The role of para social workers in rural communities in Uganda: strengthening community resilience for the protection of children

: Para social workers are widely used in African nations to address inadequate capacity in the professional workforce, but there is to date very little academic commentary on the effectiveness of their role. This article considers the potential efficacy of the para social worker model in strengthening child protection at community level in Uganda. Twenty interviews were conducted with local government officers, Civil Society Organisation staff and para social workers (ten in each of two rural areas) together with four supplementary expert interviews. Insights from community resilience were used in analysis of the data. Findings suggest that the model has considerable potential to strengthen community-level protection of children in circumstances in which the operation of formal systems is limited by resource constraints and outside interventions may struggle to gain understanding and acceptance within communities. Challenges include the potential for conflicts of interest to arise and the implications of increased reporting of child maltreatment for the response of the formal child protection system, including alternative care arrangements. Given its widespread and developing usage, further research to understand the conditions under which the PSW model is most effective and sustainable in different social, economic, political and cultural contexts is essential.


The role of para social workers in rural communities in Uganda:
strengthening community resilience for the protection of children Abstract: Para social workers are widely used in African nations to address inadequate capacity in the professional workforce, but there is to date very little academic commentary on the effectiveness of their role. This article considers the potential efficacy of the para social worker model in strengthening child protection at community level in Uganda. Twenty interviews were conducted with local government officers, Civil Society Organisation staff and para social workers (ten in each of two rural areas) together with four supplementary expert interviews.
Insights from community resilience were used in analysis of the data. Findings suggest that the model has considerable potential to strengthen community-level protection of children in circumstances in which the operation of formal systems is limited by resource constraints and outside interventions may struggle to gain understanding and acceptance within communities. Challenges include the potential for conflicts of interest to arise and the implications of increased reporting of child maltreatment for the response of the formal child protection system, including alternative care arrangements. Given its widespread and developing usage, further research to understand the conditions under which the PSW model is most effective and sustainable in different social, economic, political and cultural contexts is essential.
Page 2 of 34

Background
There are currently few empirical data on the operation of the statutory child protection system in Uganda (Child et al., 2014), but evidence that poor awareness of legislation and failures in enforcement at local level create barriers to tackling child maltreatment (Ochen et al., 2017;Kaawa-Mafigiri & Walakira, 2017). Child protection attracts very little direct budgetary provision (Walakira, et al., 2017a). In a study of one Ugandan district, less than 4 per cent of cases were deemed to have received an adequate response to initial referral (Child et al., 2014). More widely, effective implementation of statutory schemes may be unrealistic in low-resource contexts, while their approach is often inimical to that of local communities (Krueger et al., 2014). Research on the influence of community environment on reporting of child maltreatment by the general population is scarce even in industrialised nations (Wolf et al., 2018). But evaluation of INGO work demonstrates the importance of local context to the effectiveness of external interventions and points to the powerful influence of pre-existing local structures and practices (War Child, 2010).
Drawing on his work in rural Sierra Leone, Wessells (2015) argues that communitydriven ('bottom-up') approaches that encourage community ownership can help promote community engagement in formal systems and strengthen local child protection arrangements (see also Krueger et al., 2014 andWessells et al., 2012).
Similarly, Walker-Simpson (2017) concludes from a review of research in Africa that community involvement is essential to effective responses.
Efforts to strengthen the capacity of communities to identify and refer child protection concerns and resolve less serious issues at community-level have included the use of Para Social Workers (PSWs). PSWs are widely used in Africa: it is therefore essential that the very limited academic evidence-base on the effectiveness of their role is developed. Based on a qualitative study of child protection arrangements in two rural districts of Uganda, this article considers the potential efficacy of the PSW model in strengthening child protection at community level, drawing on insights from community resilience.

