Abortion in the developing world
Publication date: 1999
Abortion in the Developing World t editors AKel I. Mundigo Cynthia lndriso e ~ World liealth Organization (?',e) Vlstaar Publlcatlons (A dlVlslOn of Sage Publlcatlons lndla Pvt Ltd) New Delhl Copyright © World Health Organization, 1999 All rights reseIVed by the World Health Organization (WHO). This publication may not be reviewed, reproduced or translated, in part or in whole, without the written permission of the WHO. No part of this publication may be stored in a retrieval sys- tem or transmitted in any form or by any means-electronic, mechanical or other- without the prior permission of WHO. The editors and authors are responsible for the views expressed in this publication. The mention of specific companies or of certain manufactur,ers' products does not imply that they are endorsed or recommended by the World Health Organization in preference to others of similar nature that are not mentioned. First published in 1999 by Zed Books, London and simultaneously by Vistaar Publications (A division of Sage Publications India Pvt Ltd) M-32 Market, Greater Kailash - I New Delhi -110 048 For copyright reasons, this edition is for sale in India and South Asia only. Published by Tejeshwar Singh for Sage Publications India Pvt Ltd, phototypeset by Line Arts, Pondicherry and printed at Chaman Enterprises, Delhi. ISBN: 81-7036-743-3 (India-HB) Sage Production Team: Jaya Chowdhury, N.K. Negi and Santosh Rawat To the women, all over the world, who unselfishly agreed to be Interviewed and discuss a highly sensitive personal issue Axel I. Mundlgo Contents List of Tables List of Figures Foreword by Mahmoud F. Fathalla Preface Acknowledgements Chapter 1 Introduction Cynthia Indriso andAxel I. Mundigo PART I: The Relationship between Abortion and Contraception Chapter 2 Prevention of Pregnancy in High-Risk Women: 10 15 17 19 21 23 Community Intervention in Chile 57 Ramiro Molina, Cristian Pereda, Francisco Cumsille, Luis Martinez Oliva, Eduardo Miranda and Temistocles Molina Chapter 3 Factors Affecting Induced Abortion Behaviour among Married Women in Shanghai, China 78 GuiShi-xun Chapter 4 First-Trimester Induced Abortion: A Study of Sichuan Province, China 98 Luo Lin, Wu Shi-zhong, Chen Xiao-qing and Li Mirt-xiang Chapter 5 Abortion Practice in a Municipality of Havana, Cuba 117 Luisa Alvarez lilsquez, Caridad Teresa Garcia, Sonia Catasus, Maria Elena Benitez and Maria Teresa Martinez Chapter 6 Social Determinants of Induced Abortion in the Dominican Republic 131 Denise Paiewonsky Chapter 7 The Use of Induced Abortion in Mauritius: An Alternative to Fertility Regulation or an Emergency Procedure? 151 Geeta Oodit and Uma Bhowon Chapter 8 Determinants of Induced Abortion and Subsequent Reproductive Behaviour Among Women in Three Urban Districts of Nepal 167 A.K Tamang, Neera Shrestha and Kabita Sharma 8 t Contents Chapter 9 Cultural and Psychosocial Factors Affecting Contraceptive Use and Abortion in Two Provinces of Turkey 191 AyseAkin PART II: Quality of Abortion Care Section A: Women's Perspectives Chapter 10 Determinants and Medical Characteristics oflnduced Abortion Among Poor Urban Women in North-East Brazil 217 Chizuru Misago and Walter Fonseca Chapter 11 Induced Abortion and the Outcome of Subsequent Pregnancy in China: Client and Provider Perspectives 228 Zhou Wei-jin, Gao Er-sheng, limg lilo-ying, Qin Fei and Tang Wei Chapter 12 Between Political Debate and Women's Suffering: Abortion in Mexico 245 Maria def Carmen Elu Chapter 13 Abortion Services in Two Public Sector Hospitals in Istanbul, Turkey: How Well Do They Meet Women's Needs? 259 Aysen Bulut and Nahid Toubia Section B: Provider Perspectives Chapter 14 The Attitudes of Health Care Providers Towards Abortion in Indonesia 281 E. Djohan, R. Indrawasih, M. Adenan, H. Yudomustopo andM.G. Tan Chapter 15 Pharmacists and Market Herb Vendors: Abortifacient Providers in Mexico City 293 Susan Pick, Martha Givaudan, Suzanne Cohen, Marse/a Alvarez and Maria Elena Collado Chapter 16 Induced Abortion in a Province in the Philippines: The Opinion, Role and Experience of Traditional Birth Attendants and Government Midwives 311 Fred V. Cadelina Chapter 17 Induced Abortion in Sri Lanka: Opinions of Reproductive Healt.h Care Providers 321 P.Hewage PART III: Adolescent Sexuality and Abortion Chapter 18 Induced Abortion Among Unmarried Women in Sichuan Province, China: A Survey 337 Luo Lin, Wu Shi-zhong, Chen Xiao-qing and Li Min-xiang Chapter 19 Sexuality, Contraception and Abortion Among Unmarried Adolescents and Young Adults: The Case of Korea 346 Kwon Tai-hwan, Jun Kwang Hee and Cho Sung-nam Contents t 9 Chapter 20 Female Adolescents at the Crossroads: Sexuality, Contraception and Abortion in Mexico 368 N. Ehrenfeld Chapter 21 Induced Abortion in Dar es Salaam, Tanzania: The Plight of Adolescents 387 G.S. Mpangile, MT. Leshabari and D.l Kihwele PART IV: Research and its Implications for Policy: Conclusions Chapter 22 The Incidence and Social and Demographic Characteristics of Abortion in Colombia 407 Lucero Zamudio, Norma Rubiano and Lucy Wartenbe,g Chapter 23 Abortion in Ljubljana, Slovenia: A Method of Contraception or an Emergency Procedure? 447 Dunja Obersne/ Kveder Chapter 24 Research Methodology: Lessons Learnt 465 Axel L Mundigo Chapter 25 Policy Impact of Abortion Research 477 Axel L Mundigo and Iqbal H. Shah The Contributors 489 Index: 494 About the Editors 499 List of Tables 2.1 Sample Characteristics for Each Community 63 2.2 Abortion Risk Factors 65 2.3 Actual and Predicted Abortion Values 66 2.4 Abortion Risk Factors: Women Included in the Second Survey Compared with Women Excluded from the Analysis 68 2.5 Changes in Contraceptive Use and Pregnancy Status between the First and Second Surveys, by Community 70 2.6 Abortion Rates: AH Women and High-Risk Women, Before and After the Intervention 71 2.7 Abortion Ratios: All Women and High-Risk Women, Before and After the Intervention 71 2.8 Abortion Rates among High-Risk Women, by Contraceptive Status: before and After the Intervention 73 2.9 Abortion Ratios among High-Risk Women, by Contraceptive Status: Before and After the Intervention 73 3.1 Married Women by Age, Shanghai, 1991 80 3.2 Married Women by Education, Shanghai, 1991 81 3.3 Married Women by Occupation, Shanghai, 1991 81 3.4 Reasons for Pregnancy Tormination in the Past (%) 83 3.5 Main Reasons for Pregnancy Termination, by Age(%) 84 3.6 Main Reasons for Pregnancy Termination, by Education (%) 84 3.7 Main Reasons for Pregnancy Tormination, by Occupation(%) 86 3.8 Current Pregnancy Causes, by Type of Working Unit(%) 87 3.9 Main Reasons for Pregnancy Tormination, by Couples' Leisure Time (%) 88 3.10 The 10 Major Reasons for Not Using Contraception at the Time of the Current Pregnancy 90 3.11 The 10 Major Reasons for Contraceptive Failure for the Current Pregnancy 93 4.1 Population of the Six Counties, 1991 101 4.2 Age Distribution of a Study Sample 104 4.3 Reasons to Request Induced Abortion 105 4.4 Women Who Experienced Contraceptive Failure, by Method 106 45 Blood Loss During Operation 107 4.6 State of Bleeding< 15 Days Post-Operation 108 List of Tables t 11 4.7 Short-Torm Fever 15 Days After Abortion 108 4.8 Total Score by CES-D at Admission and at Time Intervals After Abortion 109 4.9 Percentages of Depression Before and After Induced Abortion 109 4.10 Factor Scale Obtained Two Days Before Abortion 110 4.11 Factor Scale Obtained 180 Days After Abortion 110 5.1 Summary of the Main Characteristics of the Population Studied 119 5.2 Age-Specific Fertility and Abortion Rates 121 5.3 Age-Standardized Fertility and Abortion Rates by Education, Occu- pation, Marital Status and Origin of the Partner 122 5.4 Contraceptive Knowledge 126 5.5 Contraceptive Use by Abortion Practice 127 5.6 Abortion Experience by Contraceptive Use Status 128 6.1 Characteristics of the Study Population 135 6.2 Percentage Distribution of the Study Sample by Work Status, Type of Work and Contraceptive Experience 137 6.3 Percentage Distribution of the Sample According to Selected lndica- tors of Contraceptive Use 139 6.4 Age-Specific Percentage Distribution of Women Who Had Used Methods, by Total Time Use and Methods Used 141 6.5 Percentages of Ever-Users of Contraceptives by First Method Used, Second Method Used and Method Used at the Time of Last Preg- nancy, by Specific Methods 142 6.6 Percentage Distribution of Women Who Have Used More Than One Method, According to Reasons for Discontinuation of Both First and Second Methods Used 143 6.7 Percentage Distribution of the Sample's Desire to Have More Children, According to Use of Contraception at the Time of Last Pregnancy and Intention to Use Contraception 144 8.1 Study Population, by Type of Abortion and Place of Recruitment 169 8.2 Pregnancy Status, by 'Iype of Abortion 174 8.3 Women Who Had Induced Abortion, by Number of Living Sons and 'Iype of Provider 175 8.4 Women Who Had Induced Abortion: Socioeconomic Characteristics and 'Iype of Provider 177 8.5 Women's Contraceptive Intentions, by Type of Abortion 179 8.6 Contraceptive Choices among Women Planning to Contracept, by Abortion Status 179 8.7 Prior Abortion Experience, by Current Abortion Status 180 8.8 Induced Abortion Cases: Post-Abortion Health Status at the Time of Second Follow-up 181 8.9 Induced Abortion Cases: Post-Abortion Ever-Use of Contraception, by Method 182 8.10 Induced Abortion Cases: Current Use of Contraception at the Time of Second Follow-up 183 8.11 Induced Abortion Cases: Reasons for Non-Use of Post-Abortion Contraception 183 12 t List of Tables 8.12 Spontaneous Abortion Cases: Post-Abortion Health Status at the Time of Second Follow-up 185 8.13 Spontaneous Abortion Cases: Post-Abortion Ever-Use of Contracep- tion, by Method 186 8.14 Spontaneous Abortion Cases: Current Use of Contraception at the Time of Second Follow-up 186 8.15 Spontaneous Abortion Cases: Reasons for Non-Use of Post-Abortion Contraception 187 9.1 Some Sociodemographic Characteristics of Women in Ankara and Van Provinces (in percentage) 195 9.2 Contrae<;ptive Knowledge by Province and Gender 198 9.3 Contraceptive Use Among Women in Ankara and Van Provinces (in percentage) 199 9.4 Education of Men and Women and Contraceptive Practice, Ankara and Van (in percentage) 201 9.5 Women's Work Status by Contraceptive Use and Abortion, Ankara and Van 202 9.6 Women's Self-Esteem by Contraceptive Use and Abortion, Ankara and Van 204 10.1 Sociodemographic Characteristics of Women Admitted for Induced Abortion Complications 221 10.2 Prior Pregnancy Outcomes of Women Admitted for Induced Abor· tion Complications 222 10.3 Distribution of Women Admitted for Induced Abortion Complica- tions, by Medical Characteristics 223 10.4 Distribution of2,083 Induced Abortion Cases, by Contraceptive Method Use at Month of Conception 224 10.5 Distribution of 2,074 Induced Abortion Cases, by Women's Own Statement of Method 224 13.1 Percentage Distribution of Currently Married Women by Contracep- tive Methods Currently Used, Turkey, 1988 and 1993 260 13.2 Summary of the Main Characteristics of the Women in Our Study 265 13.3 Contraceptive Knowledge and Use Among the Women Attending Both Hospitals 267 13.4 Use of Contraception Before Last Pregnancy 267 13.5 Method in Use When the Last Pregnancy Occurred Among Users 268 13.6 Quality of Care 269 13.7 Factors Relating to Satisfaction (Logistic Regression Analysis) 274 13.8 Total Cost Distribution of Pregnancy Termination Services in Both Hospitals 275 14.1 lype and Distribution of Respondents 285 15.l Summary Characteristics of the Study Group 296 15.2 Substances Recommended as Abortifacients in Role-Playing Visits 299 15.3 Medications Recommended as Abortifacients in Role-Playing Visits 299 16.1 Profile of the Birth Attendants in Our Study 313 16.2 Knowledge of Induced Abortion Methods and Methods Known (%) 315 16.3 Opinion on Induced Abortion and Reasons (%) 316 16.4 Opinion on When Induced Abortion May Be Allowed (%) 316 List of Tables t 13 16.5 Response of Birth Attendants to Women Asking for Assistance to Terminate an Unwanted Pregnancy (%) 317 16.6 The Experience of Abortion Among the Birth Attendants Themselves 318 16.7 Knowledge of Abortion Providers by Birth Attendants 319 17.1 Study Sample by Age and Profession 325 17.2 Provider Perceptions of Why Women Seek Abortion 326 17.3 Provider Perceptions of the Increase in Abortion Practice 327 17.4 Provider Perceptions of the Prevalence of Legal, Self-Induced, Unsafe, Safe and Late Abortions 327 17.5 Provider Perceptions of How People Feel About Available Contra- ceptives 329 17.6 Provider Opinions on Strategies to Reduce Unsafe Abortion 330 17.7 Provider Opinions on Best Abortion Law for Sri Lanka 332 18.1 Percentage Distribution of Sample by Age, Gestation Period, Educa- tion, Occupation, Income and Residency 340 18.2 Percentage Distribution of Sample by Contraceptive Use 341 18.3 Percentage Distribution of Sample by Number of Previous Preg- nancies and Abortions 341 18.4 Percentage Distribution of Sample by 'Ilaining and Experience of the Operator 342 18.5 Blood Loss During Operation 343 18.6 Physical Examination 15 Days Post-Operation 343 19.1 Sociodemographic Profile of the Women in Our Study 351 19.2 First Sexual Experience 354 19.3 Contraceptive Knowledge and Use at First Intercourse 356 19.4 Past and Current Contraception Use, by Method 357 19.5 Contraceptive Failure 357 19.6 Pregnancy History and Contraceptive Use Among the Sexually Active Women 358 19.7 First Abortion Experience 359 19.8 Logistic Regression of All Abortion Experiences 360 19.9 Logistic Regression of Post-Abortion Complications 362 19.10 Approval of Abortion (Full Study Sample) 363 20.1 Summary Characteristics of the Study Group 373 21.1 Demographic Characteristics of the Women (%) 390 21.2 Contraceptive Knowledge and Use, by Method (%) 393 21.3 Marital Status of the Women and Type of Relationship (%) 393 21.4 Reaction/ Advice of First Confidant (%) 395 21.5 Where and By Whom Abortions Are Done (%) 397 21.6 Advice/Treatment by the Abortionist (%) 397 22.1 Women, Aged 15-55 with Abortion Experience, as a Proportion of All Women Pregnant at Least Once, by Order of Abortion 419 22.2 Average Number of Pregnancies and Abortions for Women with Abortion Experience and Abortion Ratios, by Main Differentials, Urban Colombia, 1992 431 22.3 Annual Abortion Rates per 100 Women: Total and by Order of Abortion 443 14 t List of Tables 23.1 Characteristics of the Women in the Study Sample Compared to Women Having Abortion at the UDGO in LJubljana and in Slovenia, 1988 452 23.2 Contraceptive Use Before the Conception, by Education, Number of Children, Perception of the Method Used and Abortion Experience 453 23.3 Multiple Regression of Contraceptive Use Before Conception 455 23.4 Contraceptive Use Before Conception and at Conception, by Type of Contraception 456 23.5 Multiple Regression of Contraceptive Use at the Time of Conception 457 23.6 Discriminant Analysis of Individual Variables in the Three Groups of Contraceptive Users 459 23.7 Discriminant Analysis of Attitudes Towards Abortion in the Three Groups of Contraceptive Users 461 t List of Figures 4.1 Mean Level of SCL-90 Symptomatic Profile 113 22.1 Proportion of Women, 15-55 Years, With Abortion Experience, by Region, Urban Colombia, 1992 414 22.2 Proportion of Women, 15-55 Years, With Abortion Experience, by Education, Urban Colombia, 1992 416 22.3 Women, 15-55 Years, With Abortion Experience, by Age, Urban Colombia, 1992 417 22.4 Women, 15-55 Years, With Abortion Experience, by Number of Children, Urban Colombia, 1992 418 22.5 Women, 15-55 Years, Proportions Ever Pregnant and Who Aborted, Urban Colombia, 1992 422 22.6 Proportion of Women, 15-55 Years, With Abortion Experience, by Social Class, Urban Colombia, 1992 423 22.7 Women, 15-55 Years, With Abortion Experience, by Number of Preg- nancies, Urban Colombia, 1992 424 22.8 Women, 15-55 Years, With Abortion Experience, by Number of Preg- nancies and Region, Urban Colombia, 1992 426 22.9 Women, 15-55 Years, With Abortion Experience, by Number of Preg- nancies and Social Class, Urban Colombia, 1992 427 22.10 Women, 15-55 Years, With Abortion Experience, by Number of Preg- nancies and Education, Urban Colombia, 1992 428 22.11 Women, 15-55 Years, With Abortion Experience, by Number of Preg- nancies and Work Status, Urban Colombia, 1992 430 22.12 Women, 15-55 Years, With Abortion Experience, by Number of Preg- nancies and Abortions, Urban Colombia, 1992 434 22.13 Women, 15-55 Years, With Abortion Experience, by Number of Abortions and Children Ever-Born, Urban Colombia, 1992 435 22.14 Women, 15-55 Years, With Abortion Experience, by Number of Abortions and Education, Urban Colombia, 1992 436 22.15 Women, 15-55 Years, With Abortion Experience, by Number of Abortions and Age, Urban Colombia, 1992 437 22.16 Women, 15-55 Years, With Abortion Experience and Three or More Pregnancies, According to the Risk of Moving From One Abortion Level to the Next, Urban Colombia, 1992 438 16 t List of Figures 22.17 Age-Specific Abortion Rates, by Year of Abortion 440 22.18 Age-Specific Fertility and Abortion Rates, Urban Colombia, 1992 441 22.19 Age-Specific Abortion Rates by Regions, Urban Colombia, 1992 442 t Foreword I can never forget an encounter in my early professional practice, more than 35 years ago. I was on clinical duty in the emergency room, when a young woman was admitted in extreme distress and agony. On pelvic examination, it was with a sense of shock that I found her intestines in the vagina. It turned out that she had had a botched abortion, during which the uterus was perfo- rated and the intestines were mistaken for the products of conception and pulled down into the vagina. Her life was saved by emergency surgery but her uterus had to be removed. When she recovered, I ventured very gently to ask her why she did that to herself. I still recall hearing her weak voice answering with another question: Do you doctors understand what it means to a woman to have an unwanted pregnancy? It was only recently that we began trying to understand. The abortion issue has been the subject of intensive debate among men: legal scholars, moralists, men of religion and politicians. The voices of women have been drowned in this loud debate despite the fact that it is women whose bodies, psyche, health and life are directly concerned. Every- day, hundreds of women lose their lives in the process of trying to terminate an unwanted pregnancy under unsafe circumstances. Few studies have ven- tured to explore why women are pushed to take this dangerous course. It took conviction and courage to consider an international initiative to improve our understanding about the problem of unsafe abortion, its under- lying causes and its consequences in women's lives. The World Health Orga- nization was best suited to take this initiative, with its assets of objectivity, concern about health and the trust of member states. The Social Science Research Unit of the Special Programme of Research, Development and Research Training in Human Reproduction had the professional expertise and the reputation for research excellence essential for the success of such a major research initiative. The initiative was announced in 1989. What came as a pleasant surprise to many of us at the time was the vigorous response of the international scien- tific community, and the positive attitude of member states across all cul- tures. The Unit was overwhelmed by good research proposals from all regions of the developing world. 18 t Foreword The debate in the Cairo International Conference on Population and Development in 1994 and in the Beijing Fourth World Conference on Women in 1995 finally highlighted the need 'to recognize and deal with the health impact of unsafe abortion as a major public health concern'. The debate was not easy. Some people seem to think that if you move yourself away from a problem that you dislike, shut your ears to the sounds of pain, and close your eyes to the sights of sufferings, the problem will no longer exist. For someone concerned about women's health, and as a former director of the Special Programme, it is gratifying to see the outcome of these case stud- ies on abortion enter into the public domain. I hope it will improve the under- standing, particularly among policy makers, about the plight of women faced with difficult reproductive health choices, when they are not empowered to make decisions that respect their dignity and save their lives and their health. MAHMOUD F. FATHALLA Senior Advisor, Biomedical and Reproductive Health Research The Rockefeller Foundation; and Former Director of the UNDP/UNFPA/WHO/World Bank Special Programme of Research, Development and Research Training in Human Reproduction, World Health Organization Preface Unsafe induced abortion, particularly in countries where the practice is ille- gal, is an important contributor to reproductive ill-health and to maternal deaths, all events that are fully preventable. In developing countries where abortion is legal, the services are often concentrated in urban centres and unsafe abortion continues to exist in rural and remote areas. For women who have access to legal abortion services, induced abortion is often used as a back-up for contraceptive failure or as an alternative to effective contracep- tion, even when family planning is readily available. In other countries where family planning access is limited and health service infrastructure generally poor, induced abortion may be the only alternative for women who want to space their pregnancies or do not wish to have another child. The reasons for resorting to abortion vary from one context to another across developing as well as developed countries. The WHO Special Programme of Research, Development and Research Training in Human Reproduction, aware that induced abortion was a major reproductive health problem and the cause of untold suffering for many women around the world decided, early in 1989, to undertake a pioneering effort to understand the determinants or reasons why women resort to abor- tion in various cultural, social and service availability contexts. Another aspect of induced abortion for which information was lacking concerned the consequences of unsafe induced abortion for women, especially for their health status and emotional wellbeing. To this effect, through its Task Force for Social Science Research on Reproductive Health, the Special Programme announced a research initiative on the determinants and consequences of induced abortion aimed at developing countries. The Special Programme of Research in Human Reproduction was in a par- ticularly strong position to undertake research on such a sensitive issue- being an integral part of WHO lent the right health profile to the initiative and being a scientific research programme of worldwide renown made it pos- sible to deal with induced abortion above the usual controversies caused by religious and political ideology. In fact, when the plans for the research initia- tive were presented to the World Health Assembly, in May 1989, not a single country opposed it. , 20 t Preface The announcement of the research initiative on induced abortion coin- cided with an External Evaluation of the Special Programme which among other recommendations stated that 'reproductive health research must be placed in its social, cultural and behavioural context' noting also that 'these aspects must be given greater attention in the future'. These recommenda- tions further endorsed the importance of the initiative whose objectives were precisely to understand women's reproductive health behaviour within their own social and cultural context and, especially, to gain a better insight into the relationship between abortion and contraceptive behaviour. t Acknowledgements This book would not have been possible without the cooperation of women all over the world who volunteered to answer questions about one of the most intimate dilemmas of their reproductive lives-the decision and circum- stances under which they opted to end a pregnancy. Our thanks go to them as well as to the researchers, in particular those who wrote the chapters for this book, and their committed interviewers who often worked under extremely ~ifficult field circumstances, asking questions that required tact and patience, devotion and understanding. Our thanks go also to the Special Programme of Research, Development and Research 1raining in Human Reproduction of the World Health Organization that provided the financial and technical means to carry out the research, analyze the data, and write up the results. The Colegio de Mexico, a leading graduate training and research centre in Mexico City was host to the first Workshop on Research on Induced Abor- tion, held from 13 to 17 November 1989, that marked the beginning of the research initiative by congregating