Exploration of Relationship between Social Status and Mortality Rates in England

It is well established that health is related to Socio-Economic Etatus (SES) or deprivation status (Adler (1994)). Socio-Economic models or multiple deprivation models of health and inequalities are widely used by Public Health practitioners. Since 2000, the Department of Health has developed a pathway, which approach to determine the important influences on health within a population. A lot of researchers think social position or socioeconomic status as the fundamental cause of ill health (Davey Smith and Hart (1998)). There is now a large literature which shows that the socioeconomically disadvantaged experience higher mortality rates for most major causes of death, and this inequality exits at every stage of life-course (Turrell and Mathers (2001)). In this paper, three different approaches, Ordinary Least Square (OLS) regression, Geographically Weighted Regression (GWR) and Artificial Neural Network (ANN) (Rojas (1996)), are applied for exploring the association between the Socio-Economic status or deprivation and mortality data in England. Researchers, with the spatial data being in use, are increasingly aware of the limitation of global regression techniques, such as Ordinary Least Square regression, which by generating ’global’ outputs may mask local variations and ignoring the spatial relationship between variables (Huang and Leung (2002)). GWR, a local regression technique, has been applied in many instances to successfully reveal local relationship as well as spatial association between variables (Fotheringham et al. (2002)). However, the linear functional relationship of the OLS and GWR assumptions can not totally reflect reality. ANN, as a powerful tool and robust predictor which is capable of modelling nonlinear relationships, is used gradually in health research. This research attempts to compare with the results among three different methods to provide guidelines for future analysis of socioeconomic status and mortality.