Iatrogenic Violence? Lived Experiences of Recipients of Aid that Targets Vulnerable Children in Makete, Tanzania

Humanitarian aid does not always achieve its intended outcomes but may have unintended side-effects which harm rather than help recipient communities in the long run. In the context of aid, iatrogenic violence refers to circumstances where aid intended to benefit recipients instead causes social disruption. The relations between donor and recipient contribute significantly to the outcomes of aid. This article aims to explore the recipient side of donor–recipient relationships in humanitarian aid targeting children affected by HIV/AIDS in Makete, Tanzania. Specific objectives include exploring (i) the lived experiences of the children who are targeted by the aid, (ii) the lived experiences of related caregivers and community members and (iii) the responses of local leaders, both non-governmental organisation (NGO) officers and local government leaders. Forty-one orphaned children aged 9–18 were involved in 3 one-hour-long participatory activities. From these, 12 were selected for in-depth interviews. Four caregivers of orphaned children and 15 key informants (community leaders and NGO officers) were interviewed. Findings show that the type of aid and method of distribution are determined by the donors and do not fit well with local conditions. Examples of iatrogenic violence include conflict and division between those in the community who receive aid and those who do not, increasing dependency with a corresponding reduction in self-reliance, criteria that exclude the most vulnerable children from aid and passivity and non-participation. Donor-controlled humanitarian aid is causing unintended harm in the recipient communities.


Introduction
Humanitarian aid does not always achieve its intended outcomes. Indeed, aid may have unintended side-effects which harm rather than help recipient communities in the long run. The concept of iatrogenic violence describes a situation where an intervention by a doctor trying to heal a patient causes more harm than good. In the context of aid, iatrogenic violence refers to circumstances where aid intended to benefit recipients instead causes social disruption (McFalls, 2010).
The global structure of aid has historically caused social disruption. The harm caused by the neo-liberalist agenda of aid during the 'Washington Consensus' years has been well documented; and even under the post-Washington consensus, the introduction of outcome targets (for example the Millennium Development Goals) ensured the imposition of neo-liberal values (Mawdsley et al., 2014;Murray and Overton, 2011). The principles outlined in the Paris Declaration of 2005 -ownership, alignment, harmonisation, results-based management and mutual accountability -are intended to increase aid effectiveness with recipients seen as 'partners' (Mawdsley et al., 2014;Murray and Overton, 2011).
Civil society is an important implementer of aid at the local level. Katz (2006) outlines two contrasting views within neo-Gramscian thought on the role of civil society: on the one hand it is seen as the counterforce to neo-liberal globalisation, but on the other hand it is seen as an extension of the state (or ruling authorities). In the latter view, global civil society is seen as promoting the neo-liberal agenda, but 'under a cloak of openness' (Katz, 2006, p. 335). Mawdsley and colleagues (2014) discuss the role of civil society within the Paris Declaration framework and note that one significant goal of civil society is the strengthening of the human rights approach. Whitfield and Fraser (2010) list human rights among a series of governance conditions that were imposed additionally in the 1990s onto the shrunken state structures and further undermined the independence of recipient governments. Magesan (2013) has studied the link between ratification of human rights treaties and aid receipts and he contends that simply ratifying human rights treaties (as opposed to real changes in human rights behaviours) increases aid received per person both in the short and the long run.
The donor-recipient relationship in humanitarian aid in the HIV/AIDS context will not only affect the effectiveness of the implementation, but also the lives of those affected and infected by the disease. As Seckinelgin (2012) notes, the lived experiences of those living with the disease are seldom considered. This article aims to explore the recipient side of donor-recipient relationships in humanitarian aid targeting children affected by HIV/AIDS. Specific objectives include exploring (i) the lived experiences of the children who are targeted by the aid, (ii) the lived experiences of related caregivers and community members and (iii) the responses of local leaders, both non-governmental organisation (NGO) officers and local government leaders.