the principal investigators of the winning projects to discuss issues ofresearch planning and methodology pertaining to their forthcoming work. We are very grateful to the authorities of the Colegio de Mexico who provided us with an ideal academic environment and the nec- essary support to hold a very successful workshop. Similarly, at the conclu- sion of the research, the Universidad Externado de Colombia hosted a major conference, from 15 to 18 November 1994, that allowed the Latin American researchers to present their results in an international forum and share their experience with representative policy makers. We would like to express our gratitude to the rector and staff of the university for their unfailing support. While it would be difficult to name all the persons who at various stages contributed to this research initiative, it would be appropriate to name the technical advisors that participated in the Colegio de Mexico workshop and provided excellent lectures and research guidance to the participants: Dr Lourdes Arizpe, Colegio de Mexico, Mexico; Dr Wendy Baldwin, National Institute of Child Health and Human Development, Washington, D.C., USA; Mrs Margaret Bone, Office of Population and Surveys, London, England; Lie Gustavo Cabrera, Colegio de Mexico, Mexico; Dr Mercedes Concepcion, 22 t Acknowledgements University of the Philippines, Manila, Philippines; Dr Tomas Frejka, then Population Council, Mexico; Dr Sylvia Hartman, Pan American Health Organization, Mexico; Dr Lorenzo Moreno, Office of Population Research, Princeton University, USA; Dr Yolanda Palma, Ministry of Health and Social Assistance, Mexico; Dr Thomas Pullum, University of Texas, Austin, USA; and Dr Erica Taucher, National Institute of Nutrition, Santiago, Chile. Special thanks go also to Dr Iqbal Shah, Special Programme of Research in Human Reproduction, Geneva, Switzerland, who not only attended the Mex- ico workshop but also provided substantive advice over the several years of the initiative's progress and supervised the last details leading to the publica- tion of this book. Mr Jitendra Khanna who dealt directly with the publishers and followed up the process of publication also deserves special thanks. Ms Maud Keizer, the secretary of the Special Programme's Social Science Research Unit, who typed the manuscript and many difficult tables, also deserves our gratitude. THE EDITORS t 1 Introduction Cynthia lndriso and Axel I. Mundigo Abortion-Related Morbidity and Mortality: The Global Picture Induced abortion is an ancient practice, experienced by women of all back- grounds in every part of the world. Among the issues related to reproductive health, none has more controversial connotations than abortion nor carries a heavier burden of stigmatization, including moral and religious condemna- tion. Its exact incidence, therefore, as well as that of abortion-related mortal- ity and morbidity is still difficult to establish. Nonetheless, the data that are available demonstrate that induced abortion is very prevalent in the developing world, despite the fact that contraceptive prevalence rates have increased dramatically in the last 30 years. Although the frequency and distribution of abortion vary, there is no country where abortions do not occur. Recent estimates for developing world regions pro- vide an overall figure of around 30 million induced abortions annually, bro- ken down as follows: 3.4 million in Africa (high range estimate); 11.9 million in East Asia; between 5.2 and 12.5 million in South and South-East Asia; and between 4.4 and 6.2 million in Latin America. This means that of all births averted by either contraception or abortion, up to one-third are averted by induced abortion in Africa; up to 22 per cent in East Asia; between 11-23 per cent in South and South-East Asia; and between 21-28 per cent in Latin America (Frejka, 1993). Many estimate that the incidence of induced abortion is increasing world- wide. Reasons for this increase are attributed to a variety of changing trends worldwide, including a desire for smaller families, shifts from rural to urban res- idence, and the increase in non-marital sexual activity (Coeytaux et al., 1993). The Need for Safe Abortion In many countries in the developing world, induced abortion is illegal and therefore, largely unsafe. Of the 20 million unsafe abortions that occur each 24 t Cynthia lndriso and Axel I. Mundlgo year in the world, some 70,000 result in death, representing a case fatality of 0.4 per 100 unsafe interventions. The highest levels of case fatality are in Africa (0.6) and Asia (0.4), with a much lower rate in Latin America (0.1). Similarly, the risk of dying from an unsafe abortion is much higher in the developing world, 1 in 250 procedures, than in the developed world, where the figure is only 1 in 3,700 procedures. In fact, niost of the unsafe abortions that occur in the world today, close to 90 per cent, occur in developing coun- tries (WHO, 1994). Death is not the only tragic cost of unsafe abortion. Many more women survive the experience, only to suffer lifelong consequences of serious complications. Sepsis, haemorrhage, uterine perforation, and cervical trauma often lead to problems of infertility, permanent physical impairment and chronic morbidity. Even in countries where abortion is permitted by law, safe services are not available to all women. In the former USSR and India, for example, where abortion is allowed, many procedures are still performed outside the legal and formal health system, resulting in high numbers of unsafe abortions. Out of all illegal abortions, it is estimated that about 25 per cent occur in Latin America, 25 per cent in the former USSR, 13 per cent in India, and 10 per cent in Sub-Saharan Africa (Dixon-Mueller, 1990; Henshaw and Morrow, 1990; WHO, 1991). When restrictions on abortion are lessened, the number of abortion- related deaths and mortality decrease, owing to the greater availability of safe procedures performed by trained health professionals. In the United States, for example, death rates due to abortion fell by 85 per cent in the five years following legalization (Tietze, 1981 ). In the example of Romania, the abor- tion law was made more restrictive in 1966, and by 1984, the number of abor- tion-related deaths had increased by 600 per cent. When abortion became legally available once again in 1990, abortion mortality fell by 67 per cent in the first year (Romania Ministry of Health, 1991). Experience from other countries also confirms that legalization of abortion does not result in increased abortion rates, it only changes the conditions under which abor- . tions are performed. Despite the evidence that allowing abortion on liberal grounds reduces morbidity and mortality risks from induced abortion, only 22 per cent of the 190 countries in the world have abortion laws allowing it on request. Even countries such as Finland and the United Kingdom only offer abortion when it is justifiable for health as well as economic and social reasons. And the gap between developed and developing countries is very marked when we observe the reasons under which abortion can be legally performed. In over 80 per cent of the developed countries, for example, abortion is permitted not only to save the life of the woman, but also to preserve the physical health of the woman, to preserve her mental health, in the case of rape or incest, and when there is foetal impairment. Yet in developing countries, these reasons are much less accepted by the legal system; for example, only 26 per cent of Introduction t 25 these countries allow abortion in the case of rape or incest and only 23 per cent allow it in the case of foetal impairment. Only 6 per cent of the world's developing countries allow abortion on demand, among them Albania, China, Cuba, the Democratic Republic of Korea, Tunisia, Vietnam and most of the new countries emerging from the former USSR. But for the most part, in entire regions of the developing world abortion remains outside the law. In Africa, for example, safe abortion is not an option for most women. And it is the poorest women in all countries who bear the brunt of this unequal access to safe abortion services (United Nations, 1994). It must also be emphasized that legalization of abortion is a necessary, but not sufficient, condition for reducing the number of unsafe abortions. The many descriptions in this volume of women who resort to self-induced abor- tion or abortion performed by unskilled providers, even when it is permitted by law, attest to the need to make safe abortion services much more accessi- ble. Health care systems must learn how to respond to liberal laws with the appropriate administrative and structural support. The HRP Research Initiative In consideration of the fact that induced abortion continues to result in need- less deaths and serious illness to the poorer women in the developing world, the Task Force for Social Science Research on Reproductive Health of the Special Programme of Research, Development and Research Training in Human Reproduction (HRP) launched a major research initiative in 1989 on the determinants and consequences of induced abortion, with a focus on developing countries. Emphasis was given to projects from countries where abortion laws were restrictive, although some studies were supported in countries where abortion was legal, but services were not always accessible nor of good quality. The primary goal was to increase the knowledge base on the reasons why women seek abortion, even in contexts where family planning services are widely available and where abortion is punishable by law. Given the known difficulties of conducting large-scale surveys to assess induced abortion inci- dence, the Task Force also decided to support projects that offered new methodological insights and appeared to be feasible for application to large communities ( e.g., the study in Colombia by Zamudio). Attention to the rela- tionship between abortion and contraceptive behaviour, as well as to the con- sequences of induced abortion for the wellbeing of women were also impor- tant criteria for project selection. The decision to launch this research initiative was made during a critical time, following the 1984 Mexico World Population Conference, where abor- tion had taken centre stage in an ideologically charged public debate. Even- tually, the HRP proposal received strong and full endorsement from its 26 t Cynthia lndrlso and Axel I. Mundigo donors as well as from the World Health Assembly. More recently, the 1994 United Nations International Conference on Population and Development in Cairo and the Beijing Platform for Action endorsed at the UN Fourth World Conference for Women in September 1995 has recognized unsafe abortion as 'a major threat to the health and lives of women' and called for the promotion of 'research to understand and better address the determi- nants and consequences of induced abortion, including its effects on subse- quent fertility, reproductive and mental health and contraceptive practice as well as research on treatment of complications of abortions and post- abortion care' (United Nations, 1995: para llO[i]). It is hoped that these research studies undertaken in different political, eco- nomic, social and cultural contexts will assist in meeting these goals and pro- vide information useful for policy makers and for strengthening the public dia- logue in countries where women's health advocates are working towards legal reform and normative change in the health codes to de-penalize abortion. Part I: Between Abortion and Contraception The complexity of the relationship between contraceptive needs and behav- iour and the use of induced abortion remains, for the most part, unexplored territory in reproductive health research. For most women in the developing world, where abortion is usually illegal and unsafe, contraception would seem to offer a better fertility regulation option. Yet the research findings in this book bring to light the extensive use of induced abortion in many developing countries, even in those thought to have good family planning programmes, such as Mauritius, Mexico and Colombia. All the studies in this section link induced abortion to a sizeable gap between effective contraceptive use and childbearing intentions. In the Nepal study, unplanned pregnancy accounted for 95 per cent of induced abortion among women, but the majority of them were not using contraception. In the Dominican Republic, scarcely 25 per cent of women were using a contracep- tive method when they became pregnant unintentionally. In the Colombia study, 79 per cent of unwanted pregnancy was due to the non-use of contra- ception. Even in China, where contraception is easily and widely accessible, both Gui and Luo Lin et al., in their studies, find that non-use of contracep- tion is a primary reason for unwanted pregnancy and abortion. The other studies paint a similar picture. These studies also point to the serious unmet need for safe and effective ways to limit or space births not only among non-users of methods, but also among women who are using some form of contraception. High rates of con- traceptive failure, for example, were found in places where contraception has been legally and widely available for decades. In the China study by Gui, IUD in situ is the overriding reason for contraceptive failure; this is in a country Introduction t 27 where IUD users constitute 85 per cent of total users of reversible methods (Ping, 1995). Similarly, the study in China by Luo Lin et al. demonstrated a high method failure rate of 37 per cent; of these, an astonishing 65 per cent were IUD users. In the Cuba study by Alvarez, it was found that three out of every four women who had an abortion in the last 12 months were using a modern method of contraception. High failure rates are also reported in the Colombia and Dominican Republic studies. Also emphasized is the fact that non-married women, not just married women, have an unmet need for family planning. And as we will discuss in detail in Part III, the unmet need for contraception among adolescent girls is a serious reproductive health problem that is on the rise. In their search for explanations, the researchers in this volume challenge us to look more closely at how reproductive decisions are made and how such decisions are influenced by equally relevant, but less-researched aspects of contraceptive behaviour, such as women's and men's expressed views and concerns regarding contraceptive methods and services that are offered to them, gender relations, and the economic and cultural forces associated with sexual activity and pregnancy. Exploring the Gap Between Contraceptive Need and Use Although contraceptives have become more widely available in recent decades, a sizeable proportion of people who want to manage their fertility are not able to do so freely, effectively, and in a way that they consider safe. It is now widely recognized that an emphasis on demographic objectives during the early phases in the development of many family planning programmes worldwide, without adequate consideration of the conditions that encourage people's use of them, has helped create the existing gap between contracep- tive need and use. There is a growing body of literature that documents the failure of many family planning programmes to deliver user-friendly services, free of coercion and pressure, and respectful of the informational needs and personal preferences of the people who need to use them (Ross and Frank- enberg, 1993). Information Giving The enormous gulf between what services consider the appropriate informa- tion to be given to clients and what women and men feel they need to know in order to select the right method for their particular needs and to use it effec- tively is a major contributing factor to the existing gap between contraceptive need and use. As well illustrated by these studies, most women and men sur- veyed had not been educated about methods in a way that respects their right, as well as their capacity, to choose a contraceptive method themselves. 28 t Cynthia lndrlso and Axel I. Mundlgo Many of these studies show that the notion of method 'knowledge' com- monly applied in national surveys seems to be a poor indicator of whether women actually know how to use a method. Typically, the respondent is reputed to 'know' a method merely if she or he can state it or recognize it when it is mentioned. In fact, no or insufficient information on how to actu- ally use a method seemed characteristic of most women's experiences in these studies. In the Dominican Republic study, reported method 'knowledge' among the women is sometimes higher than that of the general population. There was also a considerable degree of prior use as well as current use of methods among the women in the sample; in fact, they had more contraceptive experi- ence, on an average, than the general population of women in the country. The author concludes, however, that one of the major determinants of unwanted pregnancy is a lack of knowledge among women and men about human reproduction and about the characteristics, correct use and possible side effects of methods. The study found, for example, that only 25 per cent of the women who have used the pill know how to use it correctly. The author points out that about 60 per cent of pill users nationwide buy their supply over the counter in pharmacies, although the study group experience indi- cates that obtaining the pills in a clinic or public hospital does not ensure that a woman will get appropriate medical screening, counselling and follow-up either. This same gap between so-called knowledge and information was found in the Turkey study by Akin. Here, authors concluded that the use of withdrawal is a primary determinant of abortion, which they attribute to, among other factors, poor knowledge of other methods of contraception. It was found that men and women rely mostly on friends and relatives for information about methods rather than on health personnel. The Colombia study also con- cludes that lack of information about more effective methods and how to use them contributes to the practice of induced abortion. Similarly, the Cuba study concludes that contraceptive failure results, in part, from the fact that women's understanding of how methods actually work is much lower than their more general knowledge of methods. In a country with such an extensive network of health services, the author discov- ered that the main source of method knowledge was a female friend; a woman's informal social network played a much more decisive role in the way she understood how the various methods work. The reason for repeat abor- tion among the women in the Nepal study was that they were unaware of the risk of conception following abortion. In Gui's study in China, among users, reasons for failure included wrong calculation of the safe period and wrong use of pills. The large majority of these Chinese women who experienced contraceptive failure said that health personnel in their local family planning units were not well-qualified and should share the responsibility for what happened; they said they received Introduction + 29 poor· guidance from them on how to master the use of various methods. Three out of every four women whose unwanted pregnancy was due to non-use of contraception declared that family planning workers in their places of residence knew nothing about their contraceptive status, or were indifferent, or did not provide any guidance. The study in China by Luo Lin et al. also attributed contraceptive failure and non-use to misinformation and lack of information. User Views on Method Safety and Effectiveness Studies about method acceptability increasingly point out the need to con- sider women's and men's perceptions and concerns about method side effects and safety. The'studies in this section help illustrate the importance of this. In Nepal, for example, reasons for non-use among women who aborted because they did not want more children included a fear of side effects, which was found to apply to all hormonal methods available, including Depo-Provera, Norplant, the pill and the IUD. Women preferred either to expose them- selves to unwanted pregnancy ( a risk they were largely unaware of) or to wait for a convenient time to undergo sterilization. The Dominican Republic study also revealed that side effects aqd health concerns are associated with the low and ineffective use of the pill as well as its discontinuation. The main reason given by women for non-use of methods in the China study by Gui was that 'contraception is troublesome'. In Turkey, Akin found that strong fears about the side effects of modern methods was a principle reason for their non-use. These studies also reveal that even though the great majority of women who seek abortion can also be very motivated to regulate their fertility, and have a high level of method knowledge and experience, their perception of the effectiveness of the method offered to them is not always in harmony with the definition of effectiveness that researchers, programme planners, and policy makers alike apply to methods. The widespread use of withdrawal in many countries is a good example of this fundamental difference in perspec- tive. Its use is completely confidential, not requiring a physical examination or contact with a clinic or pharmacy and it has no economic costs. It has the added advantage of not being associated with any proven health concern or serious side effects, particularly impairment of fertility. It also provides a flex- ibility-an immediate reversibility-that many couples seem to value highly. These are all factors that contribute to the method being 'effective' in the minds of many of the men and women who use it, while many family planning professionals tend to view this method as ineffective and therefore discour- age its use. Method effectiveness in the minds of users is also seen to be weighed against socioculturally and politically acceptable reproductive behaviour and norms. In the Dominican Republic study, for example, method use is shown 30 t Cynthia lndrlso and Axel I. Mundigo to be based on the expectation of ending the reproductive cycle, often at an early age, by means of sterilization once the desired family size has been reached. In fact, 88 per cent of the women in the sample did not initiate con- traceptive use until after the birth of their first child. Women's contraceptive practices in China reflect limitations imposed on them by the government's policy to control family size. As a result, women's choice of a method is pri- marily restricted to provider-controlled methods, mainly the IUD after first birth and sterilization after the second (Ping, 1995). As Luo Lin et al. note, not all contraceptive use is truly voluntary; some women would like to have a larger family and this feeling naturally affects use effectiveness. The Turkey study by Akin showed that ignoring women's culturally based preferences for female providers to perform IUD insertion, and for more accessible services, was found to contribute to the use of less effective methods. Resolving the disharmony between viewpoints about method effectiveness may lie in a greater recognition of the fact that fertility regulation is not a timeless strategy. These studies help to emphasize that the use of contracep- tion is not a single decision that is made only once, by women alone, usually at the beginning of their reproductive life-or worse, only at the start of mar- ried life-with only small adjustments made after that. The seeming mind- changing and ambiguity in fertility regulation behaviour in the form of 'dis- continuation' and 'switching' is actually evidence of how strongly motivated women and men are to adapt the existing reality of available contraceptive options to their real-life fertility regulation needs and preferences. In many cases, for example, women in these studies whose more traditional fertility regulation practices failed to meet their needs shifted to the use of modern methods. Many women whose first method was a modern one were seen shifting to a more traditional one when side effects were too problematic. Many women worldwide use more than one method; and they may mix or alternate methods, sometimes within short periods of time. Withdrawal, in particular, has been found to be commonly used with other methods (Rogow and Horowitz, 1995). Unfortunately, women's and men's changing contraceptive needs, mixed with incomplete information about available methods and intermittent or poor quality contraceptive supplies, bring about unwanted pregnancy. The experience of abortion following method ineffectiveness or failure subse- quently pushed many women towards the use of modern, longer-acting, and even permanent methods of fertility control. This decision was often made in spite of any preference they had for more short-term or traditional methods of fertility regulation, especially in the early part of their reproductive lives. Equally distressing is that many women, when left only with the choice between uncomfortable side effects or method failure, can become pessimis- tic about the usefulness of contraception and simply stop using any method at all. In the Mauritius study, in a country noted for its good family planning ser- vices, the author concludes that the low use of reliable meth6ds and non-use Introduction t 31 reflects the women's 'lack of confidence in all family planning methods'. Gui concludes in his study on China that women's perception of contraception as 'troublesome' also reflects their lack of confidence in the methods available and how they work. Similarly, in the Dominican Republic study, the low- income women's experience led the authors to feel that only surgical steriliza- tion and celibacy provide any real assurances against unwanted pregnancy; modern methods may work, but they are always a risky proposition. Gender Dynamics Until recently, studies of contraceptive behaviour and needs largely ignored gender dynamics in sexual relations and reproduction. Some of the studies in this section illustrate that the choice and acceptability of a fertility regulation method is strongly influenced by the preferences and pressures of other peo- ple in a woman's life, particularly her sexual partner. A primary reason for non-use of contraception in the Nepal study was opposition from a woman's husband. This study also revealed that many women were under pressure from their husbands and in-laws to get pregnant in spite of what they themselves wanted to do; they explained that their mar- riage was unstable and that their husband would bring a second wife to the household if they did not produce a child. This threat had even compelled some women to have a subsequent pregnancy immediately after a spontane- ous abortion. In Mauritius, a number of women resorted to abortion because of their husband's refusal to use a family planning method. In short, women lacked decision-making power vis-a-vis their fertility regulation behaviour. Method discontinuation was partly attributed to the husband's disapproval in the Dominican Republic study. The author notes that the three most popu- lar methods-withdrawal, periodic abstinence and condoms-demand a degree of bargaining power vis-a-vis their male partners that very few Domin- ican women possess. Information from the study sample showed that low- income men are reluctant to take on direct responsibility for the choice of a modern method or its use-such a role runs contrary to prevailing male atti- tudes in this culture. The use of withdrawal in Turkey was thought by the researchers to be a reflection of the preference, among both women and men, for men to domi- nate in decisions about the choice of fertility regulation methods (including abortion); and men had a clear preference for withdrawal. Men's knowledge of contraception, particularly modern methods, was also found to be lower than women's. Moreover, male dominance in the fertility decision-making process was linked to a woman's self-esteem. For example, women who thought that men are generally more intelligent than women, that men have a right to beat their wives, and other views that indicated low self-image tended to use effective contraception less, have fewer induced abortions, and have a higher than average fertility than women who did not agree with such views. 