In 1994 representatives of more than 180 nations met at the International Conference on Population and Development (ICPD) and approved a Programme of Action to improve reproductive health. To help in the process of defining, implementing, and evaluating strategies to carry out the ICPD program at the request of the U.S. Agency for International Development, the Committee on Population of the National Research Council organized the Panel on Reproductive Health in Developing Countries to: (1) examine the magnitude and severity of reproductive health problems in developing countries; (2) assess the likely costs and effectiveness of interventions to improve reproductive health; and (3) recommend priorities for programs and research. The panel began with the vision of reproductive health embodied in the ICPD: Every sex act should be free of coercion and infection. Every pregnancy should be intended. Every birth should be healthy. No population in the world has yet met these goals. Problems are particularly acute in developing countries: Between 20 and 40 percent of births are unwanted or mistimed, posing hardships for families and jeopardizing the health of millions of women and children. An estimated 50 million induced abortions are performed each year, with some 20 million of these performed in unsafe circumstances or by untrained providers. Almost 600,000 women each year die due to pregnancy-related causes, 99 percent of them in developing countries. Approximately 7.6 million infants die in the perinatal period each year. There are more than 333 million new cases of curable sexually transmitted diseases worldwide each year. Largely as a result of these infections, a high proportion of couples in some regions cannot conceive the children they want. Among those who have sexually transmitted infections who do achieve pregnancy, between 30 and 70 percent will transmit the infection to their infants, and many will deliver prematurely or suffer a miscarriage or stillbirth. Nearly 22 million people are estimated to be infected with the human immunodeficiency virus (HIV, the virus that causes AIDS), of whom 14 million are in sub-Saharan Africa, with rapidly increasing numbers of infected persons in South and Southeast Asia. The risk of transmission of HIV through heterosexual contact is increased two-to five-fold by infection with other sexually transmitted diseases (STDs). Although a great deal of research and experimentation with programs need to be done, there are measures available now that could have an effect on these interrelated problems. We recommend a multisectoral approach involving public services, the private sector, and policy changes. Even poor countries could make progress on the major reproductive health problems with well targeted efforts and the support of the international community.
Healthy sexuality is a vital component of reproductive health, both in its own right as an aspect of emotional and mental well-being and as a determinant of other aspects of reproductive health. Healthy sexuality should include the concept of volition and informed decision-making. Cultures differ in norms about sexuality, particularly those concerning sexual behavior of young people before marriage and women's rights to refuse unwanted sexual relations within marriage or to initiate sexual relations. But many serious health problems are caused by behavior that violates norms that are shared across cultures, such as those against sexual violence and sexual exploitation of children. Sexuality education and communication are needed to contribute to changing norms and behavior to build images of responsible sexual behavior. Sexuality education programs do not currently reach most young people and adults, and programs that do exist often provide little ,Building Solutions more than information about reproduction, contraception, and STDs. Evidence, albeit mainly from developed countries, suggests that well-designed sexuality education can reduce risky sexual behaviors. Curricula should include components on gender roles, self-esteem, decision making, sexual and domestic violence, and communication and negotiation skills. Information alone is seldom sufficient to produce changes in health-related behaviors. Programs will also need to be developed or strengthened to provide specific skills to health care providers and to individuals to improve sexual health. Sexual violence and coercion are widespread problems and have serious health consequences. High priorities in every society should be identification and removal of barriers to victims' access to the law enforcement system and creation of support services for victims. Laws against sexual and domestic violence need to be enacted, and existing laws enforced. Health services have an important role in both providing counseling and treating the victims of violence. Nongovernmental organizations may have an advantage in provision of services, but government support and reforms are needed as well. In addition to direct policies aimed at sexual violence, measures to increase women's autonomy—through higher education, opportunities for financial independence, laws guaranteeing inheritance rights and rights on the dissolution of marriage—are also likely to reduce women's vulnerability to coercion and violence. Where female genital mutilation is common, reproductive health strategies should include, at a minimum, measures to educate the public and formal and informal health care providers about its harmful effects on women's health and to enforce existing bans on the practice. Female genital mutilation is performed on some 2 million girls each year, primarily in Africa and the Middle East. There are several variants of the practice, which is typically intended as a restraint on sexual behavior. It is usually carried out in unhygienic circumstances, most often without anesthesia, and puts girls at high risk of infection and later sexual and genitourinary problems.
INFECTION-FREE SEX AND REPRODUCTION
Reproductive tract infections (RTIs) include: sexually transmitted diseases, endogenous infections that result from overgrowth of organisms normally present in the reproductive tract (such as bacterial vaginosis and candidiasis), and iatrogenic infections due to medical procedures. These infections can have severe consequences, including enhanced HIV transmission, infertility, ectopic pregnancy, and genital neoplasia. Nearly every pathogen that is sexually transmitted can also be passed on to the fetus or infant, often with tragic consequences such as AIDS, fetal wastage, premature birth, permanent neurological impairment, or blindness. To control STDs, we recommend a two-pronged approach to eliminate symptoms and reduce complications for individuals and to interrupt transmission of infections within a population. First, family planning, prenatal and general health services should include capability for management of symptomatic RTIs, since clinical encounters offer opportunities to treat infections among women who would not come into contact with specialized STD treatment settings. Second, services should be designed to meet the special needs of individuals whose behaviors are critical to sustaining STD transmission in communities, such as commercial sex workers and men with multiple sex partners. Primary prevention of STDs requires changes in personal behaviors, supported by changes in community norms. For the general population, interventions should: increase knowledge of the symptoms, signs, and consequences of STDs, encourage delay in initiation of sex among adolescents, promote use of condoms and other barrier methods among those who are sexually active in relationships that are not mutually monogamous, and identify sources of quality care for suspected infections. The campaigns that appear most successful have used a range of media, have been designed with attention to local cultural norms, and have employed audience segmentation and professional production and pretesting. Condom social marketing programs have used a range of print and broadcast media, widespread distribution, and point-of-purchase advertising to increase condom sales, even in some of the world's poorest countries. Mass media campaigns can be a valuable channel for these efforts, but alone are likely to be insufficient to catalyze widespread behavior change. Family planning, prenatal, and primary health care facilities should ensure that symptomatic individuals can obtain appropriate management of STDs. Particularly in settings where resources are very limited, the highest priority for RTI clinical services should be the case management of STDs both because these infections most frequently result in severe complications for individuals and because, unlike other RTIs, they can spread through communities. Standardized case management using currently available tools should be a routine responsibility of family planning and other reproductive health services. Management of STDs can and should be offered by every facility, program, or country that wishes to improve reproductive health. At a minimum, family planning and primary health care facilities should ensure that symptomatic women and men can obtain appropriate management of genital ulcers, discharge syndromes, and pelvic inflammatory disease. The use of locally adapted versions of standardized algorithms for syndromic management developed by the World Health Organization should help achieve this goal. These algorithms do not require laboratory support and perform well for genital ulcers in both sexes and for urethritis in men. Unfortunately, the algorithms perform less well for the syndromes that are most common among women—vaginal discharge and lower abdominal pain. The performance of these algorithms may be improved using locally appropriate means to assess behavioral risk factors. Treatment of sex partners and risk reduction counseling for infected individuals and their partners are essential to the success of STD clinical prevention services. Treatment protocols at all levels must be developed and periodically revised in light of local disease and antibiotic resistance patterns. Sentinel surveillance or special studies of etiologies of STD syndromes and antibiotic resistance patterns are needed to guide these decisions. STD screening, regardless of symptom status, and treatment as appropriate should be provided for sex workers. Screening services require the commitment of resources for etiologic laboratory testing and for targeted outreach activities. Together with treatment of symptomatic men, STD detection and treatment among sex workers are central to limiting the spread of STDs in the community. Over time the primary and secondary prevention efforts aimed at these groups should help reduce the STD burden among clients attending family planning and other health facilities. Targeted health promotion efforts should aim at reduction in number of sex partners and risky sexual practices, together with promotion of condoms and other barrier methods, and early health care seeking. Peer counseling and skill building should be tested in more settings in developing countries. Prenatal and delivery care should include syphilis screening and treatment during pregnancy and newborn prophylaxis for gonococcal eye infections. These are simple, inexpensive interventions that are highly cost-effective in most parts of the developing world. Prevention of endogenous infection requires efforts to improve women's and men's knowledge of reproductive physiology, menstrual and personal hygiene, health-seeking behavior, and adherence with prescribed Efforts should focus on reducing use of harmful intravaginal substances (i.e., douches and desiccants) and on curtailing inappropriate use of broad-spectrum, systemic antibiotics. The latter will require changing prescribing practices of both traditional and allopathic health care providers, pharmacists, and family members. Family planning and other health services should use simple, inexpensive tests of vaginal secretions for symptomatic women and provide appropriate management of endogenous RTIs. Infection prevention, consisting of simple measures such as hand washing, appropriate use of gloves, and adequate sterilization of instruments, should be a minimum standard. Prevention of iatrogenic infection requires improvement in overall quality of reproductive health services, particularly transcervical procedures. One of the most effective ways to prevent iatrogenic RTIs is to reduce the number of unsafe abortions by improving the supply of contraceptive services, promoting the use of emergency contraception, and decriminalizing abortions.
