UNIVERSITY OF GHANA REGIONAL INSTITUTE FOR POPULATION STUDIES (RIPS) RURAL-URBAN DIFFERENTIALS IN INDUCED ABORTION AMONG WOMEN IN GHANA BY GABRIEL ANANYA (10475358) THIS DISSERTATION IS SUBMITTED TO THE UNIVERSITY OF GHANA, LEGON IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE AWARD OF MA IN POPULATION STUDIES DEGREE DECEMBER, 2021 University of Ghana http://ugspace.ug.edu.gh II ACCEPTANCE Accepted by the College of Humanities, University of Ghana, Legon in partial fulfilment of the requirement for the award of Master of Arts Degree in Population Studies. SUPERVISOR DR. FIDELIA A. A. DAKE DATE: 17th December, 2021 University of Ghana http://ugspace.ug.edu.gh III DECLARATION I, Gabriel Ananya, hereby declare that except for the references made to other people’s work which have been duly acknowledged, this is the result of my own research undertaken under supervision at the Regional Institute for Population Studies, University of Ghana and that neither a part nor the whole of it has been presented elsewhere for the award of another degree. STUDENT GABRIEL ANANYA DATE: 17th December, 2021 University of Ghana http://ugspace.ug.edu.gh IV DEDICATION I dedicate this work to my parents, Mr. Christopher Ali Ayilimba Ananya (late) and Mrs. Diana Apele Ananya, as well as my wonderful nieces and nephews. Aim to perform a lot better. University of Ghana http://ugspace.ug.edu.gh V ACKNOWLEDGEMENTS I am very grateful to the Lord almighty for His grace and strength throughout this journey. This achievement would not have been possible without Him. I want to specially thank my indefatigable supervisor in the person of Dr. Fidelia A.A. Dake. I am grateful for the guidance and support you provided throughout this period. Your patience is adorable and I am privileged to have had you as a supervisor. You pushed me to limits I never thought I could ever go and I am tremendously grateful for that. God richly bless you. I would also like to express my gratitude to the entire staff of the Regional Institute for Population Studies for welcoming me with open arms whenever I needed assistance. I want to especially mention Dr. Adriana Andrea Biney and Dr. D. Y Atiglo, your support is greatly appreciated. I appreciate the support from the entire staff of the institute and some senior colleagues in the PhD class especially Charles Asabre, Afra Adomako Kwabiah, and Ruth T. Sawyer, your encouragement is appreciated. I need to say a big thank you to my family and friends who have supported me from the beginning of this journey till now. Owusu, Alex, Chris, Sandra, Debby and Faustina I am grateful for your prayers and encouragement throughout this process. Finally, I want to say thank you to my colleagues, the 2020/2021 MA Cohort at the Regional Institute for Population Studies (RIPS) especially Belinda Yayra Ofori and Patience Manubea Siaw, you guys have been of tremendous help. God bless you all!! University of Ghana http://ugspace.ug.edu.gh VI TABLE OF CONTENT ACCEPTANCE......................................................................................................................... II DECLARATION....................................................................................................................... III DEDICATION........................................................................................................................... IV ACKNOWLEDGEMENTS.........................................................................................................V TABLE OF CONTENT .............................................................................................................VI LIST OF TABLES .....................................................................................................................X LIST OF FIGURES.................................................................................................................... XI ABSTRACT...............................................................................................................................XII CHAPTER ONE …………………………….………...…………………………………........1 INTRODUCTION 1.1 Background …………….......................................................................................................1 1.2 Problem Statement ............................................................................................................... 4 1.3 Research Questions ...............................................................................................................7 1.4 Research Objectives ..............................................................................................................8 1.5 Rationale of the study.............................................................................................................8 1.6 Organization of study chapters……………………………………………………………...10 CHAPTER TWO ……………………………………………………………………..….........11 LITERATURE REVIEW 2.1Introduction ………………………………………………………………………………….11 2.2 Definition of concepts ……………………………………………………………………....11 2.3 The abortion law in Ghana ………………………………………………………………….14 2.4 Regional dynamics in induced abortion …………………………………………………….15 2.5 Place of residence and induced abortion ……………………………………………………16 2.6 Socio-economic and demographic factors influencing induced abortion … ……………….19 2.6.1 Age of woman …………………………………………………………………….19 2.6.2 Current marital status ……………………………………………………………..20 2.6.3 Level of educational attainment …………………………………………………..21 2.6.4 Current contraceptive use …………………………………………………………22 2.6.5 Ethnicity ………………………………………………………………………......24 University of Ghana http://ugspace.ug.edu.gh VII 2.6.6 Number of children ever born ……………………………………………………..25 2.6.7 Religion ……………………………………………………….…………………...26 2.6.8 Wealth Quintile ………………………………………………………….………...27 2.6.9 Exposure to media ……………………………………………………..………….28 2.6.10 Knowledge of abortion law ……………………………………………………...28 2.7 Theoretical Perspectives …………………………………………………………….29 2.8 Conceptual framework ………………………………………………………...........34 2.9 Hypotheses ………………………………………………………………………….36 CHAPTER THREE………………………………………………………...…………………..37 METHODOLOGY 3.1 Introduction ………………………………………...……………………….……………….37 3.2 Data source……………………………………………………………………….…………..37 3.3 Sample design and study sample … ……..……………………………………………….…38 3.4 Variables in the study …………………………………………………………………….….38 3.4.1 Dependent variable………………………………………………………………….……..38 3.4.2 Independent Variable……………………………………………………………………....39 3.4.3 Control variables…………………………………………….……………………….…….39 3.4.3.1 Age of woman…………………………………………………………………………....39 3.4.3.2 Current marital status…………………………………………………………….………39 3.4.3.3 Level of educational attainment……………………………………………….…………39 3.4.3.4 Current contraceptive use……...……………………………………….………………..40 3.4.3.5 Number of children ever born……………………………………………………………40 3.4.3.6 Ethnicity……………………………………...………………….……………………….40 3.4.3.7 Religion…………………………………………………………………………………..40 3.4.3.8 Wealth Quintile…………………………………………………………………………..41 3.4.3.9 Exposure to media………………………………………………………………………..41 3.4.10 Knowledge of abortion law…………………………………………………………...…..41 3.5 Variables and their categorization...……………………………………….………..……….41 3.6 Methods of analysis…...…..………………………………………………..………………..42 University of Ghana http://ugspace.ug.edu.gh VIII CHAPTER FOUR ………………………………………………………………….….……...46 ANALYSIS OF RURAL URBAN DIFFERENTIALS IN INDUCED ABORTION IN GHANA 4.0 Introduction ……………………………………………………………………………..…46 4.1 Induced abortion in Ghana………………………………………………..………………..46 4.2 Demographic characteristics of study sample …………………………………………..…47 4.3 Socio-economic characteristics of study sample ……………..……………………………49 4.4 Contraceptive use, knowledge of abortion law and media exposure among study sample…………………………………………………..….………..…..…...50 4.5 Bivariate relationships between demographic factors and socio-economic factors, contraceptive use, knowledge of abortion law, media exposure and induced abortion ………………………………………………………….…...........…51 4.5.1 Distribution of study sample by demographic characteristics and induced abortion by rural-urban disaggregation …… …………….....……………….………………....52 4.5.2 Socio-economic characteristics of study sample by rural- urban disaggregation and induced abortion ……………....……………………………….….55 4.5.3 Contraceptive use, knowledge of abortion law and media exposure among the study sample by rural-urban disaggregation and induced abortion ………..…………..57 4.6 Determinants of induced abortion among women in Ghana ……………………...….…….60 4.6.1 Determinants of induced abortion by place of residence women in Ghana ……………61 CHAPTER FIVE………………………………………………………………………………68 DISCUSSION 5.1 Introduction…………………………………………………………………………………68 5.2 Socio-economic factors and induced abortion ……………………………………………..68 5.3 Demographic factors and induced abortion ………………………………………………..70 5.4 Contraceptive use, knowledge of abortion law, exposure to media and induced abortion……………………………………………………………………….72 CHAPTER SIX………………………………………………………….…………………....74 SUMMARY, RECOMMENDATIONS, CONCLUSION AND DATA LIMITATIONS 6.1 Summary………………………………………………….……………………..….….…..74 6.2 Conclusion ………………………………………………………….………………….….76 6.3 Recommendations …….…………………………..…………………...……….…..……..76 University of Ghana http://ugspace.ug.edu.gh IX 6.4 Data Limitations …………………………………………………………………………..77 REFERENCES …………………………………………..……………………….….…..…..79 University of Ghana http://ugspace.ug.edu.gh X LIST OF TABLES Table 3.1: Variables and their categorization ……………………………………….………42 Table 4.1: Percentage distribution of induced abortion among the study sample by rural-urban disaggregation ………………………………………………………………………….…....47 Table 4.2: Percentage distribution of demographic characteristics of study sample by rural-urban disaggregation………………………………………………………………………………...47 Table 4.3: Percentage distribution of socio-economic characteristics of study sample by rural- urban disaggregation ………………………………………………………………………....50 Table 4.4: Percentage distribution of contraceptive use, knowledge of abortion law and exposure to media among the study sample by rural-urban disaggregation………………....………....51 Table 4.5: Percentage distribution of study sample by demographic characteristics and induced abortion status…………………………………………...……….……………………….…..53 Table 4.6: Percentage distribution of study sample by socio-economic characteristics and induced abortion status………………………………………………………...……………..56 Table 4.7: Percentage distribution of study sample by contraceptive use, knowledge of abortion law, exposure to media and abortion status…………………..........................................….….59 Table 4.8: Binary logistic regression showing the predictors of induced abortion among women in Ghana disaggregated by rural and urban residence ……………………………………..…63 University of Ghana http://ugspace.ug.edu.gh XI LIST OF FIGURES Figure 2.1: Social determinants of maternal health………………………………….……..….32 Figure 2.2: Conceptual framework showing the factors that influence induced abortion ……36 Figure 3.1 Analytical framework for Model 1 …...