32 t Cynthia lndriso and Axel I. Mundigo When women worked for wages outside the home, they used more modern methods and had fewer pregnancies, in both the urban and rural samples of women. The author concludes that a woman's greater control over economic resources seemed to outweigh the influence of traditional cultural values that place men in a more dominant decision-making role within the household. Several studies elsewhere have shown that social support from husbands is important for continued contraceptive use (PATH, 1989). Studies conducted in African and Islamic countries have shown that programmes involving hus- bands contribute to the successful use of modern contraceptives (cited in Trottier et al., 1994: 289). In a study of first-time use of DMPA (Depotme- droxyprogesterone) among rural women in Bangladesh, those whose hus- bands approved of family planning had significantly longer use durations than those whose husbands disapproved (Riley et al., 1994). In another example of gender relations analysis, the China study by Gui in this section considered the influence of sexual satisfaction between a couple on the non-use of contraception. Women and men who discussed their sexual life with one another 'often' and were mutually satisfied, had a lower level of non-use of contraception. Non-use was highest among women and men who 'never' discussed issues relating to their sex lives. Similarly, non-use of con- traception was nearly twice as high among women and men who 'never' dis- cussed contraception with one another compared to those couples who 'of- ten' talked about the use of methods. The author also concludes that the correct use of more traditional methods, such as rhythm, seemed to be based on the quality of inter-spousal communication about sexual and contracep- tive behaviour. For example, couples who 'sometimes' or 'never' engaged in 'discussion of how to use contraceptives' had the highest proportion of 'wrong calculation of the safe period'. Do Women Rely on Abortion as a Contraceptive Method? The fear that women tend to replace contraception by abortion and that the availability of abortion results in decreased motivation to use contraception has been expressed by a number of researchers and policy makers as well as by opponents of more liberal abortion laws. It is a belief that often serves as an obstacle to more widespread de-penalization of abortion in many places. Several of the studies in this volume provide some evidence that the reason women seek abortion is not their preference for abortion over contraception, but rather the poor quality and accessibility of suitable contraceptive meth- ods and services. Again, we see that it is actually a strong desire to control fer- tility using available contraception that actually lies behind so much method switching, discontinuation, and even apparent non-use of methods among women who experience unwanted pregnancy and then seek abortion. In the Slovenia study, for example, which looked closely at the possible relationship between available and safe abortion services and the use of 'less Introduction t 33 effective' methods, nearly all the women who had an abortion had used con- traception either steadily or during a limited period of time prior to conception and at conception. The sample was divided almost evenly bet- ween women who used a more effective method and women who used a less effective method or no method at all. The women in this latter group, how- ever, did not realize that they were exposing themselves to a high risk of get- ting pregnant at the time. This is supported by the fact that they had above average contraceptive use and half of them reported that they thought the method they were using was an effective protection against pregnancy. The author suggests that easy accessibility of abortion did not reduce a woman's motivation to use contraception; abortion was used as an emergency measure when contraception had failed. The Cuba study also emphasizes that women choose abortion with reserva- tion, even where it is legally and easily available on demand. The author found that the predominant attitude of women, regardless of whether or.not they have abortion experience, is that of ambivalence. This is in contrast to their attitude towards contraception, which is highly positive. The often agonizing moral and ethical dilemmas that women face in decid- ing how to handle an unwanted pregnancy are described throughout this vol- ume, and attest to the general reluctance of women to have an abortion, regardless of whether it is legal or not. In fact, sterilization is often the prefer- able solution of many women in these studies to any further risk of unwanted pregnancy and abortion, especially among older women. Conclusions Whatever the complexities of the relationship between contraceptive reali- ties and abortion behaviour, all the studies call for general improvement in the quality of reproductive health services as a way to reduce the need for abortion. In the Colombia study alone, the findings suggest that one-fifth of all abortions could be avoided by improving access to higher quality contra- ceptive information and services. The importance of individual and ethical counselling that respects women's and men's reproductive rights, including their ability to make their own deci- sions about which method to use and when, their sexual practices, and their concerns about side effects and health risks, cannot be underestimated. There is also an urgent need to make available a wider range of methods to allow users greater freedom to choose what suits their needs best. More fundamentally, these studies illustrate that services need to be reori- ented in a way that recognizes that women and men are responsible actors and decision makers, very much involved in their reproductive destinies. Fer- tility regulation also needs to be recognized as a dynamic process. 34 t Cynthia lndriso and Axel I. Mundigo Research Gaps In order to facilitate such a change and create reproductive health services that truly respond to people's fertility regulation needs, the following areas of research need to be explored and considered more carefully. Gender We are nowhere near a comprehensive understanding of how gender shapes reproduction in different cultural contexts and at different moments in a woman's reproductive life. Much more gender research is needed. It is a pow- erful tool for exploring the influence of male-female differences and dispari- ties in the area of fertility behaviour. Such an approach recognizes the impor- tance of power and negotiation within the sexual relationship, answering critical questions such as who has input into contraceptive decision-making and how the differences in women's and men's access to and control over resources influence fertility behaviour. Method Acceptability There is a need to identify more precisely the qualities that are believed to determine whether a contraceptive is acceptable, including both modern and traditional methods. More method-specific information is needed on what are perceived as method side effects by women and men and what constitute their health concerns within different social and cultural contexts. '\Yhat do they consider a safe and effective method, for example? It is also important in research to differentiate between the fear and anxiety felt by users, which is possibly due to a lack of clear and complete information, and their actual experience with methods. Both men and women need to be included in this research, which should focus on user views, particularly as they relate to the temporal structure of their reproductive behaviour and needs. To what degree the quality and modes of service delivery influence method acceptability must also be part of the process of identifying women and men with unmet needs. Emergency Contraception Recently there has been a surge of interest in adding 'emergency contracep- tion', defined as methods used immediately after unprotected intercourse to prevent unwanted pregnancy, to the list of fertility regulation options. The technology used, such as the Yuzpe regimen (combined estrogen-progesto- gen therapy) and postcoital insertion of an IUD, has been available for about 30 years, and both these techniques can be made easily available as part of Introduction t 35 routine services (van Look and von Hertzen, 1993). More recently, the use of mifepristone as a method of emergency contraception is being studied (Klitsch, 1991 ). The results of these studies would seem to support making such an approach more widely available, given the high rates of method failure and ineffectiveness. It is essential, however, that researchers, programme plan- ners, and policy makers all keep a firm distinction in their minds between emergency contraception and early medical abortion. A distinct moral and ethical difference in women's minds can be found between preventing unwanted pregnancy and terminating a pregnancy that has been confirmed, which should be respected. Other considerations, particularly the biomedical ones, are certainly beyond the scope of this chapter. Part II: The Quality of Abortion Care Evidence shows that the quality of reproductive health care services is a key factor in an individual's or couple's ability to initiate and sustain the use of fertility regulation methods and the essential elements of high quality care are now widely recognized (Bruce, 1990; WHO, 1995b ). High-quality care is also about accessibility: the largest number of people must be able to benefit from it, which means that they must understand how to obtain it. In the case of abortion, the quality of care received often means the differ- ence between life and death. In a context where abortion is restricted, the thin line between the two is drawn at the point of treatment for life- threatening complications. Unfortunately, most medical facilities in develop- ing countries are unprepared to deal with cases of botched abortion and those that are equipped are also usually overburdened. Many women get referred, without receiving any treatment from providers, to other hospitals a great distance away. Elu draws the conclusion that most of the Mexican women in her study could have received better care if the health providers had been more compassionate and well-trained and the facilities better equipped, regardless of whether women had 'institutional rights' to them. Elu's study in Mexico also emphasizes that 'access' to abortion cannot be resolved by physical proximity. It means an approach to the situation that answers women's own perceived reproductive health care needs in a way that is consistent with their values and way of life, with an understanding of the complex cultural values that support its existence. These studies highlight the need to focus more directly on the needs and preferences of the women who seek abortion as well as on the attitudes and skills of the providers of abortion services, in order to improve the quality of abortion care. All women who seek abortion, no matter how poor or unedu- cated, deserve to have their concerns and situations as individuals addressed with care and compassion. Similarly, the concerns and perceptions of 36 t Cynthia lndr/so and Axel I. Mund/go providers must also be explored and understood, particularly in situations where abortion is illegal and the law does not distinguish between treatment of an incomplete abortion and actually performing one. What is especially unique about this group of studies is that they illustrate that the quality of abortion care needs to be addressed, regardless of the pol- icy context. Certain fundamental elements of abortion care that profoundly affect women continue to be widely neglected, even in places where safe abortion is widely available. Women's Experiences of Abortion In a broad sense, the studies included in this section, like the studies in Part I, reinforce the failure of family planning programmes to effectively respond to women's needs and desire to control their fertility. Non-use of contraception among these women was high, as was contraceptive failure. In the China study by Zhou, for example, a number of women felt angry and betrayed because modern methods of contraception had failed them, particularly the IUD. Gender relations appears again as a significant contributing factor to explain unmet need. In Elu's study in Mexico, for example, some women never used contraception, stopped using contraception, or got pregnant at the request of their male partners, who later changed their minds or refused to take responsibility for the pregnancy. Other women became pregnant without their male partner's knowledge in the hope of strengthening an unstable relationship, but later had to abort because the man did not want the child. Understanding of Abortion Safety and Risk A characteristic that marginalizes women who seek abortion, particularly those who resort to clandestine procedures, is their apparent lack of concern or awareness of the risks involved in the procedure. Some information was revealed about the ways in which women who confront an abortion actually understand and define the notion of safety. To what degree, if any, did their perception of safety influence their decision to abort? In the countries where abortion is legal and more accessible, as in China and Turkey (and Cuba, studied in Part I of this book), medical safety and the risk of the abortion procedure emerges as an expressed concern of the women. Women wanted, and expected, the provider to be competent. It is important to note how women define competency; in Turkey, women judged a provider to be skillful if the procedure was quick and painless for them. In contrast, where abortion is restricted, a concern for safety seemed a luxury that most women literally could not afford. These women had a more fatalis- tic attitude towards the situation, accepting that without economic resources Introduction t 3 7 to pay the high price for a safe procedure, they could do nothing but submit to a clandestine procedure while praying for 'luck' to be on their side. The link between the quality of abortion care and the economic situation of women is particularly well illustrated by Elu in Mexico. The stories of the desperate attempts at self-induced abortion by married women, who risked leaving several children and a husband behind, are particularly powerful. The use of misoprostol ( also known as cytotec, a drug used to treat ulcers) by two-thirds of the women in the Brazil study by Misago is described as a cheap way for poor women to circumvent the tight restriction on legal abortion. Fortunately, most of the women sought hospital care soon after bleeding started, which resulted in a lower rate of complications than those found among women in previous studies. The Nepal study (Part I) found a positive relationship between the educa- tional level and occupational status of a woman's male partner and her use of safe abortion services. Women whose male partners had a higher educational level and higher paying jobs visited private clinics more for safer abortions, which are more expensive, than women whose partners had lower educa- tional and occupational status. Actual knowledge of different medical abortion techniques and their possi- ble risks was low to nonexistent among almost all the women, including among those who had a safe and legal procedure. The China study by Zhou, for example, revealed widespread myths among women about whether the procedure was harmful to subsequent reproductive health and its effect on future pregnancy. In the Turkey study by Bulut, most women were unsure about which method had been used for their abortion. And in the Cuba study (Part I), again in a context where safe abortion is legal and accessible, nearly 40 per cent of the women had only a vague idea or did not know what "the actual procedure for abortion was. Also, a high proportion of women expressed a fear of becoming infertile; and one in 10 women worried that an abortion might lead to death. Abortion Information and Counselling In its simplest form, counselling is the provision of information about the abortion procedure, to help a woman feel more relaxed beforehand. At its best, abortion counselling recognizes that the experience belongs to a contin- uum of events that are inextricably linked to a woman's sense of identity as a sexual being, her self-esteem and her future. As a source of in-depth psycho- logical and emotional support, such counselling helps a woman to confront her situation and her feelings about her sexual life in order to 'break the cycle of reproductive risk-taking behaviour' (Londono, 1989). As Elu illustrates in her study of poor women in Mexico, the concept of unwanted pregnancy is not as simple as it would appear. The 'desire' for a pregnancy is not necessar- ily conscious or rational and it is rarely contemplated in advance. Counsellors 38 t Cynthia lndrtso and Axel I. Mund/go must have compassion and respect for the internal and external conflicts the experience of deciding to abort an unwanted pregnancy creates in most women's lives. Neither simple counselling nor more empowering and in-depth support was offered to the great majority of the women in the studies throughout this vol- ume. Women in the Turkey study by Bulut described a lack of individual atten- tion from providers and said they felt 'rushed'. Several women pointed out the need for pre-abortion counselling services, such as screening for health prob- lems, discussion of how and why the pregnancy occurred, and information on what to expect during the abortion procedure, including information about the type of method being used. Some women also expressed a desire to have a choice of abortion technique. Pain Management Whether and to what degree women who undergo abortion experience pain is usually not a priority among abortion care providers, particularly in a con- text where the procedure is usually unsafe to begin with. In almost all the studies in this volume, concern about serious morbidity and mortality from unsafe procedures easily outweighed all other concerns among women, researchers and providers alike. Nonetheless, pain has been shown to rank high on the list of women's concerns when they confront the prospect of an abortion. It is also an aspect of abortion that can have profound implications for a woman's future sexual wellbeing. Both these points are well illustrated in the China study by Zhou, where abortion has been legal and accessible to all women for decades and there- fore the complication rate from the procedure is shown to be minimal. (See also the two studies by Luo Lin et al. in this volume). But the degree of pain women experienced was, by their own accounts, overwhelming. One woman even chose to carry a subsequent unwanted pregnancy to term to avoid the pain of another abortion. Many women also mentioned that they experienced subsequent and serious psychological anxiety towards coitus as a result of the experience. Similarly, a very high percentage of women surveyed in both hos- pital settings in the Turkey study by Bulut also experienced pain and found the procedure 'difficult', even though other medical complications, such as bleeding, were found to be minimal. Post-Abortion Family Planning Counselling and Services Women who have just had an abortion need information that facilitates informed choices about how to control their fertility, as all the studies in Part I illustrate. The urgent need to establish closer links between family planning and abortion information and services is also reflected in these studies, where Introduction t 39 post-abortion counselling was rarely offered to women to help them avoid another unwanted pregnancy. In the China study by Zhou, all the women who had an abortion, especially those with a history of abortion, very much longed for safer and more effec- tive contraceptive methods. In the Turkey study by Akin (Part I), only one-third of the women who had an abortion received counselling or infor- mation about family planning methods, irrespective of whether the proce- dure had been performed in a public hospital or by a private doctor. In the Nepal study (Part I), none of the 1,241 women who had just had an abortion were aware that they could soon be at risk of conception when they were dis- charged from the hospital/clinic. The study team noticed that clinicians made no effort to provide post-abortion family planning counselling. The women who sought abortion in the study in Turkey by Bulut expressed their strong desire to learn more about family planning methods; and the majority rated post-abortion family planning counselling as poor. An inter- esting finding with important service implications is that the more educated women in this study were more successful at 'extracting' information from their providers than less educated women. It appeared that a woman had to ask for and insist on information, which takes a certain amount of self- confi- dence. The use of modern methods six months after the abortion was shown to be much higher, 80 per cent, where family planning and abortion services were more integrated, compared to women who had abortions at the health facility where the two services were