This study addresses the application of spatial approaches to epidemiology. It pays particular attention to disease surveillance and intervention with regard to four different aspects. First, spatial epidemiological analyses start with having patient level spatial data recorded upon hospital consultation or admission. In the absence of fine level spatial data, only coarse level spatial analyses are possible. Such coarse level data from the District Health Information System (DHIS2) was used to investigate the spatial variation of the Human Immunodeficiency Virus (HIV) and Tuberculosis (TB) infections in Uganda. From a surveillance standpoint, such data mask fine-scale heterogeneity and the role of local contextual factors which derails intervention and control programmes, especially for infectious diseases, as identification of target foci of transmission and local risk factors are obscured at larger scales.
Secondly, to mitigate the effects due to coarse-level data, we developed a prototype health registry system that allows for the capture of fine-level patient spatial details. The system uses ‘pool’ spatial data that have been collected by the water distribution company in Uganda (NWSC). An Android application developed can be used in the recording of geo-referenced patient details upon admission, allowing for fine-level spatial epidemiological analyses.
Thirdly, such epidemiological analyses use more specialised tools to investigate the association between health outcomes and environmental risk factors. Using data from cardiovascular disease (CVD) and air pollution, the association between air multi-pollutants and CVD hospitalisation, and the spatial variation of these associations across Sweden was investigated using Poisson-based spatial regression models. The results provided a basis for the recommendation of more locally oriented intervention strategies concerning CVD and air pollution.
Finally, for diseases with multiple risk factors, the effectiveness of intervention lies in accurate quantification of relationships between the risk factors and the health outcome. We demonstrate this aspect by evaluating the relative spatial contribution of Linking Social Capital to elderly mortality in Sweden, evaluated at the neighbourhood level. Given the often non-linear relations between most diseases and their risk factors, we employed artificial neural networks.
The results of this study demonstrate how spatial approaches to epidemiology provide roadmaps to inform healthcare policy and resource planning through spatially-enabled records and identification of areas with elevated disease risk condition on the place-specific risk factors.
I Augustus Aturinde will be holding a PhD STUDENT seminar at KYAMBOGO UNIVERSITY, Faculty of Engineering, Department of Lands and Architectural Studies (DLAS).
This is, therefore, to inform and invite you to attend.
Venue: DLAS, KYAMBOGO UNIVERSITY
Date: Tuesday, 28th May 2019