PERCEPTION OF RISK OF HIV/AIDS INFECTION AND SEXUAL BEHAVIOUR AMONG WOMEN IN THE NORTH WEST PROVINCE, SOUTH AFRICA By DrMAEyassu MD Submitted in partial fulfilment for the requirement of the Master Degree of Social Sciences at the Department of Population Studies (North West University, Majikeng Campus). Supervisor: Dr Moses Kibet December, 2005 D ECLARATION I Melaku A Eyassu declare that, this is my own work and have never been submitted to any other university Signature ABSTRACT The association between the perception the of risk of IDV infection and sexual behaviour remains poorly understood, although the perception of risk is considered to be the first stage towards behavioural change, from risk taking to safer behaviour. Using data from a 1998 South Africa demographic and health survey, the knowledge, the perception of the risk of IDV/AIDS and the sexual behaviour of women in the North West province were analysed. Nominal regression models were implemented to examine the strength of the association between the perception of the risk of IDV/ AIDS and explanatory variables. The findings indicate that knowledge is universal and women in the North West are well informed about mode of transmission of IDV/A IDS . Married and older the women are less likely to use condoms. More than 88% of the women believe that an IDV/ AIDS status should be reported to the patients themselves and their partners. Women's perceptions were examined about whether a healthy person can have IDV/AIDS? About 30% of the women said that a healthy person could have IDV/ AIDS, but 53% said that a healthy person could not have IDV/ AIDS . With regard to the association between the perception of the risk of IDV/ AIDS and certain explanatory variables (place of residence; education; ethnicity; age at first sexual intercourse; type of sexual partners; marital status; AIDS information from newspapers, friends and partners and knowing someone with IDV/AIDS), it is shown that they have a strong association. ii ACKNOWLEDGEMENTS i would like to take this opportunity to thank all the people who have an inputs and to make this research possible. I am grateful to my supervisor, Dr Moses Kibet, of the North West University, Mafikeng Campus, for his continuous encouragement and excellent guidance. I am honoured to thank Professor Peter Mathewose, the Director of Family Medicine, University of Pretoria, for his moral support and encouragement. I would like to express my appreciation of my wife, Elizabeth Mekonnen, and my daughter, Darik, for their patience and understanding during my long-distance travels in the course period and late-night writing. I thank Ms F Velosa for professional language editing on the hard copy. My special gratitude goes to my wife for her tremendous support in typing and editing this research. My thanks go to my sister, Fanaye Eyassu, for providing of free access Internet and moral support. iii Abbreviations ABC Abstain, Be faithful , Condomise AIDS Acquired Immune Deficiency Syndrome HIV Human Immune Virus NGO Non Governmental Organisation PSU Primary Sample Unit SADHS South Africa Demographic and Health Survey SPSS Statistical Package for Social Science SSA Statistics South Africa STD Sexual Transmitted Disease iv TABLE OF CONTENTS Declaration ........... . ................ . ... . .................. .. ......... . ........ ......... .. . Abstract .. ..... ........................... .......... ... ...... .... .......... .. ..... .... ........ .. .......................... .. II Acknowledgements .. . ... .. ....... .. ........ ....... ....... ..... . ..... .... . .... . ... .. .... . .... . Ill Abbreviation ....... . . .. . ....... . . ....... .. ........ ... . . .. ... . . .... .... ...................... . IV Table of contents .. . .. . ...... ... .. . ... .... .. . .. ... ........ ... . .. . .. .......... .. ...... .. ...... . V List of table ........ . .. . .. ...... .. ..... .. ....... .. . .................. ... ........ ... .. .. .. ........ . Vl ll List of figures ............ .. ................ . .. ........ ... .. .... ... ............... ......... ...... . CHAPTER 1 Introduction ...... ................... . . .... ............ . . ... .. ...... .. . ... . ...... . . . 1 1. 1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 Aims and Objectives ............... ..... .. ........ ... .. ... .. ... ..... .. ...... .. .............. 2 1.2.1 Specific Objectives .. . ................ .. . ................... . . ... ................ . . . .... ...... 2 1.3 Problems Statement .... .. .. .. .. .. . ........... ... .... ......... ... .... .. .... .... ...... ....... .. ... . 3 1.4 Rationale .. . .................... . ..... . . .... . .. .............. .. .. ...... . ............... ... .. . 1 1.5 Limitation of Study.... .. . ... . .... .... ..... ............ ... . ..... . ... .. ... . .. .. . . . . . . 4 1.6 Organisation of Reports.. . .. . ............................ ...... ...... .. ... .. . ..... ....... . . . . . . .. 5 V CHAPTER 2 Literature review ..... . ................ ..... ....... . .... ..... ....... . .. .. . . . ..... . 6 2.1 HIV/AIDS related information ..... .. . . ......... .... ..... .......... .... ....... . .. ..... . 6 2.2 Mutual fidelity .. . ...... . .. ...... ... ... ...... . . . ........ .... .. .................... . . ..... . 8 2.3 Condom use ...... . ............. .. ....... . .... .. ... . .. ...... ........ ...... .......... . .... . 8 2.4 AIDS campaign in South Africa ..... . ..... .. .. ... ... ..... ......... . ... ... ..... ..... . . . 8 2 .5 Conceptual framework .... .... . .. . . .. . . .. ..... .. .... ....... ......... . . ... ................. . 10 CHAPTER 3 Research Design and Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 3. 1 Data . .. ... . .. .... ... . ....... . . . ................................ . ... .. ...... . .......... . . . .. ... •··· ·· 17 3. 2 Methods of analysis ..... . . . ...... . ........ ... ...... . .................... ... ............ . 18 3.3 Definitions of variables ................... ... .......... . . . . . ............ .. .. . . .. ... .. .... .. .. . ........ 19 CHAPTER 4 Results and Discussion ........ . .. . . . . ... ........ . ............ . .... ... . . .. . ..... . ...... . . 21 4.1 AIDS Knowledge & Awareness. .... .... ..... ....... . ..... ... ... ...... . .. ............ . ... .. .. .... 21 4.2 Source of Knowledge about HIV/AIDS. ...... . ..... . .. ....... . . ................. .. .......... 23 4.3 Perception of Risk ofHIV/AIDS .............. ... .............. .. ... . . . . . . . . . . . . . . . . . . . . ... 26 4.4 Opinion About Reporting HIV/ AIDS Status ............ ........ . ................. . . ... ..... 29 vi 4.5 Age at First Sexual Intercourse........ . .. . ....... .. .. ................ . . .. ......... ........ 32 4.6 Number of Sexual Partners ....... ..... .. . .... .... . ... .. . ..... . ................ ...... ..... ... 32 4.7 Relationship with Last Sexual Partner. ...... .. .... ...... ... .. ... ... ....... ................. 35 4.8 Condom Use.... .. ..... .. ..... ........... ...... . ... ... ......... ...... ... ....... .............. 38 4.9 Model Fitting Information. ......... . .................. . .. . ..... ...... .... . .. .. ...... . .. . ... 41 4.10 Goodness of Fit ............. .. ........ .... .................... . ....... .. ................... 41 4.11 Likelihood Ratio Test. ...... ............. . .. . ..... . .. ....... .... .. . ....... . ........... . 42 4.12 Discussion of Results ...... . ..... ... ... ... .... ... ... .. ........... . ....... .. ............. .. 45 CHAPTER 5 Summary, Conclusions and Recommendations .. ... . ... .. . . ....... ............. 47 5.1 Su_mmary ... ..... . ... ... ... ... ... . .. .. . ... ... ... .. . ... ... ... ... ... ... ... ... ... ... . .. ... ... . .... 47 5.2 Conclusion. ...................................... ... ..... . ...... .... .. ...... . .. .... ..... .. . .. 47 . 5.3 Recommendation ............ .. . ... . .. ......... ....... .......... .... ..... . ............... ... 48 REFERENCES.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 vii LIST OF TABLES Table 4.1: Women by Knowledge of Ways to avoid AIDS/ Urban/Rural. .... . .......... .. 22 Table 4.2: Knowledge & Source of HIV/AIDS Info/Background Char. ........ .... ... .. . 24 Table 4.3 : Perception of Risk of AIDS/Background Char. ... ...... . ... . ................ 27 Table 4.4 : Belief About Mention HIV/AIDS ............... ... ... ... .... .... . ... . .. . ............. 30 Table 4.5: First Sexual Intercourse by Specified Ages ......... .. .... .. ... ..... ............ 32 Table 4.6: Number of Sexual Partners in the last 12 Months/Background Char. . .... 34 Table 4.7: Relationship, Sexual Partner/Marital Status & Background Char. ...... .. . 37 Table 4.8: Condom Use During Last Sexual Intercourse ......... ... ..... . ...... .. ....... 39 Table 4.9:Perception of HIV/AIDS with Reference to Some Variables .......... .. ... 41 Table 4.10 Goodness of fit . . . ... . . . . . ........ .. ... ............... ...... . . . . .... . ... 41 Table 4.11 : Likelihood Ratio Test .. .. .. ... . ....................... ..... ......... . .......... 44 viii LIST OF FIGURES Figure 2.1 :Conceptual frame work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Figure 4.1: Source of knowledge of HIV/ AIDS according to background Char. ......... . . 25 Figure 4.2: Women who Know someone with HIV/AIDS/Background Char........... 28 Figure 4.3: Divulging HIV/A IDS Status to various/Urban/Rural ..... . .......... 31 Figure 4.4 : Condom Use During Last Sexual Intercourse/Background .... . .... 