Access to Formal Healthcare Services in Rural Uganda: Why Non-Medical Social Resources Matter

UvA-DARE (Digital Academic Repository)

Access to formal healthcare services in rural Uganda: why non-medical social resources matter

Musinguzi, L.K.

Publication date 2016 Document Version Final published version

Link to publication

Citation for published version (APA): Musinguzi, L. K. (2016). Access to formal healthcare services in rural Uganda: why non- medical social resources matter.

General rights It is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons).

Disclaimer/Complaints regulations If you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Ask the Library: https://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. You will be contacted as soon as possible.

UvA-DARE is a service provided by the library of the University of Amsterdam (https://dare.uva.nl)

Download date:25 Sep 2021

Access to formal healthcare services in rural Uganda: why non-medical social resources matter

LABAN KASHAIJA MUSINGUZI

Cover illustration: Village meeting taking place in the community

Access to formal healthcare services in rural Uganda: why non-medical social resources matter

ACADEMISCH PROEFSCHRIFT

ter verkrijging van de graad van doctor

aan de Universiteit van Amsterdam

op gezag van de Rector Magnificus

prof. dr. ir. K.I.J. Maex

ten overstaan van een door het College voor Promoties ingestelde

commissie, in het openbaar te verdedigen in de Agnietenkapel

op Woensdag 9 November 2016, te 10:00 uur

door Laban Kashaija Musinguzi

geboren te Mbarara, Oeganda

iii

Promotiecommissie:

Promotor: Prof. dr. R.C. Pool Universiteit van Amsterdam Copromotor: Dr. D.H. de Vries Universiteit van Amsterdam Dr. D.K. Kaawa-Mafigiri Makerere University

Overige leden: Prof. dr. R.A. van Dijk Universiteit van Amsterdam Prof. dr. A.P. Hardon Universiteit van Amsterdam Prof. dr. R. Reis Universiteit Leiden Dr. P. Bukuluki Makerere University Dr. R.P.M. Gerrets Universiteit van Amsterdam

Faculteit der Maatschappij- en Gedragswetenschappen

Summary of papers

Paper I: Linking communities to formal healthcare providers through village health teams in rural Uganda: lessons from linking social capital (Submitted and under review in Human Resources Journal)

Paper II: Community perceptions of health problems and threats to the use of social resources for access to formal healthcare in Uganda (Submitted and under review Community Development Journal)

Paper III: Improving access to formal healthcare facilities in rural Uganda through the use of motorcycle taxis (Submitted and under review BMC Health Services Research)

Paper IV: The role of social networks in savings groups: insights from village savings and loan associations in Luwero, Uganda (Accepted for publication Community Development Journal (2015) doi: 10.1093/cdj/bsv050)

Paper V: Externally Initiated Interventions and Community Level Social Bonds: An Ethnographic Study of Village Savings and Loans Associations in rural Uganda (Submitted and under review World Development)

The papers are referred to by their roman numerals (I-V)

Table of Contents Acknowledgements ...... viii

Introduction ...... 1 Current study: social resources and healthcare ...... 3 Social capital, social resources and healthcare access ...... 15 Aims and questions ...... 18

PART II ...... 28 Study setting and methods ...... 28 Overview ...... 28 The study site – Luwero district ...... 30 Data analysis ...... 35 Ethical considerations ...... 35

PART III ...... 36 Outline of papers ...... 36 Paper I ...... 39 Linking communities to formal healthcare providers through village health teams in rural Uganda: lessons from linking social capital ...... 39 Paper II ...... 75 Community perceptions of health problems, activation of social resources and threats to the use of social resources for access to formal healthcare in Uganda ...... 75 Paper III ...... 96 Improving access to formal healthcare facilities in rural Uganda through the use of motorcycle taxis ..... 96 Paper IV ...... 121 The role of social networks in savings groups: insights from village savings and loan associations in Luwero, Uganda ...... 121 Paper V ...... 142 Externally Initiated Interventions and Community Level Social Bonds: An Ethnographic Study of Village Savings and Loans Associations in rural Uganda ...... 142

PART IV ...... 178 General Discussion and Conclusion ...... 178

Overview ...... 178 Contrasting health resources (VHTs) and non-medical resources: does trust matter? ...... 181 Resource-poor, resource-limited or resource-constrained: questioning the narrative ...... 182 Theoretical implications ...... 183 Recommendations for action ...... 188 Conclusions ...... 189

Appendices ...... 194 Summary ...... 194 Samenvatting...... 197 Ethical approval letters ...... 200

vii

Acknowledgements I thank God Almighty for his grace and mercy. His mercies are new every morning. It is by his grace that I managed to finish my PhD studies.

I wish to thank people who have contributed significantly to my academic journey. I wish to begin by acknowledging the support I received from the people of Luwero district and particularly the community members of Dekabusa, a community that I now call my second home. I am deeply indebted to you for allowing me to be part of you, to learn from you, for your love and the time you always took off from your busy schedules to share your everyday life undertakings. I am particularly grateful that you allowed me to be part of this great community. The time I spent with you playing pool, board games (ludo), and/or simply sitting around the mvule tree chatting away was highly enriching. I am indebted to the members of Bajja Basaga village savings and loan association who allowed me to become a member, save money and taught me a lot about social groups in communities. To my research assistants, Grace and Rebecca, thank you.

To my supervisor Prof. Robert Pool, you are a special supervisor. From the first day we met when I appeared before you for an interview as a potential PhD student in 2011, and throughout this PhD journey, your approach to issues supposedly complex is probably the reason I have been able to finish this PhD. Your unique supervision style allowed me to gain tremendous experience and knowledge of, and insight into, things both academic and non-academic that I would otherwise never have known. You made me gain courage whenever I felt short of even my own expectations. For the greater part of my academic career I grew up believing that a supervisor is the person who tells you what to do, how to do it, when to do it and reprimands you if you fail to do as instructed. You instead have shown me that a supervisor gives you a fishing net to do your own fishing. To your credit, I believe I have mastered the art of using a fishing net to do my own fishing.

In the same vein I wish to acknowledge my co-supervisors, Dr. Daniel H. de Vries and Dr. David Mafigiri for the invaluable support and guidance without which I would never have produced this dissertation. In a special way I wish to thank Dr. de Vries for constantly and critically being

viii

there for me. It is your comments and questions that helped me reflect on a number of issues which made this work what it is today. I am forever grateful.

To the entire team at “Developing Sustainable Community Health Resources in Poor Settings in Uganda (CoHeRe)” project, Dr. Nadine Pakker, Dr. Achilles Katamba, Dr. Muhangi Denis, Marije de Groot, thank you for the great support you accorded to me. My fellow PhD Candidates on the CoHeRe project, Mr. Turinawe Benoni Emmanueil and Mr. Jude T. Rwemisisi, with whom we struggled, I am forever grateful for the useful discussions we had and the comments you made on my work which helped to shape this dissertation. You were my source of strength whenever I felt lost.

At the University of Amsterdam I was supported by colleagues and friends in the struggle. I wish to thank the pre-fielders group for the great support I received in the very first phases of my research. To our Africa group and fellow PhDs, Jonna Both, Rosalijn Both, Rutayisire Theoneste, Shabalala Fortunate, Thandeka Dlamini, Alfred Adams, Ines Faria, Emmanueil Turinawe and Jude Rwemisisi, I am forever grateful for the support. To the AMMA group, Alice, Swasti, Ngoc, Selma, Maren and Agatta, thank you.

I am also indebted in a special way to my best friends Nuria Rossell, Hanna Marisol and Carla Rodriguese with whom we developed an academic bond and a strong support network that I am sure cannot be confined to the narrow streets of Amsterdam. You made my stay in Amsterdam enjoyable and turned it into a home away from home. I cannot thank you enough.

In the same way I also wish to acknowledge Julia Challinor and your husband Piet. You made it easy for me and many other colleagues in my circles to live and enjoy the life of a student in Amsterdam. You were a constant source of support and encouragement in this seemingly lonely journey.

In the same vein, I wish to thank the University of Amsterdam for creating an environment that fosters learning. I am particularly thankful to Ms Muriel and, Mr. Oomen Janus and your team at the AISSR secretariat for diligently attending to all my administrative needs. Your patience with

PhD students, including people like me who sometimes made unreasonable demands, always amazed me. Few can fit in your shoes. Thank you.

To my employers, Makerere University and particularly the Department of Social Work and Social Administration, thank you for supporting my studies chiefly for granting me study leave to pursue doctoral studies. My colleagues in the Department of Social Work, Dr. Walakira, head of department, Ms. Nanfuka, Dr. Mafigiri, Dr. Awich, Dr. Muhangi, Prof. Asingwire, Prof. Bukuluki, Dr. Mugumya, Dr. Twesigye, Dr. Twikirize, Mr. Lumonya, Dr. Omona, Dr. Bukenya, Mr. Wandera, Dr. Kamya, Dr. Kyomuhendo, Ms. Shallotte, Ms. Mwenyango, Madam Flavia and Ssegguya – thank you.

Whenever I felt like giving up – and I felt like this a number of times – my source of strength always came from my little angels. My son, Busingye OmushaijawaYesu. I had to leave you barely a year old. You endured my absence, sometimes crying in the middle of the night asking for Dad. I can never thank you enough. My daughter, Kobusingye OmwishikiwaYesu. You came into this world when I was mid-way through this journey. It always surprised me how you could tell that “Dad is in the aeroplane.” Every aeroplane that you spotted in the skies reminded you of Dad who was somewhere far in the skies. I know that you and Busingye sacrificed a lot for me, praying for me while I was at my lowest and writing me encouraging letters, sharing with me your progress in school even when I was far from you. Whenever we talked on the phone, you always told me “Dad I will sing for you when you come back.” Sometimes I wallowed in my loneliness, missing you to the bone, but the thought of knowing that I was doing this for you made it worthwhile.

Even in my strengths and weaknesses, I would never have completed this dissertation without the support of the strongest person in my life, Adreen Kanyesigye Musinguzi, the love of my life. Adreen, thank you for being patient with me, for accepting that sometimes it was important for me to be away for long periods. You postponed your own education career so I could pursue mine. You became a father and a mother to Busingye and Kobusingye in my absence. I could write a full dissertation about you, I simply can never thank you enough. I am grateful.

To my siblings, particularly my brother Apollo Mwesigye, my academic journey would have had a still birth had you not been a pillar of strength in my life. You brought us from far. Thank you. My parents, Mr. and Mrs. Fred Kashaija, you always told us to study even when you had not enough means to facilitate our education. You created for us a strong network of friends who supported us through our education. It is by no surprise that in my study these networks of friends constitute a key component of what I refer to as social resources. Thank you. Maama, thank you for your prayers.

Last but not the least, I am grateful to the "Developing Sustainable Community Health Resources in Poor Settings in Uganda (CoHeRe)" project at the University of Amsterdam under whose auspices I received financial support from the Netherlands Science Organization (NWO), WOTRO Science for Global Development Office grant number W07.45.2009.014. If it had not been for this generous support, this work would not have been completed.

xii

PART I

Introduction As a young child growing up in a rural village in western Uganda, my siblings and I were always sent to our neighbors to borrow essentials like salt, kerosene, matchboxes to make fire or “fire” itself whenever there was need. Whenever one of us became ill, we would either be sent to our neighbors to borrow money to buy medicines or to simply ask for medicines on credit from a local clinic owned by a neighbor. Our neighbors also did the same whenever they were in need. I grew up with a belief that hospitals were for people who were on the verge of death. Phrases such as “he is very ill, he has even been taken to hospital” were common expressions whenever an individual was taken ill to hospital. When I was starting secondary education, my father approached the head teacher at a nearby secondary school for a vacancy. The head teacher gave me a bursary on condition that after every end of term I would work in the school compound to pay off the school fees. Asking for help and helping in a myriad of ways was how we lived. Our everyday existence was about giving and receiving help in all forms. However, it was not until later in life as I started to connect the dots that I learned to appreciate how these mundane things do not only make human beings human, but also, as Shipton (2007:11) put it, “constitute fibers that hold societies together” and enable people to accomplish difficult tasks such as overcoming obstacles to access healthcare and cushioning against crises such as death.

Early in my education career, we were taught about colonialism, the slave trade, primitive agricultural production and killer diseases which were known as the evils that devoured Africa. Stories about Africa that were often told in schools, and even those documented, have for the most part tended to be about poverty, helplessness, vulnerability, conflict and war, famine and hunger and general desperation. Such stories also blended within the early African anthropological literature on the study of the “other”, often encapsulated by looking at such cultures as “primitive”. Yet, as Lübbe points out, the “other” portrays “everything that is weak, bad, and inhumane” (Lübbe 2009:2-3). While this has recently changed with anthropological studies carrying titles such as “anthropology of the good” (Robbins 2013:447-462), there

remains an inclination to focus on things that do not work well, on problems and needs, rather than the opportunities, capacities, resources and strengths which already exist. Ostensibly designed to bring to the fore problems which are common place in most rural communities in Africa, such as poor access to formal healthcare, wars, hunger, diseases such as HIV/AIDS, and more recently Ebola, studies tend to interrogate that which is not working well, while appearing to mask the resources, strengths and undocumented opportunities that exist in these same communities. These resources can take the form of everyday help such as borrowing money from a neighbor to buy medicines for an ill family member or offering transport to transport an ill person to a health facility.

During fieldwork in Luwero district, I saw young men transport the ill on motorcycle taxis (boda bodas) to nearby health facilities. Boda bodas helped expectant mothers to receive timely care, and often times their absence was felt. I interacted with women who toil every day to ensure that their children are fed amidst scarcity of food, and with young children trekking long journeys in search of a nearby school, or those who never miss school despite poor schooling conditions created by officials who swindle money meant to improve schooling facilities. Some sections of the population have joined in village savings and loan associations (VSLAs), an arrangement that members use to save and accumulate money that they then use to meet their health needs. A similar arrangement existed in the form of burial groups, built on the strength of informal cooperation to counter the absence of formal social protection systems to support each other during crises. Often, people relied on their membership and access to such arrangements to mobilise support that enabled them to access formal healthcare. It is the stories behind such milestones that explain how resourceful these communities are. In this dissertation, therefore, I argue that in communities like the one where I did fieldwork, and indeed in a lot of communities in Uganda, there are resources inherent in daily interactions that enable people to solve everyday problems, including access to formal healthcare.

My experience as an ethnographer in a rural village in Luwero district, central Uganda, between 2012 and 2014, taught me that to a great extent, people were inclined to assume that problems such as illness were a normal part of life and their occurrence often times expected. Related to my experience growing up, people in this community also tended to believe that the occurrence

of such predicaments is unavoidable, and collective efforts towards addressing such crises were at the heart of these communities as earlier accounts also show (Roscoe 1901:117-130; Roscoe 1911). If a member lost a loved one, everyone would be concerned and help was provided according to one’s capabilities (Roscoe 1911). Statements such as “everyone falls sick” or “all of us shall die at one time” were used to rally collective support during such crises. An individual who reaches out to others was also thought to be generally a good person or what locally is known as “obuntu bulamu”. Community members strive to reach out to others to be seen as “abantu balamu” (plural of obuntu bulamu). The idea of buntu bulamu is also engrained in the religious beliefs when help to one’s neighbors is seen as what faith and belief in God requires. But even more succinct were phrases such as “no one is an island”, implying a sense of common responsibility towards one another. At the same time conflicts were common, often times, leading to accusations and counter accusations of witchcraft and several cases of non-compliance to the collective. Such conflicts at times appeared to affect the foundation of everyday support, to the extent that sanctions would be applied, such as withdrawal of support or being rebuked outright, using expressions such as “he never helps others, we should not help.” For most observers, this is the ethos of community, which Shipton described as “networks of indebtedness” that foster togetherness and ensure social protection and human survival (Shipton 2007).

Current study: social resources and healthcare

Social resources or healthcare resources: starting point for healthcare access The means through which people “satisfy, rectify the means of, or remove the barriers that stand in the way of designated needs” are known as resources (Donenfeld 1940:560-564). This essentially means that resources develop out of the recognition that a need, which may be economic, physical, psychological and social, exists for them (Donenfeld 1940:560-564). This means that resources can be natural resources, physical, economic, psychological depending on what need the resource is intended to satisfy. In the sociological literature, social resources have been looked at as “organized associations, institutions and attitudes, both public and private, which society has developed to satisfy its physical, psychological, economic and social needs”(Donenfeld 1940:560-564). However, beyond the sociological understanding of social

resources, this study considers the forms of everyday support as generally constituting social resources (Häuberer 2014:570-593; Bakeera et al. 2009; Donenfeld 1940:560-564). These resources may be either concrete or symbolic gestures (Häuberer 2014:570-593). They may also include financial support, offering transport to a friend or neighbor to a health facility, caring for the sick, giving information on health services, and client-health worker relationships (Bakeera et al. 2009). In some literature, social resources are regarded as a key component of social capital, which constitutes social ties, norms, trust and reciprocal relationships that people have with one another (Siisiainen 2003:183-204; Perry et al. 2008:330-336; Putnam 1995:65-78). But such resources are by no means uniform in all settings. While the temptation would be to look at anything “social” as a social resource, the understanding of social resources is context and society specific as societal means through which needs are met. In a study in eastern Uganda, Meinert’s respondents mentioned wealth, unity, learnedness, smartness, and bodily strength as key resources for a good life (Meinert 2004:11-26). The full range of resources available often determine which actions people take to improve their health (Meinert 2009). Twekolamu omulimu or even generally more organized structures of VSLAs constitute some of the means that societies have developed to meet their healthcare needs. Therefore social resources are those means that societies have developed to satisfy a range of needs. However, there is limited understanding as to how they are activated to meet healthcare needs. While such community social resources appear under threat from changes in people’s life styles, partly attributable to migration, commoditization of life and broader macro changes, “a sense of community” (Ahern et al. 1996:911-923) still exists.

Uganda’s national health policy identifies health resources as human resources (formal healthcare workers), medicines and health supplies, health infrastructure and health financing and sustainability (MoH. 2010). This classification appears to limit health resources to biomedical, formal and clinical infrastructure. Selamu et al. argue that the tendency to focus on clinical needs and gaps masks the potential of existing local resources as critical for health interventions beyond conventional biomedical resources (Selamu et al. 2015). Selamu et al. point out that resources beyond the conventional biomedical resources include social capital, access to information, and other forms of infrastructure within communities (Selamu et al. 2015). The focus on such resources is about the means, strengths, opportunities and capacity that exist in the

community. It is also, as Selamu et al. suggest, about tapping into the interactions embedded in communities where people are born, live and die, and ultimately where health is consumed beyond immediate clinical needs. The call of Selamu et al. for an understanding of the resources in communities which enable people to overcome barriers of access to formal healthcare, is related to the popular argument in the aftermath of the 1978 Alma Ata declaration, mooting health interventions which focus on resources within communities(Haines et al. 2007:2121- 2131). Even in other disciplines, earlier scholars expressed a similar argument calling for adaptation of locally appropriate and defined resources. In community development, Chambers (1983), for example, called for the need to focus on people directly affected by the problem and to place them in the driver’s seat of services delivery. Chambers’ argument was that the realities of the poor are diverse and often complex, making it hard for professional “outsiders” to define them (Chambers 1995:173-204). In his book, Whose reality counts, Chambers calls on professional development workers, academicians and researchers to reflect on their attitudes towards the poor, and advocates participatory rural appraisal techniques as the best tools for capturing people’s complex local realities (Chambers 1997). In community development also there have been calls for a communitarian approach that emphasizes the inherent capacities of communities to ensure that people’s basic needs are met and social problems solved (Midgley, 1995 p.114). In social work, the strengths-based perspective requires the practitioners to take “a roster of resources existing within and around the individual, family, or community” (Saleebey, 1996:297). The strengths perspective takes cognizant of the numerous resources available in the community. In the context of health, the World Health Organization (WHO) Action on the Social Determinants of Health reminds us that solutions to healthcare access are as social as they are medical (Irwin and Scali 2010). More recently, scholars have argued that development becomes sustainable when put in the hands of the communities themselves (Ashe and Neilan 2014). The essence of what these authors were referring to was the need to draw from the simple, basic and everyday fundamentals that are part of community and society life processes to propel development. The main criticism of the approach that emphasizes strengths and opportunities within communities, individuals and groups is that it tends to downplay real problems by overly accentuating positive thinking (Saleebey, 1996). However, as Saleebey also indicate, the leitmotif of a collection of the papers in this dissertation is to argue that reframing how we view the world about the possibilities and opportunities in communities is a critical starting pointing to

solving the real problems. In essence paying attention to strengths, opportunities and possibilities creates “access to communal resources so that they become the ticket to expanded choices and routes to change” (Saleebey, 1996:302)

However, the question remains of how to turn these ideals into locally relevant and sustainable interventions. In the subsequent sub-sections I show some of the examples of social resources and the nexus with healthcare access.

Village health teams (VHTs) and formal healthcare in Uganda The Uganda government and other actors have made strides in integrating resources that exist in communities into community interventions. In the context of health for example, Uganda’s health sector strategic framework emphasizes “the client and community” (MoH. 2010) with VHTs, an equivalent of community health workers (CHWs), as the face of community health interventions. Since their introduction into the national healthcare delivery system in 2001, VHTs have become a vehicle for delivering a range of health services at community level (MoH. 2014; MoH. 2010; MoH. 2010). The VHTs constitute a network of lay volunteers who were recruited and trained to undertake a range of primary healthcare (PHC) activities in health promotion and community mobilization for disease prevention and treatment (MoH. 2014). The VHT strategy is based on the assumption that given their close links with communities, the VHTs are best placed to link vulnerable people with formal healthcare services. As such, VHT strategy states that VHTs are the “first link between the community and formal health providers”, that helps to “link the communities to the formal health service delivery system”, “help bridge the gap that exists between un-served households and the formal health system” making the VHTs “the first contact with the health system” at community level (MoH. 2010 p.19). The national VHT strategy states further that “All health activities at community level by the government, NGOs and/or partners targeting communities shall be coordinated through VHTs” (MoH. 2010, p.19). In fact Uganda’s national healthcare system, under the decentralization policy, places the VHTs at the bottom of the health service delivery system, serving the community at the lower level (Jeppsson and Okuonzi 2000:273-289) (see figure I).

MoH

National Referral hospital

Regional Referral hospital

General hospital

Health center IV

Health center III

Health center II

Village Health Teams (VHTs) (HCI)

Figure.1:The Uganda Health System Structure. Source (MoH., 2014)

Several programs by NGOs and government outreach activities have relied on VHTs to deliver healthcare services, especially in rural areas. For example, the Uganda Malaria Reduction Strategic Plan 2014-2020 places the VHTs at the center of implementation of integrated community case management (iCCM) through diagnosis and treatment of common childhood illnesses (MoH. 2014). The Uganda ministry of health officially endorsed the iCCM as a “strategy for community health systems strengthening” which entirely relies on VHTs (MoH. 2014). From northern Uganda (Kimbugwe et al. 2014:740), west Nile region (Innocent 2007) to mid-western and south-western Uganda (Miller et al. 2014:19-4491-12-19; Strachan et al. 2015:25-015-0020-8), VHTs are a dominant community health strategy in a range of health services delivery. However, studies have raised concerns about the selection and performance of VHTs (Turinawe et al. 2015:73-015-0074-7) and that like most other programs, VHTs suffer from a lack of sustainability due to a lack of a clear remuneration structure. There have been calls for development of better referral systems if the CHWs and particularly VHTs are to link

communities to formal healthcare facilities effectively (Walley et al. 2008:1001-1007). While in general, community health workers have been credited for potentially being a vehicle that makes it easy to connect communities and formal healthcare services, in practice studies have raised concerns about financial rewards, citing high attrition rates which does not only affect their sustainability but also makes it an ineffective strategy to link communities to formal healthcare (Haines et al. 2007:2121-2131; De Zoysa and Cole-King 1983:125-130). The problems of lack of trust were emphasized during our fieldwork in Luwero, which as participants mentioned, affected the VHTs ability to link communities to formal health care.

Social roles as critical entry for healthcare Uganda, like many countries in Africa, is rich in traditional structures of social support such as the extended family system, a rich heritage of kingdoms and chiefdoms whose link to health has been documented for several years (Janzen, 1978; Katabarwa et al. 2010). Social figures such as traditional herbalists, traditional bone setters, and traditional birth attendants have made a significant contribution to the health of their communities. Their contribution has often attracted mixed attention into management of illnesses. In an ethnographic study of the role of kinship systems in control of onchocerciasis in rural Uganda, Katabarwa et al. found that the involvement of kinship structure enhanced treatment of onchocerciasis because those involved in the distribution of medicines were known and well received by their communities (Katabarwa et al. 2010:265-272). Katabarwa et al. further report that the program was not affected by attrition and that the use of kinship structures succeeded because it thrived on trust, a lack of selfish intentions and reciprocal relationships in the community which they considered to be the hall marks of social capital. Such kinship structures are effective largely because of the values of collectivism that they impart in the population. In Zaire, Janzen found that social relations defined by the extended family networks and sometimes networks of friends were critical in patients’ quest for therapy (Janzen, 1978). In Tanzania, Hausmann (2000) found that community members did have their cultural interpretation of illnesses. Hausmann showed that while the cultural interpretation of causation and treatment of malaria often contradictory to the biomedical treatment options, it was also a cultural resource for the community members. Hausmann, Mushi and Muela found that treating a ‘personalistic’ illness in Tanzania involved the help of extended

kin and that this often involved a significant amount of social pressure to comply (Hausmann, Mushi and Muela, 2000).

In addition, in several African communities the traditional birth attendants (TBAs) have for long played the social roles that community members need. Given their history in delivering mothers in their homes and communities, TBAs, also known as ‘mulerwa’ in the native Luganda language, continue to play a significant social role in delivery of maternal health services. TBAs have been hailed as an important component of the healthcare system in rural areas (Chalo et al. 2005:554-557). As part of the strategy to lower maternal mortality, TBAs in several parts of Uganda received training in the 1990s to support deliveries (Chalo et al. 2005:554-557). At the peak of HIV and AIDS, studies also called for their involvement in a range of services including delivery of HIV prevention messages and particularly in prevention of mother-to-child transmission campaigns (Bulterys et al. 2002:222-224; Kironde et al. 2004:102-103). However, at the recommendations of the WHO and Safe Motherhood Initiative, the Uganda government directed attention to the skilled birth attendants and suspended partnerships with the TBAs(Turinawe et al. 2016:1). It was also recommended that the TBAs be included in the VHTs structure to provide health services at the community level (Thompson & Tabb,2011). However, in a study we conducted in Luwero it is clear that while the community members are discouraged from accessing the services of TBAs, they remain a critical resources for engaging men in maternal health services (Turinawe et al. 2016:1). The TBAs also generally remain largely accessible, affordable and a friendly resource for most community members. Therefore the banning of the traditional social roles of such providers remains suspect as they remain largely available, if not the only available, means of access to proper healthcare for the rural and vulnerable populations. But their banning could also be understood in the general framework of limited appreciation of traditional knowledge and practices (Ochen 2014).

Other significant social roles have, over the years, been performed by traditional institutions such as the Ssenga institution, which has transmitted sexual health messages to young people, especially girls being prepared for marriage. Ssenga is a paternal aunt who traditionally has the role of preparing young girls on issues of sex and marriage. Ssengas roles are performed throughout life process of the girls’ growing up. They are recognized with gifts of goats, clothes

and other items at the time when the girl is getting married. A Ssenga is in effect a custodian of information on sex and marriage issues for young girls. Studies have documented the role of the traditional institution of Ssenga as a tool to spread and teach about sexual health (Neema et al. 2004; Muyinda et al. 2003:159-167). As a result their cultural role has been adapted to provide health information to young girls. In their study, Muyinda et al. show how the traditional Ssenga institution was adapted to provide simple but effective advice on a range of sexual and reproductive health issues in south western Ugandan communities (Muyinda et al. 2003:159- 167). Some of these roles have undergone some significant changes. In Uganda today, there has been an increase in the commercial Ssengas, who provide similar services as traditional sengas but at a cost. The concept of commercial Ssengas especially in Uganda’s capital city is a practice where women, rarely men also, are hired by young women especially the elite to perform the traditional role of Ssenga(Tamale 2005:9-36). Other avenues through which commercial Ssengas work include the media, including radio programs plus books available on the streets (Tamale 2005:9-36). It is evidently clear that the traditional institution of Ssenga is being modified and transformed. In addition, at the peak of HIV and AIDS, religious leaders and institutions were at the forefront of spreading health messages preaching abstinence, being faithful and fighting stigma (Otolok-Tanga et al. 2007). These institutions continue to do the same in most of these communities.

Social groups as entry points for healthcare Besides the traditional roles of Ssengas, TBAs and VHTs, studies have also assessed the possibility of utilizing social groups as entry points for a range of other activities. In fact the Developing Sustainable Community Health Resources (CoHeRe) project, where this study is a part, aimed at answering a broad question: Can social roles, networks and groups be identified in poor and vulnerable communities in Uganda that could serve as a source of basic health information, skills and behaviours, and link vulnerable community members to existing health resources? During fieldwork we found that a wide range of groups existed in the community. A survey conducted in the study community as part of the bigger research project, CoHeRe, revealed that about 46.8% of participants belonged to a group or association, with over 60% mentioning membership in more than one group. While some are initiated from outside of the community itself, such as the savings groups, others are initiated from within the community

such as burial groups. While some groups were formed for health purposes to serve a health related function such as community health insurance schemes called munno mubulwadde or NGO supported groups of HIV and AIDS positive members, other groups were farmers groups for agricultural promotion purposes, motorcycle taxi (boda boda) riders associations, youth groups, microfinance and savings and credit groups and social security groups such as burial groups. Other groups included women drama groups, religious groups and kinship structures that included annual clan meetings. There were often loose groups of young men who often gathered in the evening hours to play board game (ludo), pool or even to simply watch European films and football in a makeshift video hall erected in the heart of the village. Such groups and spaces provide avenues for introduction of activities and interventions that support health promoting behaviors.

While all these groups existed in the community, I wanted to analyze the savings groups, one of the critical structures which attracts a number of membership in this community. Studies have documented the potential role of village savings and loan associations (VSLAs) as spaces that enable the introduction of “other activities” such as health education, agriculture promotion, and social marketing (Rippey and Fowler 2011; Rippey and Nelson 2011; Ashe and Neilan 2014). In Mali a savings for change program used savings groups as an avenue for malaria education interventions in the communities (Ashe and Neilan 2014; Rippey and Fowler 2011). In Uganda, studies show how VSLAs have been used to scale up social marketing strategies such as the sale of solar lamps in NGO facilitated VSLAs (Rippey and Nelson 2011; Rippey and Fowler 2011). Other studies have shown how VSLAs have been targeted as an entry point for programs on child protection and vulnerability reduction (Lowicki-Zucca et al. 2014:176-181) and avenues for reaching out to people living with HIV and AIDS (Rippey and Fowler 2011). In Rwanda, there are programs aimed at linking savings groups to microfinance institutions for credit facilities (Rippey and Fowler 2011). However, the introduction of other activities, such as malaria education in Mali or social marketing in Uganda, in VSLAs operations are often engineered from outside the community with the facilitating agency taking the lead. This practice, as Rippey and Fowler show in Central America, bring about tensions and conflicts among group members and staff from the facilitating agency (Rippey and Fowler 2011). In addition, such interventions were unlikely to be sustained despite the short-term successes that

usually accompany them (Rippey and Fowler 2011). From the evidence by Rippey and Fowler, it is plausible that using savings groups as a community resource for the introduction of other activities tends to introduce the same problems experienced by other externally initiated interventions, including stifling community innovation, the creation of dependence, a lack of sustainability and lack of ownership of the interventions (Platteau and Gaspart 2003:1687-1703).

The above studies show, however, that the role of community groups as avenues for introducing health activities is not conclusive. In fact, as we have seen, while some studies suggest that these groups are important entry points for health education initiatives (Ashe and Neilan 2014),other studies show that using such groups could create more controversies about sustainability and ownership of interventions than offering lasting solutions (Rippey and Fowler 2011). However, what is critical also is that even the studies that question the use of savings groups do not entirely discount their usefulness. It appears however that the answer to these unresolved issues lies in the way in which the groups such as VSLAs interact with the existing community norms and everyday processes. In fact, as we have documented in papers (IV and V), the best way to understand the impact of VSLAs is threefold; First, to look at how they interact with processes that predate their introduction (paper IV); second, how their introduction in the community impacts on the same processes (paper V) and; third, to understand the VSLAs, and social groups in general, as part of the context in which broader community processes have a bearing.

Other studies have examined the effectiveness of community-based programs such as community-directed interventions (Ndyomugyenyi and Kabali 2010:197-205) and decentralization of health systems (Lutwama et al. 2012:41) in the context of Uganda’s healthcare. A few other studies have examined the effectiveness of project-incentive-driven interventions utilizing existing informal transport systems for increasing maternal healthcare, which face similar sustainability challenges (Pariyo et al. 2011:S10-698X-11-S1-S10). Other studies (Mafigiri et al. 2012:270-284) show the potential of laypersons in patients’ social-support networks for achieving TB treatment targets in urban areas.

So what is at stake?

While studies (Bakeera et al. 2009; Katabarwa et al. 2010:265-272; Chalo et al. 2005:554-557; Meinert 2009; Meinert et al. 2014:34-46), including earlier anthropological literature (Janzen 1978), have explored and brought to light the value of social resources, the tendency towards their utilization in health interventions appears to be the creation of new structures or systems aligned with the biomedical-oriented solutions. For example, from the early 2000s, VHTs have been trained and institutionalized as part of the formal healthcare structure intended to promote health at community level (MoH. 2010; Turinawe et al. 2015:73-015-0074-7). At the start of the program, they (VHTs) received training in biomedical approaches/strategies and their involvement in community health has tended to adopt a biomedical orientation.

A similar arrangement can be said of the early interventions towards addressing HIV and AIDS in Uganda. In the late 1980s as a result of the recognition of the weaknesses in the biomedical response to HIV and AIDS which only looked at the individual, there was increasing emphasis on care and support in the management of HIV and AIDS in Uganda (Muhangi, 2009). Consequently, key leading HIV and AIDS care agencies in the country, such as The AIDS Support Organization and AIDS Information Center, established support structures such as post- test clubs for people living with HIV and AIDS that revolutionized HIV and AIDS care (Muhangi, 2009). While this new thinking about HIV and AIDS care in Uganda built on the existing traditional system of extended family, the support structures established were, just like the VHTs, largely reliant on the support of the facilitating agencies. The support and care structures established placed emphasis on ensuring that people adhered to the biomedical practices. This meant that while the approach was intended to draw from the existing community resources, the manner and the strategy was such that they largely used a biomedical orientation.

Similarly, in the mid-1990s, as alluded to earlier, the government of Uganda supported training and partnerships with those in traditional roles like TBAs in the delivery of maternal health care(Chalo et al. 2005:554-557). The TBAs were trained to apply the knowledge of biomedicine in maternal health care. This is a similar trend with a range of other programs that are implemented at community level intended to increase access and utilization of formal healthcare services. They often tend to draw little from existing resources or at least fail to start from where the communities are. Take the case of a voucher program in eastern Uganda. In a bid to increase

access to maternal health care, a program in eastern Uganda supported the use of transport vouchers for existing informal transporters to encourage hospital deliveries (Ekirapa-Kiracho et al. 2011:S11-698X-11-S1-S11). While this voucher program utilized an existing resource, informal transport providers, the approach tended to be loaded with biomedical hegemonic power. As a result, some scholars (Haines et al. 2007:2121-2131) have raised concerns about the extent to which the interventions implemented in communities such as the use of CHWs actually draw from the resources in the community and not necessarily an extension of the biomedical system at the community level.

Therefore what appears to be at stake is that, while all these interventions show that social roles can be tapped into through training and creation of support structures and incentives, their incorporation into mainstream biomedical-defined interventions implies that they tend to lose their organic and natural character. As such, their effectiveness, or lack of it, is often seen in the light of how they integrate provision of biomedical care. There is a tendency to under look their contribution to the betterment of healthcare access on their own merit. The challenge when creating new systems or new methods of engagement is the tendency for facilitating agencies to establish an incentive-led support system (financial and technical), usually at the expense of everyday forms of support, and resource endowments that already exist in these communities. In addition, structures that are formed and facilitated by a facilitating agency external to the community are likely to target a specific problem such as stigma associated with HIV/AIDS, domestic violence, or home deliveries as opposed to general improvement in community health. As a result, the problems cascade into sustainability concerns and lack of community ownership.

While the literature suggests that there is general accord on what needs to be done to better the health of populations, there is little consensus on how it should be done to achieve better health outcomes for populations. In this regard, we know little about the health potential of non-medical resources, their nature and their influence and the processes that underlie their influence on improving access to formal healthcare for community members. This study was conceived out of the realization that there is no evidence of previous research on how such non-medical resources can be used to enable communities to get into contact with existing formal healthcare services. In particular, there is a lack of studies on the manner in which existing non-medical resources in

communities facilitate the use of formal healthcare services and get communities in contact with formal healthcare providers.

Accounting for these lacunae, I examine the health potential of non-medical resources critical for improving healthcare. I draw from the social capital framework, whose entry into health research follows a move away from individual to social- and community-level determinants of health (Campbell 2001:182-196; Eriksson et al. 2013:273-298; Eriksson 2010). Social capital provides a framework for studying the potential of existing social resources, networks and groups in healthcare access. Under social capital, I apply the concepts of social networks, social resources and social groups. I argue that there is a need for considering existing non-medical resources and the value they add to studying health at community level, and examining how their activation facilitates access to formal healthcare. Their utility, or the lack thereof, lies in the dynamics and processes of resource-seeking followed by individuals which show the influence of features of social capital as a pathway to healthcare access for rural Ugandan communities.

Social capital, social resources and healthcare access I was driven by the need to study the influence of community resources that fall outside the confines of the traditional biomedical systems. The social capital framework provides a better lens for understanding how social resources work to enable individuals and communities to access formal healthcare. The papers in this thesis draw from various conceptualizations of social capital to show that features of everyday life present better opportunities and resources for healthcare access. As alluded to earlier, community-based health programs are premised on the assumption that they utilize and strengthen existing community social resources. While concepts such as community capacity, sense of community and community control have been used in community psychology studies to predict health outcomes (Campbell 2001; Campbell et al. 2013:114-122), other studies also show that social capital influences the use of healthcare services (Kawachi and Berkman 2000:174-190; Perry et al. 2008:330-336; Bakeera et al. 2010:189-198; Hendryx et al. 2002:87-104). Interest in social capital in health research grew rapidly in the 1990s, as health research increasingly placed emphasis on the need to underscore contextual determinants of health (Campbell 2001). As such, various applications of the concept have been used to show that resources inherent to social relationships are instrumental in the

decisions that people make to access healthcare. Social capital fits within the arguments that quality healthcare is not something that can simply be allocated to people (Bishai et al. 2008:184-197). Access to quality healthcare is a function of decisions made by people in their homes and in the communities where they are born, grow and die (Bishai et al. 2008:184-197). It is in settings such as homes and communities that social interactions have a strong influence in shaping the decisions that people make to access healthcare. In some cases, delays occur at such settings due to a range of factors and hence undermine access to formal healthcare (Thaddeus and Maine 1994:1091-1110).

Two strands of social capital – as an individual and a collective attribute – have dominated academic discussions for the last three decades. While sociologists like Bourdieu and Coleman looked at social capital as an individual attribute accessed through a person’s social networks and position in the networks (Bourdieu, 1986; Coleman, 1988), political scientist, Robert Putnam, conceptualized social capital as a collective attribute (Putnam, 1993, 1995). The initial application of social capital in health literature drew heavily on its conceptualization by Robert Putnam, who assessed social capital as a collective good that is accessed and utilized to achieve collective goals (Putnam 1995:65-78). This is largely because of the general shift in the 1990s from individual-level interventions to community-level interventions in health promotion (Campbell, 2001). Putnam’s conceptualization of social capital takes into account critical aspects of everyday community life such as community norms, trust and reciprocal relationships that enable collective pursuit of collective goals (Putnam 1995:65-78). Putnam’s conceptualization largely grew out of the realization that civic participation was critical to achieve collective goals. The analytical insights I draw from the study of social groups largely hinges on the realisation that civic participation is critical for achieving collective goals. These features are critical, given that they help to explain social determinants beyond biomedical resources (Campbell 2001).

Putnam’s approach is distinct from that of sociologists like Coleman, who view social capital as resources such as information or social support, that an individual accesses based on participation in a given social structure or network (Coleman 1988:S95-S120). This conceptualization of social capital is the basis for the social networks. In paper III, I assess the role of social networks in savings groups drawing on social capital as informal networks.

The two main strands have influenced debates and discussions on social capital to the extent that some scholars criticized what they called the hype about social capital arguing that it offers nothing to broaden our understanding of factors beyond what is already known as community empowerment, community capacity and sense of community (Portes 1998). In the development literature, its main criticism is on “its coherence, potential normative effects and types of policy and practice to which it might give rise” (Bebbington, 2002:800). However, Ostrom (2000) shows how sometimes commentators have branded social capital concept as another “fad”, even when it has a significant contribution in development. Sometimes its ambiguity in explaining every social phenomena as social capital has been its main undoing.

In this dissertation, every paper attempts to apply social capital based on the local conceptions. Attention is particularly paid to its influence on individual networks, which also relate to a person’s social networks and their influence on how individuals benefit from existing social groups (paper IV); the role of community norms such as togetherness (obumu) in access to formal healthcare and how community social pressure may force people to act collectively to achieve both individual and collective goals (paper II); and the role of informal transport providers as a critical community resource in overcoming barriers to healthcare access (Paper III). In addition, the papers are tied together by a common thread based on the broader concept of social capital benefiting from its broader conceptualization to include bonding, bridging and linking forms of social capital (Woolcock 2002:20-44). While bonding social capital is derived from exclusive networks of people who share similar characteristics and needs, bridging social capital allows interactions between people who may not have regular face to face interactions. Linking social capital is made up of the resources accessed through vertical networks across power differentials, connecting people to resourceful institutions such as banks, health facilities, and political offices among others (paper I). In this regard, I show in paper V how the experiences of VSLAs influence the emergence of new bonds in off-shoot groups, which in Putnam’s view suggests that pre-existing trust can be a pre-condition for mutual cooperation (Putnam, 1993). Drawing from linking social capital I show (paper I) how external interventions create ambiguities – in this case VHTs – and even “risk unequal distribution of investments and returns” of interventions (Eriksson 2011:5611).

Aims and questions The overall aim of this study was to investigate the existing social resources, including social groups, informal networks, and informal support systems, and to explore the ways in which they serve to link and enable under-served populations in rural settings in Uganda to get into contact with existing formal healthcare resources.

Research question In order to operationalize the above aim, I was guided by the broad question:

What are the existing resources (health and non-medical) in communities that vulnerable community members rely on to access formal healthcare services in rural Uganda?

Using this descriptive question, I sought to map the social resources and networks that exist in the community by paying attention to the different kinds of relations, collective or individual, bonding or bridging, that enable people to overcome barriers to access to formal healthcare. It is these networks that provide a basis for social capital. Therefore describing the different interactions, such as homogeneous or heterogeneous, strong or weak ties, constituted the basis for further exploration of questions surrounding the utilization of social resources to overcome barriers to access to formal healthcare.

Specific questions 1. How and why do community members perceive health problems? 2. What are the existing healthcare resources in the community and how do they serve to link community members to formal healthcare services? a. Drawing from the existing VHTs, a more analytical question was posed: Why do VHTs fail to link communities with formal healthcare? 3. What are the existing mechanisms (non-medical in nature) in the community that constitute resources that people rely on to access formal healthcare? a. What is the role of informal transport providers in overcoming barriers to access to formal healthcare in rural Uganda?

4. How do social interactions build up so that people can call on others for help or feel “obliged” to support others due to previous and/or ongoing interactions? a. Why are trust and social pressure important ingredients in social interactions that enable communities to get in contact with formal healthcare providers? b. How do community members activate previous and ongoing interactions to enable them to access formal healthcare? c. What are the threats to the use of social resources which enable community members to access formal healthcare?

Throughout the dissertation, I attempt to underscore the nature and processes that underlie the influence of these resources in enabling community members to access formal healthcare. While there is indeed a range of such resources, in this study the specific focus is placed on VHTs as an example of an existing community health resource, and, social groups (mainly VSLAs), informal transport providers, and informal social networks as examples of non-medical community resources.

In paper I, I assess the role of what is already defined as health resources in the community, drawing on the example of VHTs as a health resource structure initiated in the community. In this paper, I answer the question about claims that VHTs link community members to formal healthcare facilities in rural Uganda.

In paper II, I answer the question of community perceptions of health problems, activation of social resources and threats to their use for accessing formal healthcare in Uganda. In the same line, I draw from the example of the VSLAs to answer questions on the role of social networks in distributing the benefits of VSLAs (paper III). In paper III, for example, I show that the benefits that accrue to members of VSLAs can effectively be understood by looking at the social networks within which the members belong. As a way of expanding this argument further, I show in paper IV that the VSLAs can also influence the emergence of off-shoot groups which also influence the nature of the benefits of these resources.

