Validity of commonly used indexes to estimate population attributable risk of lung cancer

Introduction
Because there are many obstacles to estimate magnitude of influence to diseases by occupational factors directly, population attributable risk (PAR) has been used as a surrogate using exposure population and relative risk indirectly. This indirect method would be hardly said to be correct, because an estimate of the population and relative risk in each investigation will vary according to circumstances.

Methods
We directly investigated exposure history of lung cancer patients in Busan metropolitan city in Korea, and calculated PAR by direct method. In the four acute hospitals in Busan 2,193 newly diagnosed patients with lung cancer between 2005 and 2009 were recruited. Five category of work relatedness (definite, probable, possible, suspicious, undetermined) were allocated to each case. Based on this work relatedness evaluation direct PAR was calculated. Indirect PARs were estimated by two methods; using cancer exposure index (CAREX index), turnover factor (TO) and 2005 census data in Busan, and using World health organization (WHO) guidelines.

Results
We found proportions of definite case, probable case were 1.7% and 5.4%, respectively. We estimate PAR of occupation factors in incidence of lung cancer by direct calculation was 7.1%. The range of PAR of each agent was -3.11% to 7.52%. Considering multiple exposure, we calculate total PAR of occupation factors was 9%.
For PAR estimated by WHO guideline, PAR for asbestos was 2%. Overall attributable fraction (AF) of lung cancer due to occupational exposure by WHO method will be shown and be compared with it by direct method.

Discussion
Through this study, we estimate PAR by each method and compare them. In the future, a broader range and continuous research is necessary to assess the actual impact of occupational exposure to incidence of cancers.