damage to health by pregnancy or labor that can cause considerable distress and preclude a normal life (Starrs, 1987; Lettenmaier et al., 1988; Tahzib, 1989). Making such estimates for any developing country is extremely difficult. In Africa, most countries do not have systematic vital registration systems. Hospital-based samples do not represent the complete population at risk because many women without access to medical facilities and those trying to evade attention die elsewhere. Sample surveys in the general population have trouble detecting rare events; and fertility surveys, though they target women, do not record cases of maternal deaths because women who have died cannot be respondents. (Recently created indirect techniques such as the "sisterhood" method may improve estimates; see Graham et al., 1989; Trussell and Rodriguez, 1990).

Despite the lack of precise data, there is little doubt that high birth rates in the absence of adequate health care produce high maternal mortality rates. In parts of Africa, difficult living conditions and inadequate nutrition

and health care, along with high fertility rates, subject women of all ages to very high risks of pregnancy-related illness and death. In 1983, an estimated 126,000 women in sub-Saharan Africa died in pregnancy or childbirth (Royston and Lopez, 1987). The maternal mortality rate (the number of women who die during pregnancy or childbearing for each 100,000 live births) is estimated to be 655 in sub-Saharan Africa, one-and-a-half times greater than the rate in Asia and over thirty times greater than that in northern Europe. Because fertility is generally higher in Africa than elsewhere and because multiple pregnancies mean multiple maternal risks, the average woman, with more than six children, will run a lifetime risk of 1 in 21 of dying as a result of pregnancy. The comparable figure for women in northern Europe is 1 in 9,850 (Graham, 1991).

Although African women as a group run greater risks of maternal mortality and pregnancy-related complications than do women in the rest of the world, their risks are clustered disproportionately in the early years of reproduction, (Senderowitz and Paxman, 1985; United Nations, 1989) or at first deliveries, which would look very much the same in statistics. In Yaoundé, Cameroon, adolescents account for only 28 percent of the obstetric population, yet they account for over 70 percent of obstetric complications (Leke, 1989). In Sierra Leone, 15-to 24-year-old women account for 38 percent of pregnancy-related complications, most of which afflict the 15–19 age group (Senderowitz and Paxman, 1985). In Addis Ababa, Ethiopia, teenagers are twice as likely to die from pregnancy-related conditions as women aged 20–24 (Swedish Save the Children Federation, 1984, cited in United Nations, 1989).

Because a mature physique is important to successful childbearing, we expect to find that young adolescents experience considerably more problems in childbearing than older ones do. Direct evidence for Africa is hard to compile, but U.S. medical findings indirectly support this contention. For example, Moerman (1982) used a longitudinal sample of 90 well-nourished girls in Ohio to measure the growth and development of the birth canal among girls aged 8 through 18. He found that the pelvis was smaller and less mature among girls with early menarche than among girls with late menarche at the same length of time after menarche (p. 528). Moerman concluded that the growth of the pelvic birth canal was associated less with gynecological development than with chronologic age (p. 532) and claimed that the immaturity of the birth canal "may have significance for obstetric risks among young teenage primiparous girls" (p. 528).

Studies of mortality among African women provide some further, often indirect, evidence. Starrs (1987) found that adolescents in Africa under the age of 15 are five to seven times more likely to die in pregnancy and childbirth than women aged 20–24. In a study of over 22,000 births in Zaria, Nigeria, Harrison et al. (1985) found that maternal mortality was

between two and three times higher for women 15 and under than for women from 16 to 29.

In developing countries in general, complications during delivery directly cause about three-quarters of all maternal deaths. The remaining quarter result from medical conditions that were aggravated by pregnancy, such as viral hepatitis, anemia, and cardiovascular disease (Herz and Measham, 1987). Several medical complications account for the majority of direct obstetric deaths and injuries: eclampsia, sepsis, obstructed labor, hemorrhage, fistulas, and unsafe abortion. In Africa, many of these complications are most severe among very young women. Two factors appear to be principally responsible: physical immaturity and the tendency, especially among young women whose pregnancies are socially disapproved, to avoid or delay treatment at an appropriate medical facility.