…………………………………………..44 Figure 3.2 Analytical framework for Model 2………………………………………………..44 Figure 3.3 Analytical framework for Model 3………………………………………………..45 Figure 3.4 Analytical framework for Model 4………………………………………………..45 University of Ghana http://ugspace.ug.edu.gh XII ABSTRACT Introduction: Globally, there are disparities in maternal health outcomes as well as the factors that influence maternal health outcomes. Some of these disparities stem from the differences in place of residence (rural or urban). These rural-urban disparities affect access to services such as safe induced abortion. In order to improve maternal health outcomes in developing countries such as Ghana, these disparities need to be addressed. Against the foregoing, this study seeks to examine the differentials in induced abortion among women in urban and rural areas in Ghana. Methods: This study used data from the 2017 Ghana Maternal Health Survey. The analytical sample includes a total of 13,176 women (weighted) who have been pregnant in the five years preceding the survey. The data was analysed using descriptive statistics, Pearson chi-square tests and binary logistic regression analysis. Results: The results show that 18.5% of urban women had induced abortion while 10.0% of rural women had induced abortion. Among rural women 33.9 % were using contraceptives while 32.6% of urban women were using contraceptives. Among urban women who were not married 35.2% induced abortion while 20.9% of rural women who were not married induced abortion. At the multivariate level, while age, level of educational attainment and ethnicity were significant predictors of induced abortion in urban areas these variables were not significant in rural areas. Also, whereas exposure to media was a significant predictor of induced abortion in rural areas it was not a significant predictor in urban areas. Conclusion: The factors that influence induced abortion in urban and rural areas differ, therefore policies promulgated to address induced abortion must be distinctive and peculiar to each area (urban/rural) in order to reap maximum results. University of Ghana http://ugspace.ug.edu.gh 1 CHAPTER ONE INTRODUCTION 1.1 BACKGROUND Abortion, according to the World Health Organization (2008) is the termination of pregnancy prior to 20 weeks' gestation or a foetus born weighing less than 500g. Also, abortion is said to be the expulsion of a foetus before it becomes viable (Hern, 1995). Abortion disrupts the full gestation of pregnancy and child birth, which when not properly done can cause death or other forms of harm to the mother. This makes abortion an integral part of maternal health, therefore improving maternal health in any country has so much to do with improving the abortion situation in that country. Abortions can be either spontaneous (miscarriage) or induced. Spontaneous abortion is the loss of pregnancy naturally before twenty weeks of gestation (Apgar & Churgay, 1993). Induced abortion on the other hand refers to the deliberate termination of a product of conception (Rigterink et al., 2013). Induced abortion can either be safe or unsafe. The World Health Organization (2008), defines unsafe abortion as a procedure for terminating an unintended pregnancy, carried out either by persons lacking the necessary skills or in an environment that does not conform to minimal medical standards, or both. Safe abortions on the other hand, are those performed in accordance with WHO guidelines and standards, thus ensuring that the risk of severe complications is minimal (WHO, 2012). Induced abortion can be done either by using medical means or non-medical means. Medical means usually involve the use of medications such as mifepristone to induce abortion whereas non- medical means can include Dilation and Curettage or Dilation and Evacuation (D&C/D&E), Manual Vacuum Aspirator (MVA) and Electronic Vacuum Aspirator (EVA). Non-medical abortions can also include some traditional methods that falls within the unsafe categorisation. Examples include drinking alcohol, drinking herbal concoctions, drinking home remedies, using herbal enema, inserting University of Ghana http://ugspace.ug.edu.gh 2 substances into the vagina, heavy massage, excessive physical activity and taking unknown tablets (Ghana Statistical Services, 2018). The incidence of induced abortion differ from region to region and from country to country. Factors such as culture and availability of health systems have the tendency to influence the incidence of induce abortion in a country. In recent years, global induced abortion rates have increased. According to the WHO (2021), 73 million abortions were induced averagely from 2014 to 2019. This is an increase from 56 million cases of abortions from 2010 to 2014 (Population Reference Bureau, 2021). Regionally, induced abortion is very prominent in Asia and Africa. Developing countries contribute about 97% of the global figures of induced abortion with Asia and Africa leading the chart (WHO, 2021). From 2010 to 2014 Africa’s induced abortion cases were 8.2 million of which 76% were unsafe. In the same period Asia, Latin America, North America and Europe recorded 39.4 million, 6.5 million, 1.2 million and 4.3 million cases respectively. Out of these cases 42%, 76%, 0.9% and 11% were unsafe induced abortions respectively. In Ghana, despite abortion being legal in situations of rape, incest, foetal abnormality/ disease or defilement of a female with a mental health challenge or when the woman’s life or mental health is at risk (Morhee & Morhee, 2010), abortion is still highly unsafe. The national abortion rate of Ghana is 44 abortions per 1000 women aged 15-49. This is equivalent to over 327,600 abortions annually (Keogh et al., 2020; Polis et al., 2020). Unsafe abortion in Ghana exposes women to maternal deaths which is currently reported to be about 310 deaths per 100,000 live births (Ghana Statistical Services, 2018) which is far from the sustainable development goal of reaching less than 70 deaths per 100,000 live births. Unsafe induced abortion has become a problem in Ghana for several reasons. According to Atakro et al. (2019), lack of knowledge of safe abortion services causes women to engage in unsafe University of Ghana http://ugspace.ug.edu.gh 3 abortion practices. In their study, Atakro et al. (2019) found that all participants had complications from unsafe abortion practices. Also, women perceived safe abortion as a religious and cultural taboo hence would rather go in for a clandestine unsafe abortion service (Atakro et al., 2019). In some instances the lack of equipment and medication in health centres have forced women to look for alternatives which are unsafe (Wodajo et al., 2017). The economic reasons why women engage in unsafe abortion practices are also sterling. The charges for safe abortion services is a major cause of unsafe abortions since some women cannot afford the safe abortion service charge (Gbagbo, 2020). All these are challenges women face in gaining access to safe abortion services and until solutions are proffered to counteract them, the canker will continue to lurk and expose women to deaths and harsh health complications. Globally, there is evidence of disparity among urban and rural women when it comes to health services, socio-economic factors and access to education. These factors have an influence on the abortion choices these women make. According to Hebert et al. (2016), access to health services is differentiated by rural-urban divide. Also, according to Guttmacher Institute (2018), in Asia, poor rural women are likely to have unsafe abortion service. The same likelihood exist in Latin America and the Caribbean (Guttmacher Institute., 2018), this is due to the unavailability of health centres and also lack of economic might to access safe abortion services. In Turkey, although the percentages of pregnancies that ended in induced abortion or unsafe abortion has reduced, the levels is still high. This is because mid-level health facilities in rural areas lack specialists to provide safe abortion services for the women (Idge et al., 2008). In rural Canada, abortion services has also reduced and nurses in these rural areas tend to resign due to the fear of being harassed (Ames & Norman, 2012). In Ghana, the problem of disparity is no different. Women in urban areas have access to better health facilities and are more economically empowered to have abortion services compared to their University of Ghana http://ugspace.ug.edu.gh 4 counterparts in the rural areas (Nyarko & Potter, 2020). Biney & Atiglo (2017), also found that, lack of finances was associated with use of harmful abortion methods among women in rural settings, while no such difference was found among urban dwellers. It is clear from the literature that different factors differentiate rural and urban women when it comes to induced abortion. Therefore, in order to reduce the gap between urban and rural women, these factors need to be thoroughly investigated, that is why this study seeks to investigate the differentials that exist between rural and urban women in Ghana when it comes to induced abortion. 1.2 PROBLEM STATEMENT Unsafe induced abortion continues to be a major reproductive health challenge in many countries, especially developing countries (Wodajo et al., 2017). This is because there is a higher risk of maternal mortality associated with unsafe induced abortions among women. The challenge of unsafe abortions made it difficult for most developing nations to achieve the Millennium Development Goals (MDG) on maternal health (MDG 5) and with the Sustainable Development Goals (SGD) just 9 years away (2030) it has become clear that it would be difficult to achieve the target on maternal health if the problem of unsafe abortions are not addressed (Wambura, 2015). Aside the high mortality rates caused by unsafe induced abortions in developing nations (Ganyaglo & Hill, 2012; Mills et al., 2008), several complications have been identified to be associated with unsafe induced abortion. Studies have found excessive bleeding, abdominal pains and damage to the uterus as some of the many complications that women who undergo unsafe induced abortions endure (Ghana Statistical Service, 2018; Guttmacher Institute, 2010). Despite the liberal legal abortion regime in Ghana, unsafe induced abortion continues to be high in the country (Oliveras et al., 2015). This paradox is best explained by findings of studies that have sought to measure stakeholders’ understanding of the law on abortion in Ghana. In such studies, University of Ghana http://ugspace.ug.edu.gh 5 it was found that physicians did not understand the abortion law in Ghana and have in many instances misinterpreted it (Kumi-Kyereme et al., 2014; Payne et al., 2013). Another challenge faced in the fight against unsafe induced abortion is conscientious objection. Conscientious objection is the right that health workers have to say no to the provision of certain services that may be in conflict with their beliefs. This challenge is corroborated by the findings of a study conducted by Awoonor-Williams et al. (2020), which sought to explore conscientious objection among clinicians and found that clinicians had little understanding of conscientious objection even though they practice it. Under such circumstances, if health care providers themselves become a problem in the delivery of health care then the path to improving health care is not just made longer but more complicated. For example, Schwandt et al. (2013) found in a study conducted in the Komfo Anokye Teaching Hospital and Korle-Bu Teaching hospitals in Ghana that nurses scold patients especially unmarried teenage girls who become pregnant and visit the hospital. This is a problem because this bad practice force these girls to seek services elsewhere and they end up going for unsafe induced abortions. In addition to these challenges, socio-economic and cultural factors also contribute to unsafe abortions among women in Ghana. Atakro et al. (2019), Biney and Atiglo (2017), and Schwandt et al. (2013), found that financial constraint and the availability of social support or the absence of it can determine whether a pregnancy would be aborted or not. The impact of financial constraint may differ in rural and urban areas which makes it important for it to be studied. In some country contexts, abortion is seen as illegal and dangerous as well as stigmatizing and a shame to families (Bleek, 1978; Hill et al., 2009). Therefore, in a liberal legal abortion regime such as Ghana’s if people still view abortion to be illegal then there is more to be done in this fight against unsafe induced abortion. The University of Ghana http://ugspace.ug.edu.gh 6 legal ambiguities existing in the conditions under which legal safe induced abortions can be offered also makes it