Agricultural land use patterns and malaria conditions in Kenya.

This paper studies the problem of malaria control in irrigation and non-irrigation areas in Kenya. Empirical results show that in both areas, households' level of awareness of malaria as a health problem, including its cause, was very high. However, attempts to trace the direct effects of malaria upon income or upon agricultural production were not statistically important. This does not imply that malaria has no consequence on household welfare. It is possible that the model equations were mis-specified--aggregate variables (total family size, total family income) and failure to quantify land in the production relationships may have contributed to these results. In addition, poor separation of malaria as a disease, from malaria as an infection, may have underestimated the effect of the disease on production. Thirdly, labour substitution (hiring or within-family substitution) was not measured in this early study, but was taken into account in subsequent research. Finally, labour requirements in the annual crop production schedules and the co-relation between these labour requirements and the pattern of adult morbidity were not longitudinally monitored. Cross-section data would bias the findings, particularly in those areas where the peak transmission season is short, where the crop grown does not require major labour input during this transmission season, and where acquisition of immunity would reduce the clinical impact of malaria upon adult labour. These vulnerabilities in the specification of the model and the data collected, probably affect the results obtained. Our empirical work raises a number of interesting and important questions which should be taken into account in future research.