India - Issues in women's health

Report No. 15328-IN India Issues in Women's Health January 25, 1996 Population ancd Hunian Resources OpeQrtiios Division South Asia Country Departnmnt II Document of the World Bank   CONTENTS EXECUTIVE SUMMARY ......................................... i I. INTRODUCTION AND FRAMEWORK ............................... 1 II. THE DIMENSIONS OF FEMALE MORTALITY AND MORBIDITY ............ 5 Mortality Patterns and Differentials ................................ 6 Causes of Death ........................................... 14 Maternal Mortality ......................................... 15 Maternal Morbidity ........................................ 19 Links Between Fertility and Maternal Mortality and Morbidity .............. 21 Other Female Morbidity ...................................... 24 III. THE HEALTH CONTEXT ..................................... 29 A. The Sociocultural Context .................................. 29 The Status of Women ................................... 29 The Consequences of Female Disadvantage ...................... 32 The Productive Dimension ................................ 38 The Poverty Dimension .................................. 39 B. Health in the Sociocultural Context ............................. 41 Traditional Systems of Medicine ............................ 41 Beliefs Concerning the Causes of Illness ....................... 42 Beliefs and Practices Related to Pregnancy and Childbirth ............. 43 IV. HEALTH SERVICES AND THEIR UTILIZATION ...................... 48 A. Government Services ..................................... 48 Family Welfare Program ................................. 48 Integrated Child Development Services Program .................. 49 This report was designed to provide an overview of women's health in India and serve as background for dialogue concerning the Bank's assistance in population and reproductive health. The Green Cover report (1993) was prepared by Frances Plunkett, Population Specialist (SAIPH, previously SA2PH), based on discussions with knowledgeable individuals in and out of Govermnent, an extensive literature review, and three background papers authored by Meera Chatterjee, Saramma Mathai, and Marcia Griffiths, Wilma Lynn and Susan Brems. The report was subsequently revised by Anne Tinker, Senior Health Specialist (HDD), based on discussions with the Government of India, a considerable volume of new data (including the National Family Health Survey), and a continuing evolution of the GOI's policy and programs. Richard Skolnik, Division Chief, supervised the report; Indra Pathmanathan, Public Health Specialist, and Catherine Fogle, Population Specialist (SA2PH), provided peer review; and Aruna Chandran, Kirrin Gill, Nandini Ooman, Edna Jonas, and Ruth Utz assisted in the report's preparation.  B. Private Sector Services ............. ... .. .. ... .. .. ... .. .. . . 49 Non-Governmental Organizations ......... . . . . .. . . . . . .. . . . . . . 49 The Modem Private Sector .......... . .. . .. . .. . .. . .. . .. . . . 50 Traditional and Local Practitioners ........ . .. . . . . . . . . . . . . . .. . 50 C. Health Care Utilization and Coverage ........ . .. . . . . . . . . . . . . . .. . 51 Utilization of Public vs. Private Sector Services ...... . . . . . . . . . . . . . 51 Coverage of Health Services for Women ....... . . . . . . . . . . . . . . . . 53 V. IMPROVING THE HEALTH STATUS OF INDIAN WOMEN ...... . . . . . . . . . . 59 A. Provision of Services .............. .. ... ... .. ... .. ... .. .. . 60 Family Welfare Program ................................. 60 Policy . ........................................... 60 Improvement of Existing Services ........................... 61 Provision of Additional Services ............................ 63 Integrated Child Development Services Program .................. 68 Coordination of Services ................................. 69 B. Demand for Services ...................................... 70 Information, Education and Communication ..................... 70 Non-Govermnental Organizations ............................ 72 Community Participation ................................. 74 C. Action Research in Support of Women's Health ..................... 74 VI. FIGURES 2.1 Population Sex Ratio, India: 1901-1991.. 5 2.2 Age-Sex Specific Death Rates, India: 1989-1991.. 7 2.3 Expectation of Life at Birth, India and Major States: 1981-1985 . . 8 2.4 Infant Mortality Rates, India and Major States: 1991-1993 ..10 2.5 M/F Ratios, Young Child (0-4) Mortality Rates, India and Selected States: 1982-1984 - 1991-1992 ..12 2.6 Female Death Rates, Ages 20-24, India and Major States: 1990-1992 . .13 2.7 M/F Mortality Ratios in the Childbearing Years, India: 1979-1991 . .14 2.8 Location at Time of Death ............ ............ 18 2.9 Fertility Rates, India and Major States: 1992-1993 ..22 2.10 Percent Currently Married Women, Ages 15-19, India and Major States: 1991 .... 23 3.1 Female Literacy Rates, India and Major States: 1991 ..35 3.2 Primary Gross Enrollment Ratios by Gender, Major Indian States: 1991 . .36 4.1 Prevalence of Methods of Contraception: 1992-1993 ..55 4.2 Current Contraceptive Prevalence Rates, Modem Methods, India and Major States: 1992-1993 ..56  VII. Footnotes ......... 77 VmI. ANNEXES 1 Family Welfare Program .83 2 Integrated Child Development Services Program .85 3 Intensified Training of Dais Program .88 4 Safe Motherhood Program .89 5 Antenatal Services to be Provided by Female Health Workers .92 6 Guidelines for Development of Information, Education and Communication Stategies .94 IX. TABLES 2.1 Age-Specific Death Rates by Sex and Residence, India: 1989-1991 ............ 96 2.2 Crude Death Rates by Sex, India and Major States: 1992 ................. 97 2.3 Expectation of Life at Birth by Sex, India and Major States: 1981-1985 ... ...... 98 2.4 Expectation of Life at Birth by Sex, India: 1970-1975, 1976-1980 and 1981-1985 .............................................. 99 2.5 Infant Mortality Rates by Residence, India: 1972-1993 ................... 100 2.6 Young Child (0-4) Death Rates by Residence, India: 1972-1992 ............. 101 2.7 Infant Mortality Rates by Sex, India and Major States: 1992 ................ 102 2.8 Infant Mortality Rates by Sex, India: 1972-1992 ....................... 103 2.9 Neonatal and Postneonatal Mortality Rates by Sex, India: Various Dates ... ..... 104 2.10 Young Child (0-4) Death Rates by Sex, India: 1972-1992 ................. 105 2.11 Young Child (0-4) Death Rates by Sex, India and Major States: 1992 ... ....... 106 2.12 Young Child (0-4) Mortality Rates, India and Major States: 1982-1984 and 1991-1992 ........................................... 107  2.13 Neonatal and Postneonatal Mortality Rates, India: 1972-1992 ..... . . . . . . . . . . 108 2.14 Death Rates by Sex, Ages 20-24, India and Major States: 1990-1992 ..... . . . . . 109 2.15 Percentage Distribution of Deaths by Age and Sex Within Major Cause-Groups, Rural India: 1989 ......................................... . 110 2.16 Distribution of Deaths by Cause, Age and Sex, Rural India: 1989 .111 2.17 Percentage Distribution of Causes of Death by Age and Sex, Urban India: 1984 . . . 112 2.18 Percent of Total Deaths Due to Maternal Causes, India and Major States (Rural): 1989 .113 2.19 Causes of Maternal Deaths: 1978-1991 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 2.20 Comparison of Maternal Death and Control Cases from the Andhra Pradesh Maternal Mortality Study: 1984-1985 .115 2.21 Measures Identified for Prevention of Maternal Deaths, Andhra Pradesh Study . . . . 116 2.22 Pregnancy Outcome Indices, Five States: 1986 ..117 2.23 Total Fertility Rates, Women Ages 15-19, India and Major States: 1992-1993 .. 118 2.24 Percent Currently Married Women, Ages 15-19, India and Major States: 1991 . . . . 119 2.25 Nutritional Status of Female Children, Two Delhi Slum Populations: 1985-1986 . . . 120 2.26 Prevalence of Anemia Among Girls by Age, Major Cities of India . ....... . . . . 121 2.27 Gynecological and Sexual Diseases ..122 3.1 Literacy Rates by Sex, India and Major States: 1991 ..123 3.2 Time Use Data of Five Pregnant Women on a Normal Day . .124 4.1 Increases in Medical Personnel and Facilities, India: 1964-1988 .125 4.2 Utilization of Health Services for Illness within the Preceding Three Months, Four States . . . . . . . . .. . .. . . . . . . . . . . . .. . .. . . . .. . . . . . . . . . . . .126 4.3 Reasons for Dissatisfaction with Government Health Services: Four States .127 4.4 Coverage of Antenatal Care, India and Major States: 1992-1993 ............. . 128  4.5 Antenatal Care: Tetanus Toxoid Coverage, India and Selected States: 1992-1993 . . . 129 4.6 Anemia Prophylaxis Coverage, India and Selected States: 1992-1993 ..... . . . . . 130 4.7 Place of Delivery and Delivery Attendance, India and Selected States: 1992-1993 ......................................... . 131 4.8 Awareness of Contraceptive Methods Among Currently Married Women of Reproductive Age, India: 1992-1993 ........ . .. . . . . . . . . . . . . .. . . . . . 132 4.9 Unmet Need for Family Planning, India and Major States: 1988 ..... . . . . . . . . 133 5.1 Allocation of Tasks for Maternal Services ....... . . . . . . . . . . . . . . . . . . . . 134 BIBLIOGRAPHY ..135   EXECUTIVE SUMMARY 1. This document presents an analysis of the main factors influencing women's health in India. It identifies reasonable strategies in light of the health burden and the cost-effectiveness and productivity gains of interventions, within the context of previous and current initiatives. It concludes that focused efforts to improve women's health will result in substantial benefits in terms of human welfare and poverty alleviation, as well as sustainable economic growth. The Dimensions of Female Mortality and Morbidity 2. While India has made considerable progress in development during recent decades, improvements in women's health, particularly in the north, lag behind gains in other areas. Despite their innate biological advantage, more girls than boys die under the age of 5, and excess female mortality persists up to the age of 30. Overall, the population sex ratio has become increasingly adverse to females. Despite this, as is the case in virtually all countries, overall death rates for females are lower than for males. 3. General morbidity levels among Indian women are reported to be high. Emerging evidence suggests that the prevalence of reproductive tract infections is considerably higher than generally recognized, and that HIV/AIDS is spreading at an alarming rate. Nutritional deficiency and iron-deficiency anemia are generally widespread among Indian girls and women. Many women do not realize their full physical growth potential, and are, therefore, at risk of obstetric difficulties and/or delivering low birthweight babies. This perpetuates a vicious cycle: a female child who survives will be disadvantaged from the start and may never escape undernourishment, with consequent detrimental effects on her own general development and reproductive capacities. 4. Maternal mortality in India is estimated at 420 maternal deaths per 100,000 live births. This implies that about 15 percent of all deaths among reproductive aged women are pregnancy related. The major medical causes of maternal death are sepsis, hemorrhage, eclampsia, and obstructed labor. The lack of appropriate care during pregnancy and childbirth, especially adequate detection and management of complications, is a major contributor to maternal mortality. According to Sample Registration System (SRS) estimates, illegal abortion is one of the five major causes of maternal deaths. Since the Medical Termination of Pregnancy Act was passed in 1971, the reported number of legal procedures has reached about 600,000 annually. However, the estimated number of illegal procedures, mostly conducted under unsafe conditions, may be at least twice, and possibly, 10 times the number of legal ones.  - ii - 5. High maternal mortality levels in India are linked to high rates of maternal morbidity, which in turn, result in good measure from high prevalence of poor nutritional status and anemia among pregnant women. The caloric and protein intake of most Indian women during pregnancy and lactation averages as much as 25-35 percent below requirements, a situation which is likely to have serious adverse consequences for their health. Anemia during pregnancy increases the risk of maternal morbidity and is a major contributor to maternal mortality. Poor maternal nutrition is also a leading cause of low birthweight babies. 6. Maternal mortality and morbidity are closely linked to fertility, since no woman is at risk of maternal death or illness unless she becomes pregnant. The total fertility rate in India is 3.4 children per woman. Childbirth closely follows marriage, which occurs early; over 30 percent of women ages 15-19 are married. Early childbearing not only causes more rapid population momentum but also poses significantly more health risks than in the mid-reproductive years. 7. Pronounced regional variaton is reflected in demographic indicators, including fertility, mortality and literacy. Substantially unfavorable levels for these indicators in the northern states of Uttar Pradesh, Bihar, Madhya Pradesh, and Rajasthan reflect well-known socio-demographic contrasts between the so-called 'Hindi belt' and the rest of India. In contrast, the southern state of Kerala has achieved nearly developed- country levels for all of these indicators. 8. The poor health of Indian women is a national and individual concern because it affects the next generation through its impact on children as well as productivity at the household level and in the informal and formal sectors of the economy. Improving women's health is also economically efficient, since, within the health sector, interventions to improve women's reproductive health are among the most cost-effective interventions available. The Sociocultural Context of Health 9. Poverty underlies the poor health status of most of the Indian population. Additionally, sociocultural factors, as well as risks associated with reproduction, help explain the relatively disadvantaged status of women's health. The status of women, particularly in north India, contributes to the unfavorable demographic indicators. Traditional north Indian society is patrilineal. Since daughters require a dowry and leave their natal households after marriage, female children are seen as a liability. Girls are married as young adolescents into households outside their natal villages, their physical movements are controlled, and their contacts with natal kin are limited. In terms of exposure to and interaction with the outside world, as well as their ability to make decisions and control resources, north Indian women are more disadvantaged than women in other regions of India.  - 111 - 10. The unfavorable situation of women in India has consequences that include discrimination in allocation of and access to resources such as health care and food, limited education of females and early marriage. In its most extreme form, son preference results in female infanticide and, more recently, sex-selective abortion. This preference is readily apparent in the relative neglect of female children compared to male children. 11. The nutritional disadvantages experienced by Indian girls and women in the intra-household allocation of food and the provision of health care are well documented. Typically, adult men and male children are fed first, followed by older women. Given the nutritional demands of childbearing and lactation, this puts women of childbearing age at particular risk. Similarly, male children and adults tend to receive earlier, more and higher quality health care than do females. Expenditure on female health problems also tends to be less, consistent with the general pattern of relatively lower investments in females. 12. Female disadvantage in India is also evident in terns of education. According to the 1991 census, only 39 percent of Indian females above age 7 were literate, as opposed to 64 percent of males. The situation was particularly bad in some of the northern states where female literacy was 21-26 percent. However, school enrollment rates for girls show a promising upward trend. Increasing education levels and improvements in health status are closely linked. The more educated a woman is, the more likely she is to want and be able to obtain modern preventive and curative health care, including contraception and immunization; schooling for her children of both sexes and delayed marriage of her daughters. 13. Among girls, age at marriage has traditionally been low but is gradually rising. However, particularly in north India, it is still early enough to cut short educational possibilities. Early marriage also exposes girls to the risk of pregnancy before their physical development is complete. This increases the probability of maternal and infant morbidity and mortality. 14. Beginning in childhood, most rural women fulfill multiple productive functions, in addition to their reproductive roles and households tasks. Women from most landowning households must carry out extensive unpaid tasks. In landless families, the income of women from agricultural or domestic labor is often the chief support of the household. The strenuous physical labor that women perform, in addition to constrained household resources and men's preferential access to food, contributes to malnutrition among Indian women. Productive responsibilities are hardest on childbearing women. Typically women must work until late in their pregnancies without special provision for rest or food and resume work before they have fully recovered. This results in a cycle of maternal depletion that undermines women's abilities to carry out both productive and reproductive responsibilities.  - iv - 15. Health Beliefs. Allopathic (western) medicine has been added to the traditional text-based medical systems also current in India and the variety of local health practitioners who are patronized. A widespread set of traditional beliefs concerning the nature of health and the causes of diseases underlie the utilization of various systems and practitioners. For example, supernatural intervention or divine retribution may be seen as the cause of several diseases, including measles, tetanus, severe diarrhea, and severe malnutrition. The possibility of complications from 'evil eye' or supernatural interference is considered to be greatest during pregnancy, at the time of childbirth and during the first year of life. 16. In India, pregnancy is not generally considered to be a condition that requires medical attention, even in the event of complications. Only a woman who becomes seriously ill can give up her normal work routine. This places pregnant women at increased risk of pregnancy-related morbidity and mortality. Certain attitudes about pregnancy jeopardize women's (and their offsprings') health. For example, certain foods are avoided and a pregnant woman's diet is further limited by beliefs concerning the detrimental effects of overeating during pregnancy. 17. In rural areas the great majority of women give birth at home, assisted by older household women and traditional birth attendants (dais). The traditionally perceived association of 'pollution' with childbirth often leads to unsanitary deliveries. This can result in high rates of sepsis, as well as neonatal tetanus in the absence of proper immunization. 18. The importance of traditional beliefs concerning health and illness has never been adequately considered by the modern medical services. In a number of areas essential to women's health, traditional beliefs are in conflict with modern medicine. Illustrative examples include beliefs about the role of supernatural forces in causing certain diseases; perception that allopathic medicines such as iron and folic acid (IFA) tablets and oral contraceptives are strong and, therefore, dangerous; and concerns about the safety of invasive procedures such as IUD insertion and surgical abortion. Traditional beliefs regarding the safety of mother and child are, in fact, largely responsible for what may appear to outside observers to be casualness about pregnancy as well as the failure to adequately prepare for a birth or to appropriately respond to complications. Such local perceptions and beliefs complicate modern treatment of health problems and may make it difficult for health workers to provide services. Health Services and their Utilization 19. Government Services. Public sector services to meet the health and nutritional needs of Indian women are provided through the Family Welfare Program (FWP) of the Ministry of Health and Family Welfare and the Integrated Child Development Services (ICDS) Program of the Ministry of Human Resources Development. The public Family Welfare Program provides services to about one half  v - of the population, and is the primary source of preventive care in India. The FWP has achieved considerable progress in mortality and fertility reduction since its inception in 1951. The program comprises family planning and maternal-child health (MCH) services, which are organized on a subcenter outreach basis. Auxiliary nurse-midwives (ANMs) are deployed to cover a population of 5,000. Cases that an ANM is unable to handle must be referred to primary or community health centers or to hospitals. However, women may also go directly to a primary or community health center or a hospital without a referral. 20. Although there is a consensus among observers and the Ministry about program difficulties, the Ministry has found it hard to modify the way in which the Family Welfare Program is implemented at state and local levels. To improve the program's ability to meet the health needs of Indian women, the following resource and organizational issues need to be addressed: inadequate allocation of resources to the northern states; limited availability of temporary contraceptive methods and other reproductive health services; poor integration among family planning, reproductive and child health services; lack of prioritization of field work routines; gaps in the knowledge and skills of staff; weak links between subcenters and referral institutions; and insufficient and/or ineffective information, education and communication (IEC) efforts. 21. India took an important step in shifting the FWP toward a reproductive health approach when it initiated the Child Survival and Safe Motherhood Program (CSSM) in August 1992. The child survival component of the program has made good progress, particularly in immunization coverage. The safe motherhood component has been slower in reaching its objectives, because it requires significant expansion of previous program activities in the area of maternal care and because its implementation requires some physical upgrading of first referral units, training of staff and procurement of equipment not usually found below the district hospital level. 22. The objectives of the Integrated Child Development Services Program include improvement of the nutritional and health status of children under 6 as well as that of pregnant and lactating women and enhancement of the capacity of mothers to look after the health and nutritional needs of their children. Delivery of the package of ICDS services, which includes nutritional supplementation, is the responsibility of anganwadi workers, each of whom cover a population of about 1,000. ICDS currently covers about 40 percent of the development blocks in India. 23. The quality of health care for girls and women under the ICDS is particularly affected by the following issues: insufficient training of staff in aspects of the program affecting women; failure to target individual malnourished children, which has especially unfavorable consequences for young females, and low coverage of the supplementary feeding component for malnourished pregnant and lactating women.  - vi - 24. Private Sector Services. The private sector, both allopathic and traditional, is the principal source of curative care in India, and with government support could play a more instrumental and effective role in strengthening reproductive and child health. Currently, the modem private for-profit health sector is largely concentrated in urban areas. In rural areas a wide range of private health care providers are found, from fully qualified medical practitioners to local healers. Most of the illness burden resulting from the high incidence of infant and child sickness and the pervasive morbidity characteristic of rural India is addressed by a variety of traditional and local practitioners, as discussed above. 25. In many instances NGO programs that provide family planning, health and related services for women have been quite effective. NGOs' successes can be attributed to program flexibility combined with skill in interpersonal communication and eliciting community support and/or participation. NGO strategies often cut across sectors to provide a wide range of services, of which health may be only one. Some NGO efforts to organize women and tackle problems of women's status directly have been remarkably successful. However, in rural areas, NGO services are only available to a small proportion of the population. 26. Health Care Utilization and Coverage. Utilization of health care in rural India is eclectic: individuals are likely to resort to different practitioners at different times, depending on the condition as well as the age and gender of the patient. Private allopathic physicians are considered to be superior to government doctors since value is expected in return for payment, the paid practitioner is likely to be more polite and concerned and their hours are generally more flexible. Many govermment services are well utilized. However, there are several areas of dissatisfaction, including: unofficial payments for nominally free services and drugs; rude and improper behavior on the part of health staff; absence of staff; lack of supplies and drugs; and long waiting time to see a doctor. 27. While there is an 82 percent coverage rate for tetanus toxoid immunization of pregnant women, almost half of all pregnant women receive no antenatal checkup, and far fewer receive the three antenatal contacts which are recommended as a minimum. Problems exist on both the programmatic and demand sides. Programs have not been successful in inforning the public of the need for antenatal care. Therefore, many women are still unaware that female health workers based at subcenters provide antenatal care. In addition, many pregnant women and their families do not perceive the need for medical attention during pregnancy. 28. The anemia prophylaris program that provides iron and folate acid tablets to all pregnant women is a key component of antenatal care. However, the prophylaxis scheme has met with bottlenecks at the field level. According to the latest survey, only about 60 percent of mothers have been receiving tablets and promotion and monitoring to ensure client compliance have been weak. In an attempt to address these problems, the  - vii - Child Survival and Safe Motherhood Program is taking measures to correct problems with supply, logistics and monitoring of compliance. Coverage levels are reported to be improving in a number of states, although they remain low in states with high maternal mortality. 29. Levels of institutional childbirth or delivery assisted by a trained provider are low; overall, only one quarter of deliveries occur in health facilities. Among the noninstitutional births, over 40 percent are attended by untrained persons, usually older female household members. As with antenatal care, households have traditionally not perceived a need for outside assistance. The cost of nominally free government services is also a constraint to the utilization of services. However, the recent large scale training of traditional birth attendants is expected to lead to more hygienic delivery practices and to improved detection and referral of complications, thereby reducing deaths due to infection and other complications. 30. In order to obtain a rapid and significant reduction in maternal mortality, hospital facilities will need to be upgraded at the level of first referral units so that they can provide effective treatment of obstetric complications and emergency care. A functioning system depends on early detection of complications and accessible and acceptable referral facilities, particularly since the most life-threatening complications develop with little warning and need immediate attention. This highlights the importance of adequate transport for women experiencing emergency complications. Community planning and involvement may be the most cost-effective way to develop transportation schemes. 