The Epidemiological Transition: Policy and Planning Implications for Developing Countries (1993)

John C.Caldwell and Pat Caldwell

The informal sector plays a vital role in the provision of health services. Families, individuals, and societies all have rules that govern the type of treatment an individual receives for a given illness. As societies modernize, health usually improves owing to greater availability of health services and to changes in attitudes and norms pertaining to women’s behavior and the value of life. In this paper we examine aspects of society and of behavior that encourage or discourage health, concentrating on the areas we know best, South Asia, sub-Saharan Africa, and to a lesser extent, the Middle East. Inevitably, the main measurement of ill-health is mortality because perceptions of illness vary across cultures and limited access to health services impedes gathering data on morbidity. Much of this paper focuses on child deaths, partly because they still form the majority of mortality in the poorer Third World societies and partly because we can locate the living carers for most dead children in contrast to the situation in the more difficult area of self-care that characterizes much of adult mortality.

The central argument of this paper is that the persons with the greatest interest in children’s health and survival, and with the greatest willingness to devote time to their protection and to care for them in sickness, are children’s mothers. Children may receive less than optimal attention both in health and in sickness because their mothers are prevented from giving

them the needed attention, lack sufficient resources from the larger family or their husbands, or lack self-confidence about their ability to care and make health decisions. Many of these elements still exist and restrict the rate of health improvement not only for the children but for their mothers as well. Moreover, these restrictions on women probably also jeopardize their husbands’ health and survival chances.

This restriction on women limits the resources available to children and younger women so that men and the elderly receive what is regarded as their rightful share. However, the main reason for controlling women is to ensure the operation of the male-dominated, ancestor-oriented, patriarchal, larger family (not necessarily defined by a common residence). Only men or the elderly should confidently make decisions. A young woman should not divert her husband’s primary obligations, duties, and affections away from his parents and brothers toward herself and her children. It is often the task of the mothers-in-law to see that this diversion does not occur. Such a structure once facilitated largely subsistence agricultural production but has retained a justification of its own as economies change.

In much of the Third World the most rapid gains against mortality can be made by giving women, especially young wives and mothers, greater confidence, more decision-making power, and greater access to resources for care and treatment. Women can be empowered as caregivers by education and by female home health visitors. They can also be empowered by social scientific “health transition” research which, as the findings become part of public knowledge, convinces men that women’s powerlessness is endangering their children, even their sons. The same end is achieved through gains made by the women’s movement in claiming greater gender equality as a question of social justice.

Women are helped in their caring mission by access to resources of such types as free advice and assistance from health visitors, and easy access to free or inexpensive health services. If these services are not free or of minimal cost, then there must be some way of ensuring that women can obtain immediate treatment for themselves or their children without having to wait for family budgetary decisions. In a fee-for-service system, women may have no way of obtaining quick access to the money necessary to pay those fees and may feel unable to take actions that would have to be covered by payment.

Women are often so immobile that the services must be close at hand. Because the traditional medical system has the loyalty of many people and is more often seen as offering complementary rather than adversarial services, the aim should usually be not confrontation with that system but rather an ever greater attempt to steer patients toward the modern system when it has the best treatment to offer, especially if the illnesses are life-threatening.

Populations of less developed countries need to consider more seriously the change, occurring in adult and old-age mortality. Males’ health is threatened by the structure of the patriarchy and by the emphasis on masculine virtues, and in this area too, gains are to be made from the feminization of care, treatment decisions, and resource control. Here, too, the findings of health transition research need to become part of the public consciousness.

All traditional societies existing before the massive imports of culture and technology that have increasingly characterized the present century have very different beliefs about the nature of illness and its causes than those espoused by modern biomedical science. At their simplest, they merely put forward an alternative explanation for each individual type of sickness. In the absence of any access to modern health services these explanations were not necessarily harmful, although in both India and Nigeria (Cantrelle and Locoh, 1990) the cure involved contacting more persons than the sufferers tended normally to do. The real danger of traditional beliefs about disease causation is that when modern health services do become available, the sick or their parents may fail to use the services or may delay using them.

These dangers are not always as great as they seem, but they do have policy implications. In both Asia and Africa, most people regard the alternative systems not as being in conflict but rather as two different but valid roads to recovery. They are willing to alternate between the two systems, and frequently employ modern health services for life-threatening complaints and traditional services for those endemic conditions for which modern medicine has little to offer. In Asia, traditional healers take a rather similar view and most Indian Ayurvedic practitioners dispense modern medicine as well as prescribing ancient herbal cures. However, they may store medicines incorrectly, prescribe inappropriate dosage levels, and misdiagnose the condition for which medicines should be given. They go along with modern practice only as far as providing medicines, but not in other behavioral advice. In Africa, traditional doctors have rarely moved into this “halfway house,” and their patients rarely receive whatever benefits modern medicine may confer. In the 1973 Changing African Family Project study of Ibadan, Nigeria, the children of traditional practitioners had higher mortality than almost any other occupational group (Caldwell, 1977).