40 ix CHAPTER ONE INTRODUCTION 1.1 Background The universal and pandemic nature of HIV/AIDS is posing huge challenges to our world in the twenty-first century. In the developing countries like sub- Saharan African countries, economical and social developments have been ~hreatened by the devastating effects of HIV/ AIDS . Southern Africa has the highest HIV/AIDS prevalence worldwide (UNAIDS 2002) with estimates of 600 to 1000 AIDS-related deaths daily in South Africa alone. An estimated 20% of the South African population of five to six million people are infected with HIV/AIDS . Recent studies predict that the epidemic could result in a loss of 17% in the Gross Domestic Product growth by 2010. In addition, South Africa has the highest rate of adult and child rape in the world, addi ng to the HIV transmission rate. It is estimated that the majority of the 3 .6 million South Africans who are currently infected with HIV will die in the next ten years unless there are successful interventions to prevent the spread of HIV. About 550 000 people will become infected each year, which will have major implications on all aspects of life in South Africa (Adler and Qulo, 1999) . Although many efforts have been made by the government, Non Governmental Organisation and different organisations to tackle the problem, the most important role to be played in the prevention of HIV/AIDS is by people themselves . Changing risky behaviour and conveying the right knowledge and attitude about HIV/ AIDS are very essential in minimising the spread of the disease. With this in mind, the researcher will try to investigate the following questions in this research : page 1 ► What are the people ' s knowledge, attitudes, and perceptions regarding about I-ITV/AIDS? ► What is the magnitude of risky behaviour for I-ITV/AIDS in South Africa, especially in the North West Province? Study area: North West province of South Africa is borders in the north by Botswana in the south by the provinces of Free State and Northern Cape, and in the north and east by the Limpopo province and Gauteng covering 118, 797sq km. The North West province was created in 1994 by the manager of Bophuthatswana, one of the former Bantustans (black homelands) and the Western part of Transvaal, one of the four former South African provinces. In 2001 the population of the North West province was estimated to be 3.8 million (out of a total of an estimated 44. 8 million people living in South Africa) sixty-five per cent of the people in the North West province live in rural areas. The majority of the Province' s residents are Tswana people who speak Setswana. Smal l groups include Afrikaans, Sotho and Xhosa people. English is spoken primarily as a second language. Most of the population belong to the Christian denomination (Figures according to census 2001 , released in July 2003). 1.2 Aims and objectives The main aim of this study is to examine the perception of the risk of I-ITV/ AIDS i"nfection and sexual behaviour among women in the North West province. To achieve this aim, the following specific objectives were pursued. 1.2.1 Specific objectives ► To establish the relationship between the perception of the risk of I-ITV/ AIDS infection and background socio-demographic factors (urban/rural residence, age and ethnicity) in the North West page 2 ► To examine the relationship between the risk of HIV/AIDS infection and intermediate socio-demographic factors ( education, marital status and age at first sexual intercourse) in the North West ► To establish the relationship between psychosocial knowledge factors (source of AIDS information, knowledge of ways in which to avoid AIDS and socio- demographic factors) ► To analyse the relationship between socio-demographic factors and sexual behaviours ► To examine the relationship between the perception of the risk of HIV/A IDS infection and sexual behaviour 1.3 Problem statement South Africa has one of the fastest growing HIV/AIDS epidemics in the world (UNAIDS 2002), with estimates of 600 to 1000 AIDS-related deaths occurring daily . If this infection trend persists, there is evidence that the economic consequences may be daunting for that region (Arndt and Lewis 2000, Bell, Devorojan and Gersbach 2003). Since the medical management of the effects of HIV/ AIDS infection is improving, there is neither a cure nor a vaccine yet. In this context, behavioural prevention is "today ' s vaccine" . Therefore it is very essential to understand the context of the risk perception in relation to sexual behaviour because it is the first stage towards behavioural change. Therefore, the researcher wi ll analyse the perception risk of HIV/ AIDS infection and sexual behaviour. 1.4 Rationale This research is relevant because it touches a present problem in South Africa, which is HIV/AIDS . The reason for choosing the North West is that most of the areas, particularly black townships and homelands, remained underdeveloped during apartheid. This may affect people's general life style regarding their perception of HIV/AIDS infection and their sexual behaviour. The North West is one of the provinces categorised as an underprivileged area. Therefore, this is a good opportunity to show the level of the risk of the perception of HIV/ AIDS page 3 infection in the area. The other reason is that few studies related to HIV/ AIDS have been conducted in the area. This research focuses on women aged 15-49 because gender issues are increasingly recognised as having a critical influence on the HIV epidemic in South Africa. Violence against women makes women vulnerable to HIV through three main mechanisms (Maria et al. , 2000). Forced sexual intercourse with an infected partner can result directly in HIV transmission. Violence may limit women ' s ability to negotiate safer sexual practice, e.g. women fear asking their partners to wear a condom as this is likely to result in violence (Wess and RaoGupta 1998). Finally, violence is associated with high risk behaviour among women (Wingood and Drelemete, 1997). Experiences of forced sex in childhood and adolescence have been shown in many studies internationally to impact negatively on sexual behaviour in later years (Jawkes et al 2001). Moreover this research could also contribute in giving information for policy makers and programme managers in tackling the HIV/ AIDS problem in the country, especially in the North West Province. 1.5 Limitations of the study Althougb there are many factors that could affect the people ' s perceptions and sexual behaviour, this study is limited in its analysis, e.g. at national level, political will and commitment to policies on HIV/ AIDS prevention and productive health can influence individual access to information and to services such as voluntary counselling and testing or the treatment of sexually transmitted diseases. These may influence people' s perception of HIV risk and sexual behaviour at community level. The ideology of masculinity and femininity may help to shape an individual ' s identity and self-concept and may thus have a strong impact on his or her sexual decision making (Gage, 1998). For instance social norms may define a "good woman" as one ignorant of sex or passive in sexual encounters, whereas a "real man" may be defined as being sexually experienced, possibly with several partners. page 4 The socialisation of young women to acquire spouses by means of sexual encounter and to give priority to male pleasure and control in sexual partnerships can contribute significantly to women's inability to negotiate when, where and ~ow sexual intercourse takes place (Family Health International, 1999). Since this study uses secondary data, it has limitations in providing the above information. The data is inadequate to provide complete information on sexual behaviour and the perception of risk, even though there are many factors that can affect people' s perceptions and sexual behaviour. This study is limited in its analysis. Factors that affect at national, community and individual levels are not studied . This study highlights the association between the perception of risk, socio- demographic and knowledge variables and sexual behaviour. However it does not demonstrate a casual link between risk perception and sexual behaviour, primarily because of the cross- sectional nature of the data. The perception of HIV risk is not static, but varies with context and over time. Since the data was collected in 1998 it may not accurately reflect the present situation of the HIV/AIDS condition in South Africa. 1.6 Organisation of the report This study focuses on women ' s attitudes, beliefs, risk perceptions of HIV/AIDS and their sexual behaviour in the North West together with the background characters. The study also includes women ' s sexual behaviour such as age at first sexual intercourse, number of sexual partners, type of sexual partners and the use of condoms. The rest of the structure of the report is explained in chapter two (the literature review findings) , chapter three (methodology), chapter four (result and discussion) and finally the summary, conclusions and Recommendations are described in chapter five . The next chapter describes the most important findings in the literature review. page 5 CHAPTER TWO LITERATURE REVIEW 2.1 HIV/AIDS related information The epidemiology of HIV/ AIDS is not the same all over the world . Although the spread of the epidemic is largely under control in the developed world, infection rates continue to ri se in countries characterised by poverty, poor health systems, and the lack of education and limited resources. Sub-Saharan African countries, i'nclud ing South Africa, comprise 70% of the proportion of the world population living with HIV/AIDS (AIDS statistics, 2002). A study conducted by Grundling, Dejager, Fourie, Ras and Grundling (2002) indicates clearly that on community and individual levels, pending the development of an effective vacci ne or cure for HIV/AIDS, sexual behavioural change is the only means of averting the continuous spread of the disease. Having the above ideas in mind, the researcher tried to compi le a survey with detailed information from library sources and Internet websites. The most devastating aspects of the HIV/AIDS epidemic compared with the other epidemics is that it usuall y affects people in their most productive years, between the ages of 15 - 49 .It is not noteworthy that the number of South Africans who died because of AIDS before they reached the age of 50 almost doubled since 1990. According to the President of the Medical Research Council ; Professor Malegapuru Magoba, AIDS is devastating South Afri ca' s most economically productive citizens - those aged between 15 and 49 with more men aged 35 and 40 dying than in any other age group in the 1990 - 2000 period . · page 6 The death rate has risen despite better health care and improvements in the quality of life of most South Africans in the ten years since 1990. Life expectancy is therefore projected to fall from 68 .2 years in 1998 to 48 years in 2010 . This has devastating implications both for the economy and the social structure of families and for their quality oflife (Sunday Times, 23 July 2000). The survey results by the Nelson Mandela Foundation (2002) confirmed that there is a high HIV prevalence rate in South Africa, in which 13.3% of all South Africans aged 15 to 