Method of estimation of radiation risks of mortality among Chernobyl clean-up workers with consideration to probable errors in cause-of-death reporting on death certificate

In the past decade the lawful use of radiation risks of non-cancer diseases estimates to solve radiation safety problems has become more significant. The highest statistical power of low dose (less than 0.2 Gy) radiation risks research is possible in the cohort of the Chernobyl clean-up workers (liquidators) registered and been under medical monitoring in the system of the National Radiation Epidemiological Register (NRER) since 1986. The liquidators total number exceeds 130 thousand people. Because of the large-scale annual collection of health data, carried out in the NRER system, verification of the diseases diagnoses and medically certified causes of death, a natural question of radiation risk estimates compatibility with consideration to probable errors in a disease diagnosis and cause-of-death reporting on death certificates arises. The work is aimed at estimating the radiation risk of mortality from diseases of the circulatory system and diseases of the digestive organs in the Russian cohort of liquidators, with account of possible errors in the disease diagnosis and certified causes of death. The size of the cohort considered for mortality analysis was 91,013 people, the average age at the Chernobyl zone entry was 33.6 years, average radiation dose was 0.133 Gy. In this study, the method of maximizing the partial likelihood function was used to assess radiation risks, with consideration of individual radiation doses to liquidators and the assessment of background mortality using the Breslow method. The specificity of the diagnosis of the primary causes of death in the NRER system significantly exceeds the level of 0.99. Considering the specificity index of 0.99, the estimates of radiation risk coefficients ERR/Gy for mortality from diseases of the circulatory system and diseases of the digestive system in the cohort of Russian liquidators do not differ from the estimates obtained in the assumption that there are no first kind errors (false positive diagnosis) in the diagnoses of the primary causes of death in the NRER database. The results obtained confirm the high stability and validity of the assessments of radiation risks of mortality from diseases of the circulatory system and diseases of the digestive system, obtained earlier from the data on liquidators registered in the NRER.