In developing countries outside sub-Saharan Africa, between one-tenth and one-third of all recent births are reported as unwanted, and the same percentages are reported to be the result of mistimed conceptions. In Africa these percentages are typically lower, but since fertility rates are high, the proportion of women and families affected by unintended pregnancies is as high as elsewhere. Reducing unwanted pregnancies promotes maternal health mainly by reducing the number of times that a woman is exposed to the risks of pregnancy and childbearing in poor environments. Children's health is also affected: unintended pregnancies are disproportionately in high-risk categories, and lower fertility results in increased family and social investments in health care, schooling, and nutrition for the planned children. To meet existing and growing unmet need for contraception, access to contraceptive services should be expanded through clinical and nonclinical channels, including postpartum care and STD prevention services. Reducing unmet need for family planning through safe access to a range of contraceptive methods is a high priority for reproductive health programs. A basic task for family planning and health programs is to support informed choice by clients. Information, education, and communication programs and improvements in counseling are still needed, even where family planning programs are well established, because of gaps in the knowledge of providers, clients, and potential clients about how to use contraceptives and the advantages and disadvantages of the methods available. Use of contraceptive pills for emergency contraception appears safe and effective for women who have unprotected mid-cycle intercourse. Information on the techniques should be provided widely to health care and family planning staff and those who may need it. Unsafe abortion remains a leading cause of maternal death. Access to safe means for abortion care, including early intervention to treat abortion complications, is needed to reduce the numbers of deaths. Even where abortion is legal, services are often low in quality, stigmatized, and access is difficult, making abortion needlessly dangerous. In those countries governments should ensure (through direct provision or regulation) adequate equipment and training for manual vacuum aspiration in the first trimester of pregnancy. Where medical supervision and surgical backup are feasible for medical abortions, the option should be available for first-trimester abortions. Health care and family planning providers will require training on medical abortion and contraindications. Where abortions are illegal, health services should ensure that women who have had septic and incomplete abortions are treated appropriately and promptly. Where the prevalence of infertility is high, as in much of Africa, measures to reduce infertility should be a high priority, including programs to control STDs, provision of aseptic abortion, and early treatment of septic abortion.
HEALTHY PREGNANCY AND CHILDBEARING
The major direct causes of maternal deaths in the developing world are hemorrhage, sepsis, obstructed and prolonged labor, septic abortion, and hypertensive disorders of pregnancy. Even among survivors, consequences of these conditions can be severe. It makes sense to consider maternal and perinatal health together, because both mother and child are affected by the direct causes of death and disability and because the interventions designed to promote maternal and perinatal health often overlap or are operationally linked. Priority should be given to providing women with essential care for obstetric complications, in particular by establishing or strengthening obstetric units at hospitals. The quality and appropriateness of skills for the management of labor should be upgraded and maintained. The major causes of maternal mortality cannot be predicted or prevented well enough during pregnancy to allow reliance on primary prevention and screening for high risk. Many previous efforts to reduce maternal mortality in developing countries have foundered because they relied solely on attempts to train traditional birth attendants, screen high-risk pregnancies, and refer women to expensive, distant, and ineffective sources of treatment. Improvements in maternal death rates will require access to facilities and trained providers and equipment in facilities that can carry out essential care of obstetric complications. Most births in developing countries take place outside health facilities, so the most effective strategy is to ensure that complications of pregnancy and delivery are recognized once they occur and that women are taken to a facility where essential care of obstetric complications of adequate quality is provided. This strategy has four elements: First, a life-threatening complication must be recognized by the woman, her family, traditional birth attendant, or others in attendance. Second, those in attendance have to decide to seek appropriate care and then, third, get the woman to an adequate facility in time. Barriers to access currently include distance, the cost or lack of transport, cost of the services, geographical or weather constraints, and perceived poor quality or attitude of the providers. Lastly, care for obstetric complications and neonatal care in the facility have to be adequate. The few existing studies of the quality of maternity care identify major deficiencies; many preventable maternal deaths are due to inappropriate or delayed care in health facilities. Most efforts to improve quality of care have focused on training—for example, training midwives in life-saving skills and interpersonal communication. Training of one cadre of workers is not enough to sustain improved practices. Programs must also train those to whom midwives are supposed to refer women and devise policies that allow trainees to put their new skills to use and improve management and supervision, information systems, logistics, and supplies. Protocols for the management of obstetric and neonatal complications are useful for medical care providers to guide and coordinate their actions, know their limits and next steps. In cities in some middle-income countries, obstetric care can be too interventionist, with potential harm for the health of mothers and infants and wasting resources. Inappropriately aggressive care in urban areas often coexists with a lack of access to obstetric care in rural areas and for the poor. Some obstetric problems may be managed or stabilized by trained midwives or other providers at a peripheral level (antibiotics for infections, sedatives for eclamptic patients) prior to referral to a site with more complete essential care of obstetric complications. How to do this effectively and for which complications are important topics for operational research. Interventions are needed to improve community awareness of, support for, and involvement in the transportation of women with obstetric complications to facilities that can deliver essential care. Families (and those who influence them) need to know signs of obstetric complications and where to seek care. Reproductive health services need to include information and education programs about early recognition of signs and symptoms of obstetric complications and when and where to seek needed help. These campaigns can draw a wide variety of media, including mass communication, face-to-face communication as part of child survival or family planning programs, and existing prenatal care. Appropriate prenatal care should include screening and treatment for syphilis, for anemia, and detection and management of pregnancy-induced hypertension. Delivery care should include neonatal prophylaxis for ophthalmia neonatorum. Postpartum care should include contraceptive counseling. Prenatal, delivery, and neonatal care provide multiple opportunities to promote reproductive health, many of which are missed opportunities when services are fragmented. Prenatal care of some kind now reaches the majority of pregnant women in developing countries and should be used to provide more effective interventions to improve both maternal and perinatal health.