49 years in the survey are HIV positive. However, HIV prevalence differs greatly by racial group: 18.4% among blacks, 7.4% among coloureds, 5.8% among whites and 1% among Indians. Black females have the highest infection rate 19% while black males have a slightly lower infection rate of 17%. Among the HIV positive individuals, only 23% report knowing their HIV status. The above figures indicate that HIV/AIDS is a major problem in South Africa. Most evaluation studies have concluded that behavioural interventions are moderately effective in bringing about the desired effects (Kelly 2000); some of which have been prominent in the context of behavioural changes and particularly the extent to which behaviour change is voluntary verses conditional on predisposing factors (UNAIDS 1990, Kelly, Parker and Lewis 2001). Many countries have promoted a three-fold behavioural prevention approach often sloganised as ABC for "Abstain, Be faithful , Condomise" (Healed, 2002). These are the three most widely promoted behavioural prevention responses to HIV/AIDS . Many researchers point out that changing sexual behaviour is very crucial to combating the HIV/ AIDS pandemic. HIV/ AIDS prevention programmes change behavioural norms mainly in three areas, namely abstinence, mutual fidelity and condom use, which are critical to slowing down and reversing the HIV pandemic. Studies in various countries have shown that the above three elements had effective results, e.g. research done in Zambia showed that the page 7 youth aged 15 - 24 who viewed the mass media campaign messages were 1.9 times more likely to practise abstinence than those who did not see the campaign. HIV prevalence now appears to be either stabilising or declining among young Zambian people. 2.2 Mutual Fidelity Mutual Fidelity refers to making stable couples as the first line of defence against HIV infection. In Nigeria the 2004 Niger bus surveys found that proportion of sexually active respondents who had sex with a non-marital partner(s) had decreased from 29% in 1998 to 20% in 2002, a decrease of 31 % .Which contributes in stabilising HIV prevalence among Nigerians. 2.3 Condom use: In Cambodia the HIV prevalence has dropped among the general population and among key high risk groups such as commercial sex workers and their clients, migrant workers, truckers and intravenous drug users while condom use has been on the increase. The policy of 100% condom use in commercial sex establishments is widely credited and PSI is the leading non governmental organization providing condoms to these establishments. 2.4 AIDS campaign in South Africa The high prevalence of HIV/AIDS in South Africa poses major challenges for both governments and civil Soul City groups, who are doing their utmost to curb the spread of the disease and help those affected by it. Farless has been said about the tremendous efforts being made year after year around the country by the local, provincial and national government agencies, as well as a myriad non-government organisations, to combat AIDS by creating awareness around the disease, page 8 promoting behaviour and providing medical , social and economic assistance to those affected and infected by the epidemic. There is no cure for AIDS . The government's strategy focuses on prevention by promoting public awareness and delivering life skills and IDV/ AIDS education. The many AIDS awareness campaigns run by the government and NGO partners such as Love Life and Soul City are new bearing fruits . There is now a high level of awareness among the youth regarding IDV/ AIDS : it is about 90%, but the pressing challenge is to ensure that this awareness translates into behavioural change. Life skills education, which incorporates IDV/AIDS education, is now a compulsory part of the school curriculum and is to be implemented fully by the end of 2003. There are many IDV/AIDS prevention programmes in South Africa, e.g. for Love Life, Soul City and Department ofEducation. The IDV/AIDS page is playing a great role in transmitting mv information. Love Life is a nationwide campaign that aims to promote a healthy sexual behaviour among adolescents, and to reduce the incidence of IDV/ AIDS, sexually transmitted diseases and teenage pregnancies. Love Life uses a widespread media campaign targeting adolescents and it offers educational , recreational and sexual health services in under-resourced areas. Soul City uses the mass media to promote awareness of health issues. Soul City has won international awards for its success in targeting education and entertainment by using popular radio and television dramas. Other programmes, such as treating sexually transmitted diseases, the reproductive health research unit, the South African vaccine initiative preventing mother-to-child transmission dealing with rape cases, treating people with HIV/AIDS, care and support of families affected by IDV/ AIDS, and lastly discrimination against people affected by IDV/ AIDS are part of the activities run by the government. page 9 2.5 Conceptual framework It is essential to understand the context of risk perception in relation to sexual behaviour because it is the first stage towards behavioural change from risk taking to safer sex. Even though health behaviour models acknowledge change of risk behaviour as the basis (Becker & Joseph, 1988; Becker, 1974), it is unclear how people's personal risk assessments relate to their sexual behaviour. Different studies have shown that individuals are more likely to underestimate than to overestimate their risk of IDV infection regardless of the nature of their sexual behaviour (Nzioka, 2001 ; Ingham & Vanzessen, 1997; Becker & Joseph, 1988). People often rationalise risk-taking behaviour using a range of socially constructed criteria that could explain the apparent mismatch between objective risks and perceived risks (Abrams et al , 1990). Risky sexual behaviour can be seen in the context of the number and types of partnerships, sexual acts and orientation (Cohen & Trussell, 1996; Dixon-Muller, 1996). Other parts of risky sexual behaviour include early age at first sexual intercourse, multiple sexual partners, unprotected sexual intercourse with "at risk" sexual partners, and untreated sexually transmitted diseases. The link between the perception of risk and sexual behaviour can be reciprocal. Individuals may perceive their risk of getting AIDS to be higher or lower depending on their previous sexual behaviour or that of their partners. In this case, risky sexual behaviour is the influencing factor on the perception of the risk The conceptual framework identifies two outcome variables : the perception of the risk of IDV infection and sexual behaviour. The conceptual framework shows the possible associations between a range of background factors at national community and individual levels, intermediate socio-demographic and psycho- social factors and the outcome factors. Even though they are not considered in page 10 this analysis, at national level, political wi ll and commitment to policies on HIV/AIDS prevention and reproductive health can influence people ' s access to information and to services such as voluntary counselling and testing or treatment of sexually transmitted diseases. This may also influence individuals ' perception of HIV risk and sexual behaviour. The importance of political will and commitment is demonstrated by Uganda' s success in controlling the spread of HIV (UNAIDS, 1998). Other national po licies and programmes such as those on education, economy, employment, marriage, divorce and gender may also influence individuals ' sexual behaviour and spread of HIV. Such policies can affect the availability of community level health services, the distribution of economic resources, transportation, communication, urbanisation, gender roles and the empowerment of women. The conceptual frame work in Figure 2 .1 briefly explains the link between perception risk of HIV/ AIDS and sexual behaviour. page 11 Figure 2.1 Sexual Behaviour in South Africa Background Intermediate Variables Outcome National level Socio Perception of Programes and . Demographic -----. Risk ~ policies on HIV/AIDS . Factors . Policies on education, Age at first sex employment, gender, Marital status eco nomy Education .~ Work status ,w ,., Community level Psycho social or Sexual Behaviour Socia l, c ultural beliefs knowledge No. of sexual partners and practices factors Type of sexual partners Gender attitudes and . Access to AIDS . Condom use . roles information Exposure to STD Network health a nd Knowledge, education infrastructure ... attitude and Region of residence, beliefs urban/rural residence Preventive Behaviour HIV testing and counsel ling Individual level Age Sex Ethnicity - Relig ion . At community level , ideologies of masculinity and femininity may help to shape an individual ' s identity and self-concept and thus have a strong impact on his/her role in sexual decision-making (Gage, 1998), e.g . social norms may define ' a good woman' as one ignorant of sex or passive in sexual encounters, whereas a page 12 ' real man ' may be defined as being sexually experienced, possibly with several partners. The socialisation of young women to acquiesce to spouses in sexual encounters and to give priority to male pleasure and control in sexual partnerships can contribute significantly to women ' s inability to negotiate when, where and how sexual intercourse takes place (Family Health International, 1999).The region of residence, and whether a person lives in an urban/rural areas, can determine the level of access to information and reproductive health services, which could influence sexual behaviour and the perception of risk. The conceptual framework includes some individual background factors, such as gender, current age, ethnicity and religion ,which may influence sexual behaviour. Evidence from literature suggests that women have fewer sexual partners than men, and women with little control over the sexual activities of their partners are vulnerable to infection by these partners (Family Health International, 1999; Reid, 1999; Inghram and Holmes, 1991). Women are two-four times more likely than men to become infected with STD after intercourse with infected partners due to physio-biological differences and their low socio-economic position (UNAIDS, 2000a., Family Health International, 1999; Cohen and Trussell, 1996). However, due to inadequate information on this data, men are excluded in this study. The