Eclampsia, or toxemia, may begin in the second or third trimester with high blood pressure, fluid retention, and protein in urine. Left untreated, full-blown eclampsia can result in extreme hypertension, convulsions, or cerebral hemorrhage. Eclampsia is largely preventable with appropriate prenatal care. Once it develops, however, it requires immediate treatment. Yet even when medical care is available, complete recovery is far from certain. In developed and developing countries 5 to 17 percent of women with eclampsia die; survivors may suffer congestive heart failure, paralysis, blindness, chronic hypertension, or kidney damage (Lettenmaier et al., 1988). Eclampsia is particularly common among young women and women having their first child. A study of women pregnant with a first child in Zaria, Nigeria, reports that 17 percent of those aged 14 or under and 10 percent of those aged 15 to 16 developed eclampsia, compared with only 3 percent of those aged 20 to 29 (Harrison et al., 1985).

Sepsis (severe infection) following childbirth or attempted abortion is common in developing countries. The probability of infection increases with prolonged labor. Infections can be compounded by endemic diseases, such as malaria and tuberculosis, and by existing conditions, such as anemia, that are most common among young mothers (Harrison et al., 1985). There is some evidence to suggest that women who suffer severe infections later face higher risks of pelvic inflammatory disease, ectopic pregnancy, infertility, and chronic pelvic pain (Lettenmaier et al., 1988). Women with uncomplicated vaginal deliveries usually do not become infected unless they have sexually transmitted diseases; but very young mothers, because they have more complicated deliveries, suffer greater risks of sepsis.

Obstructed and prolonged labor requires treatment at medical facilities that can perform surgery. Young adolescents are particularly vulnerable to obstructed labor because of their immature physiques. One of the most serious of such complications is cephalopelvic disproportion. In a study in Kasongo, Zaire, teenage women accounted for 34 percent of all births, but they accounted for 42 percent of all women with life-threatening fetopelvic disproportion and for 48 percent of women with abnormally long labor (van Lerberge et al., 1984, cited in United Nations, 1989).

Another cause of obstruction is the scar tissue that forms after a woman has been circumcised (International Center on Adolescent Fertility, 1992)—whether by simple clitoridectomy or infibulation, which involves the complete removal of the clitoris, labia minora, and labia majora, and stitching the two sides together. The scar tissue may prolong labor and create serious complications during delivery and possible brain damage to the child (Minority Rights Group, 1980, cited in Kouba and Muasher, 1985). Adolescents are especially susceptible to this condition since most first births occur among this age group. Research on the health aspects of female circumcision is hindered by the culturally sensitive nature of the subject and the secrecy that surrounds the practice. It is unclear, for example, whether the obstetric risks are not simply a function of scar tissue impediments but also of the age at which the operation is performed. This age varies from as early as the sixth day of life in Yoruba tradition to as late as shortly before the birth of the first child among the Aboh in midwestern Nigeria (Kouba and Muasher, 1985).

Hemorrhage is one of the most common causes of maternal death. A sudden hemorrhage requires immediate treatment at a health facility capable of providing blood transfusions and performing other clinical measures—treatment that is not readily available to most African women. Hemorrhaging is most common just after delivery. It can be caused by prolonged labor, early separation of the placenta from the uterine wall, a rupture or tear in the cervix or the vagina, or a poorly performed abortion; many of these are conditions to which young mothers are most susceptible (Lettenmaier et al., 1988). In some African countries, patients fortunate enough to receive a blood transfusion face the added risk of contracting HIV (human immunodeficiency virus) or hepatitis B through unscreened blood.

Fistulas, one of the most devastating obstetric injuries, are perforations between the vagina and the rectum or urethra that allow urine or feces to

leak continuously through the vagina. Women with fistulas suffer from incontinence and persistent odors caused by stale urine or excreta. Victims are frequently ostracized by both their husbands and the community. Surgical repair is possible but rarely available. Most fistulas are caused by prolonged and obstructed labor, particularly labor complicated by cephalopelvic disproportion. Very young mothers are most susceptible. One study in Kenya found that 45 percent of all victims of vesicovaginal fistulas were adolescents (World Bank, 1989, cited in Zabin and Kiragu, 1992). In northern Nigeria, of 1,443 women admitted to Ahmadu Bello Hospital requesting reparative surgery, 33 percent were 16 and under and another 22 percent were aged 17 to 19 (Tahzib, 1983).