PROGRAM DESIGN AND IMPLEMENTATION
Even in countries where fertility decline has already begun, the momentum of population growth ensures that there will be significant increases in the number of women aged 15-49 and rapid increases in the number of young people during the next several years. Just to keep up present inadequate levels of services would require substantial growth in absolute terms; to expand and improve services will require both increased resources and skilled management. Although no one configuration of reproductive health services will serve all needs, a number of potentially effective clinical and nonclinical interventions can be implemented now at different levels of the health care system. For some aspects of reproductive health, there is a good deal of experience with different types of service delivery and different scales of operation. But development of comprehensive reproductive health services will require experimentation. Research on the determinants of organizational effectiveness in provision of reproductive health care is urgently needed. Reproductive health services should concentrate on strengthening coordination, referral, and linkage among three principal service domains of reproductive health programs: STD prevention and management; pregnancy and contraceptive services; and delivery care for both mothers and newborns. Whether to integrate services at different levels cannot be decided in the abstract. Functional integration may increase the convenience of services for clients, increase the likelihood that their particular needs will be diagnosed and met, and minimize ''down time" for multipurpose providers. Its disadvantages come when providers are overloaded or insufficiently trained and supervised for some of their functions. Examples of successful functional integration of reproductive health services already exist, particularly in provision of information and counseling. HIV/AIDS information and prevention messages have been incorporated into some family planning information, education, and communication efforts. Provision of information about STDs and about danger signs in pregnancy and labor, and where to go for help, should likewise be added to the duties of every service provider who comes into contact with adult women and their families. There are also examples of successful linkage of different reproductive health services at the clinic level. In many countries, linkage with child health services is a major convenience for mothers or legitimizes what would otherwise be an embarrassing clinic visit or one considered not worth the cost. Administrative integration in public systems is often difficult to impose. Its advantages are that it allows coordination and setting of priorities across services and that it spreads the cost of overhead services across many programs. The disadvantages are that particular functions may be neglected and managers may not feel accountable for their successful performance. Effective reproductive health programs will require a focused and measurable set of objectives, adequate resources, and, sometimes, generation of demand for their services. Reproductive health is most likely to succeed if objectives are focused and managers held accountable for their achievement. Implementation of effective reproductive health programs will require significant and continuous building of local capacity in systems such as training, supervision, and management; procurement and distribution of supplies; information, education, and communication efforts; and record-keeping and evaluation. Both public- and private-sector organizations have an important role to play in increasing awareness of reproductive health services and how and when to use them—particularly recognition and treatment of RTIs and pregnancy complications. Research is needed on the determinants of demand for specific components of reproductive health services, especially how women and men come to believe treatment is needed and what motivates them to seek care in various settings.
COSTS, FINANCING, AND SETTING PRIORITIES
Financial, managerial and administrative resources for health are tightly constrained in low-income countries. Recommendations for reproductive health must be considered in light of such overall resource constraints. Reproductive health services are among the most cost-effective health investments available to both low- and middle-income countries. More research is needed at a country level on costs, as well as evaluations of effectiveness of operational scale programs. Cost-effectiveness estimates are imprecise, but even allowing for a wide margin for error, many reproductive health interventions rank high in comparison with other potential health sector investments and should receive greater priority in health sector budgets. Analyses using cost models developed for related interventions show that costs of a package of basic interventions, relative to current health care spending, can range from modest, in moderate-income settings, to significant, in low-income settings. Cost estimates vary greatly, depending on the salary costs and the degree to which personnel and infrastructure are fully utilized and are shared with other health services. Per capita costs can be high when expensive facilities are underutilized, as is often the case. Improvements in quality of services and communicating information about their availability and benefits may help achieve operation at efficient scale. Both costs and effectiveness will change over time, as service delivery organizations learn by doing and as increased demand for services changes the scale of operations. Cost-effectiveness studies are not a once-and-for-all effort to describe the best set of services, but a framework for continuous evaluation and redirection as reforms are introduced. Public-sector financing need not preclude private-sector provision of reproductive health services. Subsidies should be targeted to the poor, especially in middle-income countries and for well-established services. Improvements in reproductive health are probably best achieved by a mix of public and private finance and provision, as well as other government instruments, such as mandates and regulation. User fees are increasingly common in developing countries. While they can generate resources and spur efficiency, user fees should be implemented with caution and accompanied by monitoring and evaluation. Many of the services called for in this report are not only of community benefit rather than only individual benefit, but are also new and unfamiliar to their intended clients. Efforts to make them self-sufficient too quickly could stifle attempts to build demand. Safeguards are needed to protect access to services for the poor and services with significant public health benefits. Healthy Sexuality We use the term "healthy sexuality" to incorporate a sense of volition in sexual relations and control over one's body. Sexual autonomy is thus part of healthy sexuality, if choices are informed and responsible. Rational adults need to know the potential consequences of their actions, and one person's autonomous decisions cannot be called healthy if they are coercive to another person. Healthy sexuality is related to reproductive health in three ways. First, it is a determinant of reproductive or sexual health, in physical terms, because its lack can result in higher risk of problems such as unintended pregnancies and sexually transmitted diseases (STDs), including human immunodeficiency virus (HIV) infection. Second, healthy sexuality may be a determinant of future reproductive health, since it affects people's ability to make use of reproductive health services, and coercion and childhood sexual experience may affect adult sexual satisfaction and risky behaviors (Finkelhor, 1995). Third, healthy sexuality can be viewed as an intrinsic aspect of health, defined to include emotional and mental well-being. Although we concentrate on healthy sexuality in women, many of the arguments made here can be extended to men. In addition, although there is variation in sexual expression, many of our arguments are valid across a range of kinds of sexual expression.1 Yet, there is significant 1 At the same time, there is also a gradient of social acceptability of these other forms of sexuality, with the greatest agreement across cultures in disapproval of deviant forms of sexual behavior such as commercial sexual exploitation of children, and the least agreement about different kinds of sexual relationships between consenting adults. Cultural variation in sexual norms and practices and in the social meaning of healthy sexuality. We identify three different aspects of sexuality over which individual control is possible and on which there would be varying levels of agreement with regard to autonomy: control in the sense of protection from sexually related violence or coercion, including rape and sexual exploitation of children. This aspect of sexuality is the one most likely to find consensus across cultures. control over sexual relations within a stable union, in particular the right to refuse sexual relations, whether physically safe or ''unsafe" (that is, likely to lead to infection or to an unwanted pregnancy). The right of women to refuse sexual relations is now more openly discussed and is also being officially acknowledged, at least in principle; the Fourth World Conference on Women in Beijing in 1995 endorsed this right in its Platform of Action. control over access to sexual relations, that is, the right to seek sexual relationships. This aspect of sexuality has the least agreement across cultures, and we do not make policy recommendations in this area. Not only do these three kinds of control vary across and, at times, within cultures, they can also vary during a person's life. Norms, behavior, and their implications for reproductive health can also change over time. In this chapter we first examine sociocultural variations in these three aspects of healthy sexuality. We then look at two special problems in healthy sexuality, sexual violence and female genital mutilation. The role of public policy and interventions to promote reproductive health is discussed in the last section. We emphasize cross-cultural variations in norms and institutions and highlight the major changes in sexual behavior that are occurring.