age of a person is another factor that may influence sexual behaviour and the level of the perceived risk of HIV infection. Men and women in their teens are at increased risk of HIV infection because they often engage in unprotected sexual intercourse (Hu lton et al ... , 2000) . Sometimes there is pressure for girls to prove ~heir fertility before marriage. Similarly, boys face pressure to prove manhood by impregnating a girl or by having many sexual partners (Meekers and Calves, 1997; zioka, 2001). page 13 Although the risk of IDV infection is high among young men and women, often they do not perceive their risk to be high (Croaks et al., 1990; Cleland, 1995). Ethnicity may influence sexual behaviour through cultural beliefs and practices, e.g. the practice of levirate marriage, where a widow is remarried to one of her husband' s brothers, is still being practised in some areas of sub-Saharan Africa, despite the high prevalence of IDV (Ocholla-Ayayo, 1976, 1997; Degrees du Lou, 1999). Another risky sexual practice is that of ' wife- sharing ', which has been reported in some areas of sub-Saharan Africa, e.g. the Maasai of Kenya (Lesthaeghe, 1989). The pressure to confirm to cultural beliefs and practices may override concerns about IDV infection. Religion can influence sexual behaviour through intermediate factors such as the age at first sex, marital status, and access to information and services. It can also influence attitudes to IDV and the perception of risk. Nzioka (1996) notes in his study that religious people considered AIDS to be a disease that affectes those who transgress against God. Consequently, those who were religious perceived their risk of IDV infection to be low. Religion can also work to influence community practices and national policies. The conceptual framework assumes that the background factors operate through a range of intermediary factors to influence the perception of risk and sexual 0ehaviour. These intermediate factors may be socio-demographic (such as sexual initiation, marital status, the level of education) or psycho-social ( e.g . access to information, knowledge, attitude and beliefs), unless the first intercourse is also the start of a mutually monogamous relationship, early age at first sexual intercourse is associated with a long period of exposure to sexual activity, a higher propensity to accumulate sexual partners, and increased chances of contracting sexually transmitted diseases (Dixon-Muller and Wasserheit, 1990; Konings et al. , 1994). page 14 Marital status influences the perception of the risk of HIV infection and sexual behaviour. Whereas unmarried women may have some ability to negotiate safer sex, married women face extra challenges because of the fear of being suspected of promiscuity by their spouses, which may lead to unwanted consequences such as separation or even divorce. Married women often acquiesce in unsafe sexual practices, even if they suspect or know of their partner' s extra marital relationships (Blanc et al. , 1996). Although HIV cannot be spread through sexual intercourse in stable monogamous relationships between uninfected partners, among married women the presence and the nature of their partners ' casual or extra-marital sexual practices largely determine the risk of HIV transmission (Ahlburg et al. , 1997). The level of formal schooling may influence the perception of HIV risk and sexual behaviour, but the evidence is rather conflicting. Carael (1995) found increased casual sexual activity among those with higher schooling, but Meekers (1994) found that the association disappears when age is controlled for . Exposure to AIDS information through mass media may lead to a high level of awareness, which can in turn influence a self-assessed risk of HIV and behaviour. It has been argued that people ' s assessment of risk may depend on how much they trust the accuracy of information (Stallings, 1999). However, Prohaska et al. (1990) found that neither increased exposure to the media and greater belief in the accuracy of the media as a source of information about AIDS nor knowledge of the facts about HIV/ AIDS transmission affected people ' s perception ofrisk. UNAIDS (2000 b) notes that general awareness of AIDS is no longer important in AIDS prevention but accurate knowledge of how HIV is transmitted is important, e.g . if people believe that mosquitoes transmit the HIV virus, they may see the use of condoms as futile . Some researchers report that asymptomatic page 15 transmission of HIV is not common in the local concept of the disease (Irwin et al. ., 1991) A study of AIDS knowledge in Zimbabwe revealed that whi le all men and women had heard of AIDS 15% of men and 26% of women did not believe that a healthy-looking person could carry the AIDS virus (Central Statistical Office, Zimbabwe and Macro International, 1995). This belief can lead to exposure to HIV infection, since people are unlikely to take precautions when having sexual intercoufse with healthy-looking partners. There are numerous factors that influence sexual behaviour. A Study conducted by Sharma Arwokr, Gupta Anita, Aggraweolop ( 2001) in India showed that among metropolitan cities, women living alone in hostels and having an independent income may have a liberal life style and the chances of practising high risk sexual behaviour are greater. The study further pointed out that among 89% the hostel residents who were assessed for risks and the risk perception of HIV/ AIDS, AIDS awareness was 92.1% . Very few respondents had adequate knowledge about modes of transmission and methods of preventions. Risk of perception was poor. With many diverse of cu ltures, socio-economic and political variation in South Africa may contribute in the variation of risk of perception of HIV/ AIDS and sexual behaviour. Based on the above, the researcher attempted to establish in this research the women' s knowledge, attitudes and perceptions regarding HIV/AIDS in the North West. The next chapter describes the methodology of this research in detail page 16 CHAPTER THREE RESEARCH DESIG AND METHODOLOGY Many statisticians advise that complex survey designs need to be accounted for in data analysis in order to obtain unbiased point estimates and accurate confidence intervals (Smith 1988, Korn and Graubard, 1995, Parsons and Jang, 1997). This chapter explains the types of data and the methods of analysis in detail. 3.1 Data The study is based on secondary analyses of the 1998 South Africa Demographic and Health Survey (SADHS). The Survey collected information on various demographic and health indicators including individual characteristics, marriage and sexual activity, family planning knowledge and the use of HIV/AIDS-related knowledge, attitudes and behaviours. The South Africa Demographic and Health Survey used a two-stage stratified sample consisting firstly of selected primary sample units (PSUs) or clusters and secondly of selected households within PSUs. These PSUs were derived from census enumeration areas defined by central statistics (Statistics South Africa, SSA) for the census conducted in October 1996. All women aged 15-49 years were targeted for interviews in the selected househofds. The analysis in this study is based on women in the North West ~rovince who reported having heard of HIV/AIDS, who were sexually experienced and who responded to the question on the perception of the risk of HIV/AIDS and sexual behaviour. The study focuses only on women aged 15-49; it excluded men because of the lack of HIV/ AIDS-related information about men in the data. Although religion and sexually transmitted disease are important factors to influence sexual page 17 behaviour and the risk of IDV/ AIDS, they are not included in the study due to the lack of information. The data also lacks information on preventive behaviour, such as single partner-ship and fidelity . Otherwise, condom usage, types of partners, number of sexual partners and age at first sexual intercourse were covered in the study. Regarding the risk of the perception of IDV/ AIDS, the data lacks information on how individuals perceive themselves concerning the risk of IDV/AII?S (low, moderate and high) . This study considers how individuals perceive others regarding the risk of IDV/ AIDS by analysing their responses to the simple question: ' Can healthy-looking person have IDV/AIDS?' . This belief or perception is important because it may determine an individual ' s moves to the sexual act or to take precautions when having sexual intercourse with healthy- looking partners. Therefore, the researcher took the perception of IDV/ AIDS as the dependent variable and background characters as the explanatory variables, such as age, age of first sexual intercourse, education, marital status, residence, ethnicity, type of last sexual partner and psycho-social factor (source of information). 'fhe survey has the following strengths: ► It is a nationwide survey based on probability sampling. ► It employed face-to-face interviews. The number of women who were eligible for this study m the North West province was one thousand (1000) . 3.2 Method of analysis Data was analysed using the Statistical Package for Social Science (SPSS) and a computer software process that generates frequency, tables percentage distribution, Graphic presentations. In order to examine the nature and strength of the association between dependent and independent variables, nominal page 18 regression modell ing is used . Because the variab les have more than two components, the binary logistic is not used . The statistical significance convention measure was assessed with model fit, such as the -2log likelihood, goodness fit and the likelihood ratio test. Using the model fitting information perception ofHIV/AIDS (dependent variable) with reference to the explanatory variables (age, age of first sexual intercourse, education, marital status, residence, ethnicity, type of last sexual partner and source of information of HIV/A IDS) were fitted and the level of significance was measured . Using the goodness-of-fit the strength of the relationship between the dependent variable and the independent variables is determined . Using the l_ikelihood ratio test, the explanatory variables that have an association with the dependent variable are selected . The model equation is Y=a +~1X1 + ~2 X2 +- - - - - + ~n Xn where Y= dependent variab le X= explanatory variab les 3.3 Definition of variables ► Dependent variable: the perception of the risk of HIV/AIDS in this study is related to an individual ' s understanding or opinion about whether some one has HIV/AIDS or not. The question on perception of risk was: ' Can a healthy looking man have HIV/AIDS ?'