Abortions pose very grave health risks to adolescents in Africa. Abortion-related deaths are a major component of maternal mortality. A study at Lagos University Teaching Hospital in Nigeria found that 51 percent of maternal deaths were attributable to abortion complications (Akingba, 1971). In studies conducted in Addis Ababa, Ethiopia, and Lusaka, Zambia, abortion was identified as the cause of a quarter of all maternal deaths (Lettenmaier et al., 1988).

While abortion rates are believed to be high and rising throughout the general population (Coeytaux, 1988), there are shocking estimates of abortion rates among young unmarried women. Precise levels, of course, remain unknown. The vast bulk of research comes from hospital-based samples that reveal nothing about women who do not come in contact with medical services. Of women under age 20 in Accra who went to the hospital to deliver a second pregnancy, about a quarter had had their first pregnancy terminated by an illegal induced abortion (Janowitz et al., 1984). At the University of Calabar Teaching Hospital, 72 percent of women who presented for induced abortion were adolescents aged 13–19, the majority unmarried (Archibong, 1991). Similar observations have been made in Ghana, Kenya, Tanzania, Mali, Zaire, Liberia and Benin (see, respectively, Ampofo, 1970; Aggarwal and Mati, 1980; McKay, 1984; Burton, 1985; Nichols et al., 1987; Avodagbe, 1988). It is estimated that in Cameroon, 18 percent of all adolescent pregnancies lead to abortion (Leke, 1990; République du Cameroun, no date). In Nigeria, abortion is the preferred course of action in 90 percent of pregnancies among unmarried working women (Odejide, 1986). Another Nigerian study reports that of 1,800 never-married females ages 14–25, nearly one-half of the students and two-thirds of the others had been pregnant (Nichols et al., 1986). Nearly all bad terminated their pregnancies with an induced abortion.

In theory, abortion should be quite rare in sub-Saharan countries because it is either illegal or quite restricted (see Tables 7-1 and 7-2). Of 38

countries surveyed in 1989 by Ross et al. (1992), only 9 legally permitted abortion for reasons other than health or eugenic threat; of these 9, only Togo allows abortion on request and only Cameroon does not require the approval of the husband or a committee. Even in Zambia, where abortion is legal, a woman must obtain written consent from three physicians (Castle et al., 1990). South Africa has a similar law, the Abortion and Sterilization Act, and according to one review (Nash, 1990), only one-third of the requests for an abortion for mental reasons were approved.

One consequence of such formidable laws is to impede efforts to obtain reliable information about abortions, many of which, because of the laws, are illegal. Despite the understandable reluctance on the part of women to discuss abortions, many studies find that abortions stem from pregnancies for which paternal recognition is a problem (see, for example, Nichols et al., 1986; Nichols et al., 1987. Caldwell and Caldwell, 1988; Coeytaux, 1988; Kulin, 1988; Baker and Khasiani, 1992). In many countries, the majority of women who seek abortions are single adolescents (see Caldwell and Caldwell, 1988). Many are unemployed or underemployed (see, for example, Lema and Kabeberi-Macharia, 1990), or are desperate to stay in school. Oppong (1991) stresses that young women's opportunities for schooling, training, and employment are jeopardized by pregnancy, but

Another result of banning abortion is to make those that do occur more dangerous because illegal abortions have much higher rates of complication than legal procedures do (Zabin and Kiragu, 1992). Adolescents may seek out trained health care workers who, fearing legal repercussions, may try to create dangerous medical emergencies that warrant a legal abortion. In Zambia, at least one-quarter of all the women with abortion complications in the University Teaching Hospital in Lusaka, Zambia, had tried to induce abortions by inserting objects in the cervix such as plants or twigs, drinking detergents or gasoline, or taking large overdoses of aspirin, chloroquine, and other toxic substances (Castle et al., 1990). Private physicians sometimes induce abortion illegally by inserting a device and instructing the woman to go to the hospital to complete the procedure. The consequences can include hemorrhage, perforation, infection, and visceral and cervical injury. Left untreated, many of these complications can cause sterility or death (Zabin and Kiragu, 1992). Nearly 15,000 operations were performed in South Africa in 1987 to follow up incomplete abortion (Nash, 1990).

