Determination of reference ranges using medical laboratory date

Document Type : Original Article


Assistant professor


In order to have reliable reference ranges and to minimize the effects of variables such as the procedure of specimen collection, geographical location, dietary habits, life style and possible errors during the performance of tests, it  is recommended that each medical laboratory establish it's own reference values. Idealy, the reference ranges Should be determined by analysis of specimens from every or, from a random sample of the individuals in the population of interest. However, in terms of time and money, it is very expensive and impracticable to examine and apply the selection criteria to the entire population and therefore a random selection of a subset of individuals is recommended. As an alternative, a number of methods have been suggested where by the test results of individuals undergoing routine analysis can be used to establish reference values. These people are a pt to forma heterogeneous sample, I.e. a mixture of two distributions, one, of healthy and one, of sick individuals. For this reason and the fact that in our country almost all clinical laboratories are using reference ranges belonging to foreign sources, a special computer program was devised to analyse the data collected at Kermanshah central laboratory. A set of reference values for routine and special biochemical tests have been derived. This includes; Glucose 55-119 mg/ dl, BUN 6-20 mg/ dl, cholesterol 138-278 mg/ dl, triglyceride 57-156 mg/ dl, creatinine 0.45-1.2 mg/ dl, uric acid 2.8-7.1 mg/ dl, total bilirubin 0.39-1.2 mg/ dl, direct bilirubin 0.18-0.23 mg/ dl, calcium 8.5-10.4 mg/ dl, phosphorus 2.5-5.6 mg/ dl, sodium 138-143 mmol/ L, potassium 3.8-4.8 mmol/L, iron 68-165 ug/ dl, Hb 10.2-16.8 G/ dl, total protein 6.6-7.1 G/ dl, SGOT 9-27U/ L, SGPT 8.26 U/ L, T 3.93-12.41 ug/ dl, T3 – Uptake 23-32℅, FTI 1.3-3.8℅mg/ dl, TSH 0.36-3.13 uIU/ ml. This results were compared with the reference ranges established by kitmanufacturers. The results of glucose, cholesterol and HB estimation have also been compared with the values calculated for a random samples selected by the health survey of Iran in Kermanshah province.