Geostatistical modeling of chronic respiratory diseases and causative factors in Kandy District, Sri Lanka | |
| Author | Kumarihamy, Rajapaksha M. K. |
| Call Number | AIT Diss. no.RS-18-04 |
| Subject(s) | Geological modeling Respiratory Tract Diseases--Sri Lanka--Kandy Geology--Geographic information systems--Health aspects |
| Note | A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Remote Sensing and Geographic Information Systems |
| Publisher | Asian Institute of Technology |
| Abstract | Prevalence of chronic respiratory diseases (CRDs) is an emerging health problem in Sri Lanka. Beside proper medication, primary prevention of CRDs requires averting the level of exposure of the population to bioclimatic, environmental and socioeconomic risk factors. A system which can routinely monitor spatiotemporal trends of CRDs and their risk factors is therefore deliberate as an essential component in disease surveillance. The tremendous capabilities of geographic information systems (GIS) and spatial epidemiological approaches are providing innovative ways to harness such disease data. This study explores the potential of geo-statistical capabilities of GIS for monitoring CRDs and their risk factors in Kandy district, Sri Lanka. Daily inpatient records of CRDs were obtained from 15 different hospitals in Kandy district for the period of 2010 to 2014. Based on the patient's addresses they were geocoded and aggregated into Grama Niladari divisions (GNDs) in the study area. Average nearest neighbor analysis, Moran's I index, Getis Ord Gi* statistics and Mann Kendal trend test were employed to assess spatiotemporal patterns of CRDs. Ordinary least square (OLS) and geographically weighted regression (GWR) models were used to explore the associations between a wide array of environmental, climatic, socioeconomic risk factors and CRDs incidence at GN division level. Experts' knowledge based analytical hierarchy processes and fuzzy logic based multi-criterion spatial decision approach were used to model the magnitude of vulnerability, exposure, and risk of CRDs. A total of 44088 CRDs incidents recorded with an average annual CRDs rate of 6411 0000 population. Out of that 33384 incidents were geocoded to a relatively good matching score (76%). The results of the spatiotemporal pattern analysis revealed certain urban and semi urban areas in the study area as statistically significant hot spots. The majority emerged as intensifying, diminishing, persistent and oscillating hot spots. The main findings of the regression analysis suggested that relative humidity, proximity to roads, road density, use of firewood as a source of fuel and altitude are potential predictors of CRDs exacerbation in Kandy district. However, the strength and direction of the relationship between the predictors and CRDs are spatially heterogeneous. Thus GWR exhibited a better prediction capability than OLS models. The proposed risk mapping approaches were able to ascertain an accurate level of vulnerability, exposure, and risk. The extremely high risk areas were found in the western part of the Kandy district. The vulnerability, exposure and risk maps were designated the locations that should consider in the primary prevention activities. This study had a few limitations as well. First, this was an ecological study; the models used aggregated data at GN division level. Thus individual level exposure to the risk factors and impact of personal mobility were underestimated. Secondly, neither outpatient visits nor patients at private hospitals considered for the study. Therefore the actual amount of morbidity and mortality is likely underestimated. Thirdly, precise spatial and temporal data on risk factors were limited in the study area. All of these limitations contributed to modeling errors. Beside, this study allows health care policy makers to understand spatial characteristics of CRDs that can help to make appropriate prevention interventions aimed at reducing the burden of CRDs prevalence in Kandy district. Further, this study demonstrates the importance of geo-statistical capabilities in developing disease surveillance strategy not only for CRDs monitoring but also any other disease monitoring in the different spatial setting. |
| Year | 2018 |
| Type | Dissertation |
| School | School of Engineering and Technology (SET) |
| Department | Department of Information and Communications Technologies (DICT) |
| Academic Program/FoS | Remote Sensing (RS) |
| Chairperson(s) | Tripathi, Nitin Kumar |
| Examination Committee(s) | Sarawut Ninsawat;Samarakoon, Lal;Strobl, Josef |
| Scholarship Donor(s) | National Center for Advanced Studies in Humanities & Social Sciences (NCAS), Sri Lanka AIT Fellowship |
| Degree | Thesis (Ph.D.) - Asian Institute of Technology, 2018 |