Validation of self-reported cancer diagnoses in the Korean Atomic Bomb Survivor Cohort study

Abstract

Background: Owing to the short history of the National Cancer Registry, self-reported cancer history is an important source of morbidity in recently constructed cohorts or cohorts of the older population. This study aimed to evaluate the validity of the self–reported cancer history gathered through the Korean Atomic Bomb Survivor Cohort (K-ABC) study compared with the Korean Central Cancer Registry (KCCR). Methods: Among the K-ABC participants recruited from 2020 to 2024, self-reported cancer history from 1756 participants was linked to the KCCR. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and kappa coefficient were estimated for each cancer site. Results: Among 181 participants with a minimum of one record in the KCCR, 164 reported a history of cancer. Overall cancers and each cancer site showed high sensitivity, PPV, and kappa values of > 80 %, > 66 %, and > 0.73, respectively, except for cervical and lung cancers, which showed sensitivities of < 50 % and kappa values of 0.33 and 0.59, respectively. High specificity and NPV (≥99 %) were observed for all cancer sites. Conclusions: The validity of the self-reported cancer history of the K-ABC showed a relatively high level of agreement with the KCCR records. These findings indicate that self-reported cancer information from the K-ABC study is valid for all sexes and ages but needs consideration when used for evaluating cervical and lung cancer.