This analysis is very interesting in seeking to establish an overall R value in analysing the relationship between two different settings. However the analysis seems to inherently assume that the two settings of 'community' and 'hospital + care home' are the same size - is that right ? If so have you analysed unequal size of settings ? Another key issue is the analysis of multiple settings - e.g. community, hospital and care home. Have you been able to address this? Such an analysis would obviously important in the considering interaction between different geographic areas.