In papers II and III, I attempt to answer the question of the influence of non-medical resources on access to formal health facilities. Paper II largely draws from every day social processes including informal networks and social pressures, while paper III draws from the example of informal transport providers as a community informal social resource to show how they play a linking role in connecting communities to formal healthcare facilities.

The last question addressed in the papers concerns threats to the use of non-medical resources in healthcare. The argument is that while the need for the inclusion of locally relevant structures of support is well articulated in the literature, a lack of clear understanding of the nature of these resources has meant that little is known about threats to their use for healthcare access.

References

Ahern, Melissa M., Michael S. Hendryx, and Kris Siddharthan 1996 The Importance of Sense of Community on People's Perceptions of their Health-Care Experiences. Medical Care 34(9):911-923.

Ashe, Jeffrey, and Kyla Jagger Neilan 2014 In their Own Hands: How Savings Groups are Revolutionizing Development.Berrett- Koehler Publishers.

Bakeera, SK, M. Petzold, GW Pariyo, S. Galea, G. Tomson, and S. Wamala 2010 Community Social Capital and the use of Health Care Services in Uganda. Journal of Public Health and Epidemiology 2(8):189-198.

Bakeera, Solome, K., Sarah Wamala P., Sandro Galea, Andrew State, Stefan Peterson, and George Pariyo W. 2009 Community Perceptions and Factors Influencing Utilization of Health
services in Uganda. International Journal for Equity in Health 8(25).

Bishai, David M., Nirali M. Shah, Damian G. Walker, William R. Brieger, and David H. Peters 2008 A Social Franchising to Improve Quality and Access in Private Health Care in Developing Countries. Harvard Health Policy Review 9(1):184-197.

Bourdieu, Pierre. 1986 The forms of capital. In John G Richardson ed. Handbook of Theory and Research for the Sociology of Education. New York: Greenwood Press: 241-258.

Bulterys, M., M. G. Fowler, N. Shaffer, P. M. Tih, A. E. Greenberg, E. Karita, H. Coovadia, and K. M. De Cock 2002 Role of Traditional Birth Attendants in Preventing Perinatal Transmission of HIV. BMJ (Clinical Research Ed.) 324(7331):222-224.

Campbell, Catherine 2001 Social Capital and Health: Contextualising Health Promotion within Local Community Networks.Oxford University Press.

Campbell, Catherine 2001 Social Capital and Health : Contextualizing Health Promotion within Local Community networks.. In Social Capital : Critical Perspectives. S. Baron, J. Field and T. Schuller, eds. Pp. 182-196. UK: Oxford University Press, Oxford,.

Campbell, Catherine, Kerry Scott, Mercy Nhamo, Constance Nyamukapa, Claudius Madanhire, Morten Skovdal, Lorraine Sherr, and Simon Gregson 2013 Social Capital and HIV Competent Communities: The Role of Community Groups in Managing HIV/AIDS in Rural Zimbabwe. AIDS Care 25(1):114-122.

Chalo, RN, HM Salihu, S. Nabukera, and C. Zirabamuzaale 2005 Referral of High-Risk Pregnant Mothers by Trained Traditional Birth Attendants in Buikwe County, Mukono District, Uganda. Journal of Obstetrics & Gynecology 25(6):554- 557.

Chambers, Robert 1997 Whose Reality Counts?: Putting the First Last. Intermediate Technology Publications Ltd (ITP).

Chambers, Robert 1995 Poverty and Livelihoods: Whose Reality Counts? Environment and Urbanization 7(1):173-204.

Coleman, James S. 1988 Social Capital in the Creation of Human Capital. American Journal of Sociology:S95- S120.

De Zoysa, I., and S. Cole-King 1983 Remuneration of the Community Health Worker: What are the Options? 4(2):125-130.

Donenfeld, Martin.J. 1940 The Meaning of Social Resources. Journal of Educational Sociology 13(9):560-564.

Ekirapa-Kiracho, E., P. Waiswa, M. H. Rahman, F. Makumbi, N. Kiwanuka, O. Okui, E. Rutebemberwa, J. Bua, A. Mutebi, G. Nalwadda, D. Serwadda, G. W. Pariyo, and D. H. Peters 2011 Increasing Access to Institutional Deliveries using Demand and Supply Side

Incentives: Early Results from a Quasi-Experimental Study. BMC International Health and Human Rights 11 Suppl 1:S11-698X-11-S1-S11.

Eriksson, Malin. 2010 Social Capital, Health and Community Action: Implications for Health Promotion.

Eriksson, Malin., Lars Dahlgren, and Maria Emmelin 2013 Collective Actors as Driving Forces for Mobilizing Social Capital in a Local Community: What can be Learned for Health Promotion? In Social Capital and Rural Development in the Knowledge Society. Pp. 273-298. Edward Elgar Publishing Cheltenham, UK.

Eriksson, Malin. 2011 Social Capital and Health--Implications for Health Promotion. Global Health Action 4:5611.

Gregson, Simon, Nicola Terceira, Phyllis Mushati, Constance Nyamukapa, and Catherine Campbell 2004 Community Group Participation:: Can it Help Young Women to Avoid HIV? an Exploratory Study of Social Capital and School Education in Rural Zimbabwe. Social Science & Medicine 58(11):2119-2132.

Haines, Andy, David Sanders, Uta Lehmann, Alexander K. Rowe, Joy E. Lawn, Steve Jan, Damian G. Walker, and Zulfiqar Bhutta 2007 Achieving Child Survival Goals: Potential Contribution of Community Health Workers. The Lancet 369(9579):2121-2131.

Hausmann, Susanne,Muela. 2000 Community Understanding of Malaria, and Treatment-Seeking Behaviour, in a Holoendemic Area of Southeastern Tanzania.

Hausmann, Muela, S., Adiel Mushi k., and Ribera Muela J. 2000 The Paradox of the Cost and Affordability of Traditional and Government Health Services in Tanzania. Health Policy and Planning 15(3):296-302.

Häuberer, Julia 2014 Social Capital in Voluntary Associations: Localizing Social Resources. European Societies 16(4):570-593.

Hendryx, Michael S., Melissa M. Ahern, Nicholas P. Lovrich, and Arthur H. McCurdy 2002 Access to Health Care and Community Social Capital. HEALTH SERVICES RESEARCH-CHICAGO- 37(1):87-104.

Innocent, Komakech 2007 The Village Health Team Strategy is a'most Innovative Community Practice'Award Winner: The Experience OFA Village Volunteer Programme in Yumbe District, Uganda.

Irwin, Alec., and Scali Elena 2010 Action on the Social Determinants of Health: Learning from Previous Experiences. Social Determinants of Health Discussion Paper 1 (Debates).

Janzen, John M. 1978 The Quest for Therapy: Medical Pluralism in Lower Zaire. Berkeley.

Jeppsson, Anders, and Sam Agatre Okuonzi 2000 Vertical Or Holistic Decentralization of the Health Sector? Experiences from Zambia and Uganda. The International Journal of Health Planning and Management 15(4):273-289.

Katabarwa, M. N., P. Habomugisha, S. Agunyo, A. C. McKelvey, N. Ogweng, S. Kwebiiha, F. Byenume, B. Male, and D. McFarland 2010 Traditional Kinship System Enhanced Classic Community-Directed Treatment with Ivermectin (CDTI) for Onchocerciasis Control in Uganda. Transactions of the Royal Society of Tropical Medicine and Hygiene 104(4):265-272.

Kawachi, Ichiro., and Berkman, Lisa 2000 Social Cohesion, Social Capital and Health. In Social Epidemiology. L. Berkman and I. Kawachi, eds. Pp. 174-190. New York: Oxford University Press.

Kimbugwe, Geofrey, Maghanga Mshilla, Denis Oluka, Olivia Nalikka, Joseph Kyangwa, Stella Zalwango, Uthuman Kilizza, Munanura Turyasiima, Louis Ntambazi, and Fred Walugembe 2014 Challenges Faced by Village Health Teams (VHTs) in Amuru, Gulu and Pader Districts in Northern Uganda. Open Journal of Preventive Medicine 4(09):740.

Kironde, Samson, John Lukwago, and Richard Ssenyonga 2004 Scaling the Frontier-should Traditional Birth Attendants also be used to Provide Nevirapine for PMTCT in Uganda? African Health Sciences 3(2):102-103.

Labonte, Ronald 1999 Social Capital and Community Development: Practitioner Emptor. Australian and New Zealand Journal of Public Health 23(4):430.

Lowicki-Zucca, Massimo, Patrick Walugembe, Innocent Ogaba, and Saul Langol 2014 Savings Groups as a Socioeconomic Strategy to Improve Protection of Moderately and Critically Vulnerable Children in Uganda. Children and Youth Services Review 47:176-181.

Lübbe, A. 2009 Us Versus them: Splitting Dynamics and Turning Points in Ethnopolitical Conflict. Journal of Peace, Conflict and Development 13.

Lutwama, George William, Janetta Hendrika Roos, and Bethabile Lovely Dolamo 2012 A Descriptive Study on Health Workforce Performance After Decentralisation of Health Services in Uganda. Hum Resour Health 10(1):41.

Mafigiri, David K., Janet W. McGrath, and Christopher C. Whalen 2012 Task Shifting for Tuberculosis Control: A Qualitative Study of Community-Based Directly Observed Therapy in Urban Uganda. Global Public Health 7(3):270-284.

Meinert, Lotte 2009 Hopes in Friction: Schooling, Health, and Everyday Life in Uganda.IAP.

Meinert, Lotte, Phoebe Kajubi, and Susan Whyte Reynolds. 2014 Connections In Second Chances: Surviving AIDS in Uganda. Susan Reynolds Whyte, ed. Pp. 34-46. Durham: Duke University Press Books.

Meinert, Lotte 2004 Resources for Health in Uganda: Bourdieu's Concepts of Capital and Habitus. Anthropology & Medicine 11(1):11-26.

Midgley, James. 1995 Social Development: The Developmental Perspective in Social Welfare. London. Sage Publications Ltd

Miller, J. S., S. Musominali, M. Baganizi, and G. A. Paccione 2014 A Process Evaluation of Performance-Based Incentives for Village Health Workers in Kisoro District, Uganda. Human Resources for Health 12:19-4491-12-19.

MoH. 2014 The Uganda Malaria Reduction Strategic Plan 2014-2020 . Ministry of Health, Kampala. Uganda.

MoH. 2010 The Second National Health Policy: Promoting People's Health to Enhance Socio- Economic Development.

MoH. 2010 Village Health Team: Strategy and Operational Guidelines. Ministry of Health, Kampala. Uganda.

Muhangi, Denis. 2009 Local Government – Non-Profit Sector Partnerships in HIV/AIDS Response: Policy Narratives and Local Practices. Unpublished PhD Thesis, Faculty of Arts, Humanities and Social Sciences Queen’s University of Belfast.

Muhwezi, Wilson, Winstons., Denis Muhangi, and Firminus Mugumya 2009 Intra-Household Differences in Health Seeking Behaviour for Orphans and Non- Orphans in an NGO-Supported and Non-Supported Sub-County of Luwero, Uganda. Afr Health Sci. 9(2):109-117.

Muyinda, Herbert, J. Nakuya, Robert Pool, and James Whitworth 2003 Harnessing the Senga Institution of Adolescent Sex Education for the Control of HIV and STDs in Rural Uganda. AIDS Care 15(2):159-167.

Ndyomugyenyi, Richard, and Asaph Turinde Kabali 2010 Community-Directed Interventions for Integrated Delivery of a Health Package Against Major Health Problems in Rural Uganda: Perceptions on the Strategy and its Effectiveness. International Health 2(3):197-205.

Neema, Stella, Nakanyike Musisi, and Richard Kibombo 2004 Adolescent Sexual and Reproductive Health in Uganda: A Synthesis of Research Evidence. Vol. 14.Alan Guttmacher Institute.

Otolok-Tanga, Erasmus, Lynn Atuyambe, CK Murphey, Karin E. Ringheim, and Sara Woldehanna 2007 Examining the Actions of Faith-Based Organizations and their Influence on HIV/AIDS-Related Stigma: A Case Study of Uganda. African Health Sciences 7(1).

Ostrom, Elena.(2000). Social capital: A fad or a fundamental concept?. In P. Dasgupta, & I. Sergeldin (Eds.), Social capital: A multifaceted perspective (pp. 172–214). Washington, DC: World Bank.

Pariyo, G. W., C. Mayora, O. Okui, F. Ssengooba, D. H. Peters, D. Serwadda, H. Lucas, G. Bloom, M. H. Rahman, and E. Ekirapa-Kiracho 2011 Exploring New Health Markets: Experiences from Informal Providers of Transport for Maternal Health Services in Eastern Uganda. BMC International Health and Human Rights 11 Suppl 1:S10-698X-11-S1-S10.

Perry, M., R. L. Williams, N. Wallerstein, and H. Waitzkin 2008 Social Capital and Health Care Experiences among Low-Income Individuals. American Journal of Public Health 98(2):330-336.

Platteau, Jean-Philippe, and Frederic Gaspart 2003 The Risk of Resource Misappropriation in Community-Driven Development. World Development 31(10):1687-1703.

Putnam, Robert D. 1995 Bowling Alone: America's Declining Social Capital. Journal of Democracy 6(1):65- 78.

Rippey, Paul, and Ben Fowler 2011 Beyond Financial Services: A Synthesis of Studies on the Integration of Savings Groups and Other Developmental Activities.

Rippey, Paul, and Candace Nelson 2011 Beyond Financial Services: Marketing Solar Lamps through Savings Groups: Emerging Lessons from Uganda. Aga Khan Foundation.

Robbins, Joel 2013 Beyond the Suffering Subject: Toward an Anthropology of the Good. Journal of the Royal Anthropological Institute 19(3):447-462.

Roscoe, John. 1911 The Baganda: An Account of their Customs and Beliefs. New York: Barnes and Noble.

Roscoe, John 1911 The Baganda: Their Customs and Beliefs. London: Frank Cass.

Roscoe, John 1901 Notes on the Manners and Customs of the Baganda. Journal of the Anthropological Institute of Great Britain and Ireland:117-130.

Saleebey, Dennis. 1996 The Strengths Perspective in Social Work Practice: Extensions and Cautions. Social Work 41(3):296-305.

Selamu, Medhin, Laura Asher, Charlotte Hanlon, Girmay Medhin, Maji Hailemariam, Vikram Patel, Graham Thornicroft, and Abebaw Fekadu 2015 Beyond the Biomedical: Community Resources for Mental Health Care in Rural Ethiopia.

Shipton, Parker 2007 The Nature of Entrustment: Intimacy, Exchange, and the Sacred in Africa. .Yale University Press.

Siisiainen, Martti 2003 Two Concepts of Social Capital: Bourdieu Vs. Putnam. International Journal of Contemporary Sociology 40(2):183-204.

Strachan, D. L., K. Kallander, M. Nakirunda, S. Ndima, A. Muiambo, Z. Hill, and inSCALE study group 2015 Using Theory and Formative Research to Design Interventions to Improve Community Health Worker Motivation, Retention and Performance in Mozambique and Uganda. Human Resources for Health 13:25-015-0020-8.

Tamale, Sylvia 2005 Eroticism, Sensuality and “women’s Secrets” among the Baganda: A Critical Analysis. Feminist Africa 5(1):9-36.

Thaddeus, Sereen, and Deborah Maine 1994 Too Far to Walk: Maternal Mortality in Context. Social Science & Medicine 38(8):1091-1110.

Titeca, Kristof, and Thomas Vervisch 2008 The Dynamics of Social Capital and Community Associations in Uganda: Linking Capital and its Consequences. World Development 36(11):2205-2222.

Turinawe, Emmanueil Benon, Jude T. Rwemisisi, Laban K. Musinguzi, Marije de Groot, Denis Muhangi, Daniel H. de Vries, David K. Mafigiri, Achilles Katamba, Nadine Parker, and Robert Pool 2016 Traditional Birth Attendants (TBAs) as Potential Agents in Promoting Male Involvement in Maternity Preparedness: Insights from a Rural Community in Uganda. Reproductive Health 13(1):1.

Turinawe, Emmanueil. Benon, Jude. T. Rwemisisi, Laban. K. Musinguzi, Marije. de Groot, Denis Muhangi, Daniel. H. de Vries, David. K. Mafigiri, and Robert. Pool 2015 Selection and Performance of Village Health Teams (VHTs) in Uganda: Lessons from the Natural Helper Model of Health Promotion. Human Resources for Health 13:73-015- 0074-7.

Walley, John, Joy E. Lawn, Anne Tinker, Andres De Francisco, Mickey Chopra, Igor Rudan, Zulfiqar A. Bhutta, Robert E. Black, and Lancet Alma-Ata Working Group 2008 Primary Health Care: Making Alma-Ata a Reality. The Lancet 372(9642):1001-1007.

Woolcock, Michael 2002 Social Capital in Theory and Practice: Reducing Poverty by Building Partnerships between States, Markets and Civil Society. Social Capital and Poverty Reduction: Which Role for the Civil Society Organizations and the State:20-44.

PART II

Study setting and methods

Overview The papers in this dissertation present the main results of ethnographic study conducted under the research program, Developing Sustainable Community Health Resources in Poor Settings in Uganda (CoHeRe), between 2012 and 2014. Building on the successes of an earlier intervention study on the role of traditional Ssengas as an institution to deliver sex education for HIV and AIDS control (Muyinda et al. 2003:159-167), CoHeRe sought to “explore the possibility of

utilising such (e.g Ssengas) existing social processes to develop a sustainable community-based health resource”(Pool and Mafigiri, 2009:7). CoHeRe’s overall aim was to “contribute to the development of sustainable interventions that will mitigate the effects of the human resource deficit in healthcare and contribute to the achievement of the health-related Millennium Development Goals (MDGs) by identifying and utilising existing community resources in poor communities in Uganda to spread health information, encourage health promoting behaviours and bring vulnerable community members into better contact with existing health services” (Pool and Mafigiri, 2009:7).

CoHeRe adopted an innovative multidisciplinary approach by identifying the non-medical social roles that could potentially be customized for health promotion(Pool and Mafigiri, 2009).. In its broad approach, CoHeRe drew from the expertise of both experienced public health researchers and anthropologists – from Makerere University, Kampala and University of Amsterdam, Netherlands – whose synergies enabled the success of the program. The intervention arm of CoHeRe benefited heavily from the research component undertaken by three PhD students (the component where I fall), whose individual projects were closely related and carried out in the same community. This also meant that data collection by the three PhD projects fed into each other, effectively informing the design of the intervention project that was implemented in the same community immediately after the ethnographic fieldwork ended. The three PhD students collected data for the broader CoHeRe program and participated in setting up the intervention component of the project.

As a component of CoHeRe, I studied village health teams that fall within the broad framework of CHWs as an example of existing health resources on the one hand, and VSLAs, informal transport providers and social networks as examples of existing non-medical resources on the other. In both sets of resources (health and non-medical), I utilized a range of ethnographic methods to assess their potential for improving access to formal healthcare in Luwero district, central Uganda.

The study site – Luwero district Luwero is one of the administrative districts in central Uganda, which was severely affected by the 1981-85 guerilla war that ushered in the current government. It is located approximately 64 kilometers north of Kampala, the capital city of Uganda. Owing to the devastating effects of the 1981-85 war, the district has attracted government attention, including the creation of a special ministry in charge of Luwero affairs. Development programs such as the Luwero-Rwenzori Development Program in the office of the Prime Minister and a host of programs supported by NGOs such as Plan International, AMREF, ADRA and Busoga Trust have been implemented in the district to improve human wellbeing. Some of these programs range from health interventions to poverty and livelihood improvements. Some of these programs have promoted the organization of people into groups such as VSLAs, which were initiated by Plan Uganda in Luwero. However, the performance of NGO supported programs in Luwero have raised mixed reactions. For example a study conducted by Muhwezi et al. concluded that despite the several programs implemented by NGOs in the Luwero district, the extended family system and informal support networks of friends and relations were in a better and stronger position than NGOs in meeting the needs of vulnerable people like the orphans(Muhwezi et al. 2009:109-117). The community where fieldwork was conducted is located approximately five kilometers off Kampala-Gulu road, a major highway. I use the word community aware that it connotes various meanings but here its used more loosely to refer to a constellation of villages where study participants were drawn. Interactions in these communities is more fluid and highly diverse. The defining feature of this community is a trading center (called Dekabusa) where community members operate retail shops, beauty salons, local clinics, several bars and chapatti stalls/makeshift structures. In front of these stalls (under the tree), a group of youths, and often times adults, gather to play the board game (ludo), usually in the later hours of the day. Playing ludo appears to be the major pastime for most youth and middle-aged adults in this community. During my fieldwork, I often held informal discussions with community members during such pastimes. The video hall located in the center of the trading center epitomizes leisure in the community. Every evening, together with mostly youth and adult members, we watched videos of African films and occasionally football from major leagues in Europe on the generator- powered television set in the makeshift video hall. The trading center attracts people from other villages who come to buy medicines from the local clinics and household items from the local

shops, drink alcohol, or simply to meet friends and play ludo. It is not uncommon to find groups of children, women and sometimes men, lining up at community boreholes to fetch water, which are also important meeting points for community members. Apart from the local clinics in the villages manned by less qualified personnel, the community where fieldwork was done has no government health facility even when the government’s ministry of health structure provides for at least a health center II at parish level (as described in figure I). The nearest government health center IV is about 5-6kms away, which as I show in paper III makes it critical for community members to mobilise transport for members whenever sick. Though the community has no secondary school, a few primary schools help to serve the local population.

Most adults in the community own at least a mobile phone and a radio. These communication gadgets help to keep members informed and in touch with friends and relatives outside the community. In the months of November and December every year, a host of coffee traders usually flocks to the community from Masaka (a distant district in the West of Kampala). Their impact is always felt, not only in economic exchanges but also in social interactions. The motorcycle taxis (boda bodas), which have also become a lucrative business for young men from this community, serve to link the community with facilities located outside it. For example, through the entire fieldwork period, I observed how the boda boda riders were a key resource, ferrying people to and from the health facilities. Boda bodas are also used to transport agricultural produce to markets. Generally, residents described themselves as small-holder peasant agricultural farmers, although some still owned and ran a few shops and other businesses to supplement their meager incomes.

According to the 2014 national housing and population census, Luwero district has a total population of 456,958 persons (Uganda Bureau of Statistics, 2016). A significant proportion of the people in the study community regarded themselves as native Baganda, born in the area, but there was also a significant number from other districts including Nakasongola, Masaka, and Nakaseke. Due to the influence of migration and inter-marriage, a sizeable number of people I interacted with described themselves as non-Baganda, usually from the districts of Mbale, Mbarara, Gulu and Masaka. Whereas the community appeared ethnically diverse, Luganda was the common language spoken by both Baganda and non-Baganda. Given the characteristics of

this community, it is not surprising that social groups such as village savings and loan associations (VSLAs) are an attraction for the local community.

The study site, Luwero district, provides a good example of how people rely on each other to support access to formal healthcare by providing instrumental support in the form of access to cash or lending each other money, or even providing information support on a daily basis. While Luwero may not be generalizable to the entire country, it is plausible that other communities that share similar characteristics, as indeed most parts of rural Uganda do, could benefit from studies of this nature.

South Sudan

Kenya

Luwero Democratic Rep. district Congo

Tanzania Rwanda

Figure 2: Map of Uganda showing location of Luwero District

Data collection Ethnographic data were collected through participant observation, informal and formal interviews, and focus group discussions (FGD) from 2012 to 2014. Participant observation involved immersion in community activities (Spradley, 1980), including attending health outreach events, church and religious gatherings, community and group meetings; listening to conversations in bars; playing the board game (ludo) with youths; contributing money whenever a member of a community was sick; participating in burial activities; and playing with children. More focused observations were made during three health outreach efforts, weekly village savings and loan association (VSLA) meetings, four funerals, three wedding ceremonies, one local church fundraising event, and seven village meetings to discuss health issues that were convened by the village council executive. I observed and noted what people said and did on issues related to health, how local leaders mobilized people whenever there was a health outreach activity or a health need, and people’s responses to health outreach efforts. Issues captured during these observations were written in a diary and used as points of reflection as fieldwork progressed.

A total of 91 in-depth interviews (36 males and 55 females) and 42 FGDs (23 with females only, 15 with males only, four mixed FGD) were conducted with adult community members aged 18 and above, as part of a broader CoHeRe project. Interviews and FGDs focused on broad issues including community groups, forms of togetherness, health problems and the sources of conflicts. All the papers were based on this data. In addition, interviews with key informants were conducted with representatives from the local government, including staff at a level IV facility and the district health department, and with area NGO staff such as AMREF, Plan Uganda and Save for Health. Triangulation of methods contributed to increased validity and reliability of the findings. However, after preliminary analysis of the data, themes that emerged were followed up in subsequent interviews and discussions and analysis done for each paper.

In paper I, after preliminary analysis of the data, we conducted six additional in-depth interviews and three FGDs with VHTs, and four FGDs with community members on the perceived roles of VHTs.

In paper II, after preliminary analysis of the data, we conducted a participatory community- mapping exercise where participants mapped existing social resources, such as homes of prominent people, meeting points in the community, and infrastructure such as schools, local clinics, and sources of water. In addition, for paper II, seven ‘pile-sorting’ exercises were conducted separately with groups of women, men, and youths. In the pile sorting sessions, participants identified and listed health concerns or problems affecting access to healthcare, wrote them on slips of paper, and then put these in a pile. Each piece of paper in a pile was picked and the problems were discussed one by one and ranked in order of severity. Then, after ranking, participants discussed how community members had addressed these problems on their own. This exercise was particularly important for understanding the community’s ‘collective response’ and how people activated and utilized social resources.

For paper III, after preliminary analysis we conducted five interviews and three FGDs with boda boda riders, and four FGDs were held with community members. The additional interviews and FGDs focused on the role of boda bodas in enabling access to formal health facilities, addressing topics such as mobilization of transport services by boda bodas, rates charged by the riders, various motivations of boda boda riders, their relationship with other community members, and challenges that boda bodas and the community face.

For papers IV and V, the main insights were drawn from a one year-long participation as a member of one of the VSLAs in the community, saving money and attending weekly meetings between January and December 2013. In addition, particularly for paper IV, I followed up and interviewed on several occasions three women whose social networks I describe in the paper.

At the end of the ethnographic fieldwork, a survey was conducted with adult community members. The survey was intended to establish the prevalence of the health problems identified during ethnography. While the survey was mainly intended to inform the intervention implemented under the project, some of the papers written in this dissertation also drew from some of the analytical statistics based on the survey.

Data analysis All community interviews and FGDs were conducted in Luganda (local language), audio recorded and transcribed in English. The transcription process was highly iterative and involved reflections on the data and emerging issues that were discussed among CoHeRe team members, and were further followed up in subsequent interviews. Data was inductively analyzed in Nvivo 10 software. Themes that emerged from analysis formed the basis for writing the papers. Whenever a theme was identified, it would be discussed among CoHeRe team members and whenever possible, new analysis would be done on the issue. Detailed analyses are included in each paper.

Ethical considerations The study received ethical approval from the Amsterdam Institute for Social Science Research at the University of Amsterdam, the School of Public Health at Makerere University, and the Uganda National Council of Science and Technology. Informed written consent was received from participants before conducting audio recording interviews and FGD. Oral permission for observation was always obtained from key organizers during communal events. Pseudonyms are used to protect the privacy of participants.

References Muhwezi, Wilson, Winstons., Denis Muhangi, and Firminus Mugumya 2009 Intra-Household Differences in Health Seeking Behaviour for Orphans and Non- Orphans in an NGO-Supported and Non-Supported Sub-County of Luwero, Uganda. Afr Health Sci. 9(2):109-117.

Pool, Robert, and David, Mafigiri 2009 Developing Sustainable Community Health Resources in Poor Settings in Uganda (CoHeRe). Full Proposal, Unpublished. Submitted to WOTRO under the Global Health Policy and Health Systems Research. Spradley, James P. 1980 Participant Observation. Uganda Bureau of Statistics (UBOS). 2016 The National Population and Housing Census 2014 – Main Report, Kampala Uganda.

PART III

(Non)medical social resources and formal healthcare in rural Uganda

Outline of papers Based on the aims of the study outlined in Part I, one of the main interests of this study was to examine existing social resources in the community that serve to link community members to, or enable them to get in contact with, existing health resources. This issue is generally explored in all papers but the main emphasis is in the first three papers (Papers I, II and III). In examining the issue, I attempt to make a distinction between what is already classified as health resources by the formal healthcare system and what is not directly categorised as such but plays a significant role in enabling communities to access formal healthcare services. In this case I draw from the case of VHTs and aim to interrogate the interplay between the VHTs, the community 36

and formal healthcare providers. Our analysis in paper I questions the claims that VHTs serve to link community members with formal healthcare providers. Our analysis also reveals that VHTs are unable to link community members to formal healthcare due to existing power asymmetries, a failure by the government to equip the VHTs to act as an effective link, and a general failure to address the social determinants of health. In this paper we use the linking social capital framework and show that such community-based medical interventions have drawn little from existing social resources of mutual trust and cooperation and that, as we have shown in the case of Luwero, programs such as VHTs have undermined, instead of harnessing, effective utilization of social relations as a vehicle for health promotion. In particular, the paper draws attention to the ambiguous position that VHTs find themselves in, making it hard to disentangle themselves from the webs of power relationships. We show that the challenges lie in poor conceptualization of programs, paying little attention to community dynamics and often times a failure to learn from experiences of what already exists in communities. This paper in particular suggests a rethink in the VHT strategy, highlighting the need to take into account power structures in the community, and, more critically, to draw on existing social resources.

Paper II builds on the first one by presenting concrete scenarios where the work of VHTs may not be effective, especially given that community members appear to rely more on the ethos of community, such as borrowing money, and reaching out to the sick and bereaved in a manner that seems to suggest that issues of access to formal healthcare are not just a concern of VHTs or of a few government structures, but rather a concern for everybody. Paper III uses the example of concrete resource structure, whose establishment has nothing to do with the government or NGOs, but whose support in linking communities to formal healthcare is just as critical. The paper studies the informal motorcycle taxi drivers who, despite the fact that they are driven by profit motives, provide services that suit the community. We argue in particular that in the absence of a clear system in which the population is connected to health facilities, the motorcycle taxis have filled this gap. One central argument in this paper is that the role of social networks and the value of social relations and compliance to community norms is critical even for profit- oriented services like motorcycle taxis.

Unlike the first three papers, papers IV and V attempt to critically examine how the population interprets its social relations and utilises them for its own benefit, and we articulate how social resources permeate everyday processes of mutual support. In paper IV, for example, we show how social networks are critical for understanding the way in which the benefits of social groups in the community, as a social resource, are appropriated. My analysis in this paper identifies two critical concepts of ‘silence-in’ and ‘silence-through’ networks. My argument in this paper is then expounded further in paper V, which goes beyond the networks played out within social groups, in this case using VSLAs as an example, to the emergence of off-shoot groups, taking into account the new bonds that emerge as a result. In these two papers (IV and V) we argue that it is important to understand the social processes that particularly influence how the benefits of social resources, and particularly of informal social networks, are appropriated. These processes include the conflicts that often emerge and how they are managed as we discuss in paper V that influenced the emergence of offshoot groups, or how the informal networks that predate the introduction of social groups function to appropriate the group benefits (paper IV).

Overall, while the findings in each of these papers may be applicable only to the study setting specifically, or to other settings with similar characteristics, I hope that all the papers can be read as a contribution to the understanding of the immense value of social resources for access to formal healthcare in rural Uganda. I also implore readers to consider the papers in this dissertation as an attempt to explore the nature of social resources, their activation and limits to their use by community members for accessing formal healthcare in rural Uganda. Therefore, while there are certainly other enormous resources in these communities, the resources discussed in the five papers in this dissertation constitute the most relevant, based on community levels of engagement and utilisation.

Paper I Linking communities to formal healthcare providers through village health teams in rural Uganda: lessons from linking social capital

Authors Laban Kashaija Musinguzi1, 2: Emmanueil B. Turinawe1,2: Jude T. Rwemisisi2: Danny de Vries2: David K. Mafigiri1: Denis Muhangi1: Marije de Groot2: Achilles Katamba3: Naddine Pakker4: Robert Pool2.

1 Department of Social Work and Social Administration, Makerere University, P.O. Box 7062, Kampala, Uganda

2Amsterdam Institute of Social Science Research, AISSR, University of Amsterdam, Nieuwe Achtergracht 166, 1018 WV Amsterdam, Netherlands. Email: [email protected].

3College of Health Sciences, Makerere University Kampala Uganda. Email:[email protected].

4Institute for Global Health and Inequality/Center for Poverty-Related Communicable Diseases, Academic Medical Center, Amsterdam Corresponding author: Laban Kashaija Musinguzi. Email contact: [email protected] / [email protected]. Telephone contact: +256 701 443737

This paper has been submitted and currently under review in the Human Resources for Health Journal

Abstract

BACKGROUND: Community based programmes, particularly community health workers (CHWs), have been portrayed as a cost-effective alternative to the shortage of health workers in low-income countries. Usually literature emphasizes how easily CHWs link and connect communities to formal healthcare services. There is little evidence in Uganda to support or dispute such claims.

OBJECTIVE: Drawing from linking social capital framework, this paper examines the claim that village health teams (VHTs), as an example of CHWs, link and connect communities with formal healthcare services.

METHODS: Data were collected through ethnographic fieldwork undertaken as part of a larger research program in Luwero District, Uganda between 2012 and 2014. The main methods of data collection were participant observation in events organized by VHTs. In addition, a total of 91 in-depth interviews and 42 focus group discussions (FGD) were conducted with adult community members as part of the larger project. After preliminary analysis of the data, we conducted an additional 6 in-depth interviews and 3 FGD with VHTs and 4 FGD with community members on the role of VHTs. Key informant interviews were conducted with local government staff, health workers, local leaders, and NGO staff with health programs in Luwero. Thematic analysis was used during data analysis.

RESULTS: The ability of VHTs to link communities with formal healthcare was affected by the stakeholders’ perception of their roles. Community members perceive VHTs as working for and under instructions of “others”, which makes them powerless in the formal healthcare system. One of the challenges associated with VHTs’ linking roles is support from the government and formal healthcare providers. Formal healthcare providers perceived VHTs as interested in special recognition for their services yet they are not “experts”. For some health workers, the introduction of VHTs is seen as a ploy by the government to control people and hide its inability to provide health services. Having received training and initial support from an NGO, VHTs suffered transition failure from NGO to the formal public healthcare structure. As a

result, VHTs are entangled in power relations that affect their role of linking community members with formal healthcare services. We also found that factors such as lack of money for treatment, poor transport networks, the attitudes of health workers and the existence of multiple healthcare systems, all factors that hinder access to formal healthcare cannot be addressed by the VHTs.

CONCLUSION: As linking social capital framework shows, for VHTs to effectively act as links between the community and formal healthcare and harness the resources that exist in institutions beyond the community, it is important to take into account the power relationships embedded in vertical relationships and forge a partnership between public health providers and the communities they serve. This will ensure strengthened partnerships and the improved capacity of local people to leverage resources embedded in vertical power networks.

Key words: Village health teams, community health workers, linking social capital.

Introduction Linking people with existing healthcare services is a role of any effective and efficient healthcare delivery system. Low-income countries like Uganda that are faced with shortages of healthcare resources [1, 2], community-based approaches, and in particular involvement of less- specialized cadres in health promotion, have been advocated in global health discussions as means of linking people with existing healthcare resources [3, 4]. Following the 1978 Alma-Ata Declaration, community health workers (CHWs) gained center stage in “accelerating coverage of essential interventions particularly for the poor and underserved communities” [5]. The adoption of CHWs is partly due to suggestions that they increase coverage and cost-effectiveness of health services delivery [6], provide an alternative solution to the crisis of health workers in resource- constrained communities [7] and be flexible in participating in community health [8].

Village health teams (VHTs) fall broadly within the CHWs model. VHTs were introduced in Uganda in 2001 as part of the implementation strategy for the 2001-2006 national health strategic plan; they fall within the broader government’s health sector strategic framework which emphasizes “the client and community” [9]. The focus of VHTs as a component of health promotion is the “more active and meaningful participation in health development…” of communities [9]. In the decentralized national healthcare delivery system, where the delivery of public health services follows a tiered structure, VHTs occupy the lowest level. With the national referral hospitals at the top of the structure followed by regional referral hospitals, general hospitals, health center (HC) IVs, HCIIIs and HCIIs, the VHTs (HCIs) occupy the bottom of the ladder [9]. Given their base in the communities VHTs are, arguably, better placed to link community members to existing formal healthcare providers. The 2010 national VHT strategy that operationalizes and institutionalizes the VHTs in health services delivery is particularly emphatic about their linking role. In addition to performing major primary healthcare roles, the Ugandan national VHT strategy projects VHTs as “the first link between the community and formal health providers” serving to “link the communities to the formal health service delivery system” [10]. The national VHT strategy further states, “VHTs will help bridge the gap that exists between un-served households and the formal health system” and that VHTs will be people’s “first contact with the health system” [10].

The strategy of VHTs linking communities to formal healthcare is based on the understanding that access to formal healthcare remains a challenge for Uganda’s healthcare system. For example, whereas over 72% of the population lives within a recommended 5 kms radius of a health facility, only 44.4% of expectant mothers deliver at health facilities and only 32.4% of pregnant women attend all four recommended antenatal visits [11]. Factors that affect accessibility emanate from both demand and supply sides of health service delivery system [11- 14].

Unlike in countries like South Africa where community health workers are based at health facilities and directly answerable to the health service that selects them [15], in Uganda the VHTs took on a normative view of CHWs as community-based generalists, selected by the

people in the community, but they report to the structures of government, mainly formal healthcare service professionals.

Studies elsewhere equally demonstrate that CHWs have positive effects on health because they act as “entry points”, “links”, “bridges” or “connectors” between communities and formal healthcare services, systems and resources [4, 5, 16-18]. However, some reviews that counter such claims suggest that CHWs “have suffered from poor integration into the broader health system” and that they are not a solution to weak health systems [(19). Such reviews have tended to base their evidence on factors such as high levels of attrition, lack of motivation and poor incentives for CHWs [20-22].

Whereas there is consensus that CHWs, due to their close ties to communities, are best placed to link communities with formal healthcare providers, there is hardly any evidence of how these linkages should materialise. Haines et al. argue that one of the main challenges facing effective utilisation of CHWs is unresolved concerns about CHWs as either an extension of the authoritarian formal healthcare system into the community or CHWs as agents of community change [5]. Therefore, using linking social capital framework, we examine the claim that VHTs create “linkages”, “connections” and “bridges” between communities and formal healthcare providers. Our analysis of the networks and systems involved is informed by community members’ perception of the roles of VHTs, the formal health workers’ perceptions of VHTs and the VHTs’ perception of their own position in a rural community in Luwero District, Uganda. In- depth insight into the way VHTs form a link between communities and formal healthcare providers contributes to discussions on the effectiveness of CHWs in promoting community health in low-income countries and particularly their potential role as an alternative to the human resource deficiencies in the health sector.

Linking social capital and healthcare access Defined as a social resource accessed through a person’s networks and participation in community events, the concept of social capital became a focus of health research following calls to move away from the traditional curative healthcare services towards contextual and social determinants of health [23, 24, 25, 26]. Social capital theorists and empirical studies

project a positive link between social capital and a community’s health outcome. Ogden, Morrison and Hardee posit that social capital is critical to improving health sector outcomes because it creates bonds within communities, connects communities with each other and links them to formal health systems [25]. Ogden et al. further suggest that various forms of social capital—bonding, bridging and linking— are critical to strengthening health systems. Bonding social capital relates to exclusive solidarity among people who are alike while bridging social capital relates to inclusive solidarity between people from varying backgrounds [27].

Linking social capital, the focus of this paper, involves the accumulation of ties with individuals in power and institutions of influence usually outside the community [27]. Therefore, linking social capital constitutes “resources found in vertical relationships… between individuals in a community and institutions or individuals with access to resources beyond the community” [28, p.283]. In this case, formal healthcare facilities constitute “resources beyond the community” [28] and VHTs “conduits for feedback from the community” [25, p.4] acting as a vessel for leveraging these resources. Linking social capital is a useful framework, it serves as a lens for understanding how claims of linkages between communities and formal healthcare providers materialise or do not materialise under the influence of VHTs.

Linking social capital is based on key assumptions about the perceived linkages. First, linking social capital assumes that communities will have “access to networks or groups with relatively more power and decision-making influence” [25, p.1078]. This suggests that the starting point is to assess the “networks that connect people across explicit vertical power differentials” [27, p.2205]. Second, linking social capital is premised on the assumption that resources inherent in the vertical connections propel people’s capacity to obtain resources from formal institutions such as healthcare providers [27]. In this case, it can also be argued that understanding access and linkages to these formal institutions is an essential element of creating the capacity to “engaging power structures” [27, p.2207]. Given that health inequities in populations continue to be contingent on a lack of access to equal distribution of power [29, 30], any intervention such as VHTs that seeks to provide a way of engaging power structures and access to such resources is critical. Third, linking social capital assumes a form of “partnerships between healthcare providers or public health entities and underserved communities” [28, p.283]. Partnerships also

imply “a systems model of collaboration” in which different players support each other to achieve common goals [25]. Within this partnership there are also echelons of power, particularly the power of the professional health workers or even the power that the VHTs themselves assume after becoming VHTs in their community. Power and influence are therefore central to the discussions of linking social capital. Where power and influence are concerned, social inequalities arise. In fact, one of the earliest notable observations of the concept of social capital was based on the premise that it reproduces social inequalities in society (31). Evidence shows that interventions originating from outside the community create inequalities and asymmetries of power in communities [32, 33] and affect local innovations in managing community problems. They also increase dependence on external actors [34] and potentially facilitate elite capture [35, 36]. This means that designing interventions such as VHTs that seek to link communities to leverage resources and engage power structures of formal institutions requires taking into account the existing power and sources of influence.

Linking social capital framework enables us to see how the VHTs strategy enhances access to networks of power, effects the capacity to leverage resources from formal healthcare providers and fosters any form of partnerships among various systems. For example, some studies have shown that linking social capital can in itself be deleterious if it is not based on the existing bonding and bridging forms of capital in the community [27]. Others have argued that social capital in form of bonding social capital has an impact on how structures such as VHTs function [37]. Indeed, community-based interventions like VHTs are spearheaded on the assumption that the actors know communities better and can effectively utilize existing social bonds not only to mobilize but also link community members to formal healthcare .

Whereas some earlier studies [23] suggest a positive relationship between health and social capital, the application of social capital as a measure of understanding healthcare access has been growing rather slowly. A systematic review of 2,396 abstracts by Derose and Varda found only 21 papers with “some measure of social capital and its effects on healthcare access” [28]. Derose and Varda show that most of the research has tended to focus on the health outcomes particularly healthcare utilization with little attention to access. Therefore, linking social capital remains the least developed in health research [38]. By applying linking social capital framework

in the study of the perceived linkages between community members and formal healthcare providers, we hope to add to the development of the concept and highlight its usefulness in understanding how community based interventions like VHTs affect healthcare access for vulnerable community members.

VHTs in Luwero District The 2010 Uganda national VHT strategy specifies that VHTs should be selected by and from within the community. The recruitment of VHTs was carried out as a measure to harmonize and streamline earlier community-based tasks such as community-based growth monitors, community-owned resource persons, community drug distributors, home care providers, counselling aides, or parish mobilisers among others [10]. In fact, most of the recruited VHTs had one or more of the above responsibilities.

In Luwero District, training and recruitment of VHTs was spearheaded by an NGO, the Africa Medical Research Fund (AMREF), under what was then known as the Malaria, HIV and AIDS, and Tuberculosis (MART) project. The VHTs received training before they started working as VHTs. The training was meant to equip them with basic knowledge, skills and attitudes necessary for understanding and helping communities with their health concerns. When the project ended, the support of VHTs was left to the local government [39]. As we show later, the process of “handing over” of VHTs to the formal health structure suffered transition failures which created a series of motivational challenges. Upon selection, each VHT would be assigned at least 25-30 households implying that the number of households in a village would determine the number of VHTs to be selected per village. In the community where fieldwork was done, a total of five VHTs were selected per village.

Studies have looked at how the VHTs mobilise communities for health programmes that come into the community such as immunization and outreach activities [40, 41]. In some areas like post-conflict northern Uganda where there have been successful VHT programmes, several NGOs have implemented activities that directly provide incentives for the VHTs and health workers to perform their duties [41]. As a result, VHTs in such communities have been able to perform their duties such as home visits, community mobilisation, management of common ill

health conditions and drug distribution, follow up of expectant mothers and post-natal mothers, distribution of health commodities and malarial control [41]. Although Kimbugwe et al. found that VHTs are critical agents for bridging the gap among the community, health facilities and other key development partners in the health sector, an earlier study conducted in Yumbe in northern Uganda posits that the achievements registered by VHTs should be considered with caution, arguing that the VHT programme is a delicate venture that depends on the means by which VHTs and those who supervise them are motivated [40]. The issue of VHT motivation is an important one for their performance. Studies conducted in mid-western and south-western Uganda analysed VHT motivation, retention and performance-based incentives [20, 22]. In our study of VHTs in Luwero, we found that the selection process of VHTs was fraught with distrust, which at times seemed to have affected their legitimacy [39]. Almost all these studies, save for Turinawe et al., have been largely cross-sectional studies which provide little information about the experiences of the VHTs in providing the much needed linkage between the community and health service providers. Uganda is politically, economically and socially diverse with multiple socio-economic and political realities. Therefore, these multiple contexts impact, in a number of ways, how VHTs perform their duties. To our knowledge, there is no study using ethnographic approaches that has examined the claim that VHTs serve to link communities to formal healthcare providers.