Perhaps the greatest problem about traditional health beliefs is that they readily divert behavior into incautious patterns. In Africa there is a strong belief that the person of physical and psychological vigor cannot easily succumb to serious illness. This belief has resulted in continued high-risk behavior in such areas as sexuality, contributing to the spread of sexually transmitted diseases and AIDS (Caldwell et al., 1991). In both Africa and

Asia, people still believe that the future—or at least the timing of death—is predetermined, which may encourage incautious health behavior and weaken the belief in personal responsibility. Much sickness is regarded as a punishment for sins in this world or beyond. In sub-Saharan Africa, for example, there is often little attempt to obtain assistance during difficult childbirths because it is believed that the women can overcome the problem by admitting the adultery that caused it.

Another serious problem with regard to traditional medicine is the belief that the healer identifies the sickness at sight and prescribes the best treatment. If it does not work, then nothing will. In rural south India this practice means that illiterate women bringing sick children even to a modern physician do not overly exert themselves to explain the history of the sickness and, if the child fails to recover, frequently do not return for further treatment because they believe that the best available treatment has already been given.

On the other hand, not all traditional health precepts are in conflict with modern medicine. The historic Sri Lankan sensitivity to illness and the belief that doing nothing about it was inexcusable have fitted in well with the modern health system (Caldwell et al., 1989).

The policy implication in all but the most modern Third World societies is to avoid confrontation between traditional and modern providers. The public should be educated that the modern health system is the place of first resort for serious, sudden, or life-threatening sickness, without scorning the solace that the traditional system can provide for sufferers of rheumatism or even inoperable cancer. Where there is a well-established practice of traditional healers dispensing modern medicine, as in India, it is probably best to attempt to guide rather than prohibit this activity, given the paucity of modern doctors. The lack of confrontation is foremost a public reaction to the kind of health that is wanted and can be afforded. Politicians have moved slowly, and the modern medical profession more slowly still, in working out some kind of arrangement, but a nonconfrontational compromise is probably in the best interests of public health.

In rural south India (Caldwell et al., 1988) and much of South Asia, the mother of young children is usually still a young woman and is likely to be living with her husband’s family. She has been chosen by his parents primarily as a female member of the younger generation of the household rather than as his wife. In south India she may be living far away from her own family, and in north India this is almost inevitable. Her work and her life are controlled by her mother-in-law. The marriage must not undermine

the structure and solidarity of the larger family, and the young wife must not seem to be emotionally too close to her husband, to influence him too much, or to seek his advice too frequently. She must not try to crystallize her nuclear family from the larger family. Her children are the children of the household, and her parents-in-law have direct control over them as their grandchildren.

This family structure influences health behavior decisions. The young mother keeps the closest watch over her babies and young children because they improve her status in the family. She is usually the first to notice that her child is ill, but in few cases does she take physical action and rarely does she even tell others; rather, she waits for her mother-in-law or her husband to notice (Caldwell et al., 1988). Her position is weakened further in the lower castes by working for long hours in the fields as an agricultural laborer while her mother-in-law cares for the child and regards herself as the child’s main caregiver. The older woman would be insulted if the younger woman claimed that the child was sick and would regard it as a slur on her child-caring abilities. Consequently, taking action about sickness is delayed and the health decisions that are made are often strongly influenced by an older, uneducated woman. In parts of north India the patriarchy is so strong that most women, even the older ones, and especially the illiterates, feel little confidence in their abilities to make decisions and often take no action (Khan et al., 1989).

Education or the presence of educated women may modify these traditional structures. In south India, the older women usually regards an educated daughter-in-law as not only having rights to make health decisions, but as being more likely to make correct ones. Indeed, rural families now state that the second greatest advantage of an educated daughter-in-law is being more likely to ensure the safety and survival of the grandchildren. The young mother’s independence is reinforced by the presence of government-sponsored family-planning workers who have caused the older women to retreat. In Sri Lanka, a system of female health visitors, who develop a relationship with the young women, has strengthened the decision-making ability of young mothers. Research in Sri Lanka indicates that half of the health decisions are made by the mother without consultation; most health decisions have been made this way in Kerala since at least the last century (Sushama, 1990, citing Mateer, 1883).

African women experience situations ranging from those not very different from India—among the Luo of East Africa—to those of West Africa where women have great independence but also great responsibilities that are not provided for by a joint family budget. Whyte and Kariuki (1991) indicate that among the Luo, although the mother is the most immediate caretaker of children and is usually held responsible when the children have accidents or suffer from sickness, she is constrained by the attitudes and

traditional practices of men and family members. What is eaten, for example, and even the order in which people eat, are in many ways the most central aspects of culture. Women alone change what and how the family eats or even what they and their children eat. Yet nutritional education programs often direct themselves only at the women and avoid coming into conflict with the men and older women who make the actual decisions.

On the other hand, in West and Middle Africa, women are more independent, partly because of the prime social role of descent lineage and partly because of the existence of very high levels of polygyny, which renders each wife and her children a separate economic unit. In many settings, if a woman can find the resources, she can usually treat herself and her children as she sees fit, but the treatment may be more costly than she can afford from her farming or marketing income. In this case, she may have to seek money from her husband or the child’s father. Involving others in the decision-making results in delays and in yielding power about treatment to the person who pays. Bledsoe (1990) indicates that the West African position to be more complex when a mother has to find the resources to pay for her child’s treatment if her husband is not the father; she is also unlikely to try as hard if she is not the child’s biological mother.

