Issue 25 Editorial

This special issue of Field Exchange focuses on the food aid component of HIV related programming and was made possible through additional funding from DFID RSA. How the ENN have gathered field material for this issue marks a significant departure from our usual approach, in that a consultant (Mary Corbett) was taken on as a kind of roving researcher/correspondent. Mary visited five countries over the course of six weeks (Malawi, Zambia, Kenya and Uganda and South Africa) to meet with a wide variety of agency staff implementing HIV programming with a food aid component. Her brief was to describe these programmes and, where possible, identify lessons learnt. She was also asked to identify significant related research initiatives or findings in the countries visited. An account of the experience is captured in her editorial (page 17).

The rationale for this special issue was primarily that HIV-related programming involving food aid has been increasingly rolled out over the past few years, with a view to achieving a variety of objectives. Since much of this programming is 'cutting edge', many of the objectives have not been properly tested. Hence agencies are, effectively, learning by doing. The ENN (along with others) believes it important to document these new programming experiences to support a process of learning. Although the production of this special issue, based on 18 programme experiences, cannot be described as a comprehensive overview, it is, at least, a snap-shot of what is going on. The ENN believe that some of the lessons and cross-cutting issues to emerge from this special issue are important and indicate priorities and strategies for the future.

The agency field experiences documented are based on programmes implemented by a wide spectrum of agency types, i.e. UN, INGO, local NGO and CBO. Some of the CBOs literally started out as spontaneous community initiatives, led by a few dynamic individuals affected directly or indirectly by HIV. Over time, these organisations have professionalised, expanded, and achieved a significant profile and status with international agencies who now collaborate with them, and often provide resource support. The food aid and nutritional components of programming described here are integrated into several different programme types, i.e. DOT, HBC, CI/OVC, ARV, PMTCT and school feeding. Most of these programmes have multiple objectives for the food aid component. Furthermore, although the programme write-ups have focused on the food aid element of programming, many contain other sectoral elements as part of an overall integrated programme package. For example, the CRS programme in Dedza, Malawi (p38) not only targeted food aid to PLWHA households, but also contained an IGA element, as well as vocational training for older orphans.

There are a number of key findings from this snap-shot of agency programming:

Objectives and evidence of impact

Picture taken on World Aids Day

Multiple objectives for the food aid element of programming are invoked. For example, the 18 month food aid package as part of the PMTCT programme at St. Gabriels Hospital in Malawi (by Gertrude Kara and Mary Corbett) is meant to encourage compliance and ongoing educational support, support abrupt weaning, ensure full infant immunisation, ensure VCT for infants on reaching 18 months and provide an opportunity to target IGA. The community school feeding programme in Zambia (by Kate Vorley and Mary Corbett) is aimed at improving enrolment in HIV affected areas and dissemination of HIV related knowledge. The food aid component of Oxfam's integrated programme in Malawi (p22) is intended to break the cycle leading to individual and household crisis. While food aid may, indeed, be able to serve a myriad of objectives within different programming contexts, it is notable that almost all objectives (which either appear in programme documents or are assumed by programme implementers) are expressed in general and qualitative terms, i.e. specific quantifiable targets are not set. Thus, where the objective is to increase weight gain or survival of HIV affected individuals, the actual percentage weight gain or percentage increase in survival hoped for is never explicitly stated. As a result, achievement of objectives cannot be tested. Indeed, in some cases and contexts, it may be unrealistic to set specific targets, especially where there is little prior programme experience. However, if targets remain nebulous, then progress in understanding whether and how food aid/nutritional support is effective cannot be made. In some contexts where there is adequate infrastructure and staff capacity, e.g. DOT and PMTCT, targets should be established as a priority, so that data can be rapidly generated to test the degree to which objectives are being met. Also, by establishing targets, it is more likely that implementing agencies will invest in collecting, collating and analysing data to demonstrate programme impact.