difficult for some qualified providers to provide services (Voetagbe, et al, 2010). According to Voetagbe et al. (2010) where safe abortion services are accessible, unsafe abortion and abortion-related mortality and morbidity are reduced. However, in Ghana, access to safe abortion services in rural areas is limited because aside the lack of health facilities in these areas, there are also relatively few health professionals. This denies rural women access to safe abortion services and they are forced to look for alternatives that result in unsafe induced abortions and its accompanying complications (Voetagbe et al, 2010). Also, women in urban areas are able to afford to take care of abortion cost as compared to their counterparts in the rural areas. The issue of accessibility is not only limited to health facilities and professionals but also to education since many studies have found that women who are educated are more likely to access safe abortion services (Sundaram et al., 2012). However, here is the case that rural areas in Ghana lack the best of educational facilities to facilitate their education. Rural areas lack access to internet services hence, making it difficult to receive education through the internet (i.e., online podcasts, social media, websites and many alike). The above mentioned factors together, widen the disparities that exist between women in urban areas and rural areas in relation to access to safe abortion services. This disparity makes it difficult for Ghana to win the battle against unsafe induced abortion and until that is done the problem would still persist. Additionally, in Ghana, 27% of women who induced abortions used non-medical methods (Ghana Statistical Services, 2018). These non-medical methods included drinking alcohol, drinking herbal concoctions, drinking homemade remedies, using herbal enema, inserting substances into the vagina, heavy massage, excessive physical activity, taking tablets (exact kind unknown), to mention University of Ghana http://ugspace.ug.edu.gh 7 but a few (Ghana Statistical Services, 2018). These methods are largely unsafe and this puts women who use them at risk of maternal mortality. Moreover, the problem of stigmatization still lurks. Both service providers and service seekers face stigmatization. This act of stigmatization causes service providers to desist from providing the service to women and women also go for unsafe methods that can expose them to death (Payne et al., 2013). These challenges expose women to high risk of maternal deaths in these areas which in turn increase the maternal deaths in Ghana as a whole, hindering the achievement of the Sustainable Development goal on maternal mortality. Therefore, differentials by place of residence, that is urban and rural residence puts women who reside in rural areas in a disadvantaged position. This study therefore will seek to investigate this and have empirical evidence to ascertain it. 1.3 RESEARCH QUESTIONS Against the foregoing, this study seeks to answer the main research question, which is; What are the rural and urban differentials in induced abortions among women in Ghana? Further specific research questions are; 1. What are the levels of induced abortion among women in rural and urban Ghana? 2. What are the factors associated with induced abortion among women in rural and urban Ghana? 1.4 RESEARCH OBJECTIVES To help answer the research questions, the study aims to examine the rural and urban differentials of induced abortions among women in Ghana. This overarching aim will be achieved by achieving the following specific objectives; University of Ghana http://ugspace.ug.edu.gh 8 1. Estimate the levels of induced abortion among women in rural and urban Ghana. 2. Determine the factors associated with induced abortion among women in rural and urban Ghana. 1.5 RATIONALE OF THE STUDY The inquiry into induced abortion is not a new phenomenon in Ghana. Previous research has explored abortion in the Akan ethnic group and in the Ghanaian family (Bleek, 1978, 1990). The decline in the total fertility rate (TFR) in Ghana from 5.2 in the 1990’s to 3.9 in 2017 has been a great success for Ghana in its attempt to reduce TFR. However, one would expect that this decline would be commensurate with the contraceptive prevalence rate (CPR) in Ghana but that is not the case. Only 1 in 4 married women use modern contraceptives (25%) and CPR among women who are not married is 31% (Ghana Statistical Service, 2018). This disparity has caused some to believe that the findings of studies such as Ahiadeke (2001) and Oliveras et al. (2015) which suggest that the use of induced abortion as a birth control mechanism by some women might be what explains the inconsistency between reduction in TFR and the low contraceptive prevalence in Ghana. Previous studies have also explored contraceptive use and abortion in Ghana and reasons or factors that influence peoples decision to have abortions in Ghana (Ahiadeke, 2001; Biney, 2011; Boah et al., 2019). Few studies have also tried to explore the phenomenon of induced abortion in Ghana in the rural and urban contexts (Adjei et al., 2015; Atakro et al., 2019; Gbagbo et al., 2015; Hill et al., 2009; Kumi-Kyereme et al., 2014) but have done so in isolation. These isolated studies therefore do not give the complete picture of what is happening in some specific contexts. A study that assessed the national plan to prevent maternal mortality of selected countries from 7 regions of the world found that even though a national plan is important, where there are specific plans to tackle specific problems, the results have been better (Mirembe et al., 2010). However, to be able to develop University of Ghana http://ugspace.ug.edu.gh 9 specific plans for specific problems, the context in which the plans would be applied must be known. Therefore this study seeks to give a more comprehensive picture of induced abortion in both rural and urban context. In Ghana, though unsafe induced abortion continues to be a national problem, the rural and urban contexts and dynamics might differ in terms of methods, levels, cost, type of service providers, reasons for the abortions etc. To be able to address the challenge of maternal health (maternal mortality and morbidity), the rural and urban disparities that exist in induced abortion must be addressed. This is why this study seeks to explore rural-urban differentials in induced abortion. Also, few studies have attempted to explore differentials in induced abortion in Ghana using nationally representative data (Biney & Atiglo, 2017; Boah et al., 2019; Owoo et al., 2019). Some researchers (Sutton et al., 2019; Vibeke Rasch and Rose Kipingili, 2009) have studied rural-urban differences in the American and Tanzanian context. Sutton et al. (2019) found that women age 15 to 24 years in urban areas are more likely to induce their first pregnancies as compared to those in rural areas, therefore to be able to prevent these women from inducing their first pregnancies measures need to be put in place to prevent pregnancy at the early ages in the urban areas. Vibeke and Kipingli (2009) also found in Tanzania that 62% and 63% of rural and urban abortions were unsafe. However, more than half of the abortions in urban Tanzania were performed by an unskilled person. These findings are important guides for future studies and policy planning. This present study hopes to find the differentials in induced abortion among women in Ghana, in order to contribute to the induced abortion literature in Ghana and also help in the contextualisation of policies in rural and urban areas to attain maximum benefits. University of Ghana http://ugspace.ug.edu.gh 10 1.6 ORGANIZATION OF STUDY CHAPTERS In this introductory chapter, the background to the study, problem statement, research questions, rationale, objectives and organisation of the dissertation have been presented. The rest of the dissertation is organised as follows: Chapter Two presents the definition of key concepts and review of relevant literature on some indicators of induced abortion. This chapter further details the theoretical framework from which the conceptual framework for this study was adapted and presents the hypotheses of this study. Chapter Three presents the study design and details on the methodology involved in the study. The variables and how they are measured as well as how they were analysed were detailed. Chapter Four presents the results of the study which includes univariate analyses of the study population, the bivariate analyses and the multivariate analyses. Chapter Five includes the discussion of the results in chapter four side by side other findings in Ghana and beyond. Chapter Six concludes the study with a summary, conclusion, recommendations and data limitations. University of Ghana http://ugspace.ug.edu.gh 11 CHAPTER TWO LITERATURE REVIEW 2.1 INTRODUCTION This chapter sought to review the literature on induced abortion paying particular attention to the differentials that have been identified at the regional, country and rural- urban levels. This review relied on several sources for literature. One of the main sources was online sources. Some of the online sources included PubMed, Jstor, Research Gate, websites of organizations such as World Health Organization (WHO), Guttmacher Institute and other sources. The scope however encompasses, Ghana, Africa, Asia, America and parts of Europe. The review includes definition of concepts, the legal regime of abortion in Ghana, regional dynamics of abortion, place of residence and induced abortion, factors influencing abortion, theoretical perspectives of the study, the conceptual framing of the study and proposed hypotheses for testing. 2.2 DEFINITION OF CONCEPTS The subject of abortion sparks debates in many countries with the main arguments circling on the morality and legality of abortion. The moral debate takes inspiration from religious teachings and doctrines while the debate on legality takes inspiration from the constitution or body of laws in the respective countries. The role of tradition and personal inclination to such debates cannot be over looked in many countries especially in Africa and Asia. In sub-Sahara Africa, tradition and religion is a major part of the belief system that influences individual decision and choices. Abortion has been defined differently by people and institutions, but the core of it remains the same. Hern (1995), defines abortion to be the expulsion of a foetus before it becomes viable. WHO (2008), defines abortion as the termination of pregnancy prior to 20 weeks gestation or a foetus University of Ghana http://ugspace.ug.edu.gh 12 born weighing less than 500g. The expulsion of the foetus as captured by Hern (1995), can be either spontaneous or induced. The spontaneous expulsion of foetus is also known as a miscarriage. Apgar & Churgay (1993), define spontaneous abortion as the loss of pregnancy naturally before twenty weeks of gestation. Several factors ranging from the personal characteristics of the woman (for example, her physiological makeup/ fecundability), type of physical activity exposed to or engaged in, the income status of the women and others (Zheng et al., 2017) are associated with the risk of spontaneous abortion. Induced abortion on the other hand, is the type of expulsion caused by human efforts. United Nations (2014), sees induced abortions to be those abortions initiated by deliberate action with the intention of terminating pregnancy. Rigterink et al. (2013), puts it simple and brief by saying it is the intentional termination of pregnancy. Amies (1990), puts it slightly different and says, it is the termination of an implanted embryo by artificial measures. These artificial means include ingestion of certain plant compounds, physical trauma to the abdomen, and introduction of chemicals or sharp objects into the uterus. The definition by the United Nations (2014), states the deliberate intention dimension of induced abortion which suggest that there are other underling factors that influence ones intention to act. Amies (1990), also draws attention to the means through which this abortion is induced, thus, by artificial means. Induced abortion can further be classified into safe and unsafe or medical and non- medical/surgical. Safe abortion is that which is done based on the WHO guidelines and standards. The WHO guidelines and standard views abortion to be safe when it is performed by a qualified health professional, in an environment appropriate for such, using medical methods (WHO, 2012). Many scholars have adopted this guideline and standard definition and applied it to their work. University of Ghana http://ugspace.ug.edu.gh 13 However, it is important to note that the components of this guideline may differ based on the legal framework in specific countries. For instance, in Ghana a qualified health professional to perform abortion is a medical doctor and a midwife, however this might be different in another jurisdiction. Warriner & Shah (2006), takes the argument beyond the guidelines and standards and also beyond legality and medical aspects. They gave four (4) other components to the safety of abortion, which includes; “The decision to abort is a woman’s informed choice, made without coercion, and is free from the risk of family violence or societal stigma. The procedure is carried out in an enabling legal environment. The procedure is performed in early gestation with medically appropriate technology by an empathetic non-judgmental provider who is geographically and financially accessible. The procedure is backed up by medical services to detect and manage complications and provides information about and access to contraceptive options” Unsafe induced abortion on the other hand is deemed to be a procedure for terminating an unintended pregnancy, carried out either by persons lacking the necessary skills or in an environment that does not conform to minimal medical standards, or both (WHO, 2008). Boah et al. (2019), puts it slightly different but still maintains the core component of the WHO (2008), definition by saying “unsafe induced abortion is defined as the termination of pregnancy by a woman through the use of nonmedical method, nonmedical provider, in an environment that is not medically safe for that purpose”. Aside these characteristics that surround an abortion to warrant it to be called unsafe, there are other relevant characteristics that are acknowledged by World Health Organization (2008), and other state health institutions (Ministry of Health of the Republic of Ghana, 2012) as part of the University of Ghana http://ugspace.ug.edu.gh 14 description of abortion as unsafe. These characteristics include the absence of proper pre-abortion counselling service and the incorrect prescription of medical abortion. This suggests that the classification of induced abortion into safe or unsafe is more complex than it seems. Medical abortion as mentioned earlier has to do with abortions that make use of pharmaceutical drugs (Ministry of Health of the Republic of Ghana, 2012; Rigterink et al., 2013; Warriner & Shah, 2006). Medical abortion is suitable for pregnancies that are in the first trimester. That is, medical abortions work best when the gestation period is in its early stages. Drugs for this purposes have gone through a lot of phases, currently the most popular drugs are mifepristone and misoprostol (Appiah-Agyekum, 2018). Non-medical or surgical abortions are those abortions whereby surgical methods are used to terminate a pregnancy (Ministry of Health of the Republic of Ghana, 2012). The methods of surgical abortion include dilation and curettage (D&C), dilation and evacuation (D&E), manual vacuum aspiration (MVM) and electric vacuum aspiration (Dennis, 2012). 2.3 THE ABORTION LAW IN GHANA Ghana has one of the most liberal laws on abortion in Africa (Rominski & Lori, 2014). The practice of abortion is regulated by Act 29 of the criminal code of 1990. This act was amended in 1985 to decriminalize abortion on the conditions of rape, incest, foetal abnormality/ disease or the defilement of a female with mental health challenge. Also, the law permits abortion on grounds that the continuance of the pregnancy could risk the life of the pregnant woman or cause injury to the physical and mental health of the pregnant woman. However, this law has delayed in being translated into policy formulation and implementation (Morhee & Morhee, 2010), therefore making unsafe abortion a public health problem in Ghana. The challenge of this law has been the ambiguity in some portions of it. For example, the act permits that abortion be carried out when pregnancy puts a University of Ghana http://ugspace.ug.edu.gh 15 woman’s life at risk. The challenge here is who determines if the pregnancy puts the woman’s life at risk; is it the woman or a medical practitioner? Also, in specifying who is permitted to perform an abortion, the law states that the person must be a registered medical practitioner in gynaecology or any other registered medical practitioner. However, in Ghana not every medical practitioner is trained in abortion service provision. The question is also raised about who is a registered medical practitioner? Does it include nurses of all kinds and physician assistants or not? Another challenge is that knowledge of the abortion law in Ghana is low. According to Ghana Statistical Services (2018), only 11% of people who knew about abortion knew that abortion was legal under certain conditions in Ghana. The question then is if people do not know that the service is legal in certain circumstances how can they take advantage of it? This has implication for whether or not people seek safe abortion or induce abortion in an unsafe manner. The sensitivity of the subject of abortion expressed through stigmatization in the Ghanaian society has made the practice of abortion even under the legislations difficult (Lithur, 2004). The practice of conscientious objection by health care givers in health facilities has also affected the full implementation of the law since it makes the law non-binding to health care givers (Awoonor-Williams et al., 2018). For the abortion law to be effective and achieve its aim the law needs more clarifications, and there must be public education on the abortion law. 2.4 REGIONAL DYNAMICS IN INDUCED ABORTION As mentioned earlier, the dynamics of induced abortion vary from region to region and even from country to country. Whereas the rates of abortion in Europe and North America ranges from 15- 17 abortions per 1,000 women aged 15-49 years, that of sub-Sahara Africa is around 33 abortions per 1,000 women aged 15-49 years. In terms of rates of abortion globally, North Africa and Western Asia have the highest rate which is around 53 abortions per 1,000 women aged 15-49 (Bearak et al., 2020). University of Ghana http://ugspace.ug.edu.gh 16 The rates in Latin and Central America is around 44 abortions per 1,000 women aged 15-45 and 33 abortions per 1,000 women aged 15-49. The difference in rates in the various regions does not really give a clear picture of the abortion situation in specific countries. According to Singh et al. (2009), Europe and North America have a more liberal abortion laws hence one would expect that the abortion rates would be higher but that is not the case. In a report by the Population Reference Bureau (2021), it suggested that due to the higher contraceptive prevalence in Europe and North America, unwanted pregnancies are limited hence reducing the possibility of inducing abortion. However, in Latin America, Asia and parts of Africa were contraceptive prevalence is low and abortion is restrictive, many of the abortion cases reported are unsafe. For example, in Latin America and Africa unsafe abortion forms over 75% of the abortion cases (Population Reference Bureau, 2021). The European example gives the world a case study to draw from even though it is acknowledged that the socio-cultural dynamics are different in these regions and may be part of the reasons for the difference in rates that has been reported in these regions. 2.5 PLACE OF RESIDENCE AND INDUCED ABORTION A person’s place of abode has a great influence on the decision to terminate a pregnancy or not. The place of residence makes certain things available or not and this can affect a person’s decision to abort or not. Women in urban areas are more likely to have induced abortion because of the availability of abortion services in these areas as compared to women in the rural areas (Bankole et al., 1999). The availability of the services in these urban areas attracts health workers to urban areas as compared to the rural areas (Voetagbe et al., 2010). It can be said that health workers are attracted to urban centres more than in rural areas therefore women in rural areas might have limited qualified personnel to perform the abortion services for them when they decide to terminate a pregnancy. This leaves women who have no access to these professionals to use cruel methods to abort pregnancies University of Ghana http://ugspace.ug.edu.gh 17 which they sometimes end up paying with their lives. A Chinese study conducted in 2017, found that there is an unbalanced distribution of health workforce between rural and urban areas and that women in rural areas still face more reproductive health challenges compared to their urban counterparts (Zheng et al., 2017). In a rural study in Ghana, Hill et al. (2009) found that women in urban areas were reported to have better access to drug sellers with abortive drugs and to clinics that perform abortions and were more likely to have money to pay for these services. The finding of Gbagbo (2020), agrees with that of Hill et al. (2009) when it found that cost of abortion services varies with place and methods. Since women in rural areas are economically disadvantaged it becomes difficult for them to pay for the safe abortion services hence they rely on unsafe methods to terminate their unwanted pregnancies. Still on the matter of access, Booshehri and Dugan (2021), asserts that there is a geographic maldistribution of health facilities and health workforce. They further assert that, there is a concentration of health facilities and workforce in urban, sub-urban and affluent societies leaving the vulnerable population with inadequate health facilities. Although they make these assertions in the American context, these assertions are not far from the truth in the Ghanaian context. The major health facilities are situated in urban areas with few sub-level health facilities in the rural areas, hence if rural women are willing to even have induced abortion, there are no facilities to provide these services. The danger is that some of these women will be left with no option than to resort to unsafe methods of terminating pregnancies. Also, the distance and the road network to health centres in rural areas are in bad shape hence discourages women in the rural areas from seeking abortion services (O’Donnell, 2007). Senderowicz et al. (2019) in their abortion studies of place of residence and abortion services utilization in Mexico City found that rural women with low levels of education and are poor households disproportionately meet their abortion demands through unsafe and clandestine means. This draws attention to the fact that education plays a role in determining whether or not to terminate a pregnancy University of Ghana http://ugspace.ug.edu.gh 18 irrespective of location. In terms of the economic dimension to rural and urban dwellers and how it affects women’s abortion decision, Baruwa et al. (2021) have found that urban women are more likely to terminate pregnancy than rural women because they have better access to financial and other socio- economic resources such as health care in urban