In South Asia, as in much of the Middle East, North Africa, and East Asia, society constrains a mother by placing more emphasis on the survival of sons than daughters. This emphasis is most dangerous among children ages 1–9, who are no longer receiving the equal nourishment and antibody protection provided by breastfeeding and are not yet old enough to look after themselves. Muhuri and Preston (1991) have demonstrated lower mortality in Bangladesh among the first two sons and the first daughter, showing that a care component exists that causes differential mortality. However, it is not clear to what degree attention or neglect affects mortality, or whether the mechanisms occur through feeding, maintaining good health in other ways, or securing timely curative treatment. In sub-Saharan Africa, where there is no excess female child mortality, girls are probably more protected by a bridewealth system than would be the case under a dowry system, although it is still in many ways a man’s society. Perhaps the main factor protecting little girls in Africa is that children are very largely brought up by their mothers, who both take the responsibility and make the decisions.

What kind of care can be provided in a traditional society by a mother, and what is its likely effect? The kinds of effective care that can be offered by a mother in a traditional society include protection from accidents, protection from infectious disease, and treating a sick child in a way that will maximize the chance of recovery by allowing the child rest. They can also include elements of foresight about providing warm clothes or bedding and protection from the weather in other ways (cf. LeVine and Dixon, 1990). The provision of adequate feeding can often be achieved by the transfer of

an almost insignificant amount of food from adults to the very young. The All India Nutrition Institute decided that the prime reason for the relatively high levels of malnutrition in traditional households where neither parent was literate was that they really did not think that such tiny persons needed very much food. They were also more likely to regard breast milk as a free gift not needing supplementing even when the baby was a year old.

There is another important dimension to child care in traditional society related to absolute and relative commitment to child survival, which has been well captured in a paper by Simons (1989). In most traditional societies, mothers are not as careful about their child’s survival as they might be. They do not take as much responsibility as they might, partly because the family system does not dictate that the final responsibility is theirs and partly because they feel powerless to change the inevitable course of events. Child minding is usually very far from being the exclusive province of mothers, and in many small villages, children tend to run around together, with the community as a whole keeping an eye on them. Galel el Din (1977a; 1977b; personal communication, 1977) has described how mothers in Sudan who took such a system for granted subsequently changed as they were educated or migrated to towns and began to protect their children by intervening in the crowd and often withdrawing them into the home.

Community factors can also reduce the care a woman gives a child. In many Muslim societies, seclusion means that a woman cannot act easily to obtain treatment for herself or her children. Seclusion may be why educated women in the Middle East and North Africa do not have as low child mortality as their incomes and education would suggest; their problem is usually compounded because they do not have the education that such income levels in other parts of the world would ensure. In Bangladesh a young mother cannot take a sick child a few hundred yards across an open field to a dispensary until her husband or mother-in-law is available to accompany her. She cannot go herself to a hospital unless a curtained bicycle rickshaw can be obtained. Seclusion may explain why even in Matlab, with a hospital and four additional health centers, only 11 percent of women who die in childbirth are seen by a physician before death (Fauveau et al., 1989).

Traditional societies begin a phase of transition when they become more closely linked with the global economy and society; this phase is characterized by the growing importance of education and market economies. The traditional society passes when schools and health centers appear. Urbanization, too, and even the potential ability to migrate to a town also weaken the old system. Although family and social structure changes during these

transitions, the changes are often not sufficiently fast to make full use of the new health services.

The aim of this analysis has been to show how social and behavioral change can be directed to improving health. The optimum path in most Third World societies is dominated by the need to modify the nature of the traditional family and the position of women in it, which is not to imply that health is improved by excision from the family. Such would be least likely to be true precisely in those societies where the family is all-embracing, and the solitary widow without surviving sons in India is subject to an appallingly high-mortality risk. Nevertheless, the traditional family does provide support and even health care, although these are subject to its own priorities toward the old and those of working age, and usually toward boys. Family priorities in South Asia seem to favor most the main contributors to its survival, especially working adult males, but in sub-Saharan Africa, priority is often given to old men, who will soon become powerful ancestors. To secure its physical and competitive survival, as well as give men what they feel they deserve, the traditional family is, in most societies, strongly masculine in attitudes and priorities. The evidence from transitional societies seems to show that the faster it is feminized (usually to only a modest extent), the more rapidly does mortality fall, even in the case of males. The process has often been most rapid where women’s status has been highest in traditional or early transitional society, as in matrilineal Kerala or in Sri Lanka after the Buddhist Revival Movement of the late nineteenth century.

The role of the community is even more ambiguous. The community also can be patriarchal and oligarchic, reinforcing masculine and assertive values that are essentially anticare and, in transitional societies, monopolizing the new health care systems for the use of the powerful. In some transitional societies such as Kerala, community pressure can make the new health systems work efficiently, but these new community pressure groups appear to be an outgrowth of grass-roots, radical, and democratic tendencies that have flourished because of the weakness of older structures. They have been reinforced by modern political organization, which they have brought into existence, rather than being based on older community structures.