. The answer ' Yes ' related to the right perception, and ' o ' or 'Do not know' was considered to be wrong perception. page 19 ► Explanatory variables : these are variables that directly or indirectly influence the dependent variables, such as the background, intermediate (socio- demographic) and psycho-social factors. In this study, age, age of first sexual intercourse, education, marital status, residence, ethnicity, type of last sexual partner and source of information of HJV/ AIDS are taken as explanatory variables. Survey data on sexual behaviour has been criticised as being potentially inaccurate (Caldwell, Caldwell and Quiggen, 1989). However, Dare and Cleland 1994) argue that similar sentiments were expressed regarding ferti lity and family planning data, and these proved to be overstated. It is also argued that women understate their levels of risky sexual behaviour, while men overstate theirs (Family Health International 1999; Reid, 1999). However, in this study only women are analysed which may minimize distortion of the results. The association between the perception of the risk of HJV / AIDS with other variables may be preserved . Using the above methodology the next chapter describes the major findings and analysis of the research . page 20 CHAPTER FOUR RESULTS AND DISCUSSION 4.1 AIDS knowledge and awareness The study used a range of questions to explore the general knowledge of HIV/A IDS, especially regarding the mode of transmission of the disease and its prevention. It was not the intention of this study to examine the role of the various public awareness messages or community level networks, but rather to understand what is learnt from different sources. It is apparent that the type of sexual information people gather from different sources may have a varying influence on their attitudes and behaviour. The findings in part of Table 4.2 indicate that knowledge of AIDS is almost universal. There is no variation in age, marital status, education or ethnicity. Ninety-eight per cent of women aged 15-49 say they have heard of the disease. The data in Table 4.1 indicates that they are well informed about the way in which HIV is transmitted . More than 85 per cent of women report that it is true that living with one faithful partner, using condoms, using clean needles for injections, and avoiding sharing razor blades arc valid means of protection against the virus. Between 70-78 per cent of women know that it is not true that having a good diet, not using public toilets, avoiding touching people with AIDS, and not sharing food with a person who has AIDS are effective means of protection from contracting HIV/ AIDS .In all, urban women are more knowledgeable about HIV/A IDS than non-urban women. Women ' s misconceptions, about AIDS were negligible. The findings regarding the mode of transmission and prevention of HIV/ AIDS in this study suggest a high level of accurate knowledge. page 21 TABLE4.l PERCENTAGE OF WOME BY OWLEDGE OF WAYS TO AVOID AlDS AN D WITH MISI 'FORMATION ACCORDING TO URBAN/ RURA RES IDE 'CE. ORTH WEST PROVl CE 1998 Type of place of residence Urban Rural Total Avoid AIDS : faithful Not true 10.2 8.7 9.3 to partner Trne 88.2 83.4 85.4 Don' t know 1.4 7.8 5.3 Avoid AIDS by usmg Not trne 5.8 5.7 5.7 condom True 92.3 85.2 88.0 Don ' t know 1. 9 9. 1 6.3 Avoid AIDS: injection Not true 6.9 7.7 7.4 with clean needle True 90.8 82.2 85 .6 Don ' t know 2.2 IO.I 7.0 Avoid AIDS : avoiding Not true 85 .5 78.5 78.2 touching person True 12.3 14.5 13.6 Don' t know 2.2 12.0 8.2 Avoid AIDS: Avoid Not true 59.1 53.8 55.9 mosquito bites Trne 29.8 24 .8 28.8 Don ' t know 11.0 2 I .4 17.3 Avoid AIDS: having good Not true 74.5 67.6 70.3 diet Trne 20 .2 IS. I 17.1 Don' t know 5.3 17.3 12.6 Avoid AIDS: avoid Not true 75.1 67.4 70.4 public to ilets True 20.8 19.2 19.8 Don ' t know 4.2 13 .3 9.8 Avoid AIDS : Avoid share Not true 83.1 72 .1 76.4 food person True 13.9 14.6 14.3 Don ' t know 3.0 13.3 9.3 A voiding AIDS; avoiding Not true 11.3 7.2 8.8 sharing razor blade True 85 .9 80.9 82.8 Don ' t know 2.8 12.0 8.4 page 22 4.2 Source of knowledge about HIV/AIDS There is evidence from the findings in Table 4.2 of the penetrative influence of the mass media in the North West. Women are exposed to the formal and informal sources of HIV/ AIDS information. According to Table 4 .2, the most common source of information on HIV/ AIDS among all women in the survey is television. More than 83 per cent of the women reported that the televis ion was their source of information on HIV/ AIDS . The other major sources of information were radio (more than 82%) and health workers (80%). The television is more informative because of the audiovisual effects li kely to have huge impression. There is not much variation in the source of information by background variables, but it is important to note that for any given source of information on HIV/AIDS, access increases as the level of women' s education increases. Black women have the highest level of access to HIV/AIDS messages on television (89%), from health workers ( 86%), from the radio (85%), from newspapers (72%), and from pamphlets (68%). The magnitude of the sources of knowledge by background characteristics is presented below in Figure 4 .·1. . page 23 TA BLE 4.2 PERCENTAGE OF RESPO DE 'TS BY KNOWLEDG E OF AIDS D SOURCE OF KNOWLEDGE, ACCORDING TO BACKGRO U I CHARACTERISTI CS, NORTH WEST 1998 KNOWLEDGE OF AIDS SOURCE OF KNOWLEDGE Age 5 yrs group Knowledge of AIDS TV Radio Newspapers Pamphlets Hea!U1 worker . 15 - 19 98.2 84.9 84.9 73.9 62.8 84.2 20-24 JOO 89.8 83 .6 77.3 72.9 87 25 -29 99.4 85 86.3 72.3 72.5 83 .6 30 - 34 99.3 92.3 83.8 74 .6 68. 1 84 .5 35 - 39 JOO 83.9 85 .6 66 .1 66.9 81.4 40 - 44 98 88.5 82.3 69.8 65.3 80.2 45 -49 100 92.3 95 .3 70.8 67.7 79.7 RE IDENCE Urban 99.7 88.4 9 I .4 78.9 70.4 84.2 Rural 98.9 87.4 81.3 68.8 67. 1 83 .2 Current Marital Status Never Married 99.4 86.9 84.1 75.2 69.8 84.9 Married 99 89. 1 89.4 72.3 70 84.2 Living Together JOO 87.5 76.4 56.9 50 81.9 Widowed JOO 82.4 94 .1 70.6 58.8 82.4 Divorced 100 95.7 91.3 87 91.3 82.6 Not living together 96.7 86.2 75.9 65 .5 62.3 72 Education Level No education 95.9 81.7 7 1. 8 25.4 24 .3 71.8 Primary 98.8 86.3 81.3 6 1. 3 56.3 79.5 Secondary 99.8 88.6 87.5 8 1. 5 77.2 87.5 Higher 100 92 94 .7 92 85.3 82.7 Ethnicity Black 99.2 89.7 85.2 72.3 68.6 86.3 Colom ed 100 65.5 86.2 65 .5 55.2 58.6 White 100 68.4 86.8 8 1.6 68.4 50 Asian 100 62.5 100 100 87.5 50 page 24 FIG RE 4.1 PERCENTAGE OF RESPO DENTS BY SOURCE OF KNOWLEDGE OF HIV/AIDS ACCORDING T( BACKGROU D CHARACTERISTICS, ORTH WEST, 1998 100 ,Q "C 11:J a.:.:. 0 80 i:: Cl) Cl "C Cl) 60 j ~0 0 C: .~... 0 40 II) Cl) I:! 0 er, 20 Background Characteristics page 25 4.3 Perception of risk of HIV/A IDS Women were asked whether a person infected with the AIDS virus always showed symptoms or looked perfectly healthy. They were also asked whether they had a personal knowledge of someone who had been diagnosed with HIV/AIDS or who had died of AIDS . The result in Table 4.3 shows that about 30% of women say: ' yes ', a healthy person can have AIDS . About 35% said : ' no ', a healthy person cannot have AIDS . On the other hand, only 12% reported that they knew someone with AIDS or who had died of AIDS . The low percentage findings about knowing someone with HIV/AIDS is not related to prevalence of the disease, many literature sources showed that HIV/ AIDS is common in South Africa. Therefore, the above result might be interpreted that when people are sick or have died of HIV/ AIDS, the victim ' s disease might not be divulged to avoid certain consequences, such as social stigma, economic problems, etc. Regarding the perception of a healthy-looking person having HIV/AIDS, the findings suggest that women in the North West are less likely to take precautions when having sexual intercourse with a healthy-looking person. Rural women are more likely to report that a healthy person can have AIDS (34%) whereas 24% of urban w9men repo1t this. Black women are the highest to repo1t yes compared to other women. The highest percentage of those who reported that they did not know or did not respond to the question was found among women with no education (41 %). Surprisingly, 71 .6% of educated women reported that a healthy person could not have HIV/AIDS . Table 4.3 shows that those with higher education (26%) are the highest percentage of women who know someone with AIDS or that someone died of the disease. More than 20% of the women who were formerly in a union reported that they knew someone with AIDS or who had died of the disease. The rest of the data is pre ented below in Table 4.3 and Figure 4.2. page 26 TABLE4.3 PERCE TAGE OF WOME WHO KNOW ABOUT AIDS BY PERCEPTION OF RISK OF AIDS, ACCORDING T( BACKGROUND CHARACTERISTICS, NORTH WEST, 1998 Can a healthy person have AIDS? Know someone wit11 HIV/AIDS Age No Yes Don ' t know Total nwnber No Yes Tota.I 15 - 19 53 .7 28.7 17.7 164 92.0 8 162 20-24 56.8 29 14.2 176 86.5 13.5 171 25 -29 56.9 30 13 . l 160 91 9 155 30-34 55 .3 28.4 16.3 141 82.6 17.4 138 35 - 39 50.4 32.5 17. l 117 87.2 12. 8 117 40 - 44 51.6 32.3 16. l 93 86 .2 13. 8 94 45 - 49 41. 9 40.3 17.7 62 87.5 12.5 64 Residence Urban 64. l 24.9 10.9 357 87.l 4.9 357 Rural 46.9 34.2 18.9 556 82.2 11.8 544 Education No education 30.0 28.6 41.4 70 92.8 7.2 69 Primary 41.8 37.7 20.5 239 93. l 6.9 233 Secondary 59.6 28 .5 11.9 530 86.7 13.3 526 Higher 71.6 24. 3 4.1 74 74.0 26.0 73 E thnicity Black 51.3 32.3 16.4 833 88 12 822 Coloured 75.9 13.8 10.3 29 86. 2 13.8 29 White 92. l 5.3 2.6 38 84.2 15 .8 38 Asian 62.5 25 12.5 8 100 0 8 Current marital tatus ever married 52.8 30.5 16.7 492 88 .8 11.2 482 Married 58 31.7 10. 3 281 85.4 14.6 28 1 Living 52 .8 15.3 31.9 72 95 .7 4.3 69 togetl1er Widowed 35.3 52.9 11.8 17 70.6 29.4 17 Divorce 59. l 40.9 0 22 78.3 21.7 23 ot living 31.5 37.9 27.6 29 93.1 69 29 togetl1er page 27 FIGURE 4.2 PERCENTAGE OF WOME WHO KNOWS SOMEONE WITH HIV/AIDS ACCORDING T( BACKGROU D CHARACTERISTICS, ORTHWEST 1998 120 en C <( > 100 :f .-c: ~ (1) 80 C: 0 (1) E e0n 60 ~ ~ 0 ~ C: x:: 0 .c: 40 ~ C: (1) E .~... 