areas. According to Ouedraogo and Sundby (2014), resource influences the time a women uses to have induced abortion as well as the type of abortion a woman receives. To them, when a women has finances to take care of the cost of abortion services the abortion is done faster as compared to when women lack the finances to take care of the cost of having induced abortion. Lastly, it is important to note that cultural influence varies in rural and urban areas and this has a way of influencing women’s decision to terminate pregnancy or not. Rasch and Kipingili (2009), in their Tanzanian study found that cultural difference between urban and rural women varied with regard to the number of children they wanted to have, thus urban women wanted less children due to the urban culture they have been exposed to and rural women due to the cultural exposure also wanted more children. Due to the fact that women in urban women want smaller number of children they tend to terminate any pregnancy that comes after they have had their required number. Sutton et al. (2019), in a study conducted in USA found that there was race/ethnic disparity in rural and urban places when it comes to service delivery to black adolescents. In a situation like this these black adolescents might either tend to cruel abortion methods or may be forced to keep pregnancies they do not intend to keep. Rominski et al. (2014), have found that attitudes towards abortion in urban Ghana suggest that women face significant social and cultural hurdles to seeking safe abortion services regardless of the availability of service. However these urban women feel more empowered to admit that they have had abortion as compared to their rural counterparts. These differentials in characteristics University of Ghana http://ugspace.ug.edu.gh 19 among urban and rural women makes the study of rural-urban differentials in induced abortion and every important one. 2.6 SOCIO-ECONOMIC AND DEMOGRAPHIC FACTORS INFLUENCING INDUCED ABORTION Induced abortion is influenced by several factors ranging from the socio-demographic characteristics of women to institutional and communal factors as well as national policy factors. This section would therefore discuss some of these factors. 2.6.1 Age of a woman The motivation to carry pregnancy to term sometimes depends on the age of the woman carrying the pregnancy. A younger woman may want to delay child birth due to several challenges that is accompanied with early child birth such as health risk, financial and emotional burden of child birth and many other reasons. Women at younger ages may therefore induce abortion to save them from these challenges. Also, woman of older ages might not want to carry pregnancies to term due to several reason, some would either be because of health complications or they already have the desired number of children. However, women in their middle ages would be highly motivated to carry pregnancies to term because that is when they have many of their lives plan which includes child birth in motion. These dynamics might also differ based on jurisdictions. In a study in conducted Ghana, Sundaram et al. (2012), found that adolescents are less likely to induce abortion since they are less sexually active compared to older women. Due to their infrequent sexual activity, unwanted pregnancy that may force them to induce abortion is minimized. Younger women are also forced to induce abortion due to the stigmatization they are likely to receive from family and society including health service providers. Young women also terminate pregnancies in order to enable them continue schooling or training (Ghana Statistical Services, 2018). University of Ghana http://ugspace.ug.edu.gh 20 Additionally, a study conducted in Ethiopia found that younger women are more likely to induce abortion (Wodajo et al., 2017). This phenomenon discovered in the study by Wodajo et al. (2017) has been attributed to cultural acceptability, therefore adolescents who become pregnant premarital rely on induced abortion in order not to appear to be living outside the culturally accepted norms. According to Sully et al. (2018), there are higher rates of abortion in young women as compared to those in older ages and this was attributed to the law that allow adolescents to have an abortion in Ethiopia. Again, a study by Van Look & von Hertzen (1993) in English speaking developed countries found that abortion was higher in young ages, since young women were not ready for child bearing therefore use abortion to postponed birth. In a study in Ghana, Boah et al. (2019) found that a woman’s age is strongly associated with induced abortion and that older women are less likely to induce abortion unsafely since at older ages they would be managing the number of children they want. Mote et al. (2010) however found that older women in Hohoe in the Volta region of Ghana are more likely to induce abortion since there is lack of contraception use that results in unwanted pregnancies and induced abortions. 2.6.2 Current Marital Status Being married or not is an important factor when it comes to a woman’s decision to induce abortion. Marital status also has religious, cultural and societal undertones. A woman in a society where premarital pregnancy is frowned upon is more likely to have induced abortion compared to women who are married. A study in Cameroon, by Johnson-Hanks (2002) found that unmarried women tend to have more abortions than their married counterparts because premarital pregnancy is not condoned in the society so women who get pregnant outside the confines of marriage find it as a plausible way to save the situation. In Ghana, a study in the Volta region, one of the administrative regions in Ghana also found that unmarried women are more likely to induce abortion because they want to keep their University of Ghana http://ugspace.ug.edu.gh 21 current employment or stay in school (Mote et al., 2010). Browner (1979) in studying the abortion decision making in Columbia found that unions that are consensual took stronger steps to terminate unwanted pregnancies. However, some women in marriages tend to induce abortions just because they are uncertain about the future of the marriage due to the stress levels in that marriage. In the same study, when decisions are to be made about abortions men had more authority with some using withdrawal of economic support as a weapon to convince women. The male dominance in pregnancy termination decision making is also held by a study in Ghana (Kumi-Kyereme et al., 2014). This study reveal that when it comes to unions and abortion, the decision goes beyond the woman to other external factors. In another vein, some studies have also revealed that married women are less likely to induce abortion (Baruwa et al., 2021; Ilboudo et al., 2014). Married women are expected to have children especially in Africa where children are seen as a blessing from God, hence induced abortion is a less likely choice for a married woman when she is pregnant. 2.6.3 Level of educational attainment How well a women is educated goes a long way to influence their decision of inducing abortion. When a woman is educated she has better access to information on abortion services and about reproductive health (Sundaram et al., 2012). When women have access to information they stand at an advantage and know where to have safe abortion services. The safety of abortion service may be a better motivation to induce abortion by an educated woman. A similar assertion has been made by Baruwa et al. (2021) in terms of access but they indicated that educated women do not only have access to information but also access to health service providers and having access to such can influence a woman’s decision to induce abortion or not. A study in Mexico City has also established the association between level of education and induced abortion (Senderowicz et al., 2019). In this study, it was found that educated women are more concerned about their family size and the opportunity cost associated University of Ghana http://ugspace.ug.edu.gh 22 with child bearing and raising a child. Educated women were likely to have smaller family sizes that they can cater for hence in the event of any unwanted pregnancy they are most likely to abort it. Educated women show more dedication to aborting pregnancies, one of the reason may be that most educated women are career women and would do anything to abort a pregnancy if the pregnancy become an inhibiting factor the pursuit of their careers (Senderowicz et al., 2019). Education has some varying associations with levels of abortion in certain contexts. In an Ethiopian study, Tesfaye et al. (2014)found that the higher a woman’s level of education the lower the levels of abortion but Biney & Nyarko (2017) found the opposite in Ghana. Dias et al. (2015) conducted a study in Brazil on the association between level of education and access to abortion services and found that higher proportions of adolescent girls terminated their first pregnancy because they were in higher education. In the same study uneducated women even though did not induce abortion more than the educated women, had more unwanted pregnancies mainly due to the non-use of contraceptives 2.6.4 Current Contraceptive use One of the key components of fertility as espoused by John Bongaarts in fertility was contraceptive usage (Bongaarts, 1978). He further demonstrates how to calculate contraceptive use mathematically. This demonstrates the importance he placed on contraception use when it comes to birth control. For a woman to get to the junction of wanting abortion, contraception use or non-use might have had a role to play. Van Look & von Hertzen, (1993), says that “choices for couples who wish to limit their fertility are but three: abstinence, contraception or abortion”. This statement echoes the linkage between contraception use and abortion. However, contraceptive use is highly affected by the distribution of family planning clinics and education. Education especially determines the differentials that may exist between the various groups of a population (Addai, 1999). Again, the type of service providers are a critical part in contraceptive use since accommodating providers attracts University of Ghana http://ugspace.ug.edu.gh 23 women to come for service and unaccommodating providers scare people away. Maxwell et al. (2015) in a study to determine the use of contraceptive after abortion, found that women that were taken care of by midwives had higher odds of using contraceptives after abortion than any other providers. The training and professionalism of midwives cannot be downplayed in this finding. Contraceptive use dynamics can also change based on age because in a study in Ghana, Salifu and Mohammed (2020) found that women 15-19 years were more likely to use contraceptives as compared to women 30+ years old. This might be because younger women would want to prevent unwanted pregnancy while older women actually wants the pregnancy and are more prepared to take care of the child that would result in the pregnancy. Some women have also refused to use contraceptives due to its side effects, failure of the method or the husband’s disapproval of it (Biney, 2011). In Brekum, a study found that even though knowledge of contraception was high its usage was very low (Geelhoed et al., 2002). Rasch et al. (2000) also realized in their study in Ethiopia that only 10% of women interviewed used contraception and this was because of the limitation on the access to family planning clinics to only pregnant women or women who have ever been pregnant. In the same Ethiopia, Mekuria et al. (2019) found that a lot of single women do not use contraception after abortion due to the unmet needs of contraception. Munakampe et al. (2018) after their review of contraception and abortion practices and attitudes among adolescents in low and middle income countries