An appraisal of statistical procedures used in derivation of reference intervals

Abstract

When conducting studies to derive reference intervals (RIs), various statistical procedures are commonly applied at each step, from the planning stages to final computation of RIs. Determination of the necessary sample size is an important consideration, and evaluation of at least 400 individuals in each subgroup has been recommended to establish reliable common RIs in multicenter studies. Multiple regression analysis allows identification of the most important factors contributing to variation in test results, while accounting for possible confounding relationships among these factors. Of the various approaches proposed for judging the necessity of partitioning reference values, nested analysis of variance (ANOVA) is the likely method of choice owing to its ability to handle multiple groups and being able to adjust for multiple factors. Box-Cox power transformation often has been used to transform data to a Gaussian distribution for parametric computation of RIs. However, this transformation occasionally fails. Therefore, the non-parametric method based on determination of the 2.5 and 97.5 percentiles following sorting of the data, has been recommended for general use. The performance of the Box-Cox transformation can be improved by introducing an additional parameter representing the origin of transformation. In simulations, the confidence intervals (CIs) of reference limits (RLs) calculated by the parametric method were narrower than those calculated by the non-parametric approach. However, the margin of difference was rather small owing to additional variability in parametrically-determined RLs introduced by estimation of parameters for the Box-Cox transformation. The parametric calculation method may have an advantage over the non-parametric method in allowing identification and exclusion of extreme values during RI computation.

Protocol and standard operating procedures for common use in a worldwide multicenter study on reference values

“…Multiple regression analysis (MRA) should be performed analyte by analyte to identify factors related to variation of the test results [ 22 ]. The factors to be included in MRA are gender, age, BMI, smoking status, level of alcohol consumption, frequency and strength of physical exercise, dietary status, and, if available, ethnic group, the ABO blood group.…”

“…The presence of significant regionality in the test results can be determined by taking a ratio (SDR) of SDcntr over SDindiv; an SDR > 0.3 signifies the evidence of significant regionality in the test results [ 22 ]. However, its validity must be confirmed after adjusting for other possible confounding factors using MRA or similar procedures.…”

“…The parametric method will be used after normalizing data based on a modified Box-Cox power transformation method [ 22 ] for derivation of RIs. Secondary exclusion of inappropriate subjects should be made to remove those with abnormal results attributable to highly prevalent conditions such as the metabolic syndrome, alcohol-related hepatic dysfunction, or diabetes mellitus.…”

Reference Intervals of Reticulated Platelets and Other Platelet Parameters and Their Associations

“…12, 16 One of the possibilities to establish reference ranges is using patient data and an indirect statistical method, the Bhattacharya 11 technique. This is a robust and accepted tool for reference interval determination for analytes with an underlying Gaussian (or log-Gaussian) distribution.…”

Aproximación al patrón de normalidad de TSH para la población chilena según Encuesta Nacional de Salud 2009-2010

“…El rango de referencia normal de TSH no escapa a lo anterior y es crítico a la hora de determinar el estado de salud tiroideo de los individuos 1 .…”