In recent years, research has demonstrated that child mortality declines steeply as the mother’s education increases and that perhaps half of this effect remains when economic and other factors are controlled. This finding has elucidated the role of care in both transitional and traditional societies. We believe that it works partly through strengthening the individual internal locus of control and the commitment to child survival, making mothers feel personally responsible for what happens to their children and less likely to believe that it is inevitable (Caldwell et al., 1990). Girls who go to school identify themselves with other aspects of modernization ranging from nationalism to health centers, and modern medicine. They believe

that they are the kinds of persons who can and should go to see doctors. Lindenbaum et al. (1989) examined mortality in rural Bangladesh and concluded that much of the lower-risk behavior of girls who had been to school arose not so much from the ideas they had absorbed but from the fact that they had learned to behave as the educated should—to dress themselves and their children cleanly, to wash their hands, and so on. As schooling continues, however, it probably has additional influences. Lessons are taught about health and hygiene—the need to boil water and to wash one’s hands—as well as about modes of infection. Although most of this knowledge is probably lost, especially in nonreceptive family circumstances, some still remains. Probably the more important reason is the assault on nonscientific and nonrational thinking (Goody, 1971; Caldwell, 1982). Although students do not necessarily disown all traditional beliefs, they are taught to believe in the efficacy of modern science and may have two parallel interpretations of the world. They are much more likely to have some faith in modern medicine and to resort to it than are illiterate women. Although education and sensitivity to ill-health are not enough to produce very low levels of mortality, they tend to intensify effective maternal care, as Orubuloye and Caldwell (1975) showed. The big advance in reducing child mortality in Sri Lanka occurred when free modern health services became accessible. Cleland and van Ginneken (1989) found that much of the effect of maternal education on child survival depended on interaction with health services. They also established that there was a very considerable unexplained residual, which must largely be a measure of domestic care or health maintenance, that leads children not to become sick or have accidents in the first place.

A difficult challenge found more commonly in transitional rather than traditional societies is the rise of faith-healing Christian churches. In Nigeria, it is estimated that their membership may be as large as 25 million—one half of the Christian population and one-fourth of the national total. The growth of these churches may be related to charges for modern drugs, but they also affect behavior where there is no cost (e.g., the substitution of praying over water for boiling it). A recent study of a large town showed that modern maternity facilities were almost unused while babies were delivered in faith-healing church clinics without the benefit of sterilized instruments, boiled water, or any medicines (Adetunji, 1992). Here again, the best policy may be to avoid conflict and attempt to convince the churches that demarcation of services needs to exist.

There appears to be little doubt that when comprehensive health services are spread through a society over a period too short for social change or substantial alteration of educational levels, major mortality decline has occurred (Halstead et al., 1985; Caldwell, 1989). Two provisos should be made: (1) the successful health programs were essentially democratic and

(2) the successful services were often established in countries that were poor but had high levels of female education.

The community and its political structure affect not only a woman’s capacity to visit the health service but also the way that service works. The extent to which doctors listen to less educated or poorer women is very much the product of social attitudes. If the doctors are also drawn from a group that is economically better-off or more powerful, they may give less attention to the poor and may work neither long hours nor conscientiously. Mosley (1989) has provided examples of studies of Asian societies, where social stratification and patriarchal attitudes rendered health services ineffective for most people. Even in more egalitarian West Africa, Okediji (1975) revealed the extent to which Nigeria’s middle class monopolized the free government hospital services in the city of Ibadan.

The other side of the picture is the extent to which community attitudes and pressures make the health system work in Kerala. Popular emotions can run high and politicians quickly enter the fray. The pressures on health staff can be unfair, but they are compelled to work long and efficiently and to give due consideration to the poor. They see that drugs are ordered on time and make themselves available for emergencies outside normal hours, maximizing the access of women and children. But difficulties may arise in implementing these factors elsewhere because having adequate local control is really an agenda for political and social change extending far beyond the health service.

Democracy at the national level can ensure greater expenditure on high-quality, less expensive, and more accessible educational and health services, and greater equality for women. For this to happen, the government must have a radical and egalitarian element; otherwise, the chief thrust may be for lower taxes and the least governmental intervention. Left-wing revolutionary governments, especially in their early populist phase, can do the same. There are some interesting anomalies. Infant and child mortality decline appeared to falter under the early Pinochet regime in Chile but subsequently fell faster than ever. The explanation may have been partly the improvement of an economy that had found itself in severe difficulties, and partly the fact that external assistance and advice by donor governments determined that the experiment under way should succeed socially as well as economically. There is little doubt that in Indonesia popular pressures appear to be generated at the community level, which make health and other services work, yet in ultimate origin are autocratic rather than democratic.

Communities can influence health services in a number of ways. They can demand modern health service provision, usually more successfully in a representational democratic system. The new health system can be identified with their community, encouraging people to use it. This identification may be most successful if the community plays some role in the health

center’s administration and in the selection and employment of its staff. Perhaps more important is community pressure and spontaneous anger to ensure that the staff serve the people efficiently and without discrimination by wealth, gender, or education. Communities normally do not interfere in household care, although programs for weighing children and periodically examining their health can result in the community taking an interest in individual families and indirectly—or directly—exerting pressure.

