The Role of Social Capital in HIV Prevention

Umeå University Medical Dissertation Umeå University Medical Dissertation New Series No 1453 ISSN 0346-6612 ISBN 978-91-7459-307-5 New Series No 1453 ISSN 0346-6612 ISBN 978-91-7459-307-5

Department of Public Health and Clinical Medicine Department of Public Health and Clinical Medicine Epidemiology and Global Health Epidemiology and Global Health Umeå University, SE-901 87 Umeå Umeå University, SE-901 87 Umeå

The role of social capital in HIV prevention The role of social capital in HIV prevention Experiences from the Kagera region of Tanzania Experiences from the Kagera region of Tanzania

Gasto Frumence Gasto Frumence 2011 2011

Department of Public Health Muhimbili University Department of Public Health Muhimbili University and Clinical Medicine of Health and Allied Sciences and Clinical Medicine of Health and Allied Sciences Epidemiology and Global Health School of Public Health and Epidemiology and Global Health School of Public Health and Umeå University, Umeå, Sweden Social Sciences Umeå University, Umeå, Sweden Social Sciences Department of Public Health and Clinical Medicine Department of Public Health and Clinical Medicine Epidemiology and Global Health Epidemiology and Global Health Umeå University Umeå University SE-901 87 Umeå, Sweden SE-901 87 Umeå, Sweden

02 02 To my family – my wife Diana To my family – my wife Diana and our children Lorraine, Laura and Larry and our children Lorraine, Laura and Larry

03 03 04 04 Abstract Abstract

Background Background The role of social capital for promoting health has been extensively studied in The role of social capital for promoting health has been extensively studied in recent years but there are few attempts to investigate the possible influence of recent years but there are few attempts to investigate the possible influence of social capital on HIV prevention, particularly in developing countries. The over- social capital on HIV prevention, particularly in developing countries. The overall aims of this thesis are to investigate the links between social capital and HIV all aims of this thesis are to investigate the links between social capital and HIV infection and to contribute to the theoretical framework of the role of social infection and to contribute to the theoretical framework of the role of social capital for HIV prevention. capital for HIV prevention.

Methods Methods Key informant interviews with leaders of organizations, networks, social groups Key informant interviews with leaders of organizations, networks, social groups and communities and focus group discussions with members and non-members of and communities and focus group discussions with members and non-members of the social groups and networks were conducted to map out and characterize various the social groups and networks were conducted to map out and characterize various forms of social capital that may influence HIV prevention. A quantitative commu- forms of social capital that may influence HIV prevention. A quantitative community survey was carried out in three case communities to estimate the influence of nity survey was carried out in three case communities to estimate the influence of social capital on HIV risk behaviors. A cross-sectional survey was conducted to social capital on HIV risk behaviors. A cross-sectional survey was conducted to estimate the HIV prevalence in the urban district representing a high HIV prevalence estimate the HIV prevalence in the urban district representing a high HIV prevalence zone to determine the association between social capital and HIV infection. zone to determine the association between social capital and HIV infection.

Main findings Main findings In early 1990’s many of the social groups in Kagera region were formed because In early 1990’s many of the social groups in Kagera region were formed because of poverty and many AIDS related deaths. This formation of groups enhanced of poverty and many AIDS related deaths. This formation of groups enhanced people’s social and economic support to group members during bereavement people’s social and economic support to group members during bereavement and celebrations as well as provided loans that empowered members economi- and celebrations as well as provided loans that empowered members economically. The social groups also put in place strict rules of conduct, which helped to cally. The social groups also put in place strict rules of conduct, which helped to create new norms, values and trust, which influenced sexual health and thereby create new norms, values and trust, which influenced sexual health and thereby enhanced HIV prevention. Formal organizations worked together with social enhanced HIV prevention. Formal organizations worked together with social groups and facilitated networking and provided avenues for exchange of informa- groups and facilitated networking and provided avenues for exchange of information including health education on HIV/AIDS. Individuals who had access to tion including health education on HIV/AIDS. Individuals who had access to high levels of structural and cognitive social capital were more likely to use con- high levels of structural and cognitive social capital were more likely to use condoms with their casual sex partners compared to individuals with access to low doms with their casual sex partners compared to individuals with access to low levels. Women with access to high levels of structural social capital were more levels. Women with access to high levels of structural social capital were more likely to use condoms with casual sex partners compared to those with low levels. likely to use condoms with casual sex partners compared to those with low levels. Individuals with access to low levels of structural social capital were less likely Individuals with access to low levels of structural social capital were less likely to be tested for HIV compared to those with access to high levels. However, there to be tested for HIV compared to those with access to high levels. However, there was no association between access to cognitive social capital and being tested for was no association between access to cognitive social capital and being tested for HIV. Individuals who had access to low levels of both structural and cognitive HIV. Individuals who had access to low levels of both structural and cognitive social capital were more likely to be HIV positive compared to individuals who social capital were more likely to be HIV positive compared to individuals who had access to high levels with a similar pattern among men and women. had access to high levels with a similar pattern among men and women.

i i Conclusion Conclusion This thesis indicates that social capital in its structural and cognitive forms is This thesis indicates that social capital in its structural and cognitive forms is protective to HIV infection and has played an important role in the observed protective to HIV infection and has played an important role in the observed decline in HIV trends in the Kagera region. Structural and cognitive social capi- decline in HIV trends in the Kagera region. Structural and cognitive social capital has enabled community members to decrease number of sexual partners, tal has enabled community members to decrease number of sexual partners, delay sexual debut for the young generation, reduce opportunities for casual sex delay sexual debut for the young generation, reduce opportunities for casual sex and empower community members to demand or use condoms. It is recom- and empower community members to demand or use condoms. It is recom- mended that policy makers and programme managers consider involving grass- mended that policy makers and programme managers consider involving grassroots’ social groups and networks in the design and delivery of interventions roots’ social groups and networks in the design and delivery of interventions strategies to reduce HIV transmission. strategies to reduce HIV transmission.

ii ii Original papers Original papers

This thesis is based on the following papers: This thesis is based on the following papers:

I Frumence G, Killewo J, Kwesigabo G, Nyström L, Eriksson M, Emmelin M. I Frumence G, Killewo J, Kwesigabo G, Nyström L, Eriksson M, Emmelin M. Social capital and the decline in HIV transmission – A case study in three Social capital and the decline in HIV transmission – A case study in three villages in Kagera region, Tanzania. Journal of Social Aspects of HIV/AIDS villages in Kagera region, Tanzania. Journal of Social Aspects of HIV/AIDS (SAHARA) 2010;7:9–19. (SAHARA) 2010;7:9–19.

II Frumence G, Eriksson M, Killewo J, Nyström L, Emmelin M. Exploring the II Frumence G, Eriksson M, Killewo J, Nyström L, Emmelin M. Exploring the role of cognitive and structural social capital in the declining trends of HIV/ role of cognitive and structural social capital in the declining trends of HIV/ AIDS in the Kagera region of Tanzania – A grounded theory study. African AIDS in the Kagera region of Tanzania – A grounded theory study. African Journal of Aids Research 2011;10:1–13. Journal of Aids Research 2011;10:1–13.

III Frumence G, Emmelin M, Kwesigabo G, Killewo J, Eriksson M, Nyström L. III Frumence G, Emmelin M, Kwesigabo G, Killewo J, Eriksson M, Nyström L. Social capital and HIV risk related behaviors in Kagera region, Tanzania Social capital and HIV risk related behaviors in Kagera region, Tanzania (Submitted). (Submitted).

IV Frumence G, Emmelin M, Kwesigabo G, Killewo J, Eriksson M, Nyström L. IV Frumence G, Emmelin M, Kwesigabo G, Killewo J, Eriksson M, Nyström L. Social capital and HIV infection in Bukoba urban district, Kagera region, Social capital and HIV infection in Bukoba urban district, Kagera region, Tanzania. (Submitted). Tanzania. (Submitted).

Paper I and II are printed with the permission of the publishers. Paper I and II are printed with the permission of the publishers.

iii iii Abbreviations Abbreviations

AIDS Acquired Immunodeficiency Syndrome AIDS Acquired Immunodeficiency Syndrome CI Confidence interval CI Confidence interval CSC Cognitive social capital CSC Cognitive social capital ELISA Enyzme-Linked Immunoabsorbent Assay ELISA Enyzme-Linked Immunoabsorbent Assay FBO Faith-based organizations FBO Faith-based organizations FGD Focus groups discussion FGD Focus groups discussion HIV Human Immunodeficiency Virus HIV Human Immunodeficiency Virus IDU Intravenous drug users IDU Intravenous drug users KARP Kagera AIDS research project KARP Kagera AIDS research project MoHSW Ministry of Health and Social Welfare MoHSW Ministry of Health and Social Welfare MUHAS Muhimbili University of Health and Allied Sciences MUHAS Muhimbili University of Health and Allied Sciences NBS National Bureau of Statistics NBS National Bureau of Statistics NGO Non-governmental organizations NGO Non-governmental organizations OR Odds ratio OR Odds ratio PCA Principal component analysis PCA Principal component analysis Sida Swedish International Development Agency Sida Swedish International Development Agency SSC Structural social capital SSC Structural social capital STI Sexually transmitted infections STI Sexually transmitted infections TACAIDS Tanzania Commission for AIDS TACAIDS Tanzania Commission for AIDS TDHS Tanzania Demographic Health Survey TDHS Tanzania Demographic Health Survey URT United Republic of Tanzania URT United Republic of Tanzania VCT Voluntary, counseling and treatment services VCT Voluntary, counseling and treatment services

iv iv Administrative definitions Administrative definitions

Region One of the 26 largest geographical and administrative divisions Region One of the 26 largest geographical and administrative divisions of Tanzania governed by a regional commissioner as a political of Tanzania governed by a regional commissioner as a political leader and regional administrative secretary as the chief executive leader and regional administrative secretary as the chief executive officer. officer.

District The largest sub-division of a region governed by a district com- District The largest sub-division of a region governed by a district commissioner as a political leader and district administrative secretary missioner as a political leader and district administrative secretary as the chief executive officer. as the chief executive officer.

Division The largest sub-division of a district governed by a divisional Division The largest sub-division of a district governed by a divisional secretary. secretary.

Ward The largest sub-division of a division governed by a ward execu- Ward The largest sub-division of a division governed by a ward executive officer. tive officer.

Village The largest sub-division of a ward in the rural setting and admin- Village The largest sub-division of a ward in the rural setting and administered by a village executive officer. istered by a village executive officer.

Ten-cell unit The smallest administrative unit consisting of a group of 10 Ten-cell unit The smallest administrative unit consisting of a group of 10 households and administered by a ten-cell leader who is answer- households and administered by a ten-cell leader who is answerable to the ruling party. The ten-cell unit applies to both urban able to the ruling party. The ten-cell unit applies to both urban and rural areas and was extensively used before the introduction and rural areas and was extensively used before the introduction of a multiparty democratic system in 1995. of a multiparty democratic system in 1995.

Kitongoji/ The smallest rural based administrative unit consisting of 50-200 Kitongoji/ The smallest rural based administrative unit consisting of 50-200 hamlet households and governed by a chairperson. It replaced the ten-cell hamlet households and governed by a chairperson. It replaced the ten-cell leadership unit that was used before the introduction of a multi- leadership unit that was used before the introduction of a multiparty system. The chairperson is answerable to the local govern- party system. The chairperson is answerable to the local government. ment.

Street The smallest urban administrative unit consisting of 25-150 Street The smallest urban administrative unit consisting of 25-150 households governed by a street secretary as executive officer. households governed by a street secretary as executive officer. This administrative unit was introduced after the introduction of This administrative unit was introduced after the introduction of multiparty system and its leaders are answerable to the local multiparty system and its leaders are answerable to the local government. government.

v v Content Content

Abstract ...... iii Abstract ...... iii Original papers ...... vi Original papers ...... vi Abbreviations ...... vii Abbreviations ...... vii Administrative definitions...... viii Administrative definitions...... viii

Introduction ...... 1 Introduction ...... 1

Aims ...... 3 Aims ...... 3

Study context ...... 4 Study context ...... 4 The United Republic of Tanzania ...... 4 The United Republic of Tanzania ...... 4 The Kagera AIDS Research Project (KARP) ...... 8 The Kagera AIDS Research Project (KARP) ...... 8

Theoretical framework ...... 10 Theoretical framework ...... 10 The concept of social capital ...... 10 The concept of social capital ...... 10 Three approaches of social capital ...... 12 Three approaches of social capital ...... 12 Classifications of social capital...... 13 Classifications of social capital...... 13 Social capital and health outcomes ...... 15 Social capital and health outcomes ...... 15

Materials and methods ...... 18 Materials and methods ...... 18 A multi-methodological approach ...... 18 A multi-methodological approach ...... 18 Study area ...... 20 Study area ...... 20 Study population and sample size ...... 21 Study population and sample size ...... 21 Sampling procedures ...... 22 Sampling procedures ...... 22 Measure instruments ...... 24 Measure instruments ...... 24 Measurement of social capital ...... 25 Measurement of social capital ...... 25 The study team ...... 25 The study team ...... 25 Laboratory procedures ...... 26 Laboratory procedures ...... 26 Ethical considerations ...... 26 Ethical considerations ...... 26 Data analysis ...... 26 Data analysis ...... 26

Results ...... 30 Results ...... 30 How can the social capital in the study area be characterized? ...... 30 How can the social capital in the study area be characterized? ...... 30 What are the mechanisms through which social capital may What are the mechanisms through which social capital may influence HIV risk behaviors?...... 31 influence HIV risk behaviors?...... 31 Is access to structural and cognitive social capital associated Is access to structural and cognitive social capital associated with HIV risk behaviors? ...... 35 with HIV risk behaviors? ...... 35 Is access to structural and cognitive social capital associated with Is access to structural and cognitive social capital associated with risk of HIV infection? …………………………………………………………………………...... 38 risk of HIV infection? …………………………………………………………………………...... 38

Methodological considerations ...... 41 Methodological considerations ...... 41 Strengths ...... 41 Strengths ...... 41 Limitations ...... 42 Limitations ...... 42 vi vi Discussion ...... 43 Discussion ...... 43 Social capital and HIV infection ...... 43 Social capital and HIV infection ...... 43 Social capital and HIV prevention ...... 44 Social capital and HIV prevention ...... 44 Implications for HIV prevention strategies ...... 46 Implications for HIV prevention strategies ...... 46 The role of local governance in mobilizing social capital ...... 47 The role of local governance in mobilizing social capital ...... 47

Concluding remarks ...... 48 Concluding remarks ...... 48

The researcher ...... 49 The researcher ...... 49

Acknowledgement ...... 50 Acknowledgement ...... 50

References ...... 55 References ...... 55

vii vii

INTRODUCTION INTRODUCTION Introduction Introduction

In spite of its long history, the concept of social capital became more popular in In spite of its long history, the concept of social capital became more popular in the political and socio-economic research areas following the 1993 Putnam’s work the political and socio-economic research areas following the 1993 Putnam’s work on “Making Democracy Work: Civic Traditions in Modern Italy”. Here the on “Making Democracy Work: Civic Traditions in Modern Italy”. Here the concept was used to explain issues of local governance in Italy. Putnam defines concept was used to explain issues of local governance in Italy. Putnam defines social capital as people’s engagement in local associations, informal social net- social capital as people’s engagement in local associations, informal social networks and norms based on trust and reciprocity between community members. works and norms based on trust and reciprocity between community members. These features of social organization (trust, norms, and networks) can help to These features of social organization (trust, norms, and networks) can help to improve the efficiency of society by facilitating coordinated actions. He has ex- improve the efficiency of society by facilitating coordinated actions. He has explained several benefits to be accrued from social capital, ranging from govern- plained several benefits to be accrued from social capital, ranging from governance, health, education and neighborhood safety (Putnam, 2000). Putnam’s work ance, health, education and neighborhood safety (Putnam, 2000). Putnam’s work on social capital and governance with benefits both to individuals and the com- on social capital and governance with benefits both to individuals and the community became the beginning of a new paradigm shift that related social capital munity became the beginning of a new paradigm shift that related social capital to have positive impact particularly in the social and economic areas. to have positive impact particularly in the social and economic areas.

Studies on social capital and health outcomes are recent phenomena in public Studies on social capital and health outcomes are recent phenomena in public health research. Before 1990s the dominant paradigm focused on the association health research. Before 1990s the dominant paradigm focused on the association between behavioral and biomedical factors in treatment and prevention of dis- between behavioral and biomedical factors in treatment and prevention of diseases both at individual and community level (Campbell, 2011). The concept of eases both at individual and community level (Campbell, 2011). The concept of social capital became more prominent starting from mid 1990s when the public social capital became more prominent starting from mid 1990s when the public health arena witnessed a paradigm shift from the traditional biomedical approach health arena witnessed a paradigm shift from the traditional biomedical approach to more social and community oriented ways of determining illness behavior and to more social and community oriented ways of determining illness behavior and related interventions. This paradigm shift is seen in the advocacy by public health related interventions. This paradigm shift is seen in the advocacy by public health researchers to use social networks as important channels for conveying health researchers to use social networks as important channels for conveying health related information instead of relying on traditional information sources such as related information instead of relying on traditional information sources such as posters, radio, television and newspapers (Campbell, 2000). posters, radio, television and newspapers (Campbell, 2000).

This paradigm shift went hand in hand with the new research interest focusing This paradigm shift went hand in hand with the new research interest focusing on HIV/AIDS and social capital. At the beginning, researchers paid more atten- on HIV/AIDS and social capital. At the beginning, researchers paid more attention to the impact of HIV/AIDS on social networks and social cohesion. Pronyk tion to the impact of HIV/AIDS on social networks and social cohesion. Pronyk (2002) conducted a study on “social capital and the HIV/AIDS epidemic in rural (2002) conducted a study on “social capital and the HIV/AIDS epidemic in rural South Africa” and reported that the HIV/AIDS epidemic had created a heavy South Africa” and reported that the HIV/AIDS epidemic had created a heavy burden on the social fabric. Traditional networks, including social groups, served burden on the social fabric. Traditional networks, including social groups, served as safety net enabling communities to cope with the social and economic problems as safety net enabling communities to cope with the social and economic problems facing them. These networks supported communities by for example contribut- facing them. These networks supported communities by for example contributing money to funerals, donating food and lending money. Several other studies ing money to funerals, donating food and lending money. Several other studies have shown that the HIV/AIDS burden disintegrated and disempowered existing have shown that the HIV/AIDS burden disintegrated and disempowered existing social networks to cope with the economic burden created by the same epidemic. social networks to cope with the economic burden created by the same epidemic. The networks could no longer manage the increasing number of deaths, orphans The networks could no longer manage the increasing number of deaths, orphans

1 1 INTRODUCTION INTRODUCTION and the growing number of families headed by children (Foster, 2006; Nombo, and the growing number of families headed by children (Foster, 2006; Nombo, 2007). 2007).

Researchers in public health have developed an interest in exploring how social Researchers in public health have developed an interest in exploring how social capital in terms of informal and formal networks (structural) and norms and capital in terms of informal and formal networks (structural) and norms and trust (cognitive) are associated with the risk of HIV infection. However, there is trust (cognitive) are associated with the risk of HIV infection. However, there is continuing debate among researchers regarding the role of social capital in in- continuing debate among researchers regarding the role of social capital in influencing the HIV epidemic. In some contexts social capital, characterized by fluencing the HIV epidemic. In some contexts social capital, characterized by high level of community cohesion, has been reported to have a protective and high level of community cohesion, has been reported to have a protective and facilitating role for community responses to the HIV epidemic. In other settings, facilitating role for community responses to the HIV epidemic. In other settings, membership in social groups, such as sexual networking groups, may increase membership in social groups, such as sexual networking groups, may increase the risk of being HIV infected. Similarly those who have been excluded from the risk of being HIV infected. Similarly those who have been excluded from membership may have an elevated risk for HIV infection due to lack of collective membership may have an elevated risk for HIV infection due to lack of collective norms and values (Baum, 1999; Gregson et al, 2004; Pronyk et al, 2008). norms and values (Baum, 1999; Gregson et al, 2004; Pronyk et al, 2008).

Whether social capital is protective for HIV infection needs to be further explored Whether social capital is protective for HIV infection needs to be further explored and elaborated. There is need to understand more about people’s experiences of and elaborated. There is need to understand more about people’s experiences of being involved in organizations and social groups (structural social capital) and being involved in organizations and social groups (structural social capital) and the trust and reciprocity embedded in their interactions (cognitive social capital). the trust and reciprocity embedded in their interactions (cognitive social capital). There is limited knowledge about the mechanisms through which social capital There is limited knowledge about the mechanisms through which social capital in its structural and cognitive forms influence HIV risk behaviors. This thesis is in its structural and cognitive forms influence HIV risk behaviors. This thesis is an attempt to provide information to fill this knowledge gap. Results from a an attempt to provide information to fill this knowledge gap. Results from a qualitative case study of three communities with varying HIV prevalence is used qualitative case study of three communities with varying HIV prevalence is used to characterize social capital in terms of its structural and cognitive components to characterize social capital in terms of its structural and cognitive components and to construct a theoretical model to illustrate the mechanisms through which and to construct a theoretical model to illustrate the mechanisms through which social capital may influence HIV risk behaviors and therefore enhance preven- social capital may influence HIV risk behaviors and therefore enhance prevention. Data from a survey of social capital and HIV risk behaviors in the same three tion. Data from a survey of social capital and HIV risk behaviors in the same three communities are used to analyze the links between different forms and levels of communities are used to analyze the links between different forms and levels of individual social capital and HIV risk behaviors both for men and women. Data individual social capital and HIV risk behaviors both for men and women. Data from a cross-sectional study are used to estimate the association between social from a cross-sectional study are used to estimate the association between social capital, socio-demographic characteristics and HIV infection. capital, socio-demographic characteristics and HIV infection.

2 2 AIMS AIMS Aims Aims

The overall aims of this thesis are to investigate the links between social capital The overall aims of this thesis are to investigate the links between social capital and HIV infection and to contribute to the theoretical framework of the role of and HIV infection and to contribute to the theoretical framework of the role of social capital in HIV prevention social capital in HIV prevention

The specific aims are: The specific aims are:

• To explore and describe the mechanisms of how structural and cognitive social • To explore and describe the mechanisms of how structural and cognitive social capital may influence the transmission of HIV/AIDS (Paper I and II) capital may influence the transmission of HIV/AIDS (Paper I and II)

• To determine the association between structural and cognitive social capital, • To determine the association between structural and cognitive social capital, HIV/AIDS risk related behaviors and HIV infection (Paper III and IV) HIV/AIDS risk related behaviors and HIV infection (Paper III and IV)

3 3 STUDY CONTEXT STUDY CONTEXT Study context Study context

The United Republic of Tanzania The United Republic of Tanzania The United Republic of Tanzania was formed out of the union of two sovereign The United Republic of Tanzania was formed out of the union of two sovereign states of Tanganyika and Zanzibar. The country is bordered by Kenya and Ugan- states of Tanganyika and Zanzibar. The country is bordered by Kenya and Ugan- da to the north, Rwanda, Burundi and the Democratic Republic of the Congo to da to the north, Rwanda, Burundi and the Democratic Republic of the Congo to the west, and Zambia, Malawi and Mozambique to the south. The country’s eastern the west, and Zambia, Malawi and Mozambique to the south. The country’s eastern borders lie on the Indian Ocean (Figure 1). Tanzania mainland, where Kagera region borders lie on the Indian Ocean (Figure 1). Tanzania mainland, where Kagera region is situated, has 21 regions and is further sub-divided into 113 districts. It has a is situated, has 21 regions and is further sub-divided into 113 districts. It has a total area of 945,087 km2 of which 886,037 is land and the remaining part is cov- total area of 945,087 km2 of which 886,037 is land and the remaining part is cov- ered by water. Based on the projections from 2010 Tanzania national bureau of ered by water. Based on the projections from 2010 Tanzania national bureau of statistics, the total population of the country is about 43 million (Table 1). statistics, the total population of the country is about 43 million (Table 1).

Figure 1. Map of the United Republic of Tanzania showing Kagera Figure 1. Map of the United Republic of Tanzania showing Kagera region (see arrow). Source: URT-TDHS (2011). region (see arrow). Source: URT-TDHS (2011).

4 4 STUDY CONTEXT STUDY CONTEXT

Socio-political and economic situation Socio-political and economic situation Tanzania gained its independence from the British colony in 1961 and in 1967 Tanzania gained its independence from the British colony in 1961 and in 1967 the country adopted the socialist ideology, locally known as “Ujamaa”. The basic the country adopted the socialist ideology, locally known as “Ujamaa”. The basic traits under Ujamaa emphasized cooperation, extended family-hood, and welfare traits under Ujamaa emphasized cooperation, extended family-hood, and welfare to all members of the society (Nyerere, 1968). The implementation of Ujamaa to all members of the society (Nyerere, 1968). The implementation of Ujamaa ideology was largely characterized by communalism based on the principle of ideology was largely characterized by communalism based on the principle of collective production, nationalization of all major means of production by prac- collective production, nationalization of all major means of production by prac- ticing state economy and an equalitarian resource distribution (Nyerere, 1967; ticing state economy and an equalitarian resource distribution (Nyerere, 1967; Rodney, 1972). During this period Tanzania experienced equity in access to social Rodney, 1972). During this period Tanzania experienced equity in access to social services by providing free health and education services to all. services by providing free health and education services to all.

In the early 1980s, Tanzania’s economy experienced a heavy deterioration caused In the early 1980s, Tanzania’s economy experienced a heavy deterioration caused by poor implementation of Ujamaa policies, the war with Uganda in the late by poor implementation of Ujamaa policies, the war with Uganda in the late 1970s, the hiking of oil prices, droughts, low prices of exports, and increasing 1970s, the hiking of oil prices, droughts, low prices of exports, and increasing debt (Maliyamkono & Bagachwa, 1990; Bagachwa, 1992). In the mid 1980s the debt (Maliyamkono & Bagachwa, 1990; Bagachwa, 1992). In the mid 1980s the government of Tanzania adopted the World Bank and International Monetary government of Tanzania adopted the World Bank and International Monetary Fund’s liberal policies aiming at recovering the economy through structural Fund’s liberal policies aiming at recovering the economy through structural adjustment programmes. The liberal policies led to the demise of the Ujamaa adjustment programmes. The liberal policies led to the demise of the Ujamaa ideology when adopting the new policies of free market economy. ideology when adopting the new policies of free market economy.

The introduction of free market economy forced the government of Tanzania to The introduction of free market economy forced the government of Tanzania to abolish free access to social services by introducing a cost sharing policy. The abolish free access to social services by introducing a cost sharing policy. The government introduced user fees and social insurance schemes in the health government introduced user fees and social insurance schemes in the health sector and tuition fee in the education sector. The cost sharing policy has expe- sector and tuition fee in the education sector. The cost sharing policy has experienced several challenges. A study based on literature review and semi-structured rienced several challenges. A study based on literature review and semi-structured interviews (Schwerzel et al, 2004) showed that the government has not achieved interviews (Schwerzel et al, 2004) showed that the government has not achieved the intended goals of equity. The free market policies have increased marginaliza- the intended goals of equity. The free market policies have increased marginaliza- tion of the vulnerable poor and excluded them from accessing health services. The tion of the vulnerable poor and excluded them from accessing health services. The most affected vulnerable groups include: under-five children, pregnant women, most affected vulnerable groups include: under-five children, pregnant women, widows, elders aged 60 years old and beyond, people living with AIDS, and or- widows, elders aged 60 years old and beyond, people living with AIDS, and orphans. When one family member gets sick, close relatives are forced to sell their phans. When one family member gets sick, close relatives are forced to sell their disposable properties, borrow money from other capable relatives and friends or disposable properties, borrow money from other capable relatives and friends or get money from social groups (for those who are members) in order to pay for get money from social groups (for those who are members) in order to pay for user fees. This situation has increased impoverishment and created difficulties user fees. This situation has increased impoverishment and created difficulties for poor people to survive. Sometimes children have to drop out of schools due to for poor people to survive. Sometimes children have to drop out of schools due to lack of money to buy uniforms and stationeries (Save the Children, 2005). lack of money to buy uniforms and stationeries (Save the Children, 2005).

5 5 STUDY CONTEXT STUDY CONTEXT

Table 1. Selected demographic indicators for Tanzania and Kagera region. Table 1. Selected demographic indicators for Tanzania and Kagera region.

