Estimation of Individual Based Exposure Assessment for Indoor Radon Using the Propensity Score Method

摘要

Epidemiologic studies are the best way to examine the risk and association between residential radon exposure and lung cancer. Previous studies suggest that exposure to residential radon may contribute to the incidence of lung cancer in the general population and varying residential conditions could also affect the relationship of radon exposure and lung cancer. Many studies considering association were conducted worldwide, however few studies have shown the causal inference. To investigate the causal association between residential radon and lung cancer based on Korea living conditions, we conducted a study using propensity score method in Korea. This study includes data from 4 hospitals (Ajou university, Severance, Wonju severance, Seoul-asan) for the treatment group (Lung-cancer patients) and from 4 community-based cohorts (Seoul CMERC, Wonju KoGES, Ansung KoGES, Kyeongnam Namgaram) for the control group (Not lung-cancer patients). Total 440 (650 treatments and 5624 controls) were included in the data. Considering age, sex, smoking status, and socio-economic factors the propensity score was analyzed. The propensity score method we used was propensity score matching. The characteristic differences for covariates before and after propensity score matching is shown. Also, the distribution difference between raw data and matched data is represented in Figure 1. The propensity score distribution in the control group was matched to the treated group. The odds ratio for using propensity score matching and without propensity score were demonstrated in Figure 2. This study examined the causal association between indoor radon level and lung cancer in Korea. However, due to lack of data further study is needed after data completion.