Conceptual framework: iatrogenic violence in humanitarian aid in the HIV/AIDS context Humanitarian aid -the donor response to crises or states of emergency such as natural disasters or human-generated conflict situations -forms a particular subset of foreign aid. Hilhorst and Jansen (2010) note that the humanitarian principles of neutrality, impartiality and humanity are often used by political actors to legitimise their activities. McFalls (2010) agrees and points out that international agencies use medical metaphors to legitimise humanitarian aid because medical interventions are seen as exempt from ethical critique. He coins some of his own medical metaphors: 'therapeutic domination' (where the power differentials in the donor-recipient relations are like those in the doctor-patient relationship) can lead to 'iatrogenic violence' which he describes as 'social disruption and political violence that results from outside intervention intended to stop or prevent such violence' (McFalls, 2010, p. 319). He contends that no matter what the nature or context of the intervention 'iatrogenic violence is inherent in the formal structure of international intervention' (McFalls, 2010). Nguyen (2005) echoes McFalls' sentiment but in a more neutral manner when he asserts that donors mould social interaction and generate new forms of relations through their use of NGOs. McFalls extends his description of the violence of humanitarian aid: ' . . . therapeutic domination not only depersonalises, but decontextualises social relationships. Without any reference to culture or to history, therapeutic domination reduces social agents to human bodies ' (2010, p. 323). McFalls' use of the concept of therapeutic domination is similar to Foucault's notion of 'biopolitics' as used by both Nguyen (2005) and Duffield (2007) to mean the power to promote or undermine life.
Humanitarian aid frequently involves the defining of beneficiaries or target groups which inevitably involves inclusion and exclusion: those who are eligible for aid are defined by the criteria of vulnerability and categories of need (Hilhorst and Jansen, 2010). Such social differentiation is an example of both therapeutic domination (the aid agencies with little reference to cultural context decide who is in and who is out) and iatrogenic violence (disrupting social relationships). Other forms of iatrogenic violence include the creation of dependency and stripping local people of their agency (Hilhorst and Jansen, 2010). McFalls (2010, p. 328) describes the logic of the donors using . . . the quip 'the surgery was successful, but the patient died'. The therapeutic structure of domination also determines the mode of resistance principally as 'patient noncompliance': usually passive-aggressive, often self-destructive, and occasionally prone to apparently irrational outbursts directed at 'caregivers' . . . Hilhorst andJansen (2010, p. 1136) discuss the issue of human rights within humanitarian aid and conclude that education about human rights inadvertently leads to 'a permanent sense of dissatisfaction among refugees' and that the rights discourse has 'advanced individualisation to the detriment of community solidarity'. McFalls (2010) concludes that multiple crises have led to a permanent state of emergency which has allowed the governing of humanitarian aid to become a dictatorship (all-be-it a benevolent one) that is exempt from the challenges and critiques of normal politics. This dictatorship implies that the principles of the Paris Declaration are less visible in humanitarian aid.
An example of a seemingly 'permanent' crisis is the HIV/AIDS epidemic which is still treated as an emergency although well into its fourth decade. Duffield (1997, p. 529), commenting on 'so-called continuum thinking' (with humanitarian/relief aid at one end and development at the other), notes that government-led infrastructure projects have been superseded by partnerships with local communities that empower and build capacity. The HIV epidemic is associated with aid that provides treatment, material support and technologies to empower individuals living with HIV rather than transforming local public health care systems in Africa to provide the needed services (Nguyen, 2005). Such aid is closer to the humanitarian/relief end of the continuum. However, HIV/AIDS is also a significant policy area in development aid frameworks like the Millennium Development Goals (MDGs) (Seckinelgin, 2005) and consequently resultsbased management that is seen in other aspects of aid is also practised in the context of HIV. Other aspects of the Paris Declaration are highly visible, at least in theory, when the language of partnership, ownership and harmonisation is written into global AIDS policy (see for example the role of partners in implementing male circumcision for HIV prevention in the report by the World Health Organisation (WHO) and the Joint United Nations Programme on HIV/AIDS (UNAIDS), 2011).