20 0 ';I!. page 28 4.4 Opinion about reporting HIV/AIDS status Women who had heard of AIDS were asked for their opinion about reporting IDV/ AIDS status. They were asked whether they believed that people with IDV/ AIDS should be told of their status, whether HIV/A IDS patients should tell their pa~ners and whether diagnosed cases of IDV/ AIDS should be reported to health authorities . Over 88% of the women believed that IDV/AIDS status should be reported to the patients themselves and their partners, whi le 75% believed that IDV and AIDS should be reported to the health authorities (Table 4.4) . Less than 5% of the women reported that they did not know whether the IIlV/ AIDS status should be reported to health authorities. Place of residence plays an important role in whether women would say yes to report their IDV and AIDS status or not. Table 4.4 shows that women in urban areas are more likely to say yes than their counterparts in the rural areas. Further r:nore, rural women (7 to 10 per cent) compare to urban women are more likely to say that they do not know whether diagnosed cases of IDV and AIDS should be reported to health authorities or not. The rest of the information about reporting IDV/AIDS status is described below in Table 4.4 and Figure 4.3. page 29 TABLE4.4 Percent distribution of women who have ever heard of AIDS regarding whether they believe certain solid comments abou divulging HIV/AIDS status to various 1>eople or not, according to urban/ruraJ residence, 01ih West, 1998. Residence People witl1 AIDS be Urban Rural Total Number told status No 7. 3 5.9 6.4 59 Yes 89 .9 87.0 88.1 807 Don ' t know 2.8 7.1 5.5 50 People witl1 HIV be told of tl1eir status 0 7.0 6.3 6.6 60 Yes 90.2 86.6 88.0 806 Don' t know 2.8 7.1 5.5 50 HIV/AIDS patient tell partners 0 9.0 5.0 6.6 6.0 Yes 88 .2 86.4 87. 1 798 Don' t know 2.8 8.6 6.3 58 Diagnosis cases of AIDS be reported to health authorities 0 19.7 13.9 16.2 148 Yes 74.6 75 .2 75.0 686 Don ' t know 5.6 10.9 8.9 81 Diagnosis onset of HIV be reported to heal th authorities 0 1_8.3 14. l 15.7 144 Yes 75.8 75 .0 75 .3 689 Don tic.now 5.9 10.9 9.0 82 page 30 FIGURE 4.3 PERCENTAGE OF WOMEN DIVULGING HIV/AIDS STATUS TO VARIOUS GROUPS OF PEOPLJ ACCORDING TO URBAN/ RURAL RESIDENCE, NORTH WEST, 1998 100 90 -II) ::, 80 ,!S en en 70 C c( > 60 :c Cl D Urban C 50 _2l isl Rural ::, > 0 40 Ql -Cl I'll 30 C Ql ~ Ql a. 20 10 0 z0 3:: z0 3:: 0 0 z0 3:: 0 3:: 0 3:: 0 z 0 Z 0 C C C C C ..i.::: .-.i.::: .-.i.::: ..i.::: ..i.::: ;!--C -c -c -c -C 0 0 0 0 0 0 0 0 0 0 Women's Response page 31 4.5 Age at first sexual intercourse Table 4.5 shows women in the reproductive age group who responded their first sexual intercourse by exact age. Most of teenagers reported that they had had their first sexual intercourse by the age of 15 (9. 5%) whereas the older women between fro m 20 to 49 had their first sexual intercourse by the age of 18 (17%-20% ). The majority had had their first sexual intercourse by the age of 18 (16 %). The rest of the information is presented in Table 4.5 below. TABLE4.5 PERCE TAGE OF NEVER HA YING SEXUAL INTERCOURSE BY EXACT SPECIFIED AGES ORTH WEST, 1998 Current 15 18 20 22 25 age 15 - 19 9.5 4.8 0.6 - - 20 - 24 7.5 20.2 6.9 0.6 - 25 -29 10. l 19.0 9.5 3.2 0.6 30 -34 9.8 15.4 15.4 2.1 0.7 35 - 39 · 3.5 18.3 16.5 0.9 0.9 40-44 4.2 19.8 16.7 3.1 - 45 - 49 12.5 17.2 14. l 6.3 - 4.6 Number of sexual partners Table 4.6 shows the highest percentage of currently married women (96%) who reportedly had one sexual partner. Less than 3% of currently married women reported that they had no sexual partner. Less than 1% of currently married women reported having two or more sexual partners. The highest percentage of sexual partnership with two or more people (5%) was reported among currently married women aged 25-29. More than 87% of never married and currently non- 1)1arried women reported that they had one sexual pariner. Only 6 per cent of these women reported having no sexual partner. More than 4% reported two or more page 32 sexual partners. Only about less than 5% of black women reported that they had no sexual partner. More black women reported sexual partnerships with two or i:nore people than other ethnic groups. Place of residence did not make a major difference. page 33 TABLE4.6 PERCE TAGE OF NUMBER OF PART ERS WHO HAD SEXUAL INTERCOURSE IN THE LAST 12 MO TH B' BACKGROU D CHARACTERISTICS, ORTH WEST, 1998 NUMBER OF PARTNERS IN THE LAST 12 MONTHS Age 0 l 2 3 4 Total 15 - 19 3.6 92.7 3.6 0 0 55 20-24 6.1 89.8 2.0 2.0 0 147 25 -29 4.3 87.9 5.0 2.1 7 141 30-34 3.1 92. l 3. 1 0.8 0.8 127 35 - 39 4.7 92.5 1.9 0.9 0 106 40-44 2.4 93 .9 2.4 1.2 0 82 45 -49 4.1 93 .9 0 2.0 0 49 Residence Urban 3.8 92.l 3 1.1 0 266 Rural 4.5 90 .7 2.7 1. 6 0.5 44 1 Educational level o education 5 95 .0 0 0 0 60 Primary 4.2 90. l 2. 1 2.6 192 econdary 4.4 90.4 3.9 1.3 0 385 Higher 2.9 95 .7 1.4 0 0 70 Ethnicity Black 4 .6 90.7 3. 1 1.4 3 654 Colored 0 94.1 0 5.9 0 17 White 0 100 0 0 0 29 Asian 0 100 0 0 0 3 Current Marital tatus ever Manied 5.8 87.4 4.2 2.3 0.3 309 Manied 2.8 96 . l 0.7 0.4 0 282 Living together 2.8 90.3 5.6 0 1. 4 72 Widowed 0 100 0 0 0 8 Divorced ·o 93 .8 0 6.3 0 16 ot Ii ing 10.0 80 .0 5.0 5.0 0 20 together page 34 4. 7 Relationship with last sexual partner Table 4.7 shows that the person with whom most women had their last sexual intercourse was either their husband or regular partner. Most of the currently married women (above 52 per cent) had had their last sexual encounter with their husbands. The highest level of women who had sexual intercourse with spouses had a positive relationship as age increased, e.g. (the younger aged 20-24 52 per cent) and older aged ( 45-49 95 per cent) had had their last sexual intercourse with their husbands. This might indicate that marital unfaithfulness is less likely in older women. The younger married women were more likely to have their last sexual intercourse with a regular partner (48 .1%) than the older women, which was 5.3%.There was no remarkable variation to have sexual intercourse with their spouses or regular partners by residence. A total of 97.5% of married educated women had had their last sexual intercourse with their husband and only 2.5% with their regular partner; whereas 86. 7% of uneducated married women had had their last sexual intercourse with their husband and 13 .3% with their regular partners. Regarding ethnicity, the small numbers·in the cells for coloured, white, and Asian hamper a meaningful analysis of racial differences. With respect to unmarried women, more than 67 per cent of currently non- married women had had their last sexual intercourse with a regu lar partner. The older currently none married women aged between 40-44 (33 .3%) had had higher levels of sexual intercourse with a casual partner than the younger women 15-19 (1.9%) and between the 20-24 age groups (5 .1% ) . The high vulnerability of older unmarried people to casual partners might be the result of many factors, such as economic problems, social insecurity or sexual pleasure, etc. page 35 Regarding other background factors such as residence and education, unmarried urban women (11.7%) were more likely to have their last sexual ·ntercourse with their casual partners than their counterpart rnral women, which was 2.9% . This shows that the environment may influence the women ' s sexual behaviour. Women in urban areas might have different expDsures, temptations and pressure to have sexual intercourse with casual partners. The higher educated unmarried women (3 .8%) were less likely to have had their last sexual intercourse with a casual partner than the uneducated unmarried women (9.1% ), whereas a higher p rcentage is found in higher educated unmarried women (96 .7%),who had had their last sexual intercourse with regular partners than the uneducated counterparts (63 .7%). page 36 TABLE4.7 RELATIONSHIP WITH THE PERSON WHOM HAD SEXUAL INTERCOURSE. PERCENTAGE DISTRIBUTION OJ WOMEN WHO EVER HAD SEXUAL INTERCOURSE BY RELATIONSHIP WITH LAST PERSON WITH WHOM SHJ HAD SEXUAL INTERCOURSE, ACCORDING TO MARITAL STATUS AND BACKGROUND CHARACTERSTIC~ NORTH WES,T 1998 Currently married women Not currently married women Age Marital Regular Otl1er Total Regular Casual Otl1er Total Partner Partner 15- 19 92 .5 1.9 5.7 53 20-24 51. 9 48. l 0 27 87 .2 5. 1 7.7 117 25 -29 81.3 18.8 0 64 91.5 2.8 5.6 71 30 - 34 79.5 20.5 0 78 82. l 12.8 5.2 39 35 - 39 84.0 16.0 0 81 93.8 6.3 0 16 40-44 93.5 6.5 0 62 66.7 33 .3 0 9 45 -45 94.7 5.3 0 38 100 0 0 2 Residence Urban 79.6 20.4 0 142 82.5 11 .7 5.9 103 Rural 85 . l 14.9 0 208 91.2 2.9 6 204 Education o education 86.7 13.3 45 63.7 9.1 27 .3 11 Primary 72.9 27.l 107 87 .3 6.3 6.4 63 Secondary 84.8 15.2 158 88.9 5.8 5.3 207 Higher 97.5 2.5 40 96.2 3.8 0 26 Ethnicity Black 81.4 18.6 306 88. l 6 6 302 Colored 76.9 23. l 13 100 0 0 3 White 100 0 26 Asian 100 0 2 100 0 0 l page 37 4.8 Condom use Women were asked whether a condom was used during their last sexual i_ntercourse. This question was asked to determine the use of condoms for prevention purposes regarding the disease. As shown in table 4.8, the women who had sex with their spouse, less than 19 per cent reported condom use during their last sexual intercourse. A higher figure (24 per cent) was observed among the women who had had their last sexual intercourse with unmarried partners. Irrespective of the partners, condom use decreases with age. The younger women reported higher condom use during their last sexual intercourse than others. Among unmarried women, condom use is higher in the urban areas (23 per cent) than the rural areas (12 per cent) regardless of the type of partners. The above findings indicated that those who live in urban areas may have better access to condoms . Exposure to modernisation, mass media and technology might also help the urban women to have a better understanding of using condoms than older women. The less use of condoms in married women, as in many literature sources shows that most women did not use condoms because they loved their partners or did not want to annoy them as condoms can imply promiscuity and mistrust, and reduced sexual pleasure. As presented below m Table 4.8 and Figure 4.4, for all types of partners, the J.ikelihood of condom use during the last sexual intercourse is positively associated with an increased level of education. Due to the small numbers in the cells for coloured, white and Asian it is difficult to undertake a meaningful analysis of racial differences . page 38 TABLE4.8 USE CONDOM PERCENTAGE OF WOMEN WHO USED CONDOMS DURING LAST SEXUAL INTERCOURSE BY TYPE Ol PARTNER ACCORDING TO BACKGROUND CHARACTERISTICS, NORTH WEST, 1998 Age Last sex with Number Last sex with Ntunber Last sex with Number spouse unmarried casual partner acquaintance 15 - 19 0 l 24 50 0 l 20-24 18. 