had this to say: “Limited knowledge about sexual and reproductive health among adolescents was a significant cause of reduced access to contraception and safe abortion services, especially among unmarried adolescents. Reduced access to reproductive health services for some resulted in extreme methods of contraception and abortion such as the use of battery acid and crushed bottles. University of Ghana http://ugspace.ug.edu.gh 24 Despite all adolescents having limited access to information and services, girls faced more consequences such as being blamed for pregnancy or dealing with the effects of unsafe abortions. Parents, health workers, and teachers were cited as trusted sources of information but often received the most information from peers and other family members instead, and the girls mostly confided in their aunties, cousins and peers while the boys resorted to peers, media and even pornography” 2.6.5 Ethnicity The cultural affiliation of a person can have an influence on their decision to induce abortion or not. In a cultural setting where abortion is not seen as negative but a smart move, women would be more comfortable inducing abortion as compared to societies where it is frowned upon. In some cultural settings women are stigmatized against when they have been found to have induced abortion (Lithur, 2004). The traditional and cultural values that shape various ethnic groups have inbuilt mechanism to check premarital pregnancies by placing values on rites of passage as a way of family and community acceptance (Lithur, 2004). Sadly, these rites rather enhance abortions among the youth since they tend to induce abortions in order to avoid shame. For instance Lithur (2004), found that in the Ga tradition women who have induced abortion are given derogatory names and to some extent to include their families hence women tend to abort pregnancies using dangerous and clandestine methods. However, in the Akan tradition abortions are acceptable when they are successful but reprehensive when they are not successful (Bleek, 1990). In a study conducted by Bleek (1990), it was found that rapid succession of pregnancies is one of the reasons why women induce abortion. Addai (1999), a study on ethnicity and contraceptive use, it was found that the Ga-Adangbe ethnic group have a better contraceptive University of Ghana http://ugspace.ug.edu.gh 25 prevalence than the Akans and this may increase unwanted pregnancies in the Akan ethnic group as compared to the Ga-Adangbe ethnic group and may compel them to abort the pregnancies. The traditional perceptions of these ethnic groups when it comes to sex during and after child birth can influence their decision to induce abortion directly or indirectly. It is largely believed in the Ewe ethnic group that women are supposed to abstain from sex for 156 weeks after birth (Gaise, 1968 as cited in Addai, 1999). If women go strictly by this in the Ewe ethnic group unwanted pregnancies are minimized and the tendency to induce abortion will also be minimized. The Mole-Dagbani ethnic group however, on the matter of abstaining from sex after birth do not see it as a ritual hence expect some lapses (Gaise, 1968 as cited in Addai, 1999). With this belief the Mole-Dagbani woman are likely to engage in more sexual activities after birth which may lead to unwanted pregnancies which may necessitate induced abortion. 2.6.6 Number of children ever born The number and type (sex) of children a woman has or desires can influence the abortion decisions of that woman. The decision may not be only based on desire and influence by education but also the legal requirement under the jurisdiction a woman finds herself. In a study involving countries in North China it was found that the decision to terminate pregnancy is associated with the number of children the person has. The women were found to have induced abortion only after their first live birth (Tu Ping & Smith, 1995). The context of the study makes the findings more acceptable since under the two child per family policy people would be careful about the type of children they want and the spacing, and to do that, aborting a pregnancy after scan proves that the sex is not that which was desired by family would be a rational move. Bankole et al. (1998), also found strong association between parity and induced abortion. They found that women without children received the least percentage of abortions but women with at least a child University of Ghana http://ugspace.ug.edu.gh 26 had higher percentages of abortions. Similar finding were made by Ahiadeke (2001) when women with four or more children had higher risk of abortion largely because they have reached the pinnacle of their desired children. In a study in the United States of America (USA), women were not ready to have any children or another child because they were not ready for the process again hence resort to abortion (Finer et al., 2005). 2.6.7 Religion A person’s belief in a particular religious doctrine has a way of influencing their abortion decisions (Adamczyk, 2008). The influence of religious beliefs does not only influence the seeking to terminate pregnancy but also the one providing the services of pregnancy termination (Oppong-Darko et al., 2017). In a study exploring conscientious objection among health workers in Ghana, health workers cited religion and morals as that which makes them not want to perform an abortion procedure (Awoonor-Williams et al., (2020). In a scenario where service providers use religious conflict as a reason for not terminating pregnancy, women rely on unsafe means of abortion to help themselves. However, having said that there are some providers that are willing to carry out their professional duties even though it conflicts with their religious beliefs (Oppong-Darko et al., 2017). The magnitude of the role of religion on a person’s decision to induce abortion varies based on which religious sect one finds herself. Catholics and Muslims are less likely to induce abortion because catholic doctrines are against such acts and for the Muslims they are seen to cherish larger families therefore ceteris paribus, they are less likely to induce abortion (Baruwa et al., 2021). In Brazil, Ogland & Verona (2011) found that instead of Catholics, Pentecostals were rather vocal in their opposition to abortion which was a shocking finding because Catholics have been vocal on these matters in other jurisdictions. The findings in Brazil is attributed to what Ogland & Verona (2011), describes as the “lukewarm identification with Catholicism and the internal polarization of Catholicism”. University of Ghana http://ugspace.ug.edu.gh 27 Religion even influences people’s sexual activeness in that it can prevent people from having premarital sex that may lead to unwanted pregnancy and contemplation of abortion (Adamczyk & Hayes, 2012). The complexities and dynamics of religion and how it affects a woman’s decision to induce abortion or not makes it a good predictor of induced abortion. 2.6.8 Wealth Quintile The decision to terminate a pregnancy has always had an economic dimension since it involves payments of money for the service and also there are economic implications for keeping the pregnancy. The economic dimensions of decision to terminate a pregnancy may arise from the cost off carrying pregnancy to term and the cost of raising a child. Women who are not economically stable are likely to abort pregnancies than women who are economically stable and can cater for the children (Yi, 2011). Women who are economically stable are not likely to terminate pregnancy because they can take care of a child but also they have the means of preventing the pregnancy from coming by using contraceptive (Oppong et al., 2021). Economically empowered women have been found to have friends that they can confide in and these friends serve as a social support for them and that is factored into their abortion decision (Guiella et al., 2006). One of the factors that influence a woman’s decision to keep pregnancy is the availability of support to help raise the child, so if a woman (economically empowered woman) can get that support from friends and family then they are less likely to terminate a pregnancy. In Ghana, Payne et al. (2013), found that the higher cost involved in inducing abortion in the government hospitals scare women from the low socio-economic backgrounds who then resort to unregulated places for abortion. The women from the better socio-economic backgrounds see the government hospitals as a second choice, thus they prefer to go to private hospitals to have the abortion done. Also, some women resort to unsafe abortion methods due to financial constraints (Biney & University of Ghana http://ugspace.ug.edu.gh 28 Atiglo, 2017). These reasons makes the wealth a woman has or finds around her a good indicator when it comes to induced abortion. 2.6.9 Exposure to the media The media has been referred to in many democratic jurisdictions as the fourth realm of the state and we cannot over emphasize the power and responsibility of the media in the development of a nation. The mandate of the media is to inform, educate and entertain. It is in this mandate (especially the first two) that they influence women’s decision when it comes to issues of induced abortion. Women who have access to the media are likely to have access to abortion information so in Kenya abortion issues have been framed in media contents to help influence policy (Kafu et al., 2020). Boah et al. (2019), found in a study that women with no exposure to the media are more likely to have unsafe abortion due to the less knowledge of information they have on the matter. In a society such as Africa where matters like abortion are perceived as a taboo and cannot be openly talked about among women, the media is heavily relied upon for trusted information. Self- efficacy in abortion decision making have been attributed to the exposure of the media not just for the information it provides on abortion but also the exposure it gives especially to young women on western ideas of autonomy (Ahinkorah et al., 2020). 2.6.10 Knowledge of the abortion law One’s understanding of the law that governs his or her actions is important for the confidence and willingness to take that action or not. Similarly, a woman’s knowledge of the abortion law in their jurisdiction can go a long way to affect the decisions they take on abortion and the methods they use as well. Several studies have found that in Ghana, knowledge of abortion law is low even among health providers and the situation is no different from other Africa countries (Debela & Mekuria, 2018; University of Ghana http://ugspace.ug.edu.gh 29 Frederico et al., 2020; Voetagbe et al, 2010; Konney et al., 2009). This is a disturbing feature since if people do not know under which conditions they can receive safe abortion services then they may be at risk of going for unsafe methods even if their conditions permit a safe abortion. Contrary to that, Morhee & Morhee (2010), suggest that medical practitioners have knowledge of the abortion law but they think it lacks clarity especially in Ghana. Boah et al. (2019), also think that the reluctantly unrestrictive nature of the abortion law in Ghana should have addressed the unsafe abortion canker in Ghana but thinks that the problem has been the lack of awareness of the law. Awoonor-Williams et al. (2018), noticed that service providers and patient’s knowledge of the abortion law are important since it affects pregnancy outcomes. However, there is some misunderstanding of the legality of abortion laws in Ghana among some health workers and some think that the legalization of abortion services would rather increase abortion rates (Oppong-Darko et al., 2017). In her three recommendations for ending stigmatization on abortion, Lithur (2004), recommended that there should be a liberal interpretation of the abortion law so that people can relate with it better. This points to the fact that people’s knowledge of the abortion law can help address the stigmatization attached to abortion. 