There is evidence of education catalyzing a range of caring behavior in developing countries. We shall start with the interaction with the health system, if only because it is easier to measure, and outline our findings in rural South India (Caldwell et al., 1988), which are probably broadly similar to the situation in many countries. The young mother who had been to school probably crystallized in her own mind that her baby was sick some-what earlier than did the illiterate mother and demanded that something be done. She was more likely to be allowed to do something herself because illiterate mothers-in-law concede that educated daughters-in-law understand what is of value in the new world. If the young mother was a Muslim this concession was more likely to be made over the doubts of local religious leaders about the necessity for her taking such public action. She was less likely to accept traditional explanations for the sickness but might rely on them if the doctor’s treatment failed. The time she spent explaining the problem to the doctor was proportional to her education. The illiterate woman may fail to communicate for three reasons: (1) she thought that the doctor already understood what was wrong; (2) she lacked the words and concepts to make an easy explanation; and (3) the doctor shortened the time with the mother because he did not expect to understand her. This can be disastrous in view of Christakis and Kleinman’s (1989) estimate that at least 80 percent of the correctness of diagnoses in rural developing areas can be attributed to the histories of the complaints as presented by the parents or patients. The more educated mother is also told at greater length by the doctor what should be done. She is more likely to buy the drugs, because she may be better-off and has the determination to secure the prescribed drugs, and she is more likely to use them properly. Also, the more education a mother had, the sooner she was likely to report back that the treatment was not working. The failure of illiterates to do so arose partly from the belief, based on the traditional medical system (which contained no probabilistic element with regard to diagnosis), that the disease had certainly been identified and the best treatment already given, and partly because they did not believe that they belonged to the same world as the doctor and, in their own words, could not tell him that he had been wrong. A number of policy interventions that might improve the situation are, clearly, encouraging the original determination that something should be done, attendance at a modern health facility, interaction with the doctor, obtaining

the prescribed drugs, and reporting to the system that recovery has not taken place.

The maternal care that does not involve attendance at a health facility is much harder to measure. It has most often been assessed through specific behaviors. Bhuiya and colleagues (1990) showed that in rural Bangladesh, educated mothers were more likely to know that boiled water was safer to drink and were more likely to wash their hands after they had defecated or their children had done so. Owing to cultural proscriptions and practical difficulties, however, they were no more likely to confine the family source of water to the tube well. Chen and colleagues (1989) showed that for the Matlab area of Bangladesh, daughters received less food than sons and suffer from greater malnutrition based on standardized tests. They were also less likely to be taken to free clinics when suffering from diarrhea, which may have had a greater effect on mortality than differential feeding. In rural South India we found that educated mothers restricted other expenditures when food was in short supply, concentrating on adequate nutrition for the whole family. Our observations showed them making less distinction than uneducated mothers between old and young, or male and female, although, as Cleland and van Ginneken (1989) point out, the latter is not the usual finding in South Asia.

What follows from the above discussion with regard to policy interventions that may help women as carers both in health maintenance and in curative interaction with modern health services?

The diffusion of social science research almost certainly has a central role to play. Public debate, fueled by social science research, contributed to world fertility decline by impinging on both governments and individuals (Caldwell and Caldwell, 1986). It is similarly inevitable that sufficient debate about the cultural, social, and behavioral determinants of health will influence governments at every level. The most patriarchal government or father is disturbed to find that some boys are dying unnecessarily because their mothers have too little education or too little autonomy, both in general and specifically with regard to health decisions and behavior. Clearly, the women’s movement also has an important role to play in this area, both in securing women sufficient autonomy to make them more effective carers and in showing that some daughters are dying unnecessarily.

Lower fertility has a role to play in a sense that can be demonstrated by history rather than by comparing the fate of children in large and small families. Dramatic declines occurred in Western infant and child mortality in the first decades of the present century, following rather than preceding the movement toward the small family. Undoubtedly, a substantial compo-

nent of this decline was behavioral, although to a considerable extent this was probably also due to new biomedical knowledge and techniques. As the family became smaller, more care and resources were diverted to children and there was greater determination that all should survive. This change occurred not only in the family but also at the level of governments and national organizations. In Australia and New Zealand, and presumably elsewhere, the emotional driving force of the infant welfare movement was the national responsibility for ensuring that all the children of these unexpectedly smaller cohorts should survive to support national growth and security. The responsibility for this care was laid even more firmly on the mother. Although this shift saddled her with more responsibility, it also backed her decision-making powers by a social consensus. It might be noted that the first deliberately contrived small families in Ibadan, Nigeria, experienced extraordinarily low mortality (Caldwell and Caldwell, 1978), and it appears that this differential has also held true in China’s one-child families (Caldwell and Srinivasan, 1984). The children of sterilized mothers in rural South India have abnormally low mortality, partly because of family planning workers’ concerns for child survival (Caldwell et al., 1988).

Perhaps the most promising policy interventions on the social and behavioral side for reducing infant and child mortality are to provide mothers, especially young mothers, with more self-confidence and more decision-making power. These two matters are not identical but are usually interrelated. Part of the answer is continuing public debate and media attention to these issues and their consequences for child health. In circumstances where fees for services or transport costs are involved, women need automatic access to the family budget or some sort of funds. The trend toward charging for health services jeopardizes children’s health not only because it squeezes the family budget but also because it dramatically reduces the capacity of mothers to take quick action because they have the least access to family budgets.

One way to strengthen the position of wives and young women in the family relative to old women is the establishment of women’s groups. These can be specifically related to child health and linked with health services. They can help ensure that all women and children receive a fair deal from both their families and the health system. Such groups existed in Australia earlier in this century, but their sole purpose seems to have been putting pressure on inadequate mothers.

