Standardization of Reference Values among Laboratories of Korean Association of Health Promotion

Lee, Kap No; Yoon, Jong Hyun; Choi, Young Hee; Cho, Han Ik; Bae, Ki Woong; Yoon, Cheong Ha; Kim, Sang In

Abstract

BACKGROUND: All fifteen laboratories of Korean Association of Health Promotion were used to use different reference ranges that were mostly adapted from the published data or the recommended data by the reagent companies supplied as inserts. These created problems such as inadequate interpretation, inability of exchangeable patient data, and subsequent confusion, and all the data stored became useless particularly in the respect of health statistics. The purpose of this study was to solve such problems through defining commonly acceptable reference ranges.
METHODS
All the results performed during October 1, 2000 through September 31, 2001 were collected in preferably 'Excel' file format. All the data included the necessary information such as age and sex. The age was grouped in six; baby (0-3y), children (4-12y), adolescent (13-18y), adult (19-64y), younger elderly (65-79y), old elderly (over 80y), with references of statistics in medical informatics and WHO classification. The data were statistically analyzed with SAS 6.04 for Gaussian distribution. None of the tests showed Gaussian distribution. These procedures had been repeated twice or three times after trimming out the results lying outside three standard deviations. Though, all the tests showed non-Gaussian distribution. Subsequently, the reference ranges were defined in the range from the point of lower 2.5% to the point of higher 97.5%. And in case the lower range could be "0", the reference ranges were defined in the range of 0 to 95%.
RESULTS
The reference ranges of 56 test items were newly assigned. Among these there were eight tests that needed reference ranges by the age groups and nine tests by the sex.
CONCLUSIONS
All 15 laboratories of Korean Association of Health Promotion now have the same reference ranges that were statistically derived from their own data. This means the patient data and reference values can be exchangeable among their laboratories.