Background: The Automated Childhood Cancer Information System (ACCIS) provides a unique database for monitoring and comparing long-term prognosis of children with cancer in Europe. The 'period' method has been shown particularly useful for that purpose.
Patients and methods: The ACCIS database was used for comparative analyses of up-to-date long-term survival in four regions and 19 countries of Europe by period analysis. Overall, follow-up data on 49,503 children diagnosed between 1985 and the end of the 1990s were included. Period estimates of 10-year survival were derived for the 1995-99 period and compared with estimates obtained by cohort and complete analysis. In this article, the methodology is described and the approach is illustrated for all forms of childhood cancer combined.
Results: Long-term childhood cancer survival achieved in Europe by the end of the twentieth century is much higher than previous estimates indicated. The overall period estimate of 10-year survival was 71% for all countries combined, but it varied between 45% for Estonian children and 79% for Swedish and Finnish children with cancer.
Conclusion: Period analysis is the method of choice to monitor population-based survival. Despite major improvement during the past decades, tremendous variation in childhood cancer survival between European countries has persisted.