An Ecological Analysis of the Impact of Weather, Land Cover and Politics on Childhood Pneumonia in Tanzania

Pneumonia is the main killer of under-five children worldwide. The developing nations suffer the most. But within such countries, the spatial and temporal distribution of pneumonia cases is not uniform; yet little is known of the spatial and temporal distribution of pneumonia or the factors that might affect spatial and temporal variability. This dissertation explores the causes of spatial and temporal variation in under-five pneumonia morbidity in Tanzania.

This study uses an ecological analysis to explore weather, land cover and politics as potential drivers of the observed differences in the distribution of pneumonia. A study is at an ecological level when it examines the population-level health aspects. That is, ecological analyses in health studies evaluate groups of people rather than individuals.

The current study found that weather variables such as temperature and atmospheric pressure partially explained pneumonia variance. The strength of weather-pneumonia association varies over space and time in both seasonal elements (temporal factors) and broadly-defined climate zones (spatial factors). For example, the prevalence rate was higher in the regions with bimodal rainfall compared with the regions with unimodal rainfall, with a statistically difference 117.3 (95% confidence interval: 36.6 to 198.0) cases per 100,000. In addition, within the regions (mikoa) with unimodal rainfall regime, however, the rainy season (msimu) had lower rates of pneumonia compared to the dry season (kiangazi).

Land use and land cover also were partial drivers of pneumonia. Some land cover types?particularly urban areas and croplands?were associated with high rates of childhood pneumonia. In addition, districts (wilaya) categorized as urban land cover had high rates of pneumonia compared to those categorized as only rural.

To determine the associations between politics and pneumonia, this study compared the pneumonia cases in the administrative locations that received less central government funding with those locations that were financially rewarded for voting for the ruling party. The locations with lower funding generally had higher rates of childhood pneumonia. However, it is unclear whether these locations had higher rates of childhood pneumonia because of, or in addition, to their funding gaps.

In sum, this dissertation evaluated population-level factors affecting distribution of childhood pneumonia. Like other similarly population-level studies, this dissertation provides an understanding of the coarse-scale dynamics related to childhood pneumonia. By so doing, it contributes to the pneumonia etiology scientific literature.

That is, this dissertation contributes to the understanding of within-nation pneumonia distribution in developing nations. It is the first in Tanzania to evaluate the impact of weather, land cover and politics on childhood pneumonia. By evaluating the impact of weather and land cover, this dissertation also provides an example of non socio-economic factors affecting health inequalities. By analyzing a large landmass of two main climatic types, this dissertation also contributes appreciation of non-stationarity of temporal variations of childhood pneumonia, in addition to the commonly-evaluated spatial variations.