Background: Mortality information is essential for estimating relative cancer survival (that excludes deaths from other causes). However, sufficiently detailed mortality data are not available for all areas covered by the cancer registries (CRs) participating in the EUROCARE-3 study.
Materials and methods: Mathematical methods were used to construct complete local mortality data (life tables) for each year of age (0-99), for each year (1978-2000) and by sex, from the incomplete life tables provided by CRs, presenting the results as life expectancy at birth (LE). Socio-economic data were obtained from the United Nations (UN) and Organisation for Economic Co-operation and Development (OECD).
Results: The time and regional trends in LE provided by our estimates are closely similar to those published by the UN at the country level. According to UN data, LE (men plus women) varied from 70 years in Estonia to almost 79 years in Sweden in the period 1995-1999. LE increased markedly over the 20-year study period in most countries except Estonia and Denmark. LE correlated directly with GDP, national expenditure on health and relative survival for all cancers combined. We found that within-country LE variation was large in some countries (particularly the UK). Sweden, Iceland, Switzerland, and parts of Spain and Italy had high LE; eastern European countries had low LE.
Conclusions: Detailed area-specific life tables are essential for reliable estimation of relative cancer survival and its comparison across populations, since LE varies markedly across Europe. Where not available, life tables can be constructed to the required level of detail using mathematical approaches.