Trial Information

Importance of Cohort and Histologic Type for Secular Trends in Rising Incidence and Different Survival of Uterine Cancer -a Taiwan Population-based Registry From 1979 to 2008

Study population To evaluate the epidemiology and prognosis of uterine cancer in Taiwan, we
will conduct a nationwide analysis through linking national cancer registry database. Every
citizen in Taiwan has a life-long identification number to link individual information,
including health status. The Taiwan household registry database provided by Department of
Health will be the source population. National household registry and death certificate will
be adopted to ascertain the live status of study subjects, and provided individual
demographic characteristics. All individual linkages between databases will be conducted
according to the study protocols, i.e. databases will be linked by corresponding
identification number, name (Chinese characters) and birthday, and all data included in this
study will be analyzed without individual identification information. The agreement of
utilizing the databases in the study was obtained from the Bureau of Health Promotion in
Taiwan.

Incident and death cases ascertainment Patients with uterine cancer, including uterine
carcinoma, uterine sarcoma … etc., will be identified from computerized linkage to the
Taiwan national cancer registry with International Classification of Diseases for Oncology
Third Edition T-code C540-C543. Histological types will be identified from morphology code
in the cancer registry database. Women affected by uterine cancer and found death due to
uterine cancer (International Classification of Diseases 9th edition code 182) in deaths
certificate will be defined as death cases.

Statistical Analyses The numbers of person-years at risk of developing and dying for uterine
cancer will be calculated. Incidence rates will be calculated by dividing the number of
uterine cancer cases by the person-years at risk of developing uterine cancer. Mortality
rates will be calculated by dividing the number of uterine cancer deaths by the person-years
at risk of dying from uterine cancer. The association between mortality and age, histology,
time periods will be estimated through Kaplan-Meier survival analysis and Cox's proportional
hazards model. It is defined as significant difference statistically when p value is less
than 0.05.