Geostatistical Analysis of Health Data: State-of-the-Art and Perspectives

Abstract

The analysis of health data and putative covariates, such as environmental, socio-economic, behavioral or demographic factors, is a promising application for geostatistics. It presents, however, several methodological challenges that arise from the fact that data are typically aggregated over irregular spatial supports and consist of a numerator and a denominator (i.e. population size). This paper presents an overview of recent developments in the field of health geostatistics, with an emphasis on three main steps in the analysis of aggregated health data: estimation of the underlying disease risk, detection of areas with significantly higher risk, and analysis of relationships with putative risk factors. The analysis is illustrated using age-adjusted cervix cancer mortality rates recorded over the 1970–1994 period for 118 counties of four states in the Western USA. Poisson kriging allows the filtering of noisy mortality rates computed from small population sizes, enhancing the correlation with two putative explanatory variables: percentage of habitants living below the federally defined poverty line, and percentage of Hispanic females. Area-to-point kriging formulation creates continuous maps of mortality risk, reducing the visual bias associated with the interpretation of choropleth maps. Stochastic simulation is used to generate realizations of cancer mortality maps, which allows one to quantify numerically how the uncertainty about the spatial distribution of health outcomes translates into uncertainty about the location of clusters of high values or the correlation with covariates. Last, geographically-weighted regression highlights the non-stationarity in the explanatory power of covariates: the higher mortality values along the coast are better explained by the two covariates than the lower risk recorded in Utah.