The humanitarian aspect of aid in response to the 'crisis' of HIV/AIDS demonstrates many of the characteristics of therapeutic domination identified by McFalls (2010). The international organisations (especially UNAIDS and WHO) legitimise their interventions using the scientific rationality of medicine; the recipients -people affected and infected by HIV -are seen as passively accepting the policy interventions and biomedical interventions dominate and seem unrelated to actual lived experiences and cultural contexts of the recipients (Seckinelgin, 2012). The implication here is that while the language of global HIV/AIDS policy is partnership (ownership, alignment and mutual accountability), the reality is closer to therapeutic domination with donors controlling the nature and extent of interventions. A note about culture: HIV/AIDS is closely related to sex and death which are both taboo topics in many cultures making the cultural context extremely significant for the effectiveness of interventions. Seckinelgin (2005) outlines the clear governance structure for HIV/AIDS (unlike other 'emergency' situations) headed by UNAIDS which co-ordinates the efforts of all UN organisations and is based in Geneva, not New York, for convenient collaboration with WHO. This governance structure is influential in several of the large funds which contribute to the 'emergency' of HIV/AIDS, for example, the Global Fund for HIV, Malaria and TB; the President's Emergency Plan for AIDS Relief (PEPFAR) and of course the MDGs. In addition, several philanthropists, like Bill Gates and Bill Clinton, make significant donations to the HIV/AIDS emergency. Seckinelgin (2005, p. 359) contends that NGOs have very little agency in contributing to policy development, instead they are seen as 'close to the people' and are used merely as implementers, as 'delivery tools'.

Tanzanian context
The introduction of large amounts of aid money must be understood in the context of the evolving political economy in Tanzania and the changing balance of power between government (both central and local) and the civil society sector. After independence in 1961, Tanzania introduced a unique form of socialism known as ujamaa which was intended to enable self-reliance rather than dependence on aid from the West. However, instead of self-reliance, Tanzania has, for decades, depended heavily on donor support (Rugumamu, 1997). Michael (2004, p. 70) describes the post-independence expansion of the state in Tanzania: the proliferation of state structures effectively eliminated 'the space for autonomous civil society organization'. In the mid-1980s, with structural adjustment and the introduction of neo-liberal market forces into the socialist economy, the state was cut back and civil society began to grow once more (de Graaf, 2005). Since the 1990s, NGOs in Tanzania have mushroomed (Rueben, 2002). Green (2012, p. 312) notes that this growth in civil society in Tanzania represents a 'significant transition in the political culture' towards entrepreneurship and self-interest previously associated with capitalism. A significant portion of civil society organisations (CSOs) in Tanzania are linked to the HIV/AIDS epidemic, which hit the country hard from the mid-1980s onwards. Local NGOs frequently lack power, fail to secure funding from international donors and are obliged to cooperate with local government bureaucracy in order to function in the local environment (Michael, 2004). In addition, they have to accept what Hearn (2007Hearn ( , p. 1103) calls 'an external orientation', they are under pressure to conform to an agenda other than their own. Prior to the 1990s, district and local governments in Tanzania had strong control over service provision, but the growth of civil society has eroded this monopoly (Michael, 2004). District and local governments have attempted to reign in the CSOs, for example requiring them to submit reports on their activities, but they have few sanctions should the CSOs fail to comply. Against the background of the changes in foreign aid described above and an evolving political economy in the country, Tanzania, in 2008, received US$58 million from the Global Fund and an additional US$40 million from the President's Emergency Plan for AIDS Relief (PEPFAR) to be spent on Orphans and Vulnerable Children in the period 2008-2013. PACT (the organisation responsible for disbursing the money) is working closely with the Department of Social Welfare (DSW), UNICEF and local NGOs to establish a wide-reaching and sustainable action programme for the most vulnerable children (MVC) 1 in the country. 2

Methodology
The findings and analysis in this article are based on data collected between January 2007 and November 2008 in Makete, Tanzania. Makete District in Iringa Region is part of the pilot community-based MVC programme (overseen by DSW and UNICEF) which started in 2006 (Government of Tanzania, 2006). Makete is a remote rural area, to the northwest of Njombe, with poor transport and communication links, high rates of poverty and labour migration, high HIV/AIDS prevalence 3 and, consequently, high levels of orphaning. Makete has the country's highest number of MVC as a proportion of all children: 17.4 per cent while the average for the country is 5.3 per cent (Government of Tanzania, 2006, p. 5). MVC in Makete District are the recipients of funds from PEPFAR and the Global Fund through national NGOs and international faith-based organisations.