Findings from the ENN field work support the ubiquitous 'impression' that very little impact assessment is taking place. Most of the programmes visited had either not collected impact data, or had collected data but not undertaken analysis. In many cases this was attributed to lack of time and resources, with some agencies admitting the need for professional support (e.g. Reach Out programme in Uganda). While this is a credible excuse for CBOs and local NGOs, it is less understandable for INGOs or UN agencies. The absence of impact analysis is worrying and undoubtedly explains why there is currently very little in the published literature on this relatively new type of programming. The impact data that have been collected do appear to show some impact in terms of weight gain¹. However, much of these data and analysis lack statistical rigor and are conceptually weak. This is not to be overly critical. Impact assessment of this type of programming is methodologically complex. For example, in a food aid/ARV or DOT programme, it would be necessary to isolate the impact of food aid in terms of improving tolerance to drugs, ensuring better and longer compliance with treatment, or reducing opportunistic infection by improving nutritional status. In the case of integrated food security programming, e.g. Oxfam and CRS programmes in Malawi, it would be necessary to control for a number of external factors like rainfall/climate, agricultural inputs and initial wealth status, in order to draw conclusions regarding the role of food aid in strengthening longer term food security. Generally, there has been insufficient attention given to impact assessment methodology for emergency related programming. Apart from the methodological challenges, there are also difficult ethical considerations in impact assessment. Given the large numbers of unknowns with respect to food aid and nutritional support in the context of HIV programming, there is an urgent need to strengthen the ability of implementing agencies (especially CBOs and local NGOs) to undertake impact assessment.

Anecdotally, IP staff and beneficiaries indicate enormous value of food aid as a component of the many types of HIV related programming. Benefits cited touch on many of the objectives claimed for these types of programming, e.g. weight gain, improved well-being, less mortality, improved programme compliance, longer-term food security. With the exception of longer-term food security, there is no reason to disbelieve these findings, especially when one considers that much of the targeting is to the poorest of the poor (see section on targeting). However, we have to do better than relying on anecdotal reports of programme benefits in order to strengthen understanding of how programmes are working (what is the process underpinning impact) so that programme design can be strengthened, and so as to increase belief in this type of programming - particularly in an environment where donors are increasingly sceptical of the benefits of food aid/HIV programming.

Targeting criteria, stigma and resources

A great deal of food aid targeting within HIV programming currently takes place on the basis of proxy indicators, i.e. chronic illness, orphan containing households, single parent/widow headed households, elderlyheaded households, child headed households (child less than 18 years old) and households with children of deceased HBC clients. These categories have been employed partly to address issues of stigma which is still a major problem - particularly in Africa. However, there are important reservations around the use of proxy indicators. For example, what is the inclusion error if the purpose is to target those with HIV? Do such indicators really address issues of stigma (people aren't stupid and quickly realise that chronic illness probably means advanced stages of AIDS). Furthermore, and perhaps more significantly, many argue that HIV/AIDS cuts across all income classes so that targeting the chronically ill or orphan containing families is not an equitable means of targeting resources. Although this is somewhat contradicted by the CRS Dedza research in Malawi (see page 7), it is supported by other studies, e.g. Seaman/Petty (Field Exchange 23) and Garnier/Situmu /Watkins (WFP /REEP, this issue). In a pilot programme in Malawi (Fisher and Munk), ACF employed a food requirement/dependency ratio indicator, which the authors argued is more equitable than proxy indicators like CI or OVC. Another difficulty with the proxy indicator approach has been that, in the face of limited food aid resources, many programmes have had to make tough choices as the number of CI or orphan containing households at village level has exceeded food supply, e.g. this has created tensions and conflict for village committees charged with targeting responsibilities and implementing agencies. REEP in Kenya had to scale down food aid programming during the hunger gap period. Some agencies, recognising the inequity of targeting on the basis of proxies for HIV - especially in extremely food insecure and chronically poor areas -have used a two tier system, i.e. used proxies in conjunction with indicators of economic/food security status. Relying on the community is, clearly, key for economic or food security based targeting. Generally, in the programmes reported here, the community does target the poorest of the poor.

Targeting food aid to PMTCT/ARV/DOT programmes is, undoubtedly, an efficient means of providing nutritional support to the HIV infected. However, although consensus regarding the differential nutritional requirements and rations for those infected with HIV has not been reached, there is some agreement on the need for differential rations according to stage of disease. For example, it is believed that a 10% increase in energy intake is required to maintain nutritional status and avoid weight loss of asymptomatic individuals living with HIV, while those with AIDS related illnesses require a minimum of 20% increase in energy intake. Further research is needed on optimal protein requirements during the course of HIV disease, which may be effecincreased due to nitrogen loss associated with opportunistic infections (for the moment, current guidelines (WHO, FANTA) advise to at least the meet protein intake required of a balanced diet). Agencies like WFP need to monitor emerging consensus and expert opinion on this and adjust rations accordingly. At the same time, implementing ration differentials will be extremely challenging from a logistical perspective.

