[Methods for analysing spatial data : Using the example of skin cancer prevalence in Germany].

Abstract

Background: Studies show a spatial variation in skin cancer frequencies, but the causes have not been fully understood. When analysing spatial patterns and correlations, different approaches need to be considered, as the data have special features due to their spatial structure.

Aim of the study: Why the spatial consideration of skin cancer frequencies is important at all, and which methods of analysis are useful is considered in more detail in this article. The article provides an introduction and overview of statistical methods that are important for the spatial analysis of skin cancer frequencies.

Materials and methods: At first it is shown which descriptive methods, such as statistical smoothing, can be applied. Next, spatial cluster and regression analyses will be discussed. Testing for spatial autocorrelation will be considered.

Results: In particular, the spatial dependence of the data on neighbouring regions, which if ignored can lead to biased estimates, is of particular importance in the analysis of spatial data and requires special spatial analysis methods.

Discussion: This article provides an introduction and overview of statistical methods relevant to the spatial analysis of skin cancers.