The Para Social Worker system
Para Social Workers are one of a number of para-professional roles developed to increase professional capacity. In response to significant variations in the role of social work para-professionals (Linsk et al., 2010;Muriuki & Moss, 2016), the Global Social Service Workforce Alliance has developed competency frameworks (Interest Group on Para Professionals in the Social Service Workforce (IGPPSW), 2017). The following generic or core competencies of a para-professional social worker are identified: 'may identify children or other vulnerable populations at risk, assess needs and strengths and develop a service plan including direct support but may also include coordination of services and resources where they exist or can be brokered' (IGPPSW,12). In the study context, the PSW role was embedded in community development work and included for example HIV awareness-raising and prevention and promotion of savings groups in addition to house-to-house work with allocated vulnerable families. Much of PSWs' work aligned with that of social workers, but social workers were based in sub-county offices while PSWs worked out of their own villages or parishes and more serious cases were referred to social workers.
In a quasi-experimental study of over 500 children assessing the impact of the work of PSWs and Community Health Workers with AIDs-afflicted populations in the Côte D'Ivoire, Muriuki & Moss (2016) concluded that the model could enable significantly increased access to services and should be considered for adoption throughout sub-Saharan Africa. Evaluators of a Tanzania-based training programme concluded that PSWs must operate as part of a team including fully qualified and experienced professionals; they require authority to work effectively within community structures; the role should be formally integrated into national social services; and regular if modest remuneration is a prerequisite for sustainability (Linsk et al., 2010).
The PSW system has been in operation in Uganda for over a decade. It was initiated in response to recognition that increasing the very low ratios of professional social workers to vulnerable children was prohibitively expensive (Walakira et al., 2017b).
During implementation of the National Strategic Programme Plan of Interventions for OVC-2 between 2010 and 2016, over 6,500 PSWs were trained through Makerere University (MGLSD, 2017). The evaluation report (MGLSD, 2017) credited the initiative with being instrumental in improving identification of child protection concerns at community level, enabling children to access psycho-social support, promoting the importance of education and contributing to reductions in child labour. Children reported finding PSWs helpful in addressing some of the problems they faced. Early evaluation of the programme in Kasese district by NGOs showed increased reporting of child abuse and anecdotally suggested a reduction in prevalence (Andrews, 2017). However, the government's evaluation highlighted that coverage did not extend across the whole country and that criminal cases rarely proceeded to trial, while PSWs reported challenges relating to transport and communication. The report recommended increased in-service training for PSWs but also raised concerns about the sustainability of the system arising from the voluntary nature of the role and because operation remained dependent on external partners for coordination and oversight.

Community Resilience
A range of inter-related concepts have been used to identify and understand the collective strengths of communities, including 'community capitals' (used in areas such as rural community development in recognition that poor communities have resources unrelated to economic strengths) and 'community capacity' (concerned with communities' capacity for collective action and the use of community resources for the benefit of the community (Magis, 2010)). In the past few decades, 'community resilience' has come to the fore among a range of academic disciplines and practitioner groups in response to a shift from focusing on specific threats to building the capabilities of communities to address a broad range of challenges (Kais and Islam, 2016). Magis (2010) asserts that the notion of community capacity used in the development literature is related to that of community resilience, in particular in the emphasis on collective action and agency. However, 'community resilience' is distinctive in its focus on adaptation to change (Norris et al., 2008;Magis 2010), rendering it particularly useful for understanding social sustainability (Magis, 2010).
In the psychological literature, the notion of community resilience has emerged from work on individual resilience, which developed largely in the context of child development (Berkes & Ross, 2013). It is now pervasive in humanitarian work with children: 'Strengthening Children's Resilience in Humanitarian Action' is Principle 10 of the Minimum standards for child protection in humanitarian action (Alliance for Child Protection in Humanitarian Action, 2019), which defines resilience thus: 'Children's ability to overcome the damaging effects of adversities, their adaptive capacity to find ways to realise their rights, good health, development, and wellbeing. More generally in the humanitarian context, resilience refers to the ability of an individual, community, society or country to anticipate, withstand and recover from adversity…' (313). Community resilience has therefore been used as a theoretical lens for the purposes of this article, notwithstanding the identified weaknesses discussed below.
The literature on the adaptation of communities to adversity is limited (Poortinga, 2012;Berkes & Ross, 2013). It has primarily focused on response to disasters although there is now some attention to contexts of chronic adversity, including poverty (Poortinga, 2012;Berkes & Ross, 2013). Much of the accrued knowledge and theory applies across a broad range of contexts, but caution is required in light of the sudden onset and limited timeframe of disasters compared with chronic adversity (Norris et al., 2008).
The term community resilience is not used uniformly in the literature and there is dispute as to how the concept should be conceptualised (Patel et al., 2017). Magis' definition is used here because it was developed to cover a broad range of contexts, rather than being limited to that of disasters. Magis (2010: 402) defines community resilience as 'the existence, development, and engagement of community resources by community members to thrive in an environment characterized by change, uncertainty, unpredictability, and surprise'. She describes the recent shift in understanding of community resilience as 'paradigmatic' (404) in its recognition of the importance of active agency in resilient adaptation, which aligns with understanding of individual resilience as a process of continual adaptation to adversity (Luthar & Ciccetti, 2000), and with the focus on building on strengths rather than adopting a deficit approach (Berkes & Ross, 2013).
Berkes & Ross (2013) identify core resources for community resilience as social capital, leadership and readiness to accept change and learn, but conclude that qualities of agency and self-organisation are fundamental to enabling communities to draw on those resources to become resilient. From the literature pertinent to disadvantaged communities in South Africa, Ahmed et al. (2004) also identify the importance of leadership, together with 'solidarity and hope', 'supportive community structures' and 'selective cultural values' (390). Similar attributes are identified by Norris et al. (2008: 135), who conceptualise community resilience as arising from four sets of 'networked adaptive capacities' (emphasis in the original), while acknowledging that their definition has much in common with the notion of 'community capacity'. These are: Economic Development (including equity of resource distribution), Social Capital (including sense of community, citizen participation and leadership), Information and Communication (including trusted sources of information and skills), and Community Competence (such as community action, problem-solving skills, collective efficacy and empowerment).
Economic development is embedded in the work of the host NGO for this study (TPO Uganda) and local government community development offices holding responsibility for child protection but is not considered discretely in this article, because it is not directly relevant to the work of para social workers in response to child protection concerns.

