Problems related to the coding of multiple primary cancers.
There is growing interest in multiple primary cancers (MPs), but the lack of a universally agreed definition can both modify the results for one series and hamper comparisons among cancer registries.
The aim of this study was to compare agreement on the coding of MPs between two cancer registries, the Danish Cancer Registry and the Tuscany Tumour Registry, that adhere to different rules for accepting MPs, and to study whether coding according to common international rules (IACR) would increase the comparability.
Data on 200 patients recorded as having more than one cancer were extracted at random from the two registers.
The agreement on MP status between coders, one from each registry, using local rules and definitions on MP, was good (kappa value, 0.70).
Exclusion of 11 expected discordant cases increased the agreement (kappa=0.86).
The agreement reached with the use of the IACR rules was very high (kappa=0.80).
We conclude that registries should present data according to international rules, in particular for the study of MPs.
Registries should at least clearly indicate deviations from the agreed international standards, in order to facilitate comparisons on incidences.
Mots-clés Pascal : Tumeur maligne, Multiple, Homme, Codage, Registre, Saisie donnée, Faux négatif, Méthodologie
Mots-clés Pascal anglais : Malignant tumor, Multiple, Human, Coding, Register, Data acquisition, False negative, Methodology
Notice produite par :
Inist-CNRS - Institut de l'Information Scientifique et Technique
Cote : 96-0357071
Code Inist : 002B04A. Création : 10/04/1997.