Initial access to the participants was through an NGO working in the area and through local government officials. There were 3 participant groups: orphans and vulnerable children, caregivers and other key informants. Sixty-seven children between the ages of 9 and 18 were identified by community leaders. The children took part in 3 one-hour-long, participatory activities, with 41 children completing all 3 sessions. The activities were designed to build rapport with the research team and collect background data, for example, on the children's understanding of friendship, responsibility and becoming an adult. After observation during the activity sessions, 12 children (between the ages of 12 and 18) were purposefully selected for in-depth interviews because they were thought to be information-rich and able to articulate their experiences (see Table 1 below for details). The second group, consisting of 4 caregivers of orphaned and vulnerable children (1 grandfather, 2 grandmothers and 1 aunt), participated in in-depth interviews. There were visits to 10 households in the village, including those of all the participating caregivers and 2 of the participating children. The third group comprised 11 key informants among community leaders such as primary school teachers, care promoters, the village executive officer, NGO officers, church leaders and district officials. Data were collected through in-depth interviews and focus group discussions. In addition, programme officers of the participating NGOs were interviewed at their headquarters in Dar es Salaam. There were also interviews with programme officers at UNICEF, PACT and other relevant organisations.
All interviews were recorded and transcribed in full (Kiswahili as well as English) and the Kiswahili parts were translated into English. The transcriptions were coded with the help of Nvivo 8 and thereafter analysed using thematic network analysis (Attride-Stirling, 2001). Children were asked about the aid they had or had not received, kinship networks both near and far geographically, other social networks such as friends, church/mosque and neighbours. Adults were asked about, among other things, the criteria used by NGOs for recruitment, the nature and content of aid, the relationship between local government and civil society and the response of recipients and non-recipients to aid.

Ethics
The Commission for Science and Technology (COSTECH) granted permission to do the research in Tanzania. Rigorous research ethics, including informed consent, were followed with all participants who were ensured of their rights to anonymity and confidentiality. The informants were given the opportunity to withdraw from the study at any time should they so wish. All the child interviewees were given the opportunity to have one of the community leaders present should they so wish, but none chose to do so. The interview was the fourth meeting between the researcher and child informant, so some degree of rapport and trust had been established.

Impact of HIV/AIDS on social networks
In Makete the impact of HIV/AIDS can be seen in size of the households of the child interviewees. As shown in Table 1, 8 of the 12 selected orphans' households consisted of 2 people only, including the 3 child-headed households. A further 3 households consisted of 3 people only. The cause can almost certainly be put down to the high number of deaths experienced (largely due to AIDS) by all families as well as the high rate of migration. During the interviews, the extended family relationships and kinship networks were explored in detail, relatives living both locally and elsewhere. It was clear that many families had suffered devastating losses. The caregivers interviewed confirmed this experience. One of the caregivers, the aunt who is looking after 2 nieces and a nephew, stated: 'We were born 11; all of them are dead except two of us.' She is not particularly old; she has one (6 year old) child of her own. Another caregiver, a grandmother, has experienced the death of all of her own children: The reason I am living with my granddaughter and the other two children is death. I was blessed with five children in my marriage, two girls and three boys. . . . They are all gone; the five of them there are down.
A third caregiver had raised 6 children to adulthood, 3 had died and a fourth had been chronically ill for some months. Her youngest child is still in secondary school. Such a death rate is truly excessive -it can be understood why those remaining alive live in such small households: there is no one else left. Kinship and extended family networks have been severely weakened by deaths, shifting responsibility for childcare to grandparents and aunts. In some cases remaining adult relatives are reluctant to take responsibility for orphans and consequently there are a number of orphan-headed households exemplified by Stella, Joseph and Lucy. The holes in the extended family safety net mean that caregivers have to turn elsewhere for support. Most people in the community, however, are also poor and are not able to help with items of larger expenditure (such as school fees) or long-term needs. Communities have responded with the establishment of communal shambas for orphans (reviving a practice started under ujamaa) or with small savings and credit groups, but these meet only a fraction of the need.

