This study was conducted to evaluate the reliability of a questionnaire from an epidemiological study for nuclear power plants workers in Korea.

Among a total of 8,832 nuclear power plants workers who participated in the questionnaire survey, we selected 646 workers who repeatedly submitted the self-reported questionnaire. The reliability of the questionnaire for the categorical items was measured using kappa statistics and percentages of exact agreement, false disagreement and positive disagreement. Significance was evaluated for the continuous items by paired t-tests and intra-class correlation coefficients (ICCs).

The questionnaire was observed to be highly reliable for the smoking history (kappa=0.85), which is known as an important confounding factor in assessing the cancer risk among radiation workers. Education level as a surrogate for the socio-economic status also appeared to be highly reliable with a kappa=0.89 and 95.09% of exact agreement. The history of alcohol drinking, medical exams and diseases showed more than moderately good agreement between the first and second responses (kappa>0.4). The self-reported values for physical measurements, the smoking period and smoking amount, and the period of alcohol drinking were highly correlated between the two responses (ICC>0.7).

The questionnaire from an epidemiological study for nuclear power plants workers in Korea was found to be reliable for most items. To estimate the cancer or noncancer risk for nuclear power plants workers, we need to use the national database for incidence of diseases or death from diseases and the exposure history, and the medical records as well as the data collected from this study.