Urban agriculture and operational mosquito larvae control. mitigating malaria risk in Dar es Salaam, Tanzania

Global commitment, political will and financial support to reduce the burden of malaria, a disease which kills about one million people each year, have reached an unprecedented level. Although global malaria eradication appears to be a distant goal, there are promising efforts towards regional control and local elimination of the disease. Sub-Saharan Africa (SSA) is the region with the world’s highest malaria burden, as well as the world’s fastest growing cities. Rapid urbanisation brings enormous challenges such as increasing poverty and malnutrition, the spread of informal settlements, inadequate sanitation and lack of access to safe water, as well as health risks. Malaria transmission in urban settings, although to a lesser extent than in rural areas, remains a significant problem. In Tanzania and many other developing countries, one response to increasing urban demand for food is urban agriculture (UA), especially in backyard gardens and larger open spaces. While urban farming is an integral part of livelihoods and food security, it may also increase the risk of vector-borne diseases, including malaria, by providing suitable grounds for mosquito breeding. However, it is unclear whether agricultural land use or rather geographical factors such as topography and hydrology lead to the creation of habitats for malaria vector mosquitoes. The overall aim of this study was to contribute to a better understanding of the importance of urban agricultural land use in the context of environmental malaria control in urban SSA, and identifying potential mitigation options. All field research for this study was conducted in 2005 and 2006 in the city of Dar es Salaam, Tanzania, and was fully integrated in the operational Urban Malaria Control Programme (UMCP), managed by the Dar es Salaam City Medical Office of Health. The main goal of the UMCP is to control mosquito larvae with the application of biological larvicides by community-based resource persons. The transdisciplinary research approach used in this study was a combination of natural and social scientific approaches, and took account of the knowledge and capacities of stakeholders in the context of the UMCP. Quantitative methods included participatory mapping of administrative boundaries, cross-sectional mapping of agricultural areas, and surveys of farmer’s agricultural practices. Qualitative methods consisted of semi-structured key informant interviews and Focus Group Discussions using interview guidelines. The theoretical framework that guided the study design and data analysis was based on the concepts of risk, resilience and threat, where risk is seen as a function of threat and resilience. In order to gain a more comprehensive understanding of factors related to malaria risk (i.e. malaria infection and disease), aspects of malaria threat (i.e. presence of breeding sites with potentially malaria-transmitting Anopheles mosquitoes) and resilience to malaria (i.e. social resilience in terms of increased competence of stakeholders in the context of the UMCP) were investigated. In a first step, the current extent and characteristics of urban agriculture were assessed. This revealed that UA is widespread in Dar es Salaam, with more than 5% of the study area used for farming in 2005. Extrapolated to the total urban area, this corresponds to almost 20 km2 of land used for urban agriculture in Dar es Salaam. Furthermore, an analysis of the spatiotemporal changes of urban agricultural use showed that urban farming is a dynamic, but not a short-lived or transitional phenomenon. The overall extent of UA did not decrease during the last two decades, despite the city’s rapid growth and densification. UA thus appears to be resilient in terms of its ability to persist in the presence of high non-agricultural pressure on land use, and its ability to recover after disturbances by spatial shifts. In a second step, a participatory mapping method was developed that allowed linking the agricultural data from 2005 with mosquito larval data from the UMCP. This was necessary in order to assess correlations of these agricultural areas with Anopheles breeding sites. At the same time, this procedure facilitated comprehensive routine larval surveillance by the UMCP. The approach involved basic use of aerial imagery and Geographical Information Systems (GIS), and was validated by mapping three urban wards of Dar es Salaam, covering an area of 16.8 km2. The procedure enabled verification and correction of cognitive sketch maps drawn by UMCP field staff as guidance for their work, and therefore complete coverage of targeted areas with larval control. It proved to be practical, affordable, and requires minimal technical skill. Thus, it can be readily integrated into malaria vector control programmes, scaled up, and adapted to urban settings elsewhere in Africa. In Dar es Salaam, the participatory mapping approach became an integral part of the novel management, monitoring and evaluation system for implementing routine larviciding developed by the UMCP. Therefore, it partly contributed to the reduction of malaria transmission by the primary vector, Anopheles gambiae s.l., by 31% (95% CI=21.6-37.6%; p=0.04). In a third step, the agricultural data was linked with the UMCP larval database. This revealed that the proportion of habitats containing Anopheles larvae was 1.7 times higher in agricultural areas compared to other areas (95% CI: 1.56-1.92). Significant geographic predictors of the presence of Anopheles larvae in gardens included location in lowland areas, proximity to rivers, and relatively impermeable soils. Significant agricultural predictors comprised specific seedbed types, mid-sized gardens, irrigation by wells, and cultivation of sugar cane or leafy vegetables. Significant negative predictors included small garden size, irrigation by tap water, rainfed production, and cultivation of leguminous crops or fruit trees. Although there was an increased risk of finding Anopheles larvae in agricultural sites, breeding sites in urban agriculture accounted for less than a fifth of all breeding sites of malaria vectors in Dar es Salaam. The identified indicators of relatively high and low odds of larval presence in agricultural areas can help improve the effectiveness of larviciding interventions. In a last step, in order to explore mitigation measures implemented within the UMCP, capacity building processes on various programme levels were analysed. By applying the Multi-layered Social Resilience framework in the context of the UMCP using a qualitative approach, it was found that exchange between and within administrative levels supported resilience-building processes in terms of mosquito breeding site elimination. “Reactive” and “proactive” capacities were successfully built among programme staff. However, more potential could be tapped among local leaders and household members, by increasing their competence in eliminating breeding sites of malaria vectors. Improving the communication skills of the programme's field workers might support such processes. Together with local leaders, they could act as multipliers of sensitisation messages. In conclusion, this study showed that urban agriculture is not only widespread in Dar es Salaam, but also a potential malaria risk with regard to breeding sites for Anopheles mosquitoes. Urban farming therefore needs to be considered by integrated vector control programmes. However, it should not be overemphasized; rather farmers should be regarded as potential assets in vector control. For example, they could be involved by planting shade trees near water bodies in agricultural areas. More generally, mitigation strategies related to environmental malaria control could build on increased participation of household members within the UMCP intervention area. This may be achieved by building capacities for breeding site elimination, for example through enhanced sensitisation of household members provided by UMCP field workers. These insights, gained by conducting transdisciplinary operational research, can provide a basis for optimising malaria control and urban planning in Dar es Salaam and other malaria-affected SSA cities with comparable climatic conditions.