An effective intervention in Sri Lanka and Botswana has been female health visitors. The visitors are usually educated young women from the specific area in which they are going to work. This experience and similar experiments elsewhere have shown that the visitors should be trained adequately by some kind of health authority and infused with some expertise, self-confidence, and realization of what their role is going to be. If pos-

sible, they should be given the backing and moral support of both the local community organizations and the local health facility. They should concentrate on identifying pregnant women and giving them support through pregnancy, birth, and the infancy of the child. The system succeeds not merely by providing advice but by being active within the family. The visitor suggests decisions and offers the young mother moral support in carrying them out. The health worker may also suggest when prenatal checkups are due, may make appointments for them, and may help not only with the logistics of transport but with seeing that the woman can get away from the home. She may also do all these things with regard to the birth itself and subsequent treatment of infant sickness or securing immunization. One of the advantages of institutional birth-delivery is that for a critically short period an even stronger organization takes the decision-making power out of family hands. The health visitor may even serve as a role model for the girls and young women of the area.

The health visitor will find it hardest to be effective in matters concerning family or culture. Such areas may include the family allocation of food, the work allocation of children, and the ability to rest when sick. At their most difficult, these decisions may involve the allocation of resources between the old and the young, the extent to which access to medical treatment is permitted, or whether continuation of girls’ schooling beyond menarche should have priority over female modesty or seclusion. Community leadership and community meetings may be the mechanism for inducing such changes. In Indonesia this approach has been effective, although the emphasis is on securing community consensus on an issue rather than encouraging initiatives originating from the meetings themselves.

A different range of interventions is available to make the health system work better for mothers and children. The most important issues are the provision of adequate services for the poor, illiterate, or inarticulate; the ensuring of alternative treatments if the initial treatment is unsuccessful; and a devotion to duty that encompasses the constant availability of providers and medicines. Better training in medical and nursing schools, a better atmosphere in the health bureaucracy, an adequate surveillance system, and routine checks on patients who have not reported back—all can improve health status and conditions. Although the Kerala experience cannot be reproduced, it is clear that many of these objectives will be fully achieved at the local level, and even at the governmental and medical school levels, only if community pressures can be exerted. Local representative bodies need a voice in controlling health services.

If cost recovery is going to become a significant element in Third World health systems, there will have to be ways for ensuring that service is immediately available for sick children even if the cost is later recouped from their fathers. This difficult matter, involving both the collection of

money and the guarantee that neither the mother nor the children will suffer because of incurring such debts, is probably something that community organizations do best.

For much of the Third World we know far less about the structure of adult mortality, especially deaths outside hospitals, than we know about child mortality. Cardiovascular disease and illnesses linked to diabetes may be far more important than we thought; so may accidents and other violent deaths. There are regions such as mainland South Asia where mortality in late middle age or older, especially among men, appears to be almost inexplicably high, given the changes that have been achieved in younger mortality levels. Low levels of education and relatively low abilities for self-protection may be part of the explanation. The failure of adults to look after themselves may have causes very similar to those of the failure to care adequately for the young.

Among younger women, one obvious area for intervention is childbirth. Research at Matlab suggests that the entire excess female mortality in the reproductive-age range, which is observed throughout South Asia, can be explained by causes of death related to reproduction (Chen et al., 1974). The solution seems to be institutionalized childbirth, which overcomes the problem both of traditional birth attendants and of family unwillingness to allow modern care when difficulties with delivery arise. The fact that this female excess mortality is explained by maternal mortality does not mean that no gender discrimination is involved, for if men were subject to such risks, the resources for creating the necessary institutions would probably have been found long ago. The Matlab research shows that such mortality can also be reduced by more family planning so as to limit the number of births per woman and possibly prevent the premarital or extramarital conceptions that so often lead to fatal illegal abortions or domestic violence.

Little attention has been given to the fact that although the patriarchy diverts resources toward older men, it probably also places them in considerable danger. The emphasis on masculinity can mean a macho disregard for illness and its treatment. Our research in South India, an area characterized by a high excess of male mortality at older ages, showed only half as many adult men coming to the health center for treatment as adult women. It is probable, too, that in harsh, poor rural conditions, women, who are past reproductive age, provided they do not become widows without surviving sons, are subject to less physical adversity than old men. Farming, at this age, can be both strenuous and involve exposure to the rigors of the climate.

Finally, little attention has been given to self-inflicted damage from

deleterious behavior in the Third World, partly because it is felt that there should be some solace for a harsh life and, in any case, death will probably come early. This attitude may lead to an underestimation of the level of deaths attributable to drinking (often impure and dangerous forms of alcohol), smoking, betel chewing, addictive drugs, and violence. The issue of high-risk behavior, however, now has to be faced in sub-Saharan Africa and elsewhere because of the occurrence of AIDS.

Many Third World countries are now entering an era in which health will no longer be dominated by infectious disease and early mortality. Indeed, much of Latin America, East Asia, and Southeast Asia are far along this path. It is a common experience for researchers in the poorer regions of the world, especially in rural areas of mainland South Asia and sub-Saharan Africa, to find that most people believe that mortality has risen over the past few generations. This belief is not prevalent in the more developed and better educated societies furthest along in the epidemiologic transition, if only because their educational levels and the better national data collection systems make their populations “developed” in depending on the media and on statistical services to determine their viewpoints.