CULTURAL CONTEXT OF SEXUALITY
Volition in sexual relations can have distinct meanings, defined as conditions under which women would have an acknowledged right to refuse sexual relations: when there is the possibility of exposure to an unwanted pregnancy, when there is the possibility of exposure to an infection, especially of the reproductive tract, and when the sexual relationship in general or at any particular time is unwanted for other reasons (including physical tiredness). The third condition is a topic of much disagreement. Though the concept of rape of a spouse has been defined under legal codes in some areas, husbands have been considered to have a right to sexual relationships with their wives in most cultures, though in virtually no culture is this right unconditional in principle, even if it is sometimes in practice. Even separation or divorce may not always limit men's right of sexual access. For example, some Latin American studies suggest that once a woman has been "possessed" by a man at one point, she loses her right to refuse sexual relationships with him (Pick, Givaudan, and Aldaz, 1996). The formal and extreme form of refusal of sexual relationships with a socially sanctioned partner is the seeking of a separation or a divorce. The woman's real (as opposed to notional) access to this step is determined by a mixture of legal, social, and economic factors. Not least important is the emotional and physical support that the separated woman can expect from her natal family, friends, or others. All these factors differ widely across cultures and societies: for example, practices such as village exogamy often make it physically difficult for a girl's parents to be even aware of, let alone do anything about, their daughter's marital difficulties. Such practices are related to norms that forbid parental interference in a married daughter's life. The first act of sexual intercourse can be particularly traumatic in many cultures because it is typically the time that a woman's right of refusal is the weakest, whether such sexual activity occurs in a casual relationship or a formal union. Ignorance, a weak bargaining position, and social pressure (as well as the desire to please) create a situation in which the loss of virginity takes place under conditions similar to those that would conventionally be considered rape (see, e.g., reports of researchers in India and Algeria, in Heise, Moore, and Toubia, 1995). In some cultures in the Middle East, a "deflowering" ceremony to demonstrate the virginity of a new bride is held on the wedding night, and relatives, friends, and neighbors are invited: guests wait outside while the hymen is ruptured by sexual intercourse (or, sometimes, by hand) and a woman relative attending brings out proof of virginity in the form of a handkerchief soaked with blood. This experience is typically traumatic for women (Khattab, 1996). Norms about abstinence at specified times, such as postpartum abstinence (Caldwell and Caldwell, 1981), may not be gender neutral when sexual activity is allowed for men but denied to women. For example, between 42 and 49 percent of currently married men report a casual sexual relation in the past year in Guinea Bissau, Lesotho, and Côte d'Ivoire, while for women it is less than 20 percent in all the countries studied (Caraël, 1995 It was observed that nonregular sexual activity is generally greater among single than among currently married women, but formerly married and even currently married men report higher rates than do single men. Many Africans consider it legitimate for a man to have sexual access to cowives, mistresses, or commercial sex workers during any particular wife's postpartum period. Indeed, the ease of availability of alternative sexual partners is cited by many women as an important reason for reducing the period of postpartum abstinence. In this sense, the right of a wife to refuse sexual relations may be a double-edged weapon. When submission to sexual advances, however unwanted, is the only route to other kinds of economic or domestic security, the right to refuse sex may mean very little in practice.2 One reason for refraining from sexual activity in many cultures has been the fear of an unwanted pregnancy in the absence of other fertility control methods. This issue can be separated into two parts. First, can a woman insist on the use of contraception in such cases, and then if her partner is noncompliant, can she refuse sexual intercourse? This question in turn hinges on a woman's ability to practice contraception herself, with or without her partner's consent. If fears of unwanted pregnancy legitimize a woman's refusal of sex, then freer access to contraception can in one sense worsen women's control over their bodies: while contraceptives may free women from the burden of unwanted pregnancies, they may at the same time remove one of the few excuses accepted by men to avoid unwanted intercourse (Folch-Lyon, Macorra, and Shearer, 1981). The health implications of the lack of a right to refuse sexual intercourse are even more serious when one examines the right to refuse unsafe sex. The current evidence on this matter is scarce, but there is some evidence that in many parts of Africa, the fear of infection is slowly becoming a legitimate ground for refusing sexual relations, most likely due to the increasing prevalence and awareness of HIV infection. For example, Orubuloye, Caldwell, and Caldwell (1993) report that among the Yoruba in Nigeria, women now feel more free to refuse sex with infected partners. Similarly, Awusabo-Asare, Anarfi, and Agyeman (1993) found their female respondents in Ghana relatively free to refuse sexual relations with a partner infected with an STD. However, a refusal based purely on a partner's promiscuity, where an infection was not established, was not believed to be similarly valid. Healthy sexuality in the more positive sense of access to sexual relations can be thought of under three separate aspects: control over when sexual activity starts, control over the choice of one's sexual partner, and 2 Sexual relationships are as hierarchical as, and often mirror, other kinds of gender inequality and power structures. Increasing women's sexual autonomy may result from improvements to female status in areas that have no obvious connection with their sexuality. Control over the frequency or intensity of sexual activity. But numbers reveal little about motives; one cannot infer women's levels of autonomy from high or low levels of premarital sexual activity. Asian cultures, on the whole, are characterized by discouragement of premarital sexual behavior, and Middle Eastern cultures are generally even more insistent on premarital chastity. There has reportedly been no "sexual revolution," though political and feminist efforts have been made to encourage greater gender equality and female autonomy in spheres of life other than sexuality (Hathout, 1989). Premarital sexual control is not just an ideal in many societies: institutions and cultural practices support enforcement of this norm, although such institutions and practices are under increasing pressure to change. For example, one frequently stated reason for the traditionally early marriage of girls in South Asia and the Middle East has been the need to ensure their virginity at the time of marriage. The literature of South Asia is replete with real and fictional accounts of the methods used to confirm such virginity, as well as the opprobrium heaped on brides who fail to meet the requirement. Female seclusion is also a common way of preventing unwanted male-female interactions. Seclusion has obvious implications for the ability of women to use clinical services of all types, particularly when it is their own health needs rather than those of their children for which they seek care. Seclusion of women as a way to control their sexuality operates not just through overt seclusion, but also through norms about matters such as the correct occupations for women (see, e.g., the constraints on working class women in Naples described by Goddard, 1987). Norms about seclusion and work can also operate to limit women's ability to serve as health care and family planning providers (see, e.g., the discussion of Bangladesh by Koenig and Simmons, 1992). More often, adolescent sexual activity in traditional societies is restricted by less drastic means. For example, in South Asia, norms about female seclusion do not necessarily require young girls to be secluded from all males in the household or even in the village. Instead, premarital chastity is promoted by a cultural proscription on intrakin or intravillage marriage so that all men in a girl's village of birth are in principle her brothers and any relationship that develops is by definition incestuous. In many cultures the importance given to virginity may reflect not so much a concern about premarital sexual activity as about premarital pregnancy. Strategies to preserve the virginity of unmarried girls may therefore emphasize the latter. Whiting, Burbank, and Ratner (1986) record several such strategies, ranging from a ban on all sexual activity to relatively unlimited freedom to experiment with sexual activity that stops short of actual intercourse (see also Du-Toit, 1987). Prohibitions against male sexual activity before marriage are universally