Finally, despite the concerns that targeting PLWHA will cause problems of stigmatisation, the evidence does seem to show that CBOs and NGOs that invest in community sensitisation manage to reduce stigma. Agencies like REEP appear to have done an extraordinary job in building self-esteem of those infected and enabling them to speak freely about their status. The increase in numbers of those coming to be HIV tested in REEP programme areas is testimony to this.

Exit criteria

Provision of food aid within the context of HIV programming brings with it the enormously challenging issue of when to stop giving out food. Many of the programmes described have not invoked exit criteria, (e.g. REEP² and Shoham J³) and some beneficiaries have been food aid recipients for almost three years, e.g. Reach Out. Although there has been much discussion in the literature, and within implementing agencies, about exit criteria and discharging recipients on achieving a level of food security, the reality is that there are no simple tools to assess when that has been achieved. The only food security assessment tools widely employed are HEA and, more recently, CSI. Although HEA appears a promising tool in food security impact assessment, the approach requires extensive field based training and cannot be rolled out rapidly and on a large scale, especially if the status of individual households is to be assessed. Furthermore, as many beneficiaries are extremely poor (indeed they are targeted on this basis), the time-scale for achieving a level of food security may be entirely unrealistic for many of these programmes. In reality, what is being proposed is a form of development which, in many cases, is not something that can be squeezed into the time-frame of a PRRO or the 'attention span' of a donor food aid programme. Currently, and unsurprisingly, there are no data in the published or grey literature on proportion of food aid beneficiaries in this type of programme who are being discharged (exiting the programme). Progress in this area can probably only be made in an extremely piece-meal fashion, with implementing agencies piloting food security impact assessments in a few locations (constructing baselines and conducting subsequent assessments) and then determining what proportion of households have achieved a target level of food security following the intervention.

While discharge may be far easier for TB and PMTCT type programmes as these have a natural end point, e.g. completion of treatment or cessation of breast-feeding, adopting exit criteria may remain problematic even for these, as many beneficiaries will effectively be discharged back to the poverty and food insecurity from whence they came. Ideas of promoting food security in a 12 or 18 month time frame are probably quite unrealisable and certainly, as yet, untested.

There are also challenges about nutritional entry and exit criteria for this type of programming. Thus, in Malawi, MSF have been feeling their way with BMI cut-offs for ARV/nutritional support and have adjusted these on the basis of experience during the programme. However, MSF readily acknowledge that newly planned programmes for 7- 14 year olds on ARV will raise fresh challenges in terms of identifying appropriate anthropometric-based entry and exit criteria. We still lack experience of adult feeding programmes targeted at HIV infected individuals, especially with regard to length of time it takes for improvement, proportion expected to improve, optimal rate of improvement. This is even true for malnourished HIV positive children. Work by ACF in Malawi shows that not only were 30% of children admitted to NRUs HIV positive, but their outcome was poor, with almost half discharged failing to respond.

Integrated programming

Provision of food aid on its own to PLWHA is unlikely to have a marked impact on nutrition, morbidity, survival, food security, and general quality of life. Although this is understood by most agencies, there are still a large number of food aid and nutritionally focused HIV programmes where integrated programming is proving very difficult, (see Shoham. J, CRS Malawi, Oxfam Malawi, MSF Luxembourg) Research summarised in this issue (CTC/VALID) shows the needs of PLWHA and HIV affected households were found to cover ten different categories. These categories were ranked by participants under two groups, primary (nutritional, material, medical, economic, psycho-social, spiritual and patient care) and secondary (food security, knowledge on prevention and vocational skills). The ranking of needs reflected a prioritisation of immediate over long-term needs. More experienced and better resourced agencies like Oxfam have successfully established linkages with other programmes, e.g. IGA, but have still found it hard to strengthen linkages between HBC and health infrastructure. More medically focused agencies like MSF, which work through health care infrastructure, have struggled to integrate longer-term food security initiatives. What is clear is that fully integrated programming, whereby the multi-sectoral needs of PLWHA are addressed in a coordinated fashion, is not something that can be established easily.

