Understanding HIV/AIDS and livelihoods: new cluster analysis

Summary of published research ¹

Discussion with a group of young women in Teta, Central Province, Zambia.

A recent Overseas Development Institute (ODI) briefing paper describes the findings of longitudinal research carried out in Zambia in 1993 and 2005 to establish the effects of HIV/AIDS over time on individuals, households and communities. The research focused on two locations in Zambia, Mpongwe - close to the Copperbelt towns, and Teta, a remote rural area. Ten days were spent in the communities collecting data using participatory methods. The study was unique in that it used a 'cluster' approach to help understand how individuals and households had either adapted or disintegrated as a result of ill health, and specifically HIV/AIDS.

A cluster can consist of various households, usually, though not necessarily, living in the same geographical area. A significant benefit of the concept is that it allows the most important relationships between individuals of different generations and gender, marital and kinship statuses to be identified and understood. In the approach, data are collected from individuals, but the information is used to identify clusters and to place individuals within them. The 1993 study developed a typology of clusters according to relative resilience and vulnerability. The types formed a continuum from commercial (Cluster Type 1s) to resource-poor (Cluster Type 4s). The aim of the 2005 re-study was to trace the original clusters and see whether they had changed status within the cluster continuum and to understand what had caused the moves. More specifically, it was concerned with the role that AIDS had played in any shifts.

The study found that the category of person( s) who had died in the cluster was found to be a key determinant of the effect on the level of resilience of the cluster. If, for instance, it was a primary producer, the future of the cluster was likely to be threatened. Analysis showed that in Mpongwe, although there were 1.5 AIDS deaths per cluster in the 12 years, maize production had increased. However, this was mainly amongst the wealthier or those higher up the cluster continuum. Livelihoods had suffered in Mpongwe due to animal disease. In Teta, where AIDS deaths per cluster were only 0.72, the effects on individuals and clusters were not heavily evident in 2005. What was in evidence was the change in farming systems, with less reliance on maize and a return to more traditional crops.

Generally, factors affecting the relative vulnerability and resilience of a cluster to the impact of a death were:

• Length and degree of incapacity during AIDS-related illness
• Health status of surviving spouse
• Characteristics and stage of life-cycle of the primary producer
• Overall cluster composition
• Livelihood and agricultural production opportunities.

Another key finding of the study was that the inherent adaptability and flexibility of the matrilineal social system enables it to accommodate deaths and changing economic and demographic circumstances. For example, once a spouse dies, the widow or widower can easily become absorbed into another cluster, not only by drawing on kinship ties but also through marriage.

The ODI paper sets out a number of lessons for those making policy and designing programmes. For example;

• In programmes working with the poor, targeting should be directed at resource-poor clusters rather than poor households.
• For targeting, a broad based multi-faceted definition of vulnerability is needed, one that is not just AIDS-related.
• While national food security policy has focused on maize intensification, crop diversification remains important for household food and nutritional security.
• Diversification within and outside agriculture contributes to resilience because it allows farming systems to be adjusted depending on whether and when labour is available.
• Education about HIV/AIDS is using culturally inappropriate communication channels, particularly by failing to differentiate between HIV and AIDS.
• Appropriate messages need to be targeted at appropriate people. Thus, older women in rural Zambian societies are the channels for transmitting cultural norms about sex and sexuality.

¹ ODI (2006). Understanding HIV/AIDS and livelihoods: The contribution of longitudinal data and cluster analysis. ODI Briefing Paper, August 2006