Methods
The study was undertaken through the auspices of a Civil Society Organisation (CSO), TPO Uganda, with national reach and around a decade's experience of working through the PSW system. The author had no prior connection to TPO.
The study design aimed to capture the insights of child protection actors in three domains: local government, Civil Society Organisations and the local community, represented by PSWs. These were selected to provide insight into the interaction between the statutory child protection system, CSO interventions and community actors and to gather the views of each sector as to the effectiveness of joint working in strengthening child protection arrangements in very poor rural communities. A qualitative methodology was chosen in light of the exploratory nature of the research and in recognition of its value in understanding how child welfare systems operate in practice in low-resource contexts (Krueger et al. 2014: 53). Individual interviews were undertaken to minimise disruption to participants' work as well as gather a range of individual views.
Two rural districts (D1 and D2) in which a five-year TPO project was wellestablished were selected in consultation with the Country Director of TPO. The project involved the training and development of PSWs by TPO social workers in conjunction with the local government with the aim that the PSW system would continue once the project ended. Participants were selected and invited to take part by the district project coordinators. Practical considerations of safety and accessibility played a role in the choice of locations for interviews (primarily TPO sub-county offices) and interviewees: for example, para social workers had to travel from their villages on motor-cycles. Consequently, there is a risk of selection bias which it is not possible to assess, although the frankness with which all participants appeared to engage suggested they were strongly motivated by a desire to improve local practice. Twenty-three interviews were carried out in Uganda, comprising ten interviews in each district and three interviews with experts with national and/or international experience, including two senior TPO officials. One further contextual interview was conducted in advance of fieldwork with a representative of a different CSO with broad experience in Central East Table 1 below.

Africa. A table of participants is presented in
[ Table 1 here] Questions covered the nature of participants' role; their experience; the main child protection issues encountered; processes and mechanisms for prevention, identification, referral and resolution of child protection concerns from their perspective; how well they considered the para social worker system was working in their local area; how well they considered that local government, CSO agencies and communities worked together; any strengths, weaknesses, or challenges of current arrangements and proposals for reform they identified; and how children's voices were engaged in their work.
Interviews were recorded and transcribed verbatim before being uploaded to NVivo for analysis. Initial line-by-line coding was undertaken and codes reviewed and refined before categories developed. Memos were compiled in relation to the most populated codes and categories on an ongoing basis to aid reflection on the key themes emerging and relate those to the literature and theoretical insights.