Basic characteristic Tanzania Kagera region Basic characteristic Tanzania Kagera region Population size *43,000,000 *2,600,000 Population size *43,000,000 *2,600,000 Population growth rate 2.9% 3.1% Population growth rate 2.9% 3.1% Fertility rate 5.4% 7.4% Fertility rate 5.4% 7.4% Maternal mortality ratio 454/100,000 Maternal mortality ratio 454/100,000 Stillbirth rate 29/1000 Stillbirth rate 29/1000 Neonatal mortality rate 26/1000 31.5/1000 Neonatal mortality rate 26/1000 31.5/1000 Perinatal 36/1000 32/1000 Perinatal 36/1000 32/1000 Postneonatal 25/1000 81.6/1000 Postneonatal 25/1000 81.6/1000 Under-five mortality rate 81/1000 157/1000 Under-five mortality rate 81/1000 157/1000 Infant mortality rate 51/1000 96/1000 Infant mortality rate 51/1000 96/1000 Female literacy rate 60% 68% Female literacy rate 60% 68% Male literacy rate 69% 69% Male literacy rate 69% 69% Life expectancy (years) 51 53 Life expectancy (years) 51 53

Sources: URT (1989, 2010), TDHS (2010), MoHSW (2010) Sources: URT (1989, 2010), TDHS (2010), MoHSW (2010) *2010 projected population by Tanzania Bureau of Statistics *2010 projected population by Tanzania Bureau of Statistics

Despite the collapse of Ujamaa and the adoption of liberal economy, this ideol- Despite the collapse of Ujamaa and the adoption of liberal economy, this ideology left behind important legacy to Tanzanians including the values of solidarity, ogy left behind important legacy to Tanzanians including the values of solidarity, trust, cooperation, collectiveness and social networks which laid down a founda- trust, cooperation, collectiveness and social networks which laid down a foundation of social capital in the country. tion of social capital in the country.

The HIV epidemic The HIV epidemic Tanzania is one of the Sub-Saharan African countries greatly affected by the HIV Tanzania is one of the Sub-Saharan African countries greatly affected by the HIV epidemic. The first three HIV cases in the country were discovered in 1983 in the epidemic. The first three HIV cases in the country were discovered in 1983 in the Kagera region. Ten years later, HIV had reached all regions of Tanzania (Kwesi- Kagera region. Ten years later, HIV had reached all regions of Tanzania (Kwesi- gabo, 2001) including the rural areas thereby increasing the previously low rural gabo, 2001) including the rural areas thereby increasing the previously low rural prevalence to more than 10% (TACAIDS, 2006). Data from the Tanzania Health prevalence to more than 10% (TACAIDS, 2006). Data from the Tanzania Health Indicators Survey indicates that there has been a slight decrease in the overall Indicators Survey indicates that there has been a slight decrease in the overall prevalence for both men and women from 6.3% (2003-04) to 4.6% (2007-08) prevalence for both men and women from 6.3% (2003-04) to 4.6% (2007-08) for men and from 7.7% (2003-04) to 6.6% (2007-08) for women (TACAIDS, for men and from 7.7% (2003-04) to 6.6% (2007-08) for women (TACAIDS, 2008). 2008).

Several structural factors relating to the social, cultural and economic arenas Several structural factors relating to the social, cultural and economic arenas have been reported as responsible for spreading HIV infection in Tanzania. These have been reported as responsible for spreading HIV infection in Tanzania. These factors include: poverty, multiple partnering, family separation and gender-based factors include: poverty, multiple partnering, family separation and gender-based violence (URT, 2009). Areas with high social and economic interactions along violence (URT, 2009). Areas with high social and economic interactions along main roads, i. e. with trading centres, a large number of bars, hotels, shops and main roads, i. e. with trading centres, a large number of bars, hotels, shops and schools have also been identified as risk environments for HIV transmission schools have also been identified as risk environments for HIV transmission (Bloom et al, 2002). Cultural norms, beliefs and practices such as polygamy, wife (Bloom et al, 2002). Cultural norms, beliefs and practices such as polygamy, wife

6 6 STUDY CONTEXT STUDY CONTEXT inheritance and female circumcision are also believed to have contributed to HIV inheritance and female circumcision are also believed to have contributed to HIV transmission in Tanzania (Lugalla et al, 2004; TACAIDS, 2008). transmission in Tanzania (Lugalla et al, 2004; TACAIDS, 2008).

Like in many Sub-Saharan African countries, women in Tanzania bear the biggest Like in many Sub-Saharan African countries, women in Tanzania bear the biggest burden of HIV. Although biological factors contribute to their increased risk there burden of HIV. Although biological factors contribute to their increased risk there is a growing recognition that structural factors including gender inequalities play is a growing recognition that structural factors including gender inequalities play a role for HIV transmission. Women’s low social and economic condition makes a role for HIV transmission. Women’s low social and economic condition makes them dependent and powerless decreasing their possibilities to negotiate sexual them dependent and powerless decreasing their possibilities to negotiate sexual relations. Taboos and rigid gender roles keep women at home to carry out do- relations. Taboos and rigid gender roles keep women at home to carry out domestic activities. These practices make women unexposed to HIV education and mestic activities. These practices make women unexposed to HIV education and increase their likelihood of contracting HIV/AIDS (UNICEF, 2002). increase their likelihood of contracting HIV/AIDS (UNICEF, 2002).

HIV prevention and control HIV prevention and control Tanzania’s HIV preventive programs utilize two main approaches based on bio- Tanzania’s HIV preventive programs utilize two main approaches based on biomedical and behavioral interventions. Such programs target both the general medical and behavioral interventions. Such programs target both the general population and specific vulnerable groups. The behavioral interventions include population and specific vulnerable groups. The behavioral interventions include information, education and communications programs integrated into work information, education and communications programs integrated into work places, faith-based organizations (FBOs), community groups and schools. These places, faith-based organizations (FBOs), community groups and schools. These interventions focus on reducing risky behaviors such as number of casual sex interventions focus on reducing risky behaviors such as number of casual sex partners, condom use, intergenerational sex, commercial sex and early sexual partners, condom use, intergenerational sex, commercial sex and early sexual debut, all of which are advocated by massive media campaigns. The biomedical debut, all of which are advocated by massive media campaigns. The biomedical interventions include prevention of mother to child transmission, male circum- interventions include prevention of mother to child transmission, male circumcision, HIV counseling and testing and screening and treatment of sexually cision, HIV counseling and testing and screening and treatment of sexually transmitted infections (STI) (Kwesigabo, 2001; Lugalla et al, 2004; URT, 2009). transmitted infections (STI) (Kwesigabo, 2001; Lugalla et al, 2004; URT, 2009).

The new development in the public health discourse has raised concerns about The new development in the public health discourse has raised concerns about the ineffectiveness of the biomedical and behavioral approaches in the fight the ineffectiveness of the biomedical and behavioral approaches in the fight against HIV (Campbell, 2003). Furthermore, researchers and implementers of against HIV (Campbell, 2003). Furthermore, researchers and implementers of HIV prevention programs have put limited attention to the underlying social and HIV prevention programs have put limited attention to the underlying social and structural contextual factors influencing the HIV epidemic (UNAIDS, 2000; structural contextual factors influencing the HIV epidemic (UNAIDS, 2000; Pronyk, 2006). HIV intervention programs have also been criticized for adopting Pronyk, 2006). HIV intervention programs have also been criticized for adopting a top down approach in both design and implementation. The emerging paradigm a top down approach in both design and implementation. The emerging paradigm shift from a biomedical and behavioral approach to a more social and structural shift from a biomedical and behavioral approach to a more social and structural approach support the argument that community participation in the design and approach support the argument that community participation in the design and delivery of any intervention will lead to more effective HIV prevention programs. delivery of any intervention will lead to more effective HIV prevention programs. This means that community mobilization, participation and empowerment are This means that community mobilization, participation and empowerment are emerging in the public health discourse as the guiding principles for programs emerging in the public health discourse as the guiding principles for programs seeking to address the underlying broader social and structural factors affecting seeking to address the underlying broader social and structural factors affecting the vulnerable groups (Pronyk, 2006). However, there are a limited number of the vulnerable groups (Pronyk, 2006). However, there are a limited number of studies in the African setting that have studied the importance of community studies in the African setting that have studied the importance of community

7 7 STUDY CONTEXT STUDY CONTEXT mobilization and participation in HIV interventions. A review of epidemiological mobilization and participation in HIV interventions. A review of epidemiological data has highlighted that social mobilization through local networks has been an data has highlighted that social mobilization through local networks has been an effective approach to reduce the HIV prevalence in Uganda (Low-Beer & Sem- effective approach to reduce the HIV prevalence in Uganda (Low-Beer & Sem- pala, 2010). Social capital has emerged as an important concept of the new pala, 2010). Social capital has emerged as an important concept of the new paradigm that acknowledges the social and structural context of HIV interven- paradigm that acknowledges the social and structural context of HIV interventions. tions.

The Kagera Aids Research Project (KARP) The Kagera Aids Research Project (KARP) The research questions raised in this thesis originate from the project “Epidemi- The research questions raised in this thesis originate from the project “Epidemi- ology towards evaluation of interventions and monitoring of HIV infection in the ology towards evaluation of interventions and monitoring of HIV infection in the Kagera region of Tanzania”. The project, popularly known as Kagera AIDS Re- Kagera region of Tanzania”. The project, popularly known as Kagera AIDS Re- search Project (KARP) was initiated in 1986 as a long-term research and inter- search Project (KARP) was initiated in 1986 as a long-term research and intervention programme in which a series of population-based epidemiological and vention programme in which a series of population-based epidemiological and socio-anthropological studies have been performed. The first reported cases of socio-anthropological studies have been performed. The first reported cases of AIDS in Tanzania were identified at the Kagera Regional Hospital in 1983. Fol- AIDS in Tanzania were identified at the Kagera Regional Hospital in 1983. Fol- lowing an observation that the epidemic was already in Tanzania an agreement lowing an observation that the epidemic was already in Tanzania an agreement between the MoHSW, Muhimbili University of Health and Allied Science (MU- between the MoHSW, Muhimbili University of Health and Allied Science (MU- HAS), and Swedish International Development Agency (Sida) was reached to HAS), and Swedish International Development Agency (Sida) was reached to initiate population-based studies in the Kagera region. The overall aim of the initiate population-based studies in the Kagera region. The overall aim of the project was to determine the magnitude and spread of HIV infection in the region project was to determine the magnitude and spread of HIV infection in the region as well as to study the socio-anthropological aspects of its spread. The basic trait as well as to study the socio-anthropological aspects of its spread. The basic trait of KARP was the inclusion of inter-disciplinary collaboration between the social of KARP was the inclusion of inter-disciplinary collaboration between the social and biomedical sciences. This strategy involved a joint venture between the dis- and biomedical sciences. This strategy involved a joint venture between the dis- ciplines of epidemiology, social science, linguistic and microbiology/immunol- ciplines of epidemiology, social science, linguistic and microbiology/immunology in both Tanzania and Sweden (Kwesigabo, 2001). Since its inception, KARP ogy in both Tanzania and Sweden (Kwesigabo, 2001). Since its inception, KARP has conducted several cross-sectional studies in order to monitor HIV infection has conducted several cross-sectional studies in order to monitor HIV infection trends in the region. The baseline survey in 1987 revealed an overall prevalence trends in the region. The baseline survey in 1987 revealed an overall prevalence of HIV-infection varying from 24.2% in the urban area to 0.6% in the most remote of HIV-infection varying from 24.2% in the urban area to 0.6% in the most remote rural area (Killewo et al, 1990). Subsequent studies conducted in 1996, 1999 and rural area (Killewo et al, 1990). Subsequent studies conducted in 1996, 1999 and 2004 showed declining trends of HIV infection in the region, particularly among 2004 showed declining trends of HIV infection in the region, particularly among the women (Kwesigabo, 2001; Frumence et al, 2011). the women (Kwesigabo, 2001; Frumence et al, 2011).

KARP and other studies have found several factors to be responsible for the KARP and other studies have found several factors to be responsible for the observed downward trends. These factors included promotion of condom use, observed downward trends. These factors included promotion of condom use, health education, voluntary HIV counseling and testing (VCT), an increased health education, voluntary HIV counseling and testing (VCT), an increased awareness and knowledge about HIV/AIDS causes and prevention. An increased awareness and knowledge about HIV/AIDS causes and prevention. An increased openness that may have decreased the stigma previously associated with the openness that may have decreased the stigma previously associated with the disease was also identified to have played a role in the declining trends (Kwesi- disease was also identified to have played a role in the declining trends (Kwesi- gabo, 2001; Lugalla et al, 2004; URT, 2009). Although many factors had been gabo, 2001; Lugalla et al, 2004; URT, 2009). Although many factors had been proposed to explain the declining trends, it was still unclear how community proposed to explain the declining trends, it was still unclear how community

8 8 STUDY CONTEXT STUDY CONTEXT participation or involvement in social groups or associations may have influenced participation or involvement in social groups or associations may have influenced people’s sexual health behavior, thus contributing to the observed HIV decline people’s sexual health behavior, thus contributing to the observed HIV decline in Kagera region. This knowledge gap prompted the research interest to use the in Kagera region. This knowledge gap prompted the research interest to use the theoretical concept of social capital to investigate how social groups and networks theoretical concept of social capital to investigate how social groups and networks may positively shape people’s sexual health behaviors and ultimately contribute may positively shape people’s sexual health behaviors and ultimately contribute to reduction of HIV transmission. to reduction of HIV transmission.

The following section gives an overview of the theory of social capital by examin- The following section gives an overview of the theory of social capital by examin- ing the origins of the concept itself and describes the theoretical explanations for ing the origins of the concept itself and describes the theoretical explanations for how social capital is seen as influencing health in general and HIV in particular. how social capital is seen as influencing health in general and HIV in particular.

9 9 THEORETICAL FRAMEWORK THEORETICAL FRAMEWORK Theoretical framework Theoretical framework

The concept of social capital The concept of social capital The concept of social capital can be traced back to classical thinkers of the 19th The concept of social capital can be traced back to classical thinkers of the 19th century. According to Portes (1998) social capital originates from the works of century. According to Portes (1998) social capital originates from the works of Karl Marx and Emile Durkheim who discussed social capital in terms of com- Karl Marx and Emile Durkheim who discussed social capital in terms of community involvement and participation. In 1948 Marx wrote about class-conscious- munity involvement and participation. In 1948 Marx wrote about class-consciousness, pointing out that social capital is an important tool towards bringing about ness, pointing out that social capital is an important tool towards bringing about social change in the capitalist state. He argued that trust among workers could social change in the capitalist state. He argued that trust among workers could help in building class identity and identify common interests to fight for (Giddens, help in building class identity and identify common interests to fight for (Giddens, 1973). Durkheim (1951) provided another useful feature of social capital by point- 1973). Durkheim (1951) provided another useful feature of social capital by pointing out that group life was the basis for social cohesion. He differentiated between ing out that group life was the basis for social cohesion. He differentiated between two types of social networks based on mechanical or organic solidarity. A typical two types of social networks based on mechanical or organic solidarity. A typical pre-industrial society was characterized by social networks formed by members pre-industrial society was characterized by social networks formed by members who shared the same values, beliefs and norms, i.e mechanical solidarity. Or- who shared the same values, beliefs and norms, i.e mechanical solidarity. Or- ganic solidarity was found in the modern society where interest and roles differ ganic solidarity was found in the modern society where interest and roles differ but yet people create links and social networks based on their social class (Gid- but yet people create links and social networks based on their social class (Gid- dens, 1973; Haralambos & Holborn, 1990). dens, 1973; Haralambos & Holborn, 1990).

Other social scientists have attempted to apply the concept of social capital on Other social scientists have attempted to apply the concept of social capital on human relationships and socio-economic achievements in the society. Pierre human relationships and socio-economic achievements in the society. Pierre Bourdieu, James Coleman and Robert Putnam were the first scholars who revived Bourdieu, James Coleman and Robert Putnam were the first scholars who revived and made the concept popular as briefly discussed in the following section. and made the concept popular as briefly discussed in the following section.

Pierre Bourdieu Pierre Bourdieu Bourdieu gave the first systematic definition of social capital in the early 1980s Bourdieu gave the first systematic definition of social capital in the early 1980s (1986: 248), when he stated that it is “the aggregate of the actual or potential (1986: 248), when he stated that it is “the aggregate of the actual or potential resources which are linked to possession of a durable network of more or less resources which are linked to possession of a durable network of more or less institutionalized relationships of mutual acquaintance or recognition”. institutionalized relationships of mutual acquaintance or recognition”.

Bourdieu’s concept of social capital put emphasis on the role of social relations, Bourdieu’s concept of social capital put emphasis on the role of social relations, which is viewed through conflicts and power relations between individuals. For which is viewed through conflicts and power relations between individuals. For him, social relations empowered individuals and were seen as resources in the him, social relations empowered individuals and were seen as resources in the social struggle for better social conditions (Portes 1998; Siisiäinen, 2000). For social struggle for better social conditions (Portes 1998; Siisiäinen, 2000). For Bourdieu social capital is the existence of social networks consisting of group Bourdieu social capital is the existence of social networks consisting of group contact and membership, provision of support, access to resources and obliga- contact and membership, provision of support, access to resources and obliga- tions. Networks provide shared identity, possession of sociability skills and abil- tions. Networks provide shared identity, possession of sociability skills and ability to maintain the network. Bourdieu emphasized that social capital is created ity to maintain the network. Bourdieu emphasized that social capital is created and reconstructed by the individual. Like any other investment, community and reconstructed by the individual. Like any other investment, community members need to invest in social capital through adopting various strategies such members need to invest in social capital through adopting various strategies such

10 10 THEORETICAL FRAMEWORK THEORETICAL FRAMEWORK as participation in groups or associations (Bourdieu, 1985, 1986). According to as participation in groups or associations (Bourdieu, 1985, 1986). According to Bourdieu social capital has two basic elements. Firstly, resources in the social Bourdieu social capital has two basic elements. Firstly, resources in the social network are collectively owned by all members and not by the individual. Sec- network are collectively owned by all members and not by the individual. Sec- ondly, the type of benefits one may access through social capital differs according ondly, the type of benefits one may access through social capital differs according to the nature of the social network (Campbell et al, 2000; Macinko et al, 2001). to the nature of the social network (Campbell et al, 2000; Macinko et al, 2001).

James Coleman James Coleman In the late 1980’s, Coleman defined social capital as: “a particular kind of resource In the late 1980’s, Coleman defined social capital as: “a particular kind of resource available to an actor, comprising a variety of entities which contains two ele- available to an actor, comprising a variety of entities which contains two elements: they all consist of some aspect of social structures, and they facilitate ments: they all consist of some aspect of social structures, and they facilitate certain actions of actors … within the structure” (Coleman, 1988; p 98). certain actions of actors … within the structure” (Coleman, 1988; p 98).

According to Coleman (1990), social capital is both an individual and collective According to Coleman (1990), social capital is both an individual and collective asset. He refers to social capital as resources that an individual can acquire out asset. He refers to social capital as resources that an individual can acquire out of his or her social ties with other members in the communities. of his or her social ties with other members in the communities.

Coleman listed several mechanisms through which social capital can be gener- Coleman listed several mechanisms through which social capital can be generated including group enforcement, reciprocity and members’ expectation (Cole- ated including group enforcement, reciprocity and members’ expectation (Cole- man, 1988). This implies that individuals participate in relationships with an man, 1988). This implies that individuals participate in relationships with an expectation of getting back some rewards such as access to information or ma- expectation of getting back some rewards such as access to information or materials as a result of their participation or involvement (Coleman, 1990; Portes, terials as a result of their participation or involvement (Coleman, 1990; Portes, 1998; Macinko et al, 2001). Contrary to other scholars Coleman cautioned that 1998; Macinko et al, 2001). Contrary to other scholars Coleman cautioned that not all forms of social capital may result into positive results. Strong ties in a not all forms of social capital may result into positive results. Strong ties in a social group may for example benefit some members while prohibiting others social group may for example benefit some members while prohibiting others from accessing the same benefits (Coleman, 1990). from accessing the same benefits (Coleman, 1990).

Robert Putnam Robert Putnam The political scientist Robert Putnam has perhaps been the most influential in The political scientist Robert Putnam has perhaps been the most influential in conceptualizing and theorizing social capital. He wrote about the political and conceptualizing and theorizing social capital. He wrote about the political and economic situation in Italy and pointed out that the way society organized itself economic situation in Italy and pointed out that the way society organized itself into different social structures contributed largely towards differing levels of into different social structures contributed largely towards differing levels of achievements. He showed that regions with high civic participation had good achievements. He showed that regions with high civic participation had good performance in local government compared to regions with low level of civic performance in local government compared to regions with low level of civic participation (Hawe et al, 2000; Campbell et al, 2000). The presence of high participation (Hawe et al, 2000; Campbell et al, 2000). The presence of high level of trust and enforceable norms also increased the level of civic participation level of trust and enforceable norms also increased the level of civic participation (Portes, 1998; Campbell et al, 2000). (Portes, 1998; Campbell et al, 2000).

Putnam’s conceptualization of social capital means that a community with high Putnam’s conceptualization of social capital means that a community with high level of social capital has many social networks with high level of civic participation level of social capital has many social networks with high level of civic participation within these networks. Generalized norms of trust and reciprocal support among within these networks. Generalized norms of trust and reciprocal support among community members are also crucial (Campbell et al, 1999). community members are also crucial (Campbell et al, 1999).

11 11 THEORETICAL FRAMEWORK THEORETICAL FRAMEWORK

In summary, Bourdieu, Coleman and Putnam emphasize three important ele- In summary, Bourdieu, Coleman and Putnam emphasize three important elements of social capital theory. Firstly, social capital consists of social networks, ments of social capital theory. Firstly, social capital consists of social networks, people’s participation in organizations, and trust and reciprocity. Secondly, it is people’s participation in organizations, and trust and reciprocity. Secondly, it is a resource acquired by individuals as a result of relationships they build with a resource acquired by individuals as a result of relationships they build with others in the society. Thirdly, it consists of shared norms and values, which guide others in the society. Thirdly, it consists of shared norms and values, which guide and facilitate the functioning of the social networks and organizations for suc- and facilitate the functioning of the social networks and organizations for successful realization of opportunities and advantages. cessful realization of opportunities and advantages.

Critic against the concept of social capital Critic against the concept of social capital Many scholars have criticized that there is a lack of consensus on the basic defi- Many scholars have criticized that there is a lack of consensus on the basic definitions of social capital. They are concerned that the definitions cover many types nitions of social capital. They are concerned that the definitions cover many types of relationships at several levels making social capital vague, slippery and of relationships at several levels making social capital vague, slippery and poorly specified (Portes, 1998; Baum, 1999). In addition, Wills (2000) pointed poorly specified (Portes, 1998; Baum, 1999). In addition, Wills (2000) pointed out that Putnam’s work lack theoretical specificity by failing to show the link out that Putnam’s work lack theoretical specificity by failing to show the link between associational life, high social trust and better government. Another between associational life, high social trust and better government. Another criticism is based on Putnam’s assumption that social capital is a societal good criticism is based on Putnam’s assumption that social capital is a societal good that always yields positive results. Gilbert (2009) pointed out that even crime that always yields positive results. Gilbert (2009) pointed out that even crime gangs may involve well-organized networks with shared norms but they do not gangs may involve well-organized networks with shared norms but they do not yield positive results for society. Other identified negative results include restric- yield positive results for society. Other identified negative results include restric- tion of individual freedom and exclusion of outsiders (Portes, 1998). Also, there tion of individual freedom and exclusion of outsiders (Portes, 1998). Also, there is no consensus on how to operationalize and measure the concept of social is no consensus on how to operationalize and measure the concept of social capital. It is difficult to quantify non-observable variables such as norms and capital. It is difficult to quantify non-observable variables such as norms and shared values, which are used as proxies for social capital. shared values, which are used as proxies for social capital.

Despite these criticisms, the interest in the concept of social capital is growing Despite these criticisms, the interest in the concept of social capital is growing and many researchers have adopted it to explain and understand community and many researchers have adopted it to explain and understand community development (Portes, 1998; Hawe et al, 2000; Macinko et al, 2001). development (Portes, 1998; Hawe et al, 2000; Macinko et al, 2001).

Three approaches of social capital Three approaches of social capital The definitions of social capital are based on three approaches. The first views The definitions of social capital are based on three approaches. The first views social capital as an individual property, the second as a collective attribute and social capital as an individual property, the second as a collective attribute and the third is an integration of both individual and collective approaches. the third is an integration of both individual and collective approaches.

Individual social capital Individual social capital Bourdieu (1986), Lin (2001) and Flap (2004) view social capital as an individu- Bourdieu (1986), Lin (2001) and Flap (2004) view social capital as an individual resource that enables individuals to access resources that they could otherwise al resource that enables individuals to access resources that they could otherwise not access. An individual acquires social capital through participating in social not access. An individual acquires social capital through participating in social groups or networks and transforms the collective actions within the groups into groups or networks and transforms the collective actions within the groups into socio-economic benefits. The extent to which an individual has access to these socio-economic benefits. The extent to which an individual has access to these resources depends on the person’s connections, how much they interact through resources depends on the person’s connections, how much they interact through

12 12 THEORETICAL FRAMEWORK THEORETICAL FRAMEWORK the networks, the strengths of these networks, as well as the resources available the networks, the strengths of these networks, as well as the resources available in their connections (Sobel, 2002). in their connections (Sobel, 2002).

Collective social capital Collective social capital Putnam (1995, p 67) views social capital as a community level resource and defines Putnam (1995, p 67) views social capital as a community level resource and defines it as “features of social organization such as networks, norms and social trust it as “features of social organization such as networks, norms and social trust that facilitate coordination and cooperation for mutual benefit”. Coleman (1988) that facilitate coordination and cooperation for mutual benefit”. Coleman (1988) and Woolcock (1998) argued that social capital is a collective resource since it results and Woolcock (1998) argued that social capital is a collective resource since it results from members’ participation in various formal and informal organizations. How- from members’ participation in various formal and informal organizations. How- ever, Hyyppa and Maki (2003) claimed that even if social capital is regarded as a ever, Hyyppa and Maki (2003) claimed that even if social capital is regarded as a collective property, its impact can only be assessed at the individual level. collective property, its impact can only be assessed at the individual level.

Integrative approaches to social capital Integrative approaches to social capital The integrative approach combines features of individual and collective social The integrative approach combines features of individual and collective social capital. It is characterized by individual social connectedness, which is seen as a capital. It is characterized by individual social connectedness, which is seen as a foundation for collective social capital. Therefore social capital is viewed as a foundation for collective social capital. Therefore social capital is viewed as a resource generated from social ties and used by individuals or groups of networks. resource generated from social ties and used by individuals or groups of networks. This implies that social capital has both individual and collective benefits (Portes, This implies that social capital has both individual and collective benefits (Portes, 1998; Franke, 2005). 1998; Franke, 2005).

Classifications of social capital Classifications of social capital Social capital is commonly classified into structural and cognitive. According Social capital is commonly classified into structural and cognitive. According to Harpham and co-workers (2002), structural social capital includes the to Harpham and co-workers (2002), structural social capital includes the number of organizations individuals belong to and the extent to which indi- number of organizations individuals belong to and the extent to which individuals or members actively participate in these organizations or other social viduals or members actively participate in these organizations or other social activities. Hitt and colleagues (2002) viewed structural social capital as a basis activities. Hitt and colleagues (2002) viewed structural social capital as a basis for building relationships and cooperation as well as for facilitating mutual for building relationships and cooperation as well as for facilitating mutual benefits through collective actions. Cognitive social capital includes the re- benefits through collective actions. Cognitive social capital includes the resources accrued from people’s participation in formal and informal associations sources accrued from people’s participation in formal and informal associations including reciprocity, norms, and values. It also encompasses insights of trust, including reciprocity, norms, and values. It also encompasses insights of trust, and support and facilitates mutual benefits through collective actions (Krishna and support and facilitates mutual benefits through collective actions (Krishna & Uphoff, 2002). & Uphoff, 2002).

Structural social capital can be further sub-classified intobonding, bridging and Structural social capital can be further sub-classified intobonding, bridging and linking. Naraya and Pritchett (1997) defined bonding social capital as relations linking. Naraya and Pritchett (1997) defined bonding social capital as relations that include people from homogeneous formal and informal groups that are that include people from homogeneous formal and informal groups that are closely linked through networking. This is a type of social capital which people closely linked through networking. This is a type of social capital which people take part in daily life. It includes the family, churches, and other social groups in take part in daily life. It includes the family, churches, and other social groups in which people meet frequently (Campbell et al, 2000). Through these networks, which people meet frequently (Campbell et al, 2000). Through these networks, trust among members is created and enhanced. Also norms and codes of conduct trust among members is created and enhanced. Also norms and codes of conduct

13 13 THEORETICAL FRAMEWORK THEORETICAL FRAMEWORK are established to facilitate communication and responsibility within the social are established to facilitate communication and responsibility within the social network (Portes, 1998). network (Portes, 1998).