Although properly administered abortions are no more dangerous to adolescents than to other women, abortion is particularly lethal among young

women for three principal reasons. First, young women face severe problems of confidentiality and legal roadblocks that make them more likely to resort to illegal abortions. Using responses from 264 women requesting termination of pregnancy or treatment of abortion complications at the University Teaching Hospital in Zambia in 1985 and 1986, Likwa and Whittaker (no date) found that 60 percent were aged 15–19.

Second, few adolescents have ready cash to pay for abortions by licensed practitioners; most resort to cheaper but more dangerous methods. They may go to semiskilled practitioners or attempt self-induced abortion.

The third reason why abortions are dangerous grows partly out of the first two. Adolescents tend to postpone abortion attempts longer than older women do (Center for Population Options, 1992)—beyond the first 12 weeks of pregnancy, the period when abortion is safest (for example, Oyieke, 1986, cited in Lema and Kabeberi-Macharia, 1990). Their reasons for delaying are unclear, but they may include denial of the condition, hope that spontaneous miscarriage will occur, fear of sanction, or the expense of obtaining an abortion. Young women also tend to delay obtaining help if complications ensue after an abortion, for similar reasons. For example, the average cost of treating an illegal septic abortion in Ibadan, Nigeria, has recently been estimated to be U.S. $200, in an area where the typical monthly salary is U.S. $45 (Konje et al., 1992).

Laws that restrict access to abortion may, therefore, simply drive the practice underground: Young women attempt self-induced abortions or seek out untrained health care providers, who may work in unsanitary conditions. Caught between losing a prosperous economic future and risking their lives, schoolgirls face limited options. Gyepi-Garbrah (1985b:45) sums up their problem:

Some evidence, though sketchy, points to critical social differences between young women for whom abortion is not risky and those for whom it poses great risk. Women who encounter the most danger are more likely to be unmarried, less well-off, and perhaps less educated than others; these women are most likely to seek abortions that are cheap and therefore illegal. Most of the evidence on this point is oblique. Joseph (1979) suggests that in countries whose laws on abortion are not liberal, one needs money to obtain a safe abortion by a qualified practitioner. Lema and Kabeberi-Macharia (1990) imply that in Kenya ''highly paid women who are . . .

professionals'' are able to come up with the funds quickly to pay for abortions that are carried out both early and by private, experienced medical personnel (1990:119, 127). These key advantages sharply reduce the risks of well-to-do women. Moreover, these abortions "... may not be on any record and therefore we may never get to know the actual magnitude of abortion in Kenya" (p. 119). It is likely as well that some wealthier girls are able to finish their education because they are able to afford safe abortions.

As this discussion has revealed, the costs and risks of abortion often stem less from abortion itself than from the social and economic conditions that make safe abortion impossible. Such costs reach far beyond the individual women involved. Treating poorly performed abortions places enormous pressure on the already stretched health systems throughout the region. Coeytaux (1990) speculates as well that making abortion difficult to obtain may lead women to resist family planning in some instances. A young woman who has had an abortion is quite likely to begin to use contraception; but if she has unknowingly been made sterile by a poorly performed abortion and later tries unsuccessfully to conceive, she may attribute her infertility not to the abortion but to the contraception. Hearing of her difficulty, others may abandon contraception as well.

Sexually transmitted diseases (STDs), though less directly connected with childbirth, constitute yet another medical risk that is closely associated with sexual activity among adolescents. Left untreated, STDs can lead to pelvic inflammatory disease, miscarriage, infertility, or ectopic pregnancy. A high incidence of STDs underlies the historically high levels of infertility in many parts of Central Africa (Frank, 1983).

