Different trends in colorectal cancer mortality between age groups in China: an age-period-cohort and joinpoint analysis.

Abstract Objectives China is undergoing a rapid aging transition. The trends in age-specific mortality rates from colorectal cancer remain unknown, and a number of studies have not distinguished the age, period, and cohort effects simultaneously. Study design A descriptive study was implemented with a joinpoint regression analysis and age–period–cohort (APC) model based on the intrinsic estimator (IE). Methods Age-specific mortality rates of colorectal cancer (1987–2016) were collected by gender (men/women) and region (urban/rural). The average annual percentage change (AAPC) and relative risks in the trend were identified using joinpoint Poisson regression and APC model (IE), respectively. Results Joinpoint regression analysis revealed that the rates decreased in the younger (men aged <45 years and women aged <75 years) but increased in the older (men aged >75 years and women aged >80 years) age groups. The APC model (IE) showed that the rates increased with age and time period but decreased with birth cohorts. But from 2000 to 2005, the period effects showed a substantial decline among urban residents. From the 1910–1914 to the 1915–1919 birth cohort, mortality increased among men, and from the 1925–1929 to the 1930–1934 birth cohort, mortality increased among rural residents. Conclusions The trends in colorectal cancer mortality are different between age groups. The younger age groups show a decreasing trend, whereas the older age groups an increasing trend. Cost-effective prevention and control should be implemented more in the elderly and for older cohorts at high risk. Highlights • Through the joinpoint Poisson regression, colorectal cancer mortality decreased in the younger (men aged <45 years and women aged <75 years) but increased in the older (men aged >75 years and women aged >80 years) over the last 30 years. • Through the age–period–cohort model (intrinsic estimator), the rates increased with age and time period but decreased with birth cohorts. • Analysis based on age-specific mortality rates reveals more information than that based on age-standardized rate. [ABSTRACT FROM AUTHOR]