Bridging social capital is viewed as relations that are formed beyond the bound- Bridging social capital is viewed as relations that are formed beyond the boundaries of various social networks. These relations extend to individuals, formal aries of various social networks. These relations extend to individuals, formal and informal organizations that cut across different communities and individu- and informal organizations that cut across different communities and individuals (Naraya & Pritchett, 1997; Wallis et al, 1998; Putnam, 2000). als (Naraya & Pritchett, 1997; Wallis et al, 1998; Putnam, 2000).

According to Woolcock (2001), linking social capital involves the relationships According to Woolcock (2001), linking social capital involves the relationships between non-homogenous individuals and those who are outside their commu- between non-homogenous individuals and those who are outside their communities. It enables members to access a wider range of resources than those avail- nities. It enables members to access a wider range of resources than those available in the community. Baum and Ziersch (2003) referred linking social capital able in the community. Baum and Ziersch (2003) referred linking social capital to networks that cut across different ties and it is a vertical connection that cov- to networks that cut across different ties and it is a vertical connection that cov- ers different powers. Thus, linking social capital may help to reduce inequities ers different powers. Thus, linking social capital may help to reduce inequities by encouraging people to feel responsible for people beyond their boundaries. by encouraging people to feel responsible for people beyond their boundaries.

This thesis adopts the division of social capital into individual and collective This thesis adopts the division of social capital into individual and collective social capital. Individual social capital is further divided into its structural and social capital. Individual social capital is further divided into its structural and cognitive forms while collective social capital is divided into conventional and cognitive forms while collective social capital is divided into conventional and neighbourhood related social capital. This is in line with Rothstein and Stolle neighbourhood related social capital. This is in line with Rothstein and Stolle (2003) who define social capital both at individual and collective level. At indi- (2003) who define social capital both at individual and collective level. At individual level, they view social capital as the number of contacts a person has in vidual level, they view social capital as the number of contacts a person has in the society and the extent to which these contacts are beneficial to individuals. the society and the extent to which these contacts are beneficial to individuals. At the collective level the emphasis is on aggregated measurements of individu- At the collective level the emphasis is on aggregated measurements of individual level social contacts and networks. It also includes aggregated levels of trust al level social contacts and networks. It also includes aggregated levels of trust and the resources that are accessible collectively to all individuals in the society. and the resources that are accessible collectively to all individuals in the society. Kawachi et al (1997) underscored that collective social capital is equivalent to Kawachi et al (1997) underscored that collective social capital is equivalent to aggregated levels of individual social capital. Figure 2 depicts the different forms aggregated levels of individual social capital. Figure 2 depicts the different forms of social capital, and how they are conceptualized in this thesis. of social capital, and how they are conceptualized in this thesis.

14 14 THEORETICAL FRAMEWORK THEORETICAL FRAMEWORK

Social capital Social capital

Individual Collective Individual Collective

Neighborhood Neighborhood Structural Cognitive Conventional Structural Cognitive Conventional Related Related

Individual access: Individual access: -Informal -Informal organizations organizations -Formal -Bonding -Generalized trust -Formal -Bonding -Generalized trust organizations organizations -Participation -Bridging -Personalized trust -Participation -Bridging -Personalized trust in voluntary in voluntary activities -Linking -Reciprocity activities -Linking -Reciprocity

Aggregated levels of Individual Social capital Aggregated levels of Individual Social capital

Figure 2. Forms of social capital adapted from Nyqvist (2009) and Eriksson (2010). Figure 2. Forms of social capital adapted from Nyqvist (2009) and Eriksson (2010).

Social capital and health outcomes Social capital and health outcomes Several studies have found a positive relationship between social capital and Several studies have found a positive relationship between social capital and different dimensions of health. According to Veenstra (2000), community net- different dimensions of health. According to Veenstra (2000), community networks may provide an enabling environment in which health related information works may provide an enabling environment in which health related information could be diffused. Studies from the US (Kawachi et al, 1999; Subramanian et al, could be diffused. Studies from the US (Kawachi et al, 1999; Subramanian et al, 2002), using ‘aggregated’ individual social capital, found that living in areas with 2002), using ‘aggregated’ individual social capital, found that living in areas with high levels of social capital was strongly associated with individual well-being high levels of social capital was strongly associated with individual well-being and health. In a cross-sectional study of 75 neighborhoods in Malmö, Sweden, and health. In a cross-sectional study of 75 neighborhoods in Malmö, Sweden, Lindström et al (2004) found that ‘aggregated’ high levels of social capital (social Lindström et al (2004) found that ‘aggregated’ high levels of social capital (social participation) was significantly associated with individual self-reported good participation) was significantly associated with individual self-reported good health. Studies focusing on access to individual social capital in Västerbotten health. Studies focusing on access to individual social capital in Västerbotten County in Northern Sweden have shown that access to individual social capital County in Northern Sweden have shown that access to individual social capital increases the likelihood of having good self-rated health for both men and increases the likelihood of having good self-rated health for both men and women (Eriksson, 2010). women (Eriksson, 2010).

Even if social capital, both collective and individual, appears to influence health Even if social capital, both collective and individual, appears to influence health positively, the mechanisms of how social capital operates have been subjected to positively, the mechanisms of how social capital operates have been subjected to

15 15 THEORETICAL FRAMEWORK THEORETICAL FRAMEWORK much debate. Berkam and Glass (2001) argued that existing ties within social much debate. Berkam and Glass (2001) argued that existing ties within social networks may influence health through the social support that can be accessed networks may influence health through the social support that can be accessed by network members. These include emotional support whereby members may by network members. These include emotional support whereby members may share care, love and sympathy; person-to-person contact, which may promote share care, love and sympathy; person-to-person contact, which may promote or restrict exposure to infectious agents. Instrumental support, that may provide or restrict exposure to infectious agents. Instrumental support, that may provide material support and indirectly affecting the health of network members and material support and indirectly affecting the health of network members and information support where exchange of health information may facilitate healthy information support where exchange of health information may facilitate healthy behaviors. Getting access to material resource through networks may provide behaviors. Getting access to material resource through networks may provide opportunities for members to engage in economic activities that can directly or opportunities for members to engage in economic activities that can directly or indirectly influence health. Furthermore, Erickson (1988) suggestedsocial influ- indirectly influence health. Furthermore, Erickson (1988) suggestedsocial influence as a mechanism for social capital to influence health, where peers influence ence as a mechanism for social capital to influence health, where peers influence others through normative guidance. others through normative guidance.

Social capital may also have a negative effect on people’s health. Kawachi and Social capital may also have a negative effect on people’s health. Kawachi and Berkam (2001) pointed out that networks with strict norms may have destructive Berkam (2001) pointed out that networks with strict norms may have destructive mental health effects on members who fail to adhere to them. According to mental health effects on members who fail to adhere to them. According to Yamamura (2010), being excluded from networks may also have indirect negative Yamamura (2010), being excluded from networks may also have indirect negative effects on health. Lindstrom (2003) and Greiner et al (2004) reported that social effects on health. Lindstrom (2003) and Greiner et al (2004) reported that social participation in certain networks may encourage unhealthy behaviors, such as participation in certain networks may encourage unhealthy behaviors, such as taking alcohol and smoking. taking alcohol and smoking.

Social capital and HIV transmission Social capital and HIV transmission That social capital can be both health enhancing and health demaging is evident That social capital can be both health enhancing and health demaging is evident in studies on risk factors for HIV. Jesus et al (2002) indicated that social capital in studies on risk factors for HIV. Jesus et al (2002) indicated that social capital in its structural form may create norms and trust (cognitive social capital) which in its structural form may create norms and trust (cognitive social capital) which promote risk behaviors for HIV infection both in developed and developing promote risk behaviors for HIV infection both in developed and developing countries. Examples of networks that may have a negative impact are those in- countries. Examples of networks that may have a negative impact are those involving transactional sex and multiple partnering (Kiama, 1999; Pronyk, 2002; volving transactional sex and multiple partnering (Kiama, 1999; Pronyk, 2002; Epstein, 2007). Other studies have shown that social capital may positively influ- Epstein, 2007). Other studies have shown that social capital may positively influence and help contain the HIV epidemic by fostering positive norms of conduct ence and help contain the HIV epidemic by fostering positive norms of conduct (Pronyk, 2002; Campbell & MacPhail, 2002). In an exploratory study conducted (Pronyk, 2002; Campbell & MacPhail, 2002). In an exploratory study conducted in rural Zimbabwe, Gregson and co-researchers (2004) showed that participation in rural Zimbabwe, Gregson and co-researchers (2004) showed that participation in well functioning local community groups was positively associated with avoid- in well functioning local community groups was positively associated with avoidance of risk behaviors for HIV infection. The mechanisms have been described ance of risk behaviors for HIV infection. The mechanisms have been described to include high trust relationships, enhancing self-confidence and self-esteem. to include high trust relationships, enhancing self-confidence and self-esteem. These are regarded as important tools for making decisions and negotiating about These are regarded as important tools for making decisions and negotiating about sexual relations. Social capital has also been shown to facilitate access to infor- sexual relations. Social capital has also been shown to facilitate access to information and knowledge about HIV/AIDS and HIV testing, needed for shaping mation and knowledge about HIV/AIDS and HIV testing, needed for shaping HIV protective behavior (Jamil & Muriisa, 2004; Pronyk et al, 2008). Campbell HIV protective behavior (Jamil & Muriisa, 2004; Pronyk et al, 2008). Campbell (2003) suggested that lack of social capital may intensify the HIV epidemic since (2003) suggested that lack of social capital may intensify the HIV epidemic since

16 16 THEORETICAL FRAMEWORK THEORETICAL FRAMEWORK those outside networks may have norms resulting into less interconnectedness. those outside networks may have norms resulting into less interconnectedness. Such norms may hinder the effect of HIV interventions by encouraging for ex- Such norms may hinder the effect of HIV interventions by encouraging for example the use of illicit drugs and exessive alcohol drinking that are regarded as ample the use of illicit drugs and exessive alcohol drinking that are regarded as risk factors for HIV infection. risk factors for HIV infection.

The possible negative effects of social capital are seen also in relation to sexual The possible negative effects of social capital are seen also in relation to sexual behaviors. Associations or networks with strict rules, unaffordable fees or behaviors. Associations or networks with strict rules, unaffordable fees or contributions may exclude some community members to join. This may encour- contributions may exclude some community members to join. This may encourage excluded community members to form own groups with norms facilitating age excluded community members to form own groups with norms facilitating HIV transmission (Baum, 1999). Excluding community members from networks HIV transmission (Baum, 1999). Excluding community members from networks due to ecoomic constraints may also enhance inequity and increase the risk for due to ecoomic constraints may also enhance inequity and increase the risk for HIV in specifically vulnerable groups (Turmen, 2003). Women may more be HIV in specifically vulnerable groups (Turmen, 2003). Women may more be affected than men since their power to negotiate safe sex with their sexual part- affected than men since their power to negotiate safe sex with their sexual partners is weaker (Heise & Elias, 1995). ners is weaker (Heise & Elias, 1995).

17 17 MATERIALS AND METHODS MATERIALS AND METHODS Materials and methods Materials and methods

A multi-methodological approach A multi-methodological approach The multi-dimensionality of the concept of social capital (social interactions in The multi-dimensionality of the concept of social capital (social interactions in groups and networks, trust, norms and reciprocity) calls for employment of a groups and networks, trust, norms and reciprocity) calls for employment of a combination of qualitative and quantitative methods. The methods are seen as combination of qualitative and quantitative methods. The methods are seen as complementary allowing, the researcher to uncover the mechanisms as well as complementary allowing, the researcher to uncover the mechanisms as well as the associations between different dimensions of social capital and health in the associations between different dimensions of social capital and health in various local contexts (Harpham et al, 2002; Dudwicket al, 2006). In this thesis various local contexts (Harpham et al, 2002; Dudwicket al, 2006). In this thesis qualitative methods were used to deepen the understanding of how people’s qualitative methods were used to deepen the understanding of how people’s experiences and involvement in social networks (social capital) may influence experiences and involvement in social networks (social capital) may influence their HIV risk behaviors and quantitative methods to illustrate how this involve- their HIV risk behaviors and quantitative methods to illustrate how this involvement is distributed in the population and eventually influence health behaviors ment is distributed in the population and eventually influence health behaviors and the risk of being HIV infected. In addition, the qualitative sub-study in- and the risk of being HIV infected. In addition, the qualitative sub-study included triangulation in data collection. Focus group discussions were performed cluded triangulation in data collection. Focus group discussions were performed to capture community norms related to social capital while key informant inter- to capture community norms related to social capital while key informant interviews and documentary reviews were used to describe in detail the existing social views and documentary reviews were used to describe in detail the existing social capital in the study area (Dahlgren et al, 2007). Table 2 gives an overview of the capital in the study area (Dahlgren et al, 2007). Table 2 gives an overview of the thesis structure indicating sub-studies, overall aims, study design, study sample, thesis structure indicating sub-studies, overall aims, study design, study sample, sources of data, analytical approaches, study period and corresponding papers. sources of data, analytical approaches, study period and corresponding papers.

18 18 MATERIALS AND METHODS MATERIALS AND METHODS I I

II II III III IV IV Paper Paper Paper period Study Sept - Oct 2008 - May March 2010 July - Oct 2010 period Study Sept - Oct 2008 - May March 2010 July - Oct 2010 - - Analytical methods Case analysis theory Grounded logis Multivariable tic regression Multivariable regression logistic analysis Analytical methods Case analysis theory Grounded logis Multivariable tic regression Multivariable regression logistic analysis ------sources Data group Focus discussions informant Key interviews Semi-struc question tured naire Semi-struc ques tured and tionnaire test serological HIV for results infection sources Data group Focus discussions informant Key interviews Semi-struc question tured naire Semi-struc ques tured and tionnaire test serological HIV for results infection - - - -

sample Study mem 120 community bers informants 29 key (leaders of formal/ organizations informal and local leaders) mem community 1301 bers aged 15-64 years 3422 community members aged 15-54 years sample Study mem 120 community bers informants 29 key (leaders of formal/ organizations informal and local leaders) mem community 1301 bers aged 15-64 years 3422 community members aged 15-54 years

design Study Qualitative Quantitative: cross-sectional II) (sub-study Quantitative: cross-sectional III) (sub-study design Study Qualitative Quantitative: cross-sectional II) (sub-study Quantitative: cross-sectional III) (sub-study ------Overview of the thesis: sub-study, aims, study design, sample, data sources, analytical methods, period and Overview of the thesis: sub-study, aims, study design, sample, data sources, analytical methods, period and Aims and describe explore To the mechanisms of how and cognitive structural influ social capital may of the transmission ence HIV/AIDS the associa determine To structural tion between social and cognitive risk be capital, HIV/AIDS and HIV infection haviors Aims and describe explore To the mechanisms of how and cognitive structural influ social capital may of the transmission ence HIV/AIDS the associa determine To structural tion between social and cognitive risk be capital, HIV/AIDS and HIV infection haviors Sub- study I II & III Sub- study I II & III Table 2. corresponding papers. Table 2. corresponding papers.

19 19 MATERIALS AND METHODS MATERIALS AND METHODS

Study area Study area The study was conducted in the Kagera region situated in the northwestern cor- The study was conducted in the Kagera region situated in the northwestern cor- ner of Tanzania. The regional capital is Bukoba town. The region shares borders ner of Tanzania. The regional capital is Bukoba town. The region shares borders with Uganda to the North, Rwanda and Burundi to the West, the Kigoma and with Uganda to the North, Rwanda and Burundi to the West, the Kigoma and Mwanza regions to the South and Lake Victoria to the East. Kagera region com- Mwanza regions to the South and Lake Victoria to the East. Kagera region com- prises of seven administrative districts: Bukoba urban, Bukoba rural, Muleba, prises of seven administrative districts: Bukoba urban, Bukoba rural, Muleba, Karagwe, Biharamulo, Chato and Ngara. The 2002 National Census showed that Karagwe, Biharamulo, Chato and Ngara. The 2002 National Census showed that Kagera region had a population of about 2 million people with an annual growth Kagera region had a population of about 2 million people with an annual growth rate of 3.1% (NBS, 2003). The region has witnessed an overall decline in HIV rate of 3.1% (NBS, 2003). The region has witnessed an overall decline in HIV prevalence from 9.6% in 1987 to 3.4% in 2008 (TACAIDS, 2008). prevalence from 9.6% in 1987 to 3.4% in 2008 (TACAIDS, 2008).

As described earlier the KARP project has closely followed the HIV/AIDS epi- As described earlier the KARP project has closely followed the HIV/AIDS epidemic in the area since the late 1980’ies. Repeated cross-sectional studies have demic in the area since the late 1980’ies. Repeated cross-sectional studies have created a base for dividing the region into three prevalence zones; high, medium created a base for dividing the region into three prevalence zones; high, medium and low. In the urban high prevalence zone the prevalence declined from 24% in and low. In the urban high prevalence zone the prevalence declined from 24% in 1987, to 18% in 1993, 13% in 1996 and 8.2 % in 2004. In the peri-urban medium 1987, to 18% in 1993, 13% in 1996 and 8.2 % in 2004. In the peri-urban medium prevalence zone it declined from 10% in 1987, to 6.8% in 1996 and 4.3% in 1999 prevalence zone it declined from 10% in 1987, to 6.8% in 1996 and 4.3% in 1999 and in the rural low prevalence zone the prevalence declined from 4.5% in 1987 and in the rural low prevalence zone the prevalence declined from 4.5% in 1987 to 2.6% in 1999. The decrease in HIV prevalence was supported by a decline in to 2.6% in 1999. The decrease in HIV prevalence was supported by a decline in incidence rate; from 48/1000 in 1989 to 9.1/1000 persons at risk in 1996 and incidence rate; from 48/1000 in 1989 to 9.1/1000 persons at risk in 1996 and from 8.2/1000 in 1989 to 3.9/100 persons at risk in 1999/2000 in the high and from 8.2/1000 in 1989 to 3.9/100 persons at risk in 1999/2000 in the high and medium HIV prevalence zones respectively (Killewo et al, 1990, 1993; Kwesiga- medium HIV prevalence zones respectively (Killewo et al, 1990, 1993; Kwesiga- bo, 2001; TACAIDS, 2008). bo, 2001; TACAIDS, 2008).

For the qualitative sub-study three case communities from the three districts of For the qualitative sub-study three case communities from the three districts of Bukoba rural, Bukoba urban and Chato (Figure 3) were selected to represent Bukoba rural, Bukoba urban and Chato (Figure 3) were selected to represent high, medium and low HIV prevalence zones. In each identified zone, one ward high, medium and low HIV prevalence zones. In each identified zone, one ward was randomly selected followed by purposive selection of one community. The was randomly selected followed by purposive selection of one community. The selection of the case communities were done in collaboration with the ward lead- selection of the case communities were done in collaboration with the ward leaders who were asked to identify all existing formal and informal organizations op- ers who were asked to identify all existing formal and informal organizations op- erating in their wards to facilitate the selection of a community where all types of erating in their wards to facilitate the selection of a community where all types of organizations were represented. The selected communities were then regarded as organizations were represented. The selected communities were then regarded as “typical” communities by having an average number of both formal and informal “typical” communities by having an average number of both formal and informal organizations. organizations.

The Bukoba urban district, that represented the high HIV prevalence zone, was The Bukoba urban district, that represented the high HIV prevalence zone, was also selected for the cross-sectional survey to determine the association between also selected for the cross-sectional survey to determine the association between social capital and HIV infection. social capital and HIV infection.

20 20 MATERIALS AND METHODS MATERIALS AND METHODS

Bukoba Urban district Bukoba Urban district

Bukoba Rural district Bukoba Rural district

Chato district Chato district

Figure 3. Map of Kagera region, showing the location of studied com- Figure 3. Map of Kagera region, showing the location of studied communities. munities.

Study populations and sample size Study populations and sample size For the qualitative studies (sub-study I) 29 key informants were purposively For the qualitative studies (sub-study I) 29 key informants were purposively selected to represent community leadership, existing organizations and networks. selected to represent community leadership, existing organizations and networks. In total five community leaders, six leaders of faith based organisations and 18 In total five community leaders, six leaders of faith based organisations and 18 leaders of informal organizations were interviewed. leaders of informal organizations were interviewed.

Fifteen focus group discussions (FGDs) were conducted involving 120 members Fifteen focus group discussions (FGDs) were conducted involving 120 members and non-members of informal organizations. The FGD participants were se- and non-members of informal organizations. The FGD participants were selected to capture the wide range of experiences among community members of lected to capture the wide range of experiences among community members of existing forms of organizations and the possible influence on their health par- existing forms of organizations and the possible influence on their health particularly sexually related health behaviors. In the urban and rural communities ticularly sexually related health behaviors. In the urban and rural communities we had four FGDs in each community: one male youth group, one female youth we had four FGDs in each community: one male youth group, one female youth group aged 24-34, one male adult group and one female adults group aged 35-64 group aged 24-34, one male adult group and one female adults group aged 35-64 years. In the peri-urban community, seven FGDs were conducted comprising of years. In the peri-urban community, seven FGDs were conducted comprising of one female youth group, two male youth groups, two male adults groups and two one female youth group, two male youth groups, two male adults groups and two female adults groups. female adults groups.

21 21 MATERIALS AND METHODS MATERIALS AND METHODS

The community survey (sub-study II) focused on determining the association The community survey (sub-study II) focused on determining the association between social capital and HIV risk behaviors. The survey was conducted in the between social capital and HIV risk behaviors. The survey was conducted in the same three communities where the qualitative studies (sub-study I) had been same three communities where the qualitative studies (sub-study I) had been performed. A total of 1339 15-64 year old adults in these communities were vis- performed. A total of 1339 15-64 year old adults in these communities were visited for an interview. Of these, 1301 (97%) agreed to be interviewed. Sex repre- ited for an interview. Of these, 1301 (97%) agreed to be interviewed. Sex repre- sentation in the study sample differed between the communities. There were sentation in the study sample differed between the communities. There were more women (63%) in the urban community compared to the peri-urban (57%) more women (63%) in the urban community compared to the peri-urban (57%) and the rural (42%) communities. This is because there were more households and the rural (42%) communities. This is because there were more households with women only family members (19%) compared to households with men only with women only family members (19%) compared to households with men only (13%) in the urban community. (13%) in the urban community.

The cross-sectional study (sub-study III) was conducted in the urban district. A The cross-sectional study (sub-study III) was conducted in the urban district. A total of 3586 adults aged between 15 and 54 were interviewed. Of these, 3422 total of 3586 adults aged between 15 and 54 were interviewed. Of these, 3422 (95%) agreed to participate in the study (58% women and 42% men). (95%) agreed to participate in the study (58% women and 42% men).

Sampling procedures Sampling procedures Following written permission from both the regional and district levels the re- Following written permission from both the regional and district levels the research team visited the study area and held meetings with community leaders to search team visited the study area and held meetings with community leaders to explain the nature of the study, its objectives as well the importance of having explain the nature of the study, its objectives as well the importance of having informed consent from each participant. These leaders were also asked to assist informed consent from each participant. These leaders were also asked to assist in purposively selecting eligible participants for the FGDs as well as in identifying in purposively selecting eligible participants for the FGDs as well as in identifying and visiting formal organizations working on HIV related activities in the study and visiting formal organizations working on HIV related activities in the study area. One leader from each organization was selected, and asked to be interviewed area. One leader from each organization was selected, and asked to be interviewed as a key informant. as a key informant.

For the cross-sectional studies the researchers sought permission from the re- For the cross-sectional studies the researchers sought permission from the regional, district, and community authorities. They were informed about the purpose gional, district, and community authorities. They were informed about the purpose of the study and asked for cooperation and continuing support. Community guides of the study and asked for cooperation and continuing support. Community guides were appointed by the leaders to guide the research team in visiting households. were appointed by the leaders to guide the research team in visiting households. The research team made a list of eligible study participants and one individual was The research team made a list of eligible study participants and one individual was randomly selected to participate in the study. For both sub-studies selected indi- randomly selected to participate in the study. For both sub-studies selected individuals were asked to give informed consent to participate in the studies. viduals were asked to give informed consent to participate in the studies.

22 22 MATERIALS AND METHODS MATERIALS AND METHODS

Training of field assistants. Training of field assistants.

Focus group discussion with young men. Focus group discussion with young men.

Interviewing in privacy. Interviewing in privacy.

23 23 MATERIALS AND METHODS MATERIALS AND METHODS

Measure instruments Measure instruments

Forms and characteristics of social capital and the mechanisms through Forms and characteristics of social capital and the mechanisms through which social capital may influence HIV transmission (Paper I and II) which social capital may influence HIV transmission (Paper I and II) An interview guide for both focus group discussion (FGD) and key informant An interview guide for both focus group discussion (FGD) and key informant interviews formed the basis for collecting information on structural and cognitive interviews formed the basis for collecting information on structural and cognitive forms of social capital. The interview guide included questions based on Putnam’s forms of social capital. The interview guide included questions based on Putnam’s framework of understanding social capital. This implied discussing people’s framework of understanding social capital. This implied discussing people’s engagement in civic organizations and voluntary activities, their views on insti- engagement in civic organizations and voluntary activities, their views on institutional and organizational trust and reciprocity and strategies used to build trust tutional and organizational trust and reciprocity and strategies used to build trust and reputation. The guide also included questions about HIV and AIDS preven- and reputation. The guide also included questions about HIV and AIDS preventive activities and care seeking behavior (Table 3). tive activities and care seeking behavior (Table 3).

Most interviews and discussions were conducted in local government offices and Most interviews and discussions were conducted in local government offices and lasted between 1 and 2 hours. The FGDs consisted of 6-8 participants and they lasted between 1 and 2 hours. The FGDs consisted of 6-8 participants and they were homogeneous in terms of sex and age to enhance openness during the were homogeneous in terms of sex and age to enhance openness during the discussion. An emergent design was adopted whereby the results from the first discussion. An emergent design was adopted whereby the results from the first group determined what to focus specifically on in the forthcoming ones. All dis- group determined what to focus specifically on in the forthcoming ones. All discussions and interviews were tape recorded, transcribed verbatim and trans- cussions and interviews were tape recorded, transcribed verbatim and translated into English, to facilitate joint analysis by the research team. lated into English, to facilitate joint analysis by the research team.

Social capital and HIV risk behaviors (paper III) Social capital and HIV risk behaviors (paper III) The questionnaire used in the first survey within the KARP project in 1987 was The questionnaire used in the first survey within the KARP project in 1987 was modified and updated for use in sub-study II. This community survey aimed at modified and updated for use in sub-study II. This community survey aimed at determining the association between social capital and HIV risk behaviors. The determining the association between social capital and HIV risk behaviors. The questions were constructed to measure individual and collective social capital questions were constructed to measure individual and collective social capital and were based on indicators such as trust and reciprocity (cognitive social and were based on indicators such as trust and reciprocity (cognitive social capital) and membership in social networks and organizations and voluntary capital) and membership in social networks and organizations and voluntary participation in public activities (structural social capital). The questionnaire also participation in public activities (structural social capital). The questionnaire also included questions on HIV risk behavior: “condom use with casual sex partner” included questions on HIV risk behavior: “condom use with casual sex partner” and “testing for HIV infection”. Questions on socio-demographic and socio- and “testing for HIV infection”. Questions on socio-demographic and socio- economic factors such as sex, age, level of education, marital status, religion, economic factors such as sex, age, level of education, marital status, religion, accessibility to water source, type of building materials used in dwellings and accessibility to water source, type of building materials used in dwellings and type of assets possessed were also included in the questionnaire. type of assets possessed were also included in the questionnaire.

Social capital and HIV infection (paper IV) Social capital and HIV infection (paper IV) For the prevalence study (sub-study III) a questionnaire was designed to determine For the prevalence study (sub-study III) a questionnaire was designed to determine the association between social capital and HIV infection in the urban district with the association between social capital and HIV infection in the urban district with an observed steep decline in HIV prevalence. Like in sub-study II, the questions an observed steep decline in HIV prevalence. Like in sub-study II, the questions included in the questionnaire contained several sections: socio-demographic and included in the questionnaire contained several sections: socio-demographic and

24 24 MATERIALS AND METHODS MATERIALS AND METHODS economic characteristics and measurement of social capital. A blood sample was economic characteristics and measurement of social capital. A blood sample was collected from each consenting individual and tested for HIV-1 antibodies. collected from each consenting individual and tested for HIV-1 antibodies.

Table 3. Key issues guiding key informant interviews and focus group discussions (FGD). Table 3. Key issues guiding key informant interviews and focus group discussions (FGD).