If sexual contacts outside the marriage process are becoming more prevalent, then larger numbers of adolescents may be exposed to the risk of sexually transmitted diseases, including AIDS. In addition, the period in which adolescents are exposed to these risks appears to be lengthening. If the age at menarche is indeed declining, as happened in the West, the age at which girls begin to encounter sexual and reproductive health problems could be lower. At the other end of adolescence, age at marriage may be rising. This widening of the biosocial gap between onset of sexual maturity and marriage could increase the risk of health problems.

Some evidence suggests that adolescents are biologically more susceptible than older women to STDs. One reason for higher susceptibility is a greater exposure of cervical epithelial tissue at the opening of the vagina into the cervix in the adolescent years (Zabin and Kiragu, 1992). A second reason is that a young man or woman may have a previously unchallenged

immune system that therefore has not mobilized defenses against infection (Zabin and Kiragu, 1992). Yet how to weigh physiological factors against the complications posed by behavioral and social factors is unclear. Certainly adolescents who begin sexual relations at an early age and have multiple partners expose themselves to great risks. A recent survey reported that in a rural district in Uganda, 50 percent of women aged 13–19 living in main-road trading centers were seropositive for HIV (Wawer et al., 1991). In Monrovia, Liberia, a 1984 survey found that 23 percent of female students and 29 percent of male students aged 18–21 reported having had a sexually transmitted disease (Nichols et al., 1987), and in a Cameroon study of sexually active high-school students, one-quarter of the females and one-third of the males reported having had an STD (Mafany, 1989, cited in Zabin and Kiragu, 1992). In a 1989 survey of pregnant women in Dar es Salaam, Tanzania, women aged 13–19 were one and a half times more likely to be infected than women 30–34 (U.S. Bureau of the Census, 1992). In some areas, women contract HIV at younger ages than men, often during adolescence, because of the age gap between sexual partners. For example, a nationwide survey in Uganda during 1987–1988 found that women aged 15–19 were more than twice as likely to be infected than men the same age, and women aged 20–24 were one and a half times more likely to be infected than men the same age (U.S. Bureau of the Census, 1992).

A more far-reaching implication of the spread of AIDS for adolescents, though one less easily documented, is that heightened concern and awareness about AIDS may be exposing adolescents to greater risk of contracting the disease. Older urban males, who themselves may be HIV positive, may seek wives and sexual partners among very young girls, because they will have had fewer sexual contacts and thus are less likely to be infected (Zabin and Kiruga, 1992). This same motive may also exert downward pressure on the age of marriage for girls (Panos Institute, 1989).

Many adolescents lack basic information concerning the symptoms, transmission, and treatment of STDs. Owie (1985) reports that over one-third of adolescents in his study in Nigeria believe that treatment for an STD is unnecessary once the symptoms have disappeared, and that gonorrhea and syphilis are interchangeable names for the same disease. Drawing on a sample survey, Wilson et al. (1988) reported that 40 percent of high school students in Zimbabwe believed that AIDS in Africa is confined primarily to homosexuals. A similar proportion was unaware that HIV carriers may appear healthy. Ignorance surrounding symptoms or treatment decreases the likelihood that adolescents will seek timely treatment, and thus increases the risk of contagion and transmission.

The new DHS data leave few doubts that children born to very young mothers suffer disproportionate risks of morbidity and mortality, though whether the primary factor is the age of the mother, the high proportion of first births, or some other characteristic, is not clear. Table 7-3 shows infant mortality rates by age of mother. In 10 of the 11 countries for which DHS data are now available, infants of young mothers (aged less than 20) experience higher mortality than do children of mothers aged 20–29. In seven countries the differential exceeds 20 percent; in four it exceeds 50 percent. Child Mortality (deaths at ages 1 to 4), though lower, continues the pattern of increased risk among children of adolescent mothers. The differences in risk ranges from under 10 percent in Senegal and Togo to 25 percent and over in Botswana and Uganda.