TABLE 2-2 Males and Females Who Had Not Had Intercourse at Time of First Marriage/Partnership, Among Ever-Married Men and Women, by Current Age: in percent Age 25-29 30-34 35-39 40-49 Country or City M F M F M F M F Africa Côte d'Ivoire 35 52 24 56 23 58 29 68 Kenya 15 31 8 25 5 30 6 32 Tanzania 30 66 24 77 28 82 42 83 Lusaka 38 51 26 56 23 68 28 65 Asia Manila 36 83 29 88 21 85 24 87 Singapore 57 92 71 95 56 95 68 94 Thailand 28 98 26 96 27 95 37 99 South America Rio de Janiero 17 71 14 69 9 75 13 81 SOURCE: Adapted from Caraël (1995). weaker and less strictly enforced than prohibitions against female sexual activity before marriage. Condonement, and even encouragement, of sexual experience by young men affects the reproductive health of both men and women. Table 2-2 shows data from surveys, carried out by the Global Programme on AIDS, on sexual intercourse prior to marriage or partnership among ever-married men and women. Males and females in the Kenya survey and males in the Rio de Janeiro survey were most likely to report sexual activity prior to their first stable union. Thai, Singaporean, and Manila women reported the lowest rates of sexual activity prior to marriage or union. Quite apart from the gender inequality implied by such differences in sexual norms, in an environment in which adolescent females are denied such activity, adolescent males seeking sexual activity turn to other partners—usually commercial sex workers, as the anthropological and survey evidence from Thailand describes (see, e.g., Thongkrajai et al., 1993) or to older married women, often within the larger extended family (see, e.g., Goparaju, 1994, on India). Because their husbands have often had such sexual contacts, young married women are put at risk of acquiring STDs at a stage in life when they are culturally least able to identify or seek medical or nonmedical help for socially embarrassing conditions such as reproductive tract problems. Many popular magazine or newspaper columns on health-related matters include letters from readers that refer to female reproductive tract problems for which the letter writers are embarrassed to seek medical help (see, e.g., Basnayake, 1985). The anthropological literature also stresses the inability of young women to admit to and seek help for reproductive health problems because this kind of illness, especially if it is feared to result in infertility, can lead to legitimizing a husband's search for a new wife (see, e.g., Bourquia, 1990; Doniger, 1991); this is discussed further in Chapter 3. Many religious or cultural traditions pay at least lip service to the idea of male fidelity or to the idea of carefully restrained male infidelity. The latter approach includes norms about polygyny and about male access to other women during periods of postpartum abstinence. These social and cultural traditions have implications for the reproductive health of women in two ways: directly, if it means that they lose their sexual access to their husbands when husbands have other alternatives;3 indirectly, because the wider sexual networks increase their risk of infections. The practices of polygyny, premarital commercial sex, extramarital sex during postpartum abstinence (and during periods that males travel to cities, usually for work) are all at least partly institutionalized in the belief that males need sexual release. In contrast, very few cultures have similar beliefs or condoned practices for women who do not have access to regular sexual relations. Women who are unmarried, widowed, or separated from their husbands because of migration are usually expected to have a celibate life. The restrictions are particularly severe in Asia and the Middle East and particularly stringent in the case of widow remarriage. In South Asia in particular, universal marriage is generally prescribed, while widow marriage is generally proscribed, though a strong reform movement in India has led to a much greater tolerance of widow remarriage. Restraints on the sexual lives of women are not necessarily the cause of sexual dissatisfaction for most women in traditional cultures. Various norms and institutions serve to legitimize the restraints. Because of social conditioning, as well as women's lack of information about female sexuality, most women may be content with restricted sexual expression. Beliefs about female sexuality are strongly connected to notions of shame and honor, which seek to determine how women may express their sexuality and how their activities need constant monitoring to prevent an undue expression of such sexuality. For example, folklore refers 3 Polygyny may not always have this connotation, since wives may welcome cowives for a variety of reasons (Bledsoe and Cohen, 1993). Often to women's sexual greed or wantonness that makes them neglect home and family in the search for lovers and to the need to control this tendency (see, e.g., Constantinides, 1985). This image of women's insatiable sexuality can affect their sexual and reproductive health. In the early days of Mexico's family planning programs, men's fear that women who were not pregnant and exhausted from child care would become sexually promiscuous and cheat on their husbands was found to be a potential barrier to the success of the programs (Folch-Lyon, Macorra, and Shearer, 1981). There are also cultures in which premarital sexual activity by young girls is condoned, even encouraged, and a resulting pregnancy is welcomed; this is a pattern in much of sub-Saharan Africa (see, e.g., Caldwell, Caldwell, and Quiggin, 1989; Meekers, 1990, 1992; van de Walle, 1990). One difficulty in interpreting the meaning of this practice lies in the fluid definition of marriage that prevails in much of Africa and the Caribbean, so that sexual activity is hard to classify as either pre- or postmarital (for a review, see Bledsoe and Cohen, 1993). Marriage is a process in which cohabitation, ceremonies, and childbirth can occur in varying sequence: the issue is not so much whether a birth occurs to parents who have had a marriage ceremony, but whether the birth is considered socially legitimate. That legitimacy involves, among other things, whether the newborn has an acknowledged father, whether the adolescent parents are in a stable union, whether the union has the approval of the larger kin group, and so on. Even in cultures that have traditionally frowned on premarital female sexual activity, there have recently been strong signs of change. There are reports of increases in premarital sexual activity from all regions. Several surveys of young unmarried women report that they are now under strong social and peer group pressure to engage in premarital sex: in Thailand (see, e.g., Thongkrajai et al., 1993); Nigeria (Renne, 1993; Feyisetan and Pebley, 1989; Oyeneye and Kawonise, 1993); Senegal (Diawara, 1979); Ghana (Anarfi, 1993); Kenya (Ferguson, gitonga, and Kabira, 1988); Côte d'Ivoire (Meekers, 1990); Liberia (Taylor, 1984); North Africa (Mernissi, 1977); and India (Kapur, 1973; Savara and Sridhar, 1994). Two major reasons for an apparent rise in premarital sexual activity have been proposed. The first is that ages at first marriage have risen, so that women now have more years "at risk" of premarital sexual activity. This reason has been proposed as the major cause of the rise in premarital adolescent sexual behavior in much of Africa (Bledsoe and Cohen, 1993). Premarital pregnancies can pose an increasing social problem if they disrupt young women's schooling, their chances for desirable marriage, and other preparation for adult roles, even when age-specific fertility rates are decreasing (Bledsoe and Cohen, 1993). The second reason is the increasing "sexualization" of cultures worldwide (see, e.g., Udry, 1993; Burt, 1990). The hypothesis is that various features of modern life have greatly increased both the desire for sexual activity as well as the possibilities for such activity. Among the factors thought to be responsible for such sexualization are the mass media; increased migration and urbanization with resulting opportunities for social interaction; and increased materialism and relative or absolute impoverishment, which have led to sharp rises in commercial sex. Given the variety of factors, it is impossible to determine whether increased adolescent sexual activity is associated with increased female sexual autonomy. The term "commercial sex" covers a very wide variety of practices (Gillies and Parker, 1994), and it is difficult to quantify the extent of the behaviors, to generalize about the motivations of those who provide the sexual services, or to estimate adequately the consequences for sexual and reproductive health (Cohen and Trussell, 1996). The United Nations Global Programme on AIDS sponsored a series of household surveys of sexual behavior in developing countries, which included questions on commercial sex, defined as contacts within the last 12 months, with a nonregular partner, for which gifts or money were exchanged. This is a broad definition, and it is likely that interpretations varied among countries, that commercial sex workers were underrepresented among respondents, and that the behaviors were underreported because of social stigma or illegality (Caraël, 1995). The proportions of men reporting such contacts ranged from 1 percent in Sri Lanka to nearly 25 percent in Tanzania: "In some populations, sex in exchange for money and gifts represents an important part of sexual behavior, while in others it plays only a marginal role." (Caraël, 1995:122). Smaller, more intensive studies of commercial sex work have documented a wide variety of motivations and economic circumstances. Pickering and Wilkins (1993), in a careful study of women sex workers in the Gambia, found many from relatively well-off homes, in contact with their families, for whom commercial sex was an economic choice, an attempt to make a lot of money quickly and possibly become mistresses of rich men. Orubuloye, Caldwell, and Caldwell (1994) found that commercial sex workers in urban Nigeria tended to be better educated than the general population. Other studies have found more evidence that commercial sex is a result of families' economic hardship and very limited alternative opportunities for employment. For example, Ghanaian women migrants to the Cote d'Ivoire who engage in commercial sex are often single mothers from poor families with few alternative sources of income (Anarfi, 1993). Commercial sex workers in Thailand have little education, come from poor families, and send back a large proportion of their incomes as remittances (Archavanitkul and Guest, 1994). Sexual "autonomy" can have little positive meaning for women (or men) who feel forced to sell sexual services. Even if selling sexual services is not invariably defined as the result of coercion, women and men who do so are at high risk of sexual violence (discussed in the next section), as well as high risk of sexually transmitted diseases (see Chapter 3).