Ethical considerations
The study was funded by the Leverhulme Trust through the British Academy. Participation was anonymised save for agreement that TPO as gatekeeper organisation would be named. All participants were provided with information sheets in advance of interview. Although they all spoke English, the information sheet required by UNCST was extremely detailed, raising concerns that some PSWs might struggle to understand it fully. A representative from TPO was available to assist; a neutral interpreter was not feasible given the remote locations.
Participants also completed a consent form to indicate their understanding of the consequences of participating and their willingness to do so. The consent form sought explicit confirmation participants understood that they were free to withdraw from the study at any time and without reason and provided the author's contact details by which to do so after time for reconsideration and for communication of withdrawal without the knowledge of TPO. No participants did so.
The possibility of felt coercion is always a consideration in research such as this, where the gatekeeper organisation is relied on for access and there may also be a sense of power imbalance in relation to a Western researcher. However, the likelihood of felt pressure to attend was reduced by PSWs travelling to the interview venues. All participants appeared eager to engage and the primary concern that emerged was the possibility of unrealistic expectations and the importance of impressing on participants before interviews that the study would not result in immediate improvements for the children in their communities.
Feedback from the TPO on the utility of the final report in relation to their practice went some way to mitigate those concerns.
A final ethical concern associated with the study related to the possibility of child protection concerns arising. The information stated that although it was not anticipated that individual current child protection cases would be discussed with participants, should details suggesting that a child is at risk of harm emerge where appropriate action had not been taken, the information would be shared with the lead child protection officer for the organisation concerned. No cause for action emerged.

Results
An overarching observation from familiarisation related to the juxtaposition of data describing the community as a threat and the source of child maltreatment with data about using community resources for child protection. TPO-CD described TPO's work as an 'empowerment approach': 'look at it as self-reliance and selfefficacy for communities'. In contrast, D1-SPO said: 'the biggest, biggest challenge is resilience…communities still tend to look at these children, if they are uncles or aunties, instead of actually protecting these orphans, they are the ones taking away their property'. Professional participants also stressed the importance of defined-term NGO interventions leading to sustained improvements in the long term. Consequently, sensitising concepts from the community resilience literature were borne in mind to build categories through 'theoretically sensitive constant comparison' (Kelle, 2007). Analysis of the data relating to the role of para social workers with the qualities and resources associated with community resilience in mind led to the construction of four core categories: Leaders for change; Information and communication; Community ownership; and Agency and selforganisation (see chart at Appendix 1).

Leaders for change
PSWs are drawn from the community leadership -D1-PSW1, for example, had been a member of her local village council. D1-S/CCDO described PSWs as 'like the commanders at the village level…If the case is small, like you are not sending a child to school, they talk to the parents. The child goes to school'. PSWs described leading by example and becoming a trusted source of advice at local level: 'We still have the parents admiring us, the way we are handling their children and the way we are teaching them how to become good children' (D1-PSW3); 'I am proud of doing my work, yeah, because the community respect me…they seek advice from me when they need help' (D2-PSW1). identified as a local problem, the sub-county council can pass bye-laws to address the issue. These are communicated to the community by PSWs to ensure all members know and respect them. D1-PSW1 described 'sensitising' a woman whose gave her stepchildren less food and more work than her birth children: 'telling her that all these children are the same'. TPO-CPS cited changes in attitudes to corporal punishment as a successful example of sensitisation by PSWs, explaining: the beauty is that when you have para social workers who are part of the traditional system and they have been trained, they will be able to educate…the traditional people…and…influence the traditional way of looking at things.
Sensitisation was a central strategy in enabling sustainable community adaptation to adversity through changing social attitudes. Although D1-S/CCDO observed that 'people don't adjust immediately' and 'it is not always easy', he concluded that 'so long as there is more sensitisation, they continue learning. You will see people adjusting what you want them to be.'