The experiences of vulnerable children
The large flows of money from international donors have to be distributed, administered, monitored and evaluated according to donor specifications. Participants from each of the 3 groups noted that most of the aid given by civil society organisations (funded by international donors) consists of items which promote access to education for individual children: school uniforms, writing materials and, for secondary school students, school fees. Some organisations also give out health insurance so that the recipient is able to attend the hospital without paying user fees. Occasionally organisations help with roofing material or they may provide mattresses and bedding.
The targeting of individuals can be problematic. A 13-year-old orphaned girl, living with her maternal aunt and 3 cousins describes what happened to her: The organization has helped me twice with school uniform, but my cousins, they did not get anything. They did not say anything to me, they just looked at me. I felt terrible because I thought all of us would get. (Katie, 13 years) Children, caregivers and key informants reported that recruitment to the programmes occurs in 2 ways. First the organisation may do research itself, going from house to house to assess the needs. This research was described by the children and the caregivers as 'passing through', and it tends to raise expectations that a donation will follow. These expectations are often fuelled by promises from the organisation doing the research.
So there are certain people who went to our home and took our names and told us that a donation will be coming. But I don't know them, I didn't recognise anybody, they are not from our village. (Max 13 years) Second the organisation may work through the village government and the Village MVC committee to identify those needing assistance. Households caring for orphans may register with the village office as potential candidates for aid. A combination of these methods may also be used.
Some households fail to access aid -whether through failing to meet the criteria or because the village government has identified them as not needing assistance. Children in excluded households reported feeling distress when they see their friends receive donations: I feel terrible when the other orphans get a donation. I ask them how they do it, to get aid. Fine, they tell me, but there at the office when they receive something, I think they deleted my name. Grandmother is now fed up she has told me that even if they come and say you are wanted at the office I am supposed to refuse. For example the day before yesterday my friends went and they were given blankets, mattresses and others were given iron sheets. Two days after I overheard them saying, "There is another orphan who has come from Mbeya. But how do we know? We can say she is an orphan when in the actual sense she is not." So they ended up doubting my orphan status and the death of my mother whom we mourn. (Rosa 12 years) Stella, 18 years old at the time of the interview, had been caring for her brother since her mother died when she was 10. Her only remaining aunt (mother's sister) living in the area at the time her mother's death had rejected her because her mother died of AIDS. Because she dropped out of school in order to make charcoal and cultivate food she was not eligible for NGO programmes supplying school materials. One organisation did however give her brother school uniform and writing materials. Shortly before the interview, the organisation decided to cut the aid as her brother was now classed as having an adult caregiver: This time in June, they listed his name but said that now that I am old enough (18), this year they will only give him a pair of shoes.
She was also ineligible for assistance according to the criteria of another organisation: I wrote a letter to ask for a loan from one organization dealing with lending people money. The loan officer came to our place and asked: 'who will be your sponsor, someone who can pay back your loan in case you fail to pay back yourself?' I did not have anyone so I told them to just leave the loan if that was the case.
The girl and her brother are vulnerable in a number of ways: they have no relatives willing to help them, they have no cash income, the girl's health is poor since her lungs have been damaged by the charcoal-making when she was young, and involvement in a car accident left her with a badly broken upper arm that has limited her ability to cultivate. Yet, in spite of their obvious vulnerability, they do not meet the criteria for aid. The other children heading households, Joseph and Lucy, were both in school and received uniforms for themselves. Each had also received a once-off food parcel.

The experiences of caregivers
As mentioned above, the individualist nature of the aid can lead to conflict among the villages. Other divisions in the community may also arise from perceived inequality and exclusion. Caregivers attempted to access aid for the children in their care by registering them at the village office or, if possible, directly with the NGO. The NGO or village government officers who apply the NGO criteria for deciding who is eligible to receive a donation, in effect judge who is a 'legitimate' recipient and who is not; they have the power to include or to exclude. Registration is not a sufficient condition to be eligible for aid. Local government leaders state that they select 'the most needy', but their definition of 'the most needy' is not always shared by those attempting to access aid.
