A multilevel multimember model for smoothing a disease map of lung cancer rates

Aim of this study is to assess the effect of smoothing a hospitalization rates map, based on the assumption that they may be influenced by the neighboring municipalities, the health service organization (HSO) and environmental risk factors. To smooth rates, two different Multilevel Multimembership Models were fitted: in the first the random effects were the municipality heterogeneity, the spatial dependence of the municipalities and the local HSO; in the second we replaced the local HSO effect by the environmental risk effect. The models were applied to show the spatial rates of hospitalization for lung cancer in Apulia in the year 2006. Maps shaded with the rates obtained at the end of the smoothing procedure seem to express a geographic distribution pattern of higher or lower rates in specific areas of the region. The effect of smoothing was greater in municipalities with a more unstable Risk Adjusted Rate.