Information and communication
There was universal agreement that PSWs provide an invaluable link between isolated communities and local government institutions, strengthening the coordination of child protection work : 'the hierarchy from para social workers to the sub county to us to the district…is a strength, because information flow is clear' (D2-SW1). In D2, monthly case conference meetings were held, attended by the sub-country CDO and representatives from the religious, school and business communities. PSWs presented cases to the meetings, and participants discussed the outcome and whether the case could be closed. Cases that could not be PSWs took advantage of community gatherings such as weddings, parties and church services to make announcements and share information. Some, such as D1-PSW3, delivered advice to children in schools as to how to report concerns and how to stay safe. However, many participants expressed concern that PSWs were

Community ownership
TPO senior management expressed a clear vision of the way in which the PSW programme draws on community resources and social capital to confront challenges although, as this section demonstrates, this was not an easy vision to realise. TPO-CD explained the system was 'built on the whole community looking after a child': when we go into any community, we identify existing community capacity, existing community support structures, and we build their capacity without removing them from their normal setting.
D1-SW2 regarded one of the project's key strengths as 'that we are fully involved with the local structures. The people down there have owned that project'. A PSW 'represents the community. They are the ones who select that person' (D1-PC).
Similarly, D2-ACE stressed that 'these people are residents in their community.
They know the dynamics in the community…And they are willing to work and serve their community'. He described PSWs as embodying the 'spirit of community protection'. As members of the community they serve, familiar with community problems and dynamics from the outset and neither politically elected nor in government employment, PSWs in this study typically remained in their posts for lengthy periods, were well-known and readily trusted, enabling parents and children to approach them without fear of bias or stigma: D1-PSW4 claimed: 'People have gained peace…wherever a para social passes anywhere, you see people saying: "Hey, There is a problem in my house. Please when will you come and see me?"' However, PSWs did not all share the optimistic sense of community cohesion described above, a few citing instances in which local council officials were uncooperative or undermined their work, attributed by D1-PSW4 to 'jealousy plus ignorance'. Further, feedback at a workshop in Kampala with CSO/NGO and academic delegates to discuss the findings of the study suggests that motivation and retention are problematic in some communities.
Perhaps the most significant barrier to community ownership of child protection practice concerned the risk of corruption. Involvement of community leaders in PSW selection procedures was regarded as important to ensure community buy-in, but with the caveat that it was essential to ensure that leaders could not appoint people they knew with the expectation of benefiting themselves. Participants in both districts recounted incidents in which politicians or the police intervened to prevent cases progressing in return for votes or bribes. PSWs were threatened by the police (D1-PSW2) and families were likely to withdraw cases of defilement in exchange for money. Follow-up of cases by TPO helped to discourage such practices.
Another barrier is the limited scope for embedding child protection work within schools because of very large classes, often of over 100 children, coupled with high rates of abuse within schools. Mixed views were held as to whether better use Page 21 of 34 could be made of faith leaders, with some, such as D1-SPO, D2-C/CCDO and D1-SW2, describing considerable use of churches to disseminate key messages but some concern that some faith leaders tended to promote some harmful practices, including conducting marriage ceremonies for underage brides. Professionals considered that psycho-social support and counselling were under-valued by the poorest families, who were understandably more interested in physical resources than less tangible services.
Finally, while the limited reach of the project in targeting only the most vulnerable households within communities enables scarce resources to be carefully targeted, D2-PSW1 and D2-PSW3 considered that it had resulted in some families feeling alienated because they felt unjustly excluded from the project. Some had experienced taunting or abuse from alcoholics or in domestic violence cases and physical resistance in early marriage cases but appeared to take these in their stride: 'of course, you get scared about it…we are used to it, because I know what I am doing. I am serving tomorrow's people' (D1-PSW1).
Other concerns of professional participants related to the need for significant oversight of PSWs and for adequate capacity within the formal child protection system at local government level for robust follow-up of cases. Despite the processes to strengthen local case management systems described above, participants were alert to the risk of the community losing trust in the system as a result either of unsatisfactory resolution of more serious cases or of PSWs overstepping their remit: 'you need to continually keep reminding the para social workers of boundaries' (TPO-CD). Two participants in D2 identified problems sometimes arising when the family of PSWs were involved in a case and PSWs would reach a compromise settlement or fail to speak out.

