Asian Journal of Medical Sciences (May 2017)
Estimation of indirect reference intervals for serum thyrotropin using hospital records
Abstract
Background: Reference intervals of any biochemical analyte serve as an invaluable tool in clinical decision making. The IFCC (International Federation of Clinical Chemistry) guidelines for determining these values are not feasible in some hospital laboratory setting and have led to the development of alternative approaches. Aims and Objectives: To determine the indirect reference intervals for serum thyrotropin form the hospital records of individuals visiting a tertiary care center. Materials and Methods: In a hospital-record based, observational, cross-sectional study, data of serum TSH levels were collected from the hospital records of participants who underwent this test in the Central Clinical Laboratory, College of Medical Sciences and Teaching Hospital, Bharatpur, Chitwan, Nepal from July 2012 to June 2015. All the individuals, irrespective of their diagnoses of thyroid diseases and other possible comorbid conditions, were included in the study. Prior to the statistical analyses, partitioning was done in relation to gender, age, and ethnicity. The reference intervals for thyrotropin were established by non-parametric method. Results: Reference intervals for serum TSH best agreeing to those provided by the test kit suppliers were determined by combining the two strategies that used Tukey’s method of detection and removal of outliers, prior to the final analyses. Lower limit was best determined from the natural-log-transformed and upper limit from non-transformed TSH values with outliers removed by Tukey’s method in both. As such, for the cases with TSH in the range 0.02-98.8 mIU/L, the reference intervals were calculated as [0.31 (0.30-0.33) to 6.04 (5.97-6.12) mIU/L] and for the TSH in the range 0.102-9.99 mIU/L, [0.35 (0.34-0.37) to 5.81 (5.75-5.90) mIU/L]. Conclusion: For establishing the indirect reference intervals from the hospital records, laboratory data can be combined with information stored in clinical databases for selecting subjects fulfilling stated clinical criteria.
Keywords
