The Importance of Municipality Characteristics for Cancer Survival in Norway: A Multilevel Analysis

Discrete-time hazard models for cancer mortality in cancer patients were estimated from register and census data to find out whether various socio-economic, ideational and institutional community factors had an impact on cancer survival in Norway in the 1990s, also beyond that of the corresponding individual-level variables. Such a multilevel approach has not been employed in previous analyses of cancer survival. In addition to confirming the better prognosis for patients with high education, it was found that, among patients at the same educational level, mortality was lowest for those who lived in a municipality where the average education was relatively high. The impact of economic resources was less pronounced. While a low unemployment rate in the municipality and high individual income reduced mortality among cancer patients, a high average income had no effect. Also those who lived in municipalities where a large proportion voted with the Christian Democratic Party had an advantage, which suggests a beneficial impact of affiliation with religious communities or support for the central Christian ideas. Moreover, there was an excess mortality among patients who lived in municipalities served by a relatively small hospital that did not have any responsibility beyond the county level. These patients may have got somewhat inadequate treatment at a low level in the hospital structure, or they have perhaps not wanted, or been able to fully comply with, the recommended follow-up treatment at the highest level. Even with such factors included in the model, there was significant regional variation. Cancer survival was relatively poor, net of differences in the stage distribution, in the capital, the central parts of Southern and Western Norway, and the peripheral parts of Southern Norway.