| 초록 |
Purpose Medical laboratories carrying out all types of measurement are increasingly being asked to evaluate the uncertainties associated with their measurement results. For medical laboratories, measurement uncertainty is now an accreditation requirement as specified by standard such as ISO/IEC 17025 and ISO 15189. In order to satisfy the level of 'accredited medical laboratory', we need to evaluate the measurement uncertainties and have to provide them to certification authority such as KOLAS. For this reason, we carried out processes of measurement uncertainties by using the real test examples. Materials and Methods We searched and arranged the terms associated with ISO 15189. The CA19-9 test results of three (A, B, C) hospitals accredited by KOLAS, was the object of the measurement uncertainty estimation. We evaluated type A and, B both uncertainty estimation methods. Type A evaluation is using the repeated measurement values, (bottom-up approach). Type B evaluation is not using statistical analysis, but doing another method such as variation coefficient of internal quality control materials (Top-down approach). Results Hospital A of the expanded uncertainty on CA19-9 control 1 was $27.0{ pm}5.11U/mL$ , control 2 was $181{ pm}30.97U/mL$ , that of hospital C was $41.22{ pm}7.56U/mL$ (type A evaluation). By using the variation coefficient of intra quality control materials we could estimate the standard uncertainty, combined standard uncertainty, expanded standard uncertainty(k=2). The expanded uncertainty of A hospital was 10.7~15.8%, B hospital was 8.0~22.7%(type A evaluation). Conclusion Medical laboratories should require the estimation results of measurement uncertainties in compliance with ISO 15189. Then, we studied the concept and estimation process with real examples in measurement uncertainty and made an appropriate form to report. Medical laboratories accredited by KOLAS could evaluate measurement uncertainty by using type A and B. We need to make an effort to standardize our laboratories, which could be a chance to reduce the rate of misdiagnosis and improve the reliability in medical laboratories. |