While scepticism still abounds regarding the role of food aid in HIV programming, there is a need for well conceived and implemented pilot studies/programmes of multisectoral programming which can demonstrate how effective food aid can be within an integrated package. These experiences should next allow for analysis of how to effect integration and in which contexts good integrated programming can take place. Guidance material for implementing agencies should then follow. It may be that some form of mapping can subsequently take place at country level, to determine where integrated programming is feasible and to target resources (including food aid) on this basis. The need to rationalise food aid and nutritional support programming to situations where there can be a significant and sustainable impact (i.e. where a level of integrated programming can be assured) is key, given the logistical challenges posed by the upsurge of this type of programming in subsahelian Africa. The fact that this type of programming is so decentralised i.e. implemented through small CBOs, schools, health centres, means that costs per tonnage of food delivered is relatively high compared to more traditional types of food aid programming, e.g. general rations or supplementary feeding. Economies of scale no longer apply. Therefore, targeting food aid and nutritional support resources to locations where there is most likely to be an impact (where programming is integrated) will reduce inefficient logistics. It is also likely to have a greater influence in convincing donors that this type of programming is worthwhile.

This special issue also carries a large number of research summaries (published and grey literature) on topics related to food aid, nutrition and HIV programming. Subjects range from BMI as a predictor of survival in HIV/AIDS affected adults, to the role of community based technology to combat the impact of HIV/AIDS on food security and livelihoods. Other subjects include the impact of food aid on survival and nutritional status in Bangwe, Malawi and the prevalence of HIV in children admitted to NRUs in Malawi comparing their recovery performance to HIV negative children.

The range of research is a stark reminder of how much we don't know, as well as how many people/institutions are doing research out there. A lot of this research may not make the published literature. Indeed, the recent IFPRI conference in Durban (11th- 13th of April)⁴ received an enormous number of abstracts of ongoing research. Clearly, many of those who submitted these abstracts will need support and funding to ensure that correct research protocols are followed and that findings can be written up and disseminated.

Conclusion

There is an argument that 'while there is little empirical evidence regarding the effectiveness of food aid in responding to HIV/AIDS, this should not forestall action' and that 'a well documented learning-by-doing approach is required of building up, evaluating and disseminating experiences and lessons learned' (Gillespie S, 2004⁵). However, there may be dangers with such an approach. What this issue of Field Exchange has shown is that there are poorly described objectives and limited, if any, impact assessment of many programmes using food aid within an HIV programming context. Indeed, impact assessment in the food security sector is notoriously difficult with tools and capacity poorly developed. There are also huge challenges with regard to targeting PLWHA, not least of all whether it is equitable to target food resources on the basis of HIV status or some proxy indicator. Perhaps even more significant is the issue of when to stop food aid. Certainly, food aid is unlikely to do any significant harm (except maybe tie up implementing agency resources or undermine volunteerism where carers are targeted). However, by definition, food aid can only be seen as a shortterm resource transfer. Longer-term development objectives, which are increasingly being claimed for food aid, may in most situations be unrealistic, so that the exit criteria being invoked are unlikely to be applied. We need to watch this closely. There is something very unconvincing about claims that two or three years of food aid, in conjunction with a wreath of livelihood initiatives, will help secure a significant shift in food security.

There are also urgent issues to be addressed around poorly integrated programming, as well as the logistical challenges of this kind of decentralised food aid programming. Limited evidence (although it is probable that more could be made available) suggest that logistical costs are high compared to more standard emergency food aid programming. Furthermore, although we don't have data, it is likely that where programmes are poorly integrated with other services, impact will be limited. While there is little doubt that many very poor people are benefiting from the food aid - after all it is a resource transfer, and this explains why so many implementing agency staff and beneficiaries are positive about the programmes - the principle question we should be asking is 'could the money involved be spent more effectively'? At this stage it would seem eminently sensible to stick with small-scale pilot programmes using food aid as part of an integrated package of support, and to monitor and document the impact food aid can have in different types of programme and context. This may then place us in a better position to target food aid effectively as part of HIV programming, rather than roll out large-scale programmes and hope to learn something from them after the event. There is absolutely no reason to make food aid the proverbial 'cart' before the horse.

Jeremy Shoham (editor)

Any contributions, ideas or topics for future issues of Field Exchange? Contact the editorial team on email: office@ennonline.net

¹See field article, Nutritional support through HBC in Malawi and the research piece, WFP recipients' weight gain at Reach Out clinic, in this issue.

²See field article, REEP experiences in Western Kenya, by Mary Corbett

³See field article, WFP HIV/AIDS programming in Malawi

⁴See http://www.sahims.net/archive/specialfocus/ specialcoverage_who_consultation2.htm

⁵Kadiyala S and Gillespie S (2004). Rethinking food aid to fight AIDS. Food and Nutrition Bulletin, vol 25, no 3, pp 271-282. See research summary in this issue.