completely separate, at 36 per cent. The women's rating of overall quality of care they received, while generally low in both hospitals studied, was ranked significantly higher in the hospital where abortion and family planning services were linked. Unfortunately, an established connection between abortion services and family planning services, including systematic referral, were lacking in all the other studies, whether abortion was legal or not. There also seemed to be some hints of provider bias in the provision of methods to women who have just had an abortion and also some provider disagreement about the timing of post-abortion contraceptive use. Although none of the research focused on specific problems encountered, it is clear that the gap between contracep- tive need and use was not addressed in any effective way during this critical time in a woman's reproductive life. Other important studies have also found a serious neglect of post-abortion contraceptive service.s (Benson et al., 1992; Neamatalla and Verme, 1995). Provider Beliefs, Attitudes and Technical Competence Provider-client interaction, in addition to the technical competence of the providers, is now widely recognized as an essential component of quality care. What has been less well considered is that this same element must also be 40 t Cynthia lndriso and Axel I. Mund/go present to achieve quality abortion care. Moreover, there is an important sociocultural dimension to this interaction. While other studies in this volume also highlight the importance of pro- vider skill and views in abortion care (see, for example, Elu on Mexico and Zhou on China), the unique contribution of this group of studies is that they call particular attention to the way that the beliefs and attitudes of health care providers are brought to bear on the care and treatment of a woman with an unwanted pregnancy. They also remind us of the simple fact that health care workers in all parts of the world, regardless of legal and social sanctions, are frequently called upon to provide care for women undergoing, or who have undergone, abortion. As with the studies about women's experiences of abortion, the influence of abortion policy on the quality of services and on the health of women is evi- dent in these research projects, where abortion was a criminal act or severely restricted in each of the settings studied. In fact, almost all providers in each of the studies noted that unsafe abortion is a major health problem in their country. Belief versus Action To what degree, if any, do the personal views and religious beliefs of providers affect their treatment of women who need abortion care? A possible negative impact on the health of women is suggested by the study in Sri Lanka by Hewage, where it was found that a conservative attitude on .the part of the providers-61 per cent disapproved of contraception for non-married women- could be contributing to the occurrence of unwanted pregnancy. Similarly, 41 per cent of these providers said that abor'tion was on the increase in their country because women were 'ignoring social values'. Oearly, for some of these health providers, a more humane approach to abortion presented an ethical and moral conflict between their personal beliefs and their clients' rights and wishes. (It should be noted that these providers worked in government facilities, mainly hospitals-not the so-called 'private sector', where most illegal abortions are performed.) Similarly, other research has demonstrated that some providers have a harsh and punishing attitude towards women who have multiple abor- tions (Neamatalla and Venne, 1995). In the other three studies in this group, however, personal belief and behaviour on the part of the providers were not so easily harmonized. In fact, stated opinion was often contrary to action. Here, it is important to note that all three study samples comprised or included providers from the nonformal sector, such as traditional birth attendants (TBAs), who have the closest con- tact with women in the communities and therefore more likely to come face-to-face with requests for abortion as well as the life-threatening compli- cations of self-induced or unsafe, illegal procedures. The research conducted by Cadelina in the Philippines shows that in spite of the fact that the great Introduction t 41 majority of TBAs said they did not approve of abortion, primarily for reli- gious reasons, and laws were restrictive, a significant number of them admitted that they gave women information on abortion methods and on how to locate abortionists; some admitted to performing abortions them- selves. Similarly, in the Indonesia study by Djohan, where attitudes of both formal and nonformal sector providers generally seem opposed to abortion, again because of religious beliefs and restrictive laws, some midwives said they assisted ob-gyn doctors who performed abortions; and half of the mid- wives said they would actually perform an abortion. Also, some TBAs accom- panied women to a clinic for abortion. In Mexico, Pick shows that although market herb vendors had more stated negative views towards women who wanted an abortion than pharmacists, they were more likely to provide an abortifacient. What makes so many of these providers act contrary to strong personal beliefs and/or risk criminal punishment? In Indonesia, the willingness of qualified health care providers to finally assist a woman who requests an abortion seemed ultimately practical. It is better to help a woman than to have her go to an untrained person. In Mexico, Pick explains that provider beliefs about abortion do not have a predictable impact on their actions because the issue is surrounded by a tremendous amount of guilt, which causes a lot of ambiguity in speech and action. Offering abortifacients to women is also an important source of income for these vendors. In the Philip- pines, some of the vocal opposition to abortion expressed by the TBAs is attributed not to their actual personal opinion but to their growing reputation as abortionists, which diminishes their social standing in the community. Conclusions The obvious starting point for quality abortion care is to ensure that all women have access to a safe procedure. These studies also clearly point to the need to re-think how women who seek an abortion should be treated and the kind of information they need to be given in order to truly address their reproductive health needs, both physical and emotional ones. Individual, personal attention to each woman who seeks an abortion is essential. Counselling that explains the details of the abortion procedure can help reduce anxiety considerably and is a key factor in women's overall satis- faction with abortion services. Efforts should be .made to do whatever is pos- sible to minimize the level of pain women may experience during an abortion. As Luo Lin shows in her larger study (Part I), pre-abortion counselling can significantly lower stress and depression levels among women who have an abortion; and other studies have shown that pre-abortion psychological coun- selling and a sensitive and compassionate staff can greatly minimize the need for analgesic and anaesthetic drugs (Nasser, 1989; Stubblefield, 1989). As Zhou suggests in his study, the availability and use of appropriate anaesthetic techniques and skills of application should be improved in China. Bulut 42 t Cynthia tndriso and Axel/. Mund/go suggests that women in Turkey be given a choice of anaesthetic techniques. A choice of abortion techniques should also be offered wherever feasible. Post-abortion contraceptive counselling is also a critical factor to women's overall satisfaction with services. In Turkey, the study by Bulut concluded that contraceptive counselling before the procedure is also helpful. Such informa- tion and services must be provided in a responsible and ethical way. As Bulut points out, while IUD insertion immediately after abortion may be the easiest method for providers to offer, it must be accompanied by appropriate coun- selling. Follow-up is also essential. Integrate Services It is essential that abortion care be linked to family planning services in order to reduce the need for abortion, particularly where it is illegal or not easily accessible. In these situations, women are unlikely to get family planning ser- vices from clandestine providers and thus remain vulnerable to additional unwanted pregnancies and repeat unsafe abortion. In general, maternity hospitals or clinics that offer care to women with abortion complications resulting from unsafe procedures should also have facilities that provide these women with post-abortion family planning ser- vices. If that is not possible, upon discharge, women should be referred to family planning services where they should receive special attention. Better referral systems that alert family planning providers to the case situation should be tested and implemented, but always with the priority of preserving the privacy and anonymity of the woman. Similarly, closer attention should be paid to the provider side, including their training, attitudes and under- standing of the problem. For example, family planning and abortion provid- ers should be trained together and work collaboratively. Research has found that their training is often different, with the result that abortion providers are not necessarily well-informed about contraceptive methods and advances; and conversely, family planning providers may know little about abortion techniques and the appropriate conditions for immediate post-abortion con- traception (Neamatalla and Verme, 1995). The unique study in Chile by Molina and his colleagues gives an example of what can be accomplished when services are well-integrated. When women determined to be at high risk for abortion had individual discussions in pri- vate and in their homes about their contraceptive needs and preferences with a supportive counsellor, their use of contraception increased noticeably and abortions in the community declined. In the control study community where no intervention had taken place, abortion ratios increased by 30 per cent. Provide Wider Access to Safe Medical Abortion The obvious need for safe, accessible and affordable methods of pregnancy termination requires a closer study of the different possible approaches to Introduction t 43 safe medical abortion by biomedical scientists. In some of these studies, the suggestion for easier access to medical abortion was made by the women themselves. The use of mifepristone for medical abortion was raise(f often as a topic of discussion among both the women and providers in the: study by Zhou in China and suggested as a way to avoid the psychological harm of too much pain and reduce the possible medical risks associated with surgical proce- dures, such as dilatation and curettage (D&C). In the Mauritius study, where abortion is rul.ed a criminal act, the use of misoprostol (cytotec) to induce abortion was reported by 36 per cent of women, as high as the use of crude methods such as inserting bicycle spokes, jumping from heights, massage and herbs. The author cites this as an example of how Mauritian women will quickly adopt a new method of pregnancy ter- mination that promises to be safer and more effective than other available methods. The widespread and illegal use of cytotec was also reported in the Brazil study by Misago, where again, abortion is tightly restricted and a safe procedure is too expensive for most women. The author also notes, albeit cautiously, that the use of cytotec reduced the complications of induced abor- tion found in other studies. Another recent study on the use of cytotec in Brazil confirms that its main advantage for women is that it is relatively inex- pensive compared with other available methods. Women in this study also felt that it made the abortion experience less traumatic because it reduced the often agonizing delay between the decision to abort and the event itself, with administration possible in private without a trip to the clinic. The drug's influence in reducing the complications of unsafe abortion found in other studies was also confirmed (Barbosa and Arilha, 1993). Another recent study has found a combination of methotrexate and misoprostol to be a safe and effective alternative to invasive methods for the termination of early preg- nancy (Hausknecht, 1995). Interesting research on women's views on the experience of medical abor- tion in Stockholm revealed that nearly half the women commented spontane- ously on the relief they felt when they realized that they were going to have a medical abortion, rather than a surgical one (Holmgren, 1994). As the author describes: 'To them a surgical procedure meant that someone would operate inside very fragile parts of their bodies, which they found threatening.' This same study also found that women described an 'ethics of care'; it was impor- tant that their abortion be performed as early as possible. Improve the Provider Dimension These studies present a strong case for letting family planning workers and abortion providers at the primary care level play a role in efforts aimed at reducing the practice of unsafe abortion. In the Philippine study, for 44 t Cynthia lndriso and Axel I, Mund/go example, traditional practitioners, despite their use of unsafe methods, are more popular than medical doctors throughout the country, mostly because they are less expensive and women are assured of complete confidentiality. It is suggested that their role be utilized in reducing unsafe abortion; for exam- ple, by improving their skills in counselling women about sexuality and con- traception. In the Indonesian study, the role of the nonformal health sector providers (TBAs and PLKBs) remains central in caring for the women that the formal sector does not help. Educating and training these health workers could be an effective way to reduce levels of unsafe abortion. In the Sri Lanka study, 78 per cent of the respondents recommended the family health worker (FHW), a woman who visits households regularly, as the best person to han- dle a sex education programme at primary and secondary school levels. Pick suggests educational programmes for pharmacists and market herb vendors in Mexico, preferably as part of wider training on reproductive health and family planning. Such training could encourage providers to treat women more humanely, discourage the ineffective use of widely prescribed metrigen and the dangerous use of quinine, and emphasize the importance of provid- ing clients with accurate information about family planning methods so that repeat abortion is prevented. While these studies highlight the role of nonformal health personnel, doc- tors in the formal health sector, particularly obstetricians-gynaecologists, also need to take part in ensuring that safe abortion services are readily avail- able. Training in the management of botched abortion must also be a part of their responsibilities. It cannot be assumed that obstetricians-gynaecologists know how to perform a safe abortion. In many countries, including developed ones such as the United States, many obstetrician-gynaecologists are not trained in the application of modern abortion techniques, let alone in the management of incomplete abortion (Pick, this volume; Rosenfield, 1994). Regarding technique, the older method of D&C is more commonly used in developing country settings, although it has a higher complication rate and requires more modern and costly surgical facilities, which further limit the availability of abortion services (Sundstrom, 1993). In fact, replacing D&C with vacuum aspiration (VA) and manual vacuum aspiration (MVA) has been shown by several studies to be an easy and immediate way to make abor- tion and the treatment of incomplete abortion safer and much more accessi- ble (Baird et al., 1995; Coeytaux et al., 1993). Other studies have also demon- strated that the technique of MVA early in gestation (also referred to as menstrual regulation), can be used by specially trained paramedicals and is effective, inexpensive and accepted by women in rural areas (Sundstrom, 1993). Bulut concludes that her study in Turkey supports women's preference for menstrual regulation, with its advantages of provision by non-medical personnel and no need for anaesthesia. No matter which technique is used, provider training on how to lower psychological stress before the procedure must also be given. Introduction t 45 As these studies make clear, provider attitudes are powerful determinants of women's experience of abortion and their acceptance of post-abortion contraception. It is important to find ways to ensure that the general attitude of health personnel at all levels be sensitive to women's needs and empower- ment, regardless of the personal attitudes towards abortion of the providers themselves. Research Gaps Generally, little attention has been given to what makes women comfortable, both physically and psychologically, as they go through the experience of induced abortion. Women's views and provider skills regarding different abortion techniques and types of pain management need to be studied in var- ious settings, particularly the use of menstrual regulation. As just mentioned, safe methods of medical abortion should be thoroughly explored by biomedi- cal experts. The appropriateness and content of post-abortion family planning coun- selling and services need to be defined more carefully. Is counselling more useful to women before or after the abortion? When is it appropriate for women to start various methods of contraception following abortion? More method-specific information needs to be gathered, with close scrutiny of how the provider dimension is affecting the situation: Which methods are more commonly offered in different settings and why? Does post-abortion counsel- ling consider a woman's recent contraceptive experience leading up to the unwanted pregnancy? What is her attitude towards resuming sexual activity? Careful attention to provider-client interaction is needed, particularly the way that providers may be influencing a woman's right to free and informed choice about contraception post-abortion. Provider bias and coercion tends to be a particular problem in cases where women have had a repeat abortion. Part Ill: Adolescent Sexuality and Abortion About one-fifth of the people of the world, more than one billion, are adoles- cent, usually defined as those aged 10-19 years (WHO, 1995c). It is estimated thatmore than 50 per cent of the world's population is below the age of 25, of whom more than 80 per cent live in the developing world (United Nations, 1993). With respect to sexuality, the lower age is normally defined by puberty, while the upper age is more a sociocultural boundary than a physiological one, with marriage traditionally signalling entry into adulthood. Yet it has only been relatively recently that the health of adolescents and young people has been recognized as an important focus of attention. This has come about largely as a result of mounting evidence that they are increas- ingly sexually active and the trend for unwanted pregnancy among them is also on the rise (Fathalla et al., 1990; Senanayake and Ladjali, 1994). 46 t Cynthia lndriso and Axel I. Mundigo It is not difficult to understand why adolescent sexual activity often leads to unwanted pregnancy. Their access to sex education and family planning information and services is the exception rather than the rule in most parts of the world, and as a result, the majority lack knowledge about the basic repro- ductive functioning of their bodies and about the power and meaning of their emerging sexuality. As revealed by Ehrenfeld's qualitative study in Mexico, 'the vagina seemed not to exist' for many of the girls who had recently aborted. They were completely ignorant about female anatomy and physiol- ogy. They had little or no knowledge of contraception and most had become pregnant just after becoming sexually active. In Mpangile's study in Tanzania, a total of 89 per cent of the women, aged 17 years or less, knew nothing about either modern or traditional methods of contraception. As the author points out, this is undoubtedly related to the fact that contraceptive use by students is prohibited by school regulations and single women and adolescents are not welcome in family planning clinics in Tanzania. These findings are supported by larger studies of contraceptive use among adolescents (United Nations, 1989). Adolescents face obstacles to obtaining contraceptives such as paren- tal or spousal consent, specified minimum age, and the requirement of mar- riage-reflecting religious and cultural concerns that seek to prevent premar- ital sexual activity. The study in Korea by Kwon provides an example of how the interplay of gender relations and self-image, and sociocultural norms and expectations influences adolescent sexual behaviour as well. Although 79 per cent of the girls interviewed said they knew about contraception, more than 80 per cent of them had not used any contraception during their first sexual experience. Most had worried about getting pregnant at the time, but hesitated to use any contraception because they feared their male partner would label them a 'bad quality girl'. In fact, the non-married girls who regarded themselves as 'virtu- ous' were those least likely to use contraception, not because they were igno- rant, but because to do so would have contradicted their moral self-image. Another example is found in the Mexico study. All the girls, except one, said they had first intercourse at their boyfriend's request, and 98 per cent had not used contraception at the time. As one girl explained: 'It is not something you think about, for who can think at that moment? Love is what counts.' Also significant is the finding in this study that formal unions were established as a consequence of a girl becoming pregnant, and not from an expressed desire to formalize a relationship. The influence of sociocultural norms and expectations is also shown clearly in the China study by Luo Lin et al. Although contraceptives are freely or inexpensively available over the counter in China, 93 per cent of the unmar-. ried sample of young women seeking abortion said that neither they nor their partners had used contraception at the time the woman got pregnant. These non-married women reported that they felt too embarrassed to seek contra- ception at a pharmacy or at the local family planning clinic, illustrating how a Introduction t 4 7 strong social stigma against premarital sexual activity can effectively prevent the use of contraception. Young girls who decide to carry their unplanned pregnancy to term usually face insurmountable difficulties, particularly shame and social isolation from their family and peers, interruption of employment or careers, increased eco- nomic hardship, and a diminished opportunity for a later marriage. Not sur- prisingly, the most common solution to an unwanted pregnancy for a non-married girl in most parts of the world is an abortion. In the China study, for example, the main reason given by 93 per cent of the girls for having an abortion was the fact that they were not married. In the Korea study, the con- clusion is drawn that abortion is relied on as a strategy for dealing with unwanted, out-of-wedlock pregnancy. Indeed, all the premarital pregnancies in this sample had ended in abortion. Although precise figures are unavail- able, adolescents aged 15-19 account for at least 10 per cent of the induced abortions that occur each year throughout the world (Senanayake and Ladjali, 1994). Moreover, the studies in this section as well as information from many other sources also illustrate that it is unsafe abortion, either self-induced or performed by an unqualified abortionist, that non-married adolescent girls most often seek-whether or not abortion is legal in their country (Fathalla et al., 1990; WHO, 1995c). In the Tanzania study, nearly a third of the victims of unsafe abortion were teenagers, of whom almost half were 17 years of age or younger; about one in every four were students in primary or secondary school. Incomplete abortion was diagnosed for 72 per cent of them and post-abortion sepsis was the most common complication; almost 53 per cent of all the women also had signs of trauma to their genitals. In the Korea study, in virtually every in-depth interview and focus group session, unsafe or unhygienic abortions were reported to be widespread. Some girls said they had tried to self-induce abortion before going to a provider. The authors point out that a safe procedure is very expensive and thus unaffordable for most non-married young women in Tanzania. In the Mexico study, the wish to end the unwanted pregnancy was very clear in the minds of most of the