A metric for comparison and visualization of age disparities in cancer survival

Abstract

AIMS: Diagnostic age is an important determinant of cancer survival but the methods generally used to analyze age-group-specific survival are not developed for ready visualization of survival differences. We aim at developing a novel metric for comparing and visualizing age-group-specific survival data over different cancers, sexes, periods and countries. METHODS: The metric describes the mean absolute deviation between age-groups. The metric can be used in two variations, one showing the mean variation and its 95% confidence intervals and the other highlighting individually each age-groups distinguishing positive or negative deviations. We demonstrate the applications with age-group- specific 5-year relative survival data from the NORDCAN database RESULTS: The mean absolute deviation between age-groups for Swedish colon cancer survival declined from about 5% in 1972-1981-1% in 1992-2001 and to 1.3% in 2012-2021. Patients diagnosed before age 50 years accounted for the largest positive deviation. For acute myeloid leukemia (AML) the mean deviation increased from 4% (female) to 17% and 23%. Patients diagnosed at age below 50 years showed the largest deviations. Comparing colon cancer mean deviations between the Nordic countries, a time-related decline was observed for all, those in Sweden ending at the lowest and in Finland the highest level. CONCLUSIONS: We demonstrated the usefulness of the devised metric for summarizing age-specific survival data between cancers, sexes, periods and countries. The two variations of the metric allow a simple visual presentation of the survival experience as to deviation of the survival data, its 95%CIs and its highlighted individual age-group components.