Other studies confirm that infants born to adolescent mothers are more likely to be born prematurely and to have low birthweight (Arkutu, 1978; Ngoka and Mati, 1980, Sanghvi et al., 1983; Adedoyin and Adetoro, 1989). Low birthweight can lead to life-long neurological problems and slower development (Senderowitz and Paxman, 1985), and it is the single best predictor of neonatal mortality. Approximately two-thirds of neonatal deaths occur among infants weighing less than 2,500 grams at birth (United Nations, 1989).

Although the DHS data indicate mortality differences for children of adolescent mothers, the age of the mother may be a smaller causal factor than are primiparity, inadequate prenatal care, low socioeconomic status, and the perceived legitimacy of the birth (Haaga, 1989; National Research Council, 1989; Zabin and Kiruga, 1992).

Multivariate techniques can provide some insight into the importance of maternal age. Using data collected during the late 1970s from 34 developing countries, Hobcraft et al. (1985) examined the risk of child mortality by controlling for the potentially confounding effects of mother's education and her age at birth and for the child's sex, birth order, and birth interval. They also controlled for survival of the previous and the subsequent birth. (Whether the mother obtained prenatal care, however, was not controlled, nor was her marital status.) Even after introducing these controls, the study found that the average relative risk of mortality for children in the first month of life of mothers aged 15–19 was 24 percent higher than the risk for those of mothers aged 25–34. The relative risk of dying was even higher among children between 12 and 24 months of age, implying that environmental factors associated with early childbearing might be even more important than physiological ones. In a subsequent paper Hobcraft (1987, cited in United Nations, 1989) identified differentials in child mortality risks for women under 20 years of age for five sub-Saharan African coun

tries. His findings, reproduced in Table 7-4, show a greater risk in every country for women under 18 than for women aged 18–19. They suggest that there is indeed an independent effect of maternal age on child mortality.

Besides factors such as physical immaturity and primiparity, the mother may be unable to afford medical care. In this context, a crucial factor affecting a child's risks appears to be paternal support. Gyepi-Garbrah (1985a) contends that children born out of wedlock are at particularly high risk because many are born into poverty and suffer from malnutrition and disease, both of which can impair their intellectual and learning capabilities. Meekers (1988) provides some corroboration. Using the Côte d'Ivoire Fertility Survey of 1980–1981, Meekers finds that children born before cohabitation have higher mortality levels than children born to women who are living with men, whether or not they are the fathers.

In the tightly woven fabric of African societies, adolescent fertility has consequences beyond the repercussions for the young women and their families. Adolescent childbearing, and the illnesses and debility that sometimes accompany it, can impose heavy costs on the society itself. We take up these matters in the following sections.

Although fewer girls might drop out if less stigma were attached to being pregnant in school, in Africa teenage childbearing usually brings an abrupt halt to a young woman's formal education, as Chapter 5 has shown. A few girls find ways to resume their education after giving birth, but the overwhelming majority do not. Policyrnakers worry about the economic consequences of these dropouts from both macro and micro perspectives. At the macro level, high female dropouts rates imply lost societal investments in education, a topic too complex to assess here. At the micro level, early pregnancy can mar the socioeconomic prospects of girls who had planned to complete their educations. Low levels of education stemming from early motherhood may confine these girls to low-paying, unskilled jobs and to low socioeconomic status. However, we have no direct evidence of these connections, and recent research on teenage childbearing in the United States highlights the difficulties of documenting causal sequences (Hayes, 1987).

In fact, in some cases an early premarital pregnancy may be welcome. In an uncertain job climate, a young woman may see clear advantages in bearing a child, whether to prove fertility after a certain age (see, for example, Dynowski-Smith, 1989; République de Côte d'Ivoire, 1990) or to cement conjugal ties to a certain man. Young women often see no disadvantages to teenage childbearing, and young women and their boyfriends are often happy with a pregnancy (Dynowski-Smith, 1989; see also Agounké et al., 1990; Gohy, 1990).