31. Use of modern methods of contraception has risen from about 10 percent in 1970 to 40 percent in 1993. Contraceptive prevalence varies from 53 percent in Maharashtra to 20 percent in Uttar Pradesh and Bihar. The most striking aspect of contraceptive use in India is the predominance of sterilization, which makes up more than 85 percent of the total use of modern contraceptive methods. Lack of knowledge about and access to contraceptive methods other than sterilization reflects the historical emphasis on sterilization in the FWP. Increasing contraceptive choice, particularly methods for delaying and spacing pregnancies, should become a higher priority. 32. Despite the legalization of medical termination of pregnancy (MTP), many women continue to resort to unsafe (illegal) abortion procedures. Several factors contribute to this, including the limited number of health providers and facilities that offer MTP and the perception that clinical procedures are invasive, and therefore, dangerous. The financial difficulties involved in obtaining legal abortion services, whether from a government or private facility, also make local, unsafe alternatives more common in rural areas.  - viii - Improving the Health Status of Indian women Overview 33. The Government needs to develop a strategy on the appropriate balance between the public and private sectors for improving women's health status. Within this evolving strategy, the public sector will continue to play a crucial role in the financing and provision of essential services for women, because some of these services have characteristics of public goods and, more particularly, several provide positive externalities. In the absence of public financing, provision of these essential services for women would in many cases be below the socially optimum level. Furthermore, equity considerations argue for continued provision of subsidized services to poor women. However, not all health services - including those that are publicly funded - need to be provided by the state. The challenge for Government is to maximize the reach and breadth of reproductive health services provided by the private sector, as well as to improve their quality, through appropriate incentives and regulatory arrangements. Mechanisms need to be explored to encourage a shift from the current private sector focus on curative care to a broader approach which includes promotion and prevention. 34. Progress made under the Government program needs to be continued and expanded. Priority efforts need to be made in the following key areas: (a) Expansion and strengthening of the Safe Motherhood Program comprising improved antenatal services and obstetric care, including management of obstetric complications and increased emphasis on strengthening the referral system; (b) Provision of quality family planning services, with greater emphasis on temporary contraceptive methods, as well as expansion of their availability and acceptability through social marketing; (c) Expansion of services for the prevention, diagnosis and treatment of reproductive tract infections and sexually transmitted diseases, including HIV/AIDS; (d) Provision of appropriate technical and communication skills training for staff; (e) Public education to increase knowledge about reproduction, contraception, safe sex, safe motherhood practices, and HIV/AIDS prevention; and (f) Utilization of operational research to improve the quality and availability of services.  - ix - 35. - Consistent improvements in women's health in the medium and longer term will require changes in the disadvantaged status of Indian women that underlies existing difficulties. There is general agreement that to this end interventions to improve women's educational levels and to delay marriage and childbearing, as well as increase women's access to earning and employment opportunities, should receive priority support. 36. In order to achieve both shorter and longer term objectives, the Government of India will have to make a national commitment of considerable strength and duration. Specific Approaches to Service Provision 37. Provision of Services. In order to strengthen the Family Welfare Program policies should be modified in the following areas: broadening of population policy to give due weight to factors other than family planning that affect fertility and reproductive health; allocating resources so that the north-south gap is bridged; modifying the current reliance on sterilization in favor of temporary methods; and expanding contraceptive choices and availability. 38. To improve existing Family Welfare services measures in the following areas are required: strengthening and integrating family planning, reproductive health and MCH services; organizing fieldwork routines so they are based on priority activities and clients; providing in-service training; improving the training for and supervision of traditional birth attendants; and supporting interventions that promote improved health and nutritional status among girls. 39. ln addition, strengthening and expanding the following key services should be a priority: antenatal care and attended delivery; anemia prophylaxis; modern temporary contraceptive methods; and medical termination of pregnancy. The reduction of maternal mortality requires expanding and increasing the effectiveness of essential obstetric care by addressing four major program gaps: inadequate early detection of complications; weak linkages between communities, subcenters and referral facilities; inadequate capacity of referral facilities; and the lack of emergency transport. 40. Additional health services for women are required to reduce non-pregnancy related morbidity. For example, to deal with the widespread problem of reproductive tract infections, it will be necessary to make their diagnosis and treatment an integral part of Family Welfare services. HIV/AIDS prevention efforts must also be integrated into the existing health programs. 41. In order to strengthen the ability of the ICDS Program to provide services for women and girls, measures in the following areas are required: in-service training relating to services for women; development of strategies to ensure that nutritional  - x - supplementation reaches malnourished and micronutrient-deficient young females, in particular, as well as pregnant and lactating women; and expansion of the program to include services for adolescent girls. 42. Cooperation and coordination between FWP and ICDS, particularly between field workers, is crucial to improving the effectiveness and provision of services because of the substantial overlap in target beneficiaries and the complementarity of services provided. Measures to assure this cooperation include: joint planning; coordination of work routines; joint supervision; joint training; and coordination of formal training curricula. 43. Demand for Services. Since exposure to the mass media in rural areas, particularly for women, will remain limited for the foreseeable future, female field workers who are in daily contact with women and their families are a key channel of health communication. Programs should, therefore, give priority to communication skills training for field workers. This training should address workers' own attitudes and beliefs and promote two-way communication between workers and clients. Another important way to increase the use of services is to involve women and communities in their planning and implementation. 44. The need for increased communication efforts in the following areas is especially acute: safe motherhood, particularly about the danger signs of problem pregnancies; the use of safe medical termination of pregnancy rather than traditional unsafe practices; modem methods of temporary contraception; awareness of HIV/AIDS prevention; male support for and involvement in reproductive health; appropriate diagnosis and treatment of gynecological problems; cultural constraints to acceptance of modem contraceptive methods, such as oral contraceptives and vasectomy; and promotion of equitable treatment of girls and women. 45. NGO Involvement. Government support of NGO involvement is particularly important for intersectoral projects that include women's health components which would not be implementable within the public sector. Since NGOs have been successful in communicating with clients and mobilizing community support that facilitates effective delivery of services, NGO staff should be involved to the extent possible in the training of government workers. Additionally, since public sector family welfare services are weakest in urban areas where NGO and for-profit private sector activity are concentrated, ways to facilitate greater cooperation between the Government, NGOs and the private for-profit sector in urban areas need to be developed.  - xi - 46. Action Research in Support of Women's Health. Given the difficulties of introducing changes in large, well-established programs, action (operational) research to develop, test and/or modify program strategies and activities take on an important role, particularly where an existing approach is not working. The phased introduction of improvements in the Safe Motherhood referral system in a small number of districts should serve as action research that can guide subsequent expansion of the program. These efforts should include expanded community participation. In addition, there is a need for action research in support of women's health in the following areas: safe motherhood messages; anemia prophylaxis; reduction of unsafe abortion practices; management of sexually transmitted diseases; increased use of temporary contraceptive methods; transportation for obstetric emergencies; nutritional supplementation for pregnant and lactating women; and adherence to appropriate guidelines for assuring quality of care.   I. INTRODUCTION AND FRAMEWORK 1.1 India has made considerable socioeconomic progress in the decades since Independence, as demonstrated by improvements in general indicators such as life expectancy, infant mortality and literacy. However, indicators of women's health, particularly in the north, are lagging behind progress in other areas. The population sex ratio continues to be adverse to females and is becoming more so; excess female mortality in the younger and childbearing ages persists; and maternal mortality in rural India is among the highest in the world. The objective of this study, therefore, is to highlight the particular reproductive and socioeconomic factors which affect the health of Indian women and to identify steps which can be taken to assure the equitable impact of further health improvements resulting in benefits for women, their families and the nation as a whole. 1.2 The health of Indian women is one of a number of linked gender-related concerns that manifest themselves in demographic, nutritional, educational, and other indicators. It is increasingly clear that these indicators are innately tied to the status of women within the context of poverty. In traditional Indian society, as in many other male-dominated agrarian societies, women, especially younger women, are disadvantaged. They are typically married as young adolescents, dominated by in-laws and husbands, restricted in their physical movements, and without control over property, money or household resources. Because of dowry obligations, they are regarded as a net loss to their natal families and their labor, whether productive or reproductive, is socially devalued. This inherently inequitable social system is perpetuated through a process of socialization that rationalizes and internalizes female disadvantage. 1.3 Women's disadvantage is apparent in factors which contribute to illness and reduce access to health care, including contraception. Women typically have unequal access to food, despite obligations to engage in strenuous physical labor; not even pregnancy provides an entitlement to additional food or rest. In addition to being undernourished in terms of protein and calories, a high proportion of Indian women are also anemic, leaving them vulnerable to a variety of illnesses which decreases their capacity to work, thereby exacerbating the vicious cycle of high fertility and maternal depletion. Economic and/or social factors prevent women from deciding when they or their children need health care and from obtaining it. In such circumstances, it is inevitably female infants and young children, especially if there is an older sister in the household, who are most disadvantaged, as mortality statistics confirm. 1.4 Gender issues are confounded with the consequences of poverty: if women are generally disadvantaged, the situation of poor women is likely to be worse. It is the poor women left as heads of households, as the result of widowhood or desertion, who constitute many of the 'poorest of the poor'. 1.5 The poor health status of Indian women has substantial consequences for other family members, household economies, and the larger economy as well as for women themselves. The most obvious effects arise from high female mortality rates  during childbearing years. The mortality rates of infants who lose their mothers are many times higher than average, and the direct and indirect consequences for older children are nearly as serious. The household consequences of pervasive female ill-health are also profound. A woman's health and nutritional status affects the birthweight and viability of her newborn child, her capacity to nurse and care for the child, and her ability to provide food and care to other children and family members. Moreover, in rural households dependent on women's labor, whether on family lands or as wage labor, a woman's health, as it affects her ability to work, directly and immediately impacts the welfare of the entire household. Poor urban households tend to be even more dependent on women's wage labor. 