girls, overriding all other considerations, even the life-threatening risks of the inev- itable illegal, unsafe procedure. Even a high proportion of the Mexican girls who ended up carrying their pregnancy to term had actually wanted to abort the pregnancy. As in Korea, they did not have the money to obtain a safe abortion and most had tried unsuccessfully to self-induce an abortion using methods such as drug injections, herbal infusions, quinine tablets, or violent physical exercise-in combination or repeatedly. The exception to the common experience of physical and emotional trauma from an unsafe abortion was found in the study in China, where abor- tion has been legal and accessible to all women for decades. All the non- married women in the sample had easy access to abortion by qualified 48 t Cynthia lndrlso and Axel I. Mund/go providers under safe conditions. As a result, the medical complication rate from the procedure was minimal compared to the other studies in this section. Adding to the trend to seek unsafe abortion is the fact that adolescents often present themselves too late for abortion, when the procedure carries the greatest risk. Even where abortion is legal, it is not easy for a young non-married girl without support or finances to seek an abortion early enough. Reasons include denial, feelings of shame and guilt, inability to seek medical care without discovery, or simple ignorance about the consequences of delayed treatment. Again, an exception to this general trend was found in the China study, where the majority of young women, 64 per cent, had their abortion when gestation was between 41-60 days. This illustrates that wide- spread dissemination of health information and safe services that include young non-married women can reduce their risk of mortality and morbidity from abortion. It is significant that the topic of sexually transmitted diseases (SIDs) and HIV infection surfaced more clearly in the adolescent studies of abortion than in any of the other abortion studies in this volume. Among sexually active young people, SIDs are most frequent in those who are youngest, and appear to be increasing throughout the world, although diagnosis and report- ing is poor (WHO, 1989). In many countries, 60 per cent of all new HIV infections are among 15-24-year olds, with a female to male ratio of 2:1. The largest number of these infections is in Sub-Saharan Africa, but the biggest recent increase has been in Latin America and South and South-East Asia (WHO, 1995c). Information from other sources shows that many young peo- ple have heard about these diseases, but as with knowledge about contracep- tion, having this information does not seem to affect their reproductive behaviour. The reasons relate to the same sociocultural factors just discussed. However, it must be emphasized that young girls are especially vulnerable to SIDs and HIV/AIDS, because they tend to marry, or have intercourse, with men who have had more sexual exposure. In the Tanzania study, for example, one-third of the girls aged 14-17 years said they had been made pregnant by men who were 45 years of age or older. Added to this is an already existing biomedical risk of infection among young girls because they have an immature cervix and limited vaginal secretions, which provide less of a barrier to infection. Forced first intercourse, which is not uncommon among young girls and was reported in all the studies in this section, leads to genital trauma and cuts and thus increases the risk of infection (WHO, 1995c). Naturally, young girls engaged as sex workers are the most vulnerable to infection. They are likely to be patronized by men who in most societies have many more sexual partners than the women and thus are more likely to infect their 'clients'. In the Korea study, 14 per cent of the young women interviewed said they had gotten an SID as a result of their first sexual experience. Introduction t 49 Conclusions The Beijing Platform for Action, approved at the UN Fourth World Confer- ence on Women in September 1995, emphasized the problems for adoles- cents just discussed, stating: Counselling and access to sexual and reproductive health information and services for adolescents are still inadequate or lacking completely, and a young woman's right to privacy, confidentiality, respect, and informed con- sent is often not considered . The trend toward early sexual experience, combined with a lack of information and services, increases the risk of unwanted and too early pregnancy, HIV infection and other STDs, as well as unsafe abortion (United Nations, 1995: para 95). The actions it recommended included the implementation of education and information programmes on sexual and reproductive health issues and on STDs, including HIV/AIDS, in school curricula from the primary level; and full attention to meeting the service needs of adolescents to enable them to deal in a positive and responsible way with their sexuality (United Nations, 1995: para 108). When considering the best approaches to meeting the reproductive health needs of adolescents, it must be kept in mind that they do not comprise a homogeneous group; there is wide variation in their characteristics, such as age, cultural views of sexuality and marriage, basic education, employment opportunities and access to reproductive health information. This makes the provision of contraceptive services difficult at best. In the absence of an ideal method of contraception for adolescents, it is important to make available as wide a range of suitable methods as possible with good quality counselling services. In cases of abortion, including treatment of abortion complications, counselling both before and after the procedure is especially impor- tant-indeed, critical, for psychological wellbeing. The encouragement of sexual responsibility among boys and young men should be a priority when setting up sex education programmes. It is young girls, and not their male partners, who suffer the consequences when premar- ital sex results in unwanted pregnancy and there is therefore an urgent need to make boys aware of the consequences of their sexual actions for their female partners. Research Gaps Adolescent sexual behaviour and its determinants is very much·an under- researched area. A closer look at the gap between contraceptive knowledge and behaviour, following the examples of the studies in Part I is needed. Keeping in mind the possible influence of peers and parents, adolescents' 50 t Cynthia /ndrlso and Axel I. Mundlgo own views and perceptions about their sexuality and sexual practices, contra- ception, partner relationships, and goals for their future must be studied and understood. A critical dimension to this kind of research is gender relations, particularly the study of the attitudes and role of male partners, concerning contraception, sexual responsibility and decision-making. Profiles of the boys and 'sugar daddies' who took part in the sexual relations or fathered a child are conspicuously absent from the studies presented in this section. The way that adolescents are reacting to the threat of HIV/AIDs and other SIDs is also an important topic for research on adolescent reproductive health. Here again, a gender-based approach is crucial, not least because of a sexual double standard in most cultures that allows men, but not women, to have multiple sexual partners. Such inequality in sexual power and control often leaves young girls helpless to protect themselves against infection as well as unwanted pregnancy. In this context, understanding the sexual atti- tudes and behaviour of the male partner becomes essential to any effort aimed at reducing HIV/AIDs transmission (WHO, 1995d) Another emerging area for research is the role that coercion and sexual abuse play in the scenario of adolescent reproductive health. Each of the studies in this section revealed cases of forced sex, in the form of rape, incest, or coercion by an employer. (Other studies in this volume also reveal that rape is not uncommon, especially among young girls; see, e.g., the study in Mexico by Elu.) Other available data indicate high rates of childhood and adolescent sexual abuse in many parts of the world, which has been shown to increase a young girl's risk of getting pregnant as well as contracting infec- tion, for a variety of psychosocial reasons (Heise, 1994; Heise et al., 1994). Final Comments In all the studies in this volume, abortion clearly emerges as a prevalent and persistent threat for many women of reproductive age, regardless of their particular socioeconomic and cultural background, and the policy context of the country in which they live. The sections and chapters that follow illustrate simply and undeniably that women will adopt strategies to manage their reproductive life using whatever resources are available to them, even if such approaches mean risking their life and challenging or opposing oppressive social and legal systems that limit their reproductive rights. The evidence in this volume shows that when con- traceptive methods are not available or acceptable, or fail, people take other · steps to regulate their fertility. Women are not passive agents in their repro- ductive destinies. The lengths women will go to in order to end an unwanted pregnancy and the social networks and methods they use to solve such a critical reproductive health problem become important clues to understanding what people's true Introduction t 51 fertility regulation motivations, needs and preferences are. More research needs to be done along these lines. These studies also point the way towards creating more responsive repro- ductive health services that help women and men manage their fertility more effectively. Wider access to safe abortion and the essential elements ofhigh quality abortion care have also been clearly outlined. Even where the actual provision of safe abortion is restricted or prohibited, there is an urgent need for essential abortion-related services of high quality, such as effective and compassionate treatment of women suffering from complications of unsafe procedures and post-abortion counselling to prevent repeat abortion. An essential conclusion is that family planning and abortion services should be better linked through effective referral systems, to ensure that women who have an abortion also receive competent support to help them manage their fertility. For these improvements in the quality of care offered by reproductive health and abortion services to be successfully implemented, it is necessary to solicit and obtain the full cooperation and participation of the health care providers who have daily contact with women and their families. These stud- ies also challenge the ethics of withholding the simple medical technology and knowledge required to perform a safe abortion from these providers, who are faced with the demand for abortions and the life-threatening compli- cations of unsafe procedures every day. The Beijing Platform for Action fully endorses these conclusions, urging 'all governments and relevant intergovernmental and nongovernmental orga- nizations to strengthen their commitment to women's health, to deal with the health impact of unsafe abortion as a major' public health concern and to reduce the recourse to abortion through expanded and improved family plan- ning services . .' (United Nations, 1995: para 12.17). It calls for ready access to reliable information and compassionate counselling to women who have an unwanted pregnancy, as well as the prompt offering of post-abortion counselling, education and family planning services. It also calls for a review of laws that contain punitive measures against women who have undergone illegal abortions (United Nations, 1995: para 107[k]). None of the researchers in this volume nor the thousands of women inter- viewed in these studies have advocated abortion as a method of family plan- ning. No one has questioned that contraception remains a better alternative to abortion. But there is no perfect method and no perfect delivery system, and there is not likely to be one anytime soon. There will always be failures, whatever the reason. And millions of women will resolve these failures with the use of abortion. They should not be condemned. They should not be coerced into motherhood. And they should certainly not be forced to suffer or to die if the alternative is a relatively simple, inexpensive procedure that poses essentially no risk to their health when performed under safe condi- tions by adequately trained health personnel. 52 t Cynthia lndriso and Axel/. Mundigo t References Baird, T.L., R.E. Gringle and EC. Greenslade. 1995. MVA in the treatment of incomplete abortion: Clinical and programmatic experiences. Carrboro, North Carolina: IPAS. Barbosa, R.M. and M. Arilha. 1993. The Braziiian experience with cytotec. Studies in Family Planning 24 (4): 236-40. Benson, J., A.H. Leonard, J. Winkler, M. Wolf and K.E. McLaurin. 1992. Meeting women's needs for post-abortion family planning: Framing the questions. Issues in Abortion Care 2. Carrboro, North Carolina: IPAS. Bruce, J. 1990. Fundamental elements of the quality of care: A simple framework. Studies in Family Planning 21 (2): 61-91. Coeytaux, EM., A.H. Leonard and C.M. Bloomer. 1993. Abortion. In: M. Koblinsky, J. Timyany and J. Gay (eds), Women: A global perspective. Boulder, Co.: Westview Press, pp. 133-46. Dixon-Mueller, R. 1990. Abortion policy and women's health in developing countries. Interna- tional Journal of Health Services 20: 297-314. Fathalla, M., A. Rosenfield and C. Indriso (eds). 1990. Reproductive health: Global issues, Vol. 3, Chapter 6. In: The FIGO manual of human reproduction, 1990. United Kingdom: Par- thenon Publishing. Frejka, T. 1993. The role of induced abortion in contemporary fertility regulation. Proceedings of the International Population Conference of the International Union for the Scientific Study of Population, Montreal, Canada, Chapter 5, pp. 209-14. Hausknecht, R.U. 1995. Methotrexate and misoprostol to terminate early pregnancy. The New England Journal of Medicine 333 (9): 537-40. Heise, L. 1994. Gender-based violence and women's reproductive health. International Journal of Gynecology and Obstetrics 46 (2): 221-29. Heise, L.L. with J. Pitanguy and A. Germain. 1994. Violence against women: The hidden health burden. Washington, D.C.: World Bank. Henshaw, S.K. and E. Morrow. 1990. Induced abortion: A world review, 1990 Supplement. New York: The Alan Guttmacher Institute. Holmgren, K. 1994. Legal abortion during very early pregnancy: Women's experiences and ethical conflicts. Stockholm: Karolinska Institute. Klitsch, M. 1991. Antiprogestins and the abortion controversy: A progress report. Family Planning Perspectives 23 (6): 275-82. Londono E., M.L. 1989. Abortion counselling: Attention to the whole woman. InternationalJour- nal of Gynecology and Obstetrics (Suppl. 3): 169-74. Nasser, J. 1989. Commentary on pain management during abortion from a Latin American phy- sician's perspective. International Journal of Gynecology and Obstetrics (Suppl. 3): 141-43. Neamatalla, G.S. and C.S. Verme. 1995. Postabortion women: Factors influencing their family planning options. AVSC Working Paper (9): 1-11. New York: AVSC ( Access to \bluntary and Safe Contraception). PATH (Program for Appropriate Technology in Health). 1989. Summary of Findings from Oral Contraceptives Focus Group Discussions: Report to USAID/Cairo. Washington, D.C.: PATH. Ping, Tu. 1995. IUD discontinuation patterns and correlates in four counties in North China. Studies in Family Planning 26 (3): 169-79. Riley, A.P., M.K. Stewart and J. Chakraborty. 1994. Program· and method-related determinants of first DMPA use duration in rural Bangladesh. Studies in Family Planning 25 (5): 255-67. Rogow, D. and S. Horowitz. 1995. Withdrawal: A review of the literature and an agenda for research. Studies in Family Planning 26 (3): 140-53. Romania Ministry ofHealth.1991. Mortalitatea materna. (Unpublished document.) Bucharest: Directorate for MCH. Rosenfield, A. 1994. Abortion and women's reproductive health. International Journal of Gyne- cology and Obstetrics 46: 173-79. Introduction t 53 Ross, J.A. and E. Frankenberg. 1993. Findings from two decades of family planning research. New York: The Population Council. Senanayake, P. and M. Ladjali. 1994. Adolescent health: Changing needs. In: World report on women's health. International Journal of Gynecology and Obstetrics 46 (2): 137-43. Stubblefield, P.G. 1989. Control of pain for women undergoing abortion. International Journal of Gynecology and Obstetrics (Suppl. 3): 131-40 Sundstrom, K. 1993. Abortion: A reproductive health issue. Washington, D.C.: Population, Health, and Nutrition Department, The World Bank. Tietze, C. 1981. Induced abortion: A world review. 4th Edition. New York: The Population Council. Trottier, D.A., L.S. Potter, B.A. Thylor and L.H. Glover. 1994. User characteristics and oral con- traceptive compliance in Egypt. Studies in Family Planning. 25 (5): 284-92. United Nations. 1989. Adolescent reproductive behavior: Evidence from developing countries. Vol. II. New York: Department of Economic and Social Affairs. ---. 1993. Distribution of the world populations, 1992 Revision. New York: Dept. of Eco- nomic and Social Development, United Nations (ST/ESNSER.N134). ---. 1994. World abortion policies. New York: Department for Economic and Social Informa- tion and Policy Analysis, Population Division. ---. 1995. Beijing declaration and platform for action. Fourth world conference on women: Action for equality, development and peace. Beijing, China, 4-15 September 1995. van Look, P.V. and H. von Hertzen. 1993. Emergency contraception. British Medical Bulletin 49 (1): 158-70. WHO. 1989. The health of youth: Facts for action: Youth and STDs. Geneva, Switzerland: WHO. ---. 1991. Maternal mortality: A global factbook. Geneva, Switzerland: WHO. ---. 1994. Abortion: A tabulation of available data on the frequency and mortality of unsafe abortion. 2nd Edition. Geneva, Switzerland: WHO (WHO/FHE/MSM/93.13). ---. 1995a. Complications of abortion: Technical and managerial guidelines for prevention and treatment. Geneva, Switzerland: WHO. ---. 1995b. Quality of care: Doing things the right way. Safe Motherhood Newsletter 17 (March-June): 4-6. Geneva, Switzerland: WHO. ---. 1995c. Adolescent health and development: The key to the future. Geneva, Switzerland: WHO (WHO/ADH/94.3 Rev.I). ---. 1995d. Partners in health: Men's role in women's reproductive health. A working paper prepared for the 11th Commonwealth Health Minister's Conference, Cape Town, South Africa, December 1995. PART I The Relationship between Abortion and Contraception 2 Prevention of Pregnancy in High-Risk Women: Community Intervention in Chile Ramiro Molina, Cristian Pereda, Francisco Cumsille, Luis Martinez Oliva, Eduardo Miranda and Temistocles Molina Introduction Maternal mortality due to unsafe, illegal induced abortion is a major repro- ductive health problem affecting developing countries. When abortion is practised clandestinely, the real number of abortions that take place is diffi- cult to estimate. This is the case in most of Latin America where, with the exception of Cuba, Guyana and Barbados and Belize in the English-speaking Caribbean, abortion is not permitted (Paxman et al., 1993). Recent estimates for Latin American countries reveal that there are from three to four clandestine abortions for every 10 births in Brazil and Colombia and two for every 10 births in Peru (Singh and Wulf, 1991 ). For metropolitan Santiago, Chile, estimates based on a fertility survey show that the abortion ratio is about four per 10 livebirths-all of them performed clandestinely (APROFNCERC, 1990; Requena, 1990). In general, the objective of health authorities everywhere should be to lower abortion by increasing effective contraceptive use among couples who are sexually active and do not desire a pregnancy. In some countries where abortion is illegal, including Chile, effective family planning programmes have successfully increased the prevalence of contraceptive use, followed by a decrease in hospital reporting of abortion complications and a marked decrease in maternal mortality due to induced abortion (Molina et al., 1990). These decreases in abortion-related maternal mortality can sometimes be linked to improved access to services, including availability of intensive-care units, and better medical techniques to treat septic shock and other 58 t Ramiro Molina et al. complications. Similarly, another reason for such a decline may be the increased 'professionalism' of private clandestine abortion services that have displaced the traditional backstreet abortionists. Offering women a safer abortion in hygienic clinical environments, although still illegal, has contrib- uted to reducing infection, septic shock and other complications that tradi- tionally account for deaths due to illegal abortion (Singh and Wulf, 1993). Another factor that may explain the decline of deaths due to induced abor- tion complications, however, is simply improper hospital recording of cases of induced abortion. This occurs when the law is very restrictive, with patients as well as hospital staff punisheble by law. In such instances, it is more likely that personnel will write a different diagnosis in the file, most commonly 'spontaneous abortion'. The clandestine nature of induced abortion in many developing countries makes it a difficult subject for research. Case identification is a common problem. Moreover, it is difficult to undertake studies with cases and controls that allow a better assessment of the relationship between contraception and induced abortion. Equally difficult is to clearly demonstrate the preventive effect of family planning in lowering unsafe abortions. Studies of this nature could be very useful in strengthening family planning programmes and in developing more effective strategies for the prevention of clandestine abor- tion (World Bank, 1993). Research in the area of maternal and child health, including maternal mor- tality, has led to the identification of risk indicators that permit health staff to detect women who need special care to prevent the negative outcomes associ- ated with high-risk pregnancy, such as high parity, history of perinatal or early infant mortality, pregnancy at a young age or over 35 years. The application of a risk criteria in maternal care has made it possible to concentrate health actions and apply scarce resources to specific groups, with great success (PAHO, 1986). Similarly, studies on induced abortion have shown that there are identifi- able groups of women in a population that consistently choose abortion to regulate their fertility, among them women with poor knowledge of sexual and reproductive functions, and women who are strongly opposed to using modern contraception for a variety of reasons (Weisner, 1990). Moreover, findings from fertility surveys and descriptive epidemiological studies have identified certain variables associated with induced abortion risk. But it is very difficult to apply a predictive risk criteria to individual cases during rou- tine service work, such as primary health care visits. Data from these abortion risk studies, moreover, have not been used to identify the underlying risk fac- tors that could help predict abortion behaviour. This is because most study designs do not include case controls, or survey data based on reproductive histories are affected by memory recall problems that make past events unre- liable for predicting future behaviour (Barreto et al., 1992). Community Intervention In Chile t 59 Another approach, although more difficult, is to assess abortion risk among women of fertile age, according to information about previous unwanted pregnancies. This is based on the view that abortion behaviour is directly related to negative attitudes towards a pregnancy. Moreover; the decision of going ahead with pregnancy termination will depend on various factors, one of them being the availability and accessibility of abortion ser- vices. In countries where abortion is illegal, for example, often the woman has no other recourse but to accept her unwanted pregnancy and have the child (Viel and Pereda, 1991 ). In such situations, general information on previous unwanted pregnancies is not a good enough predictor of abortion; what is needed is information on previous unwanted pregnancies that were volun- tarily interrupted. This is a more appropriate indicator of future abortion risk, as behaviours tend to repeat themselves over a lifetime. In summary, a preventive strategy that concentrates on early detection of abortion risk, within a specified population known for its high incidence of induced abortion, would seem an appropriate, albeit difficult, means to lower abortion. We