Yes there are some organizations which passed through, they took the names of those children who are orphans and who are living in poor conditions, but we got nothing. So other children they are getting help because sometimes there will be a lot of maize in the village office. But they told me 'you are able and you have a plot to cultivate so it's not necessary for you to get this help'. But I told them 'this is not enough for me because also children depend on me'. (Grandmother caregiver) This grandmother was indeed very successful in cultivation. Elsewhere in the interview she describes giving surplus food to neighbours at times, and using some of the surplus to make pombe (the local brew) to sell for cash. However, when she compares herself to others who receive aid, she feels that she is being treated unfairly.
When people try repeatedly to access aid and fail to do so, or are treated with contempt, they may 'give up' trying or become passive in their situation. One grandmother caregiver described a situation where she felt she had been treated contemptuously. She was called to fetch some aid and then made to wait for 12 hours. She watched other people receive clothes, mattresses and iron sheets for roofing, but she was given only maize flour which she felt she did not need; what she wanted was a mattress for her granddaughter who has to sleep on the floor. When she complained she was told it was because she has health insurance, but she has to pay every time she goes to the hospital. She explained why she did not report this to the village leaders: We did not ask because all others who have health insurance have an ID. I just asked myself why they say that I have health insurance while I am still paying for all my treatments and I have not been given any ID. The reason why I think this is happening is that I don't have any relative there who can assist me to get my rights so whenever we complain nobody bothers to make a follow-up to our complaints. I keep asking myself, even today, which category the village leadership puts me in, even though they know that I have nothing at all. Aid is sent to help us but when it reaches here people change and tell us that we have health insurance. I just leave it to God.
Beyond a certain point it is no longer worth struggling to get aid, this woman has reached that point and given up.
In some cases the caregivers will manipulate the system to try and access the aid.
They take names of the children at school. The problem is, if your child is well dressed they don't give him/her any aid. Some people dress their children badly so that they look poor and get aid. (Grandmother caregiver) Rather than seeking other solutions to their problems, they are prepared to take any measure to access aid.

Local leaders' response
With the emergence of so many aid agencies, the government's role has changed from service provision only to one of greater co-ordination of a multiplicity of service providers. District government requires NGOs to report to the relevant department in the district administration, and, at stakeholder meetings, tries to organise a more equitable geographical distribution of funds. 4 UNICEF has played a significant co-ordination role in selected districts. 5 The role of local government has evolved into one of gate keeping (deciding who is eligible to benefit from aid) with a reduced responsibility for service provision (for example, establishing a communal shamba).
The way in which NGO and local government leaders interact with one another seems to provide opportunities for expediency and corruption. In the following example, the organisation had asked the village leadership to nominate a person to distribute the donations the organisation provided. A community leader described an incident that occurred during the distribution of school uniforms: The co-ordinator (the NGO officer) was with the woman who is the main distributor (appointed by village leadership). She took a pair of shoes and put it on her back under her child who is not even walking yet. Some pupils saw her and started to complain as to why the distributor would hide a pair of shoes. It was reported to the co-ordinator, it had become her habit to hide aid, but the co-ordinator refused to sack her. The children saw her hide the shoes and she is now wearing them. What kind of a leader hides things she was supposed to distribute to the needy? The way I see it the woman was somehow related to the former village chairman. That is the reason she was not sacked. They reasoned that by having that woman in the leadership, they will also benefit.
The individual nature of aid can result in conflict between villagers as is illustrated by some of the participants in a focus group discussion: P1: It does of course stimulate strife between those that have received and those that have not.
P2: The one who doesn't receive anything when others have received twice or three times from the same aid organization normally complains to those who receive and not to the distributors.
Other informants reported incidents where caregivers, frustrated at not receiving assistance, would shout at neighbours returning with wheelbarrows full of donations. The focus group participants went on to discuss their understanding of issues related to the distribution of the aid: people in ideas which are introduced by themselves, not imposed on them. (NGO Technical Officer) Direct support 'imposed' on the community (i.e. without participation) undermines the community's responsibility to solve its own problems. It may also cause the community to step back from helping those who receive outside aid. The officer also suggests that there should be collaboration with government and the grassroots community to solve the problems, rather than having the donors' solutions imposed on them.