Setting and methods This paper is based on ethnographic fieldwork conducted between 2012 and 2014 in rural Luwero, Uganda. Luwero District is located approximately 60 kilometres from Kampala, the country’s capital. Estimates put Luwero District’s total population at 405,900 persons [42].

The community where fieldwork was conducted is relatively remote, located approximately 5 kilometres off the Kampala-Gulu road, north of Kampala city. The term “community” is used here to refer to a collection of villages from which the participants were drawn. The prominent feature in this community is a trading centre called Dekabusa. With several small-scale businesses, retail shops, saloons, local clinics, makeshift video halls, bars and chapatti selling stalls, Dekabusa is an important centre of this community. Most adults in this area own a mobile phone and a radio, which keep them informed and in touch with friends and relatives outside the

community. The main means of transport, motorcycle taxis (boda bodas) have also become a lucrative business for young men from this community and help to keep the community linked to the facilities outside the community such as markets and health facilities.

The majority of the residents described themselves as small-holder peasant agricultural farmers; however, some of them still had businesses to supplement their incomes. Although a proportion of the people regarded themselves as native Baganda, due to intermarriages, the community is more ethnically diverse. The native Baganda and non-Baganda speak Luganda, the local language through which the interviews and discussions were conducted.

Data collection Data were collected by the first three authors between 2012 and 2014 as part of a bigger project, “Developing Sustainable Community Health Resources in Uganda” (CoHeRe). The ethnographic data for this paper were collected mainly by the first author and results are presented in the first person when they derive directly from his fieldwork experience. The main method was participant observation. General observation points and activities involved spontaneous interactions with people in bars, playing ludo (a board game) with youth, attending church and religious functions, contributing money whenever a member of a community was sick, participating in burial activities, playing with children, and attending community meetings and health outreach programmes. As Spradley notes, our participation was guided by activities and the physical aspects of the situations “as and when they unfold[ed] in their natural settings” [43, p.54]. Although participant observation was a continuous engagement, more focused observations were made in a total of seven village meetings, two health outreach programmes, four burial ceremonies, three wedding ceremonies, one church fundraising function and several meetings in village savings and loan associations (VSLAs). Participation in these events revealed how and what health related issues were discussed and how VHTs and local leaders mobilised people whenever there was a health outreach event. Field notes written after every day’s work were used as points of reflection and follow-up in subsequent observations, interviews and focus group discussions (FGDs).

In addition, a total of 91 in-depth interviews (36 with males and 55 with females) and 42 FGDs (23 with females only, 15 with males only, four mixed FGD) were conducted with adult community members aged 18 and above. Given that this study was conducted as part of a broader CoHeRe project, initial interviews and FGDs focused on broad issues including community groups, forms of togetherness, health problems and the roles of VHTs in addressing some health problems. After preliminary analysis of the data, and identifying the role of VHTs, we conducted a total of six in-depth interviews and three FGDs with VHTs, and four FGDs with community members on the perceived roles of VHTs. All interviews lasted between 30 minutes and one hour and were recorded with permission from the participants. In-depth interview participants were selected using a simple random procedure while VHT participants were purposively selected. Except for the FGDs with VHTs, which depended on how many VHTs were available, FGDs with community members usually involved between 10 and 12 participants and also lasted 30 minutes to over an hour. One community mapping exercise was conducted with selected community members and intended to map existing resources in the community including homes of prominent people like VHTs, available local clinics in the community, and social infrastructure such as schools.

In addition, seven pile sorts were conducted with groups of women, men and youths where participants identified the main health concerns/diseases in the community. They also ranked the identified diseases/community health problems in their order of severity and identified solutions for each. Data from pile sorts was critical in understanding the specific health concerns and how VHTs have responded to them. Key informant interviews were conducted with representatives from the local government, health workers at the existing public health facilities and district health department, and NGOs operating in the area.

Data analysis Whereas this paper is based on a total of six interviews, three FGDs with VHTs, and four FGDs with community members on the perceived roles of VHTs, the overall analysis and conclusions are based on many other situations, discussions, interviews and reflections described above. All interviews and FGDs were conducted in Luganda, audio recorded and thereafter transcribed and translated into English by the first author. The transcription process was highly iterative and

involved reflecting on the data, discussing emerging issues with CoHeRe team members and following up the data in subsequent interviews. Data was inductively analysed using Nvivo10 software. Themes that emerged from analysis include enthusiasm about VHTs, perceptions about the work of VHTs, government neglect and lack of support from formal healthcare providers and suspicions about the proper intentions of the programme. These themes emerged from analysing the perceptions of different participant categories, that is, VHTs, community members and formal healthcare providers.

Ethical statement The study received ethical approval from the Amsterdam Institute for Social Science Research at the University of Amsterdam, the School of Public Health at Makerere University number IRB00011353, and the Uganda National Council of Science and Technology. Informed written consent was received from participants before conducting and recording interviews and FGDs. Oral permission for observation was always obtained from the key organizers during communal events. Pseudonyms are used to protect the privacy of participants. Participants in FGDs were assigned respondent numbers (respondent 1 to respondent 12), which are used in this paper when verbatim quotations are used.

Results From early enthusiasm to sceptical apathy Initially, the recruitment and training of VHTs in Luwero District generated enthusiasm among the local people, especially when the first incentive-led rewards for VHTs yielded creditable results as VHTs interacted with community members and formal healthcare workers. AMREF, the NGO that trained VHTs, initially provided financial incentives and facilitated their work; for example, they routinely wrote reports and submitted them to their supervisors. They were given bicycles, T-shirts and metallic boxes for keeping drugs they were meant to distribute. The VHT strategy created a belief among community members that VHTs were there to solve all their health problems. These expectations were heightened by the initial enthusiasm with which the VHTs performed their duties under supervision from the health workers. As a result, the interactions between the VHTs, the health workers and the community members showed signs of hope. The community members believed in the VHTs’ work and took their support seriously.

One day as I moved around mobilizing people for vaccination, I came across a home that I had never visited. When I arrived, I saw this child lying down. I asked how old he was. His mother told me he was six years old. This child was six years but could not move at all; he was lame and I think the parents did not even know what to do with the child. The child looked like he was about three years old; you could not believe that he was six. So, I gave them a form and we filled it out together. I told them to take the child to a rehabilitation centre called Katalemwa. They took the child there, the doctors helped him with walking aids and with time, he gained strength and he is now doing well. I feel good that I helped that child. His parents had given up on him. This was about three years ago. (Interview with a VHT)

Nonetheless, the early bout of enthusiasm gradually waned into indifference and feelings of desperation. These shattered expectations and by the time of our fieldwork in 2012 [39], the community members appeared to have realised the powerlessness of the VHTs.

The indifference resulting from unmet expectations affected the relationship between the community, the formal healthcare workers and the VHTs, which then entered a phase of suspicion, accusations and counter-accusations. Community members “accused” the VHTs of working for the “other” and drawing salaries from the government; they also accused the government of failing to facilitate the work of the VHTs. The formal healthcare workers accused the VHTs of asking for privileges while faulting the government for attempting to use VHTs as a ploy to control people. On the other hand, the VHTs blamed the formal healthcare workers for failing to support them and accused the community of failing to appreciate their efforts and the government of abandoning them. As we show later, this also emanated from what we have called a transition failure from AMREF to the formal healthcare system. It appears to have created a ‘counter-accusation effect’, confusion and subsequently undermined the perceived linkages.

VHTs work for “other” people On June 18, 2013, a community health and personal hygiene sensitization outreach event was organized by the district health department, supported by Korean International Cooperation

Agency (KOICA) and Sahmyook University based in Korea. Fifteen female and 11 male community members attended the outreach meeting. The community members were mobilised by the local council (LCI) chairperson who was a member of the VHT. Health workers from the HCIV at Kasana facilitated the session, which lasted over one hour. As the session drew to a close, the chairperson LCI stood up to express gratitude to the facilitator. He also decided to introduce the VHTs, some of whom were not present. The chairperson had probably done this to address the concern raised about the existence and work of VHTs during the previous meeting convened on February 10, 2013 by the local council 1 (LCI) executive to discuss community concerns. When KOICA had a community hygiene programme, VHTs were instrumental in mobilisation; therefore, the chairperson was keen to make their work visible. In a discussion we had with the in-charge (the head of the health workers) at Kasana HCIV, he revealed to us that VHTs were critical mobilisation agents.

Whenever there is an outreach the VHTs are the people who mobilise us. To me that is a significant contribution. Recently we used VHTs to distribute insecticide treated mosquito nets. Now we even have a programme where we are going to teach two VHTs per village so that they can start treating basic and simple illnesses like malaria, diarrhoea, or pneumonia at the community level under the integrated community care malaria management programme.

While the health workers see the VHTs as simplifying their work of penetrating and reaching the community with health commodities, some community members perceived the VHTs as a structure that works for the “others” whom some sections of the community termed “bosses”. In one FGD we conducted with community members, participants unanimously mentioned “They (VHTs) work on instructions from their bosses. They do what they are told to do.”

The label “bosses” meant the officials at the district health office, district hospital and NGOs who use the VHTs for mobilisation during immunization and other outreach activities. Whenever the VHTs perform these occasional activities, they are paid an activity allowance. The community members seem to have interpreted the work of VHTs in this regard as an extension of formal authority as opposed to a linkage and partnership. In fact, what the community members perceived as working on the instruction of others is a function enshrined in the VHT national strategy. It states: “All health activities at community level by the government, NGOs 52

and or partners targeting communities shall be coordinated through VHTs. This includes health promotion activities, campaigns and other health events and functions” [10, p. 19].

The assumption was that by implementing all activities through VHTs it would make “the community value and continue demanding and utilizing their services” [10 p.19]. However, discussions with community members depict VHTs as powerless and their position defined more by what “their bosses” want them to do than what the community needs. Therefore, it was not surprising, as we posit in our earlier study in the same community [39, p.6], that some community members indicate that “those (VHTs) are not the ones who help us”. To some of the community members, the VHTs were simply a programme like any other programme that comes and goes. People who had this view suggested that since the VHTs were recruited to do something that they did or did not do, they should be forgotten. The participants positioned the VHTs as an object or a shop that was opened, did not serve its purpose and should be closed.

Fears that the VHTs were working for the “other” were further heightened by accusations that they were drawing salaries and other favours from the government; this is likely attributed to the earlier financial incentive-led work they did in the community.

Respondent 7: If you go out to advise someone to go to the hospital, he or she will tell you, “you are telling me, yet you are the ones who receive salaries, you get some of the money you are paid so that you take me to the health facility; you are government workers”. By the way those people take us as government workers; they think we do get salaries. (FGD with VHTs)

Although there was widespread acknowledgement that the VHTs’ capacity to perform was in favour of the “other” and not in the interests of the villages, some community members blamed their (VHTs) failures on the government’s inability and perceived lack of interest in equipping them with the necessary facilities.

Respondent 1: I wonder why they said those people [VHTs] were going to be our village doctors. Even witchdoctors have what they need to use. Now for VHTs, they are here; we

see them every day, but they have nothing (no materials) to help us with. (FGD with community members)

VHTs as a tool to control people and mask the challenges of the healthcare system Facility-based health workers mentioned that VHTs are not “experts”, yet they seem to expect preferential treatment.

They [VHTs] even want to bring patients and skip the queues! It is not possible, those people are not experts; they should simply refer people here. If they send someone to our centre, he/she should come as any other regular patients. (Interview with district health official)

The statement by the official from the district health office suggests minimal contact between the VHTs and formal healthcare workers. But this deviated from what VHTs considered an essential partnership. VHTs noted that the patients they refer should not be made to line up at the health facilities, which, to them, is a source of motivation for people to visit VHTs. These tensions, and unfulfilled and conflicting expectations have been documented elsewhere between CHWs and professional healthcare providers [15, 44].

Some health workers also questioned the idea of introducing VHTs and implied that the structure was created for other reasons besides serving to link communities to formal healthcare facilities. While questioning government commitment to make use of the VHTs, health workers wondered why the government that failed to equip the health centres with drugs, equipment and medicines would do the same for VHTs at the community level and later on be able to motivate them (VHTs) to link the community to formal healthcare facilities. A health worker wondered, “If the health facilities are not equipped with drugs, how do the VHTs refer people here!” The creation of the VHT structure, according to some health workers, was simply a ploy by the government to control people after failing to meet their cardinal obligations such as equipping health facilities.

Here [Kasana HCIV], we don’t have an operating theatre. How can that be? The doctor is here, everything, and you just cannot put up a theatre? … I think the government was not

honest in saying that they wanted to give anti-malaria drugs to the VHTs. How can they do that if the hospitals are not stocked with medicines? So, tell me, how can you say that the government is interested? (Interview with the person in-charge at Kasana HCIV)

According to the Ministry of Health, HCIVs are supposed to have an operating theatre. The revelation by the person in-charge that this HCIV lacked an operating theatre was in itself revealing of the bigger challenges facing the formal healthcare system. Whereas these challenges have been documented, the health workers’ discussion of these challenges in relation to the creation of VHTs demonstrates the hypocrisy of the government. The arguments made by health workers were that the government could have simply facilitated the health inspectors and health assistants who are already in the national healthcare delivery structure, instead of poorly facilitated VHTs to do their work. If we were to go by this argument, it would also imply that the fears of health workers are embedded in the realization that by fronting the VHT structure, the government was inhibiting the effective functioning of formal health systems. As the trend of side-lining formal health systems continues, the Ministry of Health is currently proposing creating another layer, called “the health extension workers” (HEWs) [45, 46]. If this mooted idea goes through it would then completely sidestep the positions of health assistants and health inspectors currently provided for in the formal healthcare structure. The idea that VHTs were supposed to do the work voluntarily was, according to the health workers, an attempt to shift responsibility. This relates to the phenomenon of criticising the promotion of community-based healthcare because it is seen as a means for governments to run away from fulfilling their obligations to deliver services [23].

Government neglect, lack of support and recognition from formal healthcare providers The public health services delivery structure in Uganda places VHTs at health centre I (HCI) at the community level and they serve as the first point of contact. VHTs revealed their uncertainty about their position in the structure. They mentioned that the government abandoned them by reneging on the promise to give them medicines to distribute in the community. Some VHTs mentioned that they struggle to get attention and recognition from the formal healthcare providers. In one FGD with VHTs, they mentioned that they are not recognized and the attitude of the health workers leaves them uncertain of their position in their interactions.

Respondent 7: The VHT is not taken as someone who is part of the formal health system. The health workers at the government health facilities do not respect you; so, you find that as a VHT you lose interest in following up issues. For example, we have many sick people but we no longer know them. So, most of our efforts at the moment are about latrines, and general hygiene but the issue of making follow ups at the health facilities, for example, TB or even HIV patients are no longer done because our bosses (government and district officials) are no longer helpful. (FGD with VHTs)

The attitudes of formal healthcare providers, according to VHTs, amount to humiliation that partly contributes to the lack of interest in VHT work.

The truth is that we lost interest in the VHT work, not only in this village, but elsewhere too. Those days (when VHTs had just started) you could call the health assistant to help you, but now they first ask for transport; sometimes they ask you, “Who are you to call me?” (Interview with VHT)

The VHTs’ complaints about the lack of support they receive to follow up on sick people in the community appear to push their work to perfunctory advice on basic health issues. It also means that they no longer feel obliged to act as a critical link between the community and formal healthcare workers. Health workers, particularly those meant to supervise their work, no longer do so. Failure to address social determinants of health The 2010 Uganda National Health Policy identified several social determinants of health including household income, education, status of housing and social and cultural beliefs (13). The 2006 Uganda Demographic and Health Survey showed a direct relationship between poverty and various health indicators including prevalence of diseases (47). Poverty, poor transport networks and distance to health facilities limit community members’ access to healthcare . Kiwanuka et al. defined access to healthcare as “the degree to which individuals are inhibited or facilitated in their ability to gain entry to and to receive care from the healthcare system” [14]. Factors that affect access emanate from both supply and demand sides implying that interventions aimed at bridging this gap must address both sets of issues [14].

In the community we studied, people rely mainly on motorcycle taxis (boda bodas) to physically access formal healthcare . Even when people received advice from VHTs to visit a health facility for any ailment, they still needed money and transport to enable access to a health facility. Yet the VHTs were unable to address the challenges of transport and financial resources that community members mentioned as a key challenge.

Respondent 5: We tell them about the benefits of going to the health facilities and ask them to visit the health facilities but the problem is transportation. As a VHT, I may tell the person about the benefits of going to a health facility but how will one get to the facility? ... Once, I went to someone’s home and found that the woman was at the point of delivering a child. I told her to go to a health centre. She explained that she did not have any money to get there … This showed that our only difficulty was finding a way to get to the health centre. I believe that if we had a health centre nearby it would really save u. (FGD with VHTs)

The reflections captured in the discussion with VHTs above shows that accessing a health facility for rural people requires much more than giving advice and information about the existence of a service at a health facility as is currently done by the VHTs.

During the pile sorting exercises we conducted with community members, poverty was ranked as the main healthcare concern among community members. In one of the sessions, participants argued that poverty and poor hygiene are greater health problems than the lack of health facilities. One person stated, “… you would not worry about treatment facilities if you had money to access them even if they are far away from the community.” The nearest private and public health facilities are approximately 5 km away, and it costs about USh2,500 (Ugandan Shillings, equal to US$0.70) to travel this distance by boda boda. Furthermore, public health facilities do not often have enough drugs or equipment to treat most health problems. Although they offer free medical services, it is common for drugs to be out of stock [48], leading people to seek treatment from private clinics, drug shops or pharmacies [49].

While aware that they lacked the capacity to build a health facility or buy an ambulance, people countered: “We mobilise ourselves” to solve these problems. “Mobilising ourselves” meant actively marshalling resources to overcome obstacles, including borrowing financial resources, organizing transport and caring for “unable” community members. None of the participants mentioned relying on VHTs for solving some of these problems. For the most part VHTs acknowledged their own limitations to addressing some of the main barriers that affect access to formal healthcare .

Respondent 7: There are times you find that someone is very ill but he cannot go to a health facility because he does not have money. There are also those that have lost hope in the health facilities; whenever they go there they do not get medication.

Respondent 1: Many people fail to access healthcare because they do not have money, others might have the money but fail to reach the health facility, maybe the boda bodas that usually take them are not available, especially at night.

Respondent 3: Drugs, drugs, even us that were given the responsibility of supplying drugs have never received them. When you ask someone to go to the health facility, they will ask you if you are sure that they will get drugs. You are not even sure what to tell them; in the end, people stay home. (FGD with VHTs)

In some cases, the VHTs themselves have not shown a good example of how to solve the existing health problems.

Respondent 3: Some of them take it for granted. They abuse the VHTs, even asking them whether they have them (latrines and utensil stands) at their homes. You know, some people can be stubborn and ask hard questions. They will say, “If the VHT member doesn’t have it, how can they tell me to have one?”

Respondent 1: Actually there are VHTs who do not have latrines. That’s why they have not had a lot of impact. You cannot go around telling people what to do when you do not do it yourself. 58

While community services and interventions are critical, they can hardly address most of the social determinants of health and they are not a solution to an ailing healthcare system. This suggests that attention should also be paid to factors within the broader environment than just focusing on what the VHTs are able to accomplish.

Managing various healthcare systems and beliefs Discussions with community members showed that VHTs have to deal with various health systems and community beliefs. It is important to consider how the VHTs relate to the various belief systems that they did not grow up with but also are part of given their status as VHTs. For example, on January 17, 2013, Mr. Okello, a community member, lost his son of four years. Okello’s son had been sick for some time and almost everyone in the community knew about the boys’ ill health. Initially his parents used local treatment including herbs. Realizing that the boy’s condition was not getting any better, their neighbour informed one of the VHTs (John) from whom he had received help to access doctors at a government health centre IV. When John met the mother of the boy, he asked if the boy had been taken to the hospital. The mother explained that they went to a government HCIV and showed him the medical forms. On examining the medical forms, John realized the doctors had suspected TB and referred them to a bigger hospital in Nagasaki (a neighbouring district) for further TB tests and treatment. John advised the parents to take the boy to the hospital in accordance with the doctor’s instructions. When we spoke to John, he told us the boy was not taken for further TB tests because his parents did not believe it was TB. The parents believed the boy had been bewitched by one of his relatives and subsequently decided to take him to the traditional healers. They occasionally bought him tablets from the local clinics. When asked what he thought the problem was, John said that whenever he visited, the boy’s parents always claimed they did not have sufficient funds to take the boy to the hospital, yet “the mother runs a food joint and the father is a businessman.” Surprisingly, according to John, the boy’s parents found money to take the boy to traditional healers three times before his death. In interpretation of the problem the VHT said “helping a person who has not asked for help is hard.”

Okello’s story indicates the clash between community cultural understanding of illness and

disease and the biomedical interpretation which has been documented elsewhere (50). Our interest in this case was drawn by the VHT’s point of view. For example, would the VHT have concentrated on educating the parents to understand the problem was not witchcraft but TB? Or would he have concentrated on what he did, namely telling the parents to go to the hospital where they were referred even though they did not believe in it? From the story, it appeared that by telling them to go to the hospital where they had been referred, he projected himself as a representative of the biomedical system even though they did not believe the diagnosis. As a result, the questions that arise are: Are the VHTs part of the community and attend to issues from the community’s perspective? Are the VHTs perceived as part of the biomedical system? Indeed, when faced with multiple healthcare systems, the VHTs position is challenged. But most importantly, if a VHT hits a deadlock, as happened here, i.e., if a community member does not take the VHT’s advice, what support is available from the formal healthcare system for the VHT to help him/her overcome this challenge? Already our discussions with the VHTs and the health workers suggest that such provisions do not seem to exit. In fact, we found that the health workers who are supposed to supervise the VHTs no longer ask for reports from the VHTs. In such a scenario, the position of VHTs in the “eyes” of the community is also questioned? We use the word “position” here to refer to how the VHTs are placed in the “web” of healthcare systems. The story of Okello’s son essentially summarizes the limitations faced by VHTs in serving as a link to the community members’ access to formal healthcare facilities.

Discussion In this paper, we examine the claims in Uganda’s VHT strategy and a significant reference in the literature that VHTs and CHWs in general, serve to “link”, “connect” and “bridge” communities with formal healthcare providers [4, 5, 9, 10, 17, 18, 51]. In what they describe as the “determinants of the success of community health worker programs”, Haines et al. identify several factors that determine the success of a CHWs programme [5]. These include the national socio-economic and political factors such as the political will to implement such interventions; health system factors including management and supervision and effective use of interventions; community factors that include location and infrastructure, local epidemiology, health beliefs and concepts of illness and community mobilization and empowerment; and international factors including donor policies and technical assistance [5, p. 2125]. The challenges faced by VHTs in

linking communities to formal healthcare providers lie within the interactions of various stakeholders and social determinants of health which the VHTs could not address. Our analysis reveals various challenges at the national, community and healthcare facility levels. Linking social capital suggests that understanding these interactions requires an understanding of how a systems model works and how “systematic mechanisms for partnerships across power structures” are created [25, p. 4).

VHTs in Luwero were recruited and trained with support from AMREF on behalf of the government. VHTs received financial incentives from AMREF in the early phases of the programme. People appreciated this period where the VHTs were active, wrote reports and through supportive supervision they were able to regularly interact with the formal healthcare workers. From AMREF to government structures, there was a transition failure, which culminated in “accusations” from community members and VHTs of government neglect and failure to equip the VHTs. Golooba-Mutebi discusses similar experiences with the implementation of health unit management committees (HUMC) in Mukono District, Uganda (52). Golooba-Mutebi found that HUMC were successful at a time when AMREF, implementing a World Bank project, paid allowances to members of HUMC to attend meetings and write reports. After AMREF handed the programme over to the government, according to Golooba- Mutebi, the activities of HUMC nearly grounded to a halt as they transitioned to the government formal structures [52]. In the case of VHTs, right from the moment of selection and training, the government appeared absent in guiding the processes and there was a failure to ensure that the guidelines were implemented as intended. A recent study conducted in the same community revealed that the selection of VHTs was itself dodged with patronage [39]. For example, whereas the VHT strategy states that “Political leaders such as the LC I chairperson, vice chairperson and secretary are not eligible for membership for purposes of ensuring checks and balances” [10], the researchers found that many selected VHTs were in fact elected political leaders who saw the introduction of VHTs as an opportunity to “eat big” [39]. This means that political interests drove the processes that led to the formation of the VHTs more than the needs of the community. The senior medical officer called this “politics of control”. This in itself defeated the purpose as stated in the VHT strategy, “VHTs will be accountable to the community leaders” [10, p.18). If anything, there appears to have been more of political manoeuvring than political will [39]. This

means that the process started on the wrong foot, undermining the potential partnerships envisaged in the linking social capital framework [28,, p.7).

On the basis of the mentioned transition failure, it is not surprising that VHTs were perceived by community members as working for and under the instruction of “others” or “bosses”, usually formal health workers and staff from NGOs who were implementing interventions in the community. The community interpretation of the work of VHTs as working for and under instructions of the “other” suggests that community members perceived VHTs as an extension of the formal healthcare system into the community. As Haines et al. observe such unresolved arguments rotate around perceptions of CHWs as either an extension of the authoritarian formal healthcare system into the community or CHWs as agents of community change that serve the interest of communities [5]. In our case study, we found that the constant reliance on VHTs as mobilization tools by the formal healthcare providers during immunization outreach or health awareness campaigns projected a one-way flow, where information, services, ideas and interventions were provided at the community level on behalf of the formal healthcare structure. The VHTs could not resist such occasional engagements since it usually attracts short-term remuneration. Whereas this may not necessarily be a bad thing, literature on linking social capital shows that external interventions that draw little input from the community tend to generate resentment and corruption, and instead of being beneficial, resources from these agencies become deleterious [27]. In other words, while the formal healthcare providers appear to have gained influence at the community level [25], there is little community influence implemented through VHTs on formal healthcare providers. Turinawe et al. demonstrate how community members interpreted the work of VHTs as policing, and saw VHTs as allies of a formal power structure instead of as an alliance with the community members whom they were meant to serve [39]. Community members perceived the VHTs model as a top down structure that executes instructions which contradicts the argument that a community-based healthcare programme emanates from the community [4, 23].

It is notable that VHTs were mobilised on the assumption that they are part of the community, and that they understand the local needs, health beliefs and customs, which put them on a similar pedestal to the community, as opposed to formal health workers. This in itself is the general

premise of social capital as a social resource upon which the success of interventions like VHTs depends [37, 53]. Our findings, although not necessarily contradictory to the notion that community health workers’ provide links with local communities which places them in a better position to spread health information and promote health behaviours, they do not conform to the claims that the role of VHTs links communities to existing formal healthcare services [5]. This is partly because VHTs, and by extension CHWs, may not necessarily be representative of the community [39]. However, it is evident from this case study that the contribution of VHTs was further diminished by the community’s perception of the VHTs as a representation of the external actors including NGOs and more importantly government.

Over the years, VHTs have found themselves in an ambiguous position where on one hand they are perceived by community members as representatives of a system that is to a great degree “antagonistic” to community beliefs. On the other hand, they are part of a system that has not fully appreciated and utilized their services save for occasional mobilization duties they undertake under the instruction of the health workers. The VHTs appeared torn between understanding the community where they live and the beliefs that they were socialized in and the information they are expected to pass on from the formal healthcare system. This appears not only to create tension but also implies that for the VHTs to execute the “orders” of the formal healthcare system, they had to negotiate their position carefully lest they intensify tension with the health workers. Some of the VHTs chose inactivity while others remained “on guard” and on the lookout for whenever an activity comes from the formal healthcare system or NGOs.

In countries like Pakistan the Lady Health Worker programme was hailed as effective partly because the programme invested in, among other things, maintaining linkages with existing community resources such as traditional birth attendants [5]. In South Africa, while responding to HIV and AIDS scourge, professional nurses appreciated the role of CHWs who were directly under their supervision; they were “mediators between the facility and the community” with tensions arising only on delegated sensitive tasks such as HIV counselling [15]. In Ethiopia, Kok et al. found that varied expectations between community members and the health sector workers regarding the roles of the health extension workers (HEWs) affect the motivation and satisfaction of the HEWs [54]. Our results also show that formal healthcare providers appreciated the role of

VHTs as community-based workers but were concerned about their demand for preferential treatment. Formal healthcare providers’ failure to appreciate such a basic gesture as giving preferential treatment to a patient referred by a VHT affects the motivation of VHTs. In hindsight, this presupposes a well-documented challenge to the successful implementation of CHW programmes, which is about the fear professional health workers have of losing their professional autonomy to non-professionals [5]. Our results show that the fear of losing professional power affects the perceived linkages between communities served by VHTs and the formal healthcare workers. It was not surprising in the end that the VHTs no longer took initiative in referring people to the health centres or even following up cases such as TB or HIV in the community. For example, when John was alerted about the boy who had TB, he visited them because he had been alerted and not because it was part of some routine follow up of patients in the community.

Conclusions: lessons from linking social capital The introduction of the VHT strategy overlooked community perceptions and the power of formal healthcare professionals, particularly the attitudes of healthcare providers in facilitating the capacity of the VHTs to link community members with formal healthcare workers. The findings raise concerns about the government’s willingness and readiness to facilitate the VHTs as a link between communities and formal healthcare providers. Our general impression is that assuming that the creation of a structure such as VHTs facilitates linkages between communities and formal healthcare simply juxtaposes community health with formal healthcare . It also projects a community as an “object” that can simply be “wired” to where services are. As a result, VHTs’ assumed an ambiguous position between the community and the formal healthcare system, which appears to further prove that juxtaposing community health with formal healthcare is simply an extension of the formal healthcare structure curtailing mutual linkage with the community as an integral part of the process.

However, there are lessons to draw from linking social capital framework. The main assumption under linking social capital is that linking individuals and structures across vertical relationships allows for the effective harnessing of resources that accrue once communities are connected to formal institutions of power and authority [27, 37, 53-56]. On the other hand, if communities are

not prepared to receive and absorb the information and resources that come from outside the community, then instead of supporting the local systems the inflow of these resources undermines local systems. In a study of community associations in Uganda, it was found that whereas linking social capital is critical for generating resources that are beneficial to the community, in most cases, communities may not have enough structures to absorb these resources [27]. The implementation of the VHT strategy was meant to harness these resources by creating access routes to networks of power and influence [25]. Evidence suggests that although VHTs have close links with community members, the VHT strategy was conceived with little reference to how the initial bonds in the community would facilitate the successes or failures of the programme. If the key function of linking is to leverage critical resources including information and ideas from powerful institutions outside the communities [57] such as healthcare providers, then the introduction of VHTs did little to ensure that this leverage is achieved.

As linking social capital framework shows, for VHTs to effectively act as conduits between the community and formal healthcare and harness the resources that exist in institutions beyond the community, it is important to take into account the power relationships embedded in vertical relationships. In his study of democratic governance in Italy, Robert Putnam identified the challenges embedded in the bad vertical patron-client relationships that affect democratic governance leading to nepotism, suppression and corruption [58]. Putnam later argued that linking social capital can lead to both responsive and unresponsive networks where responsive networks are likely to nourish trust and respectful ties between community members and their representatives [59]. However, for the case of VHTs, linking social capital appears to have resulted to a greater degree in unresponsive ties that affected the nourishment of respectful and trusting ties between the communities, the VHTs and formal healthcare providers.

In addition, the idea of linking social capital is that it promotes community interests in formal institutions and therefore develops the capacity to engage power structures and formal authorities [27, p. 2207]. By instituting VHTs to provide a linkage with formal healthcare providers, the intention was that community interests would be promoted at the formal healthcare facilities and that the VHTs, as community representatives. would engage the power structures for enhancing community benefit. This can be achieved by prioritizing community interests because the

process starts at the selection stage [39]. Turinawe et al. call for identifying the natural helpers through informal networks who are already effectively linked to the community. This has already been perceived to work effectively in Pakistan where the Lady Health Worker programme succeeded because of the close linkages they maintained with the traditional birth attendants [5].

Linking social capital facilitates “partnerships between the healthcare providers or public health entities and underserved communities” [28, p.7]. Derosa, Duan and Fox also argue that establishing linking social capital is a precursor to the protection against poor services and a form of accountability for the service recipients [60]. However, this also means that linking social capital facilitates trust and relationships between healthcare providers and the marginalized communities for any meaningful partnership to be realized [28]. The observed failure of the VHTs to deliver on their expectations affected the trust, enthusiasm and the relationship between the community and the VHTs. At the same time, the failure to recognize simple gestures such as preferential treatment for patients referred by VHTs affected the relationship between health workers and VHTs. In such an environment, it was difficult to engender a “system model of collaboration” upon which partnerships thrive [25, p. 1078]. In order to create mutual linkages there is a need to have processes initiated from below and particularly at the community level. Whereas the potential exists, VHTs are no panacea to weak healthcare systems. Innovative ways that seek to change the attitudes of health workers towards VHTs and equip VHTs with the necessary reduce would help reduce tension, create mutual linkages and facilitate partnerships. This should be done alongside innovative practices that seek to address social determinants of health for which the VHTs have no control.

The challenges discussed above reflect the whole problem of system failure in Uganda’s healthcare system. The system is not working well, be it at community or health facility levels or at referral hospitals. Health workers questioned the government’s commitment to equipping VHTs when the formal healthcare facilities are already poorly facilitated. In this case, the assumptions about partnerships as espoused under the linking social capital framework would only work well in a situation where the systems function. Whereas CHWs are supposed to be a cheaper option that is nearer to the community, responsive to community needs and linking communities to formal healthcare , they cannot be successful in a generally dysfunctional health

system. As the government of Uganda prepares to institute HEWs, this might simply become another failed intervention if the basic health systems failures are not addressed to enable VHTs link with the formal healthcare system. VHTs need active support from higher levels, they need materials and tools to use, they need to be supervised and the community needs to be continually prepared to understand and appreciate the position and roles of CHWs.

List of abbreviations AMREF: African Medical Research Fund. CHW(s): Community health worker(s). CoHeRe: Developing Sustainable Community Health Resources project. FGD: Focus Group Discussions. HC: Health Centre. HUMC: Health Unit Management Committee. MoH: Ministry of Health. VHT(s): Village health team(s)

Competing interests The authors declare that they have no competing interests.

Authors’ contributions All authors approve the manuscript submission. Laban Kashaija Musinguzi collected, analysed and interpreted the data and wrote the manuscript. Jude T. Rwemisis, Emmanueil Benoni Turinawe and Denis Muhangi contributed to the acquisition, analysis and interpretation of data as well as writing of the manuscript. Marije de Groot contributed to the analysis and writing of the manuscript. Danniel de Vries supervised the collection and analysis of data as well as contributed to the content of the manuscript. Achilles Katamba and David Mafigiri designed and supervised the project. Robert Pool also designed and supervised the project. In addition, he supervised data collection, and participated in data analysis and the writing of the manuscript.

Acknowledgements

This work is part of the research programme Developing Sustainable Community Health Resources in Poor Settings in Uganda (CoHeRe), which is financed by WOTRO Science for Global Development under the Netherlands Organization for Scientific Research (NWO) grant number W07.45.105.00. The authors also acknowledge all study participants and support from the Center for Social Science Research on AIDS (CeSSRA), School of Social Sciences, Makerere University, Kampala.

References

1. Chen L, Evans T, Anand S, Boufford JI, Brown H, Chowdhury M, et al. Human resources for health: overcoming the crisis. The Lancet. 2004;364(9449):1984-90.

2. Narasimhan V, Brown H, Pablos-Mendez A, Adams O, Dussault G, Elzinga G, et al.

Responding to the global human resources crisis. The Lancet. 2004; 363(9419):1469-72.

3. World Health Organization. Primary healthcare , including health system strengthening. 124th session, agenda item. 2009; 4:26.

4. World Health Organization. The world health report: 2006: working together for health. .

2006.

5. Haines A, Sanders D, Lehmann U, Rowe AK, Lawn JE, Jan S, et al. Achieving child survival goals: potential contribution of community health workers. The Lancet. 2007;369(9579):2121-

31.

6. Lehmann U, Friedman I, Sanders D. Review of the utilisation and effectiveness of community-based health workers in Africa. Global Health Trust, Joint Learning Initiative on

Human Resources for Health and Development (JLI), JLI Working Paper. 2004:4-1.

7. Lehmann U, Van Damme W, Barten F, Sanders D. Task shifting: the answer to the human resources crisis in Africa? Hum Resour Health. 2009 Jun 21;7:49,4491-7-49.

8. Bender DE, Pitkin K. Bridging the gap: the village health worker as the cornerstone of the primary healthcare model. Soc Sci Med. 1987; 24(6):515-28.

9. MoH. Health Sector Strategic & Investment Plan: Promoting People’s Health to Enhance

Socio-economic Development 2010/11 – 2014/15. Kampala: Ministry of Health Uganda; 2010.

10. MoH. Village Health Team: Strategy and Operational Guidelines. Kampala: Ministry of

Health; 2010.

11. MoH. Annual Health Sector Performance Report Financial Year 2013/2014 October,

2014 Kampala Uganda: Ministry of Health; 2014.

12. Thaddeus S, Maine D. Too far to walk: maternal mortality in context. Soc Sci Med. 1994;

38[8]:1091-110.

13. MoH. The second national health policy: promoting people's health to enhance socio- economic development. Kampala: Ministry of Health; 2010.

14. Kiwanuka SN, Ekirapa EK, Peterson S, Okui O, Rahman MH, Peters D, et al. Access to and utilisation of health services for the poor in Uganda: a systematic review of available evidence.

Trans R Soc Trop Med Hyg. 2008 11;102[11]:1067-74.

15. Schneider H, Hlophe H, van Rensburg D. Community health workers and the response to

HIV/AIDS in South Africa: tensions and prospects. Health Policy Plan. 2008 May;23(3):179-87.

16. Kowitt S, Emmerling D, Fisher E, Tanasugarn C. Community Health Workers as Agents of

Health Promotion: Analyzing Thailand’s Village Health Volunteer Program. J Community

Health. 2015:1-9.

17. Viswanathan K, Hansen PM, Rahman MH, Steinhardt L, Edward A, Arwal SH, et al. Can community health workers increase coverage of reproductive health services? J Epidemiol

Community Health. 2012 Oct;66[10]:894-900.

18. Braun R, Catalani C, Wimbush J, Israelski D. Community health workers and mobile technology: a systematic review of the literature. . 2013.

19. Tran N, Portela A, de Bernis L, Beek K, Jimenez-Soto E. Developing Capacities of

Community Health Workers in Sexual and Reproductive. Maternal, Newborn. 2014.

20. Strachan DL, Kallander K, Nakirunda M, Ndima S, Muiambo A, Hill Z, et al. Using theory and formative research to design interventions to improve community health worker motivation, retention and performance in Mozambique and Uganda. Hum Resour Health. 2015 Apr

30;13:25,015-0020-8.

21. Alam K, Oliveras E. Retention of female volunteer community health workers in Dhaka urban slums: a prospective cohort study. Hum Resour Health. 2014;12(29):10.1186.

22. Miller JS, Musominali S, Baganizi M, Paccione GA. A process evaluation of performance- based incentives for village health workers in Kisoro district, Uganda. Hum Resour Health. 2014

Apr 8;12:19,4491-12-19.

23. Campbell C. Social capital and health: Contextualizing health promotion within local community networks. In: Baron S, Field J, Schuller T, editors. Social Capital : Critical

Perspectives. UK: Oxford University Press, Oxford,; 2001. p. 182-96.

24. Eriksson M. Social capital and health-implications for health promotion. Glob Health Action.

2011 Feb 8;4:5611.

25. Ogden J, Morrison K, Hardee K. Social capital to strengthen health policy and health systems. Health Policy Plan. 2014 Dec;29[8]:1075-85.

26. Lin N. Social Capital: A Theory of Social Structure and Action, Cambridge: Cambridge

University Press; 2001.

27. Titeca K, Vervisch T. The dynamics of social capital and community associations in Uganda: linking capital and its consequences. World Dev. 2008;36[11]:2205-22.

28. Derose KP, Varda DM. Social capital and healthcare access: a systematic review. Med Care

Res Rev. 2009 Jun;66(3):272-306.

29. Marmot M, Friel S, Bell R, Houweling TA, Taylor S, Commission on Social Determinants of

Health. Closing the gap in a generation: health equity through action on the social determinants of health. The Lancet. 2008;372(9650):1661-9.

30. Irwin A, Scali E. Action on the Social Determinants of Health: learning from previous experiences. Social Determinants of Health Discussion Paper 1 (Debates). Geneva: World Health

Organisation; 2010.

31. Bourdieu P. Social space and symbolic power. Sociological theory. 1989;7[1]:14-25.

32. Adhikari KP, Goldey P. Social capital and its “downside”: the impact on sustainability of induced community-based organizations in Nepal. World Dev. 2010;38[2]:184-94.

33. Stephens C. Social capital in its place: Using social theory to understand social capital and inequalities in health. Soc Sci Med. 2008;66(5):1174-84.

34. Thurman TR, Snider LA, Boris NW, Kalisa E, Nyirazinyoye L, Brown L. Barriers to the community support of orphans and vulnerable youth in Rwanda. Soc Sci Med. 2008;66(7):1557-

67.

35. Platteau J, Gaspart F. The risk of resource misappropriation in community-driven development. World Dev. 2003;31[10]:1687-703.

36. Vollan B. Pitfalls of externally initiated collective action: a case study from South Africa.

World Dev. 2012;40(4):758-70.

37. Bakeera S, K., Wamala S, P., Galea S, State A, Peterson S, Pariyo G, W. Community perceptions and factors influencing utilization of health services in Uganda. International Journal for Equity in Health. 2009;8[25].

38. Derose KP, Varda DM. Social capital and healthcare access: a systematic review. Med Care

Res Rev. 2009 Jun;66(3):272-306.

39. Turinawe EB, Rwemisisi JT, Musinguzi LK, de Groot M, Muhangi D, de Vries DH, et al.

Selection and performance of village health teams (VHTs) in Uganda: lessons from the natural helper model of health promotion. Hum Resour Health. 2015 Sep 7;13:73,015-0074-7.

40. Innocent K. The village health team strategy is a 'Most Innovative Community Practice' award winner: the experience OFA village volunteer programme in Yumbe District, Uganda. .

2007.

41. Kimbugwe G, Mshilla M, Oluka D, Nalikka O, Kyangwa J, Zalwango S, et al. Challenges faced by village health teams (VHTs) in Amuru, Gulu and Pader districts in Northern Uganda.

Open Journal of Preventive Medicine. 2014;4(09):740.

42. Luwero District Local Government. Higher local government statistical abstract for Luwero district, 2008/2009. In press 2009.

43. Spradley JP. Participant observation. . 1980.

44. Crispin N, Wamae A, Ndirangu M, Wamalwa D, Wangalwa G, Watako P, et al. Effects of selected socio-demographic characteristics of community health workers on performance of home visits during pregnancy: a cross-sectional study in Busia District, Kenya. Global journal of health science. 2012;4(5):p78.

45. Waiswa P. Can Ethiopian health extension work in Ugandan context. New Vision. 2015 21

October 2015.

46. MoH. Health sector development plan 2015/16-2019/20. Kampala Uganda: Ministry of

Health; 2015.

47. Uganda Bureau of Statistics (UBOS)., Macro International Inc. Uganda Demographic and

Health Survey 2006. Calverton, Maryland, USA: UBOS and Macro International Inc; 2007.

48. MoH. Annual health sector performance report: Financial year 2012 / 2013 Kampala

Uganda: Ministry of Health; 2013.

49. Whyte SR. Pharmaceuticals as folk medicine: transformations in the social relations of healthcare in Uganda. Cult Med Psychiatry. 1992;16[2]:163-86.

50. Hausmann M, S., Mushi A,k., Muela R, J. The paradox of the cost and affordability of traditional and government health services in Tanzania. Health policy and planning.

2000;15(3):296-302.

51. Kowitt S, Emmerling D, Fisher E, Tanasugarn C. Community Health Workers as Agents of

Health Promotion: Analyzing Thailand’s Village Health Volunteer Program. J Community

Health. 2015:1-9.

52. Golooba‐Mutebi F. When popular participation won't improve service provision: primary healthcare in Uganda. Development Policy Review. 2005;23[2]:165-82.

53. Hendryx MS, Ahern MM, Lovrich NP, McCurdy AH. Access to healthcare and community social capital. Health Services Research, Chicago. 2002;37[1]:87-104.