Key issue Interviews FGDs Key issue Interviews FGDs HIV and AIDS related activities focusing on: health education, √ HIV and AIDS related activities focusing on: health education, √ treatment of STIs, procurement of condoms, promotion and treatment of STIs, procurement of condoms, promotion and distribution of condoms, VCT, reduction of multiple sex part- distribution of condoms, VCT, reduction of multiple sex partners, material support to vulnerable population, and duration ners, material support to vulnerable population, and duration and intensity of the HIV and AIDS activities and intensity of the HIV and AIDS activities Strategies used to build trust and reputation √ √ Strategies used to build trust and reputation √ √ Number of organizations and social groups in the community √ √ Number of organizations and social groups in the community √ √ and their activities related to care of the sick, protection of and their activities related to care of the sick, protection of vulnerable populations, support to bereavement and economic vulnerable populations, support to bereavement and economic livelihood livelihood Institutional and organizational trust and reputation √ √ Institutional and organizational trust and reputation √ √ General activities of organizations, social groups and networks √ General activities of organizations, social groups and networks √ and specific activities targeting HIV prevention and specific activities targeting HIV prevention Ranking HIV related activities according to their importance in √ Ranking HIV related activities according to their importance in √ changing sexual behavior paying particular attention to care changing sexual behavior paying particular attention to care of the sick, protection of vulnerable populations, support to of the sick, protection of vulnerable populations, support to bereavement and economic bereavement and economic

Measurement of social capital Measurement of social capital We reviewed the literature to identify indicators of social capital that could be We reviewed the literature to identify indicators of social capital that could be applicable in our local social and cultural context (Grootaert & Bastelaer, 2001; applicable in our local social and cultural context (Grootaert & Bastelaer, 2001; Narayan & Cassidy, 2001; Stones & Hughes, 2002; Harper, 2002; Grootaert et Narayan & Cassidy, 2001; Stones & Hughes, 2002; Harper, 2002; Grootaert et al, 2004; O’Briena et al, 2004; Dudwick et al, 2006). Based on this review and al, 2004; O’Briena et al, 2004; Dudwick et al, 2006). Based on this review and our experiences from the qualitative study we decided which proxy indicators of our experiences from the qualitative study we decided which proxy indicators of social capital to include in sub-study II and III. We adopted a tool consisting of social capital to include in sub-study II and III. We adopted a tool consisting of six dimensions for measuring both collective and individual social capital as well six dimensions for measuring both collective and individual social capital as well as its structural and cognitive forms namely membership in groups and or- as its structural and cognitive forms namely membership in groups and organizations, neighborhood characteristics, social networks and support, trust ganizations, neighborhood characteristics, social networks and support, trust and solidarity, collective action and cooperation, and social cohesion and inclu- and solidarity, collective action and cooperation, and social cohesion and inclusion. The details of the questions used are given in papers III and IV. sion. The details of the questions used are given in papers III and IV. The study team The study team The study team consisted of researchers and trained research assistants. The re- The study team consisted of researchers and trained research assistants. The research assistants were nurses, well experienced in the field of HIV/AIDS counseling search assistants were nurses, well experienced in the field of HIV/AIDS counseling and treatment of patients and they knew the local language and culture. The Re- and treatment of patients and they knew the local language and culture. The Re- gional Medical Office appointed one medical doctor to coordinate and oversee the gional Medical Office appointed one medical doctor to coordinate and oversee the project field work activities, two laboratory technicians, ten nurses and one driver project field work activities, two laboratory technicians, ten nurses and one driver while the Bukoba urban medical office provided one nurse for the project. Despite while the Bukoba urban medical office provided one nurse for the project. Despite

25 25 MATERIALS AND METHODS MATERIALS AND METHODS their medical background and counseling experiences the research assistants were their medical background and counseling experiences the research assistants were given additional one week training with lectures, discussions and demonstrations given additional one week training with lectures, discussions and demonstrations on pre- and post counseling. They were also trained on social capital concepts, on pre- and post counseling. They were also trained on social capital concepts, interview techniques, and how to draw, store and transport blood samples. interview techniques, and how to draw, store and transport blood samples.

Laboratory procedures (for sub-study III) Laboratory procedures (for sub-study III) Blood samples were collected aseptically in 5 ml red top vacutainers (BD, NJ, Blood samples were collected aseptically in 5 ml red top vacutainers (BD, NJ, USA) and left to clot. Sera specimens were separated after centrifugation, ali- USA) and left to clot. Sera specimens were separated after centrifugation, ali- quoted into 2 ml cryotubes tubes (Nalge Nunc International, IL, and USA) and quoted into 2 ml cryotubes tubes (Nalge Nunc International, IL, and USA) and stored at -20°C until the time for assay. stored at -20°C until the time for assay.

HIV status was determined by ELISA tests. Abbott Murex Wellcozyme anti-HIV-1 HIV status was determined by ELISA tests. Abbott Murex Wellcozyme anti-HIV-1 recombinant was used as first ELISA. Specimens with negative results underwent recombinant was used as first ELISA. Specimens with negative results underwent no further testing and were considered negative. Reactive samples were retested no further testing and were considered negative. Reactive samples were retested by second ELISA test (Dade Behring Enzygnost anti-HIV-1/2). This assay detects by second ELISA test (Dade Behring Enzygnost anti-HIV-1/2). This assay detects both HIV-1/2 infections. All samples that were reactive on first and second ELISA both HIV-1/2 infections. All samples that were reactive on first and second ELISA were regarded as positive for IgG anti HIV antibodies. Inno-Lia HIV I/II immu- were regarded as positive for IgG anti HIV antibodies. Inno-Lia HIV I/II immu- noblot assay (Immunogenetics) was used as a reference method. Discrepant results noblot assay (Immunogenetics) was used as a reference method. Discrepant results between the two ELISAs were confirmed by western blot using Inno-Lia HIV I/II between the two ELISAs were confirmed by western blot using Inno-Lia HIV I/II assay. assay.

Ethical considerations Ethical considerations The KARP has ethical clearance from the research and ethical clearance commit- The KARP has ethical clearance from the research and ethical clearance commit- tee of the MoHSW in Tanzania to implement the project “Epidemiology towards tee of the MoHSW in Tanzania to implement the project “Epidemiology towards evaluation of interventions and monitoring of HIV infection in the Kagera region evaluation of interventions and monitoring of HIV infection in the Kagera region of Tanzania”. As a requirement the project submits regular progress reports to of Tanzania”. As a requirement the project submits regular progress reports to the MoHSW and provides feedback to the Kagera regional and district authorities the MoHSW and provides feedback to the Kagera regional and district authorities about the research results. about the research results.

At the community level, ward and village/street leaders were informed about the At the community level, ward and village/street leaders were informed about the study aims, the data collection procedures and their permission was sought and study aims, the data collection procedures and their permission was sought and given. Informed consent was also sought from the study participants after given. Informed consent was also sought from the study participants after explaining to them the purpose, the methods and the data collection procedures. explaining to them the purpose, the methods and the data collection procedures. The field assistants conducted pre- and post test counseling in connection with The field assistants conducted pre- and post test counseling in connection with collecting blood samples and returning the HIV test results respectively. collecting blood samples and returning the HIV test results respectively.

Data analysis Data analysis We performed a case study analysis (Creswell, 1998) based on various sources We performed a case study analysis (Creswell, 1998) based on various sources of information to describe and characterize the social capital in the study area of information to describe and characterize the social capital in the study area (Paper I). A list of activities for each formal and informal organization was cre- (Paper I). A list of activities for each formal and informal organization was cre-

26 26 MATERIALS AND METHODS MATERIALS AND METHODS ated and each activity was given a score. The activity scores helped us to calculate ated and each activity was given a score. The activity scores helped us to calculate the average number of activities performed in each case community, which was the average number of activities performed in each case community, which was later translated into levels of interaction. A community that scored high on in- later translated into levels of interaction. A community that scored high on interaction was regarded as a community with high levels of social capital. The lists teraction was regarded as a community with high levels of social capital. The lists also helped to distingish between structural (bridging, bonding, linking) and also helped to distingish between structural (bridging, bonding, linking) and cognitive social capital. For each form of social capital, a community could score cognitive social capital. For each form of social capital, a community could score 1-3. The scores for bonding social capital were based on the number of social 1-3. The scores for bonding social capital were based on the number of social groups and networks that were formed by community members of the same groups and networks that were formed by community members of the same neighborhood and their ways of interacting. The scores for bridging social capi- neighborhood and their ways of interacting. The scores for bridging social capital were based on the number of social groups and networks that recruited mem- tal were based on the number of social groups and networks that recruited members from neighboring communities and their interaction while for linking social bers from neighboring communities and their interaction while for linking social capital the scores were based on the number of social groups and networks that capital the scores were based on the number of social groups and networks that had a direct linkage with formal organizations and how they worked together. had a direct linkage with formal organizations and how they worked together. The scores for cognitive social capital were based on the reported level of trust, The scores for cognitive social capital were based on the reported level of trust, rules and regulations guiding social groups and networks and reciprocal relations rules and regulations guiding social groups and networks and reciprocal relations in the case communities. The scoring, together with additional information from in the case communities. The scoring, together with additional information from the focus group discussions and the key informant interviews, made possible a the focus group discussions and the key informant interviews, made possible a detailed description of the role and functions of different forms of social capital detailed description of the role and functions of different forms of social capital in this specific context. in this specific context.

A grounded theory approach (Strauss & Corbin, 1990) was used to analyze the A grounded theory approach (Strauss & Corbin, 1990) was used to analyze the mechanisms of how social capital may have influenced the HIV transmission in mechanisms of how social capital may have influenced the HIV transmission in the study area (Paper II). Grounded theory is “a qualitative strategy of inquiry in the study area (Paper II). Grounded theory is “a qualitative strategy of inquiry in which the researcher derives a general, abstract theory of process, action, or which the researcher derives a general, abstract theory of process, action, or interaction grounded in the views of participants in a study” (Creswell, 2009, p interaction grounded in the views of participants in a study” (Creswell, 2009, p 13 & 229). The process involves a series of actions including emerging data col- 13 & 229). The process involves a series of actions including emerging data collection for refining interrelationships of categories and sub-categories of informa- lection for refining interrelationships of categories and sub-categories of information about certain phenomenon (Strauss & Corbin, 1990, 1998; Charmaz, 2006). tion about certain phenomenon (Strauss & Corbin, 1990, 1998; Charmaz, 2006). Table 4 shows how codes and concepts were identified and assigned line by line Table 4 shows how codes and concepts were identified and assigned line by line to segments of the text. The software Open Code facilitated the coding process to segments of the text. The software Open Code facilitated the coding process and helped in comparing codes to sort out similarities and differences as well as and helped in comparing codes to sort out similarities and differences as well as to find general patterns. Then axial coding was conducted to identify categories to find general patterns. Then axial coding was conducted to identify categories and sub-categories and to relate them to each other through an inductive think- and sub-categories and to relate them to each other through an inductive thinking. For each form of social capital (structural and cognitive) we developed three ing. For each form of social capital (structural and cognitive) we developed three main categories and three sub-categories to illustrate theoretically the general main categories and three sub-categories to illustrate theoretically the general role and mechanisms through which they may influence sexual behavior. role and mechanisms through which they may influence sexual behavior.

27 27 MATERIALS AND METHODS MATERIALS AND METHODS

Table 4. An example of line by line coding of text data and development of sub- Table 4. An example of line by line coding of text data and development of sub- categories and categories. categories and categories.

Text Code Sub-category Category Text Code Sub-category Category Rules of conduct Creates social Fosters Rules of conduct Creates social Fosters “Frankly speaking, in our group we “Frankly speaking, in our group we Exclusion of pe- pressure to strict Exclusion of pe- pressure to strict do not accept any members who do not accept any members who ople misbehaving take responsi- behavior ople misbehaving take responsi- behavior take excessive alcohol” take excessive alcohol” bility conduct bility conduct “You know Some women who Possibility for Enables socio- Serves the “You know Some women who Possibility for Enables socio- Serves the were involved in extra marital sex- women to end economic needs of were involved in extra marital sex- women to end economic needs of ual affairs for exchange of money commercialized empower- vulnerable ual affairs for exchange of money commercialized empower- vulnerable to cover some of their basic needs sex ment groups to cover some of their basic needs sex ment groups have stopped this behavior since Giving basic have stopped this behavior since Giving basic they can now get some money needs/ support they can now get some money needs/ support through social groups” through social groups” “Faith-based organizations Solidarity and Enhanced Fuels “Faith-based organizations Solidarity and Enhanced Fuels supported our group by giving caring for each togetherness religious supported our group by giving caring for each togetherness religious it a cow to keep but emphasized other emphasized norms and it a cow to keep but emphasized other emphasized norms and about solidarity and caring for values about solidarity and caring for values one another.” one another.” “Most women join into social Provide opportu- Enables socio- Increases “Most women join into social Provide opportu- Enables socio- Increases groups or networks, they do not nity for women economic women’s groups or networks, they do not nity for women economic women’s want to be lazy and dependent on Reduce depen- empower- enrolement want to be lazy and dependent on Reduce depen- empower- enrolement their husbands” dency ment their husbands” dency ment

In the second sub-study, we selected eight and nine variables that could be linked In the second sub-study, we selected eight and nine variables that could be linked to theoretical ideas on collective social capital (based on conventional and neigh- to theoretical ideas on collective social capital (based on conventional and neighborhood related variables) and individual social capital (based on structural and borhood related variables) and individual social capital (based on structural and cognitive forms). A principal component analysis (PCA) was performed on these cognitive forms). A principal component analysis (PCA) was performed on these variables separately, one for measuring collective social capital and the other for variables separately, one for measuring collective social capital and the other for individual social capital. The aim of the factor analysis was to reduce the number individual social capital. The aim of the factor analysis was to reduce the number of variables that were collected in our study and to detect or classify variables into of variables that were collected in our study and to detect or classify variables into one index or locating clusters of variables that are related to each other and used one index or locating clusters of variables that are related to each other and used them to measure social capital in our local context (Howitt & Cramer, 2003). Using them to measure social capital in our local context (Howitt & Cramer, 2003). Using SPSS, factors were extracted for eigenvalues >1 implying that the factor is account- SPSS, factors were extracted for eigenvalues >1 implying that the factor is account- ing for a greater proportion of the variance than the original variable leading to ing for a greater proportion of the variance than the original variable leading to better interpretation. Furthermore, the factors were rotated to present patterns of better interpretation. Furthermore, the factors were rotated to present patterns of loadings which are easier to interpret. Only variables with factor loadings greater loadings which are easier to interpret. Only variables with factor loadings greater than 0.3 were extracted in the factor analysis. From the patterns of loadings it was than 0.3 were extracted in the factor analysis. From the patterns of loadings it was clear that variables related to conventional and neighborhood collective social clear that variables related to conventional and neighborhood collective social capital were clumped together while those related to individual structural and capital were clumped together while those related to individual structural and cognitive social capital were also clumped together at a different level (Table 5). cognitive social capital were also clumped together at a different level (Table 5).

Using PCA, we constructed factor scores to each subject to indicate where that Using PCA, we constructed factor scores to each subject to indicate where that subject stands on the retained variables in each component. The PCA assigned a subject stands on the retained variables in each component. The PCA assigned a

28 28 MATERIALS AND METHODS MATERIALS AND METHODS value of 1 and 2 to indicate strong positive loadings, a value of -1 to strongly value of 1 and 2 to indicate strong positive loadings, a value of -1 to strongly negative loadings and a value of 0 to immediate loadings (DeCoster, 1998). The negative loadings and a value of 0 to immediate loadings (DeCoster, 1998). The generated scores were labeled according to how they were theoretically clustered generated scores were labeled according to how they were theoretically clustered or structured together in each component: conventional and neighborhood scores or structured together in each component: conventional and neighborhood scores for collective social capital and structural and cognitive scores for individual social for collective social capital and structural and cognitive scores for individual social capital. The structural and cognitive social capital scores were later divided into capital. The structural and cognitive social capital scores were later divided into three equal groups to create different levels of social capital (Paper III and IV). three equal groups to create different levels of social capital (Paper III and IV).

In determining on an aggregated level, if a community had a low, medium or In determining on an aggregated level, if a community had a low, medium or high collective social capital, we used the generated scores for conventional and high collective social capital, we used the generated scores for conventional and neighborhood related variables for each individual and calculated average scores neighborhood related variables for each individual and calculated average scores for each case community (paper III). for each case community (paper III).

In sub-studies II and III, cross-tabulations were done to describe the association In sub-studies II and III, cross-tabulations were done to describe the association of HIV risk behaviors and HIV prevalence by socio-demographic and socio- of HIV risk behaviors and HIV prevalence by socio-demographic and socio- economic characteristics. The impact of socio-demographic and socio-economic economic characteristics. The impact of socio-demographic and socio-economic factors and social capital on HIV risk behavior as measured by condom use with factors and social capital on HIV risk behavior as measured by condom use with casual partner and testing for HIV and on HIV status was analyzed using multi- casual partner and testing for HIV and on HIV status was analyzed using multivariable logistic regression analysis. Independent variables with a 95% confidence variable logistic regression analysis. Independent variables with a 95% confidence interval (CI) for the odds ratio (OR) not including 1 in the bivariate analysis were interval (CI) for the odds ratio (OR) not including 1 in the bivariate analysis were included in the multivariable analysis. Like in sub-study II, we also used a prin- included in the multivariable analysis. Like in sub-study II, we also used a principal components factor analysis in sub-study III to measure individual social cipal components factor analysis in sub-study III to measure individual social capital in its structural and cognitive forms. capital in its structural and cognitive forms.

Table 5. Variables used to measure collective and individual social capital. Table 5. Variables used to measure collective and individual social capital.

Collective social capital: Collective social capital: Conventional measure Neighborhood related measure Conventional measure Neighborhood related measure Membership in organizations Kind of neighborhood Membership in organizations Kind of neighborhood General trust Interaction with neighborhood General trust Interaction with neighborhood Trust in strangers Cooperation to influence on decisions Trust in strangers Cooperation to influence on decisions Participation in voluntary activities Participation in voluntary activities

Individual social capital: Individual social capital: Structural social capital Cognitive social capital Structural social capital Cognitive social capital Membership in organizations General trust Membership in organizations General trust Individual influence on decisions Trust in strangers Individual influence on decisions Trust in strangers Giving social support Participation in voluntary activities Giving social support Participation in voluntary activities Receiving financial support Receiving financial support Interaction with neighbors Interaction with neighbors

29 29 RESULTS RESULTS Results Results

How can the social capital in the study area be characterized? How can the social capital in the study area be characterized? We observed extensive numbers of both formal and informal organizations and We observed extensive numbers of both formal and informal organizations and networks in the study area. Formal organizations consisted of governmental, networks in the study area. Formal organizations consisted of governmental, non-governmental, community-based and faith based organisations. Informal non-governmental, community-based and faith based organisations. Informal groups comprised social groups and networks, which were not officially registered groups comprised social groups and networks, which were not officially registered but village or street leaders were aware of their operations. Our findings indi- but village or street leaders were aware of their operations. Our findings indicated that the peri-urban community with medium HIV prevalence had more cated that the peri-urban community with medium HIV prevalence had more formal and informal organizations than the high HIV prevalence urban com- formal and informal organizations than the high HIV prevalence urban community and the low HIV prevalence rural community (Table 6). munity and the low HIV prevalence rural community (Table 6).

Table 6. Number of identified social groups and networks in the case communities by Table 6. Number of identified social groups and networks in the case communities by type of organization. type of organization. *Situated in the high HIV prevalence zone, **medium HIV prevalence zone and ***low HIV prevalence zone *Situated in the high HIV prevalence zone, **medium HIV prevalence zone and ***low HIV prevalence zone

Type of organization Community Type of organization Community *Urban **Peri-urban ***Rural Total *Urban **Peri-urban ***Rural Total Informal organizations Informal organizations Adult women 12 18 3 33 Adult women 12 18 3 33 Adult men and women 5 11 12 28 Adult men and women 5 11 12 28 Adult men 1 2 - 3 Adult men 1 2 - 3 Youth men and women 1 2 1 4 Youth men and women 1 2 1 4 Youth men 1 - - 1 Youth men 1 - - 1 Workers men and women - - 1 1 Workers men and women - - 1 1 Sub-total 20 33 17 70 Sub-total 20 33 17 70 Formal organizations Formal organizations FBOs 3 6 1 FBOs 3 6 1 Community 2 2 - Community 2 2 - Government 1 2 2 Government 1 2 2 NGOs 1 1 - NGOs 1 1 - Sub-total 7 11 3 21 Sub-total 7 11 3 21 Total 27 44 20 91 Total 27 44 20 91

Furthermore, the urban community had higher scores indicating higher level of Furthermore, the urban community had higher scores indicating higher level of social interactions than the peri-urban and rural communities (Figure 1, Paper I). social interactions than the peri-urban and rural communities (Figure 1, Paper I). We found that all three case communities had high levels of bonding social We found that all three case communities had high levels of bonding social capital while linking social capital was less pronounced, especially in the rural capital while linking social capital was less pronounced, especially in the rural community. The urban and peri-urban communities had more pronounced cog- community. The urban and peri-urban communities had more pronounced cognitive social capital than the rural community (Figure 4). nitive social capital than the rural community (Figure 4).

30 30 RESULTS RESULTS

3.5 3.5

3 3

2.5 2.5

2 2 Urban Urban

Scores 1.5 Scores 1.5 Peri-urban Peri-urban 1 Rural 1 Rural

0.5 0.5

0 0 Bonding Bridging Linking Cognitive Bonding Bridging Linking Cognitive Forms of social capital Forms of social capital

Figure 4. Levels of bonding, bridging, linking and cognitive forms of social capital in Figure 4. Levels of bonding, bridging, linking and cognitive forms of social capital in high (urban), medium (peri-urban) and low (rural) HIV prevalence zone. high (urban), medium (peri-urban) and low (rural) HIV prevalence zone. The existence of informal organizations also depended on support from the local The existence of informal organizations also depended on support from the local government leadership. Sometimes problems occured within some of these or- government leadership. Sometimes problems occured within some of these organizations, which could not be resolved by group leaders and had to be reported ganizations, which could not be resolved by group leaders and had to be reported to local authorities. The most common problem reported by these social groups to local authorities. The most common problem reported by these social groups was bad debtors. The local government leaders assisted to resolve the problem by was bad debtors. The local government leaders assisted to resolve the problem by ensuring that the debts were repaid as per group’s rules and regulations. ensuring that the debts were repaid as per group’s rules and regulations.

What are the mechanisms through which social capital may What are the mechanisms through which social capital may influence HIV risk behaviors? influence HIV risk behaviors? The grounded theory analysis resulted into six categories, which describe the The grounded theory analysis resulted into six categories, which describe the general role of structural and cognitive social capital and six sub-categories that general role of structural and cognitive social capital and six sub-categories that capture the mechanisms through which structural and cognitive forms of social capture the mechanisms through which structural and cognitive forms of social capital may have influenced HIV risk behaviors (Table 7). capital may have influenced HIV risk behaviors (Table 7).

Table 7. Categories and sub-categories describing the general role and mechanisms Table 7. Categories and sub-categories describing the general role and mechanisms through which social capital may have influenced HIV risk behaviors. through which social capital may have influenced HIV risk behaviors.

Categories Sub-categories Categories Sub-categories Structural: Structural: Structural: Structural: Serves the needs of vulnerable groups Enables socio-economic empowerment Serves the needs of vulnerable groups Enables socio-economic empowerment Create new opportunities for participation Increases information, knowledge and skills exchange Create new opportunities for participation Increases information, knowledge and skills exchange Increases women´s enrolment Strengthens women´s rights Increases women´s enrolment Strengthens women´s rights Cognitive: Cognitive: Cognitive: Cognitive: Formalizes membership rules Increases openness Formalizes membership rules Increases openness Fosters strict behavioral conduct Creates social pressures to take responsibility Fosters strict behavioral conduct Creates social pressures to take responsibility Fuels religious norms and values Enhances togetherness Fuels religious norms and values Enhances togetherness

31 31 RESULTS RESULTS

Three categories describing the general role of structural social capital as well as Three categories describing the general role of structural social capital as well as the three sub-categories describing the mechanisms through which social capital the three sub-categories describing the mechanisms through which social capital may create protective environment for HIV infection support Putnam’s view of may create protective environment for HIV infection support Putnam’s view of the benefits that individuals’ participation in civic organizations and networks the benefits that individuals’ participation in civic organizations and networks may have on socio-economic development (Putnam, 1995). may have on socio-economic development (Putnam, 1995).

Serving the needs of vulnerable groups Serving the needs of vulnerable groups A majority of the community members from vulnerable groups, such as the poor A majority of the community members from vulnerable groups, such as the poor and women participated in organizations and networks to access support from and women participated in organizations and networks to access support from other members. These organizations and networks provided loans, which were other members. These organizations and networks provided loans, which were used to initiate small scale business enterprises as well as addressing basic needs used to initiate small scale business enterprises as well as addressing basic needs of members whenever in need. They also supported members by giving them of members whenever in need. They also supported members by giving them food aid and money to be able to go to the hospital for treatment if needed. food aid and money to be able to go to the hospital for treatment if needed.

Creating new opportunities for participation Creating new opportunities for participation Both new and old organizations and networks created opportunities for vulner- Both new and old organizations and networks created opportunities for vulnerable groups to participate in socio-economic activities that had a direct impact able groups to participate in socio-economic activities that had a direct impact on their daily lives. on their daily lives.

Increasing enrolment by women Increasing enrolment by women There were more women social groups and networks (47%) when compared to There were more women social groups and networks (47%) when compared to mixed groups (43%), men’s groups (6%) or youth groups (4%). The increasing mixed groups (43%), men’s groups (6%) or youth groups (4%). The increasing number of women groups may be a result of traditional gender roles in caring number of women groups may be a result of traditional gender roles in caring for sick relatives and being responsible for funerals. Women may also regard for sick relatives and being responsible for funerals. Women may also regard participation in social groups and networks as an important strategy towards participation in social groups and networks as an important strategy towards social and economic emancipation. Participation in social groups and networks social and economic emancipation. Participation in social groups and networks enabled women to provide basic needs for their families. enabled women to provide basic needs for their families.

The mechanisms for how structural social capital influenced people’s lives are The mechanisms for how structural social capital influenced people’s lives are illustrated below. illustrated below.

Enabling economic empowerment Enabling economic empowerment Participation in social groups and networks facilitated members to access finan- Participation in social groups and networks facilitated members to access financial support. Rotating give-and-take scheme, whereby members contribute cial support. Rotating give-and-take scheme, whereby members contribute money to a collective fund every month, allow one or more members to borrow money to a collective fund every month, allow one or more members to borrow a large sum of money depending on the amount available to invest in economic a large sum of money depending on the amount available to invest in economic activities. In addition to monthly contributions, the collective fund gets addi- activities. In addition to monthly contributions, the collective fund gets additional resources from other income generating activities undertaken within the tional resources from other income generating activities undertaken within the groups. Such income generating activities include buying commodities such as groups. Such income generating activities include buying commodities such as sugar, salt, cooking oil, and soaps at a whole sale price and selling them at retail sugar, salt, cooking oil, and soaps at a whole sale price and selling them at retail

32 32 RESULTS RESULTS prices and keeping goats, cows, pigs and poultry for sale. Members could also prices and keeping goats, cows, pigs and poultry for sale. Members could also apply for a loan and pay back with a small interest within a period ranging from apply for a loan and pay back with a small interest within a period ranging from three to six months. Some social groups had established a “village bank”, which three to six months. Some social groups had established a “village bank”, which was a small wooden box with three padlocks where they deposited their month- was a small wooden box with three padlocks where they deposited their monthly contributions, interest generated from loans and money generated from group ly contributions, interest generated from loans and money generated from group projects. The money deposited in the “village bank” was given to members or projects. The money deposited in the “village bank” was given to members or non-members as loans or given to members as financial support. These different non-members as loans or given to members as financial support. These different types of financial support empowered members economically by enabling them types of financial support empowered members economically by enabling them to start up small business ventures or by injecting additional capital to those who to start up small business ventures or by injecting additional capital to those who already had ongoing income generating activities. already had ongoing income generating activities.

Strengthening women’s rights Strengthening women’s rights Women’s participation in social groups and networks strengthened their rights Women’s participation in social groups and networks strengthened their rights to earn their own income. Traditionally, most women used to stay at home and to earn their own income. Traditionally, most women used to stay at home and carry out domestic activities such as fetching water, collecting firewood, washing carry out domestic activities such as fetching water, collecting firewood, washing clothes and cooking for the children and husband. By participating in social clothes and cooking for the children and husband. By participating in social groups and networks, women became empowered and able to start their own groups and networks, women became empowered and able to start their own small scale business ventures. This ultimately enabled them to be independent small scale business ventures. This ultimately enabled them to be independent and contribute to the basic household needs and enhance their respect both at and contribute to the basic household needs and enhance their respect both at household and community levels. household and community levels.