1.6 The development process, which has brought far-reaching demographic, social and economic change to India since independence, has not been as effective in improving the status of women. Ironically, development has sometimes had a negative impact on women, particularly rural women working in agriculture. 1.7 There is broad agreement that fundamental changes in the status of Indian women will require an increase in their educational level as well as expanded access to earning and employment opportunities. It would be unrealistic, however, to anticipate short-term health effects from such efforts. Given the dimensions and consequences of the health problems of Indian women, there is an urgent need for measures that result in improvements as rapidly as possible. This objective can be achieved most immediately and directly through strengthening and expanding existing health and related services, both public and private. This should be complemented by effective communication efforts aimed at increasing demand for services, supporting women's ability to utilize services, and promoting other measures that allow women to improve their own and their children's health. 1.8 Using this framework, the study proceeds as follows: (a) Chapter II uses available demographic and epidemiological data to identify dimensions and causes of female mortality and morbidity in India. (b) Chapter III summarizes, first, the implications of the Indian sociocultural context, particularly the status of women in traditional Indian society, for the health of Indian girls and women, and second, traditional beliefs and practices associated with pregnancy, childbirth and illness, as well as systems of medicine used by the variety of traditional practitioners who continue to provide much of the health care in rural India. (c) Chapter IV reviews existing public and private sector services available to meet the health needs of Indian women, as well as the coverage and utilization of those services.  (d) Chapter V outlines measures which could strengthen existing services and the demand for them, identifies additional services that would be required to improve women's health status, and indicates action research needed to develop and test new approaches. 1.9 This study gives priority to the northern 'Hindi-belt' states of Bihar, Uttar Pradesh, Rajasthan, and Madhya Pradesh. The unfavorable demographic differentials between these states, which account for almost 40 percent of India's population, and the rest of India are well documented. To achieve more rapid overall progress in improving women's health status, and to narrow the gap between the northern states and the rest of India, it will be necessary to identify the particular demand and supply constraints in the northern states, and the approaches and resources required to address them. 1.10 This study's recommendations are oriented to the needs of women in rural, rather than urban, areas. The health status of urban slum women poses as many problems as that of their rural counterparts, and are of particular concern since urban slum populations are growing more rapidly than rural populations. However, the availability of public and private health facilities and services is far more extensive in urban areas, and after having been neglected in past years, family welfare outreach in cities is now being strengthened systematically. In addition, in cities and towns physical access to services and emergency transportation does not present the major difficulties that it does in rural India. These factors are responsible for substantially lower urban mortality levels, although morbidity differentials are not likely to be as great. To the extent that the urban and rural situations differ, this study emphasizes the latter. 1.11 The broadest consideration of women's health issues is based on a life cycle approach, starting at birth and taking into account the cumulative effects of previous health status as each successive life stage is reached. However, given the extent of excess female mortality through the prime childbearing ages and high levels of maternal mortality in India, the study gives primary emphasis to women's reproductive health, as well as to some of the factors underlying excess female mortality at early ages. Other important concerns, such as the need to improve the nutritional status and health knowledge of adolescent girls before they marry and begin their childbearing careers and the health of older women who have completed their families, are given less attention than would be required in a comprehensive consideration of women's health issues. Household and occupational health issues for women are also noted, but measures required to address them, which would go much beyond the capacities of the health services, are not discussed. 1.12 This study does not deal specifically with financing issues, although at this time of financial stringency in India, recurrent cost and sustainability issues are becoming increasingly important in the health and family welfare sectors. A separate study specifically dealing with health financing has recently been undertaken by the Population and Human Resources Operations Division of the South Asia Country Department II. It  - 4 - is also generally acknowledged that many of the constraints affecting the provision of public sector health and nutrition services for women are not financial. The government's current emphasis on more efficient and effective use of existing physical and human resources is amply justified. This study, therefore, focuses on the measures necessary to address existing policy and implementation constraints, and thereby, achieve improvements in the quality, acceptability and utilization of services which will translate into improved health for Indian women. 1.13 This study focuses mainly on the public health sector, while recognizing its deficiencies and the difficulty of steering a large bureaucracy in new directions. Even though both the private voluntary and for-profit sectors have played increasing roles in recent years, and their importance is well recognized by government and donor agencies alike, the provision of family planning, preventive health, and nutrition services to the great majority of India's vast rural population, including the female half, will of necessity remain a public sector responsibility for the foreseeable future. 1.14 The study's emphasis on those aspects of services specifically relevant to women and the consequent recommendations made in the concluding chapter serve, in many instances, to reinforce the appropriateness and importance of conclusions that have also been reached from other perspectives--there is general agreement in the fairly extensive literature on India's Family Welfare and nutrition programs concerning issues and new approaches. In sum, the focus on health is an essential component of the larger effort to improve women's productivity and enhance their contribution to India's social and economic progress.  II. THE DIMENSIONS OF FEMALE MORTALITY AND MORBIDITY 2.1 For reasons that are not well understood, human females have a natural biological advantage over males. In low-mortality western populations, male death rates are higher than corresponding female rates at every age. Consequently, life expectancy at birth for females is higher than for males, and there are more females than males in the total population.' 2.2 By contrast, India is oLie of the world's few countries where males significantly outnumber females2: according to the 1991 census, for every 1,000 males there were only 929 females. What is more, since the beginning of the century the sex ratio of the population, calculated from the decennial census, has become progressively more masculine (Fig. 2.1). Figure 2.1 Population Sex Ratio, Indla: 1001-1l91 Fwmalsl per 1000 males 1000 972 804 6 0 -- - -- - - - -- - -- -- - -- -- 950 9419 040 -1 - - 030- - 920 - - 900 1001 1V11 l1021 1061 1041 101 1061 171 ttOl 1001 Sauroin ROE 1I91 t 2.3 One way to estimate the extent of female disadvantage is to compare the actual sex ratio to a sex ratio that would be expected in the absence of gender discrimination. A number is then calculated to represent the number of girls and women  - 6 - who died as a result of past and present discrimination, often called "missing women." The current estimate is that there are 35 million "missing women" in India; although estimates have ranged from 23 to 42 million (Coale 1991; Klasen 1994). 2.4 In earlier decades, India's peculiar sex ratio was generally attributed to relative undercounting of females in the census, differential migration and/or singularities of the sex ratio at birth in India. Only with the work of Visaria (1971) was it conclusively demonstrated that India's masculine sex ratio could not be ascribed to any of these factors and in fact reflected social causes characteristic of Indian society. As mortality levels in India have declined overall since 1921 (Dyson 1987), the increasing masculinity of the sex ratio suggests that women have been progressively less able to achieve mortality gains relative to their male counterparts. 2.5 The direct demographic cause of India's female-deficient sex ratio can be readily identified. Contrary to the current experience of most other populations, including developing countries of Africa, Latin America and southeast Asia, female death rates in India for all age groups from birth through ages 25-29 are higher than male rates (Table 2.1, Fig. 2.2).3 Excess female mortality is not unique to India or for that matter to south Asia. It was probably characteristic of traditional peasant societies from the Mediterranean across Asia to China before twentieth-century socioeconomic changes transformed mortality levels (Harriss 1989a; Amin and Pebley 1987). However, the severity and the persistence of excess female mortality in parts of south Asia indicate that the social causes underlying the demographic phenomena are likely to be deep-seated and difficult to address. Discrimination is also appearing in new forms with advancing technology, such as the increasing practice of sex-selective abortion, which could have an additional impact on the overall sex ratio. Mortality Patterns and Differentials 2.6 Regional, residential and age differentials in mortality are reviewed in the following paragraphs. Marked regional variation, the persistence of excess female mortality at young and childbearing ages and the worsening of male/female mortality ratios at ages 04 in a number of states emerge as key issues that must be addressed if female health status is to be improved. 2.7 Region. Pronounced regional variation is one of India's most prominent demographic characteristics. Mortality, fertility, illiteracy, and other demographic indicators are considerably higher in the northern states4 of Uttar Pradesh (U.P.), Bihar, Madhya Pradesh (M.P.), and Rajasthan than in the rest of India, reflecting underlying sociocultural contrasts between the so called 'Hindi belt' and the rest of India. On the other hand, the southern state of Kerala has achieved nearly developed-country levels for all of these. Measures of mortality provide good examples of north-south demographic differentials. In 1992, the crude death rate for U.P. was almost twice that for Kerala, 13 per thousand vs. 6 (Table 2.2), and similarly, life expectancy at birth for both sexes was  Flgure 2.2 Ag--Sx-Speolflo Death Rate, India: 1989-1 01 Deaths per 1000 populatlon 30.0 - ------------------------ 20.0 - 1 0 .0 - - - - - - - - - - - - - - - - - - - - - - - - 5 .0--- - - -- - - - -- - . . . .... 1oD ~~3 0.4 3. . & . .7 a.o SA a 1.4 w.& 1.7 0.0 N iINa 0-4 6-0 10-14 15-13 20-24 25.42 20-34 35-20 40-4* M Mal 0 Female Jouron: ROI 14 50 years in U.P. in 1981-1985, and 68 years in Kerala, or almost 20 years longer than in U.P. (Fig. 2.3 and Table 2.3). Regional differentials among females for the basic demographic indicators follow essentially the same pattern as the differentials for both sexes combined. The disparity in demographic and health indicators between the northern states and the rest of India, and the factors which underlie it, should be a major consideration in the planning of interventions intended to meet women's health needs (para. 5.8). 2.8 Urban-rural residence. As is the general experience of developing countries in recent decades, mortality is lower in urban than in rural areas of India. Urban death rates are lower than corresponding rural rates at every age, and from 1989-1991 the urban crude death rate for females was 7 per thousand vs. a corresponding rural rate of 11 (Table 2.1).  - 8 - Flgure 2.3 ExpecatIon of Life at Birth, India and Major States: 1981-1986 Thousands U.0 70.0 - B .4 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - .-- -- - -.-. 68.1 8o.8 seoe eo3 166.469 60.0 ~ ~ 0. 6. 40*~~~~~~~~~~~~~~~5L 30.0~~~~~~~~~~~~~~~~~~~~~~5. M0A KER PUN MAH KAR HAR AP GUJ WO TN IND RAJ OR MP UP fturoet ROI 1SU 2.9 Age. It is in terms of mortality at certain ages, rather than region or residence that female disadvantage in India is most apparent. Over the last several decades, Indian women have evidently been able to achieve relatively greater gains in life expectancy than have men. As a basis for comparison, the average life expectancies of men and women in the developing world are 62 and 63, and in the developed world 73 and 80, respectively (World Bank 1993). RGI life tables (Table 2.4) show that while male life expectancy was higher than female life expectancy in 1970-1975 (51 years for males vs. 49 for females), by 1981-1985 life expectancy had 'crossed over' to a slightly higher female figure (55 years for males vs. 56 for females). However, although this favorable trend is encouraging for women who survive to older ages, women under the age of 30 do not equally benefit from this advantage. The infant, child and prime childbearing age groups (ages 0-30) are at greatest mortality risk, relative to older women. The specific situations of female infants and children and women of childbearing age are discussed in more detail below. 2.10 Mortality at Younger Ages. The overall pattern of female mortality relative to male mortality at younger ages is shown in Fig. 2.2. The female mortality rate in 1989-1991 was in excess of the male mortality rate by 10 percent at ages 0-4 (28.9 vs.  - 9 - 26.3, respectively). This differential rose to 25 percent excess female mortality vs. male mortality at ages 5-9 (3.0 v.s 2.4, respectively) and declined to almost equal levels at ages 10-14. Higher excess female inortality at ages 5-9, compared to ages 0-4, presumably reflects the effect of higher male neonatal mortality in lowering the 0-4 ratio. The ratio at ages 10-14 is expected, since girls in the latter age group have survived the worst hazards of childhood but have not yet had to face the risks of childbearing. 2.11 Infant and young child mortality are among the most sensitive indicators of the health status of a population. Over the last two decades, infant and young child (0-4) mortality rates in India have declined significantly. RGI estimates indicate a fall in the infant mortality rate (IMR) from over 139 deaths per 1,000 births in 1972 to 74 in 1993 (Table 2.5), and a similar decline in the young child death rate from 57 to 27 deaths per 1,000 population age 0-4 (Table 2.6). State differentials are considerable (Table 2.7, Fig. 2.4). In 1992, Orissa and Madhya Pradesh still had IMRs over 100, while Kerala's rate of 17 was far lower than that of any other state. Young child death rates follow a similar pattern. 2.12 Given the female biological mortality advantage, the expectation is that male infant mortality rates will be higher than female. However, in India overall, the two rates are currently about equal (Table 2.8). The expected female advantage appears most clearly in Kerala. The most recent figures available (1992) show that only in the northern states of Uttar Pradesh, Rajasthan, Punjab, and Haryana, are female infant mortality rates still significantly higher than male rates (Table 2.7). 2.13 Infant mortality rates mask a basic difference between neonatal (deaths within 28 days of birth) and post-neonatal mortality. Neonatal mortality is due largely to causes such as neonatal tetanus, prematurity and congenital conditions that are not gender- specific. Post-neonatal mortality deaths, on the other hand, are by and large caused by infectious diseases whose incidence and severity are affected by immunization, health care and other controllable factors such as nutritional status. Consequently, differential treatment of male and female children (para. 3.13) will be reflected in excess female post-neonatal mortality, while the natural female advantage can be expected to appear in neonatal rates. To examine gender differences in infant mortality, therefore, it is necessary to consider neonatal and post-neonatal mortality separately. RGI does not publish neonatal and post-neonatal mortality rates by sex. However, the gender disparity between neonatal and post-neonatal mortality has been documented for a number of rural and urban study populations in north and south India; in these early studies (1965-84) male/female neonatal mortality ratios were more than one but post-neonatal ratios ranged from 0.53 to 0.80 (Table 2.9). 2.14 A comparison between male and female mortality at young ages can be obtained by examining the sex-specific death rates for ages 0-4, where the natural female neonatal advantage has less weight than it does for infant mortality (Table 2.10). The  - 10 - FRgur 2.4 Infant Mortality Rates, India and Major Stabs: 1991-193 DeaUts per 100P liv bliths 1405 I1@A 2180. -- --------------------.- . .. - . . . . . ---- . . . . . . . .. --- ---- --_ 109.0 10 .0 - 4 A ------ - - - - - - - - - - - - - - - - - - - - --.-.-.-.- 6G.OA 57.0. 8.8 .153 20A 15.0 OR MP UP RAJ IND KRP IH AP HAR WS Ow TN MAH PUN KER Sotwo: Ial 1994 extent of female disadvantage at these young ages becomes evident in state figures (Table 2.11). Only in the states of Kerala, Karnataka, Andhra Pradesh, and Maharashtra were 0-4 mortality rates higher for males than for females, as is characteristic of ratios worldwide. The recent Population Research Centre (PRC) and IIPS (1994) report on Uttar Pradesh described a more serious disadvantage for young girls than previous reports. The survey found that during the post-neonatal period, 124 girls die for every 100 boys. The most significant differential was found between the ages of 1 and 5, where 170 girls die for every 100 boys. If the nationwide 0-4 male mortality levels are taken as an index of the level of child mortality that can be achieved under existing circumstances in India, then it is clear that excess female deaths at these ages - more than 100,000 a year - must be attributed to gender discrimination.  - 11 - 2.15 Even more significant than the current female disadvantage in young child mortality rates are recent relative mortality trends for this age group. Until about 1983, female rates were declining relatively faster than were male rates. However, since then, despite continued mortality decline for both sexes, the gap in mortality rates for this age group increased rather than decreased; in other words, female disadvantage became greater, rather than the reverse. For India overall, from 1982-1984 to 1991-1992, the male/female mortality ratio for the 0-4 age group declined from 0.93 to 0.88. 2.16 The unfavorable trend in the M/F young child mortality ratio is most evident in 6 of 14 major states, in the south as well as in the north (Table 2.12, Fig. 2.5). The gap has widened the most in Orissa, Bihar and Uttar Pradesh between 1982-1992. 2.17 The reasons why the gap between male and female young child mortality did not continue to narrow after the early 1980s are not clear. The same trends are also seen in the ratio of male/female infant mortality rates since the early 1980s (Table 2.8), in spite of both neonatal and post-neonatal mortality rate declines overall (Table 2.13). Evidence from a small-scale study (Das Gupta 1987) suggests that the occurrence of increasing female mortality disadvantage at young ages in Punjab is linked to a strong son preference combined with declining fertility (para. 3.13). The implications of recent increases in female mortality disadvantage at younger ages are quite serious. 2.18 Despite recent declines in young child mortality, 34.5 percent of all deaths in India occur among children under 5. Further reductions in child mortality, therefore, are critical to improvements in life expectancy. For example, between 1941 and 1970, reductions in infant and child mortality in India accounted for 40 percent of the increase in life expectancy, compared with contributions of 14 percent, 27 percent and less than 20 percent for the age groups of 5-14, 15-44 and 45 years and above, respectively (Ruczicka 1984). An effective way to improve overall life expectancy in India, and at the same time improve the sex ratio, would be to reduce current levels of excess female young child mortality. 2.19 Mortality in the Childbearing Years. With the advent of the childbearing years, excess female mortality in India rises steeply (Table 2.1; Fig. 2.2). In percentage terms, excess female mortality in India is higher in the childbearing years than at younger ages. Female mortality in 1989-1991 exceeded male mortality by one-third at ages 20- 24, the years of highest fertility. After the bulk of childbearing has been completed and fewer women are at risk from maternal causes, the biological advantage of those women who have survived becomes evident. By 1989-1991 mortality 'crossed over' after ages 25-29, with male death rates remaining higher than female rates for the rest of the life span.  - 12 - Figure L. H/F Ratios, Young Child (0-4) Mortallt Rats, India snd Selected S:W: 1982-198 - 1991-1992 '/P RHaos 1.1S 1.07 IA. lA O - - - - -- . . . . . . . . . . . . . . . . . . . . . . . .91 O 0 1 9 - . - --- -- - -- . . . . . . . 0.34 INDIA AP ORI MP BIN Ru UP * 1982-1984 E 1991-1992 owan: Ra IO 2.20 Regional variations in mortality during the childbearing ages are substantial. At ages 20-24, when childbearing is greatest, female death rates in 1990-1992 varied from 0.9 per 1,000 female population in Kerala to 4 per 1,000 in Madhya Pradesh (Table 2.14, Fig. 2.6). With the exception of Kerala and Punjab, the male/female death rate ratios at these ages indicated excess female mortality for all the states, with the northern states having the least favorable ratios. The pattern is more favorable to females in urban areas, where excess female mortality in the childbearing years persists only through ages 20-24 (Table 2.1). 2.21 The 'Cross-Over' Transition. Demographic experience in India and elsewhere, suggests that as overall mortality levels decline, excess female mortality begins to disappear. This trend has been most clearly documented for Sri Ianka, where excess female mortality in the early 1950s was similar to the current Indian pattern. With declining mortality it is now completely eliminated. As might be expected, excess  - 13 - female mortality disappeared first at ages 10-14 and the later childbearing ages, then at the young and prime childbearing ages (Langford 1984). In India, trends in the ratios of age-sex-specific death rates over the last two decades, as well as urban trends, all suggest that a similar transition is underway, although slowly and unevenly. Overall, excess female mortality is unmistakably declining in the childbearing ages (Fig. 2.7). Figure 2.6 Female Death Rates, Ages 20-24, India and Major Stat": 190-1992 5.0 4.8 4.0~~~~~~. 3.0 -- - - - - - - - - - 3.0 - - - - .2.8.2.5 ~1U 2.7 2.7 2.0 - ----------- 1 .0~~~~~~~~~~~~~~~~~~~~~~~. 1.. - - . - - . . A 10.0 MP OR BIH UP RAJ IND AP OUJ HAR TN W MAH KAR PUN KER SIwoau ROI 1994 However, in the northern states current levels of female mortality are so high that large imnprovements will be needed before ratios reach more desired levels. As noted above (paras. 2.15-2.16), trends for ages 0-4 are currently in the opposite direction. It can be expected that excess female mortality will gradually diminish, persisting longest in the young child and prime childbearing ages and the northern states (Dyson 1987).  - 14 - However, if the transition is to be hastened, and particularly if declines in 0-4 male/female mortality ratios are to be reversed, it will be necessary for health, nutrition and other social sector programs to specifically target those groups and those regions currently at highest risk (paras. 5.8, 5.16). Flgurs 2.7 M/P Mortality Ratios In the Childbearing Yearn, India: 1979-1901 M/F Ratios 1.4 1.12~~.1 1 .2 . ..---------- ----- --- --- --- --- --- --- --- --- --- 1.07 0.S 0.80 0.8 ---------0.76------ 0.66 0.61 0.6-- 0.4--- 0.2 0 15-19 20-24 25-29 30-34 35-39 - 11979-198 81 1 1989-1991 souros: RO 1994 Causes of Death 2.22 Data on broad causes of death in rural India are available from a sample survey of deaths carried out by the Sample Registration System in rural areas, lay reporting and incomplete records of medically certified deaths in urban areas. The RGI rural data for 1989 are summarized in Tables 2.15 and 2.16 and urban data for 1984 in  - 15 - Table 2.17. These data are admittedly subject to substantial misreporting and may present a picture that has somewhat changed. However, they do shed some light on the particular problems of women's health. The rural data indicate that, leaving aside maternal deaths, deaths 'peculiar to infancy', and 'old-age deaths', about 30 percent of deaths are attributable to respiratory causes and 16 percent to circulatory disorders, the two leading causes of death. The other major causes, including accidents, account for around 10 percent of deaths each. Within each age group, the pattern does not differ markedly between males and females, except, of course, for maternal deaths in the childbearing years. About 12 percent of deaths among women in the childbearing ages (15-44) are ascribed to maternal causes, a figure much lower than that derived from demographic sources or community-based studies (para. 2.27). Figures for the percent of total deaths due to maternal causes (Table 2.18) indicate that the percentage is six times higher in the highest state, Orissa, than in one of the lowest states, Punjab. The percentages in the northern states are much higher than those in the southern states. 2.23 More specific causes of death are available from the urban medical certification of death data (Table 2.17). The pattern is not essentially different from that indicated by the rural data. At ages 5-14, more females than males appear to die of respiratory illnesses, while the reverse is true after the age of 25, perhaps due to a higher incidence of tuberculosis in men. Tuberculosis is the single most important communicable disease among both adult men and women, accounting for some 6 percent of all deaths reported by the rural survey and over 10 percent of medically certified deaths in urban areas. Although circulatory disorders account for similar proportions of male and female deaths, adult male deaths are more specifically attributed to heart attacks and female deaths to anemia. The principal discernible difference in cause of death between rural and urban areas is that a higher proportion of urban than rural women are likely to die from injuries and accidents than their male counterparts, between the ages of 5-44. 2.24 Causes of death appear to vary slightly by region. For example, coughs and disorders of the respiratory system account for a higher proportion of deaths in the northern states of Gujarat, Haryana, Madhya Pradesh, Rajasthan, Uttar Pradesh, and West Bengal, compared with the southern states of Andhra Pradesh, Karnataka, Tamil Nadu, and Kerala as well as Orissa. The extent to which this can be ascribed to climatic differences is unclear. However, since these deaths account for the largest portion of non-old-age deaths among females in high-mortality states, it may be that increased attention to respiratory infections in these areas could make a significant contribution to the reduction of female mortality and morbidity. Maternal Mortality 2.25 Our understanding of maternal mortality' in India--levels, causes, patterns--is at best incomplete and unsatisfactory. In contrast to infant mortality, for which RGI estimates are available, there is no adequate system to regularly collect for maternal  - 16 - mortality data.6 Yet even on the basis of the existing inadequate data, it is clear that maternal mortality in India is quite high. It is estimated that, although India has only about 15 percent of the world's population, at least 25 percent of the world's maternal deaths occur in India. 