There is little problem with the recognition that a need exists in these epidemiologically advanced societies to move from a health system dominated by protecting the young from infectious disease to treating the old for degenerative disease. Third World countries have been so dominated by models from the West, by the demands of their own urban elite, and by the views of the medical establishment in their capital cities that there has always been enormous pressure to accelerate such shifts. It has been expert reports, often written by teams with external membership, and the advice of such international bodies as World Health Organization (WHO), UNICEF, and the World Bank that have kept this tendency partly in check. The real challenges go behond this shift in the provision of services.

The first is to persuade the more epidemiologically advanced countries that a broader approach is needed to combat the degenerative diseases. Lifestyle changes, as well as medical provision, are necessary. This message may not be difficult to impart, given the extent to which these populations now read material originating with the Western media. The more difficult task will be to persuade market-oriented, development-focused governments and international organizations that the whole burden of the new old-age sickness should not fall on the women in Third World families and that the Western model of community assistance with the provision of subsidized hospitals, nursing homes, and old-age accommodations should, in forms suitably modified for the culture, be followed.

The second is to persuade the international health community that even in countries with life expectancies still in the range of 55–60 years of age, much of the battle against infectious disease has now been won and a grow-

ing proportion of deaths in the population over 40 years of age are now avoidable or deferable. This change has come about for two reasons. One is that the onslaught of infectious disease is being defeated for a range of reasons including the UNICEF insistence on worldwide immunization. The other is the change in age structure as fertility continues to fall in most of Asia and Latin America and begins to fall in Africa. Our ignorance of this mortality change in countries with incomplete death registration is partly the product of deficient demographic techniques for measuring adult mortality, let alone trends and rates of change, and partly a woeful lack of cause-of-death data except from the minority who die in hospitals.

The best health agenda for the Third World is reasonably clear: more investment in education, especially for females; more female empowerment, by direct intervention and through “health transition” research, which brings the social causes of unnecessarily high mortality into the public domain; more accessible and democratic health services identified with themselves by the local community; and a gradual reorientation toward attempting to reduce middle- and old-age mortality.

Adetunji, J. 1992 Health behaviour among the Yoruba of Nigeria. Ph.D. thesis, Demography Program, Australian National University, Canberra, to be submitted.

Bhuiya, A., K.Streatfield, and P.Meyer 1990 Mothers’ hygienic awareness, behaviour and knowledge of major childhood diseases in Matlab, Bangladesh. Pp. 462–477 in J.C.Caldwell, S.Findley, P.Caldwell, J.Braid, and D.Broers-Freeman, eds., What We Know About Health Transition: The Cultural, Social and Behavioural Determinants of Health. Proceedings of an International Workshop, May 1989, Health Transition Centre, Australian National University, Canberra, 2 vols.

Bledsoe, C. 1990 Differential care of children of previous unions within Mende households in Sierra Leone. Pp. 561–583 in J.C.Caldwell, S.Findley, P.Caldwell, J.Braid, and D. Broers-Freeman, eds., What We Know About Health Transition: The Cultural, Social and Behavioural Determinants of Health. Proceedings of an International Workshop, May 1989, Health Transition Centre, Australian National University, Canberra, 2 vols.

Caldwell, J.C., ed. 1977 The Persistence of High Fertility: Population Prospects in the Third World. Changing African Family Monograph. Canberra: Australian National University, 2 vols.

Caldwell, J.C. 1982 Mass education as a determinant of the timing of fertility decline. Pp. 301–330 in J.C.Caldwell, Theory of Fertility Decline. New York: Academic Press.

1989 Routes to low mortality in poor countries. Pp. 1–46 in J.C.Caldwell and G. Santow, eds., Selected Readings in the Cultural, Social and Behavioural Determinants of Health. Canberra: Australian National University.

Caldwell, J.C., and P.Caldwell 1978 The achieved small family: Early fertility transition in an African city. Studies in Family Planning 9(1):2–18.

1986 Limiting Population Growth and the Ford Foundation Contribution. Dover, N.H.: Frances Pinter.

Caldwell, J.C., and K.Srinivasan 1984 New data on nuptiality and fertility in China. Population and Development Review 10(1):71–79.

Caldwell, J.C., P.H.Reddy, and P.Caldwell 1988 The Causes of Demographic Change: Experimental Research in South India. Madison, Wis.: University of Wisconsin Press.

Caldwell, J.C., I.Gajanayake, P.Caldwell, and I.Peiris 1989 Sensitization to illness and the risk of death: An explanation for Sri Lanka’s approach to good health for all. Pp. 222–248 in J.C.Caldwell, and G.Santow, eds., Selected Readings in the Cultural, Social and Behavioural Determinants of Health. Canberra: Australian National University.

Caldwell, J.C., S.Findley, P.Caldwell, J.Braid, and D.Broers-Freeman, eds. 1990 What We Know About Health Transition: The Cultural, Social and Behavioural Determinants of Health. Proceedings of an International Workshop, May 1989, Health Transition Centre, Australian National University, Canberra, 2 vols.

Caldwell, J.C., I.Orubuloye, and P.Caldwell 1991 Underreaction to AIDS in sub-Saharan Africa. Health Transition Working Paper No 9. Canberra: Australian National University.