54. Kok MC, Kea A Z, Datiko DG, Broerse J E W, Dieleman M,Taegtmeyer M, Tulloch O. A qualitative assessment of health extension workers’ relationships with the community and health sector in Ethiopia: opportunities for enhancing maternal health performance. Hum Resour

Health. 2015 13:80. DOI 10.1186/s12960-015-0077-4

55. Ayé M, Champagne F, Contandriopoulos A. Economic role of solidarity and social capital in accessing modern healthcare services in the Ivory Coast. Soc Sci Med. 2002;55[11]:1929-46.

56. Lin N. Social capital: a theory of social structure and action. New York: Cambridge

University Press; 2001.

57. Woolcock M. Social capital in theory and practice: Reducing poverty by building partnerships between states, markets and civil society. Social Capital and Poverty Reduction: which role for the civil society organizations and the state. 2002:20-44.

58. Putnam RD. Making democracy work: Civic traditions in modern Italy. Princeton: Princeton

University Press; 1993.

59. Putnam RD. Commentary: "Health by association": some comments. Int J Epidemiol. 2004

Aug;33(4):667,71; discussion 700-4.

60. Derose KP, Duan N, Fox SA. Women's receptivity to church-based mobile mammography. J

Healthcare Poor Underserved. 2002;13[2]:199-213.

Paper II

Community perceptions of health problems, activation of social resources and threats to the use of social resources for access to formal healthcare in Uganda

Authors: Laban Kashaija Musinguzia, Emmanueil Benon Turinawea, Jude, T. Rwemisisib a Department of Social Work and Social Administration, Makerere University, P.O. Box 7062, Kampala Uganda and Amsterdam Institute of Social Science Research, AISSR, University of Amsterdam, Nieuwe Achtergracht 166, 1018 WV Amsterdam, Netherlands b Amsterdam Institute of Social Science Research, AISSR, University of Amsterdam, Nieuwe Achtergracht 166, 1018 WV Amsterdam, Netherlands

Corresponding author: Laban Kashaija Musinguzi, [email protected]/ [email protected]

Telephone contact: +31 655239901

This paper has been submitted and under review in Community Development Journal

Abstract This paper analyzes community perceptions of health problems and the use of social resources in accessing healthcare in Luwero District, Uganda, based on ethnographic data collected between 2012 and 2014. Using a social capital framework, this paper argues that rural Ugandans activate their social resources to enhance their access to formal healthcare. Borrowing money, calling on longstanding friendships, and applying social pressure to help those who are unable are all important practices in overcoming problems of poverty, inadequate health services, and distant healthcare facilities. Attitudes towards biomedicine, ideas about personal behavior, and the value placed on interconnectedness also shape how social resources are used to access healthcare. In interviews and in focus groups, people’s comments suggested that the erosion of unity and men’s lack of responsibility are threats to the use of social resources for this purpose. The research found that despite these threats, social resources are used to care for even marginal members of the community. The effectiveness of social resources in tackling the everyday obstacles to formal healthcare is premised on the fact that health problems are solved as other problems have always been solved—through social networks and reaching out to each other.

Introduction Barriers to formal healthcare in Uganda include health system failures, health worker attitudes, distance to health facilities, inadequate services, and beliefs about formal healthcare (Ensor and Cooper, 2003; Jacobs et al., 2012; Monroe et al., 2014; Ndyomugyenyi and Kabali, 2010; Kiwanuka et al., 2008; Kyomuhendo, 2003). Targeted programmatic interventions have been implemented since the 1980s to help fix the health sector, which was shattered by the state’s collapse and the misrule of the 1970s (Dodge, 1986). Notable interventions include elimination of user fees in 2001, community-directed interventions in immunization, home-based management of uncomplicated malaria, decentralization of healthcare services, and the institutionalization of village health teams (VHTs) (an equivalent to community health workers) (Ndyomugyenyi and Kabali, 2010; Xu et al., 2006; Kalyango et al., 2012; Ministry of Health, 2010a; Bakeera et al., 2009). Evidence of the effectiveness of these interventions varies. While the abolition of user fees in 2001 increased outpatient attendance at most public health facilities (Yates et al., 2006), it also paved the way for corruption and affected service delivery (Xu et al., 2006). The use of VHTs as a community-based structure can be fraught with challenges that impact their performance (Kimbugwe et al., 2014). The annual health-sector performance report for 2012/13 shows a decline in the percentage of villages with trained VHTs from 78% in 2011/12 to 55% in 2012/13 (Ministry of Health, 2013). Uganda’s health sector continues to perform dismally on key indicators. An increase in the number of health facilities appears not to have translated into adequate provision of services (Lutwama, Roos and Dolamo, 2012). Although outpatient attendance increased by over seven million people between 2011 and 2013, there was a decline in the number of women’s antenatal care visits from 34.2% to 31%, and, in the same period, only 41% of deliveries occurred in a health facility (Ministry of Health, 2013). The 2011 Uganda demographic and health survey found that lack of money and geographic inaccessibility of facilities were the two main challenges that undermine access to formal healthcare, with rural areas the most affected (Uganda Bureau of Statistics and ICF International, 2012). Remarkable capacity exists in form of social resources, which help rural people access formal healthcare, but this has remained an understudied aspect of healthcare access in rural communities in Uganda. In this paper, we analyze community perceptions of health problems, the activation of social resources, and threats to the use of social resources in cushioning barriers

to formal healthcare in Luwero District, Uganda. We draw from an ethnographic study we conducted in Luwero District, located in central Uganda, from 2012 to 2014.

Social Capital, Social Resources, and Access to Formal Healthcare The World Health Organization (WHO) Action on the Social Determinants of Health reminds us that solutions to healthcare access are as social as they are medical (Irwin and Scali, 2010). This paper responds to the WHO’s call by showing how rural communities activate social resources to access formal healthcare. After all, ill health is a social phenomenon because of its effect on community participation, and thus requires a collective response (Bakeera et al., 2009; Hausmann, Mushi and Muela,2000). The ‘collective response’ in this case is mobilization of social capital—such as networks, norms of reciprocity, and trust—that enable people pursue group goals (Putnam, 1995, 2000). The concept of social capital entered into the field of healthcare as the health research agenda began to focus on preventive services and the social determinants of health (Campbell, 2001; Ogden et al., 2014; Lin, 2001; Gregson et al., 2004). The focus on contextual factors in the mid- 1990s shaped analyses at two levels: a macro-social level that looked at social inequality, and a meso-level that attended to community-level health determinants (Eriksson, 2011), the focus of this paper. Critics of the concept of social capital posit that its nothing new to public health and that concepts of ‘community capacity’ and empowerment mean the same thing as social capital (Labonte, 1999). Yet empirical studies continue to link social capital to health outcomes, suggesting that socially cohesive communities are more likely to have better health outcomes (Gregson et al., 2014; Campbell et al., 2013; Kawachi and Berkman, 2000), and that social capital may help explain health determinants using elements ‘outside’ the scope of biomedicine (Campbell, 2001), such as networks, trust, and reciprocity. Social resources are ‘the means through which interaction take place,’ and they develop out of the recognition that a need exists for them (Donenfeld, 1940). They can come in either concrete or symbolic gestures (Häuberer, 2014). Social resources may include financial support, transportation to a health facility, caring for the sick, information on health services, and client- health worker relationships (Bakeera et al., 2009). In a study in eastern Uganda, Meinert’s respondents mentioned wealth, unity, learnedness, smartness, and bodily strength as key

resources for a good life (Meinert, Kajubi and Whyte, 2014). The full range of resources available often determine which actions people take to improve their health (Meinert, 2009). The idea that social networks play a crucial role in the health of individuals is well documented in studies on health in African countries (Read, 1966). In Tanzania, Hausmann, Mushi and Muela found that treating a ‘personalistic’ illness involved the help of extended kin, and the social pressure to comply was high compared to the treatment of ‘normal’ illness through biomedical healthcare(Hausmann, Mushi and Muela, 2000). In West Africa, Ayé et al. showed that, amidst poverty, people in Ivory Coast accessed formal healthcare services through their social networks (Ayé et al, 2002). Janzen found that social relations were critical in patients’ quest for therapy (Janzen, 1978). Writing about HIV treatment, Meinert, Kajubi and Whyte used the concept of ‘technical know-who’ to show that accessing health services was as much a matter of social connections as it was service availability (Meinert, Kajubi and Whyte, 2014). A study of patients’ experiences in eastern Uganda showed that knowing someone in a hospital setting could prove to be a useful resource for circumventing the bureaucracy of health facilities (Mogensen, 2005). Other studies however show that social networks potentially exclude people of low socio-economic status in overcoming barriers to healthcare (Bakeera et al., 2009). The argument that social networks are crucial for accessing health services is, thus, not new. However, while it appears to be true in general, it is more important for rural communities where a lack of material resources, inadequate health facilities, and poor transport infrastructure are major challenges. A study conducted in Luwero District found that the extended family system was better positioned to meet the needs of orphans than NGOs, signaling the strength of social resources (Muhwezi et al., 2009). What has not been analyzed in this body of research is how such resource networks are actively utilized by individuals, as part of everyday, community- level processes. It is not enough to posit that social resources are important without a clear understanding of how such resources are activated to overcome barriers to formal healthcare.

Study Setting Luwero District is located in central Uganda approximately 60 km from Kampala city. The population is estimated at 405,900 (Luwero District Local Government, 2009), the majority of which is native Baganda, who speak the native Luganda language. The district was a center of guerrilla war activities from 1981–1986, which greatly diminished the health infrastructure.

Since, rehabilitation efforts have established a ministry in charge of Luwero, and a host of nongovernmental organizations (NGOs) have implemented programs to improve community health. The Ugandan national health delivery system follows a tiered structure with national referral hospitals at the top, followed by regional referral hospitals, general hospitals, health centers (HC) at levels IV, III, IIs and I (VHTs) at the bottom of the ladder (Ministry of Health, 2010b). The Ministry of Health’s 2012 inventory showed that Luwero District had 40 public health facilities (1 general hospital, 3 at level IV, 36 at levels III and II), 22 NGO facilities, and 3 private for-profit facilities (Ministry of Health, 2012). The study community in Luwero district is a collection of relatively remote villages located approximately 4 km off the Kampala-Gulu road. This community has no general hospital or health centers despite provisions for such in the national health service policy. Residents in the study community sought treatment at the nearest level IV health center, located about 5km away, and the closest general hospital approximately 20 km away. Most adults own a mobile phone and a radio, which keep them informed and in touch with friends and relatives outside the community. Transport is generally by motorcycle taxis, known as ‘bodabodas,’ a lucrative business for young men that helps to keep the community accessible. Despite the poorly maintained feeder roads, bodabodas link villages to the ‘outside world’ by ferrying people to and from health facilities.

Data Collection Although all the authors collected data for this study the results presented here largely draw from the first author and are presented in the first person when they refer directly from his field experience. Ethnographic data were collected through participant observation, informal and formal interviews, and focus group discussions (FGD) from 2012–2014. Participant observation involved immersion in community activities (Spradley, 1980), including attending health outreach events, church and religious gatherings, and community and group meetings; listening to conversations in bars; playing Ludo (a board game) with youths; contributing money whenever a member of a community was sick; participating in burial activities; and playing with children. More focused observations were made during three health outreach efforts, weekly village savings and loan association (VSLA) meetings, four funerals, three wedding ceremonies, one local church fundraising event, and seven village meetings to discuss health issues that were

convened by the village council executive. I made observations and noted what people said and did on issues related to health, how local leaders mobilized people whenever there was a health outreach activity or a health need, and people’s responses to health outreach efforts. Issues captured during these observations were written in a diary and used as points of reflection as fieldwork progressed. Data gathered through observation were further discussed during interviews. In-depth interviews were held with a total of ninety-one adult community members (thirty-six men and fifty-five women, aged 18 and older). Interviews lasted from 30 to 90 minutes. Interviewees were selected using a simple random procedure. In addition, forty-five FGDs (twenty-three with only women, fifteen with only men, four with men and women, and three with only VHT members) were conducted. Except for the FGDs with the VHTs, which averaged five participants, FGDs with community members involved ten to twelve participants. One community-mapping exercise was conducted to locate existing social resources, such as homes of prominent people, meeting points in the community, and infrastructure such as schools, local clinics, and sources of water. Seven ‘pile-sorting’ exercises were conducted separately with groups of women, men, and youths. Participants identified and listed health concerns or problems affecting access to healthcare, wrote them on slips of paper, and then put these in a pile. Each piece of paper in a pile was picked and the problems were discussed one by one and ranked in order of severity. Then, after ranking, participants discussed how community members had addressed these problems on their own. This exercise was particularly important for understanding the community’s ‘collective response’ and how people activated and utilized social resources. Pile- sorting exercises lasted about 90 minutes with detailed discussions among members. Participants for both pile-sorting exercises and FGDs were usually purposively recruited a day before. Like the FGDs, ten to twelve people participated per session. Finally, interviews with key informants were conducted with representatives from the local government, including staff at a level IV facility and the district health department, and with area NGO staff . Triangulation of methods contributed to increased validity and reliability of the findings.

Data Analysis

Interviews and discussions were conducted in Luganda, audio recorded, and transcribed into English. Transcription started during data collection, was iterative, and involved reflections on the data. Subsequent interviews facilitated the pursuit of themes that emerged from these reflections. After the data was transcribed into an MSWord document, it was stored on password-secured computers and exported into Nvivo10 software for further analysis. Codes were developed based on the identified themes. Query searches were performed on these themes and related data related were coded. The themes that emerged from coding and search queries were then discussed with colleagues and formed the basis for this paper.

Ethical approval was obtained from the Ethical Advisory Board of the Amsterdam Institute for Social Science Research at the University of Amsterdam; the Institutional Review Board of Higher Degrees Research and the Ethics Committee of the School of Public Health, Makerere University; the Uganda National Council of Science and Technology; and the research desk of the Ugandan Office of the Presidency. Permission to perform the research was granted by the Chief Administrative Officer of Luwero District. Informed written consent was given by participants before conducting formal interviews and FGDs. Oral permission for observations was sought from the key organizers of community events. Pseudonyms are used to protect the participants’ privacy.

Results Community Perception of Health Problems The pile-sorting exercises revealed that community members perceive poverty, health facility availability, and community hygiene as health problems (see Table 1).

Table 1: Summary of health problems in order of severity by pile sorting 10 Females (23-30 years) * 8 Males (20-30 years) ** 5 Male & 6 Females (35+ years) 1. Poverty 1. Lack of nearby health facilities 1. Poverty 2. Poor hygiene 2. Poverty 2. Poor hygiene 3. Lack of nearby health facilities 3. Poor transport and 3. Lack of treatment facilities 4. Low literacy levels communication 4. Poor feeding among children 5. Domestic violence and adultery 4. Poor feeding and adults 6. Theft and insecurity 5. Lack of a sense of responsibility 5. Poor shelter 7. Drugs and substances of abuse in illness management 6. Drunkenness 8. Government neglect 6. Poor hygiene 7. Rude and insensitive health

9. Poor public conduct 7. Illiteracy and ignorance workers 10. Treachery and hypocrisy among 8. Poor shelter and poor sleeping residents conditions 11. Poor feeding 9. Poor environment management 12. Drunkenness Illnesses 13. Witchcraft/evil spirits 14. Bad roads/poor road network

During one exercise with young men, a debate arose over whether having treatment facilities in the community was more urgent than having an ambulance or other transport. Some suggested that the lack of an ambulance, and therefore access to health facilities, was the greater problem. Others reasoned that ambulances are limited: they cannot take everyone, they are prone to accidents and mechanical breakdowns, and, since facilities are far off, one can die before reaching the hospital. They argued that having facilities closer to the community would ensure immediate access to treatment facilities with or without transport. In the end, all agreed that the absence of nearby treatment facilities subjects them to hardship and further suffering as they need to gather money, which is rarely available, to reach distant health facilities. In another session, participants argued that poverty and poor hygiene are greater health problems than the lack of health facilities. One person stated, ‘you will not need the facilities if you have taken care of your hygiene, and you would not worry about treatment facilities if you have resources to access them even when they are far away from the community.’ The nearest private and public health facilities are approximately 5 km away, and it costs about Ush2,500 (Ugandan shillings, equal to US$0.70) to travel this distance by bodaboda. Further, public health facilities often do not have enough drugs or equipment to treat most health problems. Although public health facilities offer free medical services, it is common for drugs to be out of stock (Ministry of Health, 2013), causing people to seek treatment from private clinics, drug shops, and pharmacies (Whyte, 1992). While aware that they lacked the capacity to build a health facility or buy an ambulance, people countered: ‘We mobilize ourselves’ to solve these problems. ‘Mobilizing ourselves’ meant actively marshaling resources to overcome obstacles, including borrowing money, organizing transport, and caring for ‘unable’ community members.

Borrowing from Friends to Meet Healthcare Costs

Community members borrow money directly from neighbors and friends to meet healthcare costs in times of need. A 28-year-old mother explained:

Whenever I fall sick, I have to go to Kasana [public health center], which is about 5 km from here. If I don’t have money, I borrow from others around. You cannot fail to have someone to lend you money. Recently, my child had malaria, and I did not have any money, so I borrowed from my neighbor Ush20,000 (US$5.60), which I used to go to the hospital.

In a FGD with men, similar observations were made about borrowing from friends to meet healthcare costs. One man said:

If I am sick I can instruct my child, ‘Go tell Mr. Ssegu to lend me money so that I can go to the health center for treatment,’ and because he is my friend he will not hesitate to give me money. Yes, it happens a lot and it works. I can use my own example: just a few weeks back, I was sick. In fact, I can say I survived because of my friend’s money that I borrowed from him. I suffered with the illness for at least two months. If he had not given me the money, I think I would have died.

According to community members, there was expectation of a refund whenever money was borrowed. As such borrowing appeared to be practiced among networks of friends who had established trust in each other. Borrowing also extended to organizing bodaboda transport. In an interview, a 50-year- old man said, ‘Whenever I encounter a problem, say a child is sick at night, I always borrow bodaboda from my neighbor.’ Community members also mentioned borrowing medicines (such as paracetamol) for conditions like a headache or mild malaria. Borrowing thus not only creates access to formal healthcare, but also access to informal healthcare.

Support for ‘Unable’ Community Members

‘We mobilize ourselves’ or ‘twekolamu omulimu’ was a frequent expression that was used to show how people work together to help those unable to pay healthcare costs on their own. In a focus group with women, one participant said:

In our community here, if one of us falls sick, and we are sure that he is helpless, we approach him or her and find out what illness he/she is suffering from. If there is a need to do counseling we do it. If he/she tells us why she is not going to seek medical advice, and if that reason is financial, we mobilize ourselves and contribute money for medical bills.

Take the case of Salongo, a casual laborer whose work in people’s gardens and running errands only earned him between Ush10,000–20,000 (US$2.80–5.60) per month at the time. Salongo always did his work diligently, which endeared him to many people in the community. One day, one of his children (aged four) developed malaria. Salongo had not had any work for about a week and did not have money to take his child in for treatment. After his child had been in bed for a day, one of Salongo’s friends learned of the child’s sickness and mobilized other community members to contribute money to help take the child to the hospital. Salongo recounted:

Some community members came to see the child and asked me why I was not taking him to the hospital. I told them that I had no money at all. I had nowhere to start from.…So they went around the community, and whomever they told contributed, even a coin of 100 shillings. They mobilized Ush20,000 [US$5.60)] and gave it to me, and I took the child to the hospital.

Salongo’s story is not an isolated one. Senso, whose main source of income was vending fish, had no close relatives in the community. He developed malaria and was bedridden for a few days. The village chairperson took a logbook and went around the village collecting cash contributions from willing members, which he recorded. Recording contributors’ names was a common practice, and appeared to apply social pressure on those who did not contribute.

The activation of social networks for financial support was not confined to health, but extended to other situations. When Wadda’s house burned down, the community was called upon to contribute money towards rebuilding. Whenever a community member lost a loved one, people contributed towards funeral arrangements. Extending support to sick members to access treatment was thus seen as one of many such community obligations.

Perceptions of the Relevance of Modern Medicines and Disease Severity Participants noted that diseases like tuberculosis (TB) and malaria require visiting a health center for treatment. Diseases such as malaria were described as no longer responsive to local concoctions and only treatable with modern medicine. As one woman said:

Whenever I get malaria, I go fast to the hospital and I get treated. The moment I feel it’s malaria I go to the hospital immediately because [these days] it [malaria] is so strong that if you don’t go to the hospital you may die. Our local herbs cannot manage it.

One reason for this could be the increased exposure to health messages through VHTs and on the radio about malaria, TB, and other diseases like HIV and AIDS. The government’s distribution of insecticide-treated mosquito nets to all households with children under five and expectant mothers may also have made such an impression. If sickness was attributed to supernatural powers, emphasis was often placed on mobilizing networks of extended family, who would also be affected by evil spirits. In such cases, they are likely to reject advice or support to seek formal healthcare. Local perceptions of illnesses, and the subsequent understanding of which medicine is effective for which condition, have been documented in the literature (Hausmann, Mushi and Muela, 2000; Bode, 2011), but are rarely discussed as a resource to be tapped to enhance access to formal healthcare. Discussions with community members showed that young people were more likely to seek formal healthcare, based on their perception of modern medicine as efficacious, than older people who still favored traditional healing practices. A 52-year-old woman stated:

You see, these ways of using our local herbs are dying out, because the old people are dying. People of my age still use them, but there are other young families, like my children, do not do it. They prefer being treated in hospitals. The lifestyles have changed.

However, for some young people, parental influence is still visible in their medical care choices. A 24-year-old woman explained:

Whenever my children get sick, and my mother sees them, she might say, ‘Ehh, you don’t have mululuza or emumbwa [local herbs]. Come, and I give them to you, and you give them to the children.’ And that’s how I came to learn more about these local medicines that we use here. But, if given the chance, I would choose modern medicine because the local herbs are not very effective.

Threats to the Use of Social Resources Erosion of General Sense of Unity and Growing Intergenerational Gap There is a perceived general erosion of unity (obumu), that threatens community members’ ability to mobilize support. Participants felt that community members are becoming individualistic and unwilling to offer support. Unity is now also negotiated through groups such as village savings and loan associations (VSLAs) and burial groups. In an interview, a 48-year- old woman lamented:

No, there is no obumu at all. I can tell you, because now what we would call obumu is only—like when you have 30 members in a group—that is where there is obumu. Only among you, the 30 members, but not in the entire community. … Because those who are in groups are the ones who benefit and have obumu. But for those of us who are not in groups, we do not [have obumu], so we just live a life like that; we maneuver in whichever way we can when fate comes our way.

However, others countered this opinion, observing that people still care for one another, including those who are bad mannered, stating, for example: ‘We do not abandon someone to die even when they are poorly behaved or not in groups.’

In some discussions, the erosion of obumu was attributed to the intergenerational gap, as older members criticized the young ones for lacking the norms that hold communities together. In one FGD, a debate erupted between older and younger women (less than 35 years):

Respondent A. The elderly ladies in the area are still united. The young ones, those are problematic. Respondent B. But we the young ones, too, are also not that very bad. Respondent C. But also, (mentions respondent A), you can only say we are badly behaved if one day you fall sick and we do not care for you. That is when you will conclude that. Respondent D. Don’t get us wrong. What we are trying to say is that the people of your generation have not copied our behavior and discipline.

This perceived change in norms is also related to broader changes that have swept rural areas, particularly in the form of modernization, migration, and the commoditization of goods. While norms may be changing, this study found that members of the community indeed mobilized themselves to care for one another.

Gender, Responsibility, and Social Group Support There is a general consensus that men have negated their responsibilities for taking care of their families and providing healthcare support whenever needed. In a focus group, women gave examples:

Respondent A: During delivery, men always leave us in the houses with our children. It may happen that you get labor pain when he is not around, and there is no a nurse who can help you. Sometimes you can even produce a baby and she dies because you don’t have any help. Respondent B: Another problem is that men make us produce many children but don’t provide for basic needs. Some leave the house early in the morning to consume alcohol. Respondent C: When he comes back he asks for food, and when you don’t give it to him he beats you.

While this critique did not apply to all the men in the area, it is a narrative that is gaining currency in a community where money is scarce and where women largely rely on their husbands for provision. This is shifting due to women’s sense of empowerment, as a majority have joined groups like VSLAs to provide for their families. A 60-year-old woman related how these groups provide support to their members:

When you are in the group and you get sick you can inform your neighbor who is a member of that group about your sickness, or even inform other members of the group who may not necessarily be your neighbors, and they contribute money so you are able to seek healthcare. Those group members can take you to the health facility so that you get treated. The above observation shows that access to healthcare is can not be attained without taking into account the social determinants of health (Saha, 2014), which in this include activation of resources in groups.

Discussion Solutions to addressing challenges of healthcare access do not lie in the health sector alone (Saha, 2014). Social resources are part of the solution. Social resources in Luwero District were embedded in social networks, including longstanding friendships and neighborly relations. In these structures, community members borrowed money, mobilized transport, and offered practical help to overcome barriers to accessing formal healthcare. Borrowing money from friends to pay for hospital costs appeared to be the main means of activating personal networks, and it was mediated through trust and the ability to repay. Some studies have downplayed the importance of social capital in overcoming healthcare barriers, especially those who ‘lack human and material capital’ (Bakeera et al., 2009). Meinert et al. point out that because access to HIV services is mediated through social connections, there is a danger that people with limited social connections may be left out (Meinert, Kajubi and Whyte, 2014). In a study of pediatric consultations in Guinea-Bissau, it was established that personal connections to doctors reduced mortality risk by 48% but the risk increased for those who had no such connections (Sodemann et al., 2006). Hausman et al. question the relevancy of

social connections in healthcare access in Tanzania, arguing that ‘tight family bonds and organized support structures for paying for hospital treatments seemed to be a myth rather than a reality’ (Hausmann, Mushi and Muela, 2000). In this study, the findings suggest that the greater number of social connections held by some members does not necessarily mean exclusion for others. Less connected and poorer members also received community support because of a strong sense of interconnectedness. This interconnectedness should be seen in relation to everyday interactions in which social pressure to help others (even those who are not well connected) is high. The argument that social exclusion reduces access to formal healthcare does not fully consider interdependences with other forms of support, such as contributing money for burials (Durham and Klaits, 2002). People often mobilized to support a person based on their participation in other communal activities like burials. When the chairperson noted who had contributed to help a community member, the act created not just pressure for others to contribute but also a debt to be repaid. Shipton writes that not all debts are negative, ‘for some of those debts that never go away are the very fibers that hold society together’ (Shipton, 2007). Borrowing and lending have long been part of networks of indebtedness that hold some communities together (Shipton, 2007; Roscoe, 1911). Beyond this example from Luwero District, a study conducted in southwestern Uganda also found that participants borrowed money to cover transportation costs for HAART clinic visits (Tuller et al., 2010). Activation of networks for meeting healthcare costs ought to be seen as part of a larger process in which people reach out to one another and are indebted to one another. The activation of social resources was threatened by the erosion of a sense of unity and a perception that men were not fulfilling their responsibilities. However, such dynamics do not appear to have diminished people’s ability to rely on each other for everyday support, even though some social networks are closed groups. Saha (2014) called for an integration of healthcare interventions through savings groups and microfinance. While VSLAs and burial groups may exclude those who are not members of these groups, and are new forms of support, they do not replace longstanding informal networks of family, friends, and acquaintances. Musinguzi (2015 forthcoming) argued that through the ‘silence-in’ and ‘silence-through’ networks, VSLAs participants are able to appropriate their benefits. Social obligations towards one another do not disappear under these new arrangements, and there is, at least superficially, a

display of public solidarity whenever a person encounters a problem. Whereas ethnographic studies tend to suffer from generalizability challenges to other settings, the description of everyday processes in this study provides a starting point in exploring social resources and access to healthcare in communities with similar characteristics.

Conclusion This study demonstrates that how people activate their social resources is as important as their role in overcoming barriers to formal healthcare. The processes of borrowing money, calling on longstanding friendships, and applying social pressure to help those who are unable, as well as attitudes towards biomedicine and practices of reciprocity, are as important to enabling community health as they are to cushioning the problems of poverty and inadequate and distant health facilities. Social resources are a means of surmounting the barriers that stand in the way of healthcare access (Donenfeld, 1940), and activating social resources places access to healthcare in the hands of consumers. As Ashe and Neilan observe, it is only after it is ‘in their own hands’ that sustainability is attained (Ashe and Neilan, 2014). The incorporation of social resources into health policies is thus a valuable contribution to other efforts aimed at addressing barriers to access to formal healthcare. Understanding how social resources are involved in how people tackle healthcare obstacles, however, should not detract from the underlying issues faced by vulnerable communities. While the activation of social resources is a critical element, these resources alone are not a panacea. Understanding the role social resources therefore provides a significant starting point, and not an end in itself.

Acknowledgements The author acknowledges all study participants. We would like to acknowledge Prof. Robert Pool, and colleagues from the CoHeRe group at University of Amsterdam.

Funding statement This work is part of the research programme Developing Sustainable Community Health Resources in Poor Settings in Uganda (CoHeRe), which is financed by WOTRO Science for

Global Development under the Netherlands Organisation for Scientific Research (NWO) grant number W07.45.105.00.

References Ashe, J. and Neilan, K. J. (2014) In Their Own Hands: How Savings Groups Are

Revolutionizing Development, Berrett-Koehler Publishers, San Francisco, California.

Ayé, M., Champagne, F.,and Contandriopoulos, A. (2002) Economic role of solidarity and social capital in accessing modern healthcare services in the ivory coast. Social Science and Medicine, 55(11), 1929-1946. Bakeera, S., K., Wamala, S., P., Galea, S., State, A., Peterson, S., and Pariyo, G., W. (2009) Community perceptions and factors influencing utilization of health services in Uganda. International Journal for Equity in Health, 8(25) doi:10.1186/1475-9276-8-25

Bode, M. (2011) The transformations of disease in expert and lay medical cultures. Journal of Ayurveda and Integrative Medicine, 2(1), 14-20.

Campbell, C. (2001) Social capital and health : Contextualizing health promotion within local community networks.. In S. Baron, J. Field & T. Schuller (Eds.), Social capital : Critical perspectives. UK: Oxford University Press, Oxford, pp. 182-196.

Campbell, C., Scott, K., Nhamo, M., Nyamukapa, C., Madanhire, C., Skovdal, M., . . . Gregson, S. (2013) Social capital and HIV competent communities: The role of community groups in managing HIV/AIDS in rural Zimbabwe. AIDS Care, 25(1), 114-122. doi:10.1080/09540121.2012.748170

Dodge, C. P. (1986) Uganda—rehabilitation, or redefinition of health services? Social Science and Medicine, 22(7), 755-761.

Donenfeld, M.J. (1940) The meaning of social resources. Journal of Educational Sociology, 13(9), 560-564.

Durham, D., and Klaits, F. (2002) Funerals and the public space of sentiment in Botswana, Journal of Southern African Studies, 28(4), 777-795. doi:10.1080/0305707022000043511.

Ensor, T., and Cooper, S. (2004) Overcoming barriers to health service access: Influencing the demand side. Health Policy and Planning, 19(2), 69-79.

Eriksson, M. (2011) Embracing complexity: Towards more nuanced understandings of social capital and health. Global Health Action, 4(5964) doi:10.3402/gha.v4i0.5964

Gregson, S., Terceira, N., Mushati, P., Nyamukapa, C., and Campbell, C. (2004) Community group 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(11), 2119- 2132.

Häuberer, J. (2014) Social capital in voluntary associations. European Societies,16(4), 570-593. doi:10.1080/14616696.2014.880497

Hausmann, M., S., Mushi, A.,k., and Muela, R., J. (2000) The paradox of the cost and affordability of traditional and government health services in Tanzania. Health Policy and Planning, 15(3), 296-302.

Irwin, A., and Scali, E. (2010) Action on the social determinants of health: Learning from previous experiences. social determinants of health discussion paper 1 (debates). Geneva: World Health Organisation.

Jacobs, B., Ir, P., Bigdeli, M., Annear, P. L., and Van Damme, W. (2012) Addressing access barriers to health services: An analytical framework for selecting appropriate interventions in low-income Asian countries. Health Policy and Planning, 27(4), 288-300. doi:10.1093/heapol/czr038.

Janzen, J. M. (1978) The quest for therapy: Medical pluralism in lower Zaire. University of California Press, Berkeley CA.

Kalyango, J. N., Lindstrand, A., Rutebemberwa, E., Ssali, S., Kadobera, D., Karamagi, C., . . . Alfven, T. (2012) Increased use of community medicine distributors and rational use of drugs in children less than five years of age in Uganda caused by integrated community case management of fever. The American Journal of Tropical Medicine and Hygiene, 87(5 Suppl), 36-45. doi:10.4269/ajtmh.2012.11-0733

Kawachi, I., and Berkman, L. (2000) Social cohesion, social capital and health. . In L. Berkman, and I. Kawachi, eds, Social epidemiology. New York: Oxford University Press, pp. 174-190.

Kimbugwe, G., Mshilla, M., Oluka, D., Nalikka, O., Kyangwa, J., Zalwango, S., . . . Walugembe, F. (2014) Challenges faced by village health teams (VHTs) in Amuru, Gulu and Pader districts in northern Uganda. Open Journal of Preventive Medicine, 4(09), 740.

Kiwanuka, S. N., Ekirapa, E. K., Peterson, S., Okui, O., Rahman, M. H., Peters, D., and Pariyo, G. W. (2008) Access to and utilisation of health services for the poor in Uganda: A systematic review of available evidence. Transactions of the Royal Society of Tropical Medicine and Hygiene, 102(11), 1067-1074. doi:http://dx.doi.org/10.1016/j.trstmh.2008.04.023

Kyomuhendo, G. B. (2003) Low use of rural maternity services in Uganda: Impact of women's status, traditional beliefs and limited resources. Reproductive Health Matters, 11(21), 16-26.

Labonte, R. (1999) Social capital and community development: Practitioner emptor. Australian and New Zealand Journal of Public Health, 23(4), 430.

Lin, N. (2001). Social capital: A theory of social structure and action,. Cambridge: Cambridge University Press.

Lutwama, G. W., Roos, J. H., and Dolamo, B. L. (2012) A descriptive study on health workforce performance after decentralisation of health services in Uganda. Hum Resources for Health, 10(1), 41.

Luwero District Local Government. (2009) Higher Local Government Statistical Abstract for Luwero, 2008/2009, accessed at: http://www.ubos.org/onlinefiles/uploads/ubos/ 2009_HLG_%20Abstract_printed/Luwero%20LDG%20FINAL.pdf (23 July 2014).

Meinert, L. (2009). Hopes in friction: Schooling, health, and everyday life in Uganda Information Age Publishing Inc.

Meinert, L., Kajubi, P., and Whyte, S., Reynolds. (2014) Connections In S. R. Whyte, ed, Second chances: Surviving AIDS in Uganda, Durham: Duke University Press Books, pp. 34-46.

Mogensen, H. O. (2005) Finding a path through the health unit: Practical experience of Ugandan patients. Medical Anthropology, 24(3), 209-236.

Ministry of Health. (2010a). Village health team: Strategy and operational guidelines, Ministry of Health, Kampala Uganda. Ministry of Health. (2010b) Health sector strategic and investment plan: Promoting people’s health to enhance socio-economic development 2010/11 – 2014/15. Ministry of Health, Kampala Uganda. Ministry of Health. (2012) Health facilities inventory July 2012: Master health facilities inventory. Ministry of Health, Kampala Uganda:. Ministry of Health. (2013) Annual health sector performance report: Financial year 2012 / 2013. Ministry of Health, Kampala, Uganda. Monroe, A., Harvey, S. A., Lam, Y., Muhangi, D., Loll, D., Kabali, A. T., and Weber, R. (2014) “People will say that I am proud”: A qualitative study of barriers to bed net use away from home in four Ugandan districts. Malaria Journal, 13, 82.

Muhwezi, W., Muhangi, D., and Mugumya, F. (2009) Intra-household differences in health seeking behaviour for orphans and non-orphans in an NGO-supported and non-supported sub-county of Luwero, Uganda. African Health Sciences, 9(2), 109-117.

Musinguzi, L.K. (Forthcoming) The role of social networks in savings groups: insights from village savings and loan associations in Luwero, Uganda, Community Development Journal, doi:10.1093/cdj/bsv050.

Ndyomugyenyi, R., and Kabali, A. T. (2010) Community-directed interventions for integrated delivery of a health package against major health problems in rural Uganda: Perceptions on the strategy and its effectiveness. International Health, 2(3), 197-205.

Ogden, J., Morrison, K., and Hardee, K. (2014) Social capital to strengthen health policy and health systems. Health Policy and Planning, 29(8), 1075-1085. doi:10.1093/heapol/czt087

Read, M. (1966) Culture, health, and disease. London: Tavistock Publications.

Roscoe, J. (1911) The Baganda: An account of their customs and beliefs. New York: Barnes and Noble.

Saha, S. (2014) ‘More health for the money’: an analytical framework for access to healthcare through microfinance and savings groups, Community Development Journal, 49(4), 618- 630.

Shipton, P. (2007) The nature of entrustment: Intimacy, exchange, and the sacred in Africa. New Haven, CT. Yale University Press.

Sodemann, M., Biai, S., Jakobsen, M., and Aaby, P. (2006) Knowing a medical doctor is associated with reduced mortality among sick children consulting a paediatric ward in Guinea‐Bissau, west Africa. Tropical Medicine and International Health, 11(12), 1868- 1877.

Spradley, P. (1980) Participant Observation, Wadsworth, Cengage Learning, Belmont, CA.

Tuller, D. M., Bangsberg, D. R., Senkungu, J., Ware, N. C., Emenyonu, N., and Weiser, S. D. (2010) Transportation costs impede sustained adherence and access to HAART in a clinic population in southwestern Uganda: A qualitative study. AIDS and Behavior, 14(4), 778- 784.

Uganda Bureau of Statistics (UBOS)., and ICF International. (2012) Uganda demographic and health survey 2011. Kampala, Uganda: UBOS and Calverton, Maryland: ICF International Inc.

Whyte, S. R. (1992) Pharmaceuticals as folk medicine: Transformations in the social relations of healthcare in Uganda. Culture, Medicine and Psychiatry,16(2), 163-186.

Xu, K., Evans, D. B., Kadama, P., Nabyonga, J., Ogwal, P. O., Nabukhonzo, P., and Aguilar, A. M. (2006) Understanding the impact of eliminating user fees: Utilization and catastrophic health expenditures in Uganda. Social Science and Medicine,62(4), 866-876.

Yates, R., Kirunga-Tashobya, C., Cruz, V., Oliveira., McPake, B., Ssengooba, F., Murindwa, G., . . . Omaswa, F. (2006) The Ugandan health systems reforms: Miracle or mirage? In C. Kirunga-Tashobya, F. Ssengooba and O. Cruz Valeria, eds, Health systems reforms in Uganda: Processes and outputs UK: Health Systems Development Program, London School of Hygiene & Tropical Medicine, pp. 15-28.

Paper III Improving access to formal healthcare facilities in rural Uganda through the use of motorcycle taxis

1 Department of Social Work and Social Administration, Makerere University, P.O. Box 7062, Kampala, Uganda

2Amsterdam Institute of Social Science Research, AISSR, University of Amsterdam, Nieuwe Achtergracht 166, 1018 WV Amsterdam, Netherlands. Email: [email protected].

3College of Health Sciences, Makerere University Kampala Uganda. Email:[email protected].

4Institute for Global Health and Inequality/Center for Poverty-Related Communicable Diseases, Academic Medical Center, Amsterdam

Corresponding author: Laban Kashaija Musinguzi. Email contact: [email protected] / [email protected] Telephone contact: +31655239901

This paper has been submitted and currently under review in BMC Health Services Research

Abstract Background: Access to formal healthcare remains one of the main challenges facing the health sector in Uganda. Government reports and scholarly writings attribute this problem to poor transport networks. Although informal public transport is common in many low-income countries, little is known regarding its role and potential in improving access to formal healthcare. In cases where incentives have been used to promote informal transport providers as ambulances there have been sustainability concerns. This article examines the contribution of motorcycle taxis (boda bodas) in improving access to formal healthcare facilities in Uganda where no incentive-led external interventions exist.

Methods: Data were collected through ethnographic fieldwork conducted in Luwero District, Uganda, between 2012 and 2014. The main method was participant observation and informal conversations. A total of 91 in-depth interviews were conducted with individuals (36 men and 55 women), and 42 focus group discussions (FGDs) were conducted with randomly selected adults, aged 18 and above. Five interviews and three FGDs were held with boda boda riders (taxi drivers), and four FGDs specifically on the role of boda bodas in providing access to formal healthcare facilities were held with community members. Other participants included local government staff, local leaders, and nongovernmental organisation (NGO) staff. Thematic analysis was used to evaluate the data.

Results: We found that boda bodas are used as community-based ambulances to facilitate access to healthcare facilities. Boda boda riders offer flexible payment modalities, act as caregivers to the patients they transport, and are regarded in the community as ‘one of our own’. They offer their services as part of the wider complex social interaction and reciprocity in the community. Such dynamics enable boda boda riders to provide services according to community expectations rather than for profit. Conclusions: Even without the support of external interventions, boda bodas are a sustainable social resource for access to healthcare in remote communities of Uganda. If they were supported by a favourable policy environment, well regulated, and trained in basic patient care skills, boda boda riders could considerably alleviate challenges of access to healthcare in rural areas of Uganda.

Keywords: Motorcycle taxis; boda bodas; informal transport providers; formal healthcare; Uganda

Background Access to quality healthcare for marginalised members of society constitutes one of the ingredients of a well-organised healthcare system [1]. Quality healthcare cannot in itself be simply given to people [1]. Decisions about how to access quality healthcare are determined at the household level; within communities where people live, grow up, and associate; and at the health facility where actual medical healthcare is provided. Thaddeus and Maine suggest that access to quality healthcare is affected by delays at any one of three levels: deciding to seek care, delay in reaching a health facility, and delay in getting adequate care at the health facility [2]. High transport costs, poor transport networks, and long distances to health facilities have been identified as some of the main reasons for such delays in settings with low levels of access to formal healthcare [2]–[4]. In this paper, we discuss delays that occur between the home and health facilities due to physical accessibility factors such as travel time, location of well- equipped health facilities, transport costs, and road conditions [2]. We particularly focus on the role of informal transport providers in overcoming these challenges.

In Uganda, the national health delivery system is a tiered-structure with national referral hospitals at the top followed by regional referral hospitals, general hospitals, health centres (HCs, which are divided into levels II, III, and IV), and village health teams (VHTs) at the bottom [5]- [7]. As part of a general restructuring, and in order to increase access, the Uganda government committed to construct HC-IIs and HC-IIIs at each sub-county and parish level, throughout the country [8]. Consequently, there has been an increase in the number of health facilities and a significant improvement in the proportion of households living within a walking distance to such health facilities from 49% in 2000 [9] to 72% in 2010 [6]. In addition, the institutionalisation of VHTs, an equivalent of community health workers, to mobilise communities to seek timely treatment is also intended to address barriers within communities where healthcare seeking decisions are made [7], [10]. A range of other targeted interventions such as immunisation and HIV-testing services have been implemented through community outreach programs.

Despite these reforms, transport is still a major barrier to accessing health facilities [11]. In addition, the 2006 and 2011 Uganda Demographic and Health Surveys found that distance to a health facility remains a serious challenge limiting access [12], [13]. Transport for healthcare

delivery in Uganda can occur at three levels: self-provision, in which individuals or communities mobilize transport for patients to health facilities; direct social provision especially at higher level health facilities where subsidised ambulance services exist; and direct-market provision where private providers such as bicycles, motorcycles, and motorised vehicles are hired [14]. Unlike urban areas where access to transport services is relatively easier because of a better road network and the availability of transport providers, transport to health facilities in most of the rural areas remains largely problematic and hardly any form of organised transport system exists.

Multiple studies have highlighted the effect of poor transport on healthcare access. In a study of male involvement in maternal healthcare in rural Uganda, poor road networks and rough terrain were cited as two of several reasons that delay women’s admission to a health facility [15]. A cross-sectional study of healthcare access for older people in Uganda identified mobility as one of several factors likely to limit their physical access to health facilities [16]. In a GPS-measured distance-to-clinic study in southwestern Uganda, distance was identified as an important barrier to sustained treatment and the main reason for late presentation for HIV clinic appointments [17]. Another study on adherence and access to highly active antiretroviral therapy (HAART) in southwestern Uganda found that costs associated with transport impeded adherence [18]. Similar challenges have been reported in Tanzania and Botswana [19]. Studies conducted elsewhere also indicate that where transport services are poor, access to healthcare facilities is a challenge [3].