8 16 20 115 28.6 7 25 -29 14.5 55 13.6 81 0 2 30-34 7.7 65 9.3 54 0 5 35 -39 13.2 68 16.7 36 0 2 40-44 1.7 58 5.6 18 50 4 45 - 49 0 37 10 10 Residence Urban 5.9 118 23.4 128 23 .1 13 Rural 10.4 182 12.3 236 12.5 8 Education No education 2.6 39 12.5 16 0 l Primary 9.8 82 7.2 97 14.3 7 Secondary 10.9 138 19.3 223 25 12 Higher 4.9 41 25 28 0 l Ethnicity Black 10. l 257 15.8 354 19 21 Colored 0 11 33.3 6 White 0 27 0 2 Asian 0 2 100 l Current marital sta tus Married 9.1 276 66 .7 3 l Living together 7.1 14 5.3 67 Divorced 0 l 13.3 15 Not living 0 4 21.4 14 50 2 together page 39 FIGURE 4. 4 PERCENTAGE OF RESPONDENTS WHO USED CONDOMS DURING LAST SEXUAL INTERCOURS BY TYPE OF PARTNER ACCORDING TO BACKGROUND CHARACTERISTICS, NORTH WEST, 1998 120 C1) I..l.l :::, 100 0 I:? ..C.1.) -= iij 80 :::, X C1) C-l) Ill ra □ Last sex with spouse 60 ..J liill Last sex with unmarried partner -C1) .s:: .5 C1) 40 Ill ::::, E 0 ,:, C: 20 u0 Background Characteristics page 40 4.9 Model fitting information According to Table 4.9 the perception of HIV/AIDS with reference to explanatory variables : the -2 Log likelihood equals 928.382 and the chi-square equals 316.956 with 108 degrees of freedom the model is highly significant, meaning that the model fits the data best (this model provides the best fit) . Table 4.9 Perceptions of HIVI AIDS with reference to some variables Model Fitting Information Model -2 Log Chi- df Sig . Likelihood Square Intercept 1245.339 Only Final 928.382 316.957 108 0.000 4.10 Goodness of fit Table 4.10 shows that the Pearson chi-square is 1291.656 with 1134 degrees of freedom and highly significant at 0.001 level of significance showing a very strong relationship between the dependent variable and the independent variables (they are highly correlated/there is an association). Table 4.10 Goodness of fit Goodness-of-fit Chi- df Sig. Square Pearson 1291.656 11 34 0.001 Deviance 907.352 1134 1.000 page 41 4.11 Likelihood Ratio test Using the model fitting information perception of HIV/AIDS (dependent variable) with reference to the explanatory variables (age, age at first sexual intercourse, education, marital status, residence, ethnicity, type of last sexual partner, source of HIV/ AIDS information and know someone with HIV/ AIDS) were fitted at level of 0.01 significance (P= 0.01). Accordi~g to Table 4.11 , the model has selected only eight variables that are associated with the dependent variable. It selected (place of residence, education, ethnicity, marital status, AIDS information from News paper, AIDS information from friends, AIDS information from partners and know someone with HIV/AIDS) . Place of residence was selected by the model with -2 log likelihood equals 940 .284 and chi-square equals 11.902 with level of 0.00 significance difference. The finding suggests that there is a strong association between place of residence and perception of HIV/AIDS . Ethnicity was selected with -2log likelihood equals 946.602 and chi-square equals ~8 .22 with 0.01 level of significance. This indicates that there is a highly significant correlation between ethnicity and perception of HIV/AIDS . The finding confirms that customs, tradition and culture may contribute in affecting individual ' s perception ofHIV/AIDS. Marital status was selected by the model with -2 log likelihood of 953 .123 and chi- square of 24.74 with level of 0.01 significance. This shows that there is a significant association between marital status and perception of HIV/A IDS . page 42 Education was also selected by the model with -2 log likelihood equals 949.816 and chi-square equals 21.434 with 0.00 level of significance. This indicates that there is~ strong association between education and perception ofHIV/AIDS . Know someone with HIV/ AIDS was selected by the model with -2 log likelihood of 939.719 and chi-square of 11.336 with 0.00 level of significance. The findings confirm .that there is a significant relationship between know someone with HIV/AIDS and perception of HIV/AIDS. Regarding the source of HIV/A IDS information, source of information from News paper, friends and partners were selected by the model. This shows that there is a significant association between perception of HIV/ AIDS and the above source of HIV/ AIDS information. The model is _Y = 928 .382 +940.284 X1 +949.816 X2 +946 .602 X3+953 .123 Xii +941.127 X5 +949.475 X6 +944.479 X1+939 .719 X8 . where Y represents perception of HIV/AIDS X1 to Xs represent the predictor variables page 43 Table 4.11 Likely ratio test Likelihood Ratio Tests Effect -2 Log Likelihood of Reduced Chi- df Sig. Model Square Intercept 928.382(a) 0 0 Place of 940 .284(b) 11. 902 2 0.00 residence Education 949 .816(b) 21.434 6 0.00 Ethnicity 946.602(b) 18.22 6 0.0 1 Age at first 981.996(b) 53 .613 34 0.02 intercourse Relationship 941. 917(b) 13.535 8 0 .10 to last sex partner Marital 953. 123(b) 24.74 10 0.01 status AIDS inf. 936.087(b) 7.705 4 0.10 TV AIDS inf. 93 5.972(b) 7 .59 4 0.11 Radio AIDS inf. 941.127(b) 12.745 4 0.0 11 News paper AIDS 933.907(b) 5.525 4 0.24 inf.p aph.lets AIDS inf. 934.197(b) 5.815 4 0.21 Health worker AIDS inf. 949.475(b) 21.093 4 0 .00 Friend AIDS 944.479(b) 16.097 4 0.00 inf. partner Age 937.763 (b) 9.38 12 0 .67 Knows some 939.719(b) 11.336 2 0.00 one with HIV/AIBS page 44 4.12 Discussion of results As ment.ioned earlier in the results section, knowledge of HIV/ AIDS is almost universal and that most of the people are well informed about HIV/A IDS transmission. This might be due to the existence of effective HIV/AIDS campaigns in the province as well as better access to formal sources of information, such as television, Radio, health workers etc. Unlike HIV/AIDS awareness, the majority of the people over 53% have inadequate perception about HIV/AID. This might be due to the influence of informal source of information as indicated in the above model fit formula, such as AIDS information from friend s and partners. Theoretically, the expected effect of attending the funeral of someone one knows to have died of AIDS-related diseases on sexual behaviour, condom use or perception is not clear. When we observe a respondent has attended such a funeral , we can assume the respondent is somehow related (friend, famil y, neighbour or colleague) to the deceased. Moreover, we can assume that the respondent has some knowledge of the lifestyle of the deceased. He can use additional information that the person became infected with HIV/ AIDS because of his sexual behaviour. Therefore that may contribute on the respondent sexual behaviour and perception ofrisk ofHIV/AIDS . Education is identified as one of the determinant factors of individual perception of HIV/ AIDS and sexual behaviour. Moreover, we can identify two mutually i:einforcing effects of education. First, those who are better educated are more likely to be aware of the risk of unsafe sex and the perception of HIV/AIDS. Second, it is well documented that individuals with more education earn higher wages and incentives (Mincer 1974). Hence, ceteris Paribus, better educated individuals have more to lose from choosing unsafe sex. We should therefore expect better educated people to use condoms more often than those who are not page 45 educated. The findings in the above results show that more educated people have a poor perception about a healthy-looking HIV/ AIDS person. This shows that education does not seem to matter for a woman' s decision to be sexually active. Marriage Having sex without a condom cannot always be equated with risky behaviour. We should expect married couples to have sex more often and decrease the likelihood of condom use. This finding confirms that married couples use condoms less frequently than unmarried ones. Age Younger women are more likely to use a condom than older women. This finding reflects that the older the women, the lower the condom use. This could be explained by older women, who are less sexually active and more likely to be married and to be in permanent relationships. They are also more likely to have lower rates of partner turnover. Sexual activities The findings are in line with the commonly held assumption that early sex age and condom use indicate a high risk exposure factor. It therefore appears that delaying the sexual debut may prompt people to adopt safe sex behaviour. A Surprising finding on the above result showed that older women ( 45-49) have a high rate of earlier sexual intercourse than younger women (15 -19) . This might not be related to the risk of sexual behaviour but rather explained by other factors such as marriage at an early age in the older generation and the influences of culture and religion. Finally, the summary, conclusions and recommendations ~re presented in the next chapter. · page 46 CHAPTER FIVE SUMMARY, CONCLUSIONS AND RECOMMENDATIONS 5.1 Summary This research examined the knowledge, perception of risk of HIV/AIDS and sexual behaviour by using data from the 1998 South Africa Demographic and Heafth Survey. Nominal regression models were fitted to examine the strength of the association between the perception of the risk of HIV/ AIDS and explanatory variab_les . The findings indicate that knowledge is universal and women in North West are well informed about mode of transmission of HIV/A IDS . Married and older w0men are less likely to use condoms. More than 88% of the women believed that HIV/ AIDS status should be reported to the patient themselves and partners. Women ' s perception was examined about whether healthy people could have HIV/ AIDS and about 30% of women said "yes" and about 53% said "no". The rest of the findings are described in chapter four. 5.2 Conclusion In the North West, knowledge and awareness of HIV/AIDS are very high, but most of the women (over 53%) perceived that a healthy person could not have HIV/AIDS . On the other hand, young women and those who are not in a marital union reported high rates of risky sexual behaviour, including multiple partners and casual unprotected sexual intercourse. The author acknowledges that much remains to be done to verify these results, especially in finding and identifying variables of the perception of risk and sexual activity. page 47 5.3 Recommendations Changing the perception of the risk of IDV/ AIDS and sexual behaviour is not an easy task. It needs a collaboration effort between the individual, the community and the government. As indicated in the study, the general knowledge of IDV/ AIDS is satisfactory, but the people ' s perceptions and sexual behaviour still need to be modified. This could be done as follows : ► Providing adequate and correct information about IDV/AIDS. ► Advocating preventive behaviour, such as promotion of condom use ► Strengthening the current mass campaign oflilV/AIDS ► Supporting and educating regarding women for their rights and responsibilities in making decisions on their sexuality ► Supporting and educating women on: -how to take care of themselves; -how to practise protected sexual intercourse; -their role of taking decisions on sexual intercourse; and -how to prevent and protect IDV/AIDS, etc. ► Supporting the women who practice unsafe sex or having multiple partners for economi c resources. page 48 REFERENCES Abrams, D ., Abraham, C. , Spears, R . & Marks, D . (1990) AIDS invulnerability : relationships, sexual behaviour and attitude among 16 - 19 year olds. In: AIDS : Individual, cultural and Policy Dimensions, pp. 35-52. Edited by P . Aggleton, P. Davies & G . Hart. Falmer Press, London. Adler, G and Qulo, 0 (1999) 'HIV, AIDS and STDs' , South African Health Review, Durban: Health Systems Trust. Aggleton, P., Slutkin, G. & Davies, P. (1994) Risking everything? Risk behaviour, behaviour change and AIDS . Science 265, 341-345 . Ahlburg, D .A. , Jensen, E.R. & Perez, A.E. (1997) Determinants of extramarital sex in the Philippines. Health Transformation Review. 