It is important to stress as well that contextual factors affect the impact adolescent fertility has on educational achievement. A key example concerns differences in class and wealth. Whereas most women leave school permanently when they become pregnant, differences among social classes in ability to pay for a safe abortion or child care, not to mention advanced schooling, determine who can and cannot go on with schooling. Thus, the readily available opportunities for good education for girls from well-to-do families may discourage adolescent fertility. By contrast, the slim chances of getting a good education may, along with other factors, encourage girls from poor families to become pregnant. These girls are often least able to keep up the academic pace because demands on them for help in the household leave them little time to study. They are also least able to buy essential materials for school or to enroll in expensive private schools that scrupulously monitor students' comings and goings. Enrolling in schools with fewer qualified teachers and poorer teaching resources, they may perform poorly on national achievement tests, though we have no direct evidence for this comparison. It should not be surprising that such girls decide that ties

with men through pregnancy, may be more advantageous than continuing schooling.

Adolescent fertility may have an effect on child fosterage. Bearing children without firm paternal recognition is likely to have an especially important effect, although few surveys inquire about this matter. Lacking such data, we have used the DHS to examine the percentage of first-born children under age 5 living away from their mothers by mothers' marital status (Table 7-5). The results are mixed. In Liberia, Senegal, Togo, and Zimbabwe, unmarried women foster their small children out more than mar

ried women do. The reason may be that children from previous partnerships (whether or not these were defined as marriages) are obstacles to subsequent unions or to the resumption of education or training. Alternatively, unmarried women could be poorer than married ones and need help with child care. Page (1989) reports similar findings, although she focuses primarily on marital dissolution rather than marital status per se. On the other hand, in Ghana and Kenya, there are no appreciable differences in fostering by mother's marital status, while in Botswana and Uganda, married women aged 15–19 foster out children more than unmarried women do.

The relations of social class and rank to demographic outcomes are profoundly underdocumented in modern Africa. Ranking deeply affected social life in the precolonial era, including even the ranking of wives in polygynous marriage. Wives were explicitly ranked by the functions they performed (see Laburthe-Tolra, 1981, for southern Cameroon), by order of marriage to the household head, or by the politics of the alliance that the marriage forged (for Swazi royal marriage, see Kuper, 1986). Because polygyny has failed to decline in the fashion expected by modernization theorists (van de Walle and Kekovole, 1984), implicit or explicit ranking processes may still affect both wives and their children. The logic of ranking principles in polygynous marriages has important implications for demography. Bledsoe (forthcoming) contends that polygynous husbands are constantly assessing the advantages to themselves and the wider family that can be tapped by promoting one or another relationship. A man may take a new wife with a view to developing connections; an older wife's children may be diligently educated by virtue of her own origin or her children's intellectual capacities; another woman, whose own family is of low status, may be denied a formal union, and thus she and her children may be consigned to managing a rural farm or stripped of inheritance rights. In these ways, the opportunities available to a family's members may become sharply differentiated, and responsibilities toward particular family segments may slacken.

What these observations mean for adolescence is that the children of socially distant fathers may receive little education and may have trouble raising bridewealth (if male) or becoming a first wife (if female). Indeed, in societies with few occupational alternatives, uneducated women need to bear children to create connections they otherwise lack and to add proven fecundity as an attribute of status to their meager resources (Guyer, 1988b). Indirect evidence from western Kenya suggests that emerging patterns of premarital birth and polygynous marriage among schoolgirls may be so motivated (Division of Family Health/GTZ Support Unit, 1988). A young unmarried woman who has a child by a youthful father may find that her

best option lies in becoming a junior wife to an older polygynist who is willing to accept the child.

As we have seen, most children born before the marriage process is concluded are still actively desired by their fathers. For a child whose father avoids paternal responsibilities, however, a number of economic and social obstacles emerge. In rural areas where acknowledged parentage defines a person's status in relation to land, being without an acknowledging father almost inevitably creates hardship. However, in Cameroon, children born before marriage may be separately legitimized by their fathers. Among the Beti, these children otherwise depend on whatever ad hoc terms can be struck with their maternal kin, terms that imply a tenuous right to land (Guyer, 1984). When land is scarce or land values are rising, the claims of children with incomplete parentage may take low priority or be contested by those with stronger claims. Increased differentiation within rural populations would likely result, along with the urban migration of people who have few meaningful ties to the countryside.

