Key points are not available for this paper at this time.
This study systematically evaluated whether period analysis outperforms traditional cohort and complete analyses in estimating 5-year relative survival (RS) for childhood cancer patients using data from nine cancer registries in Taizhou, Eastern China (2009-2018). Analyses included patients under 15 years diagnosed with childhood cancers, with 5-year RS estimates from cohort, complete, and period methods compared against observed actual survival (2014-2018: 70.1%). Accuracy and robustness were assessed via deviation value (DV) and standard error (SE). Cohort analysis yielded a 5-year RS of 55.8% (DV: -14.3%, SE: 2.4), complete analysis 61.1% (DV: -9.0%, SE: 1.7), and period analysis 68.2% (DV: -1.9%, SE: 2.1). Stratified evaluations by sex, region, age at diagnosis, and cancer types confirmed period analysis as the most accurate (lowest DV across subgroups) while complete analysis showed the smallest SE values, indicating superior robustness. Cohort analysis performed the worst in both accuracy and robustness. This study is the first in China to validate that period analysis offers superior accuracy in estimating 5-year RS for childhood cancer patients overall and across various stratifications compared to cohort and complete analyses. Findings underscore the utility of period analysis in generating timely survival estimates to inform early detection strategies and cancer control policies, offering critical methodological support for advancing pediatric oncology outcomes assessment in China.
Bing et al. (Wed,) studied this question.