Sexual violence, both within and outside a formal relationship, occurs in many women's lives (Heise, Pitanguy, and Germain, 1994; Heise, 1994). Data on it are limited, in part because violence has only recently been recognized as a public health issue and an important topic of research and in part because of methodological problems, such as unwillingness to discuss or report the problem and differences among existing studies in definitions, samples, and research methods. Enough is known, however, to justify inclusion of violence against women as a serious reproductive health problem. In population-based surveys in developing countries (using various reference periods), 30 percent of women report being beaten by spouses in two Caribbean islands; between 56 and 67 percent in stratified samples in Papua New Guinea; 20 percent in Colombia; and 60 percent in Santiago, Chile (summarized in Heise, Pitanguy, and Germain, 1994:Table 1). In a recent survey in Uttar Pradesh state, India, more than one-third of men reported that they beat their wives (Martin et al., 1997). Violence against women resulting in death is seen across a range of economic and cultural conditions.4 The scope of the problem cannot be deduced from figures for homicide alone. Suicide, whether real or apparent, is often the outcome of predeath violence. The classic example is provided by the phenomenon of "dowry deaths" in parts of northern India, where young married women are often found to have died in accidents or committed suicide and where there is a very fuzzy dividing line between suicide or accidental injury and homicide. A cross-cultural survey drawn from research in Africa, Peru, Papua New Guinea, and other Melanesian islands found that marital violence was a defining feature in female suicide (Counts, 1987).5 4 Several anthropological and ethnographic reviews have documented the existence of societies and cultures in which male violence against women is not endemic, showing that current high levels in other societies are hardly inevitable (see, e.g., Gilmore, 1990; Levinson, 1989; Sanday, 1981; Counts, Brown, and Campbell, 1992). 5 Men are more likely to die from intentional injuries (homicide and suicide combined) than are women. Murray, Yang, and Qiao (1992) show death rates for males ranging from 2.5 to 10 times those for females in populations at different levels of adult mortality. Our focus here is on violence against women, though, because violence and the threat of violence against women are more directly connected to sexual coercion and other reproductive health problems than is the case for men. Gender-based violence has an obvious impact on women's control over their sexuality and therefore their sexual health. The negative consequences of violence that have a direct bearing on reproductive health include physical injuries, STDs, pelvic inflammatory disease, unwanted pregnancy, and unsafe abortion or miscarriage—as well the mental or psychological aspects of sexuality, such as depression, anxiety, and sexual dysfunction (Jenny et al., 1990; Koss, Heise, and Russo, 1994). The fear of domestic violence can make a woman unable to negotiate condom use or practice contraception, if, for example, she fears accusations of infidelity (Folch-Lyon, Macorra, and Shearer, 1981; Fort, 1989) Rape can change a woman's relationship with her partner and her family or have serious consequences for her social or economic status. The social and psychological consequences are particularly negative because most acts of violence against women, including rape, are committed not by strangers but by persons known to the woman, especially family members (Koss, Heise, and Russo, 1994). Studies in India (Paltiel, 1987), Egypt (Mashaly, Graicer, and Youssef, 1993), Kenya, Bangladesh, and Thailand (United Nations, 1989) report that females are more frequently murdered by family members, especially male partners, than by other aggressors. Although most societies do not explicitly condone violence against women, they often explicitly or implicitly support the socialization of the male psyche as one based on domination and aggression. Such gendered socialization is related to the expression of domination and aggression through violence against those perceived as weak (Heise, 1994). Even if societies do not explicitly condone violence against women, they may do little to stop it through legal channels. The concept of rape does not exist in many penal codes. In many Latin American countries, rape, even by strangers, is considered a ''crime against morality" rather than a crime against the person, like homicide. As a consequence, if the judicial system does not consider rape victims to have impeccable morals, the crime may not be prosecuted (Barricklow, 1993). Domestic violence is also neglected by the police and by the courts in many countries. For example, in Bolivia, aggressors who injure family members can only be punished by the legal system if the injuries incapacitate the victim for at least 30 days (Ford Foundation, 1992). In some Islamic countries, there are stringent requirements for corroboration from eyewitnesses to prove sexual violence against women, and women who fail to prove their complaints leave themselves open to accusations of adultery or fornication. Such laws and practices reflect social norms that condone violence against women, both within the home and outside it. A culture of violence, as well as the individual experience of violence, can breed an atmosphere of fear and tension among women that is detrimental to their sexual health, as well as their full participation in many activities of daily life. One particular kind of sexual coercion deserves separate mention, the sexual exploitation of young children. The consequences of this form of coercion are likely to be even more traumatic and long-lasting than those of violence against adult women. Child abuse is even more likely than abuse against women to be perpetrated by persons known to the young person. Evidence from the United States indicates that a history of childhood sexual abuse is associated with unhealthy sexual behavior as an adolescent or adult and greater incidence of sexually transmitted diseases (Laumann et al., 1995:Table 9.15; Browne and Finkelhor, 1986). Children in difficult circumstances—street children, orphans, refugees—are especially vulnerable to abuse (Rajani and Kudrati, 1996; Shamim and Chowdhury, 1993).6 FEMALE GENITAL MUTILATION Several authors have argued that maintaining virginity until marriage is a main purpose of the practice of female genital mutilation (El-Saadawi, 1982; Dualeh and Fara-Warsame, 1982; van der Kwaak, 1992). Female genital mutilation is the term now used by the World Health Organization (WHO) to cover a spectrum of procedures for which the older term "female circumcision" has often been used.7 The practice has been reported in more than 30 countries on the African continent, 7 in the Middle East, and 4 in Asia and in other areas to which certain ethnic groups from these countries have migrated, including Western Europe, the United Kingdom, and the United States (World Health Organization, 1994). It is estimated by the Institute of Medicine that there are currently 114 million women and girls who have been "circumcised," with 2 million new procedures performed each year (Howson et al., 1996). There are three primary kinds of procedures, with wide variation among and even within groups. The most widely practiced, often called "Sunna" circumcision, from an Arabic word meaning "tradition," consists of the removal of the prepuce or the tip of the clitoris or both. The clitoris is not completely ablated. Clitoridectomy, or excision, consists of 6 In 1996 representatives of 119 governments, and of United Nations agencies and non-governmental organizations, met in Stockholm, Sweden, for a World Congress Against Commercial Sexual Exploitation of Children. They unanimously adopted an Agenda for Action, which called for a broad range of actions, including criminalization of commercial sexual exploitation of children, education and social mobilization to inform both children and their guardians of children's rights, and programs and counseling for victims. 