2.7 THEORETICAL PERSPECTIVES The phenomenon of induced abortion has been studied using several theories from different fields of study including those developed in the field of social science. Some of the theories and models that have been used to study induced abortion include the renewal theory. The renewal theory is based on the assumption that individuals are immediately replaced as soon as they die. Potter (1972), studied induced abortion using this theory. Also, the theory of planned behaviour (TPB) has been used to study induced abortion and other related topics. TPB is a theory which explains the intentions that motivate people’s behaviour at University of Ghana http://ugspace.ug.edu.gh 30 a particular time and place. These actions or behaviour are attached to attitudes, norms and perceived power. These norms can be either subjective or societal (Ajzen, 1991). Foy et al. (2005), used the TPB theory to study induced abortion in trying to identify barriers to quality improvement initiatives. Aizen & Klobas (2013), Klobas (2011), and Moshi et al. (2020) used the same theory to study fertility and stillbirth in health care facilities. Susannah and Aniteye (2013), studied legal abortion in Ghana using the policy theory of Lipsky Street-level bureaucracy. This theory explains the role of dynamics involved by frontline workers (mostly civil and public servants) in the implementation of policies. Other theories that has been used in studying abortion are the socio-ecological health model (SEM) and the social determinants of maternal health. These two models have been widely used in social science studies. The socio-ecological health model gives description to the interaction between individuals and their surroundings that influence their health outcomes (Golden & Earp, 2012). The surrounding can stretch from immediate to distal. The immediate can be the family and friends whereas the distal can be the state or federal level. The proponents of this model suggested five (5) levels of interaction. These are individual level, family/friend/peers level, the organizational level, the community level and the policy/state level. These levels of interaction posited by the SEM model helps explain the interactions that go on in an individual’s life until the decision or action of induced abortion is taken. For instance, at the individual level the woman’s age, level of education, contraceptive use and abortion history can influence the woman’s decision to induce abortion. These characteristics at the individual level serves as a motivation or disincentive for a woman to induce abortion. At the family/friends/ peer level which others have called the interpersonal level, a woman’s marital status and the number of children she has for example would influence the decision to induce abortion or not. The opinion of a partner or the family a woman belongs would be an important factor and that may influence the outcome of a pregnancy of which induced abortion is a possibility. Similarly, University of Ghana http://ugspace.ug.edu.gh 31 at the community level the availability of social networks or support systems are likely to influence a woman’s decision to induce abortion or not. At the organizational level a woman who belongs to a religious organization where abortion is frowned upon is less likely to induce abortion as compared to a woman who has no religious affiliation. At this same level a woman in a work environment where women are not permitted to take maternal leave or face the threat of being replaced if they get pregnant, are likely to induce abortion as compared to a work environment that is more liberal to being pregnant. Lastly, in the policy level, when a country has policies in place that gives women the opportunity to induce abortion safely and legally, women would be likely to have safer abortions as compared to a place where abortion is illegal. For this reason, the abortion law in Ghana can be considered as a more structural or policy level indicator although in this study it is being asked at the individual level. Similar to the socio-ecological health model, is the social determinants of health model which has been adapted by several scholars to explain their studies in social science. One of such studies by Koroma et al. (2017) used the model to study maternal health. They adapted the social determinants health model to form the social determinants of maternal health model. The social determinants of maternal health model also explains health outcomes as a product of the influence of an external environment. It explains that an outcome is influenced by individual characteristics which are also influenced by family, who are also influenced by the community and finally the larger cultural and social context as well as the national policy context. These various levels have different influence on the individual. The framework shown in Figure 2.1 helps to explain induced abortion as an outcome variable. From the diagram, maternal health is influenced by individual attributes of women such as age, number of children, knowledge of maternal health issues, etc. Similarly, induced abortion as an outcome University of Ghana http://ugspace.ug.edu.gh 32 variable can be influenced by a woman’s characteristics such as age, educational level, contraceptive use and number of children. In Figure 2.1, the outcome variable and individual attributes are nested or influenced by a bigger bracket of family and peer influences. Induced abortion can be placed in the same context where a person’s marital status and social network/ support can influence the person to have an abortion or not. The community context of rural and urban residence with some associated factors also influence the family and the individual that influences the outcome variable which in the context of this study is induced abortion. In communities where there is access to health facilities and the cost of abortion is affordable, people are more likely to have more induced abortion than places where such access does not exist. Governance/Policies of the social determinants of maternal model depicted in the diagram shows a broader influence of community, family and individual factors on health outcome as well as the broader cultural and societal values and norms. Figure 2.1: Social determinants of maternal health Adopted from Koroma et al. (2017) as Adapted from: WHO (2011) University of Ghana http://ugspace.ug.edu.gh 33 At the governance and policy level, health infrastructure, laws, reproductive rights etc. are identified to influence the other components. Induced abortion is also affected by the availability of health infrastructure. The availability of the infrastructure can influence a woman’s decision to induce abortion or not. If the laws of a country make abortion illegal then that can affect the freedom to which it can be done as compared to a country where it is legal. In a country where there are reproductive policies that are vigorously being implemented, the abortion situation can be improved compared to a situation where there is no policy or there are but implementation is a challenge. The culture and social values of a jurisdiction has an overriding effect on so many things. For instance this model mentions religion. If a women lives in an Islamic state or a state in which abortion is considered an abomination, this may be translated into laws that may affect the family and individual on how to go about the outcome variable, in this case, induced abortion. This study adapts the social determinants of maternal health model because its application is appropriate for the study and the context in which it was applied (Sierra Leone) has similar characteristics as Ghana, although Sierra Leone with the history of conflicts may make the contexts not directly comparable. Also, its application in the study of maternal health is closely related to induce abortion and it would be appropriate to apply it to a study on induced abortion in the Ghanaian context. The model appreciates the role of community context in influencing an outcome variable and in a study where differentials by place of residence (rural or urban) is of interest it becomes appropriate to apply the social determinants of maternal health framework. University of Ghana http://ugspace.ug.edu.gh 34 2.8 CONCEPTUAL FRAMEWORK The conceptual framework in Figure 2.2 shows the various factors that influence induced abortion among women in Ghana. This framework was adapted from a study conducted by Koroma et al. (2017) in the research where quality of free antenatal and delivery services was studied. The study by Koroma et al. (2017), looked at factors that affect maternal health from the perspective of women characteristics and other external factors such as family, community and policy level factors. It was found in this study that other external factors influence maternal health aside the woman characteristics. Similarly, the framework in Figure 2.2 conceptualizes that the outcome variable which is induced abortion is influenced by individual woman’s characteristics and socio-economic and socio-cultural factors. The individual woman’s characteristics include age, current marital status, level of educational attainment, contraceptive use and number of children ever born. The age of the woman can determine her decision to induce abortion or not. For example, a women at a younger age is likely to induce abortion compared to a woman at an older age (Sundaram et al., 2012). In the framework in Figure 2.2, a woman’s marital status can determine whether they induce abortion or not. A woman who is married especially in Africa is less likely to induce abortion compared to a woman who is not in union (Baruwa et al., 2021). When a woman is educated she has better access to information and stands a better advantage of knowing where to have abortion services. When a woman uses contraceptives to prevent pregnancies she is less likely to have unwanted pregnancies that may lead to inducing abortion. Also, when a woman has many children, she is likely to induce abortion (Ahiadeke, 2001). Again, a woman who has access to the media is expected to have access to information on maternal health outcomes which in this case is induced abortion. This gives the women an advantage with regards to decision making on induced abortion. University of Ghana http://ugspace.ug.edu.gh 35 Aside the characteristics of the individual woman, she is influenced by other socioeconomic and socio-cultural factors such as ethnicity and religion. The ethnic affiliation of a woman can influence the woman’s decision to either induce abortion or not. In ethnic groups where premarital pregnancies are considered shameful, the women would do their best to prevent premarital pregnancies that may lead to the pregnancy being induced. Similarly, religion also influences the woman’s decision to induce abortion. In Figure 2.2 religion and ethnicity forms the socio-cultural factors that influence a woman to induce abortion. At the community level, a woman is either living in an urban area or a rural area and that affects their access to health facilities and health professionals. The place of residence can also affect their access to education and this influences a woman’s ability to induce abortion. The women who live in urban areas and have access to health facilities, health professionals and education are likely to induce abortion more and safely as compared to the women in the rural areas who lack access to the things their counterparts in the urban areas have. The community is made up of families where women find themselves. The wealth available to the family influences the woman’s decision to induce abortion. Ahiadeke (2001), found that poverty was one of the reasons why women induced abortion hence if a woman finds herself in s family that falls in a low wealth quintile that can influence their decision to induce abortion. Lastly, at the structural or governance level is the knowledge of abortion law. The law regulating abortion in a country has an overarching effect on the community, family and the individual woman. The knowledge of the law regulating abortion is very critical to a woman’s decision to induce abortion. For example, when a woman in Ghana knows abortion is legal on the basis of rape and incest and she gets pregnant through these means she could easily take advantage of the existence of such law to get the pregnancy terminated. University of Ghana http://ugspace.ug.edu.gh 36 Figure 2.2: Conceptual framework showing the factors that influence induce abortion 2.9 HYPOTHESES 1. The chances of induced abortion is more likely to be high in urban areas than in the rural areas. 