Methods : Capture-recapture analysis was used to obtain a nearly unbiased estimator (NUE) of the total number of WRA cases reported to KOWAS from 2004 to 2006. To do this, the 4 reporting sources were stratified into 2 categories as follows: (1) the workers’ compensation scheme (i.e. the Korea Workers’ Compensation and Welfare Service), and (2) the other 3 reporting sources (i.e. physicians’ reports). Capture-recapture analysis was performed on sex, regions and specific industries when the number of overlapping reports was ≥7. Overlapping reports, i.e. duplicate cases, were defined as individuals who had identical names, dates of birth, resident registration numbers, and employers found in both reporting categories.
Results: From 2004 to 2006, a total of 172 cases of WRA were reported by 4 reporting sources : 62 (48.8%) by workers’ compensation schemes, 31 (24.4%) by occupational physicians, 68 (53.5%) by allergy and chest physicians, and 50 (39.4%) by regional surveillance systems. The number of cases estimated by capture-recapture method was 186 (95% CI: 150.75-221.43). The completeness of worker’s compensation and physicians’ reports were 0.52 and 0.33 respectively. By industry, the reported cases and estimated cases of the manufacturing sector were 189 and 333 (95% CI : 262.62-404.27). By region, the reported cases and estimated cases in Gyeonggi and Incheon were 75 and 114 (95% CI : 83.75-143.69).
Conclusions : Capture-recapture models allow the estimation of the true number of cases and the assessment of the completeness of our surveillance data. And this method can be useful in a various public health areas.