Cantrelle, P., and T.Locoh 1990 Cultural and social factors related to health in West Africa. Pp. 251–274 in J.C. Caldwell, S.Findley, P. Caldwell, J.Braid, and D.Broers-Freeman, eds., What We Know About Health Transition: The Cultural, Social and Behavioural Determinants of Health. Proceedings of an International Workshop, May 1989, Health Transition Centre, Australian National University, Canberra, 2 vols.

Chen L.C., M.C.Gesche, S.Ahmed, A.I.Chowdhury, and W.H.Mosley 1974 Maternal mortality in rural Bangladesh. Studies in Family Planning 5(11):334–341.

Chen, L.C., E.Huq, and S.D’Souza 1989 Sex bias in family allocation of food and health care in rural Bangladesh. Pp. 247–263 in J.C.Caldwell, and G.Santow, eds., Selected Readings in the Cultural, Social and Behavioural Determinants of Health. Canberra: Australian National University.

Christakis, N.A., and A.M.Kleinman 1989 Illness behavior and the health transition in the developing world. Unpublished mimeograph, Harvard University.

Cleland, J.G., and J.van Ginneken 1989 Maternal education and child survival in developing countries: The search for pathways of influence. Pp. 79–100 in J.C.Caldwell, and G.Santow, eds., Selected Readings in the Cultural, Social and Behavioural Determinants of Health. Canberra: Australian National University.

El Din, M. 1977a The economic value of children in rural Sudan. Pp. 617–632 in J.C.Caldwell, ed., The Persistence of High Fertility: Population Prospects in the Third World. Changing African Family Monograph. Canberra: Australian National University, 2 vols.

1977b The rationality of high fertility in urban Sudan. Pp. 633–658 in J.C.Caldwell, ed.,

The Persistence of High Fertility: Population Prospects in the Third World. Changing African Family Monograph. Canberra: Australian National University, 2 vols.

Fauveau, V., B.Wojtyiniak, M.A.Koenig, J.Chakraborty, and A.I.Chowdhury, eds. 1989 Epidemiology and cause of deaths among women in rural Bangladesh. International Journal of Epidemiology 18(1):139–145.

Goody, J. 1971 Technology, Tradition and the State in Africa. London: Oxford University Press.

Halstead, S.B., J.A.Walsh, and S.Warren, eds. 1985 Good Health at Low Cost. Proceedings of a Conference held at Bellagio Conference Center, Bellagio, Italy, April 29-May 2. New York: Rockefeller Foundation.

Khan, M.E., R.Anker, S.K.Ghosh Dastidar, and S.Bairati 1989 Inequalities between men and women in nutrition and family welfare services: An in-depth enquiry in an Indian village. Pp. 175–199 in J.C.Caldwell, and G. Santow, eds., Selected Readings in the Cultural, Social and Behavioural Determinants of Health. Canberra: Australian National University.

LeVine, R.A., and S.Dixon 1990 Child survival in a Kenyan community: Changing risks over thirty years. Pp. 420–424 in J.C.Caldwell, S.Findley, P.Caldwell, J.Braid, and D.Broers-Freeman, eds., What We Know About Health Transition: The Cultural, Social and Behavioural Determinants of Health. Proceedings of an International Workshop, May 1989, Health Transition Centre, Australian National University, Canberra, 2 vols.

Lindenbaum, S., M.Chakraboorty, and M.Elias 1989 The influence of maternal education on infant and child mortality in Bangladesh. Pp. 112–131 in J.C.Caldwell, and G.Santow, eds., Selected Readings in the Cultural, Social and Behavioural Determinants of Health. Canberra: Australian National University.

Mateer, S. 1883 Native Life in Travancore. London: W.H.Allen.

Mosley, W.H. 1989 Will primary health care reduce infant and child mortality? Pp. 261–294 in J.C. Caldwell and G.Santow, eds., Selected Readings in the Cultural, Social and Behavioural Determinants of Health. Canberra: Australian National University.

Muhuri, P.K., and S.H.Preston 1991 Family composition and sex mortality differentials among children in Matlab, Bangladesh. Population and Development Review 17(3):415–434.

Okediji, F.O. 1975 Socio-economic status and attitudes to public health problems in the Western State: A case study of Ibadan. In J.C.Caldwell. ed., Population Growth and Socio-economic Change in West Africa. New York: Columbia University Press.

Orubuloye, I., and J.C.Caldwell 1975 The impact of public health services on mortality: A study of mortality differentials in a rural area in Nigeria. Population Studies 29(2):259–272.

Simons, J. 1989 Cultural dimensions of the mother’s contribution to child survival. Pp. 132–145 in J.C.Caldwell, S.Findley, P.Caldwell, J.Braid, and D.Broers-Freeman, eds., What We Know About Health Transition: The Cultural, Social and Behavioural Determinants of Health. Proceedings of an International Workshop, May 1989, Health Transition Centre, Australian National University, Canberra, 2 vols.

Sushama, P.N. 1990 Social context of health behaviour in Kerala. Pp. 777–787 in J.C.Caldwell, S. Findley, P.Caldwell, J.Braid, and D.Broers-Freeman, eds., What We Know About Health Transition: The Cultural, Social and Behavioural Determinants of Health. Proceedings of an International Workshop, May 1989, Health Transition Centre, Australian National University, Canberra, 2 vols.

Whyte, S.R., and P.W.Kariuki 1991 Malnutrition and gender relations in western Kenya. Health Transition Review 1(2):171–188.

This page intentionally left blank.