Abstract
BACKGROUND: Without standardization of medical laboratory's testing practices, there is an increase in false diagnoses when relying on test results. However, the effect of test standardization is difficult to assess numerically. This study's purpose is to quantify the effect of the standardization level of a laboratory on the prevalence of diabetes mellitus (DM) and impaired fasting glucose (IFG). METHODS: Laboratories were classified into three levels: 'highly-standardized laboratory,' 'basically-standardized laboratory,' and 'non-standardized laboratory.' Based on the results of Korean External Quality Assessment Scheme (KEQAS), the cutoff values for diagnosis of DM and IFG were recalculated, given false positive and false negative rates. RESULTS: The prevalence of DM and IFG in the population as a whole was estimated using the 2013 Korea National Health and Nutrition Examination Survey (KNHANES) database. When the prevalence of DM from KNHANES was 11.88% (95% confidence interval [CI], 10.59%-13.17%), the proportion with a systematic false error ranged from 10.91% (95% CI, 9.65%-12.17%) to 13.09% (95% CI, 11.74%-14.45%). The prevalence of IFG varied from 13.59% (95% CI, 12.25%-14.91%) to 40.49% (95% CI, 38.54%-42.43%), in contrast to 24.58% (95% CI, 22.85%-26.31%) of the reference value. The prevalence of DM and IFG tended to be over- and under-estimated more as the laboratory standardization level became lower, respectively. CONCLUSION: Our study proved that standardization of clinical laboratory tests is an important factor affecting the prevalence estimation of national disease statistics based on the simulation using KNHANES data.