Application of the chi-plot to assess for dependence in environmental epidemiology studies

Abstract

Several studies have illustrated the short-term effects of environmental factors such as air pollution on health, using a wide range of time-series regression models. However, there have been few reports of descriptive analyses based on data visualization, which can be used for a preliminary inspection of relationships between health outcomes and exposure factors. The commonly used scatterplot, displaying a large number of points, cannot adequately assess complex bivariate dependence structures. This paper presents a graphical display, the chi-plot, which, used jointly with the scatterplot, provides a powerful tool to describe a varied range of bivariate dependence structures in environmental epidemiology studies. We studied the relationship between emergency room admissions for asthma, air pollutants and pollen types in Madrid (Spain) during the period 1995–1998. Associations were depicted graphically using chi-plots and were consistent with those previously reported.