2.26 Levels. The recent National Family Health Survey estimated a maternal mortality ratio (maternal deaths per 100,000 live births) of 420 per 100,000 for 1990-1991 for India. This is lower than the earlier, community based estimates, which were derived from quite small numbers of maternal deaths in limited areas. Although the data are not adequate to support firm conclusions, one would expect a declining maternal mortality rate and ratio, given declining birth rates among higher risk younger and older women. 2.27 Assuming current Indian fertility and mortality levels, a maternal mortality ratio of 420 per 100,000 implies a maternal mortality rate (number of maternal deaths per 100,000 women of childbearing age) of about 55. This level of maternal mortality also implies that about 15 percent of all deaths among women in the childbearing ages were maternal deaths. By comparison, maternal mortality ratios in Europe are on the order of 10 per 100,000 live births. Due to the greater likelihood that she will become pregnant, combined with the greater likelihood that she will die of maternity-related causes once pregnant, the average Indian woman is almost 100 times more likely to die of a maternity-related cause than her counterpart in developed countries. 2.28 Causes of maternal mortality. Information from the RGI, hospital records and community-based data show that over 80 percent of maternal deaths are due to six major causes: anemia, hemorrhage, eclampsia, obstructed labor, sepsis, and unsafe abortion. The causes for maternal deaths, including 'abortion', are summarized in Table 2.19. The hospital and community-based data are not strictly comparable and, as noted above, the RGI rural survey data are not considered to be particularly reliable. Nevertheless, given their diverse origins, the data are fairly consistent, and the community-based field studies carried out in Andhra Pradesh (Bhatia 1988) and Karnataka (Reddy 1992) agree reasonably well. These data indicate that sepsis and hemorrhage were the leading causes of maternal death, responsible together for 33-44 percent of maternal deaths. The other major direct causes of maternal death were eclampsia and obstructed labor. About one quarter of deaths were from indirect causes, the ones specified being acute anemia, hepatitis and heart disease. Thus, maternal mortality in India follows a similar pattern to that in other parts of the developing world. 2.29 The studies cited reported that about 5-17 percent of deaths were the result of 'abortion', which refers to either induced or spontaneous abortion (miscarriage). The Andhra Pradesh and Karnataka studies indicated that about two thirds of 'abortion' deaths involved induced abortions and the other one third, miscarriages. According to the Andhra Pradesh study, roughly one half of 'abortion' deaths were due to hemorrhage and the other one half to sepsis.  - 17 - 2.30 Hospital-based studies report hemorrhage, abortion, eclampsia, and anemia as the most important causes of maternal deaths in hospitals, claiming up to three quarters of the lives lost (e.g., Sengupta and Gode 1987; Damania et al. 1988; Devi and Singh 1987; Balmur et al. 1983; Jindal et al. 1990). 2.31 The findings of the Andhra Pradesh study (Bhatia 1988), a WHO-sponsored retrospective field study of deaths among females of reproductive age carried out in Anantapur district of Andhra Pradesh in 1984-1985, merit particular attention because of the study's comprehensive design and careful execution. The study estimated that about 36 percent of all deaths among females aged 15-49 in the reference period were due to maternal causes and that these maternal deaths represented a maternal mortality ratio of 800 per 100,000 live births and a maternal mortality rate of 120 per 100,000 female population aged 15-49. This maternal mortality ratio is the highest of the available estimates for India and provides evidence of the high levels of maternal mortality in parts of India during the last decade, even in the south. 2.32 Of the maternal deaths for which detailed information was available, about two thirds were due to direct obstetric causes (Table 2.19). Sepsis accounted for a larger proportion, and hemorrhage a smaller proportion of deaths than in the other studies cited. However, the pattern is basically similar, including the high level of anemia as an indirect cause of maternal death. Some 10 percent of deaths were due to 'abortions', about two-thirds of which were induced in order to terminate unwanted pregnancies. In the case of almost one quarter of the deaths, family members were not aware of the seriousness of the woman's condition and, therefore, took no action to obtain assistance, while in another 4 percent of cases family members reported that they were aware that something was wrong but did not attempt to call a health worker or doctor. 2.33 For each case of a maternal death, the study randomly identified a woman in the same urban area or village who gave birth during the reference period and survived as a 'control case'. A comparison of the women who died and the associated control cases sheds additional light on the characteristics of those women who died and the causes of their deaths (Table 2.20). In terms of overall socioeconomic status, the two groups were similar. There were no major differences in antenatal care; in both groups, the percentage registered for antenatal care was 40-50, and the mean number of antenatal visits was actually slightly higher for mortality cases. 2.34 On the other hand, in terms of age, pregnancy history, predisposing conditions and family composition the differences between the groups were striking. As might be expected, significantly greater percentages of mortality cases had poor obstetric histories, predisposing health conditions and danger signs during the pregnancy. Although the mean number of pregnancies was slightly higher among the women who died, the mean number of living children was lower, indicating greater child loss. The women who survived had both more living children and more living sons than those who died (2.9 vs. 2.4 and 1.4 vs. 0.9, respectively). In fact, 55 percent of the women who  - 18 - died did not have a living male child, as compared to 25 percent among the control group. That women with poor obstetric histories but without sons persisted despite the risks of further efforts to have children, would seem to be a clear reflection of the pressure on Indian women to bear sons (para. 3.5). 2.35 After examination of available infonnation, expert opinion was that approximately 72 percent of the total deaths were preventable. Of the deaths which were preventable 32 percent could have been averted with proper antenatal care, while 69 percent of the deaths could have been averted with proper referral care. Other studies have also indicated that most maternal deaths in India, as in developing countries generally, are preventable (Rajaram 1989; Krishna 1989; Rao 1980). A summary of the measures identified in the Andhra Pradesh study as likely to have prevented deaths points out the interventions that are required to achieve reductions in current high levels of maternal mortality in India (Table 2.21). Adequate antenatal care, including treatment of anemia, would have prevented almost one third of the deaths; while two-thirds required appropriate care at a referral facility. 2.36 Of particular significance are the findings concerning the location of women at the time of their deaths. (Fig. 2.8) Figure 2.8 Location at Time of Death Percent (Andhra Pradesh) Died at home 41 Died at a health facility/hospital 50 Died in transit 9 Bhatia 1988 The figures indicate that about half of the women who died from maternal causes were transported to a PHC or hospital and died there. This raises a concern about the quality of facility-based care, as well as the timeliness of referral of complications and whether and how quickly women who are referred are able to get to referral facilities. There is no way to estimate the number of women who died at home because transport was unavailable or unaffordable. These findings emphasize the importance of prompt emergency referral and transport and effective management in reducing maternal mortality (para. 5.33).  - 19 - 2.37 Abortion7. Induced abortion was legalized in India in 1971 with the enactment of the Medical Termination of Pregnancy (MTP) Act. The Ministry of Health and Family Welfare subsequently undertook a systematic effort to train doctors, provide equipment and approve facilities where procedures could be carried out. The number of legal abortions in India rose rapidly after 1972, but in recent years, for reasons that are not well understood (paras. 4.30-4.32), has leveled off at about 600,000 annually (MOHFW 1992). Many legal private-sector abortions are probably not reported. Illegal abortion includes a spectrum of practices, from modem surgical procedures carried out by unapproved private practitioners to a variety of folk methods (para. 3.56). The levels of illegal abortion and associated mortality are not known with any certainty. In a hospital study cited above (Rao 1988), sepsis due to abortion was the single highest cause of death, responsible for 26 percent of 'direct' obstetric deaths and 18 percent of death from all causes. In the Andhra Pradesh and Karnataka community-based studies (Bhatia 1988; Reddy 1992), illegal induced abortions were estimated to be responsible for about 6 percent and 3 percent of total maternal deaths, respectively. 2.38 The Indian Council of Medical Research carried out a study of induced abortion in the five states of Uttar Pradesh, Rajasthan, Orissa, Haryana, and Tamil Nadu in 1983-1985 (ICMR 1989b). It indicated that for the five states combined, 6 per 1,000 pregnancies ended in a legal abortion and 13 per 1,000, or somewhat more than twice the legal rate, ended in an illegal abortion (Table 2.22). If this relationship holds for subsequent years and for the rest of India as well, the reported level of about 600,000 legal abortions in 1990 would imply a total of about 1.3 million illegal abortions annually for all of India. For the five states combined, only about 55 percent of the abortions were carried out in the first trimester, and of these only about one quarter were provided by doctors (government or private) or other health staff. A recent report supported by the Ford Foundation argues that previous research seriously underestimated the magnitude of illegal abortion and that nearly 7 million induced abortions may occur annually. This suggests that for every legal abortion, 10 more are being performed under illegal conditions. The study also estimates that abortion-related deaths are significantly higher than previously reported and account for about 15 percent of all maternal deaths (Chhabra and Nuna 1994). Extending the provision of MTP services by trained doctors in proper facilities should result in the reduction of maternal mortality associated with unsafe abortion (para. 5.23). Comprehensive family planning services, including a contraceptive choice component, should indirectly reduce maternal mortality associated with unsafe abortion by preventing unwanted pregnancies. Maternal Morbidity 2.39 While it is reasonable to expect that the morbidity profile of pregnant women will differ significantly from that of non-pregnant women, it is not adequate to infer the incidence and pattern of morbidity during pregnancy from studies of the causes of maternal deaths. During pregnancy a woman may suffer from non-life-threatening conditions which hamper her ability to function or have long-term implications for her  - 20 - post-pregnancy health and well being. She may suffer diseases common among adults in India such as tuberculosis or malaria, or develop uterine prolapse or obstetric fistulae as a result of pregnancy. Based on estimates of maternal morbidity for developing countries, at least 40 percent of pregnant women in India have a serious illness and 15 percent of women develop life-threatening complications during their pregnancies. The proportion would be even higher if women who had miscarriages and stillbirths were included. Unfortunately, there are no recent community-based studies of maternal morbidity in India. Most studies focus on the nutritional status of pregnant women, particularly anemia, while a few have investigated specific conditions such as syphilis (e.g., Upe et al. 1979). 2.40 The findings of a prospective study of 349 pregnancies in 281 women in Rajasthan, even though they relate to 1974-1980, are therefore, worth presenting in some detail. The incidence of illness correlated strongly with parity, increasing from 14 percent among primiparas to 30 percent among second paras and up to 58 percent among women of fifth or higher parity. The incidence and severity of anemia have also been found to increase with parity (ICMR 1974). Maternal morbidity in this study was comprised of: complications due to pregnancy, childbirth and puerperium, of which post- abortion complications was the single leading category (29 percent); pyrexia of unknown origin (26 percent); infectious and parasitic diseases (11 percent); respiratory diseases (8 percent); and skin diseases (8 percent). 2.41 Nutritional status during pregnancy. The generally poor nutritional status of Indian girls and women (paras. 2.56-2.59) is a vicious cycle that has particularly devastating consequences for pregnant and lactating women and their infants. Malnourished women are more likely to give birth to low birthweight babies, and if the underweight baby is a female who survives, she, in turn, is likely to continue to be undernourished throughout her childhood and adolescence, with consequent detrimental effects on her reproductive and lactating capacities, not to mention her overall development. 2.42 The caloric and protein intake of Indian women during pregnancy and lactation, particularly those of lower socioeconomic status, is likely to be grossly inadequate. A pregnant woman's initial low nutritional status is frequently aggravated by failure to supplement her diet to meet the additional nutritional demands of pregnancy (paras. 