Agency and Self-organisation
This section draws on Norris et al.'s (2008) notion of community competence as well as the qualities identified by Berkes and Ross (2013) as critical to communities' ability to utilise community resources. It highlights PSWs' contribution to community problem-solving, collective efficacy and empowerment as well as challenges to effective community action.
PSWs played a key role in organising community-level action for the protection of vulnerable children. In very poor rural areas, care of orphans and child-headed households remains reliant on neighbours providing mentorship and ensuring school attendance (D1-SPO). The PSW system effectively works to restore or strengthen traditional communitarian child-rearing practices that have come under pressure: the para social workers have actually made people get to know…the child is owned communally. Whatever the child does affects everybody in one way or another (D1-CDO).
Where children were unable to live with their birth parents, PSWs were proactive in identifying and arranging an alternative family placement within the community.
Most participants appeared unaware of the option of compulsory removal through the formal child protection system. Those senior professionals who were aware explained that the cost and logistics of attendance at the nearest Family and Children Court with Children Act jurisdiction were prohibitive for most families. But there was high resistance to resorting to babies' or remand homes far from the community because of the very poor quality of that provision and challenges for children to reintegrate into the community on leaving: 'We encourage the community. Because within the community that's where the child has been born and he knows the people, the culture, the behaviours' (D1-PC); 'The care institution comes in as a very very last resort' (D1-SPO).
Where no alternative permanent solution was available in the community, children either remained with abusive, neglectful or alcoholic caregivers with psycho-social support and indefinite monitoring or moved to live with extended family members, only for that arrangement to be no better than the care they had left. D1-PSW4's description of a case presented as successful sensitisation perhaps illustrates limited community capacity: the child defied the stepmother. The stepmother in revenge, she got a bunch of dried banana leaves and set fire in it and set it to that child. The child was totally burnt…So we entered the case. The woman was cautioned and took the child and cared for it as her very children.
Moreover, children who could not be identified as community members appeared likely to suffer from that lack of status. D2-SW1 described a case in which a child who had been thrown beside a water source was sent to the babies' home in Kampala.
Local leaders 'could not handle' the case because no-one could identify the child's family.
PSWs are overseen by local government community development staff with the ambition that 'by the time the project expires, the structures are vibrant and strong' (D1-PC). But these structures were often fragile: within the community child protection system there is no continuity...You will find like within the community there is just one person who is very active and if let's say… [the] order shifts or dies then it stops there (D1-PC).
PSWs are a voluntary workforce and this was regarded as essential (TPO-CD, TPO-CPS) for the model to be sustainable. Since they are drawn from poor communities and have many other demands on their time, their commitment must be realistic, but this could be easily compromised by the significant needs of the communities they served. Moreover, increased community awareness can create its own capacity problems. Participants (CSO-CPS, D1-CSO, D2-CSO) acknowledged an increase in reports, such as of defilement cases, consequent upon sensitisation exercises, creating additional pressure on local government structures to respond.
They were aware of the risk that if cases are not resolved effectively people will 'begin losing trust in those systems' (CSO-CPS).