7 WHO defines "female genital mutilation" to encompass clitoridectomy, infibulation, and other related practices, which vary in their severity. In 1993 the 46th World Health Assembly passed resolution WHA46.18 using the term FGM. the removal of the entire clitoris, both prepuce and glans, and may include the removal of the adjacent labia, either minora, majora, or both. Sunna circumcision and excision affect 85 percent of the women who have undergone genital mutilation (World Health Organization, 1994). The most extreme procedure, infibulation, also referred to as pharaonic circumcision, involves the removal of the clitoris, the adjacent labia (minora and majora), and the joining of the sides of the vulva across the vagina, securing them with thorns or with silk or catgut thread. A small opening is left to allow the passage of urine and menstrual blood. The infibulated vagina is forced or cut open to accommodate sexual penetration and childbirth (World Health Organization, 1994). The immediate health consequences of female genital mutilation can include infection, including tetanus and HIV, septicemia, hemorrhage, injuries to adjacent tissues, urinary retention, shock, and death (World Health Organization, 1994; Howson et al., 1996). Antiseptic techniques and anesthesia are generally not used. The sequelae of infibulation are the most serious (Howson et al., 1996). The sequelae of all procedures may be exacerbated by unsanitary conditions in which women live and give birth and lack of access to routine health care, safe surgery, and antibiotics. In some countries, nationally representative samples of women have been asked about their experience of female genital mutilation. Circumcision is nearly universal among Egyptian women (El-Zanaty et al., 1996). In the Sudan, "pharaonic circumcision" is the most prevalent type of female genital mutilation, experienced by three-quarters of all women (Department of Statistics, Sudan, 1991). In the Central African Republic, 43 percent of women aged 15-49 reported that they had been circumcised (Nguelebe, 1995). The long-term effects include loss of sexual sensitivity and sexual frigidity caused by painful intercourse, delayed menarche and cryptomenorrhea or dysmenorrhea, chronic pelvic complications, dysuria, recurrent urinary retention and kidney infection, vaginal stenosis, keloid formation, neuroma, retention cysts, and disfigurement of the external genitalia (Howson et al., 1996). Forcible sexual penetration of an infibulated woman can cause lacerations of the perineum, rectum, and urethra. Obstetric consequences range from sterility due to infection of the uterus and fallopian tubes to exposure of a fetus to infectious diseases, risk of damage to the baby's skull as it passes through the damaged birth canal, and fetal asphyxia or brain damage due to prolonged labor. An infibulated woman must be "opened" to ensure safe delivery of her child, a procedure that poses further risks to the mother and baby (Howson et al., 1996; World Health Organization, 1994).
POLICY AND PROGRAM IMPLICATIONS
Policies and programs to promote healthy sexuality on the dimensions of freedom from violence, the right to refuse unwanted sexual relations, and the ability to seek to express and to enjoy one's sexuality can be divided into three broad types: policies and programs that increase the information and knowledge base needed to promote reproductive health, including the need to collect more information, as well as the need to disseminate such information to those who need and can use it; policies and programs that provide individuals the means to achieve such healthy sexuality; and policies and programs that provide the social, legal, and community support needed to prevent sexual violence as well as to protect and treat the victims of such violence. Culturally appropriate interventions are needed at a variety of levels: social and political structures, norms, communities, families, couples, and individuals. For example, a woman's ability to refuse sexual relations with her husband may depend on the possibility of support from several of these levels. If her refusal occasions violence, she will require support from social and political structures in order to defend herself, or from her community to make clear that violence is not acceptable and that her refusal is valid. She and her husband as a couple need support in understanding her needs and the basis for her refusal. A woman's initial impulse to refuse and state her needs comes from individual characteristics such as sufficient information, self-esteem, and communication skills. The extent to which each of these levels is involved in protecting healthy sexuality is culturally and situation specific. The term "community" may not have real meaning for some people in urban areas. It may be all-important in other situations, in which neighbors and extended families are the guardians of the norms that govern sexuality. Gender also determines the relative importance of intervening at different levels. Men, with more power, have more decision-making capacity. Therefore, interventions concentrating on individual behavior change may be more effective for men than for women, whose behavior is curtailed by family, community, and societal factors. There is some experience in developing countries with programs of each type. Their effects on reproductive health and costs have rarely been evaluated, however, and the evaluations that do exist typically deal with programs in high-income countries. But each of these program and policy types warrants wider implementation and research. Research and Dissemination A high priority for research should be documenting the extent of the problems discussed in this chapter, including violence and sexual coercion, the sexual exploitation of children, and female genital mutilation, and the forms they take in particular societies. Research is also needed on the meaning of female genital mutilation for women and its consequences for health in the communities in which it is practiced. Since open discussion of sexuality is limited in most societies, information does not circulate about beliefs or behaviors. As a result, people may believe that they are the only ones who feel or behave a certain way. They may not discuss their sexual beliefs and behaviors, even with close friends or family, out of the fear that they are not "normal." Comprehensive dissemination of research to the communities where it was carried out is a powerful tool that allows people to publicly acknowledge their true beliefs and behaviors. This is the case in much of Latin America with regard to certain beliefs about sexuality. For example, in Mexico, it was assumed that sexuality education in the schools was a radical idea that would disturb the Catholic majority's sensibilities. As a result, the government was hesitant to implement a comprehensive program of sexuality education. The Mexican Institute for Research on Family and Population (IMIFAP), a nongovernmental organization, commissioned a nationwide survey that demonstrated a consensus among Mexicans that their children's formal education needed to include extensive preparation in personal development, including basic aspects of sexuality. Even more surprisingly, the survey showed that parents wanted this education to begin in elementary school (IMIFAP and Gallup Corporation, 1993). IMIFAP gave these study findings extensive dissemination through press conferences, television and radio interviews, and communication with education authorities. In addition, a film that contained the poll results was shown on national television. In meetings with parents' groups at which the film was shown and the poll results discussed, IMIFAP representatives discovered that parents felt liberated by the fact that it was considered "normal" to want assistance in giving their children information about sexuality. This example illustrates how focused research on opinions and behaviors can, if adequately disseminated, open discussion on previously taboo topics. Research findings can be used effectively to test community assumptions about sexuality and, through accurate information, open discussion that can lead to policy changes. Sexuality Education Programs In many regions sexuality education is already an accepted part of the response to high adolescent pregnancy rates and the threat of HIV infection. Approximately one-half of Latin American adolescents receive some type of sexuality education in school, much of it designed and provided by nongovernmental organizations (International Planned Parenthood Federation, 1995). In addition, governments and nongovernmental organizations in several countries have attempted various strategies to reach young people who are outside the formal educational system. Most programs provide information about reproduction, contraception, and diseases. They may also include psychosocial components related to self-esteem, gender roles, or decision making. Programs that want to address a wider range of issues in healthy sexuality might include components on gender roles, sexual and domestic violence, and rights and obligations within couples, families, and society. Concepts of gender equality would need to be introduced into other areas of education since sexuality education courses will always constitute a very small part of young people's educations. (Evidence on the effect of course duration is equivocal; Kirby, 1994.) There is also a need to reexamine the selection of target populations for sexuality education. Young people are easier to reach than adults because of their participation in the educational system. However, adults are also misinformed and require orientation on issues related to sexuality. Programs can consider reaching adults through their place of employment, their children's schools, or organizations like unions and community kitchens. A Mexican program to train parents to inform their children about sexuality found that participants had many doubts about their own sexuality that they wanted addressed (Pick et al., 1992). The Women and AIDS Program of the International Center for Research on Women is identifying innovative ways to reach adults through STD and AIDS