Socioeconomic And Environmental Covariates Of Respiratory Health In Greece

Combating respiratory diseases has been and still is one of the major challenges on human health at the global scale. According to a report by the World Health Organization in 2000, lower respiratory tract infections, chronic obstructive pulmonary disease and tuberculosis are among the ten leading causes of death worldwide. Greece is no exception to this trend and as a consequence, in coming decades, respiratory diseases will impose an enormous burden to the nation.Getting a clearer understanding of the etiology of such diseases is a critical first step to control and reduce the number of incidences. Many studies have demonstrated that respiratory health varies by gender, age, socioeconomic status and lifestyle behavior (physical activity, smoking habits and alcohol consumption). However, the majority of the studies conducted so far have adopted a classical epidemiological stance, taking no account of the variability of health outcomes and their potential determinants over space.This paper illustrates, at a national level, that spatial analysis can be applied to reveal the potential associations between the increased morbidity-mortality rates of respiratory diseases and various socio-demographic and other confounders. For the analysis, the data utilized hospital admissions due to respiratory diseases and a number of socioeconomic and demographic characteristics from the National Statistics Service in Greece at the prefecture level for 1991. Exploratory methods such as kernel estimation and the Moran's I statistic were applied in order to investigate the global trends and the geographical variability of respiratory incidences in the country. Results of the analysis confirmed the presence of spatial autocorrelation in the data, suggesting the consideration of space either by the development of a Simultaneous Autoregressive model (SAR) or by using spatial interaction terms in simple linear regression models. The application of the models showed that the geographic variation of respiratory incidences is strongly associated with socioeconomic status and levels of air pollution. We conclude that spatial analysis of health data holds promise and in combination with time series analysis can lead to the development of superior spatiotemporal methods for the study of respiratory incidences.