Comparison of spatial auto correlation analysis methods for investigating the distribution pattern of diabetes type II patients in Iskandar Malaysia

Abstract

Spatial statistics have been widely used in epidemiology studies in order to investigate and monitor the outbreak in endemic area. However, there is less application of spatial statistics in the study on built environment and diseases outbreak. It is significant to conduct such a study since non-communicable diseases have been rising over past four decades and have a strong positive relation to the built environment, particularly in the rapid urbanizing area. Therefore, this study aims to measure the geographical distribution and determine the pattern of the patients in urbanizing area. Two methods of Moran’s I and Getis-Ord G Statistics, that have been used extensively in the epidemiology are compared on diabetes type II patients data both in global and local scales. A total of 496 patients diagnosed with diabetes type II in 154 neighbourhoods of Iskandar Malaysia (IM) have been evaluated. This study compares the results of two methods based on built environment criteria. The study evaluates their applicability is such a study to identify the best method and scale to be considered in study on built environment-related epidemiology.