[Predictive value and sensibility of hospital discharge system (PMSI) compared to cancer registries for thyroid cancer (1999-2000)].
Fiche publication
Date publication
septembre 2006
Auteurs
Membres identifiés du Cancéropôle Est :
Pr VELTEN Michel
Tous les auteurs :
Carre N, Uhry Z, Velten M, Tretarre B, Schvartz C, Molinie F, Maarouf N, Langlois C, Grosclaude P, Colonna M
Lien Pubmed
Résumé
BACKGROUND: Cancer registries have a complete recording of new cancer cases occurring among residents of a specific geographic area. In France, they cover only 13% of the population. For thyroid cancer, where incidence rate is highly variable according to the district conversely to mortality, national incidence estimates are not accurate. A nationwide database, such as hospital discharge system, could improve this estimate but its positive predictive value and sensibility should be evaluated. METHODS: The positive predictive value and the sensitivity for thyroid cancer case ascertainment (ICD-10) of the national hospital discharge system in 1999 and 2000 were estimated using the cancer registries database of 10 French districts as gold standard. The linkage of the two databases required transmission of nominative information from the health facilities of the study. From the registries database, a logistic regression analysis was carried out to identify factors related to being missed by the hospital discharge system. RESULTS: Among the 973 standardized discharge charts selected from the hospital discharge system, 866 were considered as true positive cases, and 107 as false positive. Forty five of the latter group were prevalent cases. The predictive positive value was 89% (95% confidence interval (CI): 87-91%) and did not differ according to the district (p=0,80). According to the cancer registries, 322 thyroid cancer cases diagnosed in 1999 or 2000 were missed by the hospital discharge system. Thus, the sensitivity of this latter system was 73% (70-76%) and varied significantly from 62% to 85% across districts (p
Référence
Rev Epidemiol Sante Publique. 2006 Sep;54(4):367-76.