Fig. 4

Correlations between small area incidence of ACSCs and their development over time

Pearson correlations between estimates obtained from Poisson multilevel models of three Nordic capital areas: 1) small area intercepts, i.e. incidence rate ratios (IRRs) of hospitalisations for ambulatory care sensitive conditions (ACSCs) in 2000 for each small area when compared to the average of capital area they belong to (i.e. x-axis value of 1) – and 2) slopes for over time for the development of incidence of ACSCs in biennial time-periods in 2000–2015 for each small area when compared to the average development in the capital area they belong to (i.e. y-axis value of 1). Each dot represents a single small area. Models 1–2 were analysed with populations aged 45 and over – and Models 3–4 with those aged 75 and over. Each model was adjusted for gender, capital area and time as well as for selected interactions. Models 1 and 2 were also adjusted for age group and Models 2 and 4 for individual incomes (with values not adjusted for incomes presented in grey – i.e. results of Models 1 and 3)