Analysis and discussion
The small households found in Makete are an indication of deep vulnerability. Such households are prone to poverty even if they have resources like a shamba because they do not have enough people to work the land fully. They are also socially vulnerable -if one of the two members leaves or dies the survival of the other is jeopardised. These are fragile households and it is clear that they could benefit from aid.

Iatrogenic violence
The nature of the aid being given is a source of iatrogenic violence -the aid that is meant to help vulnerable members of a community is itself causing social stress (McFalls, 2010). Programmes and organisations are usually funded by Western donors who bring with them assumptions about what programmes should encompass and how they should be run. Both the nature of the aid and the method of distribution have created tensions in the community.
The individualist nature of the aid is part of the problem. In the case of programmes to help vulnerable children these usually provide uniforms and school materials for selected individual children. When local government and NGO leaders are forced (by donor requirements) to pick out individual beneficiaries, this enervates commitment to the collectivist values of the community such as the obligation to help others. The criteria seemingly benefit the lucky few, for example, the orphaned girl who was given a school uniform while her cousins in the same household received nothing. Others are excluded or deemed ineligible by the criteria causing conflict and strife among community members. The content of the aid is decided by the donor organisations acting as 'benevolent dictators' (McFalls, 2010): it is easy to measure how many school uniforms are distributed so this facilitates accountability. However, aid in the form of school uniforms -although it benefits some -excludes all those who are not in school, and those who have dropped out of school are often the neediest. Some caregivers, when they did receive aid, felt that it did not address the needs they had prioritised. This is a case of the aid not fitting the circumstances of the recipients as in the criticism levelled at aid under the Washington Consensus (Mawdsley et al., 2014). Alternatively, expressed in terms of iatrogenic violence and therapeutic domination, the 'violent' impact was the separation experienced by the girl from the family in which she lived; the exclusion of those not in school; and literal violence and conflict experienced between neighbours. The aid has been given ('the surgery was successful') and if any suffered the violence of separation or exclusion ('the patient died') the blame is certainly not attributed to the donor (McFalls, 2010, p. 328).
The method of distribution is also problematic and causes unintended harm. Recruiting teams created expectations when gathering information but not everyone who felt they were eligible received aid. The criteria for inclusion and exclusion were not clearly understood by the local community. Some households were seen to receive aid several times while others received nothing and this led to conflict in the community with neighbours shouting at each other. Other caregivers felt they had been treated with contempt and this led to passivity. The involvement of local community leaders in the distribution of aid led to opportunities for expediency causing further dissatisfaction and division in the community. All these responses closely reflect McFalls' (2010, p. 328) views of the mode of resistance to therapeutic domination as ' . . . usually passive-aggressive, often self-destructive, and occasionally prone to irrational outbursts . . . ' Another example of iatrogenic violence is caused by both the nature of the aid and the method of distribution, and that is dependency. The NGO Technical Officer who spoke about dependency noted that community members are not allowed to participate in deciding what the aid will comprise and how it will be distributed, they end up 'waiting for the organisation' -passively accepting what is given and at the same time ceasing to take the initiative in seeking solutions (Hilhorst and Jansen, 2010). As one of the caregivers described, individuals are also prepared to manipulate the system (for example, by dressing their child to look more needy) in order to access aid. Seckinelgin (2005, p. 361) calls this 'external resource dependency'. Marsland (2006) discusses the Tanzanian understanding of 'participation', in the context of ideological history, as the responsibility or obligation to contribute to the development of the country, while the Western understanding of 'participation' has provided elites with the opportunity to benefit at the expense of the marginalised (Mohan, 2002). Another aspect of dependency is that the (short term) focus on aid through NGOs replaces democratic expectations on government to make appropriate structural developments to address the peoples' needs in the longer term by the underlying structural causes of poverty. Nor does it provide an opportunity to work together to establish and nurture a project that will sustain income, well-being and independence in the long run. Seckinelgin (2005) suggests that there are long-term political implications, notably that citizens are disconnected from the state, or, as Kinsella and Brehony (2009) put it, they are 'increasingly disengaged from political decision-making'.