Increasing information, knowledge and skills Increasing information, knowledge and skills Participation in both formal and informal organizational activities provided Participation in both formal and informal organizational activities provided members’ access to various sources of information, knowledge and skills. By members’ access to various sources of information, knowledge and skills. By coming together in their weekly or monthly meetings they were informed about coming together in their weekly or monthly meetings they were informed about issues related to health including HIV/AIDS transmission and prevention meth- issues related to health including HIV/AIDS transmission and prevention methods. In addition, formal organizations, particularly FBOs provided HIV/AIDS ods. In addition, formal organizations, particularly FBOs provided HIV/AIDS education to members of social groups using group representatives or mobile education to members of social groups using group representatives or mobile workshops. Members of social groups and networks also got entrepreneurial workshops. Members of social groups and networks also got entrepreneurial knowledge and skills to start up and maintain small scale business projects. Suc- knowledge and skills to start up and maintain small scale business projects. Suc- cessful members of social groups were seen as exemplary and their groups were cessful members of social groups were seen as exemplary and their groups were regarded as role models in the community. regarded as role models in the community.

Three categories developed to describe the general role of cognitive social capital Three categories developed to describe the general role of cognitive social capital and the three sub-categories formed to describe the mechanisms through which and the three sub-categories formed to describe the mechanisms through which social capital influenced people’s behaviors can also be related to Putnam’s notion social capital influenced people’s behaviors can also be related to Putnam’s notion of social capital as creating norms of trust and reciprocity. of social capital as creating norms of trust and reciprocity.

Formalizing membership rules Formalizing membership rules All social groups and networks in the case communities had stipulated strict rules All social groups and networks in the case communities had stipulated strict rules and regulations for their members to adhere to. These rules and regulations dif- and regulations for their members to adhere to. These rules and regulations dif-

33 33 RESULTS RESULTS fered from one group or network to another. However, the most common and fered from one group or network to another. However, the most common and formalized rules included: respect fellow members, behave in a respectful man- formalized rules included: respect fellow members, behave in a respectful man- ner, keep organization’s secrets and be trustworthy. ner, keep organization’s secrets and be trustworthy.

Fostering good behavior through strict conduct Fostering good behavior through strict conduct Before joining any social groups or networks, applicants were supposed to read Before joining any social groups or networks, applicants were supposed to read the constitutions and agree to stipulated rules and regulations. These rules and the constitutions and agree to stipulated rules and regulations. These rules and regulations formed the norms and values that governed the operation of these regulations formed the norms and values that governed the operation of these organizations. Those who violated the stipulated norms and values were obliged organizations. Those who violated the stipulated norms and values were obliged to pay fines or penalties or loose their membership. Such strict rules and regula- to pay fines or penalties or loose their membership. Such strict rules and regulations contributed to maintenance of good behavior among group members. tions contributed to maintenance of good behavior among group members.

Fuelling religious norms and values Fuelling religious norms and values Faith based organizations (FBOs) encouraged members from the same neighbor- Faith based organizations (FBOs) encouraged members from the same neighborhoods to form social groups comprising five to seven neighbors and then used hoods to form social groups comprising five to seven neighbors and then used these groups as avenues for conveying educative information. FBOs also sup- these groups as avenues for conveying educative information. FBOs also supported social groups by giving them small capital or animals to start up small-scale ported social groups by giving them small capital or animals to start up small-scale business ventures. Such group activities did not only brought members together business ventures. Such group activities did not only brought members together and enhanced their interaction and support to each other but also provided an and enhanced their interaction and support to each other but also provided an important avenue for religious leaders to inculcate religious norms and values. important avenue for religious leaders to inculcate religious norms and values.

The mechanisms of how cognitive social capital influence people’s behavior are The mechanisms of how cognitive social capital influence people’s behavior are briefly discussed below. briefly discussed below.

Increasing openness Increasing openness The requirement that all members should adhere to stipulated rules and regula- The requirement that all members should adhere to stipulated rules and regulations created an environment through which members were confident and open tions created an environment through which members were confident and open to discuss misconduct that may violate these norms and values. Such practices to discuss misconduct that may violate these norms and values. Such practices were translated at the family level whereby parents became open to discuss health were translated at the family level whereby parents became open to discuss health issues including sexual related risk behaviors with their children, especially issues including sexual related risk behaviors with their children, especially teenagers. teenagers.

Increasing social pressure to take responsibility Increasing social pressure to take responsibility Many social groups had made taking loans from the group funds a compulsory Many social groups had made taking loans from the group funds a compulsory practice as a way of generating additional income to the group funds. This system practice as a way of generating additional income to the group funds. This system put pressure on members to take social responsibilities. Borrowing money from put pressure on members to take social responsibilities. Borrowing money from the group empowered members to initiate small scale business projects making the group empowered members to initiate small scale business projects making them feel responsible to provide basic needs such as school and medical fees to them feel responsible to provide basic needs such as school and medical fees to their families. their families.

34 34 RESULTS RESULTS

Increasing togetherness Increasing togetherness The social groups and networks were seen as important for bringing community The social groups and networks were seen as important for bringing community members together and the established rules and regulations enhanced collective- members together and the established rules and regulations enhanced collectiveness within the community regardless of age, sex and socio-economic status. ness within the community regardless of age, sex and socio-economic status.

Our analysis identified four major behavioral changes resulting from people’s Our analysis identified four major behavioral changes resulting from people’s involvement in social groups and networks and the norms and values that arise involvement in social groups and networks and the norms and values that arise from these interactions. The changes included decreased number of sexual part- from these interactions. The changes included decreased number of sexual partners, encouraging abstinence until marriage, decreasing opportunities for casu- ners, encouraging abstinence until marriage, decreasing opportunities for casual sex and empowering community members to demand or use condoms. al sex and empowering community members to demand or use condoms.

Is access to structural and cognitive social capital associated Is access to structural and cognitive social capital associated with HIV risk behaviors? with HIV risk behaviors?

Association between structural social capital and condom use with Association between structural social capital and condom use with casual sex partner casual sex partner Adjusted odds ratio (controlled for sex, age, marital status, wealth index, place Adjusted odds ratio (controlled for sex, age, marital status, wealth index, place of residence and level of education) show that individuals who had access to of residence and level of education) show that individuals who had access to medium and high levels of structural social capital were more likely to use con- medium and high levels of structural social capital were more likely to use condoms with casual sex partners compared to those without such access: OR=1.6, doms with casual sex partners compared to those without such access: OR=1.6, 95% CI: 1.03-2.4 and OR=1.7, 95% CI: 1.1-2.7 (Figure 5). When stratified by sex 95% CI: 1.03-2.4 and OR=1.7, 95% CI: 1.1-2.7 (Figure 5). When stratified by sex and controlling for sex, age, marital status, wealth index, place of residence and and controlling for sex, age, marital status, wealth index, place of residence and level of education, women’s access to medium and high level of structural social level of education, women’s access to medium and high level of structural social capital was significantly associated with condom use with casual sex partners capital was significantly associated with condom use with casual sex partners than women with low access (OR=2.4, 95% CI: 1.2-4.8; and OR=2.3, 95% CI: than women with low access (OR=2.4, 95% CI: 1.2-4.8; and OR=2.3, 95% CI: 1.1-4.4). However, the multivariable analysis did not show association between 1.1-4.4). However, the multivariable analysis did not show association between access to structural social capital and use of condom with casual sex partners access to structural social capital and use of condom with casual sex partners among men (Table 8). among men (Table 8).

Association between cognitive social capital and condom use with Association between cognitive social capital and condom use with casual sex partner casual sex partner The adjusted odds ratio (controlled for education, sex, place of residence, age, and The adjusted odds ratio (controlled for education, sex, place of residence, age, and marital status) show that individuals who have access to high levels of cognitive marital status) show that individuals who have access to high levels of cognitive social capital were more likely to use condoms with casual sex partners than those social capital were more likely to use condoms with casual sex partners than those with low access (OR = 1.7, 95% CI: 1.05-2.4). When stratified the analysis by sex, with low access (OR = 1.7, 95% CI: 1.05-2.4). When stratified the analysis by sex, there was no association observed between access to cognitive social capital and there was no association observed between access to cognitive social capital and condom use with casual sex partners both for women and men (Table 8). condom use with casual sex partners both for women and men (Table 8).

35 35 RESULTS RESULTS

1.8 1.8

1.6 1.6

1.4 1.4

1.2 1.2

1 1 0.8 SSC 0.8 SSC Odds ratio Odds ratio 0.6 CSC 0.6 CSC 0.4 0.4

0.2 0.2

0 0 High Medium Low High Medium Low Levels of structural and cognitive social capital Levels of structural and cognitive social capital

Figure 5. Adjusted odds ratio (OR) for condom use with casual sex partners by access Figure 5. Adjusted odds ratio (OR) for condom use with casual sex partners by access to structural social capital (SSC) and cognitive social capital (CSC). to structural social capital (SSC) and cognitive social capital (CSC).

Table 8. Bivariate and multivariable logistic regression analysis calculating odds ratio Table 8. Bivariate and multivariable logistic regression analysis calculating odds ratio (OR) and 95% confidence intervals (CI) for likelihood to use condom with casual sex (OR) and 95% confidence intervals (CI) for likelihood to use condom with casual sex partner by sex and structural and cognitive social capital (SC). Analysis limited to per- partner by sex and structural and cognitive social capital (SC). Analysis limited to persons who reported to have had sex. sons who reported to have had sex.

Sex and Level Used condom Bivariate analysis Multivariable Sex and Level Used condom Bivariate analysis Multivariable type of with casual sex logistic regression type of with casual sex logistic regression social partner social partner capital Yes No OR 95% CI OR 95% CI capital Yes No OR 95% CI OR 95% CI Females: Females: Structural SC Low 29 233 1 1 Structural SC Low 29 233 1 1 Medium 30 134 1.8 1.03–3.1 2.4 1.2-4.8 Medium 30 134 1.8 1.03–3.1 2.4 1.2-4.8 High 33 132 1.4 0.81-2.3 2.3 1.1-4.4 High 33 132 1.4 0.81-2.3 2.3 1.1-4.4 Cognitive SC Cognitive SC Low 42 308 1 1 Low 42 308 1 1 Medium 30 162 1.4 0.82-2.2 1.2 0.66-2.3 Medium 30 162 1.4 0.82-2.2 1.2 0.66-2.3 High 20 89 1.6 0.92-2.9 1.5 0.71-3.1 High 20 89 1.6 0.92-2.9 1.5 0.71-3.1 Males: Males: Structural SC Low 47 113 1 1 Structural SC Low 47 113 1 1 Medium 45 134 0.81 0.50-1.3 1.1 0.62-1.9 Medium 45 134 0.81 0.50-1.3 1.1 0.62-1.9 High 39 154 0.61 0.37-0.99 1.2 0.66-2.3 High 39 154 0.61 0.37-0.99 1.2 0.66-2.3 Cognitive SC Cognitive SC Low 21 67 1 1 Low 21 67 1 1 Medium 48 154 0.99 0.55-1.8 1.2 0.58-2.3 Medium 48 154 0.99 0.55-1.8 1.2 0.58-2.3 High 62 180 1.1 0.62-1.9 1.7 0.85-3.2 High 62 180 1.1 0.62-1.9 1.7 0.85-3.2

36 36 RESULTS RESULTS

Association between structural social capital and self-reported Association between structural social capital and self-reported HIV testing HIV testing The adjusted odds ratio (controlled for age, sex, level of education, place of resi- The adjusted odds ratio (controlled for age, sex, level of education, place of residence, and marital status), show that individuals who had access to low and dence, and marital status), show that individuals who had access to low and medium levels of structural social capital were more likely not to have tested for medium levels of structural social capital were more likely not to have tested for HIV infection than those with access to high levels: OR=1.7 (95% CI: 1.2-2.5) and HIV infection than those with access to high levels: OR=1.7 (95% CI: 1.2-2.5) and OR=1.7 (95% CI: 1.2-2.4), respectively (Figure 6). Stratified analysis by sex and OR=1.7 (95% CI: 1.2-2.4), respectively (Figure 6). Stratified analysis by sex and controlled for age, marital status, place of residence and level of education, indi- controlled for age, marital status, place of residence and level of education, indicates that women with access to medium and men with access to low levels of cates that women with access to medium and men with access to low levels of structural social capital were more likely not to have tested for HIV than those structural social capital were more likely not to have tested for HIV than those with access to the high levels: OR=2.3, 95% CI: 1.2-4.3 and OR=2.2, 95% CI: with access to the high levels: OR=2.3, 95% CI: 1.2-4.3 and OR=2.2, 95% CI: 1.3-3.7, respectively (Table 9). 1.3-3.7, respectively (Table 9).

1.8 1.8

1.6 1.6

1.4 1.4

1.2 1.2

1 1

0.8 SSC 0.8 SSC

Odds ratio 0.6 CSC Odds ratio 0.6 CSC 0.4 0.4

0.2 0.2

0 0 High Medium Low High Medium Low Levels of structural and cognitive social capital Levels of structural and cognitive social capital

Figure 6. Adjusted odds ratios (OR) for not testing for HIV infection by individual ac- Figure 6. Adjusted odds ratios (OR) for not testing for HIV infection by individual access to structural social capital (SSC) and cognitive social capital (CSC). cess to structural social capital (SSC) and cognitive social capital (CSC).

Association between cognitive social capital and self-reported Association between cognitive social capital and self-reported HIV testing HIV testing Adjusted odds ratio show that there is no association between access to cognitive Adjusted odds ratio show that there is no association between access to cognitive social capital and not tested for HIV infection at any level of cognitive social social capital and not tested for HIV infection at any level of cognitive social capital. When stratified analysis by sex, similar pattern was observed both for capital. When stratified analysis by sex, similar pattern was observed both for women and men (Table 9). women and men (Table 9).

37 37 RESULTS RESULTS

Table 9. Bivariate and multivariable logistic regression analysis calculating odds ratio Table 9. Bivariate and multivariable logistic regression analysis calculating odds ratio (OR) and 95% confidence intervals for the likelihood of not testing for HIV by sex and (OR) and 95% confidence intervals for the likelihood of not testing for HIV by sex and structural and cognitive social capital (SC). structural and cognitive social capital (SC).

Sex and type Tested for Bivariate analysis Multivariable logistic Sex and type Tested for Bivariate analysis Multivariable logistic of social Level HIV regression of social Level HIV regression capital No Yes OR 95% CI OR 95% CI capital No Yes OR 95% CI OR 95% CI Females: Females: Structural SC High 37 195 1 1 Structural SC High 37 195 1 1 Medium 42 138 1.6 0.98-2.6 2.3 1.2-4.3 Medium 42 138 1.6 0.98-2.6 2.3 1.2-4.3 Low 65 222 1.5 0.98-2.4 1.4 0.79-2.6 Low 65 222 1.5 0.98-2.4 1.4 0.79-2.6 Cognitive SC Cognitive SC High 24 101 1 1 High 24 101 1 1 Medium 36 165 0.92 0.52-1.6 0.84 0.42-1.7 Medium 36 165 0.92 0.52-1.6 0.84 0.42-1.7 Low 84 289 1.2 0.74-2.0 1.2 0.66-2.3 Low 84 289 1.2 0.74-2.0 1.2 0.66-2.3 Males: Males: Structural SC High 52 153 1 1 Structural SC High 52 153 1 1 Medium 72 129 1.6 1.1–2.5 1.6 0.95-2.6 Medium 72 129 1.6 1.1–2.5 1.6 0.95-2.6 Low 108 88 3.6 2.4–5.5 2.2 1.3-3.7 Low 108 88 3.6 2.4–5.5 2.2 1.3-3.7 Cognitive SC Cognitive SC Low 48 59 1 1 Low 48 59 1 1 Medium 86 140 1.1 0.74-1.5 1.1 0.71-1.7 Medium 86 140 1.1 0.74-1.5 1.1 0.71-1.7 High 98 171 1.4 0.90-2.2 0.97 0.56-1.7 High 98 171 1.4 0.90-2.2 0.97 0.56-1.7

Is access to structural and cognitive social capital associated with Is access to structural and cognitive social capital associated with risk of HIV infection? risk of HIV infection?

Association between structural social capital and risk of HIV infection Association between structural social capital and risk of HIV infection The findings show that the overall HIV prevalence in Bukoba urban was 9.8% (95% The findings show that the overall HIV prevalence in Bukoba urban was 9.8% (95% CI: 8.8-1.8) with women having a higher prevalence of HIV infection (11.6%; 95% CI: 8.8-1.8) with women having a higher prevalence of HIV infection (11.6%; 95% CI: 10.2-13.0) than men (7.4%; 95% CI: 6.0-8.8). After controlling for age, sex, CI: 10.2-13.0) than men (7.4%; 95% CI: 6.0-8.8). After controlling for age, sex, marital status, level of education, occupation and wealth index, the adjusted odds marital status, level of education, occupation and wealth index, the adjusted odds ratio show that individuals with access to low and medium levels of structural ratio show that individuals with access to low and medium levels of structural social capital were almost eight and three times more likely to be HIV infected than social capital were almost eight and three times more likely to be HIV infected than those with access to high levels: OR=7.7, 95% CI: 5.0-12 and OR=2.8, 95% CI: 1.8- those with access to high levels: OR=7.7, 95% CI: 5.0-12 and OR=2.8, 95% CI: 1.8- 4.5, respectively (Figure 7). Similar pattern was observed when stratified the 4.5, respectively (Figure 7). Similar pattern was observed when stratified the analysis by sex and controlled for age, marital status, level of education, occupation analysis by sex and controlled for age, marital status, level of education, occupation and wealth index. However, among men, those with low levels of structural social and wealth index. However, among men, those with low levels of structural social capital had high risk of being HIV infected than those with high levels, OR=10, capital had high risk of being HIV infected than those with high levels, OR=10, 95% CI: 4.4-23 while among women, those with low access to structural social 95% CI: 4.4-23 while among women, those with low access to structural social capital had slightly less risk of being HIV infected than those with high access, capital had slightly less risk of being HIV infected than those with high access, OR=7.2, 95% CI: 4.4-12 (Table 5, Paper IV). OR=7.2, 95% CI: 4.4-12 (Table 5, Paper IV).

38 38 RESULTS RESULTS

Association between cognitive social capital and risk of HIV infection Association between cognitive social capital and risk of HIV infection The adjusted odds ratio (controlled for age, sex, marital status, level of education, The adjusted odds ratio (controlled for age, sex, marital status, level of education, occupation and wealth index) show that Individuals whose accessibility to cogni- occupation and wealth index) show that Individuals whose accessibility to cognitive social capital is low were 2.4 times more likely to be infected with HIV than tive social capital is low were 2.4 times more likely to be infected with HIV than those in the high level OR=2.4; 95% CI: 1.6-3.5 (Figure 7). However, the asso- those in the high level OR=2.4; 95% CI: 1.6-3.5 (Figure 7). However, the association between access to cognitive social capital and HIV infection was not as ciation between access to cognitive social capital and HIV infection was not as linear as in the structural social capital, since there was no association observed linear as in the structural social capital, since there was no association observed between those who had access to medium level of cognitive social capital and between those who had access to medium level of cognitive social capital and HIV infection. The multivariable regression analysis shows that women and men HIV infection. The multivariable regression analysis shows that women and men with access to low levels of cognitive social capital were respectively 2.3 and 2.7 with access to low levels of cognitive social capital were respectively 2.3 and 2.7 times more likely to be HIV positive compared to those who have access to high times more likely to be HIV positive compared to those who have access to high level (Figure 8). level (Figure 8).

These results demonstrate that individuals, who have low participation in formal These results demonstrate that individuals, who have low participation in formal and informal organizations, have little interactions with neighbors, have insuf- and informal organizations, have little interactions with neighbors, have insuf- ficient ability to influence decisions that affect their lives and have low engagement ficient ability to influence decisions that affect their lives and have low engagement in reciprocal relations (Structural social capital) and those who have low gen- in reciprocal relations (Structural social capital) and those who have low general trust and low trust in strangers (Cognitive social capital) have increased risk eral trust and low trust in strangers (Cognitive social capital) have increased risk of HIV infection. of HIV infection.

9 9

8 8

7 7

6 6

5 5

4 SSC 4 SSC

Odds ratio 3 CSC Odds ratio 3 CSC

2 2

1 1

0 0 High Medium Low High Medium Low Levels of structural and cognitive social capital Levels of structural and cognitive social capital

Figure 7. Adjusted odds ratio for the risk of being HIV infected by individual access to Figure 7. Adjusted odds ratio for the risk of being HIV infected by individual access to structural social capital (SSC) and cognitive social capital (CSC). structural social capital (SSC) and cognitive social capital (CSC).

39 39 RESULTS RESULTS

12 12

10 10

8 8

6 High 6 High Medium Medium 4 4

Odds ratio Low Odds ratio Low

2 2

0 0 Men SSC Women SSC Men CSC Women CSC Men SSC Women SSC Men CSC Women CSC

Figure 8. Adjusted odds ratio for the risk of HIV infected by access to structural social Figure 8. Adjusted odds ratio for the risk of HIV infected by access to structural social capital (SSC) and cognitive social capital (CSC) and sex. capital (SSC) and cognitive social capital (CSC) and sex.

40 40 METHODOLOGICAL CONSIDERATIONS METHODOLOGICAL CONSIDERATIONS Methodological considerations Methodological considerations

Strengths Strengths The main strength of this study is the combination of qualitative and quantitative The main strength of this study is the combination of qualitative and quantitative methods for data collection and analysis. In the qualitative approach, we adopt- methods for data collection and analysis. In the qualitative approach, we adopted triangulation of different data sources using key informant interviews, FGDs ed triangulation of different data sources using key informant interviews, FGDs and review of documents, to enhance credibility and validity of the study (Creswell, and review of documents, to enhance credibility and validity of the study (Creswell, 2008). To capture variation, data was collected from members as well as non- 2008). To capture variation, data was collected from members as well as non- members of formal and informal organizations, leaders of social groups and members of formal and informal organizations, leaders of social groups and organizations as well as village and ward leaders. Field notes were taken and all organizations as well as village and ward leaders. Field notes were taken and all recorded information was later transcribed verbatim. Our long-term engagement recorded information was later transcribed verbatim. Our long-term engagement in the project enhanced our understanding of how social capital may exert influ- in the project enhanced our understanding of how social capital may exert influence on behavior change among community members and thereby increased the ence on behavior change among community members and thereby increased the trustworthiness of the study. trustworthiness of the study.

Hypotheses generated in the qualitative sub-study were tested in the quantitative Hypotheses generated in the qualitative sub-study were tested in the quantitative study, which increased the validity of the study. The questionnaires were devel- study, which increased the validity of the study. The questionnaires were developed following a thorough review of the literature on the measurement of social oped following a thorough review of the literature on the measurement of social capital and adjusted to fit the local context. capital and adjusted to fit the local context.

The questionnaire was administered using face to face interviews. This was con- The questionnaire was administered using face to face interviews. This was considered an appropriate to reach also illiterate respondents and to ask about sidered an appropriate to reach also illiterate respondents and to ask about sensitive issues such as HIV risk behaviors. This tool has been criticized for not sensitive issues such as HIV risk behaviors. This tool has been criticized for not being able to provide sufficient anonymity and privacy in reporting sensitive is- being able to provide sufficient anonymity and privacy in reporting sensitive issues compared to other tools like self-administered questionnaires (Gregson et sues compared to other tools like self-administered questionnaires (Gregson et al, 2002). However, a recent systematic review has shown that there is no sig- al, 2002). However, a recent systematic review has shown that there is no significant difference between types of interview tool in reporting about non-condom nificant difference between types of interview tool in reporting about non-condom use, number of sexual partner and ever had sex (Phillips et al, 2010). use, number of sexual partner and ever had sex (Phillips et al, 2010).

Our field assistants were qualified professional nurses and experienced in coun- Our field assistants were qualified professional nurses and experienced in counseling and they were trained before undertaking the field work. In addition, a seling and they were trained before undertaking the field work. In addition, a majority of them had been working within the project for more than ten years, majority of them had been working within the project for more than ten years, which increased their ability to interview about sensitive issues. which increased their ability to interview about sensitive issues.

The research team in this study represented multidisciplinary fields of social The research team in this study represented multidisciplinary fields of social science, epidemiology and biostatistics as well as microbiology and immunology. science, epidemiology and biostatistics as well as microbiology and immunology. All contributed with their specific competence in the planning of the three studies, All contributed with their specific competence in the planning of the three studies, interpretation of the findings and reviewing drafts of different papers, which interpretation of the findings and reviewing drafts of different papers, which increased the credibility of this work. increased the credibility of this work.

41 41 METHODOLOGICAL CONSIDERATIONS METHODOLOGICAL CONSIDERATIONS

Limitations Limitations The selection of informants in sub-study I was based on identification of study The selection of informants in sub-study I was based on identification of study participants by community leaders, which may have contributed to some par- participants by community leaders, which may have contributed to some participants having overly positive notions of social capital. The information on how ticipants having overly positive notions of social capital. The information on how long the formal and informal organizations had been in place was incomplete, long the formal and informal organizations had been in place was incomplete, since our informants had difficulties recalling when these organizations had since our informants had difficulties recalling when these organizations had started their operations. This limited our possibility to describe historically the started their operations. This limited our possibility to describe historically the influence of social capital on HIV risk behaviors in Kagera region. influence of social capital on HIV risk behaviors in Kagera region.

A cross-sectional design provides accurate point estimates of HIV infection. A cross-sectional design provides accurate point estimates of HIV infection. However this design is limited by its inability to allow any conclusions of causal However this design is limited by its inability to allow any conclusions of causal relationships between social capital and the risk of HIV infection. relationships between social capital and the risk of HIV infection.

In Tanzania, like other African countries, speaking openly about sexual matters In Tanzania, like other African countries, speaking openly about sexual matters particularly in the presence of children is a taboo (Killewo, 1994). In sub-study particularly in the presence of children is a taboo (Killewo, 1994). In sub-study I we therefore organized homogenous focus groups in terms of age and sex to I we therefore organized homogenous focus groups in terms of age and sex to respect the local traditions governing sexual matters. This increased the chance respect the local traditions governing sexual matters. This increased the chance that participants would openly discuss sex related issues. that participants would openly discuss sex related issues.

For Paper III, we used aggregated individual data instead of ecological data to For Paper III, we used aggregated individual data instead of ecological data to measure collective social capital. Ecological data is preferred for measuring col- measure collective social capital. Ecological data is preferred for measuring collective social capital but is rarely used (Eriksson, 2010). We are of the view that lective social capital but is rarely used (Eriksson, 2010). We are of the view that by utilizing both conventional and neighborhood related aggregated individual by utilizing both conventional and neighborhood related aggregated individual indicators; this study has sufficient strengths to measure collective social capital. indicators; this study has sufficient strengths to measure collective social capital.

There is no consensus on how to operationalize and measure the concept of social There is no consensus on how to operationalize and measure the concept of social capital. Portes (1998) underscored that it is particularly difficult to quantify some capital. Portes (1998) underscored that it is particularly difficult to quantify some of the indicators of social capital like norms and shared values. However, Stone of the indicators of social capital like norms and shared values. However, Stone (2001, P 34) suggests that all dimensions of social capital can be measured. She (2001, P 34) suggests that all dimensions of social capital can be measured. She underscores that “empirical operationalization of social capital must reflect underscores that “empirical operationalization of social capital must reflect theoretical understandings of the concept”. In this study, only indicators that theoretical understandings of the concept”. In this study, only indicators that reflect our conceptual framework were included, which we believe ensured con- reflect our conceptual framework were included, which we believe ensured conceptual validity and reduced ambiguity in our measurements of social capital. ceptual validity and reduced ambiguity in our measurements of social capital.

