Spatial models

In the previous chapters, we assumed the independence of observations, conditional on the model's fixed and random effects. The random effects considered were restricted to be independent across individuals but correlated within an individual. However, in some biomedical and epidemiological studies, group level heterogeneity might be present. For instance, health outcomes sometimes exhibit geographical patterns. Factors such as environmental exposures, socioeconomic disparities, cultural behaviors or different access to healthcare can create spatial structures where individuals in nearby areas are more similar to each other than to individuals in distant areas. This is known as spatial autocorrelation.