In some countries, motorcycles have been deliberately promoted in the health sector to provide services to underserved communities. For example, in Malawi motorcycle ambulances have been used to overcome delays in obstetric care [4]. This study found that the use of motorcycle ambulances reduced median referral time for obstetric care by two to four-and-one-half hours [4]. In Taiwan, a study of pre-hospital care found that using a basic life support motorcycle was much faster in bringing a person to a hospital than an advanced life support ambulance during emergency medical situations [20]. Motorcycles arrived earlier than ambulances in 65% of the cases, reducing response time for emergencies compared to standard ambulances [20]. A quasi- experimental study on the effectiveness of a transport voucher system in eastern Uganda found an increase in antenatal care (ANC) attendance, health facility deliveries, and postnatal care services due to the use of locally contracted informal transporters [14], [21]. The increase in

100

ANC attendance and hospital deliveries was mainly attributed to ‘motorcyclists, motivated by a financial benefit accruing from transportation of pregnant and newly delivered mothers, [who] actively mobilised women and transported them to health facilities’ [21]. A sharp decline in ANC attendance and health facility deliveries occurred when the voucher system and accompanying financial incentive for transporters was stopped [21]. This raised sustainability concerns for ANC attendance and hospital deliveries in the absence of an incentive-led intervention.

There is hardly any research exploring how boda bodas act as a bridge between people in their villages and formal healthcare facilities without an external intervention providing an incentive to riders to provide such services. As found in southwestern Uganda, a GPS-measured distance- to-clinic study reported that 72% of the participants hired motorbikes to get to the clinic [17]. Boda boda riders tend to be cast in a negative light in relation to health, as they are perceived as a risk group for HIV and AIDS and sexually transmitted infections [22], [23]. In addition, the media portrays them as a traffic danger [24], [25] and as the leading cause of accidents and injuries [26], incidents that affect budget allocations for clinical services in major hospitals in Uganda (27).

This paper examines the contribution of boda bodas in improving access to formal healthcare facilities in Uganda. The boda boda industry is attractive to young men as a source of employment. Becoming a boda boda rider requires no formal training, as long as one has money to buy a motorcycle and learns how to ride it. Using evidence from a community where there has been no external intervention, such as a voucher transport system, we show that boda boda riders constitute a critical social resource for health promotion in rural communities. We use the terms ‘boda bodas’ to refer to the riders of motorcycle taxis and ‘boda boda riders’ for emphasis where necessary. The paper is based on ethnographic fieldwork conducted between 2012 and 2014 in Luwero District, central Uganda, and is part of the Developing Sustainable Community Health Resources in Uganda (CoHeRe) project.

Methods Study setting

101

Luwero District is located in central Uganda, approximately 60km from Kampala city along Kampala-Gulu road. The majority of the residents are native Baganda and speak Luganda. The district was a centre of guerrilla war activity between 1981 and 1986, which negatively impacted the health infrastructure. As part of the rehabilitation efforts, the president appointed a minister of state for Luwero Triangle in the office of the prime minister. A host of NGOs including Plan Uganda, AMREF, and Association Francois-Xavier Bagnoud (AFXB) have also implemented health support programmes in the district. The Ministry of Health partnered with AMREF to spearhead the identification, training, and support supervision of VHTs, and rehabilitation efforts have boosted the health infrastructure to now include more than 40 public health facilities, 22 NGO private non-profit facilities, and 3 private for-profit facilities [28].

The study community is a collection of relatively remote villages located approximately 5km off the Kampala-Gulu road. Community residents seek treatment at the nearest HC-IV, located about 5km away, while the closest general hospital is 20km away. Most adults own a mobile phone and a radio, which keep them informed and in touch with friends and relatives outside the community. Transport to and from this community is generally by walking, bicycle, and boda boda, the latter having gained popularity only in the last decade. Boda bodas link villages to the ‘outside world’ and help to transport people to and from health facilities despite the poorly maintained feeder roads.

Data collection Between 2012 and 2014, the first three authors lived in the community, participated in everyday activities, and held informal and formal interviews and focus group discussions (FGDs) with community members. Participant observation involved immersion in community activities, engaging in informal conversations with boda boda riders at their operating stages and while riding on their motorcycles, and after work when they played the popular board game ludo. These interactions were a form of ‘deep hanging around’ [29] with the boda bodas. Other community activities, such as church and religious gatherings, community meetings, socializing in bars, visiting the sick, and burials also constituted key opportunities for observations. As Spradely advises, interactions involved people in the situations as and when they unfolded in their natural settings [29]. Observations focused on what people said and did in relation to health

102

issues, how local leaders mobilised people whenever there was a health outreach program or a health need, and people’s response to health outreach efforts. Information gathered during these observations was used for discussion and reflection with the research team as fieldwork progressed.

Data gathered through observation were also discussed during interviews. A total of 91 in-depth interviews were held (with 36 men and 55 women), and 42 FGDs were conducted with randomly selected adults, aged 18 and above. These interviews and FGDs focused on a range of topics including health-related problems that the community encountered, and solidarity obligations at the community level. Within the interviews, an array of actors including boda boda riders, VHTs, and community leaders were mentioned as some of the resourceful people in health matters in the community.

To understand the influence of boda boda riders, five more interviews and three FGDs were held with boda boda riders, and four FGDs were held with community members on the role of boda bodas in enabling access to formal health facilities. The main topics in these interviews and FGDs included the mobilisation of transport services by boda bodas, rates charged by the riders, various motivations of boda boda riders, their relationship with other community members, and challenges that boda bodas and the community face. The insights presented in this paper are based mainly on these specific interviews and FGDs, but we also pay attention to the larger context described above and the analysis developed from the 91 interviews and 42 FGDs to illustrate the role of boda bodas in overcoming physical access barriers. Therefore, our conclusions are hinged on the study and observations of many other cases and situations in which boda bodas acted as a critical resource for access to healthcare.

Participants in the in-depth interviews and FGDs were selected using a simple random procedure. Interviews and FGDs generally lasted between 30 and 90 minutes. FGDs with community members usually involved between 10 and 12 participants. Key informant interviews were conducted with representatives from the local government including staff at Kasana HC-IV and the district health department, health assistants, and NGOs operating in the district. One staff member, the medical officer in-charge of Kasana HC-IV, was interviewed more than once on

103

general issues concerning health in the district and specifically on the role of motorcycle taxis. Triangulation of methods contributed to increased validity and reliability of the findings.

Data analysis Interviews and discussions were conducted and audio recorded in Luganda, and then transcribed in English. The transcription process, which started during data collection, was highly iterative and involved reflections on the data. Emerging issues were further investigated during subsequent interviews. After the data had been transcribed into a word processing document, it was password secured on computers and exported into NVivo10 software for further analysis. Codes were developed based on the emerging themes. When a theme was identified through reading the texts, a search query on a particular theme was performed and issues related to that theme identified. The emerging themes from coding and search queries were then discussed with colleagues and formed the basis for this paper.

Ethical clearance The study received ethical approval from the Ethical Advisory Board of the Amsterdam Institute for Social Science Research at the University of Amsterdam; the Institutional Review Board of Higher Degrees Research and Ethics Committee at the School of Public Health, Makerere University number IRB00011353; the Uganda National Council of Science and Technology; and the research desk of the Office of the Presidency in Uganda. Permission to carry out the research was also granted by the Chief Administrative Officer, Luwero District. Informed written consent was provided by participants before conducting formal interviews and FGDs. Oral permission for observations was always sought from the key organisers or leaders of community events. Only the members of the research team were able to access the data. Pseudonyms are used in this paper to protect the privacy of participants.

Results Boda bodas as community-based ‘ambulances’ From the community to the first contact with a health facility, there are ambulatory services. The health workers we interacted with mentioned that community members have to find their own means to get to the first health facility of contact.

104

At the community level there is no organised transport mechanism to ensure that when someone falls sick, he/she is transported to the health facility. This depends on the family’s capacity either to hire a boda boda or motor vehicle. There is no standard means. We have an ambulance at the centre but it’s not meant to pick up people from the communities.…The health system is set up in such a way that ambulances do not pick people from their homes. (Interview with official in charge of Kasana HC-IV, Luwero)

This quote illustrates the poor linkage between communities, where decisions to seek healthcare are made, and formal healthcare facilities. It also points to potential challenges in the referral system between the community and healthcare facilities. In some cases, VHTs are used to mobilize people in the community during health outreach efforts, and to advise people when to seek formal healthcare. However, VHTs also expressed that they are unable to address the problem of physical access.

Community members rely on informal transport, mainly boda bodas, to get from their homes to health facilities, though some walk or use bicycles. In an interview with the medical official in charge of Kasana HC-IV, he said:

Most of the patients we get here, especially emergencies like mothers coming to deliver, come on boda bodas. The riders are very good agents in enabling the community access healthcare. …On average, I would say about 40% of patients we see every day are brought by boda bodas.

The boda boda riders described their services as gap-fillers and community-based ambulances. In a FGD, one rider said:

Well, I would say we act as ambulances. For example, if a person gets in an accident, we right away come in to help. The fact is we do not have ambulances around. Therefore,

105

boda bodas transport people from this community to the health facilities. Even if it’s 2 a.m. or 4 a.m., we take these patients to health facilities.

The challenges experienced in accessing and utilizing facility-based ambulances also meant that boda bodas often step in to transport people referred from one health facility to another. We established that services at public health facilities are meant to be free, but those who use the ambulance are often required to fuel it due to the financial constraints the facility faces that affect their capacity to keep the ambulance operational. The Kasana medical officer told us:

In the case of referral, we encourage the caregivers to fuel the ambulance. Whereas the government pays the driver’s salary, as a facility the financial resources we have are not enough; we struggle to ensure that the ambulance is functional. So, we encourage the community to contribute fuel. …You know, sometimes you are in conflict with policies. The policy says all services should be free but when it comes to the resource envelope, there is nothing that shows that you can provide this service totally free.

Discussions with boda boda riders corroborated the views expressed by the officer above. In an FGD, a rider said:

At times, you may arrive at the health facility and you need to be transferred to another facility; the ambulance is available and you have the money to fuel the ambulance. However, you may find that the driver is not around. So you remain with two options: to look for that driver or to look out for someone else to take you to that facility. To solve the problem and work within our small budget, you opt for a boda boda.

As a form of community-based ambulance, boda bodas were reported as a critical force that offers emergency services, especially at night. A male community member described how he had called upon boda bodas in the middle of the night:

106

A few months ago, one of my family members fell sick at night. I called up my neighbour who has a boda boda. When he came, we supported the patient to sit on the boda boda and rode up to the hospital. About two weeks ago, that child you see there [points to one of the children playing in the compound] developed malaria at night. I mobilised a boda boda from my neighbour and we took him to the hospital at night. We were lucky God helped him and he got well.

Similarly, a boda boda rider recounted, during an FGD, how he used to wake up at night to transport a patient, and how this continued for more than six months. Failure to secure the services of the boda boda at night might be costly, especially for expectant mothers. In an FGD with VHT members, one person recounted the following story:

In our village, one time, a woman was about to deliver; the family tried to reach out to the boda boda but his phone was off. So they failed to take her to the facility and she delivered at home. When we [VHTs] reached the home in the morning, we found that she had lost a lot of blood. Her mother-in-law told us she did not have money [to hire a vehicle] to take her to the hospital, and the rider she knew was unavailable. So we called another boda boda and she was taken to the hospital. On reaching the hospital, we were told that she was about to die. She was admitted for two weeks.

Boda bodas offer flexible payment terms The overriding reason for the use of boda boda is that they were considered easily available, affordable, and offered flexible payment terms, for example, by transporting patients on credit. One boda boda rider explained this to us in an FGD:

What happens is that when I take you to the hospital, I don’t ask you for the money immediately. The reason is that I know since it is an emergency, you may not have [the money]. But when you return from the hospital, you pay me for the service I rendered.…But I give you ample time to enable you to get the money. Indeed, people also know that when such a thing happens, they are meant to pay us after they get better.

107

Some boda bodas also offered free services to some members of the community they deemed unable to afford the costs and to those with whom they have had a long-standing relationship. The ability of a boda boda rider to offer free or flexible payment depended on whether they owned a motorcycle. While some riders owned their own motorcycles, others rented them out at a daily rate of Ushs5,000–10,0001 per day. Community members pointed out that those who own a motorcycle find it easy to take people to hospitals for free or at most only ask for fuel. However, those who rent motorcycles have obligations to pay daily rates, which often leave them with limited flexibility to offer free services. However, it was not always clear who owned and who rented a motorcycle. Therefore, pressure to offer flexible payment was high for both motorcycle owners and those who rented.

In some cases, community members collected money and hired boda bodas for those who were judged as unable or not in position to afford the charges. James, a 35-year-old father of four, described how he was helped by friends when one of his daughters developed a fever from malaria:

When my daughter fell ill, I was not around. My wife went around in the community looking for transport to take her. When she was there talking to people, some community members said, ‘Oh! Your daughter is sick! Okay, let’s mobilize transport and take her.’ So, they collected money, hired a motorcycle that took her to the hospital where she spent about three days.

Boda bodas as caregivers and the role of community relations According to community members, the reason why boda boda riders feel compelled to offer flexible payments is mainly because they are part of the community and the help they offer is offered in response to community expectations. They are bound by the norms that hold the community together. The social benefits and interactions for help experienced by community members in their everyday lives also applied to services offered by boda boda riders. One rider explained this during an FGD:

1 This is equivalent to US$1.36–US$2.72 based on the Bank of Uganda foreign exchange rate, on 21 September 2015, of US$1=Ushs3,666. 108

You know, the people we offer services to are normally people we know; they are fellow residents and in most cases they stay with our families and act like guardians when we go to work. … My child can go to his home and ask for water or food and they offer it to him or her when I am away. So, giving a free lift to such a person when they are sick is like a sign of appreciation to the person. Indeed, after giving such a person a free lift, he will develop even more love for my children and continue helping them and treating them well.

Given these pre-existing relations, it was not surprising to learn that sometimes when boda boda riders transport patients to the health centres, they also act as caregivers. Take the example of Dick, a 38-year-old man whose child developed malaria while he away at his work place. He called his neighbour, a boda boda rider, to take the child to the hospital. The boda boda rider took the child to the hospital with the mother and gave them Ushs7,000 (US$1.9) to cover the hospital costs. The boda boda rider also urged other community members to visit the child at the hospital, and he brought the family food every morning of the three days the child was admitted. According to boda boda riders, health workers sometimes ‘mistakenly’ perceive boda boda riders as the primary caregivers of the patients they transport. This was particularly common with expectant women, as boda bodas riders often took on the responsibilities of their husbands. One rider stated: By the way, when we take those women [to the health facilities], the health workers think that we are their husbands and they start asking us to buy some of the things in the delivery kit.

Being part of the community means that boda boda riders are compelled to behave and relate as the community expects, especially when it comes to helping others, and they do so as not to risk the wrath of the community. Community members insisted that it is almost impossible for a boda boda rider to refuse to offer help to a sick person even when that person does not have money. One woman in an FGD told us:

109

[Sure], we can decide to hate that boda boda rider. We can discriminate against him [if he misbehaves]; even if I have the money, I would not sit on his motorcycle, even if it is the only one around. By the way, even if I am not sick and I ask you to take me, and you refuse because I have little money, people would stop using your motorcycle.

The riders we talked to also felt this form of pressure; therefore, they are often ‘forced’ to prioritise the health of people in the community. A boda boda rider confirmed the above woman’s sentiments:

Haaah, my dear, even a sick person will hate you for that [not taking him/her to the hospital for free], and if that person dies and people know that you refused to take him, you cannot even attend the burial – the whole community will hate you.

Perceptions of boda bodas We found that for some people it is costly to hire a boda boda, especially if one is not a friend or a neighbour to a boda boda rider if assistance is needed at night. One community member explained this in an interview:

One time, it happened to me deep in the night. The first boda boda rider I rang asked me for UShs100,000 (US$27.27) to transport me and the sick child to the hospital. I rang a second one who then asked me for UShs20,000 (US$5.45). It was at night and they always have reasons, including fear of being attacked by robbers along the way or risking their lives.

The safety of boda boda riders was mentioned by both the riders and community members as a serious concern. They are often targeted and killed by thieves. Stories of boda boda riders dying at the hands of thieves were common, as were reports of riders aiding crime and as sources of insecurity [30]. Such stories appear to have raised suspicions among the community members about the riders who work at night. Community members who argued that boda boda riders are likely to be thieves attributed the problem to lack of regulation of the industry, as anyone can

110

join and leave at will. In an FGD with male community members, the following opinions were expressed:

Respondent 1: It is the only business within this country where someone does not need qualifications. If someone buys his own motorcycle, whether through selling his land or otherwise, he can start riding.

Respondent 8: The government has not done much to set up requirements one has to fulfil before he gets into this business. It is not known what someone must have to join the business – even a thief who stole chickens and sold them can buy a motorcycle and join the business. Someone who rapes women can also easily get into the business if he gets a motorcycle. If requirements such as having a driving permit or passing a driving test were in place, then this business would have the right people.

Such suspicions might appear to contradict earlier statements that boda boda riders are known in the community. To address the problem of knowing who is in the business, the boda bodas have formed an association with a known leadership structure. The members are registered with district authorities. However, from statements made in interviews, it appeared that the association merely was a tool to ensure tax compliance, rather than focused on regulating the business and ensuring safety. Calls for ensuring the safety of boda bodas have been made more in major towns especially Kampala city than in rural areas.

Boda boda riders have also been accused of spreading diseases, including HIV and AIDS. When Ebola broke out in Luwero District in 2012, the index case was a boda boda rider whom some community members believed to have caught it from a passenger and then spread it to his family members.

Boda boda riders have also been accused of luring young girls into sex and abandoning them after impregnating them, perceptions that conflict with their role in health promotion. In an FGD with men from the community, one participant said:

111

A boda boda rider can easily impregnate girls. He does not stop with one girl. You always hear that the boda boda guy has impregnated about 10 girls. There is one I know who has impregnated four girls; he lives around here.

An interesting discussion arose during an FGD with women: some participants argued that boda boda riders are more likely than other men in the community to lure young girls into sex and spread diseases. In response, others claimed that sexual intercourse between old men and young girls (usually teenage girls) is a common practice within the community and it would be unfair to point fingers at the boda boda riders only. This discussion within the FGD took place in a context of a larger discussion in the community where community members decried the high levels of teenage pregnancy, making the discussion even more appealing to the participants in the FGD. Some participants countered that young girls themselves lure men into sex. Some health workers also argued that boda boda riders can embody both negative and positive images, and pointed out that negative news spreads faster and thus overshadows their positive contributions.

At times there is a tendency among some sections of the community and policy makers to perceive boda bodas negatively or as a problem that needs to be fixed. Boda boda riders are targeted for HIV- and STI-prevention programs because they are perceived as a risk and an at- risk category. During fieldwork we learnt of circumcision and HIV programs that specifically targeted boda boda riders as the most at-risk people in the community.

Boda boda riders are also perceived by some as lacking in skills to adequately help patients when transporting them to health facilities. In our interview with the health inspector for Luwero District, he said:

I have my reservations about boda bodas because the riders are not skilled in carrying patients; they are not skilled in offering first aid. So, even if they could take that person to hospital, the patient could easily die along the way because he did not receive first aid. The basic training is missing. The issue of knowing how to carry a patient depending on that patient’s condition is important.

112

What the inspector called ‘reservations’ might be instead seen as opportunities for improving and equipping boda bodas with basic skills in handling patients they transport.

Discussion Boda bodas’ potential as an informal and sustainable resource for overcoming physical access barriers to formal healthcare is not well documented in the academic literature. We have documented how boda bodas support community members to overcome challenges of geographical inaccessibility and poor road infrastructure in rural areas. We found that boda bodas were popular in this rural area because they were convenient, served as caregivers, offered credit, and sometimes waived transport charges. While the riders are known in the community and accessed on cell phones in case of emergencies, a story of a woman who could not access the boda boda at night and cases of people voicing concerns about using the boda bodas at night because they are not regulated suggest that their use should be made more reliable for the members.

The flexible payments offered by boda bodas and occasional fee waivers enable people to cope with transport costs in a low-income setting where the majority depend on seasonal subsistence farming. Similar flexibility in payments among traditional healers has been reported [32]. In contrast, formal healthcare facilities rarely offer flexible payment [31], [32]. Where such flexible payments have been reported, they are usually related to existing social relations with the private providers of such services [33]. Although boda bodas are ostensibly a profit-driven private enterprise, they tended to work on a voluntary basis when it came to providing their services.

In addition, our finding of the importance of flexible payments supports certain arguments about models of healthcare delivery, especially the notion that healthcare delivery systems, even those that are for-profit, should not have financial gain as their sole motive [1]. This finding and the fact that boda bodas value community solidarity recall the distinction between intrinsic and extrinsic motivations. Our results suggest that the boda bodas are as much influenced by intrinsic motivations, for example, a sense of moral and community duty, as they are by extrinsic motivations, such as earning an income or avoiding condemnation [34].

113

Boda boda riders live in communities where they are known; they have families, children, spouses, and relatives. Therefore, they encounter social pressures to assist. The community may impose sanctions on those who fail to offer flexible payment for their services. These social pressures and sanctions can be equated to ‘voice’, the assumption that people can exit or withdraw their services if they are unhappy with service providers [35]. Whereas this was true for boda boda riders and certainly relevant, it may not explain why some people go further and chastise riders who refuse to offer flexible payment or refuse to transport a sick person who lacks money. It appears that the boda bodas are also subject to existing community norms. When they ‘refuse’ to take a person for free or offer credit, community members may ‘mark’ them as badly behaved and withdraw other support, including support typically extended in everyday interactions or when crises such as death occur. Therefore, their services are intertwined with community life, and related to connections and solidarity with community members. The importance of connections and solidarity has been emphasised by several scholars in other settings [31], [32], [36]-[38].

While programs and resources the world over have been invested in increasing community participation through such interventions as training laypersons as community health workers [34], there are limited examples of deliberate attempts to harness transport resources. In Uganda, emphasis appears to have been placed on the construction of local health facilities to reduce the distance between communities and formal healthcare. With exception of a few targeted interventions like in eastern Uganda [14], [21], as far as we know, there is no program that has sought to tap into boda bodas as a critical resource in the country. Even when studies identify transport as a key problem, they still recommend unsustainable costly measures, advising that ‘treatment programs should consider subsidizing transport’ and that ‘transport costs can also be reduced by setting up a more extensive network of facilities’ [19]. Another study has called for a review of the threshold established by the Ministry of Health, now set at 5km to health facilities in rural areas [39]. Results from studies in eastern Uganda have already shown that sustaining such efforts is a challenge [14], [21].

Concerns about the sustainability of incentive-led transport voucher programs reinforces more general concerns about externally initiated interventions that often affect not just innovation but

114

also change existing reward and motivation dynamics. Our results show that boda bodas provide a useful transport resource with no external support, and that community members access and utilise this resource because it is able to meet local conditions by offering flexible payments. This finding implies that for informal transport systems to be sustained there is no need for an incentive-based external intervention. The value of boda bodas as a resource lies in what communities are already doing, such as the activation of social networks that predate incentive- based interventions. It appears more realistic to concentrate on identifying what people already do to overcome barriers to access to health facilities, and to strengthen these means for sustainable results. The starting point is for example to appreciate that even in the difficult of circumstances community members are often willing to use available means to travel long distances in search of better healthcare [40]. The perception that boda bodas lack patient-care skills or that they are dangerous could be more of a signal of authorities’ failure to harness their potential than an actual problem. For example, proposals to increase the safety of boda boda riders and their passengers, such as use of helmets and reflective jackets, trainings on road safety and patient-handling, and certification, have largely concentrated on urban areas [24] with little focus on the riders in rural areas. Where boda boda riders have been involved, they have been looked at as a risky group that has to be addressed [22] rather than as a resource to be harnessed. The problems the boda boda riders face generally signal a failure on the part of government to mainstream this critical resource. Our findings suggest that if approached as a resource and regulated, the boda boda transport system could potentially be a sustainable solution to the challenges of transport as a barrier to accessing formal healthcare.

Conclusion Concentrating on transport costs as a limitation masks the great resources that already exist in the community, which people utilise to overcome barriers related to physical access. Boda bodas are a community resource that facilitates critical linkages to the health system. Interventions that have tended to focus on correcting the negatives often fail to harness and build on the resources the communities already use to access health facilities. We recommend that the Ministry of Health should champion the training of boda boda riders in basic patient care without introducing any incentives to riders. To actualise this, there is need for a policy environment that

115

acknowledges the role of boda bodas, and enhances their strengths while minimizing the challenges they face including security concerns. This would not only contribute to the attainment of greater health outcomes, by addressing delays and increasing access to formal healthcare, but also could become a springboard for meeting the recently launched sustainable development goal number 3 which seeks to promote healthy lives and wellbeing for all people of all ages [41].

Abbreviations

ANC: Antenatal care FGD: Focus group discussion HC: Health centre NGO: Nongovernmental organisation VHT: Village health team

Competing interests The authors declare that they have no competing interests.

Authors’ contributions All authors approve the submitted manuscript. Laban Kashaija Musinguzi collected, analysed, and interpreted the data and wrote the manuscript. Jude T. Rwemisis, Emmanueil Benoni Turinawe, and Deni Muhangi contributed to the acquisition, analysis, and interpretation of the data as well as to the writing of the manuscript. Marije De Groot contributed to the analysis and writing of the manuscript. Daniel De Vries supervised the collection and analysis of the data, and contributed to the content of the manuscript. David Mafigiri designed and supervised the project. Robert Pool designed and supervised the project and supervised data collection, and participated in data analysis and in the writing of the manuscript.

Acknowledgments This work is part of the Developing Sustainable Community Health Resources in Poor Settings in Uganda (CoHeRe) research program, which is financed by WOTRO, Science for Global

116

Development, under the Netherlands Organisation for Scientific Research (NWO), grant number W07.45.105.00. The authors also acknowledge all of the study’s participants and the support from the Center for Social Science Research on AIDS (CeSSRA), School of Social Sciences, Makerere University Kampala.

References

1. Bishai DM, Shah NM, Walker DG, Brieger WR, Peters DH. A social franchising to improve quality and access in private healthcare in developing countries. Harvard Health Policy Review. 2008;9(1):184-97.

2. Thaddeus S, Maine D. Too far to walk: maternal mortality in context. Soc Sci Med. 1994;38(8):1091-110.

3. Fiagbe P, Asamoah D, Oduro FT. Assessing the role of transport in the achievement of maternal mortality reduction in Ghana. International Journal of Business and Management. 2012;7(5):p256.

4. Hofman JJ, Dzimadzi C, Lungu K, Ratsma EY, Hussein J. Motorcycle ambulances for referral of obstetric emergencies in rural Malawi: Do they reduce delay and what do they cost? International Journal of Gynecology & Obstetrics. 2008;102(2):191-7.

5. Jeppsson A, Okuonzi SA. Vertical or holistic decentralization of the health sector? Experiences from Zambia and Uganda. Int J Health Plann Manage. 2000;15(4):273-89.

6. Ministry of Health (MoH). Health sector strategic & investment plan: promoting people’s health to enhance socio-economic development 2010/11–2014/15. Kampala: Ministry of Health Uganda; 2010.

7. MoH. Village health team: strategy and operational guidelines. Kampala: Ministry of Health; 2010.

8. MoH. The second national health policy: promoting people's health to enhance socio- economic development. Kampala: Ministry of Health; 2010.

9. MoH. Health sector strategic plan 2000/01–2004/05. Kampala: Ministry of Health; 2000.

10. Bakeera SK., Wamala SP., Galea S, State A, Peterson S, Pariyo GW. Community perceptions and factors influencing utilization of health services in Uganda. Int J Equity Health. 2009;8(25).

11. Lutwama GW, Roos JH, Dolamo BL. A descriptive study on health workforce performance after decentralisation of health services in Uganda. Hum Resour Health. 2012;10(1):41. 117

12. Uganda Bureau of Statistics (UBOS), Macro International Inc. Uganda demographic and health survey 2006. Calverton, MD: UBOS and Macro International Inc; 2007.

13. Uganda Bureau of Statistics (UBOS), ICF International Inc. Uganda demographic and health survey 2011. Kampala, Uganda: UBOS and Calverton, MD: ICF International Inc.; 2012.

14. Pariyo GW, Mayora C, Okui O, Ssengooba F, Peters DH, Serwadda D, et al. Exploring new health markets: experiences from informal providers of transport for maternal health services in Eastern Uganda. BMC Int Health Hum Rights. 2011;11 Suppl 1:S10,698X-11-S1-S10.

15. Kakaire O, Kaye DK, Osinde MO. Male involvement in birth preparedness and complication readiness for emergency obstetric referrals in rural Uganda. Reproductive health. 2011;8(12):1-7.

16. Wandera S, Ojiambo., Kwagala B, Ntozi J. Determinants of access to healthcare by older persons in Uganda: a cross-sectional study. Int J Equity Health. 2015;14(26):1-10.

17. Siedner MJ, Lankowski A, Tsai AC, Muzoora C, Martin JN, Hunt PW, et al. GPS-measured distance to clinic, but not self-reported transportation factors, are associated with missed HIV clinic visits in rural Uganda. AIDS. 2013;27(9):1503-8.

18. Tuller DM, Bangsberg DR, Senkungu J, Ware NC, Emenyonu N, Weiser SD. Transportation costs impede sustained adherence and access to HAART in a clinic population in southwestern Uganda: a qualitative study. AIDS and Behavior. 2010;14(4):778-84.

19. Hardon AP, Akurut D, Comoro C, Ekezie C, Irunde HF, Gerrits T, et al. Hunger, waiting time and transport costs: time to confront challenges to ART adherence in Africa. AIDS Care. 2007;19(5):658-65.

20. Lin C, Chang H, Shyu K, Liu C, Lin C, Hung C, et al. A method to reduce response times in prehospital care: the motorcycle experience. Am J Emerg Med. 1998;16(7):711-3.

21. Ekirapa-Kiracho E, Waiswa P, Rahman MH, Makumbi F, Kiwanuka N, Okui O, et al. Increasing access to institutional deliveries using demand and supply side incentives: early results from a quasi-experimental study. BMC Int Health Hum Rights. 2011;11 Suppl 1:S11,698X-11-S1-S11.

22. Nyanzi S, Nyanzi B, Kalina B. Contemporary myths, sexuality misconceptions, information sources, and risk perceptions of Bodabodamen in Southwest Uganda. Sex Roles. 2005;52(1- 2):111-9.

23. Nyanzi S, Nyanzi B, Kalina B, Pool R. Mobility, sexual networks and exchange among bodabodamen in southwest Uganda. Cult Health Sex. 2004;6(3):239-54.

24. Anena H and Natabaalo G. Boda boda menace: where do solutions lie? Part two. Daily Monitor Newspaper. 2015, June 27. Retrieved from

118

http://www.monitor.co.ug/SpecialReports/Boda-boda-menace-solutions-lie/-/688342/2766000/- /1392hwkz/-/index.html.

25. Ssegawa M. The menace that is bodas. Daily Monitor Newspaper. 2015, August 26. Retrieved from http://www.monitor.co.ug/artsculture/Reviews/The-menace-that-is-bodas/- /691232/2846126/-/acfbsdz/-/index.html.

26. Naddumba E. A Cross-sectional retrospective study of boda boda injuries at Mulago Hospital In Kampala-Uganda. East and Central African Journal of Surgery,2007, 9(1). Retrieved from https://tspace.library.utoronto.ca/handle/1807/52625.

27. Kigera J, Nguku L, Naddumba E. The impact of bodaboda motor crashes on the budget for clinical services at Mulago Hospital, Kampala. East and Central African Journal of Surgery 2010, 15(1), 57-61.

28. MoH. Health Facilities inventory July 2012: master health facilities inventory. Kampala: Ministry of Health Uganda; 2012.

29. Spradley JP. Participant observation. Belmont, CA: Wadsworth, Cengage Learning; 1980.

30. Larubi P, Robert. Kayihura vows to regulate bodabodas. Chimpreports. 2015, April 1. Retrieved from http://www.chimpreports.com/kayihura-vows-to-regulate-bodabodas/.

31. Muela SH. Community understanding of malaria, and treatment-seeking behaviour, in a holoendemic area of southeastern Tanzania. Diss. phil.-nat. Univ. Basel, 2000.

32. Muela SH, Mushi AK, Muela RJ. The paradox of the cost and affordability of traditional and government health services in Tanzania. Health policy and planning. 2000;15(3):296-302.

33. Rutebemberwa E, Pariyo G, Peterson S, Tomson G, Kallander K. Utilization of public or private healthcare providers by febrile children after user fee removal in Uganda. Malar J. 2009 Mar 14;8:45,2875-8-45.

34. Maes K. Community health workers and social change: an introduction. Ann Anthropol Pract. 2015;39(1):1-15.

35. Sharma B. Voice, accountability, and civic engagement: a conceptual overview. Overseas Development Institute. Commissioned by Oslo Governance Centre, Bureau for Development Policy. Oslo: UNDP; 2008.

36. Ayé M, Champagne F, Contandriopoulos A. Economic role of solidarity and social capital in accessing modern healthcare services in the Ivory Coast. Soc Sci Med. 2002;55(11):1929-46.

37. Meinert L, Kajubi P, Whyte S.R. Connections. In: Whyte SR, editor. Second chances: surviving AIDS in Uganda. Durham, NC: Duke University Press; 2014. p. 34-46.

119

38. Meinert L. Hopes in friction: schooling, health, and everyday life in Uganda. Charlotte NC: Information Age Publishing Inc; 2009.

39. Ettarh R, Galiwango E, Rutebemberwa E, Pariyo G, Peterson S. Spatial analysis of determinants of choice of treatment provider for fever in under-five children in Iganga, Uganda. Health Place. 2011;17(1):320-6.

40. Ndyomugyenyi R, Neema S, Magnussen P. The use of formal and informal services for antenatal care and malaria treatment in rural Uganda. Health policy and planning. 1998;13(1):94- 102.

41. United Nations. Transforming our world: the 2030 Agenda for Sustainable Development A/RES/70/1, Seventieth session. 2015. retrieved from: http://www.un.org/ga/search/view_doc.asp?symbol=A/RES/70/1&Lang=E.

120

Paper IV The role of social networks in savings groups: insights from village savings and loan associations in Luwero, Uganda

Author: Laban Kashaija Musinguzi

This paper was accepted for publication in Community Development Journal. doi:10.1093/cdj/bsv050.

121

Abstract Studies of village savings and loan association (VSLAs) programmes in several African countries portray these initiatives as spaces that increase financial access for the poor, improve livelihoods, and provide members with social capital. Little is known, however, about their impact beyond increasing financial access. This paper shows that the benefits that accrue (or do not) from membership in VSLAs are mediated through networks of friendships and other social relations that predate the introduction of VSLAs. Based on ethnographic research on VSLAs conducted between 2012 and 2014 in Luwero District, Uganda, this paper examines women’s experiences in VSLAs, how social networks influence their decision to join a VSLA, and how VSLAs provide women an opportunity to exercise agency through utilizing their social networks in their community. In this way, they are able to challenge structural barriers to financial autonomy and control at the household level. This research shows that female participants utilize two kinds of networks in VSLA spaces: ‘silence-in’ and ‘silence- through’ networks. The findings underscore the influence of friendships and family relations in shaping the impact of externally initiated micro- level programmes such as VSLAs beyond financial inclusion, livelihood, and poverty reduction. The decision to join a VSLA, the desire for financial autonomy, the struggle against power dynamics, and unintended consequences are all negotiated within the VSLAs space through social networks.

122

Introduction

The microfinance movement has long emphasized women’s access to credit. Some scholars have argued, in response, that it is not just access that counts but also control over funds (Kabeer, 1997; Kantor, 2003). Formal microfinance institutions hardly reach the poorest, the majority of whom are women, in most African countries (Lønborg and Rasmussen, 2014; Lowicki-Zucca et al., 2014). In order to address the limitations of mainstream microfinance, CARE International launched the village savings and loan association (VSLAs) model in 1991 in Niger (Allen, 2006). Organizations like Oxfam America, Freedom from Hunger, Plan International, and other international and local organizations have made such savings groups a central part of their development strategy, expanding access in over sixty-five countries world-wide with over ten million members (Ashe and Neilan, 2014, p. 9). Women constitute close to 70 percent of VSLA membership worldwide (Allen and Panetta, 2010; CARE, 2011).

AVSLA is ‘a time-bound accumulating savings and credit association’ in which a group of people contribute to and receive disbursements from a group fund (Anyango et al., 2007, p. 12). Many studies have assessed the targeting, participation, and financial inclusion of VSLA members (Hendricks and Chidiac, 2011; Lønborg and Rasmussen, 2014; Lowicki-Zucca et al., 2014); whether VSLAs help members build assets, reduce poverty, and improve livelihoods (Allen and Hobane, 2004; Adam et al., 2014); and assessed their performance, sustainability, and replicability (Allen and Hobane, 2004). In Uganda, a few studies have focused on microfinance and women’s empowerment (Lakwo, 2006), savings groups as a vulnerability-reduction strategy (Lowicki-Zucca et al., 2014), and VSLAs as an empowerment intervention to reduce HIV and AIDS through education (Ssewamala et al., 2010). Some studies (Lønborg and Rasmussen, 2014) suggest that VSLAs, like mainstream microfinance, do not necessarily include the poorest of the poor. A randomized control study of VSLAs in Ghana, Malawi, and Uganda found little evidence to suggest that participation in VSLAs improves participants’ involvement in community activities even though intra-household decision-making power might improve (Karlan et al., 2012).

While VSLAs have been praised as a ‘catalyst for enhanced social capital, improved gender relations, women’s leadership, and community social and economic development’ (Allen and Panetta, 2010, p. 2), few analyses document how members actually use the social capital inherent 123

in VSLAs or how they use VSLAs to challenge existing structural and cultural barriers and expand their social networks. This is partly because developing qualitative indicators to measure non-financial benefits is difficult (Vermaak, 2001; van Rooyen, Stewart and de Wet, 2012); analyses of women’s participation in development have been pre-occupied with numbers than qualitative assessment (Goetz and Gufta, 1996). In addition, significant variations within communities make the adaptation of VSLAs dependent on context (Bouman, 1995). However, there is increasing emphasis on documenting the outcomes of development interventions related to social relationships and networks (Dowla, 2006; Ibargu¨en-Tinley, 2014). In line with this focus, I argue that women’s participation in externally initiated interventions such as VSLAs is shaped by their existing social relationships. The aim of this paper is to show that the benefits– and unintended consequences – of participation in VSLAs are found in the networks of friendships and social relations that predate the VSLA. These social networks influence women’s decisions to join VSLAs and to challenge structural barriers, and they enable women to expand and maintain friendship networks. By drawing on three case studies of women in VSLAs, this paper enhances our understanding of the value of social networks in relation to development interventions.

Social networks, social capital debates

I use the concept of social networks as described in the literature on social capital. Putnam (1995, p. 16) defines social capital as ‘dense networks of social interaction’ generated through civic participation. These social networks can be ‘formal’, emerging as individuals deliberately choose to join groups, or ‘informal’, evolving ‘naturally and casually’ among friends (Häuberer, 2014, p. 3). They can also be categorized as strong or weak ties, and as homogenous or heterogeneous networks (Cattell, 2001). Strong ties are described as ‘bonding’ social capital, the close networks that develop between relatives, friends, families, and neighbours (Woolcock and Narayan, 2000). Weak ties are depicted as ‘bridging’ social capital: the value extends into distant networks (Putnam, 2000; Woolcock and Narayan, 2000; Perkins, Hughey and Speer, 2002). Social resources inhere in such networks, and are linked to resource exchange and membership in groups (Bourdieu, 1986; Lin, 2001). I consider social networks as social resources, which include intangible aspects of everyday interactions, whether symbolic or concrete, such as seeking assistance or advice (Häuberer,, 2014). Access to such resources is influenced by the

124

quality of a person’s networks; these may be expanded by group membership. Interpersonal bonds and relationships constitute critical resources that offer the possibility of strong cooperation and collective action for individuals (Ibargu¨en-Tinley, 2014, p. 231). As Titeca and Vervisch (2008) point out, groups are a building block of social capital.

Power and inequality influence the distribution of social resources in social networks (Bourdieu, 1986). Krishna and Goldey (2009, p. 189) write: ‘the informal network of cooperation revolves around the powerful people who often create difficulty or refuse to cooperate with those who go against them’. They further claim that networks tend to enhance patronage relations. Development initiatives can become vulnerable to elite capture, where powerful people in the community use their positions to leverage resources for personal gain (Platteau and Gaspart, 2003). In this way, social networks may be constraining (Cleaver, 2005) especially if they are seen as being influenced from outside the community (Krishna and Goldey, 2009).

Despite such criticism, many studies have concluded that social networks of family, friends, and community members are beneficial (Campbell, 2001), especially for under-resourced communities (Narayan and Pritchett 1999; Woolcock and Narayan, 2000). In Bangladesh, women’s informal networks were found to be important drivers of contraception use (Gayen and Raeside, 2010). In Rwanda, a study of savings and credit associations found that networks of friends, harnessed through trusting face-to-face relations, reduced members’ feelings of isolation (Benda, 2012). A study of VSLAs and livelihood in Ghana also found that women’s social networks positively impacted child survival (Adams, Madhavan and Simon, 2002).

The study setting

Ethnographic fieldwork was conducted between 2012 and 2014 in a rural community in Luwero District, central Uganda. Luwero District is located approximately 60 km north of the capital, Kampala. The term ‘community’ here refers to a collection of villages from which the participants were drawn. Although many regard themselves as Baganda (native inhabitants of the district), the influence of migration and intermarriage has led to greater ethnic diversity. Estimates put the district’s population at 405,900 persons (Luwero District Local Government, 2009). Reconstruction following the 1981 – 1986 war, which ushered the current government to power, saw an influx of nongovernmental organizations (NGOs). Plan Uganda and Community

125

Vision, two local NGOs based in Luwero, introduced VSLAs to Luwero in 2006/ 2007 (Plan Uganda, 2011).

Methods

Data collection followed an iterative, triangulated process. The main method was participant observation, which included attending VSLA meetings, and religious gatherings and burials; and playing games with children and board games with youths. Following the suggestions of Spradley (1980, p. 54), I engaged in activities as and when they unfolded. I joined a VSLA group, Bumu, and participated in weekly meetings from January to December 2013. As a member, I saved money, and I documented members’ interactions during meetings, the type of information that was shared, and how meetings were conducted. Community conflicts were observed, for example, when VSLA members engaged in argumentative exchanges while distributing VSLA savings. I also observed discussions between VSLA members and non- members in everyday interactions. I kept a journal of group meetings, daily observations, and interactions with people.

Information gathered through observations was then explored in a total of ninety-one interviews (thirty-six men, fifty-five women) and forty-two focus group discussions (FGDs), involving ten to twelve participants each (twenty-three with women, fifteen with men, and four mixed). Interviews and FGDs lasted from between 30 and 90 min. Participants were selected using qualitative sampling techniques (purposive, snowball, and theoretical sampling). Participants were both members and non-members of VSLAs, adult women and men (aged eighteen and above), local leaders, NGO staff, and district and local government officials. Data collection generally followed an iterative approach beginning with broad topics as entry points, followed by a back-and-forth investigation of themes (Glaser, 1978: 45). The flow of questions asked during interviews was flexible and followed themes as they emerged from the data. This iterative approach allowed for ongoing analysis, as data from interviews and FGDs were further explored in subsequent interviews and discussions. I held subsequent interviews with three women – Jane, Sanyu, and Jalia – whose experiences I describe in detail in this paper. While these case studies illustrate the central argument, my conclusions about the role of social networks are based on many additional cases and situations. However, given that this was a micro-level study, it may

126

not be easily generalizable beyond communities with similar characteristics. Findings should be interpreted with this limitation in mind.

Formal interviews were conducted in the respondents’ preferred language, mainly Luganda. All interviews and FGDs were audio recorded and transcribed into English. The transcription process was highly iterative, allowing identification of interesting issues and emerging themes. After the data were transcribed, they were imported into NVivo software for further analysis. Following Braun and Clarke (2006), coding was done for the entire dataset, giving attention to each data item in order to identify patterns.

The study received ethical approval from the Amsterdam Institute for Social Science Research at the University of Amsterdam, the School of Public Health at Makerere University, and the Uganda National Council of Science and Technology. Informed written consent was received from participants before conducting interviews and FGDs. Oral permission for observation was always obtained from the key organizers during communal events. Pseudonyms are used to protect the privacy of participants and the VSLA groups.

Using personal networks to join VSLAs

I used my personal networks to join a VSLA. In January 2013, I contacted my friend, Robert, a group leader in one of the ten VSLAs in the community at that time. Asking whether it would be possible for me to become a member, Robert replied, ‘Oh yes .. . we have had some three people leave the group this year – you can come and replace them’. I joined Robert’s group, called Bumu, which included twenty-five people (three men and twenty-two women).