42 42 DISCUSSION DISCUSSION Discussion Discussion

This thesis supports the argument that social capital can be protective to HIV infec- This thesis supports the argument that social capital can be protective to HIV infection. Access to structural and cognitive forms of social capital is an important tion. Access to structural and cognitive forms of social capital is an important strategy for community members to adopt safer sexually related behaviors. There- strategy for community members to adopt safer sexually related behaviors. There- fore including social capital in the design and delivery of HIV and AIDS interven- fore including social capital in the design and delivery of HIV and AIDS interventions strategies will enhance effectiveness in the fight against HIV epidemic. The tions strategies will enhance effectiveness in the fight against HIV epidemic. The distribution of collective social capital differed between our three case communities, distribution of collective social capital differed between our three case communities, which represented varying HIV prevalence. The urban community, initially char- which represented varying HIV prevalence. The urban community, initially characterized by high HIV prevalence and later on declining HIV trends, had the high- acterized by high HIV prevalence and later on declining HIV trends, had the high- est collective social capital. This implies that the collective social capital that ini- est collective social capital. This implies that the collective social capital that initially was regarded as fueling the epidemic could be mobilised and become an tially was regarded as fueling the epidemic could be mobilised and become an important avenue for creating conducive environments for adopting safer health important avenue for creating conducive environments for adopting safer health behaviors. However, social capital can also expose community members to increased behaviors. However, social capital can also expose community members to increased risk of HIV infection by excluding individuals from exisiting networks or by being risk of HIV infection by excluding individuals from exisiting networks or by being part of networks with high-risk interactions. Thus programme managers and part of networks with high-risk interactions. Thus programme managers and policy makers need to consider these negative aspects when deciding how to include policy makers need to consider these negative aspects when deciding how to include social capital into new strategies for HIV prevention. social capital into new strategies for HIV prevention.

Social capital and HIV infection Social capital and HIV infection The studies illustrate that access to high levels of structural and cognitive forms The studies illustrate that access to high levels of structural and cognitive forms of social capital can be protective to HIV infection. Access to structural social of social capital can be protective to HIV infection. Access to structural social capital increases the likelihood for reducing HIV related risk behaviors. Partici- capital increases the likelihood for reducing HIV related risk behaviors. Partici- pation in formal and informal organizations, individual ability to influence deci- pation in formal and informal organizations, individual ability to influence decisions, interactions with neighbors and relatives play a positive role for reducing sions, interactions with neighbors and relatives play a positive role for reducing HIV infection among community members. This effect of structural social capi- HIV infection among community members. This effect of structural social capital was observed among both men and women although the influence of struc- tal was observed among both men and women although the influence of structural social capital was greater for men than women. Our findings conforms with tural social capital was greater for men than women. Our findings conforms with Putnam’s conceptualization of social capital referring to high level of commu- Putnam’s conceptualization of social capital referring to high level of community participation in civic organizations that results in positive aspects of com- nity participation in civic organizations that results in positive aspects of community life, which in this case include avoidance of HIV infection (Putnam, 1995). munity life, which in this case include avoidance of HIV infection (Putnam, 1995).

Our findings also show that access to high levels of cognitive social capital is as- Our findings also show that access to high levels of cognitive social capital is associated with less risk of HIV infection. This is also in line with Putnam’s defini- sociated with less risk of HIV infection. This is also in line with Putnam’s definition of social capital, suggesting trust as one of the main features of social capital tion of social capital, suggesting trust as one of the main features of social capital that facilitates the coordination and the attainment of various community and that facilitates the coordination and the attainment of various community and individual benefits Bandura (1996). Israel et al (1994) has further described that individual benefits Bandura (1996). Israel et al (1994) has further described that trust within social groupings leads to generalized levels of perceived self-efficacy trust within social groupings leads to generalized levels of perceived self-efficacy or ‘empowerment’ that may influence individuals to engage in health-protective or ‘empowerment’ that may influence individuals to engage in health-protective behaviors. behaviors.

43 43 DISCUSSION DISCUSSION

Social capital and HIV prevention Social capital and HIV prevention Structural (bonding, bridging, linking) and cognitive social capital was most Structural (bonding, bridging, linking) and cognitive social capital was most pronounced in the communities situated in the urban and peri-urban areas with pronounced in the communities situated in the urban and peri-urban areas with observed declining HIV trends, compared to the community with low HIV observed declining HIV trends, compared to the community with low HIV prevalence situated in the rural area. People’s participation in formal and infor- prevalence situated in the rural area. People’s participation in formal and informal organizations and the norms of trust and reciprocity embedded in their in- mal organizations and the norms of trust and reciprocity embedded in their interactions seem to have played a positive role in changing community members’ teractions seem to have played a positive role in changing community members’ behaviors. The organizations have been used as meeting points for community behaviors. The organizations have been used as meeting points for community members to engage in activities that empower them, and where norms, values, members to engage in activities that empower them, and where norms, values, trust and reciprocal relations also helped to shape sexual behaviors. Active par- trust and reciprocal relations also helped to shape sexual behaviors. Active participation in formal and informal organizations enabled members to access loans/ ticipation in formal and informal organizations enabled members to access loans/ money, through weekly or monthly contributions from other members and to money, through weekly or monthly contributions from other members and to initiate their own income generating projects. These economic activities were initiate their own income generating projects. These economic activities were particularly important for women allowing them to engage in small projects, particularly important for women allowing them to engage in small projects, decresing the need for transactional sex relationships. Economic empowerment decresing the need for transactional sex relationships. Economic empowerment also enhanced women’s confidence to negotiate use of condoms with their sexu- also enhanced women’s confidence to negotiate use of condoms with their sexual partners. The values and norms developed within the networks played an al partners. The values and norms developed within the networks played an important role by for example demanding members to avoid engaging in multi- important role by for example demanding members to avoid engaging in multiple sexual relations and encouraging young generation to abstain from sexual ple sexual relations and encouraging young generation to abstain from sexual relations until marriage. Other studies have also suggested that social capital may relations until marriage. Other studies have also suggested that social capital may empower vulnerable community members to develop a consciousness that may empower vulnerable community members to develop a consciousness that may enhance their capacity to take control over important aspects of their lives includ- enhance their capacity to take control over important aspects of their lives including those related to health (Wallerstein, 1992; Campbell, 2003). ing those related to health (Wallerstein, 1992; Campbell, 2003).

Formal organizations, particularly faith-based organizations, were found to play a Formal organizations, particularly faith-based organizations, were found to play a vital role in nurturing social capital in the study setting. They supported the com- vital role in nurturing social capital in the study setting. They supported the community to fight against further spread of HIV infection by forming social groups munity to fight against further spread of HIV infection by forming social groups that also became entry points for HIV/AIDS related health education focusing on that also became entry points for HIV/AIDS related health education focusing on encouraging abstinence and faithfulness. Putnam also pointed out the indirect role encouraging abstinence and faithfulness. Putnam also pointed out the indirect role of religious organizations when describing how they served American civic life by of religious organizations when describing how they served American civic life by providing social support to community members and at the same time inculcated providing social support to community members and at the same time inculcated moral values encouraging unselfishness (Putnam, 2000). In South Africa, religious moral values encouraging unselfishness (Putnam, 2000). In South Africa, religious leaders themselves have emphasised their role in educating the young generation leaders themselves have emphasised their role in educating the young generation about sexuality. However, the taboo of speaking about sexuality in public, in the about sexuality. However, the taboo of speaking about sexuality in public, in the African context, makes it difficult for some religious leaders to organize life skills African context, makes it difficult for some religious leaders to organize life skills programme for their followers (Eriksson et al, 2011). programme for their followers (Eriksson et al, 2011).

This thesis indicates that access to structural and cognitive social capital increased This thesis indicates that access to structural and cognitive social capital increased the likelihood for both men and women to use condoms with casual sex partners. the likelihood for both men and women to use condoms with casual sex partners. The structural social capital provided avenues for people to come together and The structural social capital provided avenues for people to come together and

44 44 DISCUSSION DISCUSSION openly discuss HIV related issues including use of condoms. Furthermore, par- openly discuss HIV related issues including use of condoms. Furthermore, participation in social groups empowered people economically and enhanced their ticipation in social groups empowered people economically and enhanced their confidence thereby enabling particularly women and the poor to negotiate about confidence thereby enabling particularly women and the poor to negotiate about condom use with their sexual partners. Our analysis showed that women with condom use with their sexual partners. Our analysis showed that women with access to high and medium levels of structural social capital had an increased access to high and medium levels of structural social capital had an increased likelihood of using condom with casual sex partners as compared with those with likelihood of using condom with casual sex partners as compared with those with access to low levels. Gregson et al (2011) reported similar findings from Zimbabwe access to low levels. Gregson et al (2011) reported similar findings from Zimbabwe indicating that participation in community groups such as rotating credit socie- indicating that participation in community groups such as rotating credit societies and burial societies influenced women to adopt protective behaviors against ties and burial societies influenced women to adopt protective behaviors against HIV infection by enhancing self-efficacy. This is in line with a South African study HIV infection by enhancing self-efficacy. This is in line with a South African study by Campbell et al, (2002), that examined civic participation as a proxy for un- by Campbell et al, (2002), that examined civic participation as a proxy for understanding community influences on HIV infection. They showed that women derstanding community influences on HIV infection. They showed that women belonging to sports clubs had a higher likelihood to use condoms with casual belonging to sports clubs had a higher likelihood to use condoms with casual partners than non-members. partners than non-members.

The studies also showed the potential influence of cognitive social capital on sex- The studies also showed the potential influence of cognitive social capital on sexual risk behaviors among network members. Discussions about HIV risks created ual risk behaviors among network members. Discussions about HIV risks created norms that influenced others to take precaution and to use condoms with casual norms that influenced others to take precaution and to use condoms with casual sex partners. A systematic review of social capital and health found a stronger sex partners. A systematic review of social capital and health found a stronger influence of cognitive than of structural forms of social capital on health (Harpham influence of cognitive than of structural forms of social capital on health (Harpham et al, 2004; Phongsavan et al, 2006). A study among male partners of female sex et al, 2004; Phongsavan et al, 2006). A study among male partners of female sex workers in the Dominican Republic reported that pro-condom norms within male’s workers in the Dominican Republic reported that pro-condom norms within male’s social networks explained the high level of condom use, thereby providing an entry social networks explained the high level of condom use, thereby providing an entry point for HIV prevention efforts (Barrington, 2009). However, other studies have point for HIV prevention efforts (Barrington, 2009). However, other studies have reported contradictory results showing that non-use of condom especially among reported contradictory results showing that non-use of condom especially among permanent sex partners is a sign of trust. Men in particular connote use of condom permanent sex partners is a sign of trust. Men in particular connote use of condom as lack of trust and intimacy (Boulton, 1995; Middelthon, 2001). as lack of trust and intimacy (Boulton, 1995; Middelthon, 2001).

In this thesis people with access to low and medium levels of structural social In this thesis people with access to low and medium levels of structural social capital had a higher likelihood of not being tested for HIV than those with access capital had a higher likelihood of not being tested for HIV than those with access to high levels. This indicates that participation in formal and informal organiza- to high levels. This indicates that participation in formal and informal organizations may lead to increased knowledge about the importance of testing for own tions may lead to increased knowledge about the importance of testing for own benefit but also for protecting others. Such participation may result in enhanced benefit but also for protecting others. Such participation may result in enhanced self-efficacy, reduce the HIV/AIDS related stigma and influence both women and self-efficacy, reduce the HIV/AIDS related stigma and influence both women and men to test for their HIV status. A study from India (Sudha et al, 2009) has dis- men to test for their HIV status. A study from India (Sudha et al, 2009) has discussed the role of HIV/AIDS related stigma in preventing community members cussed the role of HIV/AIDS related stigma in preventing community members to test for HIV. In this study people who had high levels of structural social to test for HIV. In this study people who had high levels of structural social capital (members of associations) and high levels of cognitive social capital (high capital (members of associations) and high levels of cognitive social capital (high norms of reciprocity between neighbors and high trust to health care providers) norms of reciprocity between neighbors and high trust to health care providers) had less stigmatising attitudes and were more likely to test for HIV. had less stigmatising attitudes and were more likely to test for HIV.

45 45 DISCUSSION DISCUSSION

Social capital and increased HIV risk Social capital and increased HIV risk This thesis also highlighs that social capital can become a risk factor for HIV This thesis also highlighs that social capital can become a risk factor for HIV infection. The adherence to strict behavioral membership rules may deny some infection. The adherence to strict behavioral membership rules may deny some community members to participate in informal organizations. Many of these community members to participate in informal organizations. Many of these informal organizations also demand regular contributions of money which put informal organizations also demand regular contributions of money which put strong economic pressure on poor people. These requirements may lead to the strong economic pressure on poor people. These requirements may lead to the formation of new groups characterized by less strict rules, which may become formation of new groups characterized by less strict rules, which may become potential risk groups for HIV infection. It may also increase the gap between rich potential risk groups for HIV infection. It may also increase the gap between rich and poor, as the poor will automatically exclude themselves because of inability and poor, as the poor will automatically exclude themselves because of inability to pay. Bourdieu’s (1986) argued that social capital can be used by the well-to-do to pay. Bourdieu’s (1986) argued that social capital can be used by the well-to-do people for transmitting resources among themselves and in doing so; they may people for transmitting resources among themselves and in doing so; they may exclude members with lower resources from joining such resourceful networks. exclude members with lower resources from joining such resourceful networks. Campbell (2000, 193) emphasized that “social capital is not a homogenous re- Campbell (2000, 193) emphasized that “social capital is not a homogenous resource that is equally created, sustained, and accessed by all members of a source that is equally created, sustained, and accessed by all members of a particular community”. Despite the fact that access to different forms of social particular community”. Despite the fact that access to different forms of social capital may be unequally distributed based on gender, age and socio-economic capital may be unequally distributed based on gender, age and socio-economic status, we still argue that access to social capital has been mainly beneficial for status, we still argue that access to social capital has been mainly beneficial for the most vulnerable groups in the study area. the most vulnerable groups in the study area.

Networks for social gatherings among the youth were seen as potential risk en- Networks for social gatherings among the youth were seen as potential risk environments in the study setting. Discos attracted many youths as part of recreation vironments in the study setting. Discos attracted many youths as part of recreation and enjoyment. Since condoms were scarce particularly in the rural areas and and enjoyment. Since condoms were scarce particularly in the rural areas and youth usually drink alcohol during these night gatherings, their risk of having youth usually drink alcohol during these night gatherings, their risk of having unprotected sex was likely to increase. A study from Kenya (Njue et al, 2011) also unprotected sex was likely to increase. A study from Kenya (Njue et al, 2011) also reported that ��young men and women who were exposed to night clubs and por- reported that ��young men and women who were exposed to night clubs and por- nography in video halls had an increased risk of being HIV infected since it en- nography in video halls had an increased risk of being HIV infected since it encouraged liberal sexual attitudes and unprotected sex. couraged liberal sexual attitudes and unprotected sex.

Implications for HIV prevention strategies Implications for HIV prevention strategies There is increasing evidence that the conventional approach to health education There is increasing evidence that the conventional approach to health education using posters, televisions, radios, school lessons, fliers, seminars and workshops using posters, televisions, radios, school lessons, fliers, seminars and workshops has major limitations in reaching marginalized communities to reduce the spread has major limitations in reaching marginalized communities to reduce the spread of HIV infection. Campbell (2000) and Beeker et al (1998) advocate for a para- of HIV infection. Campbell (2000) and Beeker et al (1998) advocate for a paradigm shift from these formal channels towards a more comprehensive approach digm shift from these formal channels towards a more comprehensive approach involving local communities in promoting their own health. Social capital has involving local communities in promoting their own health. Social capital has been shown to be an important resource for creating health enabling environ- been shown to be an important resource for creating health enabling environments through its positive influence on economic development, reducing crimes ments through its positive influence on economic development, reducing crimes and mortality rate and enhancing democratic processes and health (Kawachi et and mortality rate and enhancing democratic processes and health (Kawachi et al, 1997; Reid, 2000; Pronky et al, 2008). This thesis provides additional support al, 1997; Reid, 2000; Pronky et al, 2008). This thesis provides additional support for a paradigm shift by showing that strong social capital in its structural and for a paradigm shift by showing that strong social capital in its structural and

46 46 DISCUSSION DISCUSSION cognitive forms has been an important determinant of the community response cognitive forms has been an important determinant of the community response to the HIV epidemic, particularly among the most vulnerable women and the to the HIV epidemic, particularly among the most vulnerable women and the poor in Kagera, Tanzania. This implies that policy makers and programme man- poor in Kagera, Tanzania. This implies that policy makers and programme managers should start considering including social capital in HIV interventions and agers should start considering including social capital in HIV interventions and implementation strategies to remove the barriers observed by using the conven- implementation strategies to remove the barriers observed by using the conventional approach. Policy makers and programme managers need to identify and tional approach. Policy makers and programme managers need to identify and recognize the existence of informal organizations such as social groups and recognize the existence of informal organizations such as social groups and networks and their functions at the local level. They should involve leaders of networks and their functions at the local level. They should involve leaders of these organizations in the design and delivery of health promotive HIV interven- these organizations in the design and delivery of health promotive HIV intervention activities and in conveying health information to all group and network tion activities and in conveying health information to all group and network members. members.

HIV/AIDS messages should also be conveyed directly to members of social groups HIV/AIDS messages should also be conveyed directly to members of social groups and networks through their monthly or weekly meetings. This may lead to dis- and networks through their monthly or weekly meetings. This may lead to dis- semination of health information to other family members, neighbors and the semination of health information to other family members, neighbors and the community at large. I believe that when social groups or community members community at large. I believe that when social groups or community members participate in the design and implementation of HIV interventions they are more participate in the design and implementation of HIV interventions they are more likely to be successful since participation increases community members’ power likely to be successful since participation increases community members’ power and control over their daily lives. Furthermore when community members are and control over their daily lives. Furthermore when community members are responsible for developing own norms and values, they are more likely to adopt responsible for developing own norms and values, they are more likely to adopt them and protect themselves from HIV infection (Levine et al, 1993; Reid, 2000). them and protect themselves from HIV infection (Levine et al, 1993; Reid, 2000).

The role of local governance in mobilizing social capital The role of local governance in mobilizing social capital To mobilise strong structural and cognitive social capital requires supportive To mobilise strong structural and cognitive social capital requires supportive local leadership. Social groups and networks face operational problems such as local leadership. Social groups and networks face operational problems such as conflicts among members due to delays in paying back loans. Local leaders are conflicts among members due to delays in paying back loans. Local leaders are expected to interfere and resolve such conflicts, and support these organizations expected to interfere and resolve such conflicts, and support these organizations to survive and function well. Tanzania has adopted a decentralization policy, to survive and function well. Tanzania has adopted a decentralization policy, which allows decision-making and planning on community development activi- which allows decision-making and planning on community development activities to start at the grassroots level. Local governance comprising of ward, village/ ties to start at the grassroots level. Local governance comprising of ward, village/ street and hamlet leaders may utilize their autonomy to play a linking role and street and hamlet leaders may utilize their autonomy to play a linking role and create an enabling administrative environment for collaboration between policy create an enabling administrative environment for collaboration between policy makers, programme managers, social groups and networks in their respective makers, programme managers, social groups and networks in their respective areas. Low-Beer and Sempala (2010) argue along the same line suggesting that areas. Low-Beer and Sempala (2010) argue along the same line suggesting that governance can support the mobilization of resources for HIV prevention pro- governance can support the mobilization of resources for HIV prevention programs based on social networks. grams based on social networks.

47 47 CONCLUDING REMARKS CONCLUDING REMARKS Concluding remarks Concluding remarks

Firstly, the findings from this thesis add to Putnam’s definition of social capital Firstly, the findings from this thesis add to Putnam’s definition of social capital showing that community cohesion, resulting from high levels of individual par- showing that community cohesion, resulting from high levels of individual participation in civic organizations, leads to positive aspects of community life. In- ticipation in civic organizations, leads to positive aspects of community life. In- dividual participation in bonding, bridging and linking organizations (struc- dividual participation in bonding, bridging and linking organizations (structural social capital) was associated with the adoption of positive community tural social capital) was associated with the adoption of positive community norms, values, trust and reciprocal relations (cognitive social capital), which in norms, values, trust and reciprocal relations (cognitive social capital), which in turn led to behavioral change among community members. turn led to behavioral change among community members.

Secondly, participation in civic organizations generated opportunities for vulner- Secondly, participation in civic organizations generated opportunities for vulnerable social groups, such as women and the poor, to access economic resources able social groups, such as women and the poor, to access economic resources such as loans, which empowered them economically, giving them confidence, such as loans, which empowered them economically, giving them confidence, self-esteem and decision-making power to negotiate over daily life, including self-esteem and decision-making power to negotiate over daily life, including their sexual relations. their sexual relations.

Lastly, behavioral change and empowerment enabled community members to Lastly, behavioral change and empowerment enabled community members to adopt safer sexual behaviors, thus contributing to the observed declining trends adopt safer sexual behaviors, thus contributing to the observed declining trends in HIV infection. in HIV infection.

There is no doubt that considering structural and cognitive social capital in the There is no doubt that considering structural and cognitive social capital in the national and district HIV prevention programs and strategies will increase effi- national and district HIV prevention programs and strategies will increase efficiency and effectiveness in the fight against new HIV infections especially in ciency and effectiveness in the fight against new HIV infections especially in settings where resources to implement HIV interventions are limited. settings where resources to implement HIV interventions are limited.

48 48 THE RESEARCHER THE RESEARCHER The researcher The researcher

The first contractual employment with the MoHSW in 1997 changed my ambition The first contractual employment with the MoHSW in 1997 changed my ambition to become an administrator. When I was working as a transport management to become an administrator. When I was working as a transport management consultant, I had an opportunity to travel all over the country and visited many consultant, I had an opportunity to travel all over the country and visited many hospitals, health centers and dispensaries. During these visits I observed the hospitals, health centers and dispensaries. During these visits I observed the impact of the HIV epidemic. I observed the suffering from AIDS related diseases, impact of the HIV epidemic. I observed the suffering from AIDS related diseases, met orphans and saw dilapidated houses left unattended because the owners had met orphans and saw dilapidated houses left unattended because the owners had died from AIDS. I became concerned and asked myself why HIV is spreading died from AIDS. I became concerned and asked myself why HIV is spreading more rapidly in some societies than in others. more rapidly in some societies than in others.

In 2003 I was involved in a multi-country study on “The functioning of health In 2003 I was involved in a multi-country study on “The functioning of health systems in the era of HIV/AIDS in Tanzania, Zambia and South Africa”. How- systems in the era of HIV/AIDS in Tanzania, Zambia and South Africa”. How- ever, this study did not answer my inquiry about the spread of HIV. In 2006 I ever, this study did not answer my inquiry about the spread of HIV. In 2006 I was introduced to the Kagera AIDS research project (KARP) that assessed the was introduced to the Kagera AIDS research project (KARP) that assessed the magnitude, spread, risk factors as well as the community response and social magnitude, spread, risk factors as well as the community response and social impact of the HIV epidemic in the Kagera region of Tanzania. I felt that this was impact of the HIV epidemic in the Kagera region of Tanzania. I felt that this was something for me and asked the KARP principal investigator if there was a pos- something for me and asked the KARP principal investigator if there was a possibility for me to pursue my PhD studies within the project. My request was ac- sibility for me to pursue my PhD studies within the project. My request was ac- cepted and in early 2007 I had the opportunity to meet the whole KARP research cepted and in early 2007 I had the opportunity to meet the whole KARP research team to discuss my research interest. I was asked to consider formulating a PhD team to discuss my research interest. I was asked to consider formulating a PhD proposal focusing on the role of social capital in influencing the observed HIV proposal focusing on the role of social capital in influencing the observed HIV declining trends in Kagera. This was actually the first time I heard about the declining trends in Kagera. This was actually the first time I heard about the concept of social capital. However, after reviewing the literature, I realized that concept of social capital. However, after reviewing the literature, I realized that social capital formed part of my original interest of exploring the spread of HIV social capital formed part of my original interest of exploring the spread of HIV in Tanzania. in Tanzania.

As a PhD student at the division of Epidemiology and Global health in Umeå, I As a PhD student at the division of Epidemiology and Global health in Umeå, I began my research journey focusing on the role of social capital for HIV preven- began my research journey focusing on the role of social capital for HIV prevention. I had to move from my early specialization in political science and public tion. I had to move from my early specialization in political science and public administration to the field of public health and attend courses in epidemiology, administration to the field of public health and attend courses in epidemiology, biostatistics and qualitative methods. The new knowledge upgraded my research biostatistics and qualitative methods. The new knowledge upgraded my research capacity in general and gave me opportunities to explore further public health capacity in general and gave me opportunities to explore further public health related problems and to find solutions to solve them. I agree with the greatest related problems and to find solutions to solve them. I agree with the greatest astrophysicist, Albert Einstein who said: “The significant problems we have astrophysicist, Albert Einstein who said: “The significant problems we have cannot be solved at the same level of thinking we were at when they were cre- cannot be solved at the same level of thinking we were at when they were created”. ated”.

49 49 ACKNOWLEDGEMENTS ACKNOWLEDGEMENTS Acknowledgements Acknowledgements

This work is a product of contributions from several people and organizations. This work is a product of contributions from several people and organizations. It is a result of long term collaboration between the government of Sweden through It is a result of long term collaboration between the government of Sweden through its international development agency (sida) and the government of Tanzania. its international development agency (sida) and the government of Tanzania. Different partners have been involved in the Kagera AIDS research project in- Different partners have been involved in the Kagera AIDS research project including the Department of Public Health and Clinical Medicine, Epidemiology cluding the Department of Public Health and Clinical Medicine, Epidemiology and Global Health from Sweden, Department of Epidemiology and Biostatistics, and Global Health from Sweden, Department of Epidemiology and Biostatistics, Microbiology and Immunology and Department of Development Studies from Microbiology and Immunology and Department of Development Studies from Muhimbili University of Health and Allied Sciences. Other partners included the Muhimbili University of Health and Allied Sciences. Other partners included the Kagera regional hospital and the Kagera regional medical office. The financial, Kagera regional hospital and the Kagera regional medical office. The financial, material and technical support for this project were received from the Swedish material and technical support for this project were received from the Swedish government through sida, Umeå Centre for Global Health Research through the government through sida, Umeå Centre for Global Health Research through the Swedish Council for Working Life and Social Research, the Swedish Research Swedish Council for Working Life and Social Research, the Swedish Research School for Global Health and the government of Tanzania. School for Global Health and the government of Tanzania.

I wish to express my very special thanks to everyone who in one way or another I wish to express my very special thanks to everyone who in one way or another contributed towards the accomplishment of this work. Since it is not possible to contributed towards the accomplishment of this work. Since it is not possible to express my gratitude to everyone in person and in public but I would like to take express my gratitude to everyone in person and in public but I would like to take this opportunity to mention some of people who deserve special recognition: this opportunity to mention some of people who deserve special recognition:

Associate Professor Lennarth Nystrom, main supervisor. Words cannot express Associate Professor Lennarth Nystrom, main supervisor. Words cannot express my gratitude to him for taking up the challenging work of supervising me despite my gratitude to him for taking up the challenging work of supervising me despite his heavy workload. Very special thanks for his generous support, advice and his heavy workload. Very special thanks for his generous support, advice and professional statistical guidance, especially by taking time to make a detailed professional statistical guidance, especially by taking time to make a detailed review of the tables, figures, references and formatting the manuscript. His con- review of the tables, figures, references and formatting the manuscript. His constructive criticisms, challenges and his teaching on data analysis using bivariate structive criticisms, challenges and his teaching on data analysis using bivariate and multiple logistic regressions techniques helped me to improve and shape my and multiple logistic regressions techniques helped me to improve and shape my quantitative research skills. Our relationship transcended the academic sphere quantitative research skills. Our relationship transcended the academic sphere and included social interactions; support and attention that made me feel at home and included social interactions; support and attention that made me feel at home when I am in Umea. when I am in Umea.

Professor Maria Emmelin, co-supervisor. I am deeply grateful for the stimulat- Professor Maria Emmelin, co-supervisor. I am deeply grateful for the stimulat- ing and interesting scientific discussions that we have had during the whole ing and interesting scientific discussions that we have had during the whole period of undertaking my PhD studies. Her field visits in Kagera region, Tanzania period of undertaking my PhD studies. Her field visits in Kagera region, Tanzania were very productive in improving the research instruments. Many thanks go to were very productive in improving the research instruments. Many thanks go to her for sharing her theoretical and methodological knowledge with me in a her for sharing her theoretical and methodological knowledge with me in a friendly but learning atmosphere. Her open criticisms and challenges and her friendly but learning atmosphere. Her open criticisms and challenges and her competence in qualitative research methodology has inspired me a lot and made competence in qualitative research methodology has inspired me a lot and made a lot of difference to my thinking and writing. I always regarded her as my mom a lot of difference to my thinking and writing. I always regarded her as my mom

50 50 ACKNOWLEDGEMENTS ACKNOWLEDGEMENTS because of the social support and care she gave me when I was in Umea and because of the social support and care she gave me when I was in Umea and later on in Malmo. later on in Malmo.