2. Urban areas are more likely to have level of educational attainment as a significant predictor of induced abortion than rural areas. Adapted from Koroma et al., (2017) Structural Determinants Knowledge of abortion law Community Context Place of residence - Urban - Rural Family Context Family context Wealth quintile Individual woman’s characteristics Age Current marital Status Level of educational attainment Current contraceptive use Number of children ever born Exposure to media Socio - economic and cultural factors Ethnicity Religion Maternal health outcome Induced abortion - No - Yes University of Ghana http://ugspace.ug.edu.gh 37 CHAPTER THREE METHODOLOGY 3.1 INTRODUCTION This chapter outlines the methodology employed in this study. The chapter includes information on the source of data, sample design and sample size, variables considered for the study, including their measurement and categorisation and how the data were analysed. 3.2 SOURCE OF DATA The data used for this study is from the Ghana Maternal Health Survey (GMHS) conducted in 2017. The GMHS is a nationwide survey conducted by the Ghana Statistical Service (GSS) with technical support from Inter City Fund (ICF). The data collection process was in two phases. The first phase included listing of households and screening for deaths that has occurred in the various households since January 2012. The second phase included the confirmation of the reported deaths in the households. Also, verbal autopsy interviews were conducted using the verbal autopsy questionnaire and women aged 15-49 were also interviewed using the women questionnaire. The process of data collection started on 15th June, 2017 and ended on the 12th October, 2017. The survey was a cross- sectional survey that took place in all the then ten regions of Ghana with the aim of accessing levels of maternal deaths in the country as well as identifying the causes of maternal and non-maternal deaths in the country. This GMHS is currently the most recent survey on maternal health issues with a nationwide coverage and provides in-depth data on maternal health in Ghana. University of Ghana http://ugspace.ug.edu.gh 38 3.3 SAMPLE DESIGN AND STUDY SAMPLE The sampling frame for this study was the frame generated from the 2010 Population and Housing census conducted in Ghana. In this frame the 10 administrative regions of Ghana were divided into smaller units called enumeration areas (EAs). The survey used stratified sampling to select samples from the sample frame. This was done by dividing regions into urban and rural areas. EAs were then selected using probability proportional to size from urban and rural areas. Household listing was then done in these EAs and the resulting household listing served as sampling frame for the selection of a sample of 30 households to be interviewed. A total of 27,000 households were selected nationwide and 26,324 households were interviewed. Out of the households interviewed 25,062 interviews were completed. However, in this study a total weighted sample of 13,176 women (from an unweighted sample of 13,419) would be used. This sample was extracted from data from women who have been pregnant in the five years preceding the survey (2012 to 2017). 3.4 VARIABLES IN THE STUDY The dependent variable for this study is induced abortion and the independent variable is place of residence. The control variables are age of woman, current marital status, level of educational attainment, current contraceptive use, number of children ever born, ethnicity, religion, wealth quintile, exposure to media and knowledge of Ghana’s abortion law. 3.4.1 DEPENDENT VARIABLE The dependent variable in this study is induced abortion. It was computed from the sample of women who have been pregnant in the five years preceding the survey. These women who have been pregnant in the five years preceding the survey were those who indicated having one or more live births, induced abortions, miscarriages, and still births from 2012-2017. Among this group of women, those University of Ghana http://ugspace.ug.edu.gh 39 who had one or more live births, miscarriages and stillbirths within that period were categorized as “No = 0”, indicating they have not had an induced abortion in that period while all those who have had one or more induced abortions in the same period was categorized as “Yes = 1”, thus, they have had an induced abortion. 3.4.2 INDEPENDENT VARIABLE The independent variable is place of residence. It was treated as a dichotomous variable which was categorised into two; urban and rural. Urban was coded as “1” and rural was coded as “2”. These categories were maintained for this study. 3.4.3 CONTROL VARIABLES The control variables for this study are; 3.4.3.1 Age of woman This was the age of the women at the time the data was being collected. The original variable was a continuous variable measuring age in completed years. This variable was recoded into a categorical variable. The categories are “< 20 =1”, “20-29 =2”, “30-39 =3” and “40-49 =4”. 3.4.3.2 Current marital status This variable was originally a categorical variable with the categories being currently married, living with a man and not in union. The original codes for the variable were “currently married =1”, “living with a man =2” and “Not in a union = 3”. However, for this study the variable was recoded into “Married= 1”, “cohabiting = 2” and “not married =3”. University of Ghana http://ugspace.ug.edu.gh 40 3.4.3.3 Level of educational attainment This variable was originally a categorical variable with six categories mainly primary, middle, JSS/JHS, Secondary/Technical /Vocational/Commercial, SSS/SHS/Technical/Commercial and higher. These categories were coded as “primary education =1”, “middle =2”, “JSS/JHS =3”, “Secondary/Technical/Vocational/Commercial=4”, “SSS/SHS/Technical/Vocational/Commercial = 5”and “Higher = 6”. For the purpose of this study the variable was recoded into “No education = 0”, “Primary =1”, “Middle/JHS = 2”, “Secondary = 3” and “Higher = 4” with the secondary category encapsulating the secondary/Technical/Vocational/Commercial and SSS/SHS/Technical/Vocational/Commercial categories. 3.4.3.4 Current contraceptive use This variable was used as a dichotomous variable that was coded as “Yes =1” and “No = 2”, when women were asked if they or their husbands were doing something to prevent pregnancy. This variable and its categorization were maintained for this study. 3.4.3.5 Number of children ever born The number of children a woman has ever had was measured as a discrete variable in the original data. This was then recoded into a categorical variable with the following categories, “None = 0”, “1=1”, “2 - 4 =2” and “5+ =3”. 3.4.3.6 Ethnicity This variable was a categorical variable with nine categories. However, for the purposes of this study these categories were recoded into five new categories. The new categories are “Akan= 1”, “Ga/Dangme = 2”, “Ewe = 3”, “Mole-Dagbani = 4” and “others = 5”. The other category is made up of ethnic groups such as Guan, Grusi, Gurma, Mande and other smaller ethnic groups. University of Ghana http://ugspace.ug.edu.gh 41 3.4.3.7 Religion This variable was a categorical variable with ten categories coded as “Catholic=1”, “Anglican= 2”, “Methodist=3”, “Presbyterian=4”, “Pentecostal/Charismatic=5”, “other Christian=6”, “Islam=7”, “Traditionalist/Spiritualist=8”, “no religion=9” and “others=96”. These categories were recoded into new categories which are “No religion = 0”, “Orthodox =1” made up of women whose religious affiliation are Catholic, Anglican, Methodist and Presbyterian, “Pentecostal/Charismatic = 2”, “other Christian =3”, “Islam =4” and “Traditionalist/Spiritualist = 5” which is made up of people whose religious affiliation are traditionalist/ spiritualist and others. 3.4.3.8 Wealth Quintile This variable measured the wealth of the household the respondent belongs to taking into consideration their urban and rural dwelling. These categories included “lowest=1”, “second=2”, “middle=3”, “fourth=4” and “highest=5”. For the purpose of this study this categories were renamed into “poorest=1”, “poor=2”, “middle=3”, “rich = 4” and “richest=5”. 3.4.3.9 Exposure to media This variable was computed from three variables that sought to find out the number of times respondents read newspaper, listened to radio and watched television. The responses for these original variables were at least once a week, less than once a week and not at all. These variables were used to compute a new variable called exposure to media. Respondents who responded at least once a week to all the three variables were coded to be “exposed to media= 1” and all others were coded as “Not exposed to media =2” 3.4.3.10 Knowledge of abortion law This variable sought to know if respondents had any knowledge about the legality of abortion in Ghana. This variable had three categories that included “Yes =1”, “No =2” and “don’t know=3”. For the University of Ghana http://ugspace.ug.edu.gh 42 purpose of this study and also to ensure that the knowledge of abortion law was unequivocal, the last two were recoded into “No = 2” and “Yes =1” was maintained to have a dichotomous variable. 3.5 VARIABLES AND THEIR CATEGORIZATION Table 3.1 shows the variables used for this study and how they were categorised. It also shows codes assigned to the various categorizations. Table 3.1: Variables and their categorizations D DEPENDENT VARIABLE C CATEGORIZATION I Induced abortion [ [No = 0] [Yes = 1] I INDEPENDENT VARIABLE C CATEGORIZATION P Place of residence [ [Urban = 1 ] [Rural = 2] C CONTROL VARIABLES C CATEGORIZATION A Age of woman [<20 =1] [20-29 = 2] [30-39 = 3] [ 40-49 = 4] C Current marital status [Married = 1] [Cohabiting = 2] [Not married = 3] L Level of educational attainment [No education = 0][Primary = 1][Middle/JHS = 2][Secondary =2] [ [ Higher = 3] Current contraceptive use [No = 1][Yes = 2] Number of children ever born [None = 0] [1= 1] [2-4 =2] [ 5+ = 3] Ethnicity [Akan = 1] [Ga/Dangme = 2][Ewe = 3][Mole-Dagbani = 4][Other = 5] Religion [No religion = 0][Orthodox = 1][Pentecostal/Charismatic = 2] [ [Other Christian = 3] [Islam = 4] [Traditionalist/Spiritualist = 5] Wealth Quintile [ [Poorest = 1] [Poor = 2] [Middle = 3][Rich = 4] [Richest = 5] Exposure to media [ [Exposed to media = 1][Not exposed to media = 2] Knowledge of abortion law [ [Yes = 1] [No = 2] 3.6 METHODS OF ANALYSIS The data was analysed at three levels. These levels are the univariate, bivariate and multivariate levels using the IBM SPSS (Statistical Package for the Social Sciences) package, version 25. At the various levels, tables were used to show results from the analysis. Percentage distributions were used to show the percentage distributions for the variables at the univariate level. At the bivariate level the University of Ghana http://ugspace.ug.edu.gh 43 associations between the dependent variable and independent variable as well as the control variables were explored. At this l