believe that such a strategy is worth pursuing, particularly if applied to such populations of low socioeconomic level and in countries where abortion is illegal and unsafe. This project set out to do exactly that. Study Objectives The main objective of this study is based on the hypothesis that by improving family planning services, and by focusing them on women identified as having a high risk of induced abortion, it is possible to achieve a significant reduction in abortion rates in areas known for their high abortion incidence. We would test this hypothesis in low-income urban communities where we knew induced abortion rates to be high. The basic study design consisted of a 'before' survey to identify high-risk women, followed by an intervention. The intervention we were planning included improved services in two communities; and in one of these two, an additional effort to visit women, identified as having a high risk of abortion, in their homes. A third community would serve as a control area. An 'after' measurement would follow to verify whether we had succeeded or not. In general, our aim was to determine the preventive effect of a sustained family planning intervention on women previously identified as having a high risk of abortion. We were also interested in comparing women who were exposed to different levels of abortion risk, and observing their behaviour when they had been subjected to different degrees of intervention intensity. One of the main challenges arising from our research objectives was pre- cisely how to identify women with high abortion risk. Therefore, our first requirement was to develop a simple predictive instrument that would permit us to assess such risks. Next, we had to apply the instrument to all women of 60 t Ramiro Molina et al. fertile age in the communities selected for the intervention to identify those with high risk of abortion. To achieve this initial goal, it was first necessary to undertake a detailed retrospective survey that would provide us with the nec- essary information on their reproductive histories, including abortion, to build the predictive instrument. Research Methodology To meet our stated objectives, we selected for our study the three communi- ties of Pincoya, Cortijo and Quinta Bella, all situated in the northern sector of Santiago, the capital of Chile. The three communities had comparable char- acteristics, including: low socioeconomic level; similar population profiles; equal access to public means of transportation; similar primary health care facilities provided by the Ministry of Health; and sufficient geographic sepa- ration to prevent contamination of the intervention effect across communi- ties, either through personal communication or by seeking health care ser- vices in a neighbouring community. We realized from the start that this would be a long project, taking several years, and would require careful planning for its various stages. We divided the project activities into five clearly distinct phases, which we explain next and which correspond, approximately, to the structure of this chapter. Phase I included the planning of the project, the selection of the three communities to be included, and the development of questionnaires and sampling designs for the retrospective survey, the first field activity. This phase included carrying out the 'before' retrospective smvey that was neces- sary to document the existing abortion and fertility patterns of these communities. Phase II consisted of developing a predictive instrument to detect abortion risk based on the information obtained from the retrospective survey. The predictive instrument was tested and then applied to all three communities. As a result, women were classified into two groups: those with high risk of abortion (HRA); and those with low risk of abortion (LRA). During Phase III, we designed the intervention and selected the two com- munities where it would be applied, and the third that would be the control area. To facilitate the identification of women with different abortion risks, an epidemiological map was prepared that included all households in these communities. Phase IV was devoted to the implementation of the intervention in the two selected communities. The intervention period lasted 18 months. Finally, Phase V assessed the impact of the 18-month intervention by means of a post-intervention, or 'after' survey. This second survey covered the same women that were surveyed originally in all three communities. The study con- cluded with the analysis of a vast amount of information. Community intervention In Chile t 61 Phase I: The First Survey Our first activity was to develop and test the survey questionnaire, which was developed following a detailed analysis of previous survey interviews utilized in abortion studies conducted in Chile. During this phase, we paid particular attention to items directly related to abortion issues (Armijo and Monreal, 1986; Faundes et al., 1968; Pereda, 1986; Requena, 1965). The study questionnaire was tested three times by trained staff, both in a community and in a hospital context, with women who did or did not have a clinically confirmed abortion. Our first draft took 45 minutes to apply. Tests using double blind approaches were conducted in one community with women who had a known background of abortion, but with the interviewers unaware of what the abortion status of the woman was. The information obtained in this test matched very closely the hospital data and we considered the results very reassuring. The final survey questionnaire was greatly improved as a result. Once the field data collection phase started, the completed questionnaires were entered into computers immediately following the field interview and supervisor verification. This procedure facilitated the correction of informa- tion. Errors were thus easily rectified when detected by the data codifiers or during the process of data cleaning, because it was easy to return to the origi- nal case, if necessary, to clarify the error. We used trained interviewers, mostly nurse-midwives, who worked intensely to complete the fieldwork in the shortest possible time. One decision we took was to interview on Satur- days and Sundays, which helped to complete the survey in a shorter time. On weekends it was also easier to contact women who were absent from the household during the week. Two coders entered the information into com- puters and they were, in turn, supervised by a data manager. One of the inno- vative aspects we included during the fieldwork was a mobile office that was parked near the blocks where the interviewers were deployed. This 'office' was an old school bus we rented from the university for the duration of the fieldwork. It allowed us to achieve substantial reductions in the field costs of the project. The field supervisors used this office to do the first data checking and to assess on the spot the progress of the fieldwork. Sampling From the start we decided to conduct the study in a lower-income area of Northern Santiago where clear physical demarcations, such as a large ceme- tery, separate specific communities within this large urban area. The three communities of Pincoya, Cortijo and Quinta Bella were randomly selected. To calculate the sample size needed for the initial survey, an incidence of 70 induced abortions per 1,000 women of reproductive age (WRA) was 62 t Ramiro Molina et al. estimated for the previous two years in the general area where we planned to conduct the study. This figure was estimated on the basis of discharges due to abortion complications from two hospitals that cover the area. A decreasing assumption, to 55 abortions per 1,000 WRA, was estimated for the second survey, which represented the change we expected after 18 months of inter- vention. The result was a sample size of 2,000 women to be interviewed in each community, or a total sample of 6,000 women. To our surprise, the ini- tial survey found a rate of 98 abortions per 1,000 WRA, based on abortions reported for the previous two years before the interview. To select 2,000 women in each community, local area information derived from the 1982 Census and made available by the National Institute of Statis- tics (INE) was used. The census tract data included detailed maps for each of the three communities we had selected. Although the existing maps were helpful, we had to redraw them with updated information that permitted us to identify every house in these communities. Within each community, the blocks were randomly selected, then the houses in each block were equally selected, until 2,000 households were obtained; thus providing us with the necessary total to interview 6,000 women of reproductive age. To avoid problems of information exchange within households, we decided to interview only one woman per household. To replace households without a woman of fertile age, we made a second, and even a third selection of households from blocks not included in the original sample, following the same random procedure. With these successive rounds, a total of 6,020 households were selected and 5,842 women were actually interviewed. Thus we arrived at 97.6 per ceht of our intended total sample. From this group we eliminated 12 women who provided incomplete information and 828 women who were not sexually active (had never engaged in sexual intercourse) and had as a consequence no abortion risk. Therefore, we ended up with a study sample of 5,014 women. The Interview Process The woman to be interviewed in each household was identified following a procedure recommended by the Survey Research Center of the United States. A page listing each woman in the household, in descending order of age, was attached to each interview questionnaire. In each case, the woman to be interviewed was pre-selected from the household list utilizing a table of random numbers. A different random numbers series was used for each household. This procedure prevented the interviewer from choosing who to interview or from applying other selection criteria. A further advantage was that it did not over-enumerate women who stayed at home during the day: women who worked outside the home were left a message in writing with a specific time and date when the interviewer would return. Community Intervention In Chile t 63 Forty interviewers participated in the field phase. All were female and most of them were midwives, except for two social workers and three high school teachers. They received a 10-day training course from the principal investigator and five field supervisors, all of whom had ample survey experi- ence. The same team conducted the second survey. Some Results from the First Survey Some of the sample characteristics for each community are shown in Table 2.1. One of the important findings of the first survey was that women did report their abortion experience. When they were asked what type of abor- tion they had, however, most women, 72 per cent, reported their abortions as spontaneous. This proportion was more than twice the highest proportion reported in the literature for a population (Singh and Wulf, 1993). This meant that there was a clear under-reporting of induced abortion, most prob- ably due to fear of punitive action. As noted earlier, a similar declaration hap- pens in hospitals for the very same reason. 'fable 2.1: Sample Characteristics for Each Community Variables Pincoya Cortijo Quinta Bella Average age (in years) 32 33 33 Marital status: % legally married 67 74 69 % cohabiting 18 12 14 Marital duration: > lOyears 38 39 39 Years of schooling (average) 8 9 9 Residence in Santiago: 10 years or more (%) 95 95 95 Housing: House or apartment (%) 81 91 84 Rooms, dilapidated(%) 19 9 16 Living children 2.5 2.3 2.2 Total number of women 1,613 1,620 1,674 We had expected this situation, because abortion is a criminal act in Chile. Therefore, our decision of first asking if the woman had ever experienced an abortion, ancl then for the type, seemed the correct method of inquiry. Since it was obvious that most reported abortions were not or could not be sponta- neous, we simply decided to put all reported abortions into one cate- gory-'abortion'-without attempting to separate events that could not be identified one way or another. 64 t Ramiro Molina et al. Phase II: Developing the Predictive Instrument During this phase of our study, we set out to develop a predictive instrument that would be sensitive to abortion risk. The first criteria was to analyze the abortion history of the women interviewed during the first survey. We had to use the past as the best predictor of future behaviour and, therefore, try to identify which variables or factors were closely associated with any previous abortion. As noted, the data were biased towards abortions reported as spon- taneous, although this was a systematic bias that applied equally to all three communities. For women with or without abortion history, key independent variables were compared. To establish a better basis for comparison with the results of the second smvey, which was to take place two years later, we decided on a second criteria: to consider a two-year reference period for abortion for both surveys. This made sense since the intervention that would follow the first survey was expected to last two years. In this way the 5,014 women selected for inclusion in our analysis during the first survey were divided into three groups: (a) Those without history of abortion-3,492; (b) Those with history of abortion two years or more before the survey (before 1987)-1,270; (c) Those with history of abortion during the two years preceding the survey (1987-88)-252. The classification into these three subgroups allowed us to learn about abortion during two reference periods: (J) abortion that occurred in the more distant past, that is, before 1987; and (2) more recent abortions, during 1987 and 1988. We wanted to see whether background variables (e.g., age, marital status, etc.) for these subgroups exhibited different trends given that the predictive risk could be differentially affected by time or changes since the occurrence of the last reported abortion. We found that the variables analyzed for each group showed the same effects over time. Nevertheless, the associative tendencies of the statistically significant variables were more evident in the group with more recent abor- tion. Given these results, a third criteria for the analysis was adopted; the 5,014 women were separated into two groups: (a) Women with history of induced, spontaneous or non-specified abortion in the past two years-252; (b) Women without history of abortion in the past two years (includes women with longer history of abortion in the past)-4,762. A comparative univariate analysis was perfoi:med on these two subgroups to find the significant variables that were later included in the multivariate analysis necessary for the development of the predictive instrument. Multivariate Analysis A first selection of significant variables associated to abortion risk was made. Almost all the variables in the questionnaire were first considered, each of Community Intervention In Chile t 65 them independently, and some in combination, in order to analyze the differ- ences between the women in group (a) and group (b), as defined earlier. The odds ratio (OR) of the most significant variables was calculated with limits of confidence of 95 per cent. To select the significant variables a value of p smaller or equal to 0.05 was accepted. After this initial step, a logistic regression was performed. The logistic regression including 4,907 women was performed using SAS software. The regression model selected the variables that were closely asso- ciated with the risk of abortion, with a p value ofless than 0.10. We eliminated 107 cases because of incomplete information on some variables required by the regression model. The logistic regression showed that when nine variables from those selected in the previous analysis were sequentially introduced into the model, they fulfilled the conditions of increasing, in a significant way, the association with abortion in the last two years. These variables and their categories were: age, below 35 years; marital status, cohabiting; union duration, less than 10 years; previous unions, several; sterilization, not contemplating; contracep- tion, used mostly traditional methods; use of IUD, never used; housing, living in single rooms, or shared or dilapidated houses; fertility, three or more chil- dren, the last under five years of age (see Table 2.2). Thble 2.2: Abortion Risk Factors Mzriable Age Partner relations Sterilization Previous use of methods Marriage duration Previous marital cohabitation Previous IUD use Housing type Fertility level • significant at the 0.05 level. •• significant at the 0.001 level. Item Beta Coefficient Under35 o.os•• Cohabiting 1.40** No 1.36** Some spacing••• 0.40* 10 years or less 1.18** Yes 0.58** No 0.42* Poor, dilapidated 0.31* 3-10 children alive, last born under 5 years 0.32* J.1zriable Number 1 2 3 4 5 6 7 8 9 • ** methods include monthly injectables, rhythm, condom, withdrawal and vaginal douche. With these nine variables, we then had the elements of a predictive instru- ment which, with the assistance of a regression model, allowed us to sort out the women in the sample according to their abortion risk. The category that included women with high risk of abortion (HRA), included those cases whose probability of having an abortion was higher than average for the sam- ple, which was found to be 5.03 per cent. As a result, 1,911 women were classi- fied as being in the high-risk category and 3,056 fel1 into the low-risk category 66 t Ramiro Molina et al. Table 2.3: Actual and Predicted Abortion Values Predictive Value Abortion Status High Abortion Low Abortion Total Risk Risk Actual Value With abortion experience* 184 63 247 No abortion experience 1,727 2,993 4,720 Total 1,911 3,056 4,%7 Note: The predictive value was obtained by the application of the nine variables shown in Table 2.2. * Refers to abortion experience during the previous two years. (Table 2.3). For all three communities we now had an estimate of which women had a high or a low risk of induced abortion. The women at high risk of abortion were distributed according to their community of residence as follows: 36 per cent of the high risk women (681) lived in Pincoya, where they represented 42 per cent of the total sample of all women of reproductive age in that community; another 30 per cent (581) Jived in Cortijo, where they represented 36 per cent of the total sample of women of reproductive age in that community; and 34 per cent (649) lived in Quinta Bella, where they represented 39 per cent of the sample of women of reproductive age in that community. High-risk women living in Pincoya were selected to receive 'full' intervention; those in Cortijo, 'partial' intervention; and those in Quinta Bella, the control community, nothing at all. Phases Ill and IV: The Intervention The Planning Phase During the third phase of the study we planned the intervention and selected Pincoya and Cortijo as the communities to be intervened utilizing a random procedure. Both communities were provided with additional staff resources for their clinics, including a part-time physician, a midwife and an auxiliary nurse, the latter two full-time. In addition to improvements in service quality, a wider choice of contraceptive methods was made available. The only differ- ence between the two communities was that the 'full' intervention in Pincoya was designed so that all women classified as having a high abortion risk were visited in their homes by a specially trained social worker to discuss their fam- ily planning and related reproductive health needs. As already noted, in Quinta Bella, the 'control' community, there was no intervention at all. The Implementation Phase The 'full' and 'partial' interventions in Pincoya and Cortijo, respectively, were initiated in September 1989 and lasted 18 months, ending in February 1991. The actual intervention constituted the fourth phase of our project. Community Intervention in Chlte t 67 In Pincoya, each woman classified as having a high risk of abortion was identified by name and address from a confidential list obtained from the computer listings. A social worker visited each one to see if they wished to discuss any reproductive health needs, such as information and services to regulate their fertility. (Women who did not work at home were visited on Saturdays and Sundays.) The women were also informed of the improved family planning services being made available in the community clinic. The social worker focused on motivation and basic education on contraception, and arranged for appointnents at the local family planning clinic, if requested. The midwife assigned by the project to the Pincoya clinic registered compli- ance with all appointments arranged by the social worker. The social worker and midwife worked closely together. The service improvement in the second community, Cortijo, was imple- mented as in Pincoya, but no special effort was made to advertise it. Word of mouth was expected to be sufficient for the purposes of the study. Phase V: The Second Survey The second, or 'after', survey was carried out 24 months following the com- pletion of the first survey and it marked the final, or fifth phase, of the study. Interviewers were able to locate 73 per cent of the women included in the first survey: 3,588 women out of the 4,967 in the first survey, for all three communities. The loss to follow-up in the three communities was: 16.9 per cent in Pincoya; 32.3 per cent in Cortijo; and 31.2 per cent in Quinta Bella. Most of the loss was due to change of address, as many people in these communities are migrants and tend to move around in the various communities in the periphery of Santiago. Additionally, the presence of 'allegados', or people who do not have a stable home and come to live temporarily with relatives or friends, added to the problem. There was a rather high proportion of such cases by the time of the second survey. Also, the population in these commu- nities is young and prone to fluctuation. While 3,588 women were interviewed in the second survey, the analysis included only 2,991 women because we excluded 597 women due to incom- plete information (284) or due to absence of sexual intercourse between sur- veys (313). Moreover, to observe the change in behaviour among women who were true potential users of contraception between the first and second sur- veys, we had to create a sub-sample that eliminated an additional 735 women who had been sterilized at the time of the first survey or who had become sterilized or pregnant by the time of the second survey. This resulting sub-sample of women 'exposed' to pregnancy and abortion was considerably reduced to 2,256 . In each of the communities, the breakdown for the various analysis groups was: in Pincoya, we interviewed 1,086 women, eliminated 233 who were 68 t Ramiro Molina et al. either sterilized or pregnant; and of the remaining 853 'exposed' women, 43 per cent (366) were found to have a high risk of abortion. In Cortijo, we inter- viewed 936 women, eliminated 276, and of the remaining 660 'exposed' women, 35 per cent (233) were in the high-risk category. Lastly, in Quinta Bella, we interviewed 969 women, eliminated 226, and of the 743 'exposed' women, 38 per cent (285) were found to have a 'high risk' of abortion. The criteria used for the classification of women with high risk of abortion in the second survey was exactly the same used in the first survey (women identified as having high risk of abortion in the first survey maintained that classification in the second). However, we felt this approach could introduce a bias in our analysis. To assess the potential for such bias, the profile of the nine main abortion risk factors of the women lost to follow-up (1,916) was compared with that of the women we were able to re-interview (2,991). The factors that behaved differently were: age, previous contraceptive use, mar- riage lasting less than 10 years and fertility level. These differences, although marked, were not significant, except for the fertility variable, which did show a significant difference. Later, we used these factors to adjust abortion rates in the before-after comparisons, and thus reduced the possibility of any bias (Table 2.4). In the main analysis itself, for each risk factor, a significant difference could be observed among the three communities. However, differences are much smaller between women included and excluded from re-interview in the second survey, within each community. As a result, we feel that Tobie 2.4: Abortion Risk Factors: Women Included in the Second Survey Compared with Women Excluded from the Analysis Variable* Pin.coy