Professor Japhet Killewo, co-supervisor. I am greatly indebted to him for trust- Professor Japhet Killewo, co-supervisor. I am greatly indebted to him for trust- ing me and accepting my request to join the PhD program within the Kagera ing me and accepting my request to join the PhD program within the Kagera AIDS research project. I really enjoyed and learned a lot from all our joint field AIDS research project. I really enjoyed and learned a lot from all our joint field visits to the case communities where he tirelessly worked with me into the night visits to the case communities where he tirelessly worked with me into the night hours of the day. His professional guidance during field work, data cleaning and hours of the day. His professional guidance during field work, data cleaning and analysis played a crucial role in my research training. I am grateful to his wise analysis played a crucial role in my research training. I am grateful to his wise and interesting conversations especially about epidemiological knowledge, which and interesting conversations especially about epidemiological knowledge, which has helped me to improve and enriched my scientific thinking as a researcher in has helped me to improve and enriched my scientific thinking as a researcher in the public health field. I enjoyed his working style which promotes self-confidence the public health field. I enjoyed his working style which promotes self-confidence and autonomy. From him I learned what it means to be a good mentor for and autonomy. From him I learned what it means to be a good mentor for younger researchers. younger researchers.

Dr Malin Eriksson, co-author. I enjoyed her detailed review and constructive Dr Malin Eriksson, co-author. I enjoyed her detailed review and constructive comments during preparation of data collection tools, field work, analysis and comments during preparation of data collection tools, field work, analysis and writing of the publications and the thesis. I am most grateful to her for allowing writing of the publications and the thesis. I am most grateful to her for allowing me to share her theoretical and methodological knowledge on social capital is- me to share her theoretical and methodological knowledge on social capital issues, which enriched the quality and writing of the publications and the thesis. sues, which enriched the quality and writing of the publications and the thesis. Her social support when she joined me in the field in Kagera, Tanzania and dur- Her social support when she joined me in the field in Kagera, Tanzania and during the whole period of staying in Umea is highly appreciated. ing the whole period of staying in Umea is highly appreciated.

Dr Gideon Kwesigabo, co-author, Dean of School of Public Health and Social Dr Gideon Kwesigabo, co-author, Dean of School of Public Health and Social Sciences. I would like to thank him very much for his encouragement, advice and Sciences. I would like to thank him very much for his encouragement, advice and fruitful discussions which we had during the whole period of undertaking my fruitful discussions which we had during the whole period of undertaking my PhD studies. He spared time to work with me during data analysis and writing PhD studies. He spared time to work with me during data analysis and writing of the manuscripts despite his very busy schedule with school management tasks, of the manuscripts despite his very busy schedule with school management tasks, teaching and travel workload. I enjoyed his constructive and valuable epidemio- teaching and travel workload. I enjoyed his constructive and valuable epidemiological and statistical comments, which improved my skills in writing the manu- logical and statistical comments, which improved my skills in writing the manuscripts. I highly appreciate his administrative and material support as it facili- scripts. I highly appreciate his administrative and material support as it facilitated the accomplishment of this work. tated the accomplishment of this work.

Dr Sabrina Moyo, co-author, coordinator of laboratory work for the project. I Dr Sabrina Moyo, co-author, coordinator of laboratory work for the project. I am very thankful to her for ensuring that the adopted HIV testing strategies meet am very thankful to her for ensuring that the adopted HIV testing strategies meet the required and acceptable standards. I am also indebted to her for providing the required and acceptable standards. I am also indebted to her for providing the technical input into the writing process that pertains to the HIV testing strat- the technical input into the writing process that pertains to the HIV testing strategies and for being in charge of supervising the laboratory technicians who were egies and for being in charge of supervising the laboratory technicians who were testing the samples. I have learned a lot from her how collection, transportation testing the samples. I have learned a lot from her how collection, transportation and testing of blood specimens finally translate to HIV status of individuals. and testing of blood specimens finally translate to HIV status of individuals.

51 51 ACKNOWLEDGEMENTS ACKNOWLEDGEMENTS

Professor Stig Wall, former head of the division Epidemiology and Global Health. Professor Stig Wall, former head of the division Epidemiology and Global Health. I thank him very much for allowing me to join the unit to undertake my PhD I thank him very much for allowing me to join the unit to undertake my PhD studies. studies.

Professor Lars Weinehall, head of the division Epidemiology and Global Health. Professor Lars Weinehall, head of the division Epidemiology and Global Health. I am very thankful to him for the administrative and social support as well as I am very thankful to him for the administrative and social support as well as encouragement he provided during the whole period of my PhD training. encouragement he provided during the whole period of my PhD training.

Anna-Karin Hurtig, coordinator, Swedish Research School for Global Health. Anna-Karin Hurtig, coordinator, Swedish Research School for Global Health. Her consideration in providing financial support to enable me attend some PhD Her consideration in providing financial support to enable me attend some PhD courses and seminars within and outside Umea as well as visiting Malmo for courses and seminars within and outside Umea as well as visiting Malmo for consultation with my supervisor is highly appreciated. consultation with my supervisor is highly appreciated.

Nawi Ng, Anders Emmelin and John Kinsman and Miguel San Sebastian Nawi Ng, Anders Emmelin and John Kinsman and Miguel San Sebastian Chasco, my teachers in epidemiology and public health courses. I am grateful for Chasco, my teachers in epidemiology and public health courses. I am grateful for having an opportunity to share knowledge and experiences with them, which having an opportunity to share knowledge and experiences with them, which enabled me to develop more interest on epidemiology and public health field. enabled me to develop more interest on epidemiology and public health field.

Birgitta Astrom, administrator. Her competent administrative support over the Birgitta Astrom, administrator. Her competent administrative support over the years such as organizing for accommodation and travels and arrangement for years such as organizing for accommodation and travels and arrangement for dissertation is highly appreciated. dissertation is highly appreciated.

Karin Johansson, Lena Mustonen, Sabina Bergsten, and Susanne Walther, Karin Johansson, Lena Mustonen, Sabina Bergsten, and Susanne Walther, thank for their administrative support and guidance. thank for their administrative support and guidance.

Jerzy Pilch and Goran Lonnberg, their technical computer assistance and ensur- Jerzy Pilch and Goran Lonnberg, their technical computer assistance and ensuring that I work with up to date computer tools and programs made my study life ing that I work with up to date computer tools and programs made my study life very simplified. I thank them very much. very simplified. I thank them very much.

Hussein Kidanto, Isabel Goicolea, Stephen Maluka, Barnabas Njozing, Felix Ki- Hussein Kidanto, Isabel Goicolea, Stephen Maluka, Barnabas Njozing, Felix Ki- sanga, Rose Laisser, Hailemariam Lemma, Fredinah Namatovu, Nkulu Kalengayi sanga, Rose Laisser, Hailemariam Lemma, Fredinah Namatovu, Nkulu Kalengayi Faustine, Fredrik Norstrom, Anticona Cynthia, Hernandez Alison, Ruano Ana Faustine, Fredrik Norstrom, Anticona Cynthia, Hernandez Alison, Ruano Ana Lorena, Elli Nur Hayati, Hendrew Gekawaku Lusey, Utamie Puji Lestari, Fatwa Lorena, Elli Nur Hayati, Hendrew Gekawaku Lusey, Utamie Puji Lestari, Fatwa Sari Tetra Dewi, Yien Ling Hii, former and current PhD colleagues and friends. Sari Tetra Dewi, Yien Ling Hii, former and current PhD colleagues and friends. Their social support and interesting discussions and encouragement about PhD Their social support and interesting discussions and encouragement about PhD studies have meant a lot to me over the years. I am grateful for everything we shared studies have meant a lot to me over the years. I am grateful for everything we shared during the whole period we have been together in Umea, Stockholm and other during the whole period we have been together in Umea, Stockholm and other places. places.

Dr George Kichau and his family, thank you very much for their generosity, Dr George Kichau and his family, thank you very much for their generosity, kindness, solidarity, cooperation and social and material support, all of which kindness, solidarity, cooperation and social and material support, all of which

52 52 ACKNOWLEDGEMENTS ACKNOWLEDGEMENTS made me feel at home during the whole period of staying in Umea. made me feel at home during the whole period of staying in Umea. The Kagera medical and regional hospital officials at various times:Dr Pius Tubeti The Kagera medical and regional hospital officials at various times:Dr Pius Tubeti (former regional medical officer), Dr. Juma D. Nyakina (regional hospital (former regional medical officer), Dr. Juma D. Nyakina (regional hospital medical officer in charge),Mr . Herman E. Kabirigi (regional health officer and medical officer in charge),Mr . Herman E. Kabirigi (regional health officer and acting regional medical officer),Mr . Renatus Ndyetabura (regional hospital ac- acting regional medical officer),Mr . Renatus Ndyetabura (regional hospital ac- countant), Honoratha Mutembei (regional hospital matron), Justus Bengesi countant), Honoratha Mutembei (regional hospital matron), Justus Bengesi (regional hospital secretary), for their continuous and tirelessly support and co- (regional hospital secretary), for their continuous and tirelessly support and cooperation at all stages of undertaking this work. I am grateful indebted for release operation at all stages of undertaking this work. I am grateful indebted for release of the regional hospital staff even at times when they were seriously facing shortage. of the regional hospital staff even at times when they were seriously facing shortage.

Former and current field coordinators; Lilian Kakwesigabo, Dr. Simeon Nya- Former and current field coordinators; Lilian Kakwesigabo, Dr. Simeon Nya- benda, and Dr. Khairoonisa Pathan. I highly appreciate for their excellent per- benda, and Dr. Khairoonisa Pathan. I highly appreciate for their excellent performance in administering and coordinating field activities, which was very formance in administering and coordinating field activities, which was very crucial in making this piece of work a success. crucial in making this piece of work a success.

Kagera regional hospital laboratory officials,Audax Ambakileki, Rukaiya Rush- Kagera regional hospital laboratory officials,Audax Ambakileki, Rukaiya Rush- aka, Joanita Kamando and Amina Bagoka. I thank them for organizing and aka, Joanita Kamando and Amina Bagoka. I thank them for organizing and storing the blood samples according to the specified standards. storing the blood samples according to the specified standards.

Field assistants; Anastazia Ngemera (team leader), Anacletus Kalugwa, Gloria Field assistants; Anastazia Ngemera (team leader), Anacletus Kalugwa, Gloria Obadiah Kagaruki, Jesephina Lutanjuka, Jesca Daudi, Laurencia Rugemalira, Obadiah Kagaruki, Jesephina Lutanjuka, Jesca Daudi, Laurencia Rugemalira, Leonardina Mbaiya, Philomena Lugaimukamu, Specioza Ruhikula, Winifrida Leonardina Mbaiya, Philomena Lugaimukamu, Specioza Ruhikula, Winifrida Maxmilian and Zelda Rutalemwa. Drivers were Mustafa, Medard and Bitumbe. Maxmilian and Zelda Rutalemwa. Drivers were Mustafa, Medard and Bitumbe. I sincerely thank them for hard working and effective spirit of team work, some- I sincerely thank them for hard working and effective spirit of team work, sometimes they worked under extremely difficult conditions including rain season but times they worked under extremely difficult conditions including rain season but they never despaired. I want them to know that I enjoyed working with them and they never despaired. I want them to know that I enjoyed working with them and that their contribution in accomplishing this thesis is highly appreciated. that their contribution in accomplishing this thesis is highly appreciated.

Aida Mkumbo, data entry clerk, I highly appreciate for her hard working during Aida Mkumbo, data entry clerk, I highly appreciate for her hard working during official and non-official working hours in entering and cleaning data. She made official and non-official working hours in entering and cleaning data. She made my research work easy for producing the final data on time. my research work easy for producing the final data on time.

Dr. Mughwira Mwangu, head, department of Development Studies, his admin- Dr. Mughwira Mwangu, head, department of Development Studies, his administrative support and accepting to allocate my teaching responsibilities to my istrative support and accepting to allocate my teaching responsibilities to my fellow academic staff despite the existing problem of shortage of staff has con- fellow academic staff despite the existing problem of shortage of staff has contributed to timely and successful accomplishment of this work. He always inspires tributed to timely and successful accomplishment of this work. He always inspires me with his wise and leadership ways of making rational decisions. me with his wise and leadership ways of making rational decisions.

Professor Angwara Kiwara, Drs Ave Maria semakafu, Tumaini Nyamhanga, Professor Angwara Kiwara, Drs Ave Maria semakafu, Tumaini Nyamhanga, Eunice chomi, Mr Amani Anaeli and Derick Chitama and Miss Glads Mahiti, Eunice chomi, Mr Amani Anaeli and Derick Chitama and Miss Glads Mahiti,

53 53 ACKNOWLEDGEMENTS ACKNOWLEDGEMENTS my fellow academic staff members. I thank them for wise decision of taking care my fellow academic staff members. I thank them for wise decision of taking care of my teaching and academic responsibilities during my absence either in the of my teaching and academic responsibilities during my absence either in the field or writing of this thesis. Such decision has shown a high level of solidarity field or writing of this thesis. Such decision has shown a high level of solidarity that will remain in my memory forever. Without such valuable decisions, it would that will remain in my memory forever. Without such valuable decisions, it would have been difficult for me to accomplish this work. have been difficult for me to accomplish this work.

MUHAS vice-chancellor, Professor Kisali Palangyo, deputy vice-chancellor, MUHAS vice-chancellor, Professor Kisali Palangyo, deputy vice-chancellor, academic, research and publications, Professor Eligius Lyamuya, Director re- academic, research and publications, Professor Eligius Lyamuya, Director research and publications, Professor Aboud Muhsin, former and current HIV/TB search and publications, Professor Aboud Muhsin, former and current HIV/TB Sub-programme coordinator, Professor Fred Mhalu and Mohammad Bakari Sub-programme coordinator, Professor Fred Mhalu and Mohammad Bakari and staff at Vice-chancellor’s office. I sincerely thank them for administrative and staff at Vice-chancellor’s office. I sincerely thank them for administrative and material support and granting permission to accomplish research activities and material support and granting permission to accomplish research activities and the writing of this thesis. and the writing of this thesis.

Amerdeus Ringia Samki, Senior medical illustrator, MUHAS. I highly appreciate Amerdeus Ringia Samki, Senior medical illustrator, MUHAS. I highly appreciate for taking time to draw the graphic for the cover story. for taking time to draw the graphic for the cover story.

Special thanks goes to my parents Frumence Luca Msoffe and Leonia Claudi Special thanks goes to my parents Frumence Luca Msoffe and Leonia Claudi Mchome who raised me, took me to the school and taught me the importance of Mchome who raised me, took me to the school and taught me the importance of education, otherwise without them I could not have reached where I am today. education, otherwise without them I could not have reached where I am today.

Lastly, my deepest and heartfelt gratitude goes to my lovely family, particularly Lastly, my deepest and heartfelt gratitude goes to my lovely family, particularly my wife, Diana Martin Kasonga and our children Lorraine, Laura and Larry. my wife, Diana Martin Kasonga and our children Lorraine, Laura and Larry. I owe my loving appreciation to my wife, Diana, since it was not easy to be away I owe my loving appreciation to my wife, Diana, since it was not easy to be away from the family for a long period, her love, patience and understanding, encour- from the family for a long period, her love, patience and understanding, encouragement and constant prayers made my dream come true. Most importantly, my agement and constant prayers made my dream come true. Most importantly, my sincere appreciation goes to her for assuming fatherly role to our children when sincere appreciation goes to her for assuming fatherly role to our children when I was away. To our lovely children, Lorraine, Laura and Larry, their love and I was away. To our lovely children, Lorraine, Laura and Larry, their love and constant prayers to their dad’s studies made invaluable contributions. They missed constant prayers to their dad’s studies made invaluable contributions. They missed me most of the time when I was away but still they never complained. It is because me most of the time when I was away but still they never complained. It is because of all this that I dedicate this thesis to my family. of all this that I dedicate this thesis to my family.

54 54 REFERENCES REFERENCES References References Bagachwa, M.S.D. (1992). Background, evolution, essence and prospects of current economic Bagachwa, M.S.D. (1992). Background, evolution, essence and prospects of current economic reforms in Tanzania, pp 19-43. In Bagachwa, M.S.D., Mbelle, A.V.Y., Van Arkadie, B. (Eds), reforms in Tanzania, pp 19-43. In Bagachwa, M.S.D., Mbelle, A.V.Y., Van Arkadie, B. (Eds), Market reforms and parastatal restructuring in Tanzania. Economics Department and Market reforms and parastatal restructuring in Tanzania. Economics Department and Economic Research Bureau, University of Dar es Salaam, Dar es Salaam, Tanzania. Economic Research Bureau, University of Dar es Salaam, Dar es Salaam, Tanzania. Bandura, A. (1996). Self-efficacy in changing societies. Cambridge: Cambridge University Bandura, A. (1996). Self-efficacy in changing societies. Cambridge: Cambridge University Press. Press. Baum, F.E., Ziersch, A.M. (2003). Glossary: Social capital. Journal of Epidemiology and Com- Baum, F.E., Ziersch, A.M. (2003). Glossary: Social capital. Journal of Epidemiology and Com- munity Health. 57:320–323 munity Health. 57:320–323 Baum, F. (1999). Social capital: Is it good for your health? Issues for public health agenda. Baum, F. (1999). Social capital: Is it good for your health? Issues for public health agenda. Journal of Epidemiology and Community Health. 53:195-196. Journal of Epidemiology and Community Health. 53:195-196. Beeker, C., Guenther-Grey, C., Raj, A. (1998). Community Empowerment Paradigm Drift and Beeker, C., Guenther-Grey, C., Raj, A. (1998). Community Empowerment Paradigm Drift and the Primary Prevention of HIV/AIDS. Social Science and Medicine. 46: 831-42. the Primary Prevention of HIV/AIDS. Social Science and Medicine. 46: 831-42. Berkam, L.F., Glass T . (2001). Social integration, social networks, social support and health. Berkam, L.F., Glass T . (2001). Social integration, social networks, social support and health. In Berkam LF & Kawachi I (Eds). Social epidemiology (pp 137-173). New York: Oxford In Berkam LF & Kawachi I (Eds). Social epidemiology (pp 137-173). New York: Oxford University Press. University Press. Bloom, S.S., Urassa, M., Isingo, R., Ng’weshemi, J.J., Boerma, J.T. (2002). Community effects Bloom, S.S., Urassa, M., Isingo, R., Ng’weshemi, J.J., Boerma, J.T. (2002). Community effects on the risk of HIV infection in rural Tanzania. Sexually Transmitted Infections. 78:261–266. on the risk of HIV infection in rural Tanzania. Sexually Transmitted Infections. 78:261–266. Boulton, M.J., McLean, R.F., Jlart, G. (1995). Gay men’s accounts of unsafe sex. AIDS Boulton, M.J., McLean, R.F., Jlart, G. (1995). Gay men’s accounts of unsafe sex. AIDS Care.7:619-630. Care.7:619-630. Bourdieu, P. (1985). The social space and the genesis of groups. Social Science Information. Bourdieu, P. (1985). The social space and the genesis of groups. Social Science Information. 24:195-220. 24:195-220. Bourdieu, P. (1986). The forms of social capital. In Richardson, J. (Ed). Handbook of theory and Bourdieu, P. (1986). The forms of social capital. In Richardson, J. (Ed). Handbook of theory and research for the sociology of education. New York: Greenwood Press. research for the sociology of education. New York: Greenwood Press. Campbell, C., Wood, R., Kelly, M.K. (1999). Social capital and health. London: Health Educa- Campbell, C., Wood, R., Kelly, M.K. (1999). Social capital and health. London: Health Educa- tion Authority. tion Authority. Campbell, C. (2000). Social capital and health: contextualizing health promotion within local Campbell, C. (2000). Social capital and health: contextualizing health promotion within local community networks. In Baron, S., Field, J., Schuller, T. (Eds). Social capital: critical community networks. In Baron, S., Field, J., Schuller, T. (Eds). Social capital: critical perspectives. Oxford: Oxford University Press, p 182-196. perspectives. Oxford: Oxford University Press, p 182-196. Campbell, C., MacPhail, C. (2002). Peer education, gender and development of critical con- Campbell, C., MacPhail, C. (2002). Peer education, gender and development of critical consciousness: participatory HIV prevention by South African youth. Social Science and sciousness: participatory HIV prevention by South African youth. Social Science and Medicine. 55:331-345. Medicine. 55:331-345. Campbell, l.C., Williams, B., Gilgen, D. (2002). Is social capital a useful conceptual tool for Campbell, l.C., Williams, B., Gilgen, D. (2002). Is social capital a useful conceptual tool for exploring community level influences on HIV infection? An exploratory case study from exploring community level influences on HIV infection? An exploratory case study from South Africa. AIDS Care. 14:41–54. South Africa. AIDS Care. 14:41–54. Campbell, C. (2003) ‛Letting them die’ why HIV/AIDS prevention programmes fail. Oxford: Campbell, C. (2003) ‛Letting them die’ why HIV/AIDS prevention programmes fail. Oxford: James Currey. James Currey. Campbell, C. (2011). Embracing complexity: towards more nuanced understandings of social Campbell, C. (2011). Embracing complexity: towards more nuanced understandings of social capital and health: commentary. Global Health Action. 4:3402. capital and health: commentary. Global Health Action. 4:3402. Charmaz, K. (2006). Constructing grounded theory. Thousand Oaks, CA: Sage. Charmaz, K. (2006). Constructing grounded theory. Thousand Oaks, CA: Sage. Coleman, J.S. (1988). Social capital in the creation of human capital. American Journal of Coleman, J.S. (1988). Social capital in the creation of human capital. American Journal of Sociology. 94:95-121. Sociology. 94:95-121.

55 55 REFERENCES REFERENCES

Coleman, J.S. (1990). Foundations of social theory. Cambridge: Belknap Press of Harvard Coleman, J.S. (1990). Foundations of social theory. Cambridge: Belknap Press of Harvard University Press. University Press. Creswell, J.W. (1998). Qualitative inquiry and research design: Choosing among five ap- Creswell, J.W. (1998). Qualitative inquiry and research design: Choosing among five approaches. London: Sage Publications. proaches. London: Sage Publications. Creswell, J.W. (2009). Research Design: Qualitative, Quantitative, and Mixed Approaches. Creswell, J.W. (2009). Research Design: Qualitative, Quantitative, and Mixed Approaches. Thousand Oaks, CA: Sage. Thousand Oaks, CA: Sage. Creswell, J.W. (2008). Educational research planning, conducting and evaluating quantitative Creswell, J.W. (2008). Educational research planning, conducting and evaluating quantitative and qualitative research. New Jersey, USA: International Pearson Merrill Prentice Hall. and qualitative research. New Jersey, USA: International Pearson Merrill Prentice Hall. Dahlgren, L., Emmelin, M., Winkvist, A. (2007). Qualitative methodology for international Dahlgren, L., Emmelin, M., Winkvist, A. (2007). Qualitative methodology for international public health. Umea, Sweden. Umea University. public health. Umea, Sweden. Umea University. DeCoster, J. (1998). Overview of factor analysis. Available at http://www.stat-help.com/ DeCoster, J. (1998). Overview of factor analysis. Available at http://www.stat-help.com/ notes.html. Retrieved October 4th 2011 notes.html. Retrieved October 4th 2011 Dudwick, N., Kuehnast, K., Jones, V.N., Woolcock, M. (2006). Analyzing social capital in context: Dudwick, N., Kuehnast, K., Jones, V.N., Woolcock, M. (2006). Analyzing social capital in context: a guide to using qualitative methods and data. Washington DC, US: World Bank Institute. a guide to using qualitative methods and data. Washington DC, US: World Bank Institute. Durkheim, E. (1951). Suicide: A study in sociology. Glencoe, IL: The Free Press. Durkheim, E. (1951). Suicide: A study in sociology. Glencoe, IL: The Free Press. Einstein, A. (2000). The expanded quotable Einstein. New Jersey, US: Princeton University Einstein, A. (2000). The expanded quotable Einstein. New Jersey, US: Princeton University Press. Press. Epstein, H. (2007). The invisible cure: Africa, the west and the fight against AIDS. London: Epstein, H. (2007). The invisible cure: Africa, the west and the fight against AIDS. London: Viking/Penguin. Viking/Penguin. Eriksson, M. (2010). Social capital, health and community action: Implications for health Eriksson, M. (2010). Social capital, health and community action: Implications for health promotion. PhD thesis, department of Public Health and Clinical Medicine, Umeå Univer- promotion. PhD thesis, department of Public Health and Clinical Medicine, Umeå Univer- sity, Umeå, Sweden, p 83-91. sity, Umeå, Sweden, p 83-91. Eriksson, B. (1988). The relationnal basis of attitudes. Social structures: a network approach. Eriksson, B. (1988). The relationnal basis of attitudes. Social structures: a network approach. S. Berkowitz. New York. Cambridge University Press. S. Berkowitz. New York. Cambridge University Press. Eriksson, E., Lindmark, G., Haddad, D., Axemo, P. (2011). Involvement of religious leaders in Eriksson, E., Lindmark, G., Haddad, D., Axemo, P. (2011). Involvement of religious leaders in HIV prevention, South Africa. Swedish Missiological Themes. 99:2. HIV prevention, South Africa. Swedish Missiological Themes. 99:2. Flap, H.D. (2004). Creation and returns of social capital, p 2-24. In Flap H.D., Volker, B. (Eds.): Flap, H.D. (2004). Creation and returns of social capital, p 2-24. In Flap H.D., Volker, B. (Eds.): Creation and returns of social capital. London: Routledge. Creation and returns of social capital. London: Routledge. Foster, G. (2006). Children who live in communities affected by AIDS. The Lancet. 367:700- Foster, G. (2006). Children who live in communities affected by AIDS. The Lancet. 367:700- 701. 701. Franke, S. (2005). Measurement of Social Capital: Reference Document for Public Policy Franke, S. (2005). Measurement of Social Capital: Reference Document for Public Policy Research, Policy Initiative project, Canada. Research, Policy Initiative project, Canada. Frumence, G., Eriksson, M., Killewo, J., Nyström, L., Emmelin, M. Exploring the role of cog- Frumence, G., Eriksson, M., Killewo, J., Nyström, L., Emmelin, M. Exploring the role of cognitive and structural social capital in the declining trends of HIV/AIDS in the Kagera region nitive and structural social capital in the declining trends of HIV/AIDS in the Kagera region of Tanzania – A grounded theory study. African Journal of Aids Research 2011;10:1-13. of Tanzania – A grounded theory study. African Journal of Aids Research 2011;10:1-13. Giddens, A. (1973). The class structure of advanced societies. London: Hutchinson. Giddens, A. (1973). The class structure of advanced societies. London: Hutchinson. Gilbert, L. (2009). Analyzing the dark side of social capital: organized crime in Russia. In Cox, Gilbert, L. (2009). Analyzing the dark side of social capital: organized crime in Russia. In Cox, M. (ed). Social capital and peace-building: Creating and resolving conflict with trust and M. (ed). Social capital and peace-building: Creating and resolving conflict with trust and social networks. New York: Routledge. social networks. New York: Routledge. Gregson, S., Zhuwau, T., Ndlovu, J., Nyamukapa, C. (2002). Methods to reduce social desir- Gregson, S., Zhuwau, T., Ndlovu, J., Nyamukapa, C. (2002). Methods to reduce social desir- ability bias in sex surveys in low-development settings: experience in Zimbabwe. Sexually ability bias in sex surveys in low-development settings: experience in Zimbabwe. Sexually Transmitted Disease. 29:568-585. Transmitted Disease. 29:568-585.

