Misclassification of breast cancer as cause of death in a service screening area

Abstract

Objective: The aim of this study was to assess the misclassification of cause of death for breast cancer cases, and to evaluate the differential misclassification between cases detected in an organized screening program and cases found in current clinical practice.

Methods: All deaths occurring between 1999 and 2002 within breast cancer cases were linked to hospital discharge records. Death certificates and latest available hospital discharge notes were classified into various categories. We created a classification algorithm defining which combinations of categories (of death certificates and hospital discharge notes) suggested the probability of misclassification and the need for an in-depth diagnostic review. Questionable cases were reviewed by a team of experts in order to reach a consensus on cause of death. Based on our algorithmic classification and diagnostic review results, the agreement between original cause of death and that resulting from the assessment process was analyzed stratifying for every variable of interest.

Results: According to death certificates, breast cancer was the cause of death in 66.9% of subjects, and after assessment this figure changed to 65.7%. The misclassification rate was 4.3% and did not differ significantly between screen-detected (4.7%) and non-screen-detected (4.3%) cases. Higher misclassification rates in favor of false positivity (cause of death wrongly attributed to breast cancer in death certificates) was observed for subjects with multiple cancers (6.5% vs. 1.9%), with no admission in the year before death (4.6% vs. 2.4%) and with an unknown cancer stage (4.9% vs 2.4% or 2.3%).

Conclusions: The cause of death misclassification rate is modest, causing a slight overestimate of deaths attributed to breast cancer, and is not affected by modality of diagnosis. The study confirmed the validity of using cause-specific mortality for service screening evaluation.