It is not clear whether premarital births are considered more problematic in urban or rural areas. Longmore's observations for South Africa a generation ago (1959) suggest that urban fathers made clear distinctions between their children by legitimate versus nonlegitimate unions. In The Gambia, however, administrators of an adolescent fertility survey limited their study to Greater Banjul because they considered premarital pregnancies to adolescents a more significant problem in the major urban area than in rural areas (Gambia Family Planning Association, 1998). Greater tolerance for premarital fertility in urban areas, if this is indeed the more predominant pattern, may stem less from the assumed power of forces of modernity or instability in urban areas than because of the differing importance of fatherhood in rural and urban property systems.

Aside from inheritance issues, the more extreme results of ranking may account for cases of neglected housegirls, street children, and young prostitutes. The consequences of having no acknowledged father, or a loosely attached one, can be disadvantages for children that may subsequently pass from one generation to the next. A young woman who has only a marginal attachment to a man, who comes from a family of low status, or who has little education may have considerably more trouble supporting a daughter with unrecognized paternity through an extended period of "maidenhood" than would a better-endowed partner of the same man. A child with weak kinship supports clearly falls low in the ranking hierarchies and is forced into coping mechanisms that may include early sexual relations and early childbearing outside of marriage. Some literature, moreover, has suggested

that households headed by women create more such households in the next generation, although the issue is controversial (see Peters, 1983, 1984; Kerven, 1984; and Handwerker, 1991).

As adolescent girls increasingly attend secondary school or work, whether in the formal or the informal sector, they are exposed for longer periods to the risk of early, unsanctioned pregnancies that can impede their prospects for success as adults in the rapidly changing economies in which they live. A pregnancy that might fit into a usual sequence of events leading to marriage does not fit when the father is a schoolmate or an older man with little interest in supporting the young woman or her child. If a girl had wanted to remain in school, her expulsion is a double loss: first of scarce national resources allocated to increase the modest number of secondary school graduates; and second of opportunity for the young woman herself. Now, having a child and probably being unable to complete school, her choices may be limited. Her appeal in the urban marriage market diminished, she may become the second or third wife of an older man, perhaps one living in a rural area, or she may take up an informal, less prestigious union with an educated man.

Yet it is impossible to reach confident conclusions about the lifetime effects of adolescent fertility in general and unsanctioned adolescent fertility in particular. Because most pregnancies of adolescents are highly desired, it may be disastrous for young women in many situations to postpone childbearing past the teen years. Especially for married teens, their childbearing capacities may be highly suspect in the absence of a birth, but even unmarried teens who have not given birth might be cast both as bad prospects as wives and as possibly infertile. Becoming pregnant while in school has overwhelmingly negative consequences for adolescents who want to continue their education and whose educational aspirations are supported by their families. To stay in school, they must obtain an abortion that is likely to entail great medical risk. Nevertheless, for only a small minority of young women in Africa will dropping out of school curtail their life chances. The vast majority undoubtedly find that the need to find a suitable husband and begin a family outweighs the risk of losing opportunities for further education and career training. It may even outweigh the possible health risks of early childbearing. For such women, the social and economic consequences of not bearing children are immense.

But just as some urban schoolgirls may be better off becoming pregnant and dropping out of school, many young mothers at the other end of the spectrum—rural, uneducated, married ones—may have problems that re-

ceive litle public attention, precisely because their reproductive patterns are viewed as normal.

These observations underscore the fact that society exerts strong influences on the consequences of adolescent childbearing. In both the rural and the urban configurations we have examined, health outcomes are products of social forces: The problems associated with one model stem from pressures to begin marriage and childbearing early, whereas the deleterious outcomes of pregnancy to unmarried schoolgirls frequently stem from the risks they take to avoid detection and the sanctions that would follow. The unfortunate irony is that regardless of whether young mothers bear their children in rural areas or urban ones, they may receive inadequate health care: In rural areas few health care options exist; and in urban areas, though the availability of medical care is greater, the social condemnation they face often prevents them from seeking help.