56 56 REFERENCES REFERENCES

Gregson, S., Terceira, N., Mushati, P., Nyamukapa, C., Campbell, C. (2004). Community group Gregson, S., Terceira, N., Mushati, P., Nyamukapa, C., Campbell, C. (2004). Community group participation: can it help young women to avoid HIV? An exploratory study of social capi- participation: can it help young women to avoid HIV? An exploratory study of social capital and school education in rural Zimbabwe. Social Science and Medicine. 58:2119-2132. tal and school education in rural Zimbabwe. Social Science and Medicine. 58:2119-2132. Gregson, S., Mushati, P., Grusin, H., Nhamo, M., Schumacher, C., Skovdal, M., Nyamukapa, Gregson, S., Mushati, P., Grusin, H., Nhamo, M., Schumacher, C., Skovdal, M., Nyamukapa, C., Campbell, C. (2011). Social Capital and Reduced Female Vulnerability to HIV infection C., Campbell, C. (2011). Social Capital and Reduced Female Vulnerability to HIV infection in Rural Zimbabwe. Population Development Review. 37: 333-359 in Rural Zimbabwe. Population Development Review. 37: 333-359 Greiner, K. A., Li, C., Kawachi, I., Hunt, D. C., Ahluwalia, J. S. (2004). The relationships of Greiner, K. A., Li, C., Kawachi, I., Hunt, D. C., Ahluwalia, J. S. (2004). The relationships of social participation and community ratings to health and health behaviors in areas with social participation and community ratings to health and health behaviors in areas with high and low population density. Social Science and Medicine. 59: 2303-2312. high and low population density. Social Science and Medicine. 59: 2303-2312. Grootaert, C., Bastelaer, T. (2001). Understanding and measuring social capital: a synthesis Grootaert, C., Bastelaer, T. (2001). Understanding and measuring social capital: a synthesis of findings and recommendations from the social capital initiative. Washington DC: The of findings and recommendations from the social capital initiative. Washington DC: The World Bank. World Bank. Grootaert, C., Narayan, D., Nyhan, J.C.V., Woolcock, M. (2004). Measuring social capital. An Grootaert, C., Narayan, D., Nyhan, J.C.V., Woolcock, M. (2004). Measuring social capital. An integrated questionnaire. Washington: The World Bank. integrated questionnaire. Washington: The World Bank. Haralambos, M., Holborn, M. (1990). Sociology: themes and perspectives. 3rd. ed. London: Haralambos, M., Holborn, M. (1990). Sociology: themes and perspectives. 3rd. ed. London: Unwin Hyman. Unwin Hyman. Harper, R. (2002). The measurement of social capital in the United Kingdom. London: Office Harper, R. (2002). The measurement of social capital in the United Kingdom. London: Office of the National Statistics. of the National Statistics. Harpham, T., Grant, E., Thomas, E. (2002). Measuring social capital within health surveys: Harpham, T., Grant, E., Thomas, E. (2002). Measuring social capital within health surveys: Key issues. Health Policy and Planning. 17:106-111. Key issues. Health Policy and Planning. 17:106-111. Harpham, T., Grant, T., Rodriguez, C. (2004). Mental health and social capital in Cali, Colum- Harpham, T., Grant, T., Rodriguez, C. (2004). Mental health and social capital in Cali, Colum- bia. Social Science and Medicine. 58:2267-2277. bia. Social Science and Medicine. 58:2267-2277. Hawe, P., Shiell, A. (2000). Social capital and health promotion. A review. Social Science and Hawe, P., Shiell, A. (2000). Social capital and health promotion. A review. Social Science and Medicine. 51:871-885. Medicine. 51:871-885. Heise, L. L., Elias, C. (1995). Transforming AIDS prevention to meet women’s needs: a focus Heise, L. L., Elias, C. (1995). Transforming AIDS prevention to meet women’s needs: a focus of developing countries. Social science and medicine. 40:931-943. of developing countries. Social science and medicine. 40:931-943. Hitt, M.A., Lee, H.U., Yucel, E. (2002). The importance of social capital to the management Hitt, M.A., Lee, H.U., Yucel, E. (2002). The importance of social capital to the management of multi-national enterprises: Relational networks among Asian and Western firms. Asia of multi-national enterprises: Relational networks among Asian and Western firms. Asia Pacific Journal of Management.19:353-368. Pacific Journal of Management.19:353-368. Howitt, D., Cramer, D. (2003). An introduction to statistics in psychology: A complete guide Howitt, D., Cramer, D. (2003). An introduction to statistics in psychology: A complete guide for students (revised 2nd ed.). Hemel Hempstead: Prentice Hall. for students (revised 2nd ed.). Hemel Hempstead: Prentice Hall. Hyyppa, M.T., Maki, J. (2003). Social participation and health in a community rich of social Hyyppa, M.T., Maki, J. (2003). Social participation and health in a community rich of social capital. Health Education Research. 18:770-779. capital. Health Education Research. 18:770-779. Israel, B., Checkoway, B., Schulz, A., Zimmerman, M. (1994). Health education and commu- Israel, B., Checkoway, B., Schulz, A., Zimmerman, M. (1994). Health education and community empowerment: conceptualizing and measuring perceptions of individual, organiza- nity empowerment: conceptualizing and measuring perceptions of individual, organizational and community control. Health Education Quarterly. 21,149-170. tional and community control. Health Education Quarterly. 21,149-170. Jamil, I., Muriisa, R. (2004). Building social capital in Uganda: the role of NGOs inmitigating Jamil, I., Muriisa, R. (2004). Building social capital in Uganda: the role of NGOs inmitigating HIV/AIDS challenges. Paper presented at the International conference in Toronto 11–14 HIV/AIDS challenges. Paper presented at the International conference in Toronto 11–14 July. http://www.istr.org/conferences/toronto/workingpapers/jamil.ishtiaq.pdf. Retrieved July. http://www.istr.org/conferences/toronto/workingpapers/jamil.ishtiaq.pdf. Retrieved September 2009. September 2009. Jesus, M., Garcia Callejaa, G., Walkera, N., Cuchib, P., Lazzaric, S., Ghysa, P.D., Zacariasb, F. Jesus, M., Garcia Callejaa, G., Walkera, N., Cuchib, P., Lazzaric, S., Ghysa, P.D., Zacariasb, F. (2002). Status of the HIV/AIDS epidemic and methods to monitor it in the Latin America (2002). Status of the HIV/AIDS epidemic and methods to monitor it in the Latin America and Caribbean region. AIDS. 16 (suppl 3):3–12. and Caribbean region. AIDS. 16 (suppl 3):3–12.

57 57 REFERENCES REFERENCES

Kawachi, I., Kenedy, B., Louchner, K., Prothrow-Stith, D. (1997). Social capital, income inequal- Kawachi, I., Kenedy, B., Louchner, K., Prothrow-Stith, D. (1997). Social capital, income inequal- ity and mortality. American Journal of Public Health. 87:1491–1498. ity and mortality. American Journal of Public Health. 87:1491–1498. Kawachi, I., Kennedy, B., Glass, R. (1999). Social capital and self-rated health: A contextual Kawachi, I., Kennedy, B., Glass, R. (1999). Social capital and self-rated health: A contextual analysis. American Journal of Public Health. 89:1187-1193. analysis. American Journal of Public Health. 89:1187-1193. Kawachi, I., Berkman, L. F. (2001). Social ties and mental health. Journal of Urban Health. Kawachi, I., Berkman, L. F. (2001). Social ties and mental health. Journal of Urban Health. 78:458-467. 78:458-467. Kiama W. (1999). Where are Kenya’s homosexuals? AIDS Analysis Africa. 9:9–10. Kiama W. (1999). Where are Kenya’s homosexuals? AIDS Analysis Africa. 9:9–10. Killewo, J., Nyamuryekunge, K., Sandstrom, A., Bredberg-Raden, U., Wall, S., Mhalu, F., Killewo, J., Nyamuryekunge, K., Sandstrom, A., Bredberg-Raden, U., Wall, S., Mhalu, F., Biberfeld, G. (1990). Prevalence of HIV-1 infection in the Kagera region of Tanzania: a Biberfeld, G. (1990). Prevalence of HIV-1 infection in the Kagera region of Tanzania: a population-based study. AIDS. 4:1081-1085. population-based study. AIDS. 4:1081-1085. Killewo, J., Sandstrom, A., Bredberg-Raden, U., Mhalu, F., Biberfeld, G., Wall, S. (1993). In- Killewo, J., Sandstrom, A., Bredberg-Raden, U., Mhalu, F., Biberfeld, G., Wall, S. (1993). In- cidence of HIV-1 infection among adults in the Kagera region of Tanzania. International cidence of HIV-1 infection among adults in the Kagera region of Tanzania. International Journal of Epidemiology. 23:528-536. Journal of Epidemiology. 23:528-536. Killewo, J. (1994). Epidemiology towards the control of HIV infection in Tanzania with spe- Killewo, J. (1994). Epidemiology towards the control of HIV infection in Tanzania with special reference to the Kagera region. PhD thesis. Department of Public Health and Clinical cial reference to the Kagera region. PhD thesis. Department of Public Health and Clinical Medicine, Umea University, Umea, Sweden, p 37-38. Medicine, Umea University, Umea, Sweden, p 37-38. Krishna, A., Uphoff, N. (2002). Mapping and measuring social capital through assessment of Krishna, A., Uphoff, N. (2002). Mapping and measuring social capital through assessment of colective action for concerving and developing watershed in Rajasthan, India, p 85–88, colective action for concerving and developing watershed in Rajasthan, India, p 85–88, 115–124. In Bastelaer, T.V. (Ed). Role of social capital in development: an empirical assess- 115–124. In Bastelaer, T.V. (Ed). Role of social capital in development: an empirical assessment. New York: Cambridge University press. ment. New York: Cambridge University press. Kwesigabo, G. (2001). Trends of HIV infection in the Kagera region of Tanzania 1987-2000. Kwesigabo, G. (2001). Trends of HIV infection in the Kagera region of Tanzania 1987-2000. PhD Thesis. Department of Public Health and Clinical Medicine, Umea University, Umea, PhD Thesis. Department of Public Health and Clinical Medicine, Umea University, Umea, Sweden, p 26-32. Sweden, p 26-32. Levine, O., Britton, P., James, T., Jackson, A., Hobfoll, S., Lavin, J. (1993). The empowerment Levine, O., Britton, P., James, T., Jackson, A., Hobfoll, S., Lavin, J. (1993). The empowerment of women: A key to HIV prevention. Journal of Community Psychology. 21:320-334. of women: A key to HIV prevention. Journal of Community Psychology. 21:320-334. Lin, N. (2001). Social capital: A theory of social structure and action. Cambridge: Cambridge Lin, N. (2001). Social capital: A theory of social structure and action. Cambridge: Cambridge University Press. University Press. Lindström, M. (2003). Social capital and the miniaturization of community among daily and Lindström, M. (2003). Social capital and the miniaturization of community among daily and intermittent smokers: A population based study. Preventive Medicine. 36:177-184. intermittent smokers: A population based study. Preventive Medicine. 36:177-184. Lindström, M., Moghaddassi, M., Merlo, J. (2004). Individual self-reported health, social Lindström, M., Moghaddassi, M., Merlo, J. (2004). Individual self-reported health, social participation and neighborhood: a multilevel analysis in Malmö, Sweden. Preventive participation and neighborhood: a multilevel analysis in Malmö, Sweden. Preventive Medicine.39:135-141. Medicine.39:135-141. Low-Beer, D., Sempala, M. (2010). Social capital and effective HIV prevention: Community Low-Beer, D., Sempala, M. (2010). Social capital and effective HIV prevention: Community responses. Global Health Governance. 4:1-18. responses. Global Health Governance. 4:1-18. Lugalla, J., Emmelin, M., Mutembei, A., Sima, M., Kwesigabo, G., Killewo, J., Dahlgren, L. Lugalla, J., Emmelin, M., Mutembei, A., Sima, M., Kwesigabo, G., Killewo, J., Dahlgren, L. (2004). Social, cultural and sexual behavioral determinants of observed decline in HIV (2004). Social, cultural and sexual behavioral determinants of observed decline in HIV infection trends: lessons from the Kagera region, Tanzania. Social Science and Medicine. infection trends: lessons from the Kagera region, Tanzania. Social Science and Medicine. 59:185-198. 59:185-198. Macinko, J., Startfield, B. (2001). The utility of social capital in research and health determi- Macinko, J., Startfield, B. (2001). The utility of social capital in research and health determinants. Milbank Quarterly. 79:387-427. nants. Milbank Quarterly. 79:387-427. Maliyamkono, T.L., Bagachwa, M.S.D. (1990). The second economy in Tanzania. Ohio: Ohio Maliyamkono, T.L., Bagachwa, M.S.D. (1990). The second economy in Tanzania. Ohio: Ohio University Press. University Press.

58 58 REFERENCES REFERENCES

Middelthon, A. (2001). Interpretations of condom use and nonuse among young Norwegian Middelthon, A. (2001). Interpretations of condom use and nonuse among young Norwegian gay men: A qualitative study. Medical Anthropology Quarterly. 15:58-83. gay men: A qualitative study. Medical Anthropology Quarterly. 15:58-83. Narayan, D., Pritchett, L. (Eds). (1997). Cents and sociability: Household income and social Narayan, D., Pritchett, L. (Eds). (1997). Cents and sociability: Household income and social capital in rural Tanzania. Washington DC: World Bank. capital in rural Tanzania. Washington DC: World Bank. Narayan, D., Cassidy, M.F. (2001). A dimensional approach to measuring social capital: de- Narayan, D., Cassidy, M.F. (2001). A dimensional approach to measuring social capital: development and validation of a social capital. Inventory. Current Sociology. 49:59-102 velopment and validation of a social capital. Inventory. Current Sociology. 49:59-102 NBS (National Bureau of Statistics). (2003). Population census 2002 report. United Republic NBS (National Bureau of Statistics). (2003). Population census 2002 report. United Republic of Tanzania, Dar es salaam, Tanzania. of Tanzania, Dar es salaam, Tanzania. NBS (National Bureau of Statistics). (2008). Tanzania HIV and malaria indicator survey NBS (National Bureau of Statistics). (2008). Tanzania HIV and malaria indicator survey report 2007-08. United Republic of Tanzania, Dar es Salaam. report 2007-08. United Republic of Tanzania, Dar es Salaam. Njue, C., Voeten., H.A.C.M., Remes, P. (2011). Porn video shows, local brew, and transac- Njue, C., Voeten., H.A.C.M., Remes, P. (2011). Porn video shows, local brew, and transactional sex: HIV risk among youth in Kisumu, Kenya. BMC Public Health. 11:635 tional sex: HIV risk among youth in Kisumu, Kenya. BMC Public Health. 11:635 Nombo, C.I. (2007) When AIDS meets poverty: Implications for social capital in a village in Nombo, C.I. (2007) When AIDS meets poverty: Implications for social capital in a village in Tanzania. AWLAE SERIES No. 5. PhD thesis, The Netherlands, Wageningen Academic Tanzania. AWLAE SERIES No. 5. PhD thesis, The Netherlands, Wageningen Academic Publishers. Publishers. Nyerere, J. (1967). Freedom and Unity. London & Nairobi: Oxford University Press Nyerere, J. (1967). Freedom and Unity. London & Nairobi: Oxford University Press Nyerere, J. (1968). Freedom and socialism. Dar es Salaam & New York: Oxford University Nyerere, J. (1968). Freedom and socialism. Dar es Salaam & New York: Oxford University Press. Press. Nyqvist, F. (2009). Social capital and health: Variations, associations and challenges. PhD Nyqvist, F. (2009). Social capital and health: Variations, associations and challenges. PhD thesis, Åbo Akademi University, Åbo, Finland. thesis, Åbo Akademi University, Åbo, Finland. O’Briena, M.S., Burdsala, C.A., Molgaard, C.A. (2004). Further development of an Australian- O’Briena, M.S., Burdsala, C.A., Molgaard, C.A. (2004). Further development of an Australian- based measure of social capital in a US sample. Social Science and Medicine. 59:1207–1217. based measure of social capital in a US sample. Social Science and Medicine. 59:1207–1217. Phillips, A.E., Gomez, G.B., Boily, M., Garnett, G.P. (2010). A systematic review and meta- Phillips, A.E., Gomez, G.B., Boily, M., Garnett, G.P. (2010). A systematic review and meta- analysis of quantitative interviewing tools to investigate self-reported HIV and STI associ- analysis of quantitative interviewing tools to investigate self-reported HIV and STI associated behaviors in low and middle-income countries. International Journal of Epidemiol- ated behaviors in low and middle-income countries. International Journal of Epidemiol- ogy. 39:1541-1555. ogy. 39:1541-1555. Phongsavan, P., Chey, T., Bauman, A., Brooks, R., Silove, D. (2006). Social capital, socio- Phongsavan, P., Chey, T., Bauman, A., Brooks, R., Silove, D. (2006). Social capital, socio- economic status and psychological distress among Australian adults. Social Science and economic status and psychological distress among Australian adults. Social Science and Medicine. 3:2546-2561. Medicine. 3:2546-2561. Portes, A., Landolt, P. (1996). The downside of social capital. The American Prospect. 26:18- Portes, A., Landolt, P. (1996). The downside of social capital. The American Prospect. 26:18- 21. 21. Portes, A. (1998). Social capital: its origins and applications in modern sociology. Annual Portes, A. (1998). Social capital: its origins and applications in modern sociology. Annual Review of Sociology. 24:1-24. Review of Sociology. 24:1-24. Pronyk, P. (2002). Social capital and the HIV/AIDS epidemic in rural South Africa: The new Pronyk, P. (2002). Social capital and the HIV/AIDS epidemic in rural South Africa: The new magic bullet? London: Department of Infectious and Tropical Diseases, London School of magic bullet? London: Department of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine. Hygiene and Tropical Medicine. Pronyk, P.M. (2006). Development, finance, social capital and HIV/AIDS: The intervention Pronyk, P.M. (2006). Development, finance, social capital and HIV/AIDS: The intervention with microfinance for AIDS and gender equity. PhD thesis. Department of Infectious and with microfinance for AIDS and gender equity. PhD thesis. Department of Infectious and Tropical Diseases. London School of Hygiene and Tropical Medicine, UK. Tropical Diseases. London School of Hygiene and Tropical Medicine, UK. Pronyk, P.M., Harpham, T., Morison, L.A., Hargreaves, J.R., Kim, J.C., Phetla, G., Watts, C.H., Pronyk, P.M., Harpham, T., Morison, L.A., Hargreaves, J.R., Kim, J.C., Phetla, G., Watts, C.H., Porter, D.J. (2008). Is social capital associated with HIV risk in rural South Africa? Social Porter, D.J. (2008). Is social capital associated with HIV risk in rural South Africa? Social Science and Medicine. 66:1999-2010. Science and Medicine. 66:1999-2010.

59 59 REFERENCES REFERENCES

Pronyk, P.M., Kim, J.C., Abramsky, T., Phetla, G., Hargreaves, J.R., Morison, L.A., Watts, C., Pronyk, P.M., Kim, J.C., Abramsky, T., Phetla, G., Hargreaves, J.R., Morison, L.A., Watts, C., Busza, J., Porter, J.D. (2008). A combined microfinance and training intervention can Busza, J., Porter, J.D. (2008). A combined microfinance and training intervention can reduce HIV risk behaviour in young female participants. AIDS. 22:1659-1665. reduce HIV risk behaviour in young female participants. AIDS. 22:1659-1665. Putnam, R., Princeton, N.J. (1993). Making democracy work. New Jersey: Princeton Univer- Putnam, R., Princeton, N.J. (1993). Making democracy work. New Jersey: Princeton Univer- sity Press. sity Press. Putnam, R. (1995). Bowling alone: American’s declining social capital. Journal of Democracy. Putnam, R. (1995). Bowling alone: American’s declining social capital. Journal of Democracy. 6:65-78. 6:65-78. Putnam, R. (ed). (2000). Bowling alone: The collapse and revival of American community. Putnam, R. (ed). (2000). Bowling alone: The collapse and revival of American community. New York: Simon and Schuster. New York: Simon and Schuster. Reid, E. (2000). Incorporating social capital into development practice: The HIV epidemic. Reid, E. (2000). Incorporating social capital into development practice: The HIV epidemic. Development Bulletin. 52:21-23. Development Bulletin. 52:21-23. Rodney, W. (1972). Tanzanian ujamaa and scientific socialism.African Review. 1:61-76. Rodney, W. (1972). Tanzanian ujamaa and scientific socialism.African Review. 1:61-76. Rothstein, B., Stolle, D. (2003). Social capital in Scandinavia. Scandinavian Political Studies. Rothstein, B., Stolle, D. (2003). Social capital in Scandinavia. Scandinavian Political Studies. 26:1-26. 26:1-26. Save the Children UK. (2005). The unbearable cost of illness: Poverty ill health and access to Save the Children UK. (2005). The unbearable cost of illness: Poverty ill health and access to healthcare – evidence from Lindi rural district in Tanzania. London, Save the children. healthcare – evidence from Lindi rural district in Tanzania. London, Save the children. Schwerzel, L., Munga, M., Laterveer, L. (2004). Equity implications of health sector user fees Schwerzel, L., Munga, M., Laterveer, L. (2004). Equity implications of health sector user fees in Tanzania: do we retain the user fee or do we set the user f(r)ee? Analysis of literature in Tanzania: do we retain the user fee or do we set the user f(r)ee? Analysis of literature and stakeholder views (research report). http://www.repoa.or.tz/documents_storage/ and stakeholder views (research report). http://www.repoa.or.tz/documents_storage/ Research_and_Analysis/Equity_Implications_User_Fees_Health.pdf. Retrieved 30th Research_and_Analysis/Equity_Implications_User_Fees_Health.pdf. Retrieved 30th April 2011. April 2011. Siisiäinen, M. (2000). Two concepts of social capital: Bourdieu vs Putnam. Department of Siisiäinen, M. (2000). Two concepts of social capital: Bourdieu vs Putnam. Department of Social Sciences and Philosophy, University of Jyväskylä. Paper presented at ISTR Fourth Social Sciences and Philosophy, University of Jyväskylä. Paper presented at ISTR Fourth International conference “The third sector: for what and for whom?”. Trinity College, Dub- International conference “The third sector: for what and for whom?”. Trinity College, Dub- lin, Ireland July 508, 2000. lin, Ireland July 508, 2000. Sobel, J. (2002). Can we trust social capital? Journal of Economic Literature. 40:139-154. Sobel, J. (2002). Can we trust social capital? Journal of Economic Literature. 40:139-154. Stone, W. (2001). Measuring social capital: towards a theoretically informed measurement Stone, W. (2001). Measuring social capital: towards a theoretically informed measurement framework for researching social capital in family and community life. Research Paper No. framework for researching social capital in family and community life. Research Paper No. 24. Australian Institute of Family Studies, XL Printing. 24. Australian Institute of Family Studies, XL Printing. Stone, W., Hughes, J. (2002). Social capital: Empirical meaning and measurement validity. Stone, W., Hughes, J. (2002). Social capital: Empirical meaning and measurement validity. Research Paper No. 27, Australian Institute of Family Studies, Melbourne. Research Paper No. 27, Australian Institute of Family Studies, Melbourne. Strauss, A., Corbin, J. (1990). Basics of qualitative research: Grounded theory procedures and Strauss, A., Corbin, J. (1990). Basics of qualitative research: Grounded theory procedures and techniques (1st ed.). Newbury Park, CA: Sage. techniques (1st ed.). Newbury Park, CA: Sage. Strauss, A., Corbin, J. (1998). Basics of qualitative research: Techniques and procedures for Strauss, A., Corbin, J. (1998). Basics of qualitative research: Techniques and procedures for developing grounded theory. Thousand Oaks, CA: Sage. developing grounded theory. Thousand Oaks, CA: Sage. Subramanian, S.V., Kim, D.J., Kawachi, I. (2002). Social trust and self-Rated health in US Subramanian, S.V., Kim, D.J., Kawachi, I. (2002). Social trust and self-Rated health in US communities: A multilevel analysis. Journal of Urban Health. 79(Suppl 1):21-34. communities: A multilevel analysis. Journal of Urban Health. 79(Suppl 1):21-34. Sudha, S., Carla, Z., Srikrishnan, A. K., Carl, L., Go, V.F., Suniti, S., David, C. (2009). Asso- Sudha, S., Carla, Z., Srikrishnan, A. K., Carl, L., Go, V.F., Suniti, S., David, C. (2009). Asso- ciations between social capital and HIV stigma in Chennai, India: Considerations for pre- ciations between social capital and HIV stigma in Chennai, India: Considerations for prevention intervention design. AIDS Education and Prevention. 21:233-250. vention intervention design. AIDS Education and Prevention. 21:233-250. TACAIDS (Tanzania commission for HIV/AIDS). (2003). National multi-sectoral strategic TACAIDS (Tanzania commission for HIV/AIDS). (2003). National multi-sectoral strategic framework on HIV/AIDS 2003–2007. Dar es Salaam, Tanzania. framework on HIV/AIDS 2003–2007. Dar es Salaam, Tanzania.

60 60 REFERENCES REFERENCES

TACAIDS (Tanzania commission for HIV/AIDS). (2006). UNGASS indicators country report TACAIDS (Tanzania commission for HIV/AIDS). (2006). UNGASS indicators country report January 2003 - December 2005. Dar es Salaam, Tanzania. January 2003 - December 2005. Dar es Salaam, Tanzania. TACAIDS (Tanzania commission for HIV/AIDS). (2008). UNGASS country progress report TACAIDS (Tanzania commission for HIV/AIDS). (2008). UNGASS country progress report January 2005 - December 2007. Dar es Salaam, Tanzania. January 2005 - December 2007. Dar es Salaam, Tanzania. Turmer, T (2003). Gender and HIV/AIDS. International Journal of Gynaecology and Obstet- Turmer, T (2003). Gender and HIV/AIDS. International Journal of Gynaecology and Obstet- rics. 82:411-418. rics. 82:411-418. UNAIDS. (2000). Innovative Approaches to HIV Prevention; Selected Case Studies. Geneva, UNAIDS. (2000). Innovative Approaches to HIV Prevention; Selected Case Studies. Geneva, Switzerland. Switzerland. UNAIDS. (2004). Report on the global AIDS epidemic: Executive Summary. Geneva. UNAIDS. (2004). Report on the global AIDS epidemic: Executive Summary. Geneva. UNICEF. (2002). HIV/AIDS education: A gender perspective, tips and tools. New York. UNICEF. (2002). HIV/AIDS education: A gender perspective, tips and tools. New York. URT (The United Republic of Tanzania). (1989). Tanzania population projections (1989–2025). URT (The United Republic of Tanzania). (1989). Tanzania population projections (1989–2025). Available at http://www.tanzania.go.tz.html. Retrieved 30th September 2011. Available at http://www.tanzania.go.tz.html. Retrieved 30th September 2011. URT (The United Republic of Tanzania). (2009). National multi-sectoral HIV prevention URT (The United Republic of Tanzania). (2009). National multi-sectoral HIV prevention strategy 2009–2012: Towards achieving Tanzania without HIV. Dar es Salaam, Tanzania. strategy 2009–2012: Towards achieving Tanzania without HIV. Dar es Salaam, Tanzania. URT (The United Republic of Tanzania). MoHSW (Ministry of Health and Social Welfare). URT (The United Republic of Tanzania). MoHSW (Ministry of Health and Social Welfare). (2010). Health sector performance report profile 2010 update. Dar es Salaam, Tanzania. (2010). Health sector performance report profile 2010 update. Dar es Salaam, Tanzania. URT (The United Republic of Tanzania), NBS (National Bureau of Statistics). (2011). Tanza- URT (The United Republic of Tanzania), NBS (National Bureau of Statistics). (2011). Tanza- nia demographic and health survey 2010. Dar es Salaam. Tanzania. nia demographic and health survey 2010. Dar es Salaam. Tanzania. Veenstra, G. (2000). Social capital, SES and health: an individual-level analysis. Social Science Veenstra, G. (2000). Social capital, SES and health: an individual-level analysis. Social Science and Medicine. 50:619–629. and Medicine. 50:619–629. Wallerstein, N. (1992). Powerlessness, empowerment and health: implications for health Wallerstein, N. (1992). Powerlessness, empowerment and health: implications for health promotion programmes. American Journal of Health Promotion. 6:197–205. promotion programmes. American Journal of Health Promotion. 6:197–205. Wallis, A., Crocker, J.P., Schlechter, B. (1998). Social capital and community building, part 1. Wallis, A., Crocker, J.P., Schlechter, B. (1998). Social capital and community building, part 1. National Civic Review. 87:253–271 National Civic Review. 87:253–271 WHO. (2009). Global health risk: Mortality and burden of disease attributable to selected WHO. (2009). Global health risk: Mortality and burden of disease attributable to selected major risks. Geneva: World Health Organization. major risks. Geneva: World Health Organization. Wills, G. (2000). Putnam’s America. The American Prospect. 11:17. Wills, G. (2000). Putnam’s America. The American Prospect. 11:17. Woolcock, M. (1998). Social capital and economic development: Towards a theoretical syn- Woolcock, M. (1998). Social capital and economic development: Towards a theoretical synthesis and policy framework. Theory and Society. 27:151–208. thesis and policy framework. Theory and Society. 27:151–208. Woolcock, M. (2001). The place of social capital in understanding social and economic out- Woolcock, M. (2001). The place of social capital in understanding social and economic outcomes. Canadian Journal of Policy Research. 2:1–17. comes. Canadian Journal of Policy Research. 2:1–17. Yamamura E. (2010). Different effects of social capital on health status among residents: Yamamura E. (2010). Different effects of social capital on health status among residents: evidence from modern Japan. Journal of Socio-Economics. 40:475–479. evidence from modern Japan. Journal of Socio-Economics. 40:475–479.

61 61