Discussion
The rural districts in this study, like many similar communities, face considerable adversity from the impact of persistent poverty and changing demographics arising from high birth rates coupled with relatively low life-expectancy. These circumstances have placed significant pressure on community arrangements for the welfare and protection of children. The introduction of PSWs as part of a community development approach to address the range of challenges in such contexts is a widely-used response on which there is as yet very little academic commentary. Some of the identified characteristics of community resilience have been used in this article to examine the potential for the PSW model to strengthen community-level protection of children in circumstances in which the operation of formal systems is limited by resource constraints and outside interventions may struggle to gain understanding and acceptance within communities.
The findings suggest that perhaps the most significant attribute of the PSW model is its potential to regenerate community resources through the engagement of community members themselves in line with Magis' (2010) definition of community resilience. PSWs can provide significant additional capacity to the child protection system at little cost. The availability of PSWs has significantly increased the speed and volume of reports of child protection cases, through their proximity to community members, availability outside office hours and knowledge of local services as well as expansion of the workforce. Their authority to mediate and solve problems within families and to deal with less serious child protection cases directly eases pressure on the professional workforce. Moreover, in the longerterm actual prevalence may fall with changing community conceptualisations of child abuse, as indicated in Andrews' study (2017). As reported, some participants identified falling prevalence in some categories of abuse, suggesting the PSW programme has been effective in enabling positive and sustainable community adaptation.
PSWs in the two study districts acted as change agents, offering leadership and education to the community, providing trusted sources of information and problem-solving skills, and enabling community self-organisation. In this way the PSW model meets the need identified by Krueger et al. (2014: 53) to develop systems that are not merely culturally sensitive but culturally appropriate: that is, 'that genuinely recognise and integrate local understanding, values, aspirations and capabilities.' One of the characteristics of the African communities engaging in this study that stands out to a researcher from a very different cultural background is the enduring centrality of the notion of community itself, which renders the concept of community resilience helpful in analysis, while likely reflecting values that are taken for granted by study participants. PSWs in this study found their work deeply rewarding, expressing high levels of motivation, enormous commitment to serving their community and pride in the status and respect they gained. However, restricting aspects of the work of PSWs to those households identified as the most vulnerable may undermine community coherence, while scholars in Uganda such as Ochen et al. (2017) have suggested a more general attention to the implementation of children's rights would be preferable to boost prevention work.
Further, inadequate community resources for responding to the most serious child protection concerns impacted on community resilience because of the scarcity of good-quality alternative placements. Adoption or fostering outside the extended family or tribe is rare in Ugandan culture and the vast majority of children's homes are not licenced by the government . Understandable reluctance to send children away from the community coupled with limited alternative care capacity within the community has led to a situation in which participants described most cases as indefinitely ongoing. The need for family support and child protection services to be provided within the context of extended families and communities has also been identified by Krueger et al. (2014) in their mapping exercise of child protection arrangements in five West African countries. Poor provision for alternative care is an issue highlighted as of particular urgency in Uganda by Walakira et al. (2017b): Uganda is developing a National Alternative Care Action Plan (NACAP) 2016-2021 to address reliance on institutional care.
It is also important to note the much wider conceptualisation of child protection than that which is familiar to Western practitioners which is reflected in the remit of the PSW role. This mirrors the placement of child protection within the community development sector of local government institutions, which in itself arises from recognition of the central significance of poverty for parenting capacity and practices. While economic development initiatives have not been considered here, they are central to Norris et al.' (2008) networked adaptive capacities and it is clear that changing parental attitudes to issues such as physical punishment, education, child labour and early marriage will only crystallise into accepted norms of behaviour when families have the luxury of being able to look beyond their immediate needs to their children's futures.
Further issues to be addressed include the question of whether the role should be voluntary: while in the study settings this was regarded as a practical necessity to maintain services in the long-term, modest remuneration was regarded as essential for sustainability in Linsk et al.'s (2010) Tanzanian study (2010) and would provide greater status and insurance against exploitation for PSWs. Similar dilemmas in relation to balancing fiscal considerations against community ownership govern decisions as to whether PSWs should provide universal services in their communities or work only with households identified as particularly vulnerable.

Limitations
Community households and children were not consulted, due to the logistical complications of travel and language barriers: greater attention should be paid in future work to the potential for PSWs to empower children themselves and to how PSWs negotiate traditional hierarchies within households. In addition to issues of selection bias addressed above, the extreme remoteness and deprivation of communities should be noted. Although described as village leaders, demonstrating high levels of commitment and providing unique insight into their role, only seven PSWs were interviewed with varying levels of education, and those in D2 were notably less articulate than those in D1.

Conclusions
At a time when protracted crises, poverty and the AIDS epidemic have together strained the capacities and cohesion of traditional community arrangements in many African nations, the research reported here suggests that the PSW model is potentially a powerful vehicle for the development of community resilience, embodying Lachman's (2002) vision for poverty-afflicted nations in which child protection is 'enmeshed within the services of a community, and where the participation of the community itself and children is seen as paramount' (591).
Identified attributes of resilient communities, including equitable distribution of resources, networks of relationships to support social capital and the importance of sense of community, collective efficacy and empowerment may help guide decision-making in response to these issues in individual communities.
Given its widespread and developing usage it is imperative that more research is undertaken to understand the conditions under which the PSW model is most effective and sustainable in different social, economic, political and cultural contexts and how best it can work to support formal or statutory services delivered by local government institutions.