I had earlier learnt from conversations with people in the community and officials from Plan Uganda that when VSLAs were fi t initiated, Plan Uganda and Community Vision approached local village council leaders, and asked them to hold community meetings at which the VSLA approach was explained. The importance of group leaders in micro-enterprise development for women has been emphasized elsewhere (Ibargu¨en-Tinley, 2014). In the VSLA methodology, all groups begin by selecting a committee, including a chairperson, treasurer, and secretary. Members build a fund through a minimum weekly savings of UGX1,000 (approximately USD0.33) (based on a dollar-shilling exchange rate as of 24 April 2015 of 1USD ¼ 2956.28 UGX) and a maximum of UGX5000 (USD1.69). 127

At every meeting I attended, the chairperson read out each member’s name, received their money, and noted the contribution in their savings book with a stamp; contributions were tracked for one year. A cross was marked in the book for members who saved less than the minimum. The total number of stamps consequently determined how much money each member received at the time of distribution. Requests for loans were made to the chairperson, and had a 10 percent monthly interest rate. All other VSLAs in the community operated by similar procedures. The money saved and interest earned was shared in December 2013; the group then reconstituted in early 2014. Money was kept in a metallic box with three padlocks whose keys were kept by three different members. This was meant as a safe-guard against robbery, which had happened to another VSLA in the community, and as a precaution against any one member betraying the trust of the others in the group. In this way, both issues of trust and far-sightedness informed VSLA practice.

VSLA membership and shifting network dynamics

Through my membership in Bumu, I met multiple female VSLA participants, one of whom was Jane, aged thirty. We frequently conversed, and she told me that she felt her marriage would not last long. Her husband, a youth leader in the village, was not a member of any VSLA. Although he opened a generator-operated hair salon for her, she had no control over the income generated; it is widely documented that women have little to no control over household income (Goetz and Gufta, 1996; Mayoux, 2010). Jane yearned for some degree of financial independence, which led her to join Bumu in 2008:

I joined the savings group because whenever I want the money, no one will ask me what I am going to use it for. They [the group] can only tell me that ‘Maybe today Jane will not take money because someone else wants it’, or when the money is insufficient, but not ask me what I am going to use it for. There is no one putting pressure on me, because I know it is my personal savings.

Jane described how tired she had become of begging money from her husband, and stated that saving through the VSLA offered her an alternative and financial autonomy. Although she did not belong to any other groups like burial associations, Jane said she no longer begged her husband or anyone else for money.

128

Jane’s relationship with her husband turned sour when, according to her, he confessed to having an extramarital affair and fathering a child in the process; this culminated in their separation at the end of 2013. ‘It was too much, I could not stand him’, she told me, when I visited her at a rented one-roomed house where she relocated with her two children. Two months later, she reunited with her husband after, according to her, he became a born-again Christian and received counselling from her friends. In September 2014, Jane separated from her husband again, and used her own money to rent a house where she lived with her children. ‘What do I need from a man? I can live on my own’, she told me, using a conceited tone to emphasize her independence. Jane explained that she had also developed a network of friends through Bumu; she said she was comfortable with her life and would only go back to her husband on her own terms.

Jane’s story not only exemplifies the search for financial autonomy, but also shows how VSLAs constitute a different form of network; rather than ‘traditional’ networks of friends and family, VSLAs foster ‘formally constituted’ networks that arise out of a deliberate attempt to join groups (Häuberer, 2014). A study in Tanzania on the impact of savings groups revealed a similar shift in the use of such groups, away from relying on the home, friends, and banks for savings and borrowing purposes (Allen, 2009). This research in Luwero district shows, further, that the presence of VSLAs affects pre-existing social networks.

VSLA membership and the use of silence-in networks

Jalia, a twenty-nine-year-old woman, was born and lived in Busoga region, eastern Uganda, for most of her life until 2010 when she got married in Luwero. Although Basoga (people from Busoga region) share certain cultural norms with Baganda, it was not easy for Jalia to fit in a new community. According to her, from the time she got married, she and her husband had many conflicts over the welfare of her child from a previous relationship. Village rumours that her husband’s family practiced witchcraft did not help. She initially excluded herself from most social activities in the community, save for obligatory events like burials. Her mother-in-law was supportive but could hardly discuss with her the challenges she had with her husband or the rumours circulating in the community, leaving Jalia isolated. She avoided sharing her everyday challenges and family secrets with others because, she said, ‘you will tell people something and as soon as you have left, they will tell other people and problems start’.

129

Jalia told me that one day in 2011, a neighbour told her about a VSLA and asked her to join. She joined, but hesitantly, because even though she ‘wanted to belong somewhere’, she did not know how the VSLA would make a difference in her life. After participating in the VSLA meetings for a while, she realized that in addition to helping her save money, the meetings offered her a space to confide in trusted friends, providing her ‘counselling about my marriage and other problems’ she faced in her home. She mentioned that the advice given during VLAs meetings proved insightful in managing her home and taming conflicts with her husband:

Old women in our group sometimes give ideas, and then you pick something to learn. There is one particular woman who counsels members especially us, the young marrieds. She tells us that marriage is about patience. One time she told us, ‘If a man quarrels, put water in your mouth – you will not be able to answer any word, [so] if he had an intention of beating you up after the quarrel, he will not [because] you have not replied’.

VSLA meetings provided Jalia a space to address everyday challenges, with other women. While Ibargu¨en-Tinley (2014) gives examples of how a leader in a women’s micro-enterprise in Mexico confronted the abusing husband of one of the members, Jalia learnt from her VSLA friends to instead deploy silence in challenging her husband’s power. Scott’s (1985) concept of ‘weapons of the weak’, in which subjugated people use non-confrontational means to contest their repression, is illuminating in this case. Jalia always looked forward to VSLA meetings to ‘chat out’ stress, relieve her burdens, and meet friends. It mattered less whether she had money to save. The meetings were a forum to discuss not only financial activities, but also ‘common problems affecting the community and family’ (Saha, 2014, p. 619). Older mother figures offered a source of solidarity and strength for the younger members.

Putnam (1995) defines social capital as those features of social life – such as networks, norms, and trust – that allow pursuance of collective goals. As Harriss (2001) notes, trusting relations develop within a context, and the case of VSLAs shows that these meetings facilitated such trust, and thereby generated more social capital for members. Studies on bonding social capital also show that interactions and connections in groups generate solidarity and social capital for their members (Saha, 2014). Jalia’s story illustrates a form of silence-in networks where network members teach each other to deploy silence to overcome challenges of everyday life. Women in Jalia’s group taught each other how to deploy silence in their homes to challenge the authority of 130

their husbands, solidify their marriages, reduce conflicts in homes, and expand personal friendship networks.

VSLA membership and the use of silence-through networks

In August 2012, I met and interviewed Sanyu, a thirty-five-year-old woman. During our first meeting she told me about her daughter Teddy, aged ten, who was physically challenged and confined to a wheelchair. Teddy’s head was big and looked much heavier than the rest of her body. Sanyu said that Teddy’s head ‘became deformed after birth’, and that, with no immediate attention, her condition deteriorated. A belated operation was finally done three months later, which stopped the head from bulging but confined her to a wheel chair. During our subsequent interactions, I realized that Sanyu was concerned not only about her disabled child but also the lack of support from her husband.

Sanyu explained that she became pregnant while she was still at school; the father was discouraged from marrying her by his parents because it would affect his education. He eventually convinced his family and they agreed to the marriage. Teddy’s deformity triggered rumours from her husband’s family and other community members that Sanyu had committed adultery while she was pregnant. Her husband became resentful, as did his family, and he started withholding support she needed for their daughter to get treatment. When it was time for Teddy to go to school, her husband refused to provide support. Although Sanyu intermittently operated a small nursery school at her home, the money was hardly enough to cover Teddy’s needs. Sanyu faulted her husband:

The big challenge is that her father is not cooperative. I remember there was a school in Kampala where she would have gone and would be in the boarding section. However, he said, ‘That money is too much’. There were several other offers for help but he never followed them up. He always claimed he had no money. He left the burden to me alone.

Sanyu said she had grown weary of her husband’s failure to fulfil his obligations. On several occasions, she contemplated leaving him but the thought of lacking any fallback position frightened her. One of her friends advised her to join a VSLA:

131

One day, I told my friend that I wanted to separate from my husband. She told me, ‘Be patient. Such things happen in marriages, you just have to be strong’. She sat me down and said, ‘Do you know that I am an old woman. If I had prepared for my old age I would not have suffered with marriage’. She told me she had divorced her first husband but because she had no home, she had to remarry and regretted it. She told me to join a VSLA and plan my future.

When Sanyu decided to join the VSLA, her husband opposed the idea, arguing that VSLAs were ‘useless’ and had no direct benefits. Although not all the men I met shared this belief, it was common for men to scorn VSLAs and to discourage their wives from joining. However, Sanyu understood her husband’s refusal, which was based on his assumption that women who join meet other women who influence them into promiscuity. ‘He told me that when women go for such things they start being adulterous’, Sanyu said in one of the our interviews. She also saw his reaction in relation to the rumours that Sanyu had been adulterous, leading to Teddy’s deformity. When her husband told her that the groups were ‘useless’, she knew that convincing him to grant her permission would be futile. She decided to join secretly:

I went through one of my neighbours who was a member. We used the name of my friend’s daughter because my father-in-law is also in that savings group. I secretly make my savings without their knowledge, including their son [her husband]. Whenever I receive my savings money, I go very fast to our home, buy a cow, and keep it there. Now, I am planning to sell my cows and buy land in Kasana town and build [own home].

The behaviour of Sanyu’s husband should not be seen as unique. Lønborg and Rasmussen (2014) cite husbands’ refusal to approve their wives’ participation in VSLAs as one of the reasons for women not joining VSLAs. Rah- man (1999) also found similar tensions and frustrations in homes that were attributable to women’s involvement in Grameen Bank activities. How- ever, what these studies fail to show are the subtle but effective steps, the ‘weapons of the weak’, that women like Sanyu take to overcome these barriers, avert family conflicts and achieve personal goals. This is what I have called ‘silence-through’ networks where through network members Sanyu was able to effectively utilize silence of her friend to save secretly.

132

The role of a network of friends is important to note. Sanyu exploited her networks to join the VSLA and trusted that her friend would not reveal her secret. In England, Campbell (2001) reports that individuals found networks of friends with face-to-face interactions more rewarding than simply being an active participant in an official group. In this case, the VSLA did not directly offer benefit to Sanyu; it was through her friendship networks that Sanyu was able to benefit from the VSLA. What Sanyu’s story reveals is that the networks outside the VSLA are critical in influencing decisions to join a VSLA and the subsequent benefit that accrue. By 2012, Sanyu had been saving secretly for four years without her husband’s knowledge. When I visited her again in September 2014, she told me she had sold her cows, bought land, and started constructing a house. Sanyu was confident that when she eventually separates from her husband, she will be able to manage on her own because of the silence she deployed through networks.

Sustaining VSLAs membership and unintended consequences of social networks

The desire for financial autonomy can have unintended consequences, as participating in a VSLA creates an additional demand for some women who have no constant flow of income. While some who are unable to participate must leave their VSLA, others receive advice from their friends focusing on how to stay in the group. According to a community development officer (CDO) in Luwero, some women went to great lengths to sustain their membership:

A female member joined the VSLA but did not have the money to save and sustain membership. She became financially constrained and decided to take advantage of the coffee traders in the area and offered sex for money. When she was caught, she confessed to her husband that she wanted money for VSLAs. In another case, a man had bought his wife five gomesi [traditional dresses]; she kept selling one after another so that she could save in the VSLA. By the time they brought the case here, all the five gomesi had gone.

The CDO later told me that she had learnt that women were influenced by their friends to engage in such acts. This highlights the possible costs resulting from the influence of friendship networks, but it also implies that women remain vulnerable within broader cultural boundaries. In such cases, men’s behaviour, perceived as irresponsible, pushes women to join VSLAs. In an FGD, women discussed at length how men neglect their duties and responsibilities, such as paying school dues for their children and providing for the home’s general upkeep.

133

Borrowing and loan repayment was also mediated through social networks involving group leaders. Throughout my membership in Bumu, I witnessed how the group leaders influenced decisions on borrowing and loan repayment for their own benefit and that of their friends. For example, Aidah was given a loan without fulfilling certain requirements because she was a personal friend of Maria, the group’s treasurer. They described themselves as best friends; Aidah had even attended to Maria for a month when she was bedridden after an accident.

In their discussion of Oxfam’s Savings for Change initiative in Mali, Ashe and Neilan (2014) note a similar pattern. Although procedures for accessing loans were followed, such as everyone publicly stating why they needed the loan, it was usually the case that someone with higher status would have easier access than someone with lower status. Social networks also smoothed out when certain members defaulted on loans. In my interview with Loyce, a forty-eight-year-old woman and a member of a VSLA, she described how one of the members who borrowed money and failed to repay it was not sanctioned because he was a friend of the group’s chairperson. Cheating and default payments on loans are sustained by the interrelatedness of the community; leaders and those who default are also part of the community, with spouses, children, and relatives. Punishing them would also mean risking being seen as punishing their kin.

These networks that predate the introduction of VSLAs work to filter which benefits accrue from VSLAs and for whom. For some, these networks make it possible to gain more of the benefits of VSLAs. Women in VSLAs were aware of the structural challenges they faced, including the power men have over them. Simply starting a VSLA in the community may not remove these barriers. However, through silence-in and silence-through networks women were able to navigate such barriers, by saving secretly or diversifying into other economic ventures. Madina, a VSLA member, started making beads, based on the advice of her relative in Kampala city. She then introduced the idea to other women in her VSLA as a way of supplementing their income. After making the beads, they are collected and sold in Kampala through Madina’s relative at UGX1500 (0.60USD) making an average of UGX6000 (2.4USD) per month which money is then used to sustain membership in VSLA. When the bead-making business prospered, it was expanded to other villages and picked up by other women in other villages involved in VSLAs; as one member said, ‘Those of us who learnt went on teaching people and many kept joining us’.

134

By diversifying into other activities such as bead-making to sustain their membership, women in these VSLAs were also creating a space they could control outside of men’s influence.

Whereas studies (Karlan et al., 2012) show that VSLA participants expand or start businesses using loans received from VSLA, the stories of women having sex for money or those making beads to sustain their membership shows that a different dynamic to sustain membership is as significant as the need for loans to start businesses.

Conclusion

Externally initiated interventions are not implemented in a vacuum. Development organizations that roll out savings schemes provide women opportunities to save and borrow, but also to develop organizational and leadership skills that allow them to contribute to matters of community interest (Ashe and Neilan, 2014). This means that benefits beyond financial inclusion are part of the savings model, but how such benefit are appropriated and distributed remains to be closely studied. This research shows that social networks are at the heart of how benefits beyond financial inclusion are appropriated.

Social networks influence women’s decisions to join a VSLA, challenge structural barriers (such as male power), and expand their personal networks. The findingssuggest that VSLA membership is driven by the need for financial autonomy; most of the women were less concerned with how much they saved than with retaining control over whatever they saved. These findings support earlier studies that posit that it is not just access but control that matters (Kabeer, 1997; Kantor, 2003; Mayoux, 2010). The cases presented in this study suggest that social networks ultimately facilitate financial control, which in turn influence network dynamics.

The experiences of women in VSLAs and the significance of social networks and social capital in understanding these experiences should not detract from the underlying vulnerability faced by women participants to VSLAs. The desire for financial autonomy should also be looked at in relation to other factors such as mistreatment or lack of support from spouses. Women in this community remain vulnerable to structural barriers to financial autonomy, and limited or no income flows; attempts at sustaining VSLA membership appear to exacerbate the vulnerability of some. Attaining financial independence would allow women to successfully address structural

135

barriers. Whereas the three case studies show that VSLA membership can help women negotiate power structures, participation itself may not be panacea for participants’ vulnerability.

Empirical studies continue to be inconclusive. Kim et al. (2007) report the results of a randomized control trial that showed a positive impact of microfinance programmes on the reduction of intimate partner violence in South

Africa, but Izugbara’s (2004, p. 82) study found that micro-lending schemes in south-eastern Nigeria did not enable women to ‘challenge and overcome the more complex set of constraints that mediate women’s access to resources, power and opportunities in society’. Karlan et al. (2012) found little evidence to suggest that VSLAs membership improves participants’ involvement in community activities. The cases discussed here support Izugbara’s (2004) findings in part since some networks can be disempowering. In this study, women’s networks are their main weapons against structural barriers. Women deployed ‘weapons of the weak’ in the form of silence-in (Jalia’s story) and silence-through (Sanyu’s story) networks to negotiate power. No doubt, VSLAs also play a catalysing role in development by allowing horizontal learning as participants ‘share experiences and ideas with peers’ (Toomey, 2011). VSLAs provide women a space where they can interact with friends (in the case of Jalia); where they can turn to in case things fail to work out with their spouses (in the case of Jane); or a space where they can challenge mistreatment by becoming financially independent (in the case of Sanyu); or reach out to each other to diversify their incomes (in the case of Madina).

The findings support studies that suggest that participation in micro- enterprises generates cohesion, solidarity, and bonding social capital for their members (Gregson et al., 2004; Thabethe, Magezi and Nyuswa, 2012; Ibargu¨en-Tinley, 2014). While interventions such as VSLAs have a significant influence on community life, women’s ability to negotiate space cannot be simply explained by the presence of an external intervention. Changes in the lives of women should be measured by paying attention to how they access intangible resources and how they deploy networks to solve everyday problems. Social networks influence the decision to join interventions, challenge cultural boundaries, and search for financial independence. As they subtly challenge men, this has long-term implications for the future of gender relations at the community level. Future studies should compare benefits of VSLAs membership to members’ use of formal savings channels such as banks and microfinance institutions. 136

Acknowledgements

The author acknowledges all study participants and the useful comments from the three anonymous reviewers on the initial manuscript. I also acknowledge support from CoHeRe project at University of Amsterdam and Center for Social Science Research on AIDS (CeSSRA), School of Social Sciences, Makerere University Kampala.

Funding

Laban Kashaija Musinguzi is with the Department of Social Work and Social Administration, Makerere University, PO Box 7062, Kampala, Uganda, and Amsterdam Institute of Social Science Research, AISSR, University of Amsterdam, Nieuwe Achtergracht 166, 1018 WV Amsterdam, the Netherlands.

References Adams, A. M., Madhavan, S. and Simon, D. (2002) Women’s social networks and child survival in Mali, Social Science & Medicine, 54 (2), 165 – 178. Allen, H. (2006) Village savings and loan associations-sustainable and cost-effective rural finance, Small Enterprise Development, 17 (1), 61 – 68. Allen, H. (2009) Impact and Programme Evaluation of Plan and UHIKI‘s Joint VSL Programme in Tanzania, Plan/VSL Associates, Plan Tanzania. Allen, H. and Hobane, P. (2004) Impact Evaluation of KupfumaIshungu, CARE International, Arusha, Tanzania. Allen, H. and Panetta, D. (2010) Savings Groups: What Are They? Overview, The SEEP Network, Washington, DC. Anyango, E., Esipisu, E., Opoku, L. et al. (2007) Village savings and loan associations: experience from Zanzibar, Small Enterprise Development Journal, 18 (1), 11 – 24.

137

Ashe, J. and Neilan, K. J. (2014) In Their Own Hands: How Savings Groups Are Revolutionizing Development, Berrett-Koehler Publishers, San Francisco, Calfornia. Benda, C. (2012) Community rotating savings and credit associations as an agent of well-being: a case study from northern Rwanda, Community Development Journal, 48 (2), 232 – 247. Bouman, F. J. A. (1995) Rotating and accumulating savings and credit associations: a development perspective, World Development, 23 (3), 371 – 384. Bourdieu, P. (1986) The forms of capital, in John G. Richardson, ed., Handbook of Theory and Research for the Sociology of Education, Greenwood Press, New York, pp. 241 – 258. Braun, V. and Clarke, V. (2006) Using thematic analysis in Psychology, Qualitative Research in Psychology, 3 (2), 77 – 101. Campbell, C. (2001) Social Capital and Health: Contextualizing Health Promotion within Local Community Networks, LSE Research Online, accessed at: http://eprints.lse.ac.uk/473/1/300- social_capital.pdf (10 May 2014). CARE (2011) Closing the Gap: Microfinance in Africa: State-of-the-Sector Report, CARE International. accessed at: http://www.care.org/sites/default/files/documents/ MF-2011-CARE- Access-Africa-Closing-the-Gap.pdf (19 February 2015). Cattell, V. (2001) Poor people, poor places, and poor health: the mediating role of social networks and social capital, Social Science and Medicine, 52 (10), 1501 – 1516. Cleaver, F. (2005) The inequality of social capital and reproduction of chronic poverty, World Development, 33 (6), 893 – 906.

Dowla, A. (2006) In credit we trust: building social capital by Grameen Bank in Bangladesh, Journal of Socio-Economics, 35 (1), 102 – 122.

Gayen, K. and Raeside, R. (2010) Social networks and contraception practice of women in rural Bangladesh, Social Science and Medicine, 71 (9), 1584 – 1592. Glaser, B. G. (1978 silence in) Theoretical Sensitivity, Sociology Press, Mill Valley, CA. Goetz, A. M. and Gufta, R. S. (1996) Who takes the credit? Gender, power, and control over loan use in rural credit programs in Bangladesh, World Development, 24 (1),45 – 63. Gregson, S., Terceira, N., Mushati, P. et al. (2004) Community group participation: Can it help young women to avoid HIV? An exploratory study of social capital and school education in rural Zimbabwe, Social Science & Medicine, 58 (11), 2119 – 2132. Harriss, J. (2001) Depoliticizing Development: The World Bank and Social Capital, Left Word Books, New Delhi. Häuberer, J. (2014) Social capital in voluntary associations, European Societies, 16 (4), 570 – 593.

138

Hendricks, L. and Chidiac, S. (2011) ‘Village savings and loans: a pathway to financial inclusion for Africa’s poorest households’, paper presented at the Global Micro Credit Summit, 14 – 17 November 2011, Valladolid, Spain. Ibargu¨en-Tinley, L. M. (2014) Social capital development in a rural women’s micro-enterprise in Mexico: insights on leadership, trust and cooperation, Community Development Journal, 50 (2), 229 – 243. Izugbara, C. O. (2004) Gendered micro-lending schemes and sustainable women’s empowerment in Nigeria, Community Development Journal, 39 (1), 72 – 84. Kabeer, N. (1997) Reversed Realities: Gender Hierarchies in Development Thought, Verso, London. Kantor, P. (2003) Women’s empowerment through home-based work: evidence from India, Development and Change, 34 (3), 425 – 445. Karlan, D., Thuysbaert, B., Udry, C. et al. (2012) Impact Assessment of Savings Groups: Findings from Randomized Evaluations of CARE Village Savings and Loans Associations in Ghana, Malawi and Uganda. Final Report. Innovations for Poverty Action (IPA). New Haven, USA. Kim, J. C., Watts, C. H., Hargreaves, J. R. et al. (2007) Understanding the impact of a microfinance-based intervention on women’s empowerment and the reduction of intimate partner violence in South Africa, American Journal of Public Health, 97 (10), 1794 – 1802. Krishna, A. P. and Goldey, P. (2009) Social capital and its ‘downside’: the impact on sustainability of induced community-based organizations in Nepal, World Development, 38 (2), 184 – 194. Lakwo, A. (2006) Microfinance, Rural Livelihoods, and Women‘s Empowerment in Uganda, African Studies Centre, Leiden University, Netherlands. Lin, N. (2001) Social Capital: A Theory of Social Structure and Action, Cambridge University Press, New York. Lønborg, J. H. and Rasmussen, O. D. (2014) Can microfinance reach the poorest: evidence from a community-managed microfinance intervention, World Development, 64, 460 – 472. Lowicki-Zucca, M., Walugembe, P., Ogaba, I. et al. (2014) Savings groups as a socioeconomic strategy to improve protection of moderately and critically vulnerable children in Uganda, Children and Youth Services Review, 47 (2), 176 – 181. Luwero District Local Government (2009) Higher Local Government Statistical Abstract for Luwero, 2008/2009, accessed at: http://www.ubos.org/onlinefiles/uploads/ubos/ 2009_HLG_%20Abstract_printed/Luwero%20LDG%20FINAL.pdf (23 July 2014). Mayoux, L. (2010) Reaching and empowering women: towards a gender justice protocol for a diversified, inclusive, and sustainable financial sector, Perspectives on Global Development and Technology, 9 (3 – 4), 581 – 600. 139

Narayan, D. and Pritchett, L. (1999) Cents and sociability: household income and social capital in rural Tanzania, Economic Development and Cultural Change, 47, 871 – 897. Perkins, D., Hughey, J. and Speer, P. W. (2002) Community psychology perspectives on social capital theory and community development practice, Community Development Society, 33 (1), 33 – 52. Plan Uganda (2011) Village Savings and Loan Association (VSLA) Approach; Plan Uganda‘s Experience November, Plan Uganda, Kampala, Uganda. Platteau, J. and Gaspart, F. (2003) The risk of resource misappropriation in community-driven development, World Development, 31 (10), 1687 – 1703. Putnam, R. D. (1995) Bowling alone: America’s declining social capital, Journal of Democracy, 6 (1), 65 – 78. Putnam, R. D. (2000) Bowling Alone: The Collapse and Revival of American Community, Simon and Schuster, New York. Rahman, A. (1999) Micro-credit initiatives for equitable and sustainable development: who pays?, World Development, 27 (1), 67 – 82. Saha, S. (2014) ‘More health for the money’: an analytical framework for access to healthcare through microfinance and savings groups, Community Development Journal, 49 (4), 618 – 630. Scott, J. C. (1985) Weapons of the Weak: Everyday Forms of Peasant Resistance, Yale University Press, New Haven, CT. Spradley, P. (1980) Participant Observation, Wadsworth, Cengage Learning, Belmont, CA.

Ssewamala, F. M., Ismayilova, L., McKay, M. et al. (2010) Gender and the effects of an economic empowerment program on attitudes toward sexual risk-taking among AIDS-orphaned adolescent youth in Uganda, Journal of Adolescent Health, 46 (4), 372 – 378.

Thabethe, N., Magezi, V. and Nyuswa, M. (2012) Micro-credit as a community development strategy: a South African case study, Community Development Journal, 47 (3), 423 – 435.

Titeca, K. and Vervisch, T. (2008) The dynamics of social capital and community associations in Uganda: linking capital and its consequences, World Development, 36 (11), 2205 – 2222. Toomey, A. H. (2011) Empowerment and disempowerment in community development practice: eight roles practitioners play, Community Development Journal, 46 (2),181 – 195. van Rooyen, C., Stewart, R. and de Wet, T. (2012) The impact of microfinance in sub-Saharan Africa: a systematic review of the evidence, World Development, 40 (11), 2219 – 2262.

Vermaak, N. J. (2001) Rural financial schemes’ contribution to community development, Community Development Journal, 36 (1), 42 – 52. 140

Woolcock, M. and Narayan, D. (2000) Social capital: implications for development theory, research and policy, The World Bank Research Observer, 15 (2), 225 – 249.

141

Paper V Externally Initiated Interventions and Community Level Social Bonds: An Ethnographic Study of Village Savings and Loans Associations in rural Uganda

1 Department of Social Work and Social Administration, Makerere University, P.O. Box 7062, Kampala, Uganda

2Amsterdam Institute of Social Science Research, AISSR, University of Amsterdam, Nieuwe Achtergracht 166, 1018 WV Amsterdam, Netherlands. Email: [email protected].

3College of Health Sciences, Makerere University Kampala Uganda. Email:[email protected].

This paper has been submitted and currently under review in World Development

142

Abstract Although studies have examined externally initiated interventions as important spaces for the generation of social capital, there is little evidence of how community social processes influence disproportionate distribution of social resources. Evidence show that externally initiated groups generate various social, symbolic and material resources for community members, who access these resources differently: some benefit while others are disadvantaged. Through ethnographic study of village savings and loans associations (VSLAs) in rural Luwero, Uganda, this paper sought to answer two questions. First, what are the social processes and mechanisms in the community that are affected by externally initiated interventions? And second, how do these processes (re)shape the social bonds in this community and enable disenfranchised community members to challenge the negative effects of externally initiated interventions? The findings show that VSLAs reinforce existing community differences. Through community norms of general “goodness” obuntu bulamu, the capacity of members to curb cheating is undermined. The perceived inability to confront cheaters is overcome through the creation of offshoot groups which emerge as unforeseen outcomes of unequal power relations in VSLA structures. Within offshoot groups, new loyalties (trust) are (re)shaped. Although the offshoot groups provide a new form of social bonding, they do not allow the members to gain new power within the community hierarchy forcing them to operate in secret. The results contribute to ongoing debates on the unintended potential negative consequences of externally initiated interventions, VSLAs in particular, in sub-Saharan Africa.

Key words: externally initiated interventions; Village savings and loans associations; social processes; social capital; rural; Uganda

143

Introduction Externally initiated groups generate various social, symbolic and material resources for community members, who access these resources differently: some benefit while others are disadvantaged. Through community social benefit processes based on traditional norms, existing power asymmetries and social inequalities are reinforced and challenged: enhanced because those who benefit from externally initiated interventions gain additional influence and power; challenged because those who are disadvantaged devise means to create opportunities outside the externally initiated interventions.

This paper presents the findings of an ethnography that examined externally initiated groups, village savings and loans associations (VSLAs), in the Luwero district, central Uganda. The ethnographic research was a component of a research project2 that sought to document community health resources in rural Uganda. In the process of collecting the data, it became apparent to the authors that externally initiated interventions were not without fallout. Therefore, the authors sought to answer two questions directed at determining how externally initiated interventions shape community-level social bonds using the example of VSLAs. First, what are the social processes and mechanisms in the community that are affected by externally initiated interventions? And second, how do these processes (re)shape the social bonds in this community and enable disenfranchised community members to challenge the negative effects of externally initiated interventions? We pay particular attention to what happens when people are “cheated” out of VSLAs. Using a social capital framework as conceptualized by Robert Putnam (1993, 1995), this paper provides evidence of how externally initiated interventions (re)shape community-level social bonds as disenfranchised VSLAs members create off-shoot groups wherein various forms of loyalties (such as trust) are established. Putnam (1995:664-665) conceptualized social capital as those features of social life such as networks, norms, and trust that allow pursuance of collective goals. This conceptualization emphasises the benefits that accrue from people’s “participation or civic engagement in local community networks” (Gregson, Terceira, Mushati, Nyamukapa, & Campbell, 2004:2121) to which programs like VSLAs contribute. We tease out the community dynamics that help explain how off-shoot groups affect social bonds. The authors hope that this description of an understudied

144

phenomenon will contribute to ongoing debates on the unintended potential negative consequences of externally initiated interventions, VSLAs in particular, in sub-Saharan Africa.

Externally initiated interventions, microfinance and VSLAs Due to a growing consensus in the late 1980s for participatory approaches in social services delivery in low- and middle-income countries, interventions initiated by the state and non-state actors have focused on group-based models (Gregson et al., 2004). Participatory group-based interventions are justified on grounds that they ‘‘increase the voice’’ of the people (World Bank, 2000:129), benefit the poor (Thorp, Stewart, & Heyer, 2005), provide avenues “where different social resources accumulate” (Häuberer, 2014:2), generate social capital (Feigenberg, Field, Pande, Rigol, & Sarkar, 2014; Janssens, 2010) and engender virtues of civic participation and collective action (Cleaver, 2007; Putnam, 2000).

The burgeoning literature examining externally initiated interventions also indicates that they create inequalities and asymmetries of power within communities (Krishna & Goldey, 2009; Stevens, 2008), increase dependence on external actors (Thurman et al., 2008), undercut trust and reciprocal relations within communities (Vollan, 2012: 763), and potentially facilitate elite capture and patronage (Dasgupta & Beard, 2007; Labonte, 2011; Lund & Saito-Jensen, 2013; Fritzen, 2007; Platteau & Gaspart, 2003; Vollan, 2012), making it hard to generate social capital through external interventions (Ostrom, 2000).

Nonetheless, efforts to “induce” social capital through externally initiated interventions continue. In Uganda, such efforts date back to the colonial period when the British protectorate government adopted the first Native Authority Ordinance3 in 1919 to help “communalise” local communities (Grischow & McKnight, 2008). In addition, when a neo-liberal policy agenda started taking shape in the 1980s, the reduced role of government in direct services provision opened space for non-state actors to initiate programs in communities in Uganda. As global attention focused on addressing poverty, coupled with limited access to credit facilities by the poor, several interventions, notably microfinance, were proposed for low income countries following the success of the Grameen Bank model (Dowla, 2006; Rahman, 1999). In Uganda,

145

emphasis on microfinance coincided with the UN declaration of access to financial services as a human right in 2005 and the country’s 2005/2006 general elections. Uganda’s development agenda shifted from poverty reduction to “prosperity for all” coalescing around the provision of microfinance to individuals within community groups plus government-engineered savings and credit cooperatives (SACCOs) (Golooba-Mutebi & Hickey, 2010; Hickey, 2013).

Microfinance interventions (e.g., micro-credit and micro-savings) have often been studied in relation to poverty reduction, community development and empowerment (Benda, 2012; Imai, Gaiha, Thapa & Annim, 2012; Imai, Arun, & Annim, 2010; Izugbara, 2004; Lønborg & Rasmussen, 2014; Mayoux, 2001; Thabetha, Magezi & Nyuswa, 2012; Weber & Ahmad, 2014). In 2011, this journal ran a special issue that addressed the impact, sustainability and outreach of microfinance (Hermes & Lensink, 2011). However, concerns abound on how microfinance studies are conducted and the effectiveness of microfinance interventions. Some studies claim that group-based microfinance create stocks of social capital through regular meetings and repeated interactions (Anderson, Locker & Nugent, 2002; Ostrom, 1994); others, argue that microfinance groups are likely to exclude and exploit the poorest and that impact can be polarizing even at household level (Copestake, 2002; Rahman, 1999). This is in addition to the crises in “India, Morocco, Bosnia, Pakistan, Nicaragua and Nigeria where thousands are over- indebted” (van Rooyen, Stewart & de Wet, 2012:2250) leading to concerns about a potential “mission drift” (Mersland & Strøm, 2010).

The microfinance approach we examine, the VSLAs, appear designed to address these concerns. Although VSLAs can be seen as one form of microfinance (Adams, Muhammed, & Boateng- Kwakye, 2014), they usually lie below the regulatory radar (Allen, 2006). VSLAs are accumulating savings and credit associations whose membership ranges between 15 and 30 (Anyango et al., 2007). VSLAs operate on the principle of self-managing with no external capitalization, but rely on “high level organizational structure, disciplined procedures and a basic form of record keeping” processes (Allen, 2006:63). Such processes are usually absent in rural areas, which calls for external facilitation, usually by NGOs. Various agencies have scaled up CARE’s 1991 VSLA model. Numerous examples of VSLAs as well as community banking models in several African countries indicate the potential of savings groups in promoting social

146

development based on trust, solidarity and social cohesion (Bähre, 2007; Biekpe, 2004). It is well documented how VSLAs affect targeting and participation (Lowicki-Zucca, Walugembe, Ogaba & Langol, 2014); financial inclusion (Adams et al., 2014); education achievement (Ssewamala et al., 2010); child protection, food security and nutrition (Abubakari, Sadik & Keisan, 2014; Brunie, Fumagalli, Martin, Field& Rutherford, 2014); and, social capital generation (Feigenberg et al., 2014). It is also clear that once VSLAs have been formed, attrition occurs (Lønborg & Rasmussen, 2014). Less understood are the dynamics surrounding this attrition, more so what happens when people opt out or are “cheated” out of VSLAs.

Evidence on VSLAs and social capital in sub-Saharan Africa remains scanty. In a systematic review of the evidence on the impact of microfinance in sub-Saharan Africa, van Rooyen et al. (2012) found no study on the impact of microfinance on social cohesion and that a lot of evidence dealt with micro-credit than micro-savings. Plausible explanations include; difficulty in developing qualitative measurement indicators for non-financial benefits (van Rooyen et al., 2012; Vermaak, 2001); historical pre-occupation with numbers than quality of contribution in development (Goetz & Gufta, 1996); and skewed evidence in favor of Asian communities where microfinance movement originated (Feigenberg et al. 2014; van Rooyen et al., 2012; Weber & Ahmad, 2014). In addition, “Micro-savings schemes are also newer and there is less evidence of its effectiveness” (van Rooyen et al., 2012:2259). The growing popularity of VSLAs in Africa opens prospects for research to inform decision-making as external actors scale up innovations that increase financial access for the poor. VSLAs research is also needed to understand community variations that make program adaptation context specific (Bouman, 1995). Our study contributes to this knowledge by exploring understudied yet important social dimensions of community dynamics, and social bonds that emerge from VSLAs. In Uganda the first VSLAs were initiated by CARE International in West Nile region in 1999 and later in Kamuli and Luwero districts in 2006/2007 by Plan Uganda. By the time this study was conducted there were approximately 700 VSLAs in rural Luwero, with an average membership of 15-30.

Studies of externally initiated community groups in Uganda have paid limited attention to how VSLAs influence community-level bonds. Katungi, Machethe, and Smale (2007:188), claim that

147

group-based approaches should be encouraged “…for their positive externalities in strengthening community networks.” Katungi et al. (2007) argument partially contradicts Titeca and Vervisch (2008:2219) who claim that successful groups start with strong bonds of social capital to benefit external linkages because externally initiated interventions bring new ideas and information to the community, which can be detrimental to bonding social capital. Bonding social capital refers to social resources accessible through homogenous groups, usually comprised of family and kinship networks (Putnam, 2000; Woolcock & Narayan 2000). This calls for more research into these processes.

Social capital and community associations Most definitions of social capital draw on the work of Pierre Bourdieu, who linked “resource exchange to membership in a group” (Bourdieu 1986: 248). Although Bourdieu viewed social capital as a resource in relation to power, conflict and daily class struggles, Putnam’s conception of social capital placed more emphasis on social networks enhanced through voluntary associations, trust and reciprocity. The social networks dimension of social capital popularized by Putnam has been adopted by various writers arguing that social capital can be generated, nurtured (Muriisa & Jamil, 2011) and induced through group processes (Krishna & Goldey, 2009). The social network dimension also suggests that networks can be formal, emerging as individuals make a deliberate choice to join groups, or informal, evolving naturally and casually among family members and friends in the community (Häuberer, 2014:3).

As noted earlier, bonding social capital thrives in homogeneous, inward-looking groups (Häuberer, 2014; Ferlander, 2007; Putnam, 2000) and more generally from a sense of common interest. Members in such groups are bonded together by adherence to group norms and shared identity fostered through regular face-to-face interactions (Anderson et al., 2002; Larsen, 2010). In the context of group-based interventions, interactions are fostered through group meetings. Feigenberg et al. (2014) found that group meetings influenced social capital among the microfinance clients in India. Stewart (2005) observes that microfinance groups thrive on creation and maintenance of bonds among members.

148

Evidence also indicates that bonding groups are exclusive and emerge against oppression, as was the case with cooperative associations formed in 1913 to counteract colonial exploitation in Uganda (Kyazze, 2010). Kabeer, Sudarshan and Milward (2013), show how organisations’ efforts in mobilizing women against exploitation by powerful actors led to shared identity and bonding. In the context of externally initiated interventions, and social capital literature, the argument is that community associations created by external agencies need to transform bonding into bridging and linking ties (Perkins, Hughey, & Speer, 2002; Titeca & Vervisch 2008; Vervisch & Titeca, 2010; Woolcock & Narayan, 2000). Bridging social capital refers to social resources accessible through networks outside the homogeneous groups (Putnam, 2000; Häuberer, 2014). Linking social capital are “networks that connect people across explicit vertical power differentials” (Krishna & Goldey, 2009:2207). In the context of rural communities, it is apparent that groups initiated from “outside” are more likely to provide linking and bridging connections, while groups that emerge from the “inside” provide bonding connections. However, externally initiated interventions can also influence emergence of bonds encapsulated as resistance against exploitation. However, the success of these bonds will depend on how the central features of social capital such as trust and power influence interactions among members.

Trust is critical because it acts as the “wires” through which associational activity and bonding relationships become meaningful for the network members. The idea that trust relates to associational life implies a link between “passive emotional sentiments” expressed as trust and “behaviours that produce familiarity” (Larsen et al., 2004:65). Larsen et al. further argue that such a perceived link is very important for bonding social capital because it provides a basis upon which members rely on each other to accomplish their goals (p. 69). Two forms of trust, particularized and generalized, have been distinguished in social capital literature (Putnam, 2000; Bjørnskov, 2006:2). Putnam (2000) notes that particularized trust arises from face-to-face interactions in community groups, which spills over into generalized trust. Generalized trust is trust between people who do not have regular contact or those with whom people have no face- to-face interactions (Bjørnskov, 2006).

Unlike Putnam (1995), who viewed social capital in a more positive light, as a producer of “civic engagement,” Bourdieu portrayed it as producing and reproducing inequalities. Bourdieu (1986,

149

1989) understood social capital as one of the scarce resources whose access is negotiated in daily practices within which people with social connections gain access because they possess a degree of power reflected in their position in the community. Some studies have made similar conclusions. Krishna and Goldey (2009), in their study of community based organisations in Nepal, found that social capital induced through groups produced power and social inequalities, as some members were elevated above others. Cleaver’s (2005) ethnographic study in Tanzania and Stevens’ (2008)4 study of social capital and inequalities in health in New Zealand argued along similar lines. The power differences then form the basis of social relationships and become part of a dynamic process in which individuals negotiate their way out of the power imbalances (Boissevain, 1974). Negotiations against power imbalances includestrategies employed by marginalized people such as “use of implicit understandings and informal networks…to avoid any direct, symbolic confrontation with authority” (Scott, 1985: xvi). Strategies of resistance could also be well organised from external actors in which case, they are less about traditional “weapons of the weak” (Scott, 1985) and more of “weapons of the organised” (Kabeer et al., 2013). Marginalized groups draw from these weapons to circumvent power asymmetries common in participatory approaches (Lund & Saito-Jensen, 2013). These efforts collectively underpin the significance of human agency in challenging inequalities and recognizes variations in human experience and influences on sources of social capital (Cattell, 2004; Cleaver, 2007).

Although social capital as a concept has been generally well received and applied in various disciplines, it remains contested. In the development literature, its main criticism is on “its coherence, potential normative effects and types of policy and practice to which it might give rise” (Bebbington, 2002:800). Cleaver (2005) suggests that building social capital may be in futility for the poorest who experience clusters of disadvantage. Ostrom (2000) shows how sometimes commentators have branded social capital concept as another “fad”, even when it has a significant contribution in development. While discussing the social capital debates at the World Bank, Bebbington, Guggenheim, Olson and Woolcock (2004) provide a detailed account of how the internal discussions about social capital shaped its place in development studies, influenced debates for future research policy and practice and the subsequent focus on quantitative measures of social capital (see Bebbington, 2004, for more on debates). Given this

150

influence in the development of the concept, it is not surprising that earlier thinking and writing about social capital employed top-down statistical surveys (Campbell & Gilles, 2001). In Uganda, social capital studies (see Bakeera et al., 2010; Hassan & Birungi, 2011; Katungi et al., 2007) continue to employ quantitative proxy measures, which tend to overlook the everyday social processes. Calls for qualitative studies (Morvant-Roux, Guérin, Roesch & Moisseron, 2014) and long-term ethnographic accounts of access of microcredit and savings underscore the need to articulate everyday social processes. It is the articulation of these processes on which this paper contributes.

Setting and methods The paper is based on ethnographic fieldwork conducted between 2012 and 2014 in a rural community in Luwero district, central Uganda. The community where fieldwork was conducted is relatively remote, located approximately five kilometres off the Kampala-Gulu road, north of Kampala city. The term “community” is used here to refer to a cluster of villages from which the participants were drawn. The pronounced feature in this community was a trading centre called Dekabusa. With several small-scale businesses, retail shops, saloons, local clinics, makeshift video hall, bars, food stalls and makeshift structures, Dekabusa trading centre constituted a defining feature of this community. At the food stalls, under a tree, a group of youths, and often adults, regularly gathered to play a board game (ludo), which provided a major pastime for many youths and middle-aged adult males. Every evening, a generator-powered television set in the makeshift video hall relayed videos of African films and occasionally football matches of major European leagues. The activities in the trading centre attracted people from various villages who came to drink alcohol, play ludo, buy a few household items or simply to meet friends.

Most adults in this area owned a mobile phone and a radio, which helped to keep them informed and in touch with friends and relatives outside the community. The main means of transport, motorcycles, kept the community more accessible despite the poorly maintained feeder roads. The majority of the residents described themselves as smallholder peasant farmers, although some owned shops or small businesses to supplement their incomes.

151

Although a significant proportion of the people regarded themselves as native Baganda5, the influence of migration and inter-marriages contributed to the community’s ethnic diversity. The majority spoke Luganda, the native language of the Baganda. Recent estimates put the district’s total population at 405,900 in 2009 (Luwero District Local Government, 2009).

Data collection Although the three first authors all carried out research in Luwero between 2012 and 2014, the ethnographic data for this paper were collected mainly by the first author and results are presented in the first person when they derive directly from his fieldwork experience. In line with calls for micro-level qualitative studies to understand; everyday social processes that generate social capital (Campbell & Gilles, 2001); how people use microcredit services (Morvant-Roux et al., 2014), and different forms of elite capture in participatory programs (Lund & Saito-Jensen, 2013), we adopted a broad exploratory ethnographic approach. This approach enabled maximization of “possibilities of obtaining data and leads for more data…where and how to locate a rich supply of data” (Glaser, 1978;45). This involved being present and listening to conversations in bars, playing ludo, attending religious gatherings, participating in burial activities, playing with children, and attending community meetings. During these activities, I documented how people cooperated, observed who participated in what activities, and witnessed community conflicts, for example when VSLAs members engaged in verbal exchange when sharing out savings.

