Related Experiment Videos
Net survival: comparison between relative and cause-specific survival estimates
E Crocetti1, V De Lisi, L Gafà
1Registro Tumori Toscano, Unità operativa di epidemiologia clinica e descrittiva, Centro per lo studio e la prevenzione oncologica, Firenze. e.crocetti@cspo.it
Epidemiologia E Prevenzione
|November 7, 2001
Summary
Cause-specific survival and relative survival show good agreement for many cancers, including stomach and breast cancer. Adjustments for specific cancer types, like colorectal and uterine cancers, are needed for accurate comparisons using mortality data.
Area of Science:
- Oncology
- Epidemiology
- Biostatistics
Background:
- Comparing cause-specific survival (CSS) and relative survival (RS) is crucial for cancer research.
- The Hakulinen method provides a standardized approach for calculating RS.
- Cancer registry data from Tuscany were utilized for this comparative analysis.
Purpose of the Study:
- To compare the estimates derived from cause-specific survival (CSS) and relative survival (RS) using the Hakulinen method.
- To assess the agreement between CSS and RS across various cancer sites.
- To identify necessary adjustments for accurate survival estimations.
Main Methods:
- Analysis of cancer registry data from Tuscany.
- Calculation of cause-specific survival (CSS) and relative survival (RS) via the Hakulinen method.
- Inclusion of specific death certificate codes to reconcile discrepancies between CSS and RS.
Main Results:
- Very good agreement between CSS and RS was observed for stomach, lung, and skin melanoma.
- Good agreement was found for female breast cancer.
- Adjustments for deaths from other large bowel cancers and unspecified intestinal sites were required for colon and rectal cancer analysis.
- Inclusion of deaths from unspecified uterine cancer was necessary for cervix and corpus uteri cancer analysis, with CSS remaining higher.
Conclusions:
- Cause-specific survival (CSS) can be utilized with overall mortality data, provided that deaths from cancers prone to misclassification are included.
- Careful consideration of mortality data reliability and cause-of-death coding criteria is essential for comparative studies using CSS.
- Adjustments enhance the comparability of CSS and RS, particularly for specific cancer types like colorectal and uterine cancers.