a Cortijo Quinta Bella lnclu- Exclu- p lnclu- Exclu- p Inclu- Exclu- p ded ded Value ded ded Value ded ded Value Age 57.1 67.6 0.0 53.1 61.1 0.001 57.4 64.0 0.01 Partner 17.9 18.0 0.93 11.3 12.6 0.44 12.5 16.7 0.01 Sterilization 90.8 89.9 0.58 84.2 89.9 0.00 87.7 90.4 0.09 Previous use 42.2 43.3 0.89 46.3 47.4 0.65 47.4 45.2 0.39 Marriage: > 10 years 26.6 32.1 0.02 29.7 35.4 0.01 29.9 31.6 0.45 Cohabitation 16.3 16.9 0.76 11.0 12.0 0.33 10.8 13.3 0.11 No IUD 27.1 33.0 O.Ql 27.7 36.8 0.00 29.9 34.2 0.06 Housing 17.5 22.6 0.01 9.4 8.5 0.52 14.2 17.7 0.05 Fertility 19.4 19.2 0.89 18.1 14.0 0.03 17.3 18.6 0.51 Total 1,086 527 936 684 969 705 Note: 'Included' women are those re-interviewed in the second survey; 'excluded' women are lost to follow-up cases for whom background characteristics data were available from the first survey. • For details on variable names and the meaning of each item selected as risk factor please refer to Table 2.2. Community Intervention In Chile t 69 comparisons between the first and the second survey are valid even though some of the women interviewed in the first survey could not be included in the second. Evaluating the Intervention Before and After Effects Our second, or 'after' survey was designed as a comparison of the fertility his- tory of women two years 'before' the first survey (1987-88) to that of two years 'after' the first survey (1989-90). The 18-month intervention period itself fell within this second two-year period. For example, the proportions of abortions in the three communities in the survey 'before' the intervention were compared to the proportions 'after' the intervention, adjusted by the differences in risk factors in the model. Each woman was considered as an observation unit, and these data were entered in an adjusted logistic regression model that considered as dependent variable 'Y' the pregnancies during the previous 24 months. Two 'dummy' variables to represent the three communities were created, and then added as independ- ent variables together with the nine core variables from the original model. Analysis of the interaction of the model was excluded because our interest was to study, by controlling other variables, the relationship between the women's behaviours in each community. The comparison of the proportion of abortions in the three communities found in the second 'after' survey was done in the same way as the compari- son between proportions found in the first and second surveys, but with a modification: the responses obtained in the first survey were included as a control variable. That is to say, the dependent variable that was used in the 'before' analysis was now included as an independent variable in the 'after' analysis. Contraceptive Use In the first survey, women were asked about current use of contraception at the time of the interview. In the second or 'after' survey, current use included use of contraception during the entire period of observation, that is, during the two-year period between the first and the second survey. Given this change in the reference period, we anticipated an increase in contraceptive users in the second survey. Table 2.5 shows a decrease in women using temporary methods in the full-intervention community of Pincoya, from 66.9 per cent to 60.3 per cent, but a marked increase in women using permanent methods, from 2.2 per cent to 13.6 per cent during the two-year period between surveys. Described another way, at the time of the first survey, 7.6 per cent of the women in 70 t Ramiro Molina et al. Tuble 2.5: Changes in Contraceptive Use and Pregnancy Status between the First and Second Surveys, by Community Community and Status at Second Survey: Contraceptive and Pregnancy Contraceptive First Status of Women Classified According to Status in Status Survey the First S!'rvey First Survey Total Users Non-Users Sterilized Pregnant No. (%) No. (%) No. (%) No. (%) No. (%) Pinooya: Users 727 (66.9) 537 73.9 89 12.2 84 11.6 17 2.3 Non-users 253 (23.3) 86 34.0 141 55.7 18 7.1 8 3.2 Sterilized 24 ( 2.2) 24 100 Pregnant 82 ( 7.6) 32 39.0 22 26.8 22 26.8 6 7.4 All women 1,086 (100) 655 60.3 252 23.2 148 13.6 31 2.9 Cortijo: Users 648 (69.2) 427 66.0 85 13.1 115 17.7 21 3.2 Non-users 171 (18.3) 49 28.7 99 57.9 15 8.8 8 4.6 Sterilized 50 ( 5.3) 1 2.0 49 98.0 - Pregnant 67 ( 7.2) 50 74.6 16 23.9 - 1 1.5 All women 936 (100) 526 56.2 201 21.5 179 19.1 30 3.2 Quinta Bella: Users 665 (68.6) 470 70.7 99 14.9 78 11.7 18 2.7 Non-users 191 (19.7) 52 27.2 122 63.9 10 5.2 7 3.7 Sterilized 55 ( 5.7) 55 100 Pregnant 58 ( 6.0) 31 53.3 22 37.9 4 6.9 1 1.7 All women 969 (100) 553 57.1 243 25.0 147 15.2 26 2.7 Pincoya were pregnant; at the time of the second survey 26.8 per cent of these same women had chosen sterilization. The most dramatic change, in all three communities, in the period between surveys, was the increase in the number of women who opted for female ster- ilization. For all three communities, it jumped from 4.3 per cent to 15.8 per cent in this short period. As a result, there were fewer women pregnant in the second than in the first survey, in all three communities. Total contraceptive prevalence, both the use of permanent as well as other methods, increased from 69.1 to 73.9 per cent in Pincoya; and from 74.5 to 75.3 per cent in Cortijo. It decreased, however, from 74.3 to 72.3 per cent in Quinta Bella, the control community, despite the strong increase in female sterilization in that community during the same period. Although Pincoya showed the largest increase in contraceptive use, it was, to a large extent, because of a switch to sterilization. Abortion Rates and Ratios for Women of Reproductive Age Abortion rates and ratios are presented for the total sample and for high-risk women in all three communities, in Tables 2.6 and 2.7. Community Intervention In Chile t 71 Tobie 2.6: Abortion Rates: All Women and High-Risk Women, Before and After the Intervention Community Abortion 1,000 Women, 15-49 Years Before After Al/Women Abortions Rate Abortions Rate %Change Pincoya 64 58.9 26 23.4' -59.4 Cortijo 43 45.9 21 22.4 -51.2 Quinta Bella 54 55.7 37 38.2 -31.4 Total (All) 161 53.8 84 28.1 -47.7 High-Risk Women Pincoya 50 112.6 14 31.5 -72.0 Cortijo 28 93.6 9 30.1 -67.8 Quinta Bella 41 119.9 20 58.5 -51.2 Total(HRW) 119 109.6 43 39.6 -63.9 Tobie 2.7: Abortion Ratios: All Women and High-Risk Women, Before and After the Intervention Community Abortion Ratios, per 1,000 Pregnancies Before After All Women Ratio Ratio % Change Pincoya 135.9 130.0 -4.3 Cortijo 115.9 128.8 +11.2 Quinta Bella 134.0 207.9 +55.0 Total (All) 154.8 155.3 +0.3 High-Risk Women Pincoya 165.0 109.4 -33.7 Cortijo 140.7 97.8 -30.4 Quinta Bella 172.2 200.0 +16.1 Total(HRW) 193.8 134.4 -30.7 Abortion rates decreased by 59 per cent in Pincoya, the community with full intervention. In Cortijo, where the intervention was only partial, the rates also decreased by 51 per cent And in the control area of Quinta Bella, rates decreased by 31 per cent. These decreases in abortion rates are in line with our initial expectations and support our study hypothesis (Table 2.6). With respect to abortion ratios, shown in Table 2.7, the impact of the inter- vention is even more evident. Abortion ratios decreased by 4 per cent in Pincoya; but increased by 11 per cent in Cortijo and by 55 per cent in Quinta Bella. The increase in the abortion ratio in Cortijo, however, is simply a reflection of a larger relative decline in pregnancies than in abortions, although both experienced a decline. As already noted (Table 2.5), in all 72 t Ramiro Mo/Ina et al. three communities the number of pregnancies declined substantially during the period of the intervention, probably due, at least in part, to the popula- tion of women becoming older. The decrease in Pincoya was 58 per cent; in the other two communities, 56 per cent. Tests of significance show that differences between the community with full intervention and the control community are significant, as are the differences between the community with partial intervention and the control area. How- ever, the differences between the communities with full and partial interven- tion are not statistically significant. The implication of these findings is important for policy, because it shows that the added cost of individual home visits, while effective, may not be justified. Abortion Rates and Ratios for High-Risk Women For each of the three communities, we also looked at the women we had clas- sified as having a high risk of abortion. The lower half of Tables 2.6 and 2.7 show the abortion rates and ratios for these women. In Pincoya, abortion rates decreased by 72 per cent; in Cortijo, by 68 per cent; and in Quinta Bella, by 51 per cent. Abortion ratios decreased by 34 per cent in Pincoya and by 30 per cent in Cortijo, but increased 16 per cent in Quinta Bella. As in the group comprising all women of reproductive age, the truly significant differences were between the two communities that received the intervention and the one that did not, but not between the two intervened communities. Abortion Risk and Contraception Our final analysis aims at documenting the complex inter-relationship between abortion, abortion risk, and contraceptive use by level of interven- tion in each of the three communities. This is perhaps the most important step in the analysis of our data, as it provides the most direct answer to the central hypothesis of our study regarding the impact of family planning on abortion. To fully understand the effects of the intervention in terms of behavioural outcomes, we had to consider the different levels of abortion risk among women of reproductive age as well as their contraceptive status. To this effect, as noted before, we eliminated from the analysis pregnant women whose pregnancy was advanced, or who had been sterilized or had a hysterectomy in both the first and second surveys. The purpose was to have a population truly exposed to the risk of pregnancy and abortion, even though this substantially reduced our study sample. The women who received the full intervention of home visits and services in Pincoya, and who, as a consequence, adopted effective methods of contra- ception, showed the largest decrease in abortion rates, 83 per cent, and in abortion ratios, 52 per cent (Tu.hies 2.8. and 2.9). In contrast, women in Community Intervention in Chile t 73 Tobie 2.8: Abortion Rates among High-Risk Women, by Contraceptive Status: Bdore and After fhe Intervention Communities Using Contraception Not Contraception Women Abortions Rate Women Abortions Rate Pincoya: Before 278 28 100.7 88 10 113.6 After 286 5 17.5 80 6 75.0 %change -82.6 -34.0 Cortijo: Before 196 14 71.4 37 7 189.2 After 186 4 21.5 47 4 85.1 %change -69.9 -55.0 Quinta Bella: Before 226 21 92.9 59 14 237.3 After 227 10 44.1 58 7 120.7 %change -52.5 -49.1 Note: Rates are per 1,000 women of reproductive age. Tobie 2.9: Abortion Ratios among High-Risk Women, by Contraceptive Status: Bdore and After the Intervention Communities Using Contraception Not Using Contraception Women Abortions Rate Women Abortions Rate Pincoya: Before 170 28 164.7 42 10 238.1 After 63 5 79.4 23 6 260.9 %change -51.8 +9.6 Cortijo: Before 114 14 122.8 20 7 350.0 After 42 4 95.2 15 4 266.7 %change -22.5 -23.8 Quinta Bella: Before 158 21 141.9 32 14 437.5 After 52 10 192.3 17 7 411.8 %change +35.5 -5.9 Note: Ratios are per 1,000 pregnancies. Pincoya who received the full intervention, but who did not change their con- traceptive behaviour, had a significantly lower decrease in abortion rates of 34 per cent; and an increase in abortion ratios of 10 per cent. The abortion rate decreases are statistically significant when either Pincoya or Cortijo, the intervened areas, are compared with the control area, Quinta Bella. When comparing the 'before/after' behavioural changes in each group, the most significant effect was the shift to more effective contraceptive use in the two intervened communities. Among non-users of contraception there was 74 t Ramiro Molina et al. also a clear decrease in abortion risk in these communities, although not as pronounced as among women using contraception. In Pincoya, there is a greater difference between the decrease in abortion rates between women using and not using contraception than in the other communities (Table 2.8). The pattern of change in abortion ratios among users of contraception matches that of the abortion rates, although among non-users, the pattern is less clear (Table 2.9). When considering relative risk, we found a greater increase in the risk of abortion among non-users of contraception in the community with full inter- vention. The increase in risk is particularly high among women who were classified as having a 'high risk' of abortion, but who did not adopt effective contraception. The increase in relative risk appears to be lower in the par- tially intervened and control communities, which may be explained by the fact that we did not search out and visit the women with high risk of abortion in these areas as we did in the fully intervened community where such risk applied to just a few women. These findings generally confirm that women who do not use contraception have a higher risk of abortion. Discussion and Policy Implications Abortion, regardless of its legal status, is a reality for many women facing unwanted pregnancy. Despite the increased outreach of family planning programmes, abortion continues to exist and has played an important role in fertility declines. Nevertheless, the relationship between abortion and con- traception continues to be poorly understood. Our study, starting with the development of an instrument to identify and predict abortion risk, followed by an intervention designed to apply the instrument in three low-income communities, including a control group, has demonstrated that increasing the prevalence of contraceptive use among women identified as having a high abortion risk does reduce its incidence. The predictive instrument developed to identify abortion risk is perhaps the most innovative aspect of our study. It is easily applicable by health staff with basic training in primary health care and community work. Its applica- tion allows concentrated efforts within segments of a population where pri- mary prevention is needed the most, such as among women of low socioeco- nomic level in developing countries where abortion is illegal and its incidence is high. Because of the difficulties in obtaining totally reliable data on induced abortion, we decided on the alternative strategy of classifyihg all types of abortion, including both spontaneous and induced, under a single category. The levels of spontaneous abortion reported were far too high to be realistic anyway. And by placing all abortions into a single category, we did not under-enumerate the level of induced abortion, which would have risked a major bias. Community Intervention In Chile t 75 Although we were aware that improved methods for measuring induced abortion prevalence have been suggested, such methods require greater ano- nymity than was possible in our intervention study, which necessitated home visits to high-risk women. We felt this was true even though we were not con- tacting them about their past abortion behaviour but rather trying to help them prevent repeat abortion. Our results show that more personalized family planning services, in asso- ciation with risk detection, can succeed in lowering abortion. Definitely the population who received the benefit of the full intervention and who subse- quently increased their use of effective contraception showed the clearest decline in abortion rates, These behavioural changes could be observed, even in a relatively short period of time, which had been one of our main concerns when we planned the study. It is possible, therefore, to suggest that such approaches be incorporated in discussions concerning family planning strat- egy, particularly when the goal is to lower abortion. A risk approach can be more effective than a more typical clinic-based family planning programme in contexts such as those of the communities studied. In experimental situations involving human populations there is always the danger of external interference that can introduce unexpected biases as a result of changes in the larger national or local context. Such changes are often beyond the control of the project staff. We were lucky that there were no major changes in the family planning policy of the Ministry of Health of Chile during the intervention period between September 1989 and February 1991. Although the project did coincide with the discussion of an amendment to eliminate an article in the national health legislation that allowed preg- nancy interruption for therapeutic reasons. The amendment was passed but it is doubtful that this became an issue for discussion among the women in the communities we studied. Another factor that facilitated the study was the important and constant institutional and professional support provided by the health area director in charge of the clinics servicing the communities that were intervened. We were also fortunate that there were no changes in local.health service policies during the period in any of the clinics serving the three communities studied. There was one possibility of an external source introducing unexpected bias, however. One of our own project staff, the social worker assigned to home visits in Pincoya, was a highly motivated woman. She was able to make an average of three visits to each woman classified as having a high risk of induced abortion and she was able to establish a very close relationship with most of them. As the intervention continued, she became an influential per- son in the community and encouraged pregnant women to continue their pregnancies, which eventually resulted in an increase in live births. She also promoted contraception very strongly. Local women may have decided against a clandestine abortion to avoid feeling ashamed in front of this very friendly, yet strong, community worker. Was there also a fear that she niight 76 t Ramiro Mo/Ina et al. denounce them to the authorities if she found out that they had an illegal abortion? This could be a possibility, but it does not seem probable under the circumstances. However, this potential bias cannot be verified. In Cortijo, the partially intervened community, there was no social worker to undertake per- sonal motivation or community work and the outcome of the intervention was, as we have already noted, less successful in terms of reducing the abor- tion rate and ratio. In short, the personal dimension in intervention work is important, albeit difficult to measure. In conclusion, we believe our study has shown that an effective interven- tion, of improved family planning services with personalized inputs, directed at women with high risk of abortion, can be successful in lowering abortion incidence in populations of low socioeconomic status, especially in contexts where abortion is illegal. We may also suggest that by applying a risk strategy in communities with high incidence of illegal induced abortion, it may be possible to lower mater- nal mortality due to complications of unsafe abortion (Sotelo, 1993). The cost of implementing a programme of this type more widely could be recovered from savings obtained from eliminating hospitalizations due to abortion complications. After this initial experimental experience, we recommend that our inter- vention model be tested in other settings. The application of the predictive instrument to identify abortion risk in communities different from those studied in Chile might validate further our results and permit this approach to be extended to primary health care services in other countries where abor- tion is illegal, particularly in the Americas. t References APROFA/CERC. 1990. Encuesta de fecundidad en area metropolitana de Santiago, Final Report, unpublished, Santiago: Chile. Armijo, R. and T. Monreal. 1968. Epidemiologia del aborto en Chile. Communication presented to the Chilean Health Society, unpublished, Santiago, Chile. Barreto, T., O.M. Campbell, L. Davies, V. Faveau, V. Filippi, W. Graham, M. Mamdami, C. Rooney and N.F. Toubia. 1992. Investigating induced abortion in developing countries: Methods and problems. Studies in Family Planning 23: 159-70. Faundes, A., G. Rodriguez and O. Avendano. 1%8. The San Gregorio experimental family plan- ning program: Changes observed in fertility and abortion rates. Demography 5 (2): 836-45. Maramatsu, M. 1988. Japan, In: Paul Sachdev. /ntemational Handbook on Abortion. New York: Greenwood Press, pp. 293-301. Molina, R., C. Pereda, E Cumsille et al.1990. Prevenci6n el embarazo en mujeres con alto riesgo de aborto. In: M. Requena (ed.). Aborto inducido en Chile. Santiago de Chile: Edici6n Sociedad Chilena de Salud Publica, lmpresora CRECES Ltda. Panamerican Health Organization (PAHO). 1986. Manual sobre enfoque de riesgo en la atenci6n materno-infantil. PALTEX: Para ejecutores de prograrnas de Salud, No. 7. Paxman, J.M., A. Rizo, L. Brown and J. Benson. 1993. La epidemia clandestina: La practica del aborto ilegal en America Latina. Perpectivas lntemacionales en Planificacion Familiar. Numero Especial, pp. 9-15. Community Intervention In Chile t 77 Pereda, C. 1986. Factors and predictors of discontinuation of oral contraceptives and IUDs in public health clinics of Santiago, metropolitan area, Final Report to WHO/HRP, Santiago, Chile. Requena, B,M. 1965. Social and economic correlates of induced abortion in Santiago, Chile. Demography 2:2. --. 1990. Aborto inducido en Chile. Santiago, Chile: Sociedad Chilena de salud publica, Impresora CRECES Ltda. Singh, S. and D. Wulf. 1991. Calculo de los niveles de aborto en el Brasil Colombia y Peru a base de datos hospitalarios y de encuestas de fecundidad. Perspectivas lntemacionales en Plani- ficaci6n Familiar. Numero Especial, pp. 14-23. ---. 1993. The likelihood of induced abortion among women hospitalized for abortion com- plication in four Latin American countries. International Family Planning Perspectives 19(4). December: 134-41. Sotelo, M. 1993. El enfoque de riesgo y la mortalidad materna: Una perspectiva latinoamericana (informe especial). Boletin de la Oficina Panamericana de la Salud. 114 ( 4): 289-301. Viel, B. and C. Pereda. 1991. El embarazo no deseado. Resultados de un proyecto apoyado por HRP/OMS, (82046). Boletln de la Asociaci6n Chilena de Protecci6n de la Familia No. 1/12 Enero Diciembre. Weisner, H.M. 1990. Comportamiento reproductivo y aborto inducido en mujeres Chilenas de sectores populares. Una perspectiva antropol6gica. In: M. Requena (ed.). Aborto inducido en ·Chile.Santiago, Chile: Edici6n Sociedad Chilena de Salud Publica, Impresora CRECES Ltda. World Bank. 1993. Inversion en salud. USA: Oxford University Press. 