The poorer cancer survival among the unmarried in Norway: Is much explained by comorbidities?

Studies from Norway and other countries have shown that the unmarried have poorer cancer survival than the married, given age, tumor site and stage at diagnosis. The objective of this investigation was to assess the importance of comorbidities for this difference, using disease indicators derived from the Norwegian Prescription Database (NorPD) and information on cancer and sociodemographic characteristics from various other registers, all of which cover the entire Norwegian population. Discrete-time hazard models for cancer mortality up to 2007 were estimated for all 22,925 men and 21,694 women diagnosed with 13 common types of cancer in 2005–7. There were 4898 cancer deaths among men and 4187 among women. Controlling for sociodemographic factors and tumor characteristics, the odds of dying from cancer among never-married men relative to the married was 1.56 (CI 1.41–1.74). The corresponding estimates for widowed and divorced were 1.16 (CI 1.05–1.28) and 1.27 (CI 1.15–1.40). For women, the odds ratios for these three groups were 1.47 (CI 1.29–1.67), 1.10 (CI 1.01–1.20) and 1.14 (CI 1.02–1.27). Several of the 24 indicators of diseases in the year before diagnosis were associated with cancer survival, but their inclusion reduced the excess mortality of the unmarried by only 1–5 percentage points, or about 10% as an overall relative figure. Similar results were found when the four most common cancers were analyzed separately, though there were some differences between them in the role played by the comorbidities. It is possible that important comorbidities are inadequately captured by the included indicators, and perhaps especially for the unmarried. Such concerns aside, the results suggest that the marital status differences in cancer survival to little extent are due to comorbidities (and the few disease risk factors that are also captured), but rather to various other “host factors” or to treatment or care.