education, including training for traditional women's groups in Africa and holding group informational sessions at STD treatment clinics. Education programs can also have an impact at the family and community levels by training parents to provide sexuality education directly to their children and by involving community leaders in establishing goals for family life and sexuality education programs. Published evaluations of sexuality education pertain mostly to the United States (Kirby, 1994; Kirby et al., 1994; Mauldon and Luker, 1996). A recent review concluded that HIV and sexuality education programs did not promote earlier initiation of sexual relations nor lead to more partners or more frequent sex among young people (Kirby, 1994). Some curricula were associated with delayed initiation of sexual intercourse. Results of evaluations of the effects of HIV and sexuality education programs on the use of condoms and other contraceptives are equivocal. The programs that have consistently appeared successful in meeting their goals have a number of common characteristics, including: a theoretical grounding in social influence or social learning theories, focus on specific behaviors, instruction on social influences and pressures, and activities to practice specific skills and to increase young people's confidence in their skills. We have not found evaluations dealing specifically with the influences of sexuality education on attitudes and behavior related to violence and sexual coercion. Health Services The concepts behind healthy sexuality need to be integrated into reproductive health service provision. Service providers have access to clients' attention and trust and should receive training that allows them to move beyond providing contraceptive services to explore the factors that might influence clients' misuse of contraceptives or exposure to STDs. In male family planning clinics in Brazil and Colombia, the staff found that men will participate if clinics offer a wide range of services not directly related to family planning, such as sexuality counseling, sexual dysfunction treatment, and infertility treatment (Rogow, 1990). Two Brazilian experiments in incorporating women's concerns about sexuality into reproductive health care demonstrated that this type of model attracts clients and removes obstacles to effective contraceptive use (Bruce, 1990). A more comprehensive approach that integrates concerns about sexuality with reproductive health needs could help women achieve healthy sexuality and improve reproductive health markers, such as consistent contraceptive use (Dixon-Mueller, 1993). (See the description of an information and counseling program at a family planning clinic in Brazil, in Diaz, 1996.) Access to and enjoyment of sexuality in both men and women is conditioned by their physical reproductive health. Reproductive tract infections, impotence, and the strain of repeated childbearing greatly limit the expression of sexuality. Health and family planning services need to take account of the concerns of their clients on these issues, but they must do so in a culturally sensitive way. Only then will women be able to shed the "culture of silence" (Khattab, 1992) in which they treat reproductive tract problems as part of a normal existence or else are too embarrassed to seek help for such problems even if they are perceived to be abnormal. Health services need to play a stronger role in detecting domestic violence and sexual abuse of children, counseling and treating victims, and interventions. The health system may be the only public service with which victims come into contact (Belsey, 1996). Information and Communication In many different parts of the world, media campaigns have been successfully used to spread information about AIDS, as well as to increase awareness of family planning, which may lead to positive behavior changes. Television and radio are widely accessible and influential, particularly among young people. The commercial media already manipulate sexual images in many ways. It is therefore critical to analyze the media's role in formulating norms and attitudes related to sexuality and to use mass media to promote images of healthy sexuality. Extensive research would be necessary to determine how the mass media could best promote healthy sexuality. Campaigns could focus on such previously neglected issues as domestic violence and, where relevant, female genital mutilation. Collaboration with entertainers who are interested in promoting healthy messages about sexuality should be an important part of interventions directed at young people. All interventions have to be designed with respect for religious beliefs and local culture. This is particularly important in designing reproductive health strategies in places where female genital mutilation is common; work should include research on its meaning to local women and measures to educate the public and formal and informal health care providers about its harmful effects on women's health. Some countries have implemented education programs and media campaigns to discourage female genital mutilation. These programs may be directed at the general population, specifically at mothers of young daughters, or to medical practitioners, including midwives who traditionally perform the procedure (Kiragu, 1995). Increased Educational, Economic, and Organizational Opportunities for Women A principal barrier to women's control over their sexuality is their dependence on men (or on their children) for economic survival. If women have no alternative, they will enter into unions in which they exchange sexual availability for support on their partners' terms. They will have few options for escaping such unions if their partner is abusive or unfaithful. Therefore, one key to women's achieving control over their sexuality is more economic participation and control in their households and in the larger community. In addition, it has to be socially acceptable both for women to delay marriage and to leave abusive unions. A comprehensive reproductive health strategy should include development efforts that assist women in gaining access to education and finding employment. This type of intervention will need to work through social and political structures and societal norms that restrict or facilitate women's economic autonomy. Legal and Policy Changes Legal and policy changes to promote healthy sexuality can include reform of laws related to sexual and domestic violence, policies that include sexuality education in public schools, reform of family laws to increase the rights of women to property and inheritance, and enforcement of laws against cultural practices harmful to reproductive health (such as female genital mutilation). Legal changes promoting healthy sexuality can help foster changes in public attitudes, and changing attitudes in turn make enforcement of the laws more likely and more effective. Governments around the world have already taken some important steps toward policies favorable to healthy sexuality. For example, Mexico, Colombia, and Chile passed laws in 1994 making sexuality education obligatory. In India, a recent law stipulates that if a woman dies within 7 years of marriage from "unnatural" causes (including accidents and suicide), the police are obliged to investigate the possibility of a dowry- or marriage-related death. The police departments in many parts of the country also have special units to deal with complaints about domestic violence, although there has been much criticism about how they really operate. In Brazil, there are now 200 police stations staffed entirely by women to deal with domestic violence (Ford Foundation, 1992). Female genital mutilation is now outlawed in some African countries (Kenya, Senegal, and the Central African Republic). However, even where it is outlawed, the practice may continue in secrecy, with those suffering complications inhibited from seeking help (World Health Organization, 1994). In 1994 the United Nations 47th World Health Assembly adopted a resolution to encourage all countries to "establish national policies and programs that will effectively, and with legal instruments, abolish female genital mutilation. … and other harmful practices affecting the health of women and children." Nongovernmental organizations have contributed in vital ways to improve health services and health policies and laws in most parts of the world. In Africa, a network of organizations is working to change public attitudes about female genital mutilation and to encourage government leaders to speak out against the practice and to enforce laws against it where they exist. Other organizations have focused on the issue of violence against women, providing counseling and support and practical help to victims and working to change public attitudes and law enforcement. In Zimbabwe, for example, a group called Musasa has conducted training sessions for the police and other government agencies (Stewart, 1996). There is a need for improved communication both among nongovernmental organizations and with governments on issues affecting women's health and sexuality. Such dialogue will enable them to be more effective agents for policy change.