7 (supplement), 467-479. Ajayi. A. , Marangu, L. , Miller, J. (1991). Adolescent sexuality and fertility in Kenya: a survey of knowledge, perceptions, and practices. Study of Family. Planning 22, 205-216. Akwara. · P. A. Madise NJ. Perception of Risk of HIV/AIDS and Sexual behaviour in Kenya. Journal Bio social science (2003), 35, 385-411. Arndt, C. , and J.D. Lewis (2000)'The macro implications of HIV/AIDS in South Africa; A preliminary assessment.. , South African Journal of Economics 68(5), 856-87. Barnett, Tony, and Alan Whiteside (2002) AIDS in the Twenty-First Century. Disease and Globalisation (New Work, N.Y. ; Palgrave Macmillan) Becker, M . H . (1974) The health belief model and sick role behaviour Health Educatio;i Mono 2, 4 . page 49 Becker, M .H . & Joseph, J. G. (1988)" AIDS and behavioural change to reduce risk: A" review. American Journal of Public Health 78, 394-410. Blanc, A. , Wolff, B ., Gage, A., Ezeh, A., Neema, S. & Ssekamatte-Ssebuliba, J. (1996) Negotiating Reproductive Outcomes in Uganda. Macro International, Inc., Calverton, MD. Bob Mash (2000). Handbook of Family Medicine. PP 174-5. Bongaarts, J. (1995) Global Trends in AIDS Mortality . Research Division Working Paper 80, Population Council, New York. Caldwell, J. C. , Caldwell, P . & Quiggin, P. (1989). The social context of AIDS in sub-Saharan Africa. Population Development Review 15, 185-234. Caldwell. , J. C., Orubuloye, I.O. & Caldwell. , P. (1999) Obstacles to behavioural change to lessen the risk of HIV infection in the African AIDS epidemic: Nigerian research. Health Transformation Review 9, 113-124. Camel., M . (1995) Sexual behaviour. In : Sexual Behaviour. In : Sexual Behaviour and AIDS in the Developing World, pp. 75-123 , Edited by J. Cleland & B. Ferry. Taylor and Francis, London. Central Statistical Office (Zimbabwe) & Macro International (1995) Demographic and Health Survey 1994. Central Statistical Office, Harare, and Macro International, Columbia. MD. Cleland, J. (1995) Risk perception and behavioural change. In : Sexual Behaviour and AIDS in the Developing World, pp. 157-192. Edited by J. Cleland & B. Ferry. Taylor and Francis, London. page 50 Cohen, B. & Trussell. , J. (EDS) (1996) Preventing and Mitigating AIDS in Sub- Saharan Africa Research and Data Priorities for the Social and behavioural Sciences. National Academy Press, Washington, DC. Dare, 0 .0 . & Clrland, J.G. (1994) Reliability and validity of survey data on sexual behaviour. Health Transformation Review 4 (supplement), 93-110. Degrees Du Lou, A. (1999) Reproductive health and AIDS in sub-Saharan Africa. J;>roblems and prospects. Population (an English Selection) 11, 61-87. Dixon-Mueller, R. (1996) The sexuality connection in reproductive health. In: Learning about Sexuality : A Practical Beginning, pp. 137-157. Edited by S. Zeidenstein & K. Moore Population Council and International Women ' s Health Coalition, New York. Dixon-Mueller, R. & Wasserheit, J. (1990). The Culture of Silence: Reproductive Tract Infections among women in the Third World . International Women ' s Health Coalition, New York. Fami ly Health International (1999) A Dialogue Between the Sexes: Man, Women and AIDS Prevention. http ://www.Family Health International. organisation. United Kingdom Gage, A.J. (1998) Sexual activity and contraceptive use: the components of the decision making process. Study of Family Planning. 29, 154-166 Grundling, J. , Fourie, L, and Dejager (2000). A Research driven HIV/AIDS Strategy for Tertiary Institutions. Technicom Pretoria, South Africa. page 51 Harison, A, Smit Myer!. Prevention of HIV/ AIDS in South Africa: A review of behaviour change interventions, evidence and options for the future . South Africa Journal Science 2000; 96: 285 - 290. Hulton, L.A. , Cullen, R . & Khalokho, W.S. (2000) Perceptions of the risks of sexual activity and their consequences among Ugandan adolescents . Study of . Family Planning. 31, 35-46. Idele, P.A. (2002) Perception of Risk and Sexual Behaviour Change among Women and Men in Kenya. Unpublished PhD Thesis, University of Southampton, United Kingdom. Ingham, · R . & van Zessen, G. (1997) Towards an alternative model of sexual behaviour: from individual properties to international processes. In: Sexual Interactions and HIV Risk: New Conceptual Perspectives in European Research, pp. 83-99. Irwin, K. , Bertrand, J. , Mibandumba, N,. Mbuyi , K. , Muremeri, C ., Mukoka, M ., Munkolenkole, K. , Nzilambi, N. , Bosenge, N. & Ryder, R. (1991) Knowledge, attitudes and beliefs about HIV infection and AIDS among healthy factory workers and their wives, Kinshasa, Zaire. Social Science and Medicine. 32, 917-930. Kelly, K. (2000) ' Communicating For Action : A contextual evaluation of youth response to HIV/AIDS . Site monitoring and evaluation project for the beyond Awareness Campaign, Department of Health , Pretoria, South Africa. Kenijumba, Ragmat Gameldien, Laudon Myer. Sexual Risk Behaviours are influenced by knowing someone with HIV/ AIDS . South Africa Medical Journal Forum 2004; 94 (7): 522-23. page 52 Korn, E .L. & Graubard, B.I. (1995) Analysis of large health surveys : accounting for the sampling design. Journal of Royal Social Series Article 158, 263-295 . Lesthaeghe, R.J. (1989) Production and reproduction in sub-Saharan Africa: An overview of organising principles. In : Reproduction and Social Organisation in Sub-Saharan Africa, pp . 13-59. Edited by R.J. Lesthaeghe. University of Californ~a Press, Berkeley, CA Levy, Amnon (2002) 'A lifetime portfolio of risky and risk-free sexual. ' Journal of Health Economics 21, 993-1007 Meekers, D. (1994)Sexual Initiation and premarital child bearing in sub-Saharan Africa. Population Study 48, 47-64. Meekers, D. & Calves, A (1997) Main girlfriends, girlfriends, mamage, and money : the social context of HIV risk behaviour in sub-Saharan Africa. Health Transformation Review (supplement), 361-375. Jylincer, Jacob (1974) Schooling, Experience, and Earnings (New York and London: National Bureau of Economic Research and Columbia University Press) Nelson Mandela Foundation (2002) (Pretoria, South Africa: South African National HIV Prevalence, Behavioural Risk and Mass Media) Nzioka, C. (1996) Lay perceptions of risk of HIV infection and the social construction of safer sex: some experiences from Kenya. AIDS Care 8, 565-579. Nzioka, C. (2001) Perspectives of adolescent boys on the risks of unwanted pregnancy and sexually transmitted infections: Kenya. Reproductive Health Matters 9, 108-117. page 53 Ocholla-Ayayo, A.B.C. (1997) HIV/AIDS ri sk facto rs and changing sexual practice in Kenya. In: African Famjlies and the Crisis of Social Change, pp . 109- 124. Ocholla-Ayayo, AB.C. & Schwarz, R .A. (1991) Report on Sex Practices and the Spread of STDs and AIDS in Kenya. University of Nairobi, Nairobi . Pfeffermann, D., Skjnner, C.J. , Holmes, DJ., Goldstein, H . & Rasbash. J. (1998)Weighting for unequal selection probabilities in multilevel models. J. Roy. Stati_stical Social. Science Series 60, 23 -40. Philipson, Thomas, and Richard A. Posner (1993) Private Choices and Public Health: the AIDS Epidemic in an Economic Perspective (Cambridge: Harvard University Press) Preston-Whyte, E . (1994) Gender and the lost generation: The dynamics of HIV transmission among black South African teenagers in Kwa-Zulu, Natal Health Transformation Review 4, (supplement), 241-255 . Prohaska, T.R. , Albrecht, G., Levy, J. A. , Sugrue, N. & Kim, J. (1990) Determinants of self-perceived risk for AIDS . Journal of Health and Social Behaviour 31, 384-394. Reid, E . (1999) Placing Women at the Centre of the Analysis. United ations Development Programme (UNDP) : HIV and Development Programme 6. Rao, Vijayendra, Indrani Gupta, Michael Lokshin, and Smarajit Jana (2003) ' Sex workers and the cost of safe sex: The compensating differential for condom use in Calcutta.' Journal of Development Economics Forthcoming Sharma A.K. , Gupta A, Aggarwal op. HIV/AIDS related knowledge, risk perception, attitude and sexual behaviour of working women staying in hostels. Indian Journal Dermatol Venereol Lepsol 2001; 67: 21-24 . page 54 Smith, T.M.F. (1988) To weight or not to weight, that is the question. In: Bayesian Statistics, pp . 437-451. Edited by J.M. Bernado, .H. Degroot, D.V. Lindley & A.F.M. Smith. Oxford University Press, Oxford. South Africa Demographic and Health Survey Data (1998) . Central Statistical Office South Africa. Stallings, R. A. (1990) Media discourse and the social construction of risk of Social Problem 37, 80-95 . UNA IDS (1990) 'Trends in HIV incidence and prevalence: atural course of the epidemic or results of behavioural change ', Geneva UNA.IDS (1998) A measure of success in Uganda: The value of monitoring both HIV prevalence and sexual behaviour. UNAIDS, Geneva UN AIDS (2000a) Report on the global HIV/ AIDS epidemic. Joint United ations programme on HIV/AIDS. UNAIDS, Geneva UNAIDS (2000b) Programme monitoring and evaluation indicators. UNAIDS, Geneva page 55