Data gathered through participant observations and informal interactions were triangulated through interviews and focus group discussions (FGDs). Participants would broadly talk about community problems such as poverty, ill-health and cooperation. When it became clear that social groups and informal networks were popular means of cooperation, subsequent observations, interviews and discussions focused on explanations of these processes and how they unfolded. For example, people initially mentioned several ways in which they helped each other through social groups such as burial groups, VSLAs, church groups, etc. We also found informal networks that linked individuals across social space. Our attention was drawn by the popularity of VSLAs as spaces that connect people in the community. The more we studied

152

VSLAs, the more we heard about endemic cheating6. Initially it was not clear how cheating affected bonding in the community or what exacerbated cheating in the VSLAs. Through participation as a member of a VSLA and deep hanging around in the community, it became clear that cheating was embedded in everyday social processes and traditional norms of goodness (obuntu bulamu), and that attrition took place in the VSLAs. Throughout the paper, we analyze these processes that give cheating and subsequent “resistance” practical relevance as explications of community-level bonding.

Subsequently, a total of 91 in-depth interviews (36 with males, 55 with females) were conducted with adult community members, aged 18 and above. They lasted 30-60 minutes. We ceased to sample participants after saturation was reached. In addition, 42 FGDs were conducted (23 with females, 15 with males and four mixed). Each FGD had between 10-12 participants, purposively recruited. Seven key informant interviews were conducted with representatives from the local government and NGOs operating in the area. Participants for in-depth interviews were selected following a theoretical sampling procedure. Sampling decisions were based on the emerging categories and themes in the data.

To ascertain how groups functioned, I joined a VSLA, called Bajja Basaga, participated in the weekly savings meetings, from January to December 2013. I saved money, documented members’ interactions during meetings, the type of information shared, and how meetings were conducted. I also observed and listened to what members and non-members said about VSLAs in their everyday interactions. I kept a diary of group meetings, and this provided points for reflection on what was going on in the community and VSLAs, which were in turn discussed with the rest of the research team in the field to help make sense of the data. This iterative approach allowed initial analysis to take place during data collection as emerging ideas were further explored in subsequent interviews and discussions.

Data analysis All interviews and FGDs were conducted in the vernacular Luganda, audio recorded and transcribed in English. Transcriptions were stored on a password secured computer and only shared with the research team. The transcription process was iterative and, reflections on

153

emerging issues were followed up in subsequent interviews. Transcriptions were imported into NVivo10 for coding and analysis. Key themes that emerged were reasons for joining VSLAs, benefits and challenges of VSLAs membership. Through a process of refinement sub-themes were identified: cheating and the creation of offshoot groups. When a theme was identified, a query search was done in the entire data set and text related to that theme was examined further to ensure that all the key ideas were covered. This process continued until saturation. These themes then formed the analysis and writing process.

Ethical statement The study received ethical approval from Ethical Advisory Board of the Amsterdam Institute for Social Science Research at the University of Amsterdam, the Institution Review Board of Higher Degrees Research and Ethics Committee of School of Public Health, Makerere University, the Uganda National Council of Science and Technology (UNCST) and the research desk of the Office of the Presidency, Uganda. Permission to carry out the research was also granted by the Chief Administrative Officer of Luwero District. Informed written consent was received from participants before conducting formal interviews and FGDs. Oral permission was always sought and received from the key organizers for observations during communal events. Pseudonyms are used in this paper to protect the privacy of participants.

Results Becoming a member of a VSLA- Bajja Basaga There were about ten VSLAs in the community, meeting weekly. Early January 2013, through a friend, Mr. John, who as a leader of Bajja Basaga, knew that three people had left the group arranged for me to replace one of them. On my first day as a member I was given the savings book of the member I had replaced. There were 25 members in total. From that day on, attending Bajja Basaga meetings became my weekly routine and I always looked forward to hearing stories about the goings-on in the village that were shared during the meetings. Members endeavoured to attend in person, and those who could not either registered their apologies through the chairperson or sent someone to save for them.

154

During most of the Bajja Basaga meetings I attended, it was clear that VSLAs provided opportunities beyond the financial benefits for members. January 30, 2013 was a normal routine VSLA meeting. I arrived 20 minutes late but found that the meeting had not yet started. Five members were present. As was the custom in these meetings, people arrived one by one until it was clear that no additional members would join. As we sat together waiting, we talked about events in the community. One of the members said, “My friends we need to contribute money for Wasswa’s [a community member] house which got burnt.” She requested all of us to contribute. I contributed UGX5007(USD0.17). I asked her what had happened and she said, “Wasswa had just returned from the bar, he knocked over a lamp that sparked fire. The Jerri cans in his house made it easy for the fire to spread. I decided to get members to contribute since we are already in this group.” Those who had not contributed made pledges and implored others to do the same.

However, as she collected the money, some members expressed concern about theft: stories of people “disappearing” with others’ savings had featured prominently during interviews and informal discussions. I learnt days later that the money was never delivered to Wasswa but no one including those who contributed appeared interested in following up why money did not reach Wasswa. This corroborated stories that I heard during interviews and FGDs about “cheating” in groups.

Cheating in groups, “goodness” (obuntu bulamu) and VSLA leaders VSLAs operated in such a way that money saved and interest earned through loans to members at an interest rate of 10% per month was shared at the end of the year. December was the month that most VSLAs savings were shared and distributed based on how much each member saved and accrued interest. Sunday December 22, 2013 was a day set for the Bajja Basaga meeting to share accumulated savings. This meeting had been postponed at least twice over the past two and a half weeks due to lack of money. The meeting that was scheduled to start at 10am was delayed until 3pm. When I asked the treasurer of the group why, she said that some people who had borrowed money had promised to bring it that day but had not. Everywhere I went people told me how bitter they were about the manner in which their savings were being “eaten” by members who fail to pay back borrowed money. “We have been cheated so much,” Loyce told

155

me visibly frustrated as I sat in her shop to chat. “I can never join those groups. You can see how they [the members] are crying; they have been cheated… I am far better than most of these people crying,” said Jane, who had come to buy groceries from Loyce’s shop. Jane, in her mid- fifties, had become one of my close friends. On several occasions, whenever we conversed about groups, she would tell me how she disliked savings groups because of the cheating.

As I strolled in the trading centre killing time until the meeting started, I kept meeting people who were preoccupied by the loss of their money. I learnt that another VSLA had postponed sharing the money to January 2014 because the borrowed money had not been returned. “I will always keep my money in my own pocket, no more saving groups,” said the local council chairperson from a neighbouring village, who looked visibly distressed. Those nearby laughed sarcastically and some said, “ehh, even you chairperson?”. They could not imagine the chairperson being cheated. At the meeting venue the chairperson and the treasurer told us that our savings were less by UGX2,000,000(USD674.94). As suggestions on how to share the little that remained were being discussed, two female members, Loyce and Regina, almost fought over what each should take. There was no formula for calculating what was due to each member. The chairperson haphazardly developed a formula and distributed the money, even though he was himself a defaulter. After the meeting, Loyce told me that she was planning to leave the group, so were Regina and five other members.

Regina and other five left the group. Membership of Bajja Basaga dropped from 25 in 2013 to 18 members in 2014. However, Loyce remained in the group and reasoned that with time things will improve and cheating will reduce. She and several others who keep in the VSLAs despite the cheating often cited reasons such as lack of alternatives accentuating the popular notions about the inaccessibility of mainstream banking and microfinance services. For others it was about the ease in accessing loans in case of emergency. For some, it was financial independence, which meant that they were, with some level of certitude, inclined to remain in the VSLAs. Take for example Nalunkuma, a female aged thirty. After relocating from Kampala city in 2007, her husband (non-VSLA participant) opened for her a saloon business but she had no control of the money generated from the business. When she decided to join the VSLAs she was never worried about the cheating.

156

I joined the savings group because whenever I want the money, no one will ask me what I am going to use it for. They (group) can only tell me that “maybe today Nalunkuma will not take money because you (mentions interviewer) want it”, or when the money is insufficient but not asking me what I am going to use it for.

Given that majority of VSLAs participants were females, Nalunkuma’s revelation could also be seen in the context of the patriarchal society. For members like Nalunkuma, it was about the feeling of autonomy and control. For others, cheating erased that independence and would prefer money in their pockets especially males. What is critical is that even when members acknowledged cheating in VSLAs, addressing the problem seemed to be the greatest challenge. Discussions with community members revealed two main issues. First, people who cheat are either group leaders or friends of group leaders who take advantage of their positions. Second, members who cheat feel they cannot be punished because they are part of the community. The latter relates to the pressure of maintaining obuntu bulamu or harmony in the community, which meant that the defaulters could not be punished. Below we expound on these two problems and their effect on social bonds.

During the discussion at Loyce’s shop before the meeting, the chairperson from the neighbouring community told me, “The problem here is people fear confronting the members who cheat because they are also members of the community: we do not confront them since they are our own people (bantu baffe).” His point was that community members are so inter-linked that if one community member confronted another then he/she would be considered a bad person by the rest of the community and as someone who falls short of buntu bulamu. For him, the idea of obuntu bulamu is the “engine” that drives and maintains community bonds and the feeling of togetherness during crises such as deaths. Group members do not wish to antagonize their social relations and therefore choose to maintain obuntu bulamu.

Not everyone agreed with the chairperson’s view: they reasoned that people who cheat are devoid of the ethos of obuntu bulamu. Concerns about erosion of obuntu bulamu featured frequently during FGDs and interviews.

157

There is no more obuntu bulamu. People no longer trust one another. In earlier days, one would buy land even without formal agreements but today there should be witnesses and written agreements otherwise after a few years someone can tell you to leave his land, saying “I do not know you”. (Participant, FGD with men May 10, 2013).

Cheating in VSLAs also illustrated erosion of obuntu bulamu. Some community members argued that community leaders and their friends have taken advantage of their position and a general erosion of obuntu bulamu to cheat in VSLAs. When VSLAs were initiated in Luwero, Plan Uganda and Community Vision8 approached local leaders (local council chairpersons) first, to mobilise community members for meetings during which the VSLA approach was introduced. Those who were interested formed groups of 15-30 members. Initial training and support supervision for about one year and six months was provided to constituted VSLAs. At the beginning all VSLAs selected a management committee comprised of a chairperson, a treasurer, and a secretary as required by the model. The leaders were responsible for the day-to-day management of the VSLAs including keeping records, money and convening meetings, during which members built their fund through minimum weekly saving contribution of UGX1,000(USD0.337) and a maximum of UGX5,000(USD1.687). VSLA leaders were literate and had some level of education, and held other leadership positions such as being members of village health teams, local council executive committee members or church leaders. These positions held prior to the initiation of VSLAs meant that they already had some degree of influence in the community. Because leaders were at the fore of mobilizing other members to form and join groups, they were often perceived as the ones who brought them to the community.

Leaders maintained their influence in the VSLAs and their operations. Whenever there was training, they represented the group even when other members were not aware of them doing so. From June 3 to June 7 2013, a week-long refresher training supported by Plan Uganda brought together people from multiple VSLAs in Luwero. For Bajja Basaga, only the chairperson and the vice-chairperson attended the training. When I asked fellow members if they knew about the

158

training, no one did, and no one ever discussed the training during our savings meetings. I found out about the training during a meeting I had with the Health Assistant at Luwero sub-county. Information hoarding by group leaders appeared to be a common way of maintaining influence. When Bajja Basaga registered with Luwero district local government in October 2013, the chairperson of the group, who had been informed by his contacts at the local government that it would be easier to attract other partners once the group was registered, spearheaded the process. This was the case when an international NGO called Ecumenical Church Loan Fund International (ECLOF) with a branch in Luwero district visited us looking for organized groups to give members loans, at a low interest rate. When I asked the chairperson, he told me that one of the reasons he quickened the process of registration was to target ECLOF money. Apparently this information had been relayed to him through his contacts that ECLOF was looking for registered groups to give them loans at a low interest rate. No other member of Bajja Basaga knew about ECLOF’s offer. The chairperson knew about ECLOF but kept the information to himself. The members of Bajja Basaga believed that the chairperson had brought ECLOF, and this made him important and the group dependent on him. When he defaulted in December 2013, no member questioned him. I later learnt that he had defaulted on three previous occasions.

Evidence of cheating in VSLAs by the leaders and their close associates was corroborated during FGDs and in-depth interviews as excepts from an FGD with male residents reveal below.

Moderator: Has cheating ever happened? All: Yes R2: Just here in Kalembwe (neighbouring village), not even a month has passed, the treasurer of the group kept all the money that members were saving. He used the money for his own needs and when the time came for sharing, he did not have a coin. Even the time they gave him to mobilise the money passed and there was no money. At the end of the day, every member got about UGX70,000(USD23.62) less than what they expected. Until now he tells them “I will save and pay”. …and such cases scare away those that would have joined.

159

In an interview with Loyce, a female resident aged 48 and a member of a VSLA group, she revealed how one of the members who borrowed money from their VSLAs and failed to repay could not be sanctioned because he was a friend to the chairperson of the group.

The [defaulter] even put his piece of land as security9 and said we should sell it, but when we got buyers, he quarrelled so much and at the end of the day he did not give us the money and we ended up not selling the land...you see this is a community leader and a close friend to the chairperson of our group…so the chairperson in away also protected the defaulter.

What Loyce and other community members talked about could also be seen in relation to the interrelatedness of the community. Leaders of the VSLAs were also part of the community with spouses, children and relatives in the community. Punishing them would also mean risking being seen as punishing their close kin. Our discussions with people like Loyce and everyday observations revealed the underlying differences in the community relations between people generally regarded to have influence (abawaggulu) and those regarded as having limited or no influence (abawansi). This community differentiation existed in everyday socialization but became prominent when in externally initiated interventions.

Having influence (abawaggulu) and not having influence (abawansi): community differences on bonding People use the terms abawansi and abawaggulu to describe a degree of social difference in the community. Abawaggulu usually know how to read and write, possess some assets and have contacts in far places like Kampala, the capital. Take an example of Sande, a male aged 55 educated up to ordinary level of education. He became a local contact for NGOs like Plan Uganda and AMREF. He was also a member of the village health teams (VHTs). His home is used as an immunisation outreach post conducted once a month. Whenever there are conflicts in the community, he is the first one to be consulted as was seen when a community member was accused of witchcraft. He is also a chairperson of the parents teachers association of the only primary school in the community. He owns a big chunk of land in the community. He has

9 160

contacts beyond the community. When external programs are initiated such as VSLAs groups, abawaggulu generally assume leadership positions. This gives them additional power to control the group’s activities and contacts with an NGO supporting the group or a government office. They use all options, including how they distribute benefits to the community, to keep this power. Sande influenced the selection of his son, Senje, as a member of a VHT because of the perceived benefits. His wife was a treasurer in one of the VSLAs in the community. It is no wonder that friends of such VSLA leaders are usually in a better position to get services. Take Zimula, male aged 35. His association with the chairperson of the village made him a “de facto” beneficiary of the agricultural inputs distributed under government-run National Agricultural Advisory Services program, which the chairperson headed at the village level. This appeared to have been a reward when Zimula nominated him a few years ago to become the village chairperson. Bonding between the leaders and their friends is further harnessed in the group and extends to protecting the friends when they borrow money and fail to return it. However, the concept of abawaggulu is sometimes fluidly used, and the clear demarcation of who are the actual abawaggulu can be blurred. Not many people wished to be identified as such for the obvious reason that people often associate abawaggulu with “eating”, connoting some degree of privileged position that usually assumes one is above others. Abawaggulu can also imply people at higher authority levels including staff in local government and political heads at the national level.

Abawansi are usually illiterate, poor, timid and hold no position of influence in the community. They are made to always look up to abawaggulu for support services and information on what is happening outside the community. When externally initiated groups are formed, the abawansi are mobilized to join. They may have expectations of the group, but when these are not met, they do not openly question the leaders, but rather grudgingly accept what they get (or do not get). In such circumstances, it was hard to enforce sanctions on VSLA cheaters. When VSLA members realize that they cannot enforce sanctions on cheaters, two things happen: first, people lose trust in the VSLA groups; second, because they want to keep obuntu bulamu, they opt to start their own loose groups. The idea that people form loose groups to keep obuntu bulamu contrasted what we alluded to about its erosion. Whereas the general feeling was that obuntu bulamu eroded, a proportion of the population still made efforts to preserve it even if it meant

161

circumventing confrontation with cheaters in VSLAs. On several occasions, I witnessed people act solidary during communal events such as burials to be seen as possessing obuntu bulamu. This appears to differ from Allen’s (2006:67) observation about “spontaneous replication without external support” of VSLAs based on the assumption that people learn the VSLA model and apply it to start their own VSLAs. Instead what members in this community did was start off-shoot groups in opposition to the “cheating” in VSLAs.

Creation of offshoot groups to avoid cheating in VSLAs: Namu, Mildred and Agnes’s story It is a common precept that people find it easy to bond and associate with those who possess trust. Trust is an important ingredient in social life, and provides a basis for people to join VSLAs groups, stay or leave them. Although people may not always leave a VSLA group because of cheating, as Loyce, Nalunkuma and other stories show, others do because they feel unable to enforce any form of accountability. Regina left, together with five other members of Bajja Basaga at the beginning of 2014. When I met her early 2014, she was looking for a VSLA to join hoping that “may be things will be different.” Those who left, some gave up on savings, others crisscrossed between VSLAs while others started offshoots.

Offshoot groups are informal loose groups, which members referred to “our arrangements,” established by people who were either cheated or scared off VSLAs due to cheating. Their loose nature meant that it was difficult to establish the exact number created. Estimates from discussions with community members indicate that the offshoot groups ranged between five and seven in total in a village with a membership of four to six persons. One main characteristic of the offshoot groups was not only their informality but also the manner in which members contrasted their operations from VSLAs. Members always emphasized how “they did not function as other groups.” Members saved whenever they received money, sometimes following agreed procedures. Unlike VSLAs, offshoot groups did not involve lending and borrowing to members. Savings varied from as low as UGX500 (USD0.17) to as high as UGX10,000 (3.37). Members received their savings at an agreed date, usually after six months or, as was always the case, when need arose. Members of the offshoot groups knew each other or were related kin or shared similar activities such as making mats. Offshoot groups were kept as a secret from non- members. It was interesting how through deep hanging around offshoots were discovered. Unlike the VSLAs that had names like Bajja Basaga, offshoot groups had no particular names. 162

Participants preferred to call them “our arrangements that are different from VSLAs.” There were no fixed days for meetings but only when it was necessary. Each offshoot group had a treasurer who kept the records and the money of the members. Below we provide two examples of (Namu, Mildred and Agnes) offshoots groups rich in trusting relations.

I first met Namu during an FGD in which she voiced her opposition to VSLAs. We became friends that she would call me on the phone if she had not seen me for a while. Namu is 50 years old, studied up to primary six10 and is a widow. She moved to Dekabusa when she married in 1985. When her husband died about five years ago, life became hard and she relied more on her sons, one of whom became my friend during fieldwork. Through regular interactions, Namu told me she had initially wanted to join groups to save money and receive the support that group members shared. However, scared off the VSLAs by the cheating, she, along with female friends in the village, started their own loose savings group. This group, which Namu says has no name, was started early 2012 with five members who were still active during my fieldwork. Namu was chosen as a group treasurer. Whenever a member got money, it was given to her and details recorded in a book. She was hesitant to call it a group, because, she said, it did not operate like other groups she had seen. But, she was quick to say that it operates on the basis of trust. With the exception of Namu, other members of this loose savings group had previously joined a VSLA but withdrew after being cheated. As she explained:

… other members were in (VSLA) groups but when they cheated them they feared, got frustrated and they consulted me,…and they said that “let’s start our own group” and that’s how we started…we don’t have a name for it… but our book records show that so and so brought this amount of money which they entrusted me to keep. They first tested me in the first year. When they realised that I did not cheat them, even the second year they trusted me to keep it. I even asked them to get another person, but they refused and said “we want you,” some of them call me “mother.”

Namu’s experience contrasts the routines of VSLAs but rhymes with flexibility of rural community life. When Namu talked about saving money any time a member gets it, emphasis

163

was placed on the flexibility with which their loose savings group operated compared to constituted VSLAs. “If I get money today and I feel like saving it, I do not have to wait for a particular day” she said. A member could save any amount of money unlike in VSLAs that had a fixed minimum saving amount of UGX1,000(USD0.33) and maximum UGX5,000(USD1.68) per week. The members’ saving could be accessed as and when need arose.

This year our member got a problem. Her husband was jailed, he knocked someone on motorbike, yet he had no money. She approached me as an individual and said “lets share the money.” After consulting other members they accepted to break the container and each got their money, after we began a fresh.

Although Namu felt that people in VSLA groups were cheated, she was also quick to observe that her limited financial inflows would not have allowed her to make regular savings in a VSLA. Namu decided to keep away from any VSLAs that had strict saving routines.

When she said some of the members refer to her as a mother, she was emphasizing the bond members have for each other. In traditional Buganda, the mother figure represents not just a bond but also respect and trust. This means that when starting savings groups, the interested individuals look for features in potential members that relate to familiar characteristics. The members of her group knew each other before they started their loose savings group. The catalyst was the cheating in existing VSLAs.

The story of Mildred and Agnes depicts a similar pattern. Mildred and Agnes, both residents of the same village, started an in-house savings group in which they saved as members of the same kin. Mildred and Agnes had never joined a VSLA group. They feared being cheated in the VSLAs, which inspired them to start their own savings group. The members of Mildred and Agnes’s group were involved in making mats which earned them between UGX2,000(USD6.54) and UGX15,000(USD4.90) per month. This gave them a common interest in addition to the kin relations that the members shared. They saved the money they earned from selling the mats. Like Namu, Mildred and Agnes’s group did not have a name. Outside the members, no community member knew about this group and how it functioned as they revealed during the discussion;

164

Moderator: Tell me more about this arrangement… Agnes: Yes, we don’t cheat ourselves, Mildred: It’s ours Agnes: But those other groups (VSLAs) can cheat you. There are people who saved in some group (VSLAs) which I know, but when it was time to get the money they did not get and the group even died out. Moderator: What made you start this group? Mildred: We wanted to have somewhere to save from because you cannot do something valuable out of that one mat, but if I get like Ush 15,000 in a month, I can do something with that money. Agnes: And like the other groups (VSLAs) that we told you about they have their problems…they despise, cheat you…. We are not like them (VSLAs)…

Mildred and Agnes attributed their susceptibility to cheating in VSLAs to their religious beliefs, Seventh Day Adventist (SDA), which emphasizes prayer for those who cheat or exploit others. They mentioned that none-SDAs have the option to consult native doctors who can punish cheaters. Mildred: Like for us here they know we are not supposed to abuse or do anything, they even say “a thief who steals at an SDA person's home is not scared of any consequences,”…for us when they steal anything you just call others and pray. Will a person be scared because you prayed? Agnes: They fear to cheat others (none-SDA) because they can easily consult traditional doctors, but for us we do not engage in those things.

Whereas not all the six members of this offshoot group were SDAs, they shared kin relations. Where social trust is central, members are careful not to allow those they cannot trust into their loose savings group. This is also partly a precautionary measure against cheating in VSLAs. It is also because determining trust is a gradual process. The members took pride in this trust and boasted of how this arrangement was insulated from cheating. Mildred and Agnes compared their arrangement with the operation of VSLAs, and mentioned that they were drawing from the weaknesses inherent in the VSLAs to start their informal loose groupings.

165

Discussion and conclusion The growing interest in community groups as “building blocks” for social capital has generated interest in the social science and social capital literature. Although studies have analysed the potential of community groups for building social capital, examining the social processes inherent in communities and how these processes “meet” with the externally initiated groups has lagged behind. After describing the VSLAs approach, we examined the concepts of obuntu bulamu, and offered a glimpse into the existing community norms in Luwero District that often sustain and challenge asymmetries of power in the context of externally initiated groups. In contrast to most statistical and quantitative studies on social capital, our case results contribute to debates about how a microanalysis of community dynamics and processes provides a useful framework for understanding consequences of externally initiated interventions.

From our case results, three conclusions can be drawn. First, we found that VSLAs served to reinforce existing community differences. When VSLAs were introduced in Luwero District, local leaders were the first contact in the community. Because these local leaders are already defined as powerful in the community itself, they maintained influence in externally initiated groups. One of the dangers of their influence was that it undermined the capacity of members to impose sanctions and curb cheating in the VSLAs. As we have documented in the results, the processes that allow cheating to continue and affect accountability lie within the experiences and realities of the community itself. The internal dynamics of various VSLAs mirrored existing power relations between group leaders and ordinary members, fueling untamed cheating. This is because when interventions are initiated, they do not operate in a vacuum, but rather are “regulated” within existing norms of obuntu bulamu that often times enhance, as much as they defuse the power structures and strengthen people’s informal networks. In some ways, this corresponds with the rich literature on patronage in different contexts in Africa such as gatekeeper politics in South Africa (Beresford, 2015); community development associations in various countries (Lund & Saito-Jensen, 2013; Titeca & Vervisch, 2008; Vervisch & Titeca, 2010; Vollan, 2012); and, peace and governance in Sierra Leon (Labonte, 2011). Platteau and Gaspart (2003:690) also note that “community-based development risks creating and reinforcing an opportunistic rent-seeking elite…” Therefore, in line with patronage literature, VSLAs reinforced existing power asymmetries within the community. In some ways, this finding is

166

nothing new as the above studies suggest. However, when we apply this knowledge into how VSLAs participants who felt a deep sense of “cheating” circumvent the powers inherent in VSLAs initiatives, different dynamics emerge.

In the debates on microfinance, and VSLAs in particular, our case results can be understood as a response for “a call for in-depth and longitudinal empirical investigations that carefully characterise forms and outcomes of elite capture” (Lund & Saito-Jensen, 2013:111). Whereas elite capture is indeed an endemic problem in several parts of rural Africa, our case results reinforce the need for studies to pay close attention to how power shapes the emergence of social bonds (Grischow & McKnight, 2008). This is also important because power is not static. In communities like the one we studied where threats of sanction are low, partly due to inherent norms that reinforce feelings of powerlessness, elite control is likely to prevent opportunities for VSLAs participants. As indicated in the literature on patronage relations in Africa (Labonte, 2011; Lund & Saito-Jensen, 2013), people with no power tend to fear losing the few benefits they get should they confront the elites. We found this evident in this community as well. Our analysis of VSLAs suggest that the formation of offshoot groups is a cautious attempt to circumvent or break free of patronage relations. It’s a form of “keep distance” strategy (Bourdieu, 1989). When this happens, new social bonds are enacted outside constituted VSLAs. Lønborg and Rasmussen’s (2014) suggestions on when and why people exit VSLAs does not tell us what they do when they exit VSLAs, which relates to our second conclusion.

A second conclusion is that because of failure to impose sanctions on those who cheat, community members responded in two ways to the power asymmetries. First, being non- confrontational to those who cheat because the cheaters were considered as “our own people” (bantu baffe) confining resistance to traditional weapons of the weak (Scott, 1985). Second, choosing to start offshoot groups, depicting weapons of organised (Kabeer et al., 2013). The members in offshoot groups understand the community constraints on addressing cheating, and know that it would be a futile effort, so they cut out space in smaller loose savings groups, where they can shine as offshoots of a bad system. They are also able to circumvent powerful elite (Lund & Saito-Jensen, 2013) in the VSLAs. However, the challenge remains that the members in offshoot groups enjoy their new found loyalties without contributing to the resolution of the

167

problems of cheating in the VSLAs. If the argument that social capital is “a product of people’s choices” is true (Bähre, 2007:35), then what is happening in the study community is not an erosion of social capital but an indicator that it evolves under different settings. This also relates to observations that “social capital opens up some opportunities while restricting others” (Ostrom, 2000:176). Contrary to Titeca and Vervisch’s (2008) argument that groups develop bonding social capital first to benefit from other types of social capital, our analysis indicates that externally initiated VSLAs fall short of bonding, but influence the emergence of stronger bonding ties in offshoot groups. The stories of Namu, Mildred and Agnes are particularly illuminating in this regard. For the practitioners in the VSLAs movement, this particular finding is likely to be new. Whereas it is commonly asserted that VSLAs operate on the basis of mutual trust since members are usually from the same neighborhood, our results suggest that in the context of increasingly popular VSLAs, those who get cheated are inclined to form offshoot groups, at least in the part of the community we studied. The scale, pace and level of informality in offshoots shows that processes that emerge at local level occur at a much slower rate and operate different from those that emerge under external influence. The outcomes are unforeseen, and the gains appear negligible. On this basis, it is difficult to concretely know whether these processes represent isolated moments in the terrain of inequality or are significant antecedents to the later changes.

Third, given that the “tipping point” may not be easily identified since the processes are not only small, discrete but also dependent on a constant reshaping by the community norms of say obuntu bulamu, the offshoot groups can then be clearly understood in the context of existing community dynamics. The formation of offshoot groups suggest a lack of trust in VSLAs membership as people turn to their informal networks and form alliances based on the mutual bonds of close ties (Putnam, 2000). In the offshoot groups, the role of trust is not only renegotiated, it is also enhanced. Those who trusted Namu with their money were dismissive of the externally initiated VSLAs as a space for trust to thrive. Members of various offshoots we talked to emphasised how they could not cheat each other. The way Namu referred to the members of her group as her “daughters” shows the level of bonding that the members of her loose savings group considered important. In Buganda, the title of “mother” carries a significant amount of respect and a sign of strong bonds. The establishment of offshoots based on mutual

168

and strong “motherly” trusting bonds depicts two issues; first, it helped to insulate the members in offshoot groups from their perceived vulnerability experienced in the VSLAs; second, the adoption of familiar cultural norms in the operation of offshoots suggests an attempt to bring change without threatening the established norms.

Finally, with concerns about cheating in groups, the future of VSLAs remains unclear in Luwero, just like in other communities. As pointed out, in the community studied, VSLAs had been in existence five years prior to our fieldwork implying that offshoots were newly created. With increasing popularity of VSLAs, the temptation to link them to financial institutions is real (Allen, 2006) as attempts to link Bajja basaga to ECLOF show. In some countries, like Rwanda, such processes have already started. However, as the evidence from this study suggests, this process needs to be cautiously instigated. From practitioners point of view, while the tendency is to develop monitoring and evaluation indicators that measure the success of the externally initiated groups, this paper shows that it is also the emergence of social bonds in spin-offs, unintended outcomes and other consequences that should be monitored. Qualitative techniques would enable such detailed analysis especially to document what people do when they exit VSLAs to guide decisions for scaling up.

Given the limited geographical scope of this study, there is a need for caution while extrapolating the results to other settings. However, the results in this paper can be taken as a piece of evidence for practitioners and researchers investigating what happens when people are “cheated” out of VSLAs. When interpreting these results in the broader VSLAs framework, it is important to observe that even though the VSLAs may be attractive to some participants, it influences other processes of bonding. Understanding these processes is central to decision-making processes aimed at scaling up innovations to increase financial access for the poor. This paper contributes original ethnographic data and analysis to discussions on community processes that affect VSLAs and related issues such as the effect of VSLAs on gender relations. It also provides a starting point towards addressing questions such as what keeps members in VSLA groups despite the cheating?

169

NOTES

1. Ethnographic research was part of the Dutch government (WOTRO) funded programme at the University of Amsterdam. This programme, Developing Sustainable Community Health Resources in Resource-Poor Settings (CoHeRe) aims to contribute to the development of sustainable interventions that mitigate the effects of the human resource deficit in healthcare by identifying and utilising existing community resources in poor communities in Uganda.

2. Native Authority Ordinance. This was a form of regulation that empowered the local chiefs, who ruled on behalf of the colonial administration, as the British doctrine of indirect rule. The NAO therefore empowered the local chiefs to collect revenue, impose taxes and prevent crime on behalf of the colonial administration at the time. The NAO was revised in 1922 to create “communal labour” The NAO and subsequent regulatory changes gave powers to the local chiefs and helped to “retribalise” or “communalise” the communities (Grischow & McKnight, 2008). 3. Stevens (2008) used Bourdieu’s social theory to understand social capital and inequalities in health in New Zealand. Stevens’ study was conducted among three distinct deprived and non- deprived neighborhoods and concluded that social capital is not a “thing” that can be measured and assigned to individuals as a variable, but an aspect of daily negotiations.

4. Baganda are the people from Buganda, situated in southwest Uganda The norms, beliefs, traditions and customs of the Baganda are referred to as Kiganda culture. The language spoken by Baganda is Luganda

5. Cheating here included people who defaulted on their loans obligations which meant some VSLAs participants borrowed money and never returned it or returned less what they had borrowed. Most people borrowed beyond what they had as savings. One case was reported where the treasurer connived with thieves who stole the metallic box that had money and they shared the money. The group later collapsed. I witnessed a case where an illiterate member who did not know how to keep track of her savings did not know how much to claim.

6. Exchange rate as of 13/05/2015. Source: http://www.oanda.com/currency/classic-converter

7. At the time the VSLAs were started, Plan Uganda worked through a local NGO called Integrated Community Effort for People Living with AIDS. Later, when this NGO collapsed 170

Plan Uganda partnered with Community Vision, another local NGO based in Luwero to continue with the program in the different sub-counties where Plan Uganda was operating.

8. When I followed up this case, I was informed that the defaulter put his property after failing to pay but that it appeared he was simply hoodwinking the members. Even after failing to recover the money he still remained in the group. Another case was mentioned where the defaulter also put a property security whose value was less than what he had borrowed but they could not sell it anyway.

9. Six years of elementary school.

171

References Abubakari, A., Sadik, B.B., & Keisan, Y. (2014). Impact of village savings and loans associations on nutritional status of under-five children: a case study in Sissala west district of upper west region. Pakistan Journal of Nutrition, 13(7)390-396.

Adams, A., Muhammed, S. A., & Boateng-Kwakye, S. (2014). Village savings and loan associations and livelihood of people in rural communities in the Bole district of Ghana: A case of village savings and loans associations supported by JAKSALLY youth group. International Journal of Development Research 4(1) 118-126.

Allen, H.(2006). Village savings and loan associations-sustainable and cost-effective rural finance. Small Enterprise Development Journal, 17(1):61-68.

Anderson, C.L., Locker, L., & Nugent, R. (2002). Microcredit, social capital and common pool resources. World Development, 30(1)95—105.

Anyango, E., Esipisu, E., Opoku, L., Johnson, S., Malkamaki, M., & Musoke, C.(2007). Village savings and loan associations: Experience from Zanzibar. Small Enterprises Development Journal, 18(1)11-24.

Bakeera, S.K., Petzold, M., Pariyo, G.W., Galea, S., Tomson, G., & Wamala, S.(2010). Community social capital and the use of health care services in Uganda. Journal of Public Health and Epidemiology, 2(8):189-198.

Bähre, E.(2007). Reluctant solidarity death, urban poverty and neighbourly assistance in South Africa. Ethnography 8(1) 33–59. DOI: 10.1177/1466138107076136.

Bebbington, A.J.(2002). Sharp Knives and Blunt Instruments: Social Capital in Development Studies. Antipode 34, 800–803.

Bebbington, A.(2004). Social capital and development studies 1: Critique, debate, progress? Progress in Development Studies, 4,343-349.

Bebbington, A., Guggenheim, S., Olson, E., & Woolcock, M.(2004). Exploring social capital debates at the World Bank. The Journal of Development Studies 40(5) 33-64. DOI: 10.1080/0022038042000218134.

Benda, C. (2012). Community rotating savings and credit associations as an agent of well-being: a case study from northern Rwanda. Community Development Journal, 48 (2) 232–247.

Beresford, A. (2015). Power, patronage, and gatekeeper politics in South Africa. African Affairs, 114/455, 226-248.

Biekpe, N.(2004). Financing small businesses in Sub-saharan Africa. Journal of African Business 5(1):29-44. Doi: 10.1300/j156v05n01_03.

172

Bjørnskov, C.(2006). Determinants of generalized trust: A cross-country comparison. Public Choice, 130 (1):1—21. DOI 10.1007/s11127-006-9069-1.

Boissevain, J.(1974). Friends of friends: Networks, manipulators and coalitions. Basil Blackwell. Great Britain.

Bouman, F. J. A. (1995). Rotating and Accumulating Savings and Credit Associations: A Development Perspective. World Development, 23 (3) 371-384.

Bourdieu, P. (1985). Distinction: A social critique of the judgment of taste. London, Routledge.

Bourdieu, P. (1986). The forms of capital. In John G Richardson ed. Handbook of Theory and Research for the Sociology of Education. New York: Greenwood Press: 241-258.

Bourdieu, P. (1989). Social space and symbolic power. Sociological Theory, 7, 14-25.

Brunie, A. Fumagalli, L. Martin,T. Field, S. & Rutherford, D. (2014). Can village savings and loan groups be a potential tool in the malnutrition fight? Mixed method findings from Mozambique, Children and Youth Services Review 47: 113–120

Campbell, C., & Gilles, P.(2001). Conceptualizing social capital for health promotion in local communities: A micro-qualitative study. Journal of Community and Applied Social Psychology, 11, 329-346.

Cattell, V. (2004). Having a laugh and Mucking in together: Using Social Capital to Explore dynamics between structure and agency in the context of declining and regenerated neighborhoods. Sociology, 38 (5):945-963

Cleaver, F. (2007). Understanding agency in collective action. Journal of Human Development, 8(2)223-244, DOI: 10.1080/14649880701371067.

Cleaver, F. (2005). The inequality of social capital and the reproduction of chronic poverty. World Development, 33(6)893-906, DOI:10.1016/J.worlddev.2004.09.015.

Copestake, J. (2002). Inequality and the polarizing impact of microcredit: Evidence from Zambia’s copperbelt. Journal of International Development, 14(6), 743–755.

Dasgupta, A., & Beard, V.A. (2007). Community driven development, collective action and elite capture in Indonesia. Development and Change, 38(2):229-249.

Dowla, A. (2006). In credit we trust: Building social capital by Grameen Bank in Bangladesh. The Journal of Socio-Economics 35, 102—122. Doi:10.1016/j.socec.2005.12.006.

Feigenberg, B., Field, E., Pande, R., Rigol, N., & Sarkar, S. (2014). Do group dynamics influence social capital gains among microfinance clients? Evidence from a randomized experiment in urban India. Journal of Policy Analysis and Management, 33,(4), 932–949.

173

Ferland, S. (2007). The importance of different forms of social capital for health. Acta Sociologica, 50, 115—128. Fritzen S.C. (2007). Can the Design of Community-Driven Development Reduce the Risk of Elite Capture? Evidence from Indonesia. World Development Vol. 35, (8)1359–1375

Glaser, B.G. (1978). Theoretical Sensitivity. Sociology Press, Mill Valley, California.

Goetz, A.M., & Gufta, R.S.(1996). Who Takes the Credit? Gender, Power, and Control Over Loan Use in Rural Credit Programs in Bangladesh. World Development, 24 (1) 45-63.

Golooba-Mutebi, F., & Hickey, S. (2010). Governing chronic poverty under inclusive liberalism: The case of Northern Uganda Social Action Fund. Journal of Development Studies, 46(7), 1216- 1239.

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 capital and school education in rural Zimbabwe. Social Science & Medicine,58, 2119-2132. Doi: 10.1016/j.socscimed.2003.09.001.

Grischow, J. D., & McKnight, G. H. (2008). The power of social capital: Historical studies from colonial Uganda and the Gold Coast. Canadian Journal of African Studies/La Revue canadienne des e’tudes africaines 4298-128.

Hassan, R., & Birungi. P.(2011). Social capital and poverty in Uganda. Development Southern Africa, 28(1),19-37, DOI: 10.1080/0376835X.2011.545168

Häuberer, J. (2014). Social capital in voluntary associations. European Societies, 16 570-593, DOI: 10.1080/14616696.2014.880497.

Hermes, N., & Lensink, R. (2011). Microfinance: Its Impact, Outreach, and sustainability World Development 39,(6) 875–881.

Hickey, S. (2013). Beyond the poverty agenda? Insights from the new politics of development in Uganda. World Development, 43:194-206.

Imai, K.S., Gaiha, R., Thapa, G., & Annim, S.K. (2012). Microfinance and poverty—a macro perspective, World Development, 40(8)1675-1689.

Imai, S. K., Arun, T., & Annim, S. K. (2010). Microfinance and household poverty reduction: New evidence from India. World Development 38 (12)1760–1774

Izugbara C. O. (2004) Gendered micro-lending schemes and sustainable women’s empowerment in Nigeria. Community Development Journal 39 (1)72–84.

174

Janssens, W. (2010).Women’s Empowerment and the Creation of Social Capital in Indian Villages. World Development, 38 (7)974–988. doi:10.1016/j.worlddev.2009.12.004

Kabeer, N., Sudarshan, R., & Milward, K. eds.(2013). Organising women workers in the informal economy: beyond the weapons of the weak. Zed Books. London &New York.

Katungi,E., Machethe, C., & Smale, M.(2007). Determinants of social capital formation in rural Uganda: Implications for group-based agricultural extension approaches, African Journal of Agricultural and Resource Economics, 01(2)167-190.

Krishna, A. P., & Goldey, P.(2009). Social capital and its “downside”: The impact on sustainability of induced community-based organizations in Nepal. World Development 38, 184– 194.

Kyazze, L.M.(2010). Cooperatives: The sleeping economic and social giants in Uganda. CoopAFRICA Working Paper; No.15. International Labour Office. - Dar es Salaam Tanzania.

Labonte, M. T. (2011). From patronage to peacebuilding? Elite capture and governance from below in Sierra Leone, African Affairs, 111/442, 90-115. Doi: 10.11093/afraf/adr073.

Larsen, J.(2010). The social vagina: labia elongation and social capital among women in Rwanda. Culture, health and sexuality: An International Journal for Research, Intervention and Care 12, 813—826.

Larsen, L., Harlan,S.L., Bolin,B., Hackett, E.J., Hope, D., Kirby, A.,… Wolf, S. (2004). Bonding and bridging: Understanding the relationship between social capital and civic action. Journal of Planning Education and Research 24, 64-77. DOI: 10.1177/0739456X04267181.

Lønborg, J.H., & Rasmussen, O.D. (2014). Can Microfinance Reach The Poorest: Evidence from a community-managed microfinance intervention. World Development, 64,460–472

Lowicki-Zucca, M. Walugembe, P. Ogaba, I. & Langol, S.(2014). Savings groups as a socioeconomic strategy to improve protection of moderately and critically vulnerable children in Uganda. Children and youth services review, 47:176-181.

Lund, J.F ., & Saito-Jensen, M. (2013). Revisiting the issue of elite capture of participatory initiatives. World Development, 46:1104-112. http://dx.doi.org/10.1016/j.worlddev.2013.01.028.

Luwero District Local Government.(2009). Higher local government statistical abstract for Luwero district, 2008/2009. Retrieved from; http://www.ubos.org/onlinefiles/uploads/ubos/2009_HLG_%20Abstract_printed/Luwero%20LD G%20FINAL.pdf. Accessed on 23/07/2014.

Mayoux, L. (2001). Tackling the down side: Social capital, women’s empowerment and microfinance in Cameroon, Development and Change 32, 435–64.

175

Mersland, R., & Strøm, R. Ø. (2010). Microfinance Mission Drift?, World Development,38 (1)28–36. doi:10.1016/j.worlddev.2009.05.006.

Morvant-Roux, S., Guérin, I., Roesch, M., & Moisseron, J. (2014). Adding Value to Randomization with Qualitative Analysis: The Case of Microcredit in Rural Morocco, World Development, 56, 302–312, . http://dx.doi.org/10.1016/j.worlddev.2013.03.002.

Muriisa, K. R., & Jamil, I.(2011). Addressing HIV/AIDS challenges in Uganda: Does social capital generation by NGOs matter?, Sahara-J: Journal of Social Aspects of HIV/AIDS, 8, 2-12.

Ostrom, E. (1994). Constituting social capital and collective action. Journal of Theoretical Politics, 6(4), 527-562.

Ostrom, E.(2000). Social capital: A fad or a fundamental concept?. In P. Dasgupta, & I. Sergeldin (Eds.), Social capital: A multifaceted perspective (pp. 172–214). Washington, DC: World Bank.

Perkins, D., Hughey,.,& Speer, P.W.(2002). Community psychology perspectives on social capital theory and community development practice. Community Development Society, 33,(1) 33-52

Platteau, J. P., & Gaspart, F.(2003). The risk of resource misappropriation in community-driven development. World Development, 31, 1687–1703. doi:10.1016/S0305-750X(03)00138-4

Putnam, R. D. (1993). Making democracy work. New Jersey: Princeton University Press.

Putnam, R. D. (1995). Tuning in, tuning out: The strange disappearance of social capital in America. Political Science and Politics, 28(4), 664–683.

Putnam, R. D. (2000). Bowling alone: The collapse and revival of American community. Simon and Schuster, New York.