3.48-3.49). The direct consequences are inadequate weight gain during pregnancy and delivery of low birthweight babies. In a study of pregnant women in U.P., average weight gain over the entire pregnancy averaged 6.9 kg, ranging from 6.3 kg among the poorest group to 7.7 kg among the best off (Bhardwaj et al. 1990). Other studies have reported similar figures (Tripathy et al. 1987). By comparison with the average Indian weight gain of about 7 kg., the average in Thailand and the Philippines is close to 9 kg (Krasovec and Anderson 1991) and in developed countries about 12 kg.  - 21 - 2.43 Anemia during pregnancy. Anemia in pregnancy, which is characteristic of many developing countries, increases the risk of maternal morbidity and mortality and is also associated with low birthweight babies and poor lactational capacity. International data indicate that some 40 percent of severely anemic women are at heightened risk of heart failure and fatal hemorrhage during childbirth. Maternal causes of death data for India (Table 2.19) also indicate that severe anemia is a major underlying cause of maternal death. 2.44 While anemia is prevalent in Indian women of reproductive age it is even more acute in pregnant women, due to higher iron requirements which increase about five-fold during pregnancy (Hallberg 1988). A variety of studies have shown that a high percentage of pregnant women in India are anemic, particularly in the last trimester. One study in rural districts of Gujarat and Maharashtra (Christian et al. 1989) found that about 90 percent of pregnant women had hemoglobin levels below 11 g/dl and were anemic by WHO standards. Other studies have found similar levels of anemia in pregnant women (K.N. Agarwal 1984; Raman 1988; D.K. Agarwal et al. 1987; Raman 1980; Sheshadari et al. 1989). Links between Fertility and Maternal Mortality and Morbidity 2.45 Discussion of maternal mortality and morbidity is incomplete without consideration of the underlying cause, namely fertility--no woman is at risk of maternal death or illness unless she first becomes pregnant. Childbearing poses particularly high risks for women when it is too early, too closely spaced, too late, and/or too frequent. 2.46 In most developing countries, fertility decline has taken place first and most rapidly at higher parities and older ages as couples begin to terminate childbearing deliberately after the birth of a certain number of children. India is a case in point. Marital fertility at ages 35-49 has fallen by more than half over the last two decades; in both rural and urban areas, women age 35 and above now contribute only 10 percent or less to total fertility. A reduction in higher parity births in India is also indicated by the decline in the fertility of married women from an average of about 6 children in 1970 to the most recently estimated total fertility rate (TFR) of 3.4 children per woman. Fertility levels vary widely among the states, with the highest TFR, 4.8 in Uttar Pradesh (Fig. 2.9, Table 2.23). 2.47 The risk of too-early pregnancy is also declining because the age at which women marry is rising. (In India, exposure to the risk of pregnancy is almost entirely limited to married women, so that the risk of childbearing by adolescent girls is governed by the age at which girls are married.) In 1972, 41 percent of women were married at ages 15-19. Mean age at marriage for women8 has been rising steadily for several decades, and from 1971 to 1981 increased from 17.2 to 18.3 years. Relatively speaking,  - 22 - Figure 2.9 Fertility Rates, India and Major States: 1992-1M TFR SAO 1.F00 -4,U ...... .. .. .. . ............ ... .. .. . .. .. . . ... . . . .. . . . 4A0 4.0S 3.st S . 3.68 a.- - - - 1*3~~2 2.32 2.32 2.II-LU 1 S sAo . . ........... em e_-2--- - moo ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ 2 UP BIH HAR MP RAJ IND WUJ OR WB PUN MAH KAR AP TN KER SowIn: IIPS 194 however, female age at marriage in India is still low. The proportion of currently married women ages 15-19 in 1991, by state is shown in Table 2.24 and Fig. 2.10. As expected, the northern states had higher percentages married, and in Rajasthan, Madhya Pradesh, Bihar, and Andhra Pradesh, the proportion of girls 15-19 who were married was still over 40 percent.  - 23 - Figure 2.10 Percent Currently Marrled Women, Ages 15-19, India and Malor Stabs: 1991 46 42 42 4 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 so . - - - -- - - - - - - . 2-. 25 10 10 10- - -- - MP RAJ AP BIH MAH UP IND HAN WE GUJ KAR OR TN KEN PUN So: ROI 194 2.48 Spacing of births is determined by a combination of factors related to breastfeeding, nutritional status and exposure to the risk of conception, for all of which data are limited. Although there is insufficient evidence to show that short birth intervals increase the risk of maternal mortality, many researchers have hypothesized this association (Govindasamy et al. 1993). Short birth intervals do, however, have a clearly demonstrated association with increased rates of infant mortality and low birthweight. The imnportance of ensuring birth intervals of at least three years underscores the need for increased use of temporary contraceptive methods (paras. 5.20-5.22), since shorter intervals may pose a danger to the health of the mother and the survival of both the infant and older sibling(s). 2.49 Although maternal mortality data are not comparable to the fertility and mortality data available for India, in general, both maternal mortality rates and ratios across regions reflect the differentials evident in the fertility data. Maternal mortality  - 24 - rates are higher in high-fertility states because women are having more births. Maternal mortality is also higher in these states because more births are to high-risk mothers at young ages and older ages/high parity (as well as reflect poorer obstetric care). The continuing fertility decline results in declines in both maternal mortality rates and ratios, although the impact on rates is substantially greater than on ratios (Fortney 1987). 2.50 The proportion of women falling into low age, high age and high parity maternal risk categories is declining, and will continue to decline as fertility falls and the age at which women marry rises. However, the extent of childbirth at early ages, higher parity and later ages in some states continues to pose significant and avoidable maternal risks for Indian women as well as for their infants. Accelerated efforts to increase age at marriage and to reduce the number of high-parity births can more rapidly reduce the numbers of women at risk from these causes. Other Female Morbidity 2.51 If our knowledge of mortality and morbidity in pregnancy is inadequate and incomplete, for female morbidity generally it is almost non-existent. Most data on non- pregnancy-related morbidity among women are from hospitals, clinics, or selective or focused studies, and thus, do not provide information on morbidity rates in the population overall. The sections that follow deal, first, with diseases and conditions that are not gender-specific, and second, with those problems that are entirely or primarily confined to women. 2.52 Infectious and Other Diseases. The micro studies that are available generally paint a picture of high morbidity and malnutrition among girls and women (e.g., Chidambaram et al. 1986; Rao et al. 1973). Among children, a higher proportion of female morbidity is likely to be due to those diseases that are also major causes of death, such as respiratory infections (Devadas and Kamalanathan 1985) or diarrheal diseases (Cohen 1987; Pettigrew 1987). Several disorders emerge as widely prevalent among non-pregnant women: iron deficiency anemia; parasitic infestations; respiratory infections, including pulmonary tuberculosis and reproductive and urinary tract infections. In addition, some diseases, e.g., malaria, leprosy, filaria, are endemic in certain areas. 2.53 The few studies that compare the health among women and men in the same household report a higher prevalence of illness among women (Duggal and Amin 1989; Jesudason and Chatterjee 1979). However, a gender difference in the incidence of disease has not been found, leading some researchers to suggest that higher female morbidity is the result of less medical attention given to female illnesses (para. 3.19). 2.54 Both the gaps in the data and the higher disease burden of females are well illustrated by an examination of data on malaria in India (Chatterjee 1991). These data consist largely of numbers of cases detected and deaths reported by primary health centers (PHCs) on the basis of the number of blood smear examinations. Although these  - 25 - data are gathered for the purposes of malaria surveillance, they are not random. For passive case detection, blood smears are usually collected from patients visiting PHCs, among whom women are under-represented. For active case detection, outreach workers (usually male multipurpose workers) contact women in their homes, so that women are possibly, but not necessarily, over-represented. Some studies of the PHC data that give breakdowns by age and gender are available (Tewari et al. 1984). When slide positivity rates are computed, they are found to be higher among females than males (Ohlin 1984). Two interpretations of these findings are possible: (a) that malaria prevalence is indeed higher among females than males; or (b) that women who actually have malaria are more likely than men to come to PHCs, presumably because women would be brought to a PHC when illness is more advanced and they or their family members are more convinced of the need for treatment. In either case, the need for the health services to give increased attention to women's illnesses--in terms of data collection as well as patient identification and treatment--is clear. 2.55 Nutritional Deficiency. Protein-energy malnutrition is widespread among Indian girls, boys and women. This is documented in a variety of studies from all areas of the country, and both rural and urban populations (e.g., Basu 1989b; Srikantia 1989a; Gopalan 1985; Ghosh 1985; NNMB 1980a and 1980b). Poor nutritional status is already evident among females during infancy; it persists through childhood and tends to increase with age. For example, in a study of two Delhi slum populations, Basu found that only 50-65 percent of female infants below the age of 1 year were either normal or mildly malnourished, while for 5-9 year-old female children the percentage of normal and mildly malnourished fell still further, to about 30-55 percent (Table 2.25). 2.56 Anthropometric data also show that many women do not realize their full growth potential (NNMB 1980a and 1980b). For example, between 12 percent and 33 percent of 20-24 year-old women in surveyed states had heights below 145 cm, and between 15 percent and 29 percent weighed less than 38 kg. Below this height or weight women are at risk of obstetric difficulties and/or of delivering low birthweight babies. The percentages of younger women who are of low weight or height are higher because they may not have completed their adolescent growth spurt. Most women in north India become pregnant before reaching full maturity. This conveys considerable risk to mother and child. 2.57 The data on the relative nutritional levels of males vs. females, as indicated by various anthropometric measures, are more ambiguous than might be expected, given the strong indications of inequalities in food intake (paras. 3.14-3.16). A number of studies have found that malnutrition is more frequent and/or more serious in female children (Das Gupta 1987; Kielmann et al. 1978; Chen et al. 1981; Levinson 1974). However, it should be noted that the data available from the National Nutrition Monitoring Board (1980a and 1980b) did not, by and large, indicate that the nutritional status of female children was significantly worse than that of males. Other recent studies also have not reported a female disadvantage (Basu 1990; Srikantia 1989).  - 26 - 2.58 The National Nutritional Monitoring Board surveys (NNMB 1980a and 1980b) have documented low intakes of vitamin A and iron among girl children and adolescents. Vitamin A deficiency, which has been firmly linked to high mortality and morbidity in children, is also likely to be an underlying cause of high levels of respiratory and genito-urinary tract infections in women. With the onset of menarche, young girls are susceptible to anemia if dietary iron intakes fail to compensate for blood loss. Other studies also have established iron and calcium deficiencies among female children, adolescent girls and adult women (Harriss 1986). Iron deficiency anemia, in particular, is a major health problem of Indian women from an early age, and its widespread prevalence in India is well established. A nation-wide study found that over 95 percent of 6-14 year-old girls in the east (Calcutta) were anemic, around 65 percent in the south central area (Hyderabad), 57 percent in the northern areas (New Delhi), and about 20 percent in the south (Madras) (Table 2.26). The prevalence of anemia among women aged 15-24 and 25-44 years followed similar patterns and levels. The implications of anemia in pregnancy and its contribution to maternal death are discussed above (para. 2.43). Anemia also increases susceptibility to diseases such as tuberculosis and reduces energy for daily activities, whether household chores and child care or agricultural labor. A severely anemic individual is taxed by most physical activity, including walking at an ordinary pace. 2.59 Gynecological Problems. Of those health problems that are largely or entirely confined to women, the most obvious are gynecological disorders. Nationwide prevalence estimates are not available. The reproductive health problems of women have been studied at the community level even less than has general morbidity, due to the difficulties in diagnosing gynecological problems at a field level. A few population-based studies, influenced by clinical experience, have focused only on specific disorders, e.g., cervical or uterine cancer (Wahi et al. 1972; Garud et al. 1983; Mali et al. 1968), or vaginal discharges and genital infections (Bali and Bhujwala 1969). A study of women in two villages in Maharashtra (Bang et al. 1989) provides data on the prevalence of gynecological diseases, although the study was carried out in a tribal area, and therefore, may not be typical of rural Maharashtra, let alone of India generally. Some 55 percent of the women studied had 'gynecological complaints', the most prevalent was related to problems with menstruation, vaginal discharge or burning on urination. In addition, many complained of two non-specific but related symptoms, low back pain and lower abdominal pain. On clinical examination, astonishingly high levels of disease were identified: 92 percent of the women were found to have one or more gynecological or sexually transmitted diseases, with an average of 3.6 diseases per woman. Notably, almost all women who reported symptoms were found to have a gynecological disorder, but so did 85 percent of those who were symptom free. Half of this morbidity resulted from infections of the genital tract. The prevalence of disease found in the study is shown in Table 2.27. The most frequent specific problems were bacterial vaginitis (62 percent), cervicitis (49 percent) and dysmenorrhea (58 percent). Some 7 percent of the women had primary or secondary infertility, and 11 percent were infected with syphilis.  - 27 - 2.60 Of the women studied, only 8 percent had ever had a gynecological examination or treatment, an indication of both the women's failure to recognize their need for treatment and/or their inability to obtain it, as well as an apparent lack of awareness of the magnitude of the problem. There is a paucity of women doctors able to detect or treat gynecological problems among rural women who are particularly reluctant to approach male doctors about gynecological or sexual disorders. Female nurses and paramedical field staff are not trained to deal with these problems, contributing to the almost total absence of care. Laboratory capacity for diagnosis and treatment is scarce. 2.61 A disease burden, to the extent indicated by this study, inevitably results in a host of complications for the women: difficulties in occupational and domestic work (chronic backache was reported by over 30 percent); fetal wastage (miscarriages and stillbirths); infections of their newborns, which are acquired in the birth canal; sterility; sexual disorders; anxiety; and stress. In addition, the study indicated a significant association between use of contraception and subsequent gynecological problems, such as menstrual and cervical disorders and pelvic inflammatory disease. The fact that a woman's ability to contracept can be jeopardized by pre-existing gynecological disorders that are exacerbated by the use of contraceptives links women's general health problems to maternal health and underlines the importance of improving the ability of the health services to meet women's reproductive health needs beyond those associated with pregnancy and delivery (para. 5.26). HIV/AIDS is a rapidly emerging problem which will significantly affect the health of Indian women and require increasingly greater attention by the health system. 2.62 Back pain, a symptom regarded as 'non-specific' by the medical profession, has been called 'the feminine affliction' and traced to a variety of conditions and underlying causes (Shatrugana et al. 1990). The authors based their analysis on a study of women hospitalized for fractures. The majority of fractures were related to osteoporosis, a condition resulting from calcium and related nutritional deficiencies and associated with menopause. While most of the fractures among women over 40 were the result of minor trauma or occurred 'just during walking,' among younger women 18-39, most occurred as a result of work-related accidents. The authors concluded that several factors contributed to these problems: (a) the upbringing of young girls, which requires the adoption of postures that are detrimental to bone integrity; (b) early and continuous nutritional deprivation among women; (c) early and repeated pregnancies and years of lactation; and (d) engagement in work which entails poor posture, is high risk and/or is also low-paying and does not allow women to improve their diets. 2.63 Occupational Health Problems. Little information is available on the occupational health problems of women because the great majority are employed in the unorganized sector in a large variety of occupations. The majority of poor rural women work as agricultural laborers, which exposes them to specific health difficulties. Long hours spent standing in water while weeding and transplanting rice, essentially a female occupation, increases susceptibility to vaginal infections, infectious and parasitic diseases,  - 28 - insect bites, arthritis, and rheumatism. Back pain and osteoarthritic complaints are also common among women undertaking harvesting tasks and other stoop agricultural labor. Most agricultural and vending tasks, as well as domestic tasks, such as fuelwood and water collection, involve carrying heavy headloads which can cause spinal problems. 2.64 Cooking, an almost exclusively female occupation that takes up a substantial portion of most women's time, is particularly hazardous to Indian women because of cultural constraints which require that kitchens be located indoors. Kitchens are rarely adequately vented, so that women are commonly exposed to levels of smoke particles and pollutants which are hazardous to health. Concentrations of such pollutants averaging one hundred times the level deemed acceptable by the WHO were measured in rural kitchens in one study (Smith et al. 1983). Wood smoke causes lung problems which, in turn, place a strain on the heart. The incidence of this problem was found to be the same among men and women by a 15-year hospital-based survey. However, its etiology was traced to tobacco smoking among men, in contrast to kitchen smoke inhalation among women. The age of onset of the disease was lower among women, and nearly all the women affected were from low-income groups. As noted above, respiratory diseases are a leading cause of death among women and girls over 5 years of age. There is also evidence linking impaired fetal development, low birthweight, and perinatal death to maternal exposure to pollutants, particularly in the presence of anemia, an almost universal condition among Indian women. 2.65 The health hazards of some additional occupations can be illustrated briefly. For example, women working in the carpet industry suffer from ankyloses and chronic postural defects which may result in difficult pregnancies or even subsequent sterility. Workers who roll bidis (indigenous cigarettes) are exposed to tobacco dust and are susceptible to problems such as tuberculosis, asthma, allergies, backaches, and rheumatic complaints. In coir, jute and cashew-nut processing, cotton and tea plucking, rubber tapping, and the textile industry, exposure to toxic chemicals and physical stress is substantial. Workers in the garment and embroidery industries complain of chronic back pain and eye problems due to poor physical and lighting conditions in the work environment (Chatterjee 1987; Ghosal and Chakraborti 1987). All of these problems can be further aggravated by malnutrition, anemia, frequent childbearing, and/or long working hours. 2.66 Social Health Issues. Indian women are exposed to crimes such as rape, burning and beating at unusually high rates. Although these problems almost certainly contribute to 'accidents and injuries' being the leading causes of death among women aged 15-34 (Table 2.16), specific data on their incidence and resultant morbidity are conspicuously absent. Whether accidental or intentional, burns are a major cause of hospital admissions in urban hospitals (Karkal 1985). High rates of alcoholism among men, common among poor populations in tribal, rural and urban India, can also be seen as a woman's health problem because of the link between alcohol consumption and domestic violence.  - 29 - III. THE HEALTH CONTEXT A. THE SOCIOCULTURAL CONTEXT 3.1 In India, where families live in poverty and where health infrastructure is poor, males as well as females suffer. However, there are particular risks that women face because of their reproductive biology, and maternal mortality rates in India are among the world's highest. Furthermore, age and gender specific mortality rates indicate that girls and women under 30 years of age are additionally disadvantaged by sociocultural factors. Significant disparities exist between the north and south. In much of southern India, impressive progress has been made in improving education and health for the population overall, particularly for women. However, the female mortality disadvantage relative to males not only persists but is worsening in some northern states. For this reason, and because the northern states comprise a major portion of India's population, the discussion that follows focuses on north India. The Status of Women 3.2 In traditional societies, personal and household circumstances as well as access to key social resources are structured, to a great extent, by family, kin and marriage relationships. These can vary greatly--in fact, the kinship systems of Hindu north India, Hindu south India and Muslim India represent three quite different types of systems. What is common to all of these systems is that they define gender in social terms. That is, what is considered to be the proper role, function and behavior of women, and who controls their productive and reproductive capacities, are integral aspects of larger social systems. 3.3 The position of women (their 'status') in these systems can be assessed in terms of two related sets of criteria (Basu 1989a; Dyson and Moore 1983). First, the possibilities available to women are conditioned by their exposure to and interaction with the outside world. Second, the situation of women is further determined by their capacity to make decisions both inside and outside their households (what has been termed their 'autonomy'), which includes their ability to (a) control their own physical movements; (b) acquire, retain and dispose of earnings and/or property; (c) control to some extent their reproductive careers (e.g., the choice of a husband and use of contraception); and (d) associate with their natal kin. 3.4 In the light of these criteria, the position of north Indian women is notably poor. Traditional Hindu society in rural north India, is hierarchical, patrilineal, patrilocal, and strongly rnale dominated. The implications of this for women can be seen in terms of marriage. Ideally, for north Indian Hindus marriage within socially acceptable (i.e., caste) boundaries requires a woman be married to someone to whom she is not related on either her father's or her mother's side, and who lives outside her natal village. Marriages are alliances in which young women and men have no say. Wife-  - 30 - givers are socially and ritually inferior to wife-takers, necessitating the provision of dowry. Conversely, because girls will require a dowry when they are married and thereafter will be lost to their natal families, female children are generally considered more of a burden to their parents than sons. 3.5 Thus, north Indian Hindu brides entering their husbands' households are strangers in a strange place9. They are controlled by older females in the household and their behavior reflects on the honor of their husbands and the larger patrilineal group. Restrictions on their personal movements often amount to their seclusion ('purdah'). Additionally, emotional ties between spouses are considered to constitute a potential threat to the solidarity of the patrilineal group. Hence, the northern system is associated with the segregation of the sexes in general and with limited communication between spouses in particular, a circumstance that has direct consequences for the adoption of family planning and other 'modem' health behavior. And, as is typical of such cultural traditions, the norms are internalized by those who are disadvantaged by them. That is, a young Indian bride is socialized to believe that her own wishes and interests are subordinate to those of her husband and his family. In such circumstances, a newly married young woman's primary duty, and virtually her only means of improving her position in the hierarchy of her husband's household, is to bear sons. 3.6 The traditional circumstances of women in north India are brought into clearer focus by comparison with the situation in south India. In south India, the preferred customary marriage of a daughter is to her mother's brother (i.e., to her maternal uncle), or failing that, to her mother's brother's son (i.e., her cross-cousin). Even if the preferred kinsman or woman is not available, others nominally standing in desired relationships will be given preference. Such marriages would be unthinkable in north India. The consequences of these contrasting marriage systems for women are substantial. 3.7 In south India, since men are likely to marry women to whom they are related, either actually or nominally, the strict north Indian distinction between patrilineal and marital relatives is not found. Men are as likely to have social, economic and/or political ties with other men to whom they are related by marriage as by descent, and women are likely to be married into familiar households near to their natal homes. There is no ritual or social distinction between the families of the bride and bridegroom as there is in the north. Marriages have typically been characterized by bride price rather than dowry, although any large marriage exchange among close relatives is considered inappropriate. Since no premium is put on controlling their physical movements, women are much more likely to retain close relationships to their natal kin, and affective ties between spouses are culturally accepted. 3.8 It should be noted that over the last several decades marriage patterns in south India have changed significantly and rapidly. In consequence of a set of social, economic and demographic changes, the number of marriages among close relatives is  - 31 - declining and bride price has given way to a dowry system akin to that in the north (Caldwell et al. 1988)'