Therapeutic domination
The fact that both the nature of aid and the method of its distribution have been determined by donors is an example of therapeutic domination. People's attempts to access aid are controlled by the village leadership applying the criteria of PEPFAR and the Global Fund. They have the power to judge that one person is a legitimate recipient of aid, has the right to benefit while another is ineligible. All households caring for an orphan can register at the village office and this makes them potential, but not necessarily actual, recipients of aid. However, aid for orphan households has come to be seen as a right, with the leaders -who know the circumstances of the residents -obliged to distribute accordingly. One of the grandmothers quoted above referred to the fact that she did not have a relative to assist her to get her rights and she saw that as the cause of the contempt she was shown. The language of rights has been brought in by international aid agencies and Marriage (2006) contends that that this language is often used to facilitate denial about what is really happening in aid programmes in order to protect and sustain the flow of money. The language of rights has also inadvertently led to a sense of dissatisfaction (Hilhorst and Jansen, 2010).
Exclusion or being judged as illegitimate to receive aid may be due to criteria or the aid itself being inappropriate for the recipient community. As the 12-year-old girl expressed it, her orphan status and the death of her mother were being questioned. When the loan officer visited the home of the 18-year-old girl, he exposed her destitution -in the full sense of the concept in Africa: lack of family and lack of resources. As personhood is defined in terms of family and kin relationships, he was effectively questioning her personhood. Her only defence in the face of such humiliation was to tell him to just leave. When the aid agency that had been helping her brother withdrew their aid because she had turned 18 and so was counted an adult, it was a case of lack of recognition of her true situation, rejection of her ongoing vulnerability. The grandmother, who waited 12 hours for a sack of maize meal when what she really needed was a mattress for her granddaughter, felt she had been treated with contempt -her real needs had not been understood or listened to and her anger was brushed off with a lie about her access to free medical aid. These are examples of iatrogenic violence stemming from therapeutic domination -the donors, backed by the power of the funds they supply, decide on criteria and aid that do not fit the situation on the ground.
The aid in question here is targeting children made vulnerable by the HIV/AIDS epidemic. It clearly fits under several of the MDGs and should theoretically avoid the characteristics of the Washington consensus that have been so heavily criticised, such as large power-imbalances between donors and recipients (Mawdsley et al., 2014). Instead it features many aspects of therapeutic domination: the power imbalance inherent in the doctor-patient relationship is evident in the donor-recipient relationship here (McFalls, 2010). This imbalance or lop-sidedness confirms that HIV/AIDS funding in this case fits more closely with humanitarian aid than with post-Washington consensus conditions (McFalls, 2010;Seckinelgin, 2012). In other words, more than 3 decades into the epidemic, HIV/AIDS is still seen as an 'emergency' situation requiring relief aid. While the short-term needs of some vulnerable individuals may be met, the requirements for long-term survival in this particular socio-cultural environment and the underlying causes of vulnerability are not being addressed (Seckinelgin, 2005).
The current perspective allows the aid agencies to avoid ethical questioning about their methods, they are taken as 'doing good': the 'benevolent dictatorship' in action (McFalls, 2010). The findings presented above give examples of iatrogenic violencealthough the aid is intended to help children and support caregivers, the experience of children is often exclusion, and the experience of caregivers is often conflict.

Conclusion
This article has explored the recipient side of donor-recipient relationships in humanitarian aid targeting children affected by HIV/AIDS by studying the lived experiences of children, caregivers and community members as well as the responses of local leaders in Makete, Tanzania. The findings reveal numerous examples of social disruption and unintended harm caused by the aid itself, for example, conflict between those who receive aid and those who do not, criteria that exclude the most needy, passive responses to contempt and an increase in dependency with a corresponding decline in self-reliance. These examples of iatrogenic violence reflect therapeutic domination by donors: donors, with little or no participation by community members, decide on the type of aid and the method of distribution that do not fit the local conditions. Donor-recipient relationships in the context of aid to vulnerable children are generating inadvertent side-effects reminiscent of those that generated criticism under the Washington Consensus.