Rahman, A.(1999). Micro-credit initiatives for equitable and sustainable development: who pays? World Development 27, 67-82.

Scott, J. C. (1985). Weapons of the Weak: Everyday Forms of Peasant Resistance. New Haven: Yale University Press.

Ssewamala, F,M., Ismayilova, L., McKay, M., Sperber, E., Bannon, W., & Alicea, S. (2010). Gender and the Effects of an Economic Empowerment Program on Attitudes Toward Sexual Risk-Taking Among AIDS-Orphaned Adolescent Youth in Uganda, Journal of Adolescent Health, 46, 372–378

Stevens, C.(2008). Social capital in its place: Using social theory to understand social capital and inequalities in health. Social Science and Medicine, 66, 1174-1184.

176

Stewart, F. (2005). Groups and Capabilities, Journal of Human Development, 6 (2) 185-204, DOI: 10.1080/14649880500120517.

Thabethe, N., Magezi, V., & Nyuswa, M. (2012). Micro-credit as a community development strategy: A South African case study, Community Development Journal, 47(3),423-435.

Thorp, R., Stewart, F., & Heyer, A. (2005). When and how far is group formation a route out of chronic poverty. World Development, 33(6): 907–920

Thurman, T. R., Snider, L.A., Boris, N.W., Kalisa, E., Nyiraminyoye, L., & Brown, L.(2008). Barriers to community support of orphans and vulnerable youth in Rwanda. Social Science and Medicine 66(7), 1557—1567.

Titeca, K., & Vervisch, T. (2008). The dynamics of social capital and community associations in Uganda: Linking capital and its consequences. World Development 36 (11)2205–2222.

Van Rooyen, C., Stewart, R., & De Wet, T.(2012). The impact of microfinance in sub-Saharan Africa: A systematic review of the evidence. World Development, 40 (11), 2219-2262. http://dx.doi.org/10.1016/j.worlddev.2012.03.012

Vermaak, N.J. (2001). 'Rural financial schemes' contribution to community development. Community Development Journal, 36(1) 42-52

Vervisch, T., & Titeca, K. (2010). Bridging community associations in post-conflict Burundi: the difficult merging of social capital endowments and new ‘institutional settings’. The Journal of Modern African Studies, 48 (3), 485-511. doi:10.1017/S0022278X10000327

Vollan, B. (2012). Pitfalls of externally initiated collective action: A case study from South Africa. World Development 40 (4), 758–770. doi:10.1016/j.worlddev.2011.09.016.

Weber, O., & Ahmad, A. (2014). Empowerment through microfinance: the relation between loan cycle and level of empowerment. World Development , 62, pp, 75-87.

Woolcock, M.,& Narayan, D.(2000). Social capital: Implications for development theory, research and policy. The World Bank Research Observer,15, 225-249. Oxford University Press.

World Bank.(2000). World development report 2000/ 2001. New York: Oxford University Press.

177

PART IV

General Discussion and Conclusion Overview Getting communities in contact with formal healthcare remains a major health challenge. Several interventions have been implemented to ensure that they do. While the implementation of some interventions such as CHWs, and in Uganda, VHTs, have had some significant challenges (Turinawe et al. 2015:73-015-0074-7; Innocent 2007; Kimbugwe et al. 2014:740), nonetheless lessons have been learnt. Unlike the earlier focus on biomedical interventions, it is now almost universally acknowledged that solutions to addressing the challenges of healthcare access are as social as they are medical (Irwin and Scali 2010). In particular, following the 1978 Alma Ata Declaration, the need for community involvement, control and participation has become one of the most pronounced strategies in health promotion interventions. The challenge remains, however, of how to turn this knowledge into locally relevant interventions that take into account existing social resources. While calling for a negotiation between universalism and cultural relativism, Bukuluki advocated for the adoption of such concepts like cosmopolitan localism that accommodate global discourses while maintaining a strong focus on the local context in which people are born, live and grow (Bukuluki 2013:1-7). However it remains a rather difficult terrain in terms of how much of local and global should be adopted. Besides the difficulty is not whether something is local or global but rather whether it actually works.

A leitmotif of this dissertation, espoused in all five papers, is that social resources are critical in overcoming barriers of access to formal healthcare and improving access to formal healthcare in Luwero district. As I have attempted to argue throughout the dissertation that solutions to the challenges of healthcare access do not lie in the health sector alone (Saha 2014:618-630). Solutions also lie in the communities where people are born, live, grow and socialize and where decisions for healthcare access are made (Irwin and Scali 2010; Thaddeus and Maine 1994:1091- 1110). I have argued that social resources are part of the solution. Social resources in Luwero district were embedded in social networks, including longstanding friendships and neighborly relations. In these structures, community members borrowed money, mobilized transport, and

178

offered practical help to overcome barriers to accessing formal healthcare. Borrowing money from friends to pay for hospital costs appeared to be the main means of activating personal networks, and it was mediated through trust and the ability to repay. This runs counter to the Uganda ministry of health’s narrow definition of health resources as formal healthcare workers, medicines and health supplies, health infrastructure and health financing and sustainability (MoH. 2010). This classification by the Uganda ministry of health appears to limit health resources to biomedical, formal and clinical infrastructure. Such an attitude tends to underplay the potential of existing local approaches in addressing everyday problems which is partly attributed to the general lack of appreciation of local knowledge that exist in communities. As (Ochen 2014:239-251) show, there is limited interface between local approaches and national and international efforts in utilising local solutions to a range of problems including healthcare access challenges.Yet, as Selamu et al. argue, the tendency to focus on clinical needs and gaps masks the potential of existing local resources beyond the biomedical (Selamu et al. 2015). Biomedical approaches are often more inclined to favour incorporation rather than dialogue and collaboration with local resources and stakeholders(Ventevogel 1996). As we have seen in paper I, Uganda introduced village health teams to link communities to formal healthcare facilities. Given their links with the community, VHTs were perceived to be an embodiment of community resources. However, as we show in paper I, the introduction of the VHT strategy overlooked community perceptions and the power of formal healthcare professionals, particularly the attitudes of healthcare providers in facilitating the capacity of the VHTs to link community members with formal healthcare services. Turinawe et al. show that there were also problems at selection of the VHTs (Turinawe et al. 2015:73-015-0074-7). As a result, the VHTs strategy suffered the fate of most other externally initiated interventions: an inability to take into account and utilize existing community resources. In this sense, we argue in paper I that the creation of the VHTs was an extension of the formal healthcare structure into the community that simply reinforced beliefs about the juxtaposition of community health with formal healthcare. We argue that even when the creation of VHTs is not seen as a problem in itself, it certainly has not been effective in addressing the problems of access to formal healthcare in rural areas of Uganda, and particularly in Luwero district. In fact, as we have seen in paper I, VHTs members could not help whenever the problem was lack of money or transport to a health facility.

179

On a more general level, the findings suggest that paying attention to existing resources inform of social groups, informal transport networks and longstanding values of friendship networks is critical for understanding how communities navigate direct and indirect costs of access to formal healthcare. One of the main threats to such resources (discussed in paper II) relate to broader changes that have swept rural areas, particularly in the form of modernization, migration, and the commoditization of goods. While to a certain degree such changes are a threat to the value attached to informal norms of cooperative and collective behavior, in the community where fieldwork was conducted, community members still relied on their informal networks of friends, family and social relations in the community to access everyday help, support, comfort and information. One explanation for this is that despite these changes, there is hardly any significant infrastructure and social protection mechanisms that have been extended to rural communities. As earlier mentioned, a study conducted in Luwero district found that despite the excessive social pressures on the extended family system, they remain a credible source of support for most people to address their everyday challenges.(Muhwezi et al. 2009:109-117)

“Twekolamu omulimu”, a novel perspective on embeddedness of formal healthcare seeking? I have shown in paper II how people in the community who were perceived unable to meet costs of healthcare received support from other community members. In paper II, we argue that the processes of borrowing money, calling on longstanding friendships, and applying social pressure to help those who are unable, as well as attitudes towards biomedicine and practices of reciprocity, are as important to enabling community health as they are to cushioning the problems of poverty and inadequate and distant health facilities. Social resources are a means of surmounting the barriers that stand in the way of healthcare access (Donenfeld 1940:560-564), and activating social resources places access to healthcare in the hands of consumers. As we have shown, people often mobilized “ourselves” or “twekolamu omulimo” to address critical problems including costs of healthcare access. This support was often in form of money to meet the healthcare costs when they themselves or their children were sick. We found this a critical notion that captures the processes that bring together people for a common cause but also formed a means of social control involved in the efforts of mobilization.

180

Power and social control: VSLAs as new ways of defining inclusion and exclusion in mobilizing resources for healthcare access The findings of the study open up a critical discussion about the mobilization of resources for seeking healthcare. We have demonstrated in paper II that mobilization of resources for healthcare seeking must be seen as embedded in an everyday life where strategies for collective mobilization of resources and helping those in need is part of the social fabric. The introduction, and apparent popularity of VSLAs in these villages suggest that VSLAs are taking over the obligations that would otherwise belong in the domain of the extended family system. People are not only expanding their networks in VSLAs, but membership in VSLAs gave participants an added advantage in mobilization of resources for meeting costs of access to healthcare. These are the new domains of power and spaces of social support that programs need to pay attention to.

Contrasting health resources (VHTs) and non-medical resources: does trust matter? One important observation is that the effectiveness of resources for accessing healthcare is influenced by their origin, whether externally initiated as we have seen with the VHTs, or endogenous as is often the case with community support networks. Also important to note, especially for the externally initiated resources, is how they influence, or are influenced by, existing forms of community networks. As we have seen, for example, when the VSLAs were initiated, their interaction with preexisting social networks influenced how people made decisions to join and ultimately how the benefits are appropriated (paper IV). The unintended consequences of cheating were also due to the way in which the VSLAs were shaped by preexisting networks and community processes of general goodness (buntu bulamu) (paper V). Paper V shows that when VSLAs were introduced, existing norms of goodness meant that it was difficult to punish defaulters, which consequently facilitated cheating. As a result, those who felt cheated out of the groups started their own off-shoot groups based on preexisting norms of trusting relations. These findings relate to observations by Robert Putnam, who argued that people’s decisions to join and benefit from social capital inherent in social groups is based on the idea that trust is a precondition for social capital (Putnam, 1993). This is contrary to, for example, arguments by Woolcock (2001) that trust is a consequence of social capital.

181

The example of VSLAs shows that trust is a precondition for successful interventions and particularly interventions that seek to adapt local resources. And trust here meant simple everyday rubrics such as a sense that a particular structure exists to serve their interests. In the case of VHTs, community members perceived their introduction as an extension of the national formal healthcare structure and therefore did not perceive the VHTs as working for them. As paper I shows, community expectations of the VHTs were high at the beginning of the program but gradually waned as expectations were not met. People then lost their trust that the VHTs could provide useful help, and this was further affected by the perception that the VHTs were working under instruction from the formal healthcare system.

Resource-poor, resource-limited or resource-constrained: questioning the narrative A significant number of interventions in healthcare in Uganda is initiated with little input from the community which some studies (Ochen, 2014) conducted in Uganda have attributed to a general lack of appreciation of local approaches and knowledge . Yet often, rural communities are labeled as “resource-poor”, “resource-limited” or “resource-constrained.” The use of such labels implies that we are studying communities starting from a point of weakness. Studies often look at obstacles to various health issues, for example HIV adherence, and barriers to access to healthcare. This then implies that we look at resources in the community in order to overcome obstacles, but not as resources in their own right which can propel greater health outcomes. It also means that attention is placed more on what does not work than on the resources, strengths and undocumented opportunities that exist in these same communities. In this dissertation (see papers III and V), community members often times drew from their own resources to address some common problems. These processes tend to go undocumented. Therefore, when we use blanket labels like “resource-poor settings”, it creates an unnecessary comparison between the communities we study and a certain level of resource endowment assumed to exist elsewhere, usually in the developed world. As a result, our attention to resources becomes overshadowed by our attention to obstacles. In whose interests is it to use labels such as resource-poor, resource- limited or resource-constrained when describing communities? One possible answer is that it is easy to have a category that “fits” many communities that can be classified as facing difficulties. Again labeling communities as “resource-limited” or “resource poor” and we study them as resource poor takes us back to the same problems faced by earlier anthropologists who studied

182

“primitive societies” because of limited appreciation of the resources in those communities. And as (Lübbe 2009) shows, the discussion becomes “Us against Them”. As I have attempted to show, there is still remarkable capacity in the communities. People often said “we mobilize ourselves”, by contributing money to a needy person or using informal transport facilities to access healthcare facilities. This may not mean that they actually do this all the time, but it is interesting to note that they do rely on social resources. In paper II, I have shown how people go as far as mobilizing money to take people, otherwise unable, to hospital. Perhaps we also need to note that such efforts are small-scale, and may not be appealing if one is interested in big numbers and large-scale operations.

The other issue is that communities have different levels of resource endowments. Bourdieu outlines different types of resources that include social capital, human capital, natural, symbolic capital and various other forms of capital(Bourdieu 1989:14-25). Tautological as it sounds, a community that is rich in human capital may be poor in natural resource endowment or even social capital. Would this community be regarded as a resource-poor setting? I always hear my friends in Africa referring to western countries as lacking social collectivism and therefore deficient in community social capital. Whereas it is not my intention to pursue this logic, I would argue that such communities would be regarded as resource-poor communities by my African colleagues, who by all measures tend to get dismayed at how impersonal and individualistic people are in Europe or in the developed world in general.

Theoretical implications This dissertation has largely borrowed from the concepts of social resources and social capital. From the social capital framework we can draw some basic implications. First, the experiences of the VHTs, and their perceived limited influence in terms of contributing towards increased connection between communities and formal health providers, shows that social capital engineered from outside can have a negative impact in terms of utilizing existing resources. This is consistent with Eriksson’s study of social capital and access to health in Sweden. Eriksson found that social capital can lead to unequal distribution of investments (Eriksson 2011:5611). While Eriksson studied a developed community, in a study of community associations in northern Uganda, Titeca and Vervisch also found that linking social capital can hinder instead of 183

enabling community members to benefit from resources that come from outside the community (Titeca and Vervisch 2008:2205-2222). From our study, the plausible explanation is that interventions from outside the community fail to build trust in the community or even allow trust to emerge. VHTs were essentially seen as coming from outside the community. The assumption is always that trust already exists, and that whatever comes from outside the community should fit in naturally, or be propelled by this trust. Trust in itself is a significant resource. So while the answer is not that there is a lack of trust, it is evident that trust can be affected by programs initiated from outside the community. As time progresses, as we have shown in paper IV, community members re-evaluate their trust in the initiated programs and reconstitute into trusting networks outside externally-induced interventions.

The examples (VSLAs and VHTs) illustrated in the dissertation therefore show how resources initiated from outside can produce various results on the basis of how their interface with existing norms, networks and practices influences, and/or is influenced by, levels of trust. The findings show that what makes VHTs in the community ineffective is that community members had lost trust in their work, viewing it simply as an extension of the formal healthcare system into the community which only served to entrench the hegemonic power of formal healthcare providers over the community members. In the broader sense, this suggests the hegemonic power of biomedical systems over local and systems found outside the confines of traditional formal health care system.

Trust and power are central features of social capital. Our findings fit in the argument that for interventions to be effective there is a need to look at how trust and power are played out in the community. This critical issue was neglected during the implementation of the VHT strategy in rural Uganda. Power and trust affect reciprocal exchanges and as we have seen, the community members felt they had little influence on what the formal healthcare workers do with patients referred to health facilities. Yet, for the VSLAs, participants felt they had power in exercising their choice to leave the groups and form lasting bonds in more informal networks (paper V).

I have shown in the various papers that what makes non-medical resources effective is their influence and distribution of trust, reciprocal social networks and bonds. I have also shown that

184

VHTs, and what could generally be regarded as health resources introduced into the community, become ineffective, because their introduction is usually devoid of the existing features of social capital and the way in which accrued benefits are appropriated. For example, there was limited trust as the VHT program progressed, community members had no influence over the way the needs of the community were addressed by formal healthcare providers; and those in power largely influenced what the healthcare workers did via VHTs in the communities. We also found (paper IV) that lack of trust among members of the VSLA, due to cheating in groups, led to the emergence of off-shoot groups that appear to have had a profound impact on bonds. We also show that reciprocal social networks are at the heart of the manner in which the benefits of interventions like VSLAs are appropriated among members. The initial resources such as bonds of together, obumu and ethos of community such as obuntu bulamu also act against the effectiveness of interventions that seek to tap into their very existence. This suggests the limitation of relying on the local resources as defined by the community. However, I have labored to argue that it is rather that such resources are understood in how they operate and influence various outcomes. The theoretical framework, social capital, used in the papers in this thesis considers contextual and community-level determinants of health as critical in understanding the health of people. In particular, two strands of social capital, individual and collective social capital, have dominated debates on the subject. I have premised my arguments on both strands. From the individual social capital strand I have paid attention to the role of individual social networks, built actively by individuals, who benefit from inherent social resources (Bourdieu, 1986). I have shown in this thesis that community-based interventions become successful if preceded by a clear assessment of existing social resources, particularly social networks within which the benefits of such interventions are filtered to members (paper IV). Those networks which predate the introduction of interventions promise the best chance of success (paper IV).

For example, the emergence of community-level bonds in VSLAs shows that the emergence of off-shoot groups relates to Putnam’s idea of trust as a precondition for collective pursuit of goals (Putnam 1995) as opposed to trust as a consequence of interventions. In this case, therefore, I agree with Putnam’s conceptualization of trust. In a similar contrast, one of the reasons why the VHTs did not fair well was that there was a loss of trust in what they were doing. Community

185

members perceived the VHTs as working for the “other” and not for them, and perceived the formal healthcare providers as wielding a lot of power and influence on the VHTs. This loss of trust implies that whereas a program can start with trust, it is not easy to sustain it, and that communities are continually reevaluating their trust. Therefore, while it is critical to consider trust at the beginning of an intervention, as demonstrated in the case of the offshoot groups, and even in the initial implementation of the VHTs program, it is equally important that trust is maintained throughout an intervention’s interactions and engagements. One way that this trust can be maintained is through regular face-to-face reciprocal exchanges, which Campbell (2001) found more rewarding in England than membership of constituted groups. This study shows that while trust is critical at the beginning of any interaction, it is also critical to maintain it.

Within social networks inhere social resources. We have shown (paper II) that these social resources are critical for health promotion in communities like the one we studied where access to formal institutions is limited. The findings in this paper show that individuals acted collectively to help individuals who were considered unable to fend for themselves, using a popular phrase “we mobilize ourselves” to help. Erikson found similar results in communities in Sweden where collective social capital created supportive and health-enhancing environments for women(Eriksson 2011:5611). These observations then build naturally into the collective social capital largely associated with political scientist Robert Putnam (Putnam 1995:65-78; Putnam 1993; Putnam 2004:667-71; discussion 700-4).

The collective social capital strand has been presented in the literature as if to suggest that it is entirely different from individual social capital. In a community like the one we studied, individual networks benefit close and distant members, but more importantly, individual networks can be used to address a range of health-related problems. Eriksson (2011, page 7) argues that “collective social capital connects to the community development approach of health promotion.” This then means that “health promotion programs that build on community development” aim at building community capacity to address a range of diseases and not a specific disease or condition. I agree that in studies of community, no specific or single intervention can be detached from a range of other processes that happen at community level. It

186

is for this reason that some programs have called for a holistic approach rather than a single strand. Collective social capital is informative in this regard.

Implications for further research

One of the limitations of this study is that while we studied processes in the community that enable community members to access formal healthcare, we did not necessarily study the factors at the other end of the spectrum—the formal health facilities. While the study has shown that there is potential for existing social resources to improve formal healthcare access, it is critical, before their adoption, to understand the attitudes of formal healthcare workers. For example, learning from the attitudes of formal healthcare workers towards the VHTs in the community, it is important to understand their perception of the use of social resources. We found that in some cases, the boda boda are perceived in a negative light, which may hamper their integration. Therefore, for these resources to be effective, it is important to study processes from the other end of the spectrum.

Experience tells us that interventions must be made as simple as possible. Providing insights into existing social ills is one of the critical aims of medical anthropology. Even in the 1980s, there were concerns about the need to keep behavioral research as simple as possible (Foster, 1987). As Foster noted, researchers need to outgrow the need for sophistication and adopt interventions that address the simple, basic realities of everyday life. In fact, from this study, it is clear that public health and medical anthropology would benefit a great detail by paying attention to mundane aspects of everyday life, solving simple problems from there. We have already shown in this study that a big problem like lack of physical access to a health facility can be solved by a neighbor with a motorcycle and about 2,000Ushs (0.58USD) as money for fuel. Yet, it is not uncommon to hear interventions introducing grandiose ideas such as paying the riders to transport women to deliver in hospitals. Such interventions die a natural death because they cannot be sustained. So what we learn from this study, and medical anthropology in general, is to keep things simple and address the real needs without any increase in medical expenses.

187

Recommendations for action From the study findings, some specific recommendations can be made.

The findings of the study have shown that while the VHTs are community based and supposedly understand the communities from which they are drawn, their efforts to link communities to formal healthcare are largely hampered by the power and attitude of the formal healthcare providers, lack of equipment and facilities to do their work and general lack of trust in their services. Based on this, instead of introducing another layer such as the Community Health Extension workers (CHEWs), it is better to rethink the whole process by advocating structures which are already available and performing. One of these structures, as we have seen, is the boda bodas, which ferry people from the community to formal healthcare facilities. Given that the riders are already linking and connecting people to formal healthcare, it is plausible and makes economic sense to simply provide them with enough information and training in patient care skills rather than providing incentives to the riders to act as a link to the health facilities. However, to do this, there is a need to create an effective policy environment that addresses critical concerns such as security and safety. In addition, the boda boda riders could be equipped with basic knowledge on patient handling so they can take effective control of the process from an informed point of view.

In the same vein, social resources are critical. The study findings have shown incredible potential inherent in social groups as entry points into communities. A case in point is the manner in which VSLAs distribute benefits to members through social networks. In this case, an understanding of social resources also requires an understanding of the influence of everyday social interactions, identifying the critical nodes in the interaction process and ensuring that such nodes are given health information to spread and influence the behavior of others. Of critical importance in such a discourse is the way in which the social networks influence, and are influenced by, trusting relationships. Given that trust relationships can change, as we have seen in paper V, it is also important that interventions are aware of such changes.

From my experience and my fieldwork, I notice the value of ethnographic fieldwork in underscoring such processes. While some issues appear obvious, others escape notice. It is important that interventions at community level begin to consider an expensive but rather 188

important approach of ensuring that accounts of everyday conditions in communities are taken into consideration in any intervention implemented in the community. This is particularly important for interventions that consider the use of social resources, which often times require understanding and monitoring of everyday local processes. I found the ethnographic approach useful as it helps to uncover some of the hidden aspects of everyday life.

Conclusions Despite growing recognition that drawing from community resources adds value to ownership of interventions, in contemporary Africa, and particularly in Uganda, few, if any, draw from these resources. As we have seen with the VHTs, while the concept is good, it appears that it was initiated with little consideration of existing power asymmetries which have consequently affected its implementation and sustainability. In contrast with what is defined in the literature as social resources embedded in social interactions, most interventions appear to draw little from such resources. In Uganda today, to implement any meaningful development intervention in a community, planners and technocrats in government ministries get together with development partners to convene planning meetings, draft action plans and set up implementation committees at national, district, and local levels of government. Proposals are then drafted and sent to donors or development partners in Europe or America, and recently China, who appraise and determine the worth of the program. However, such arrangements that draw little, if anything, from community resources have been criticized as potentially inducing dependence on external actors(Thurman et al. 2008:1557-1567), undermining the usefulness of community resources by undercutting trust and reciprocal relations within communities (Vollan 2012:758-770), and facilitating elite capture and patronage (Lund and Saito-Jensen 2013:104-112; Platteau and Gaspart 2003:1687-1703; Labonte 1999:430; Fritzen 2007:1359-1375). It has also been argued that such interventions are likely to be disempowering to people by curtailing innovation and creativity in addressing everyday problems (Thurman et al. 2008:1557-1567; Vollan 2012:758- 770; Chambers 1995:173-204).

In the papers presented, a recurring theme emerges that resources do exist and that labeling such communities as resource-limited or resource-constrained may in itself be a limiting approach towards understanding the role of existing, unenumerated social resources. Therefore, in

189

conclusion, it is clear that communities are indeed resource rich and endowed in various ways. The starting point is to acknowledge that a range of resources exists. Informal networks and social groups enable people to mobilize both monetary and non-monetary resources to overcome barriers of access to formal healthcare. However, the existence of such resources does not take away the obligation of government in service provision. It simply calls for programs to recognize their existence and utilize them in designing interventions.

I have shown in the various studies that even when it is acknowledged that community resources exist, interventions draw little from them. And even when interventions such as VHTs are implemented in the community, community members still rely on such resources as their informal support networks and social groups to access formal healthcare facilities and to get in contact with formal healthcare providers. Some resources such as informal transport providers have received negative publicity and yet they constitute a critical resource for enhancing access to healthcare. I advocate a holistic approach to the dynamics of community health that takes into account the role of non-medical resources such as social groups, informal transport providers and the nature of norms, obligations, trust and reciprocal relationships that build and bind community members together. On the whole, non-medical social resources matter.

190

References

Bourdieu, Pierre 1989 Social Space and Symbolic Power. Sociological Theory 7(1):14-25.

Bukuluki, Paul 2013 Negotiating Universalism and Cultural Relativism in Peace and Development Studies. International Letters of Social and Humanistic Sciences(10):1-7.

Chambers, Robert 1995 Poverty and Livelihoods: Whose Reality Counts? Environment and Urbanization 7(1):173-204.

Donenfeld, Martin,J. 1940 The Meaning of Social Resources. Journal of Educational Sociology 13(9):560-564.

Eriksson, Malin. 2011 Social Capital and Health--Implications for Health Promotion. Global Health Action 4:5611.

Fritzen, Scott A. 2007 Can the Design of Community-Driven Development Reduce the Risk of Elite Capture? Evidence from Indonesia. World Development 35(8):1359-1375.

Innocent, Komakech 2007 The Village Health Team Strategy is a'most Innovative Community Practice'Award Winner: The Experience OFA Village Volunteer Programme in Yumbe District, Uganda.

Irwin, A., and Elena. Scali 2010 Action on the Social Determinants of Health: Learning from Previous Experiences. Social Determinants of Health Discussion Paper 1 (Debates).

Labonte, Ronald 1999 Social Capital and Community Development: Practitioner Emptor. Australian and New Zealand Journal of Public Health 23(4):430.

Lübbe, A. 2009 Us Versus them: Splitting Dynamics and Turning Points in Ethnopolitical Conflict. Journal of Peace, Conflict and Development 13.

191

Lund, Jens Friis, and Moeko Saito-Jensen 2013 Revisiting the Issue of Elite Capture of Participatory Initiatives. World Development 46:104-112.

MoH. 2010 The Second National Health Policy: Promoting People's Health to Enhance Socio- Economic Development.

Ochen, Eric Awich 2014 Traditional Acholi Mechanisms for Reintegrating Ugandan Child Abductees. Anthropology Southern Africa 37(3-4):239-251.

Platteau, Jean-Philippe, and Frederic Gaspart 2003 The Risk of Resource Misappropriation in Community-Driven Development. World Development 31(10):1687-1703.

Putnam, Robert. D. 1993 Making Democracy Work: Civic Traditions in Modern Italy. Princeton: Princeton University Press.

Putnam, Robert D. 1995 Bowling Alone: America's Declining Social Capital. Journal of Democracy 6(1):65- 78.

Putnam, Robert. D. 2004 Commentary: "Health by Association": Some Comments. International Journal of Epidemiology 33(4):667-71; discussion 700-4.

Saha, Somen 2014 ‘More Health for the Money’: An Analytical Framework for Access to Health Care through Microfinance and Savings Groups. Community Development Journal 49(4):618- 630.

192

Thaddeus, Sereen, and Deborah Maine 1994 Too Far to Walk: Maternal Mortality in Context. Social Science & Medicine 38(8):1091-1110.

Thurman, Tonya Renee, Leslie A. Snider, Neil W. Boris, Edward Kalisa, Laetitia Nyirazinyoye, and Lisanne Brown 2008 Barriers to the Community Support of Orphans and Vulnerable Youth in Rwanda. Social Science & Medicine 66(7):1557-1567.

Titeca, Kristof, and Thomas Vervisch 2008 The Dynamics of Social Capital and Community Associations in Uganda: Linking Capital and its Consequences. World Development 36(11):2205-2222.

Turinawe, Emmanueil. Benon., Jude. T. Rwemisisi, Laban. K. Musinguzi, Marije. de Groot, Denis. Muhangi, Danniel. H. de Vries, David. K. Mafigiri, and Robert. Pool 2015 Selection and Performance of Village Health Teams (VHTs) in Uganda: Lessons from the Natural Helper Model of Health Promotion. Human Resources for Health 13:73-015- 0074-7.

Ventevogel, P. 1996 Whiteman’s Things: Training and Detraining of Healers in Ghana; Health, Culture and Society. Studies in Medical Anthropology and Sociology.Het Sphinuis Publishers.the Netherlands.

Vollan, Björn 2012 Pitfalls of Externally Initiated Collective Action: A Case Study from South Africa. World Development 40(4):758-770.

193

Appendices

Summary The Uganda government, and her development partners, have implemented a number of interventions at community level to improve access to formal healthcare services. Some of these interventions such as the government policy of construction of health centers in every parish, increasing medical supplies and building a strong health infrastructure take on a biomedical perspective. Other interventions such as the training of lay persons as village health teams, the formation of HIV and AIDS post-test clubs, supporting the establishment of community health insurance schemes, and community health outreaches for immunization or safe male circumcision entail an infusion of biomedical and socially-oriented approach. Initiatives, such as community-based microfinance, have also come up, but these have often been considered to have no immediate and direct impact on access to formal healthcare services. Adoption of community-based interventions for promoting health is based on the knowledge that illness is a social phenomenon, whose management requires collective community action. It is also based on the recognition that community-based interventions tap into existing systems of cooperation and norms of interpersonal trust and reciprocal relations that inhere in community structures. As a result of this knowledge, there has been a growing consensus that solutions to healthcare access in the majority of low-income countries are as social as they are medical.

However, there is limited consensus on how to turn this knowledge into locally relevant solutions that enable vulnerable populations to access formal healthcare. In particular there is limited information on how structures and processes in the community perceived to have no direct impact on healthcare access could in effect provide the much needed solutions to problems of healthcare access. This lacunae raises an important question, do non-medical resources matter in enabling vulnerable community members access formal healthcare.? The aim of my study therefore was twofold; first, I sought to examine why interventions initiated from outside the community with a significant biomedical bias fail to link vulnerable communities to formal healthcare services; and secondly, to underscore why non-medical community resources matter in enabling community members access formal healthcare services.

194

To study these processes, I conducted ethnographic fieldwork between 2012 and 2014 in a rural community in Luwero district, central Uganda. I used participant observation in the naturally occurring activities. I conducted a total of 91 in-depth interviews with community members, 42 focus group discussions, seven pile-sorting exercises and several informal everyday interactions with community members. The process of data collection was highly iterative and so was the analysis. Data were processed and analyzed using a qualitative data analysis software Nvivo10.

Drawing on village health teams (VHTs) as an example of a biomedical-oriented intervention conceived and implemented by actors external to the community, we found that VHTs in Luwero have not been effective in linking communities to formal healthcare. At the initiation of the VHT program, community members took the advice of VHTs seriously, and the VHTs themselves were enthusiastic about their work because of the incentive-led motivations. The referrals by the VHTs to the formal healthcare facilities worked well. However, interest in their work waned and the VHTs lost the trust of community members. Their biomedical orientation created a sense of a policing healthcare system and an extension of a formal healthcare structure than a supportive resource in the community. Yet, even with the biomedical orientation, their inability to address the wider social determinants of health and a lack of support from the formal healthcare providers affected their work.

People’s longstanding friendships, membership in savings and credit associations, connections with informal transport providers and tendency to mobilize each other to help disadvantaged community members were found to be key resources in overcoming problems of poverty and accessing distant health facilities. These are what I have called non-medical social resources whose effectiveness for healthcare access lies in their ability to trigger activation of localism/grassroots initiatives to meet the costs of healthcare. Terms such as “twekolamu omulimu”, translated as “we mobilize ourselves”, were used to refer to the sense of collective mobilization of resources for meeting the costs of healthcare access especially for those members considered unable to meet their own health related costs (papers II and III). The idea of “twekolamu omulimu” unfolds a novel perspective on the embeddedness of formal health care seeking practices in rural Uganda. Community members in village savings and loan associations got support from their membership in the groups and mentioned that such savings groups are

195

taking over the obligations that would otherwise belong to the domain of the extended family. We found that to understand and appreciate the role of non-medical resources in enabling community members’ access to formal healthcare, it is critical to understand how the everyday networks of embeddedness influence the appropriation of the benefits of interventions and emerging social processes (paper IV and V).

Although these processes lie outside the traditional boundaries of the formal healthcare system, they nonetheless present opportunities for a people-centered approach towards improving access to formal healthcare. Unlike the seemingly conventionally held view that community-based interventions draw from existing norms of trust and mutual support at community level, the experiences of VHTs in the community we studied in Luwero district show that community- based medical interventions initiated externally from the community draw little from existing social resources of mutual cooperation. Moreover, it was found that in some instances, programs such as village health teams (VHTs) are ineffective tools for linking communities to formal healthcare and they potentially undermine, instead of harnessing, effective utilization of social relations as a vehicle for healthcare access.

The leitmotif of a collection of papers in this dissertation is that communities are already doing enough to access formal healthcare services and that access to formal healthcare can be improved through simple, locally crafted solutions spontaneously initiated by community members as part of the daily practices and long-standing traditions of mutual support. While some resources, such as informal transport providers, have tended to attract negative publicity, and others such as VSLAs are fraught with endemic cheating perpetuated by attempts to preserve community norms of general goodness (buntu bulamu) they constitute a critical resource for overcoming some barriers to healthcare access. As obligations for support to healthcare access begin to shift from the traditional extended family system to institutions such as VSLAs or burial groups, it is critical to begin a process of rethinking community-based strategies for promoting access to formal healthcare in rural areas. The findings of this study suggest that the existence of such social resources in the community is in direct contradiction to the labels often given to such communities as resource-limited, or resource-constrained.

196

Samenvatting Samen met partners van ontwikkelingsorganisaties heeft de Ugandese regering interventies geïmplementeerd op gemeenschapsniveau om de toegang tot formele gezondheidszorg te verbeteren. Sommige van deze interventies hebben een biomedisch perspectief, zoals het overheidsbeleid om gezondheidscentra voor elk parish (gemeenteniveau) te bouwen; het vergroten van beschikbare medische voorraden; en het opbouwen van een sterke gezondheidszorg infrastructuur. Andere interventies, zoals de het opleiden van leken tot village health teams (VHT’s); het samenstellen van HIV en AIDS na-test groepen; het stimuleren van de oprichting van een zorgverzekeringsstructuur in de gemeenschap; en hulpverlening voor immunisatie of het uitvoeren van veilige besnijdenis voor mannen brengen zowel een medische als sociale benadering met zich mee. Initiatieven zoals Mikro financiering zijn ook ontwikkeld, maar deze worden meestal niet beschouwd als direct impact hebbend op de toegang tot formele gezondheidszorg. Het opnemen van community-based interventions ter promotie van gezondheidsbevordering is gebaseerd op de wetenschap dat ziekte een sociaal fenomeen is waarbij de bedrijfsvoering collectieve actie van een gemeenschap vereist. Daarnaast is het gebaseerd op de erkenning dat community-based interventions bestaande systemen van samenwerking aanboren, evenals normen van interpersoonlijk vertrouwen en wederzijdse relaties die doordrongen zijn in de structuren van een gemeenschap. Door deze kennis is er een groeiende consensus ontstaan over het feit dat in de meeste laag-inkomenslanden oplossingen voor problemen met toegang tot gezondheidszorg zowel in het sociale als in het medische domein liggen.

Ondanks deze bestaande kennis bestaat er minder consensus over hoe dit moet worden omgezet in lokaal relevante oplossingen die het mogelijk maken dat kwetsbare bevolkingsgroepen toegang krijgen tot de gezondheidszorg. Er is met name weinig informatie over hoe structuren en processen in een gemeenschap, die niet gezien worden als directe impact hebbend, in feite de nodige oplossingen kunnen bieden. Het doel van mijn onderzoek is daarom tweeledig: als eerst tracht ik te onderzoeken waarom interventies geïnitieerd van buiten de gemeenschap met een belangrijke biomedisch effect er niet in slagen om kwetsbare gemeenschappen te koppelen aan formele gezondheidzorg; als tweede onderschrijf ik waarom niet-medische bronnen in een gemeenschap er toe doen om toegang tot zorg mogelijk te maken voor leden uit de gemeenschap.

197

Om de bovengenoemde doelen te onderzoeken heb ik van 2012 tot 2014 etnografisch veldwerk uitgevoerd in een gemeenschap op het platteland van Luwero district, in centraal Uganda. Ik maakte gebruik van participerende observatie in dagelijks voorkomende activiteiten, processen en evenementen zoals bruiloften, begrafenissen en dorpsbijeenkomsten. Daarnaast heb ik 91 diepte interviews uitgevoerd met leden van de gemeenschap, 42 focus groep discussies gedaan en zeven pile-sorting exercises. Informele alledaagse interacties met de leden van de gemeenschap waren ook belangrijke momenten in het proces van data verzameling. Zowel het proces van dataverzameling als transcriptie en analyse was zeer iteratief. Nadat alle data was getranscribeerd is dit geïmporteerd in Nvivo10 software voor verdere analyse. Voortbordurend op de village health teams als voorbeeld van een biomedisch gerichte interventie bedacht en uitgevoerd door actoren buiten de gemeenschap, hebben wij bevonden dat VHTs in Luwero niet effectief zijn geweest in het koppelen van de gemeenschap aan de formele gezondheidszorg. In het begin van het programma namen de bewoners het advies van de VHT’s serieus en waren VHT’s zelf enthousiast over hun werk, mede omdat de NGO die de VHT’s rekruteerde hen een maandelijkse toelagen aanbood. Eén van de hoofdtaken van de VHT’s was om de dorpsbewoners die zorg nodig hadden middels verwijsbriefjes te verwijzen naar de formele gezondheidzorg. De verwijzingen door de VHT’s naar de formele gezondheidszorg werkten goed zolang de VHTs hun maandelijkse toelage ontvingen. Echter, ook al deden de VHTs hun werk goed, de biomedische benadering creëerde eerder een gevoel van een optredende politie als verlengde van de formele gezondheidszorg dan als een ondersteunende functie voor de gemeenschap. Daarnaast bleek dat het gebrek van VHT’s om de bredere determinanten van gezondheid aan te pakken en het gebrek aan support vanuit de formele gezondheidzorg het werk van VHT’s beïnvloedden. Uiteindelijk verloren de VHTs het vertrouwen van de gemeenschap en nam hun invloed af om kwetsbare mensen aan de formele gezondheidszorg te koppelen.

Sleutelbronnen in het omgaan met problemen als armoede en de toegang tot de gezondheidszorg waren de langdurige vriendschappen van mensen, lidmaatschap in spaar- en kredietverenigingen, contacten met informele transport aanbieders en de gewoonte om bij elkaar te komen en hulp te geven aan kansarme mensen in de gemeenschap. Dit zijn wat ik de niet-medische sociale

198

bronnen heb genoemd. De effectiviteit van het bevorderen van de toegang tot gezondheidszorg ligt in het activeren van initiatieven om kosten gerelateerd aan het bereiken van gezondheidszorg te overbruggen. Termen als “twekolamu omulimu”, vertaald als ‘wij mobiliseren onszelf’, werden gebruikt om te verwijzen naar het idee van het collectief mobiliseren van zulke bronnen vooral voor de leden waarvan gedacht werd dat zij deze kosten moeilijk konden betalen (papers II). Het idee van “twekolamu omulimu” openbaart een nieuw perspectief op hoe de gebruikte manieren om formele gezondheidszorg te bereiken zijn verankerd in de plattelandscultuur van Oeganda. Leden van een gemeenschap die lid zijn van een spaar- en kredietvereniging krijgen ondersteuning van hun vereniging en zij gaven aan dat deze verenigingen de verplichtingen overnamen die anders zouden behoren tot de familie. Ons onderzoek laat zien dat om de functie te begrijpen en waarderen van niet-medische sociale bronnen die leden van een gemeenschap in staat stellen gezondheidszorg te bereiken het belangrijk is om inzicht te krijgen in hoe alledaagse netwerken van verankering invloed hebben op het toe-eigenen van de voordelen van de interventies en in de opkomende sociale processen (artikel IV and V).

Hoewel deze niet-medische bronnen buiten de traditionele grenzen van de gezondheidszorg vallen, laten ze wel mogelijkheden zien voor een persoonsgerichte aanpak om de toegang tot de formele gezondheidszorg te verbeteren. Anders dan de schijnbaar conventionele opvatting dat community-based interventions worden gebaseerd op bestaande normen van vertrouwen en wederzijds support op lokaal niveau, laat de ervaring met VHT’s in Luwero zien dat medische community based interventions die van buiten geïnitieerd zijn, weinig gebaseerd zijn op sociale bronnen van onderlinge samenwerking. Bovendien werd in sommige gevallen zelfs gevonden dat programma’s zoals de VHTs ineffectief zijn in het koppelen van gemeenschappen aan de formele gezondheidszorg. In plaats van gebruik te maken van, ondermijnen zij eerder het effectieve gebruik van sociale relaties als een middel om toegang tot zorg te bereiken. De rode draad van de artikelen in dit proefschrift is dat gemeenschappen toebedeeld zijn met bronnen die, mits goed bestudeerd en toegepast, essentieel kunnen zijn voor het verbeteren van de toegang tot gezondheidszorg in landelijke gebieden van laag-inkomens landen. De toegang tot formele gezondheidszorg kan verbeterd worden door simpele, lokaal vervaardigde oplossingen, die in het dagelijks leven van leden van een gemeenschap spontaan geïnitieerd worden. Terwijl

199

sommige bronnen van sociale interactie, zoals informele transportaanbieders, veelal negatieve aandacht krijgen en anderen, zoals spaar- en kredietverenigingen, beladen zijn met oplichting die in stand gehouden wordt door gemeenschapsnormen van algemeen goeddoen (buntu bulamu) in een omgeving met weinig overheidssupport, zijn juist dit de middelen die van cruciaal belang zijn voor het overwinnen van barrières in de toegang tot de gezondheidszorg. Het vertrouwen van mensen op spaar- en kredietvereniging wijst op een potentieel paradigma shift weg van het traditionele familie systeem. Zo’n paradigma shift vraagt om een proces van heroverweging van community-based interventions ter promotie van toegang tot formele gezondheidszorg in landelijke gebieden. Het bestaan van niet-medische bronnen in een gemeenschap is in directe tegenspraak met de labels die meestal gegeven worden aan zulke gemeenschappen zoals resource-limited (beperkte bronnen) of resource-constrained (beperkte hulpbronnen).

Ethical approval letters

UNIVERSITY OF AMSTERDAM 200

AISSR Ethical Advisory Board

Amsterdam Institute for Social Science Research (AISSR)

APPROVAL FORM

Date 26 October 2012

Contact person Karen Kraal

Application number: 2012-SSC-2432

Name of research: Generating and sustaining social capital in a resource constrained rural community in Uganda.

Name of principal investigator: Laban Musinguzi Kashaija (PhD) Supervisor: Prof. Robert Pool

Faculty: Humanities & Behavioural Sciences

The AISSR Ethical Advisory Board and the Faculty Ethics Committee have granted approval for this study. The decision is recorded in the official portal of the Faculty Ethics Committee.

On behalf of the Ethical Advisory Board Karen Kraal (coordinator)

201

Uganda National Council f or Science an and Technology (Established by Act of Parliament of the Republic of Uganda)

Our Ref: SS 3029 23rd January 2013

Mr. Laban Kashaija Musi nguzi Department of Social Works and Social Administration Makerere University P 0 Box 7062 Kampala

Dear Mr. Musinguzi,

RE: RESEARCH PROJECT, "GENERATING AND SUSTAINING SOCIAL CAPITAL IN A RESOURCE CONSTRAINED RURAL COMUNITY IN UGANDA" ·

This is to inform you that the Uganda National Council for Science and Technology (UNCST) approved the above research proposal on 10th January 2013. The approval will expire on 10th January 2014. If it is necessary to continue with the research beyond the expiry date, a request for continuation should be made in writing to the Executive Secretary, UNCST.

Any problems of a serious nature related to the execution of you r research project should be brought to the attention of the UNCST, and any changes to the research protocol should not be implemented without UNCST's approval except when necessary to eliminate apparent immediate hazards to the research participant( s).

This letter also serves as proof of UNCST approval and as a reminder for you to submit to UNCST timely progress reports and a final report on completion of the research project.

for: Executive Secretary 202