3 Factors Affecting Induced Abortion Behaviour Among Married Women in Shanghai, China Gui Shi-xun Introduction The family planning programme implemented by the government of China in the late 1970s and then strengthened during the 1980s has been very success- ful in Shanghai in at least one sense: the total fertility rate in 1990 reached the very low level of 1.23 children per woman. Moreover, in 1993, Shanghai expe- rienced negative growth, with a birth rate of 6.5 per thousand, and a mortality rate of 7 .3 per thousand, which resulted in a rate of natural increase of -0.8 per thousand (for the population with de jure household registration). Pro- jections for the next century show a continuing negative rate of natural increase for Shanghai (Gui Shi-xun et al., 1994). Despite the high use of contraception in Shanghai, the municipality has traditionally had a higher rate of induced abortion than reported for other parts of China. According to results from the First In-depth Fertility Survey of China (State Statistical Bureau, 1986), the abortion rate for Shanghai Municipality in 1983 was 119.5 per thousand women, a much higherrate than Hebei (50.9 per thousand) and Shaanxi (38.3 per thousand), the other two provinces included in the survey (see also: Gao Er-sheng et al., 1988). While there has been a decline i~ the induced abortion rate for women in Shanghai in recent years, its level still remains fairly high: 68 per thousand in 1991 (Research Team on Induced Abortion, 1994). Undoubtedly, the extensive use of induced abortion, which is legal in China, plays an important role in maintaining such a low level of fertility. This would indicate that there is some urgency for the family planning authorities to lower the rates of induced abortion and the number of unwanted pregnancies. factors Affecting Induced Abortion BehaVlour In Shanghai t 79 Objectives The main purpose of our study was to identify the factors that affect the deci- sion to have an induced abortion among married women in both the urban and surrounding rural areas of Shanghai. We also wanted to identify approaches that could lead to a reduction of the induced abortion rate, espe- cially repeat abortions, to protect women's health. To this effect a survey of social, psychological and demographic factors affecting unwanted pregnancy among married women was carried out in Shanghai in 1991. This chapter presents a preliminary analysis of the findings of this survey. Research Methodology The survey of 10 sites started on 22 April 1991. Six sites in the urban area of Shanghai included: the International Peace Maternal and Child Hospital; the No. 1 Maternal and Child Hospital of Shanghai; the No. 1 People's Hospital of Shanghai; and the Maternal and Child Hospital in Luwang, Changlin and Putuo districts. The survey also included four sites in rural areas: the Mater- nal and Child Hospital in Jiading County, in Chuangsha County and in Song- jian County; and the Center Hospital of Fengxian County. All these hospitals were selected as part of a plan laid out in collaboration with the Chinese Fam- ily Planning Technology Center. A simple random sampling procedure was used to select the women for the survey. The potential subjects were all women who registered at any of the hospitals selected requesting an induced abortion. Women had to have de jure household registration in Shanghai, that is, legal residence in the munici- pality. Based on the statistical data on the estimates of total induced abor- tions for married women in Shanghai in 1990, we arrived at a sample size of 2,806 women ( a. = 0.05). The ratio between urban and rural areas was 65.6/ 34.4, which determined how the sample size was divided. The final study population was reduced to 2,765 women, representing a 98 per cent response rate; 2,760 men were also interviewed. For these subjects, 2,644 women, or 95.6 per cent, were cases with gestations of 'equal to or less than 90 days' (first trimester); and for 121 women, or 4.4 per cent, gestations were 'equal to or more than 91 days' (second trimester). All field personnel, including supervisors and interviewers, were female medical workers who received training in how to conduct face-to-face inter- views. Subjects were interviewed just before and after the induced abortion. In order to guarantee confidentiality and to ensure the quality of the inter- views, a list of married women who had registered for abortion was used to identify the addresses of the study subjects. As a result, in the actual question- naire used for the survey, the name and address of the subjects was not recorded, which ensured complete anonymity. 80 t Gui Shl-xun Study Results Abortion Incidence Among Married Women Abortion by Age Women aged 25-29 years represented the largest proportion, nearly a third, among the 2, 765 married women who had an induced abortion, followed by the 30-34 age group. These two age groups represented 61 per cent of the women in the sample (Table 3.1 ). Given that we knew exactly the proportion that our sample represented of all the women who had an abortion in Shang- hai, we decided to estimate abortion rates by age for the total Shanghai Munici- pality. One problem was the lack of information on married women, by age, according to de jure household registration of Shanghai, in mid-1991. In order to circumvent this problem, the proportion of the 20-49-year-old female population on 1 July 1990, derived from the fourth census of Shanghai, was used. The census data on the number of married women was used to calculate the denominator; the induced abortion rate for all married women with de jure household registra- tion in 1991 was used to obtain the total number of induced abortions in 1991. This total was 70.32 times larger than the sample size. Next, we took the number of women in each age group in our sample, all of whom had an abortion, and expanded each group 70.32 times. This approach gave us the total abortions by age. Then that number was divided by the number of married women for each age group and the results were estimated age-specific abortion rates for married women in Shanghai in 1991. The resulting age-specific rates are shown in Tobie 3.1. They reveal that while abortions are proportionally higher in the 25-34-year age group, the actual abortion rate is much higher for the very young women. The total abortion rate, that is, the average number of abortions a woman would have if she experienced the abortion pattern of these women over her lifetime, would be 2.3 abortions. Age Group Proportion Rates Table 3.1: Married Women by Age, Shanghai, 1991 • 20-24 25-29 30-34 35-39 40-44 45-49 17.5 32.1 29.4 16.1 4.8 0.1 195.7 114.7 75.3 47.9 21.1 1.0 • Women who had an abortion (N = 2,765). • • Total abortion rate. Abortion and Education Total 100.0 ••2.3 The sample of women was highly educated. The majority, 80.8 per cent, had completed junior and senior middle school; only 10 per cent had finished Factors Affecting Induced Abortion Behaviour In Shanghai t 81 primary school or less. Another 10 per cent had completed training school, college or a higher level. Utilizing a similar procedure as for age-specific rates, we used data from the fourth census of Shanghai to estimate abortion rates for these educational categories (Table 3.2). It is clear from these data that the higher the educational level, the more likely a woman is to end an unplanned pregnancy by abortion. Table 3.2: Married Women by Education, Shanghai, 1991 * Education Level Illiterate Near Primary Senior Junior Training College Total Illiterate H. School H. School School or More 3.0 6.3 Proportion(%) 0.8 Rates 44.2•• 25.2 • Women who had an abortion (N = 2, 765). • • Includes illiterate women. Abortion and Occupation 40.0 62.6 40.8 95.5 6.2 2.9 100.0 134.4 122.2 The women who had the highest proportion of induced abortion were pro- duction and transportation workers; the second highest were technicians; and the third highest were service workers. We also estimated induced abortion rates by occupation, which are shown in Table 3.3. What this reveals is that there is a high correlation with our findings by educational level: women in higher level occupations, such as technical, business, office and service jobs, have the highest rates of abortion. Table 3.3: Married Women by Occupation, Shanghai, 1991 * Occupa- Techni- Mana- Office Business- Atten- Agricul- Produc- Other Not Total tion cian ger Worker man dant ture tion Worker Working Worker Proportion (%) 15.6 1.0 8.7 8.6 9.7 3.0 45.5 0.1 7.8 100.0 Rates 79.4 53.6 149.2 92.5 64.1 19.8 622 292.6 345.6 • Women who had an abortion (N = 2,765). Abortion and Residence The proportion of women with established urban residence (registered) among our sample of 2,765 married women was 72.3 per cent. Those with rural registration amounted to 27.7 per cent. The corresponding abortion rates for these groups are 77.8 per thousand and 51.2 per thousand, respec- tively. In short, in Shanghai, married urban women have higher abortion rates than married rural women. 82 t Gui Shl-xun Abortion and Health Induced abortion, especially repeated induced abortions, may impact nega- tively on health and have long-term adverse effects. Of the 1,446 married women who had experienced one previous induced abortion, 11.8 per cent believed the operation had been harmful to their health, and among this group one-third claimed the harm was serious. Of 350 women who had expe- rienced two previous induced abortions, 16.1 per cent responded that their operation had been harmful; and among them, a half claimed the harm was serious. Similar responses were given by women with three or more previous abortions, yet, despite this concern they returned for still another abortion as a result of their current unwanted pregnancy. Reasons for Abortion Reasons for Abortion In the Past A large majority of induced abortions are the result of women's wishes to ter- minate unwanted pregnancies resulting either from contraceptive failure or from non-use of contraception, which accounted for 96.3 per cent of all abor- tions in our study sample. The rest were for health or other personal reasons. Among the 2, 765 women in our sample, nearly half were requesting an abortion for the first time, 47.7 per cent. The other half of the sample, 1,446 women, had experienced one or more abortions in the past. Among these women, the distribution of these earlier abortions was: 69.4 per cent had experienced one previous abortion; 24.2 per cent had two; 4.9 per cent had three; 1.0 per cent had four; and 0.5 per cent had five or more. One woman had experienced seven induced abortions in the past. Among the 1,004 women who had one previous abortion, the main reason for the unwanted pregnancy was non-use of contraceptive methods (54.8 per cent). As the number of previous abortions increases, however, contraceptive failure becomes the dominant reason for seeking abortion. Women with more abortion experience tend to be older and less likely to continue risking unwanted pregnancy; however, given the poor quality of methods available, their use of contraception does not always prevent an unwanted pregnancy (Table 3.4). Reasons for the Current Abortion Of the 2,765 pregnant women requesting abortion at the time of the study, 24.7 per cent had not used contraception at the time of conception. A very large proportion, 75.3 per cent, reported that contraceptive failure had led them to seek an abortion. This led us to explore the reasons for not using Factors Affecting Induced Abortion Behaviour In Shanghai t 83 Table 3.4: Reasons for Pregnancy Termination in the Past (%)• Causes Induced Abortions 1 2 3 4 5 6 7 (N) (1,004) (350) (71) (14) (5) (1) (1) Not using 54.8 28.1 14.1 10.0 16.6 0 0 Failure 45.1 71.7 84.8 85.0 66.7 100.0 100.0 No answer 0.1 0.2 1.1 5.0 16.7 0 0 Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 • Women who had one or more previous abortions (N = 1,446). contraception in a city such as Shanghai, where services are widely available. We also looked at possible explanations for such a high rate of contraceptive failure. Reasons for Not Using Contraception To understand why one-fourth of the couples were not using contraception at the time they became pregnant, we decided to see what factors may have had an effect on their behaviour. We included in our analysis the background characteristics of the women and their husbands, such as their economic and social position, sexual behaviour, and general living and working environment. Overall, we selected 29 variables, which were grouped into five main fac- tors for analysis. The first factor was personal background and included the age of the women and their husbands. The second factor measured economic circumstances and included the occupation of each woman and her husband, within the six months before the current pregnancy, as well as the average net household income during the previous year. The third factor reflected social conditions and included the women and men's educational level, de jure resi- dence registration, leisure time during the week of conception, duration of marriage, fertility history, induced abortion history, and knowledge of contra- ception. The fourth factor reflected the women and men's contraceptive and sexual behaviour, and included: whether the couple discussed the choice of contraceptive methods; mutual sexual satisfaction; the frequency of sexual intercourse per week during the six months before the current pregnancy; and who initiated sexual intercourse during the period of conception. The fifth factor defined the living/working environment and included: their area of residence within the six months prior to the current pregnancy; the type of work unit; attitude of neighbours towards abortion; attitudes of colleagues towards abortion; and whether the family planning programme staff in the woman's district and work unit knew about their contraceptive status before the current pregnancy. The results showed that the proportion of women who became pregnant due to non-use of contraception varied significantly by each of the five main 84 t Gui Shi-xun factors described above. For example, except for women aged 45-49 years and men aged 50 years or more ( a relatively small group), the younger the subjects, the higher the proportion who became pregnant due to non-use of contraception (Table 3.5). Table 3.5: Main Reasons for Pregnancy Termination, by Age(%)* Age Group Contraception 20-24 25-29 30-34 35-39 40-44 45-49 Women Non-use 49.0 25.7 18.3 11.9 11.4 25.0 Failure 51.0 74.3 81.7 88.1 88.6 75.0 Total 100.0 100.0 100.0 100.0 100.0 100.0 Men Non-use 67.6 31.1 19.3 16.2 11.1 16.7 Failure 32.4 68.9 80.7 83.8 88.9 83.3 Total 100.0 100.0 100.0 100.0 100.0 100.0 * Women who had an abortion (N = 2,765) and the men concerned (N = 2,760). 50+ 25.0 75.0 100.0 It is also quite interesting that in almost all age groups, men tend to report higher proportions of non-use than the women. This difference was particu- larly noticeable among the younger age groups. Does this mean that younger men are less aware of their wives' contraceptive choices and use? We do not have the answer to this question, but it seems from our data that as couples get older, they tend to discuss their contraceptive decisions and choices more openly with each other. Women with only primary education were the most likely to fall into the non-use category. In fact, the proportion of women who became pregnant due to non-use of contraception was highest in the middle or lower levels of the educational scale. Moreover, except for the illiterate group, the lower the men's educational level, the higher the percentage of women not using con- traception. Contraceptive failure, on the other hand, seemed to be more common among the illiterate or the more highly educated women (Table 3.6). Table 3.6: Main Reasons for Pregnancy Termination, by Education(%)* Education flliterale Near Primary Junior Senior Under College No Illiterate College or More Answer Women Non-use 21.7 28.0 35.3 27.5 22.9 10.5 17.3 0 Failure 78.3 72.0 64.7 72.5 77.1 89.5 82.7 0 Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 0 Men Non-use 0 38.2 29.7 28.4 24.4 13.9 13.8 0 Failure 100.0 61.8 70.3 71.6 75.6 86.1 86.2 100.0 Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 • Women who had an abortion (N = 2,765) and the men concerned (N = 2,750). Factors Affecting Induced Abortion Behaviour In Shanghai t 85 The educational level of the men seems to have an important influence on contraceptive behaviour; among the more educated men the proportions of 'non-use' as a reason for the pregnancy termination of their wives are very low. The proportion of women who became pregnant due to non-use of contra- ception was higher when they and the men had rural de jure household regis- tration; 33.3 per cent and 32.4 per cent, respectively. The proportions were lower among women and men with urban de jure household registration; 21.4 per cent and 22.1 per cent, respectively. Women and men who were not working during the six months prior to the pregnancy were the highest proportions not using contraception: 47.9 per cent and 37.5 per cent, respectively. Among specific occupational groups, women engaged in agriculture, forestry, animal husbandry and fishery were the highest proportions not using, at 27.4 per cent; for the men, the percent- age was 27.9 per cent. These proportions were above average, perhaps indica- tive of their rural, more traditional habits and values. Second among non- users was the group of women and men in productive activities and transpor- tation; at 25.7 per cent and 27.1 per cent, respectively. It is clear that the rela- tionship between non-use and occupation is a close one and that non-use as a reason for pregnancy termination is lower among men and women with higher paying jobs, such as those in technical or office positions. At the same time, men and women in these kinds of jobs use more efficient methods but also experience higher rates of contraceptive failure. They seek abortions because they fear that their job situation could be negatively affected by a pregnancy that is not permitted by the one-child policy (Tobie 3.7). The patterns of non-use by type of work unit are shown in Table 3.8. As already noted, the proportions not using contraception were particularly high among non-working women and men and also for men and women working in individual or private working units, where they do not have to report to family planning field supervisors often. Such units are small-scale and have emerged as a result of a "more open market oriented policy in China. Women and men working in state enterprises or in collective units had lower non-use levels, probably because of better family planning surveillance by local cad- res. This was confirmed when we explored the degree of family planning inputs in relation to the use of contraception. We found that the proportion of women not using contraception was higher where supervision by the family planning programme staff was lower. Of the 683 women not using contracep- tive methods, 71. 7 per cent said the family planning programme monitors did not know they were 'not using'. Another 12.4 per cent who were not using methods said the family planning workers knew it, but were indifferent. The remaining 15.9 per cent of non-users explained that the family planning workers were opposed to their not using contraception, but did not offer them any guidance. Interestingly, the more leisure time the women and men had during the month before the pregnancy, the higher the percentage of non-use of Table 3.7: Main Reasons for Pregnancy Termination, by Occupation (%)• Occupation Technician Manager Office Businessman Attendant Agricultun Production Other No Job No Answer Worker Worker Worker Women Non-use 17.6 25.0 17.0 23.6 20.1 27.4 25.7 0 47.9 0 Failure 82.4 75.0 83.0 76.4 79.9 72.6 74.3 100.0 52.1 100.0 Total 100.0 100.0· 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Men Non-use 16.3 19.4 17.8 30.0 26.2 27.9 27.1 100.0 37.5 30.0 Failure 83.7 80.6 82.2 70.0 73.8 72.1 72.9 0 62.5 70.0 Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 • Women who had an abortion (N • 2,764) and the men concerned (N = 2,730). (Occupations for one woman and 20 men were 'not clear'.) 'lllble 3.8: Curnnt Pregnancy Causes, by 'lype of Working Unit(%)• 7j,pe of Working Units State Collective Individual Private Joint ~ntwTs Other No EmpWJment Not Clear Women Non-use 18.5 27.3 40.0 40.0 26.8 20.8 47.9 0 Failure 81.5 72.7 60.0 60.0 73.2 79.2 52.1 100.0 Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Men Non-use 20.6 31.4 35.8 41.7 21.7 29.5 37.5 16.7 Failure 79.4 68.6 64.2 58.3 78.3 70.5 62.5 83.3 Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 • Women who had an aborti()fl (N = 2,762) and the men concerned (N = 2,744). (fypes of working unit for three women and six men were 'not clear'.) 88 t Gui Shi-xun contraception. This percentage was higher for the women and men who had an average leisure time of 36-42 hours every week, 42.0 per cent and 45.8 per cent, respectively (Table 3.9). This indicates, as in previous tables, that men and women who are not working tend to have a greater tendency not to use contraception. It seems also that people with 36 or more hours of 'leisure time' may actually be 'unemployed'. Tobie 3.9: Main Reasons for Pregnancy Termination, by Couples• Leisure Time (%)• Leisure Time 0 1-7 8-14 15-21 22-28 29-35 36-42 43+ Not (Hours per Week) Clear Women Non-use 18.0 18.5 20.7 26.0 34.0 33.3 42.0 41.1 75.0 Failure 82.0 81.5 79.3 74.0 66.0 66.7 58.0 58.9 25.0 Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Men Non-use 26.4 20.9 21.1 25.0 28.2 29.7 45.8 40.9 16.7 Failure 73.6 79.1 78.9 75.0 71.8 70.3 54.2 59.1 83.3 Total 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 • Women who had an abortion (N = 2,761) and the men concerned (N = 2,732). (The leisure time of four subjects and 18 husbands was 'not clear'.) As expected, the proportion not using contraception was higher (50.6 per cent) among childless women than among women who had children (22.2 per cent). The less knowledge of contraceptive methods women had, the higher the rate of non-use. The result showed that among 2,082 women who experi- eneed contraceptive failure, the average number of methods known was 4.8. Among 683 women not using contraception, the same average was much lower: 3.6 methods. Similarly, women and men who talked more frequently about their choice of contraceptive methods were the least likely to report non-use as the reason for an abortion. The proportions not using any contraceptive method were: 15.9 per cent for couples who 'often' talked about the use of methods; 22 per cent for couples who 'sometimes' talked about this; 24.8 per cent for couples who 'seldom' talked; and 29.9 per cent for couples who 'never' discussed contraception. The frequency of sexual intercourse was directly correlated with non-use of contraception. Couples who had sex more frequently had a higher proportion of women not using contraceptives. The actual proportions not using were: 22.9 per cent for couples who had sexual relations 'once a week'; 24.2 per cent, for 'twice a week'; 26.6 per cent, for '3 times a week'; 28.0 per cent, for '4 times a week'; 63.6 per cent, for '5 times a week'; and 69.2 per cent for '6 times a week'. This finding is interesting in that it runs contrary to expectations; one would assume that couples who are more sexually active would be most wor- ried about pregnancy protection. Factors Affecting Induced Abortion Behaviour In Shanghai t 89 Women and men who declared to have lower levels of sexual satisfaction were more likely to be non-users of contraception. Those who discussed their sexual life 'often' and were mutually satisfied, had a lower level of non-use, 20.4 per cent ( measured during the six months prior to pregnancy). Women and men who responded 'seldom' and 'sometimes' to their frequency of dis- cussion about sexual life had somewhat higher levels of non-use, 22.5 per cent and 23.7 per cent, respectively. Among women and men who 'never' dis- cussed issues relating to their sex lives, non-use was highest (30.4 per cent). A Closer Look at Contraceptive Non-Use We wanted to explore whether not using contraception at the time the preg- nancy occurred was part of an established behavioural pattern or a one-time occurrence. For this line of inquiry, we decided to look into the main reasons for not using contraceptive methods in the two years prior to the current pregnancy. We presented 15 possible reasons to 793 women ( a sub-set of our main sample) who had a previous abortion and who had not used contraception to avoid that earlier pregnancy, provided the pregnancy occurred within the two years preceding the interview. We asked these women to rank each reason according to its applicability. The reasons included: (a) desire to have a child according to the family planning policy; (b) desire to have a birth even though it is not permitted by the family planning policy; (c) lactation makes preg- nancy impossible; ( d) not aware of contraceptive methods; ( e) no 1confidence in contraceptive methods; (/) husband not cooperative; (g) contraception may impair health; (h) contraception may affect sexual ability; (i) contracep- tive use interferes with sexual enjoyment; (j) contraception is troublesome; (k) induced abortion is not harmful to health; (l) after induced abortion women can be taken care ofby the government or work units; (m) no contra- ception available; (n) unmarried; (o) other reasons. The results showed that for these women with previous abortion experi- ence and who, for the last two years, had not used contraception, the three major reasons for non-use were: 'unmarried' (32.5 per cent); 'protected by lactation' (24.8 per cent); 'did not know any methods' (17.3 per cent). For 124 women who had two previous induced abortions, the two main reasons were the same: 'unmarried' (23.4 per cent); 'protected by lactation' (19.4 per cent). The other main reasons for this second group were 'desire to have a birth according to the family planning policy' and 'method seen as troublesome' (16.9 per cent). For 13 women who had three previous abortions, the first major reason was the same as for the other two groups: 'unmarried' (23.1 per cent). The other primary reasons were 'desire to have a birth according to the family planning policy' and 'consider that contraception may impair health' (15.4 per cent, each). Of
Looking for other reproductive health publications?
The Supplies Information Database (SID) is an online reference library with more than 2000 records on the status of reproductive health supplies. The library includes studies, assessments and other publications dating back to 1986, many of which are no longer available even in their country of origin. Explore the database here.