The implementation of common reference intervals

Abstract

Introduction Reference intervals set by individual laboratories are commonly highly variable without good reasons based on population, pre-analytical or analytical differences. This variability is a major contributor to differences in interpretation when patients attend different laboratories. A solution is to adopt common reference intervals where this is appropriate. The key factors to achieve this are agreed intervals to adopt and the processes for local acceptance. Methods Mechanisms available for laboratories to validate external reference intervals for use include the following: analysis of a number of samples from normal subjects; performance of a local reference interval study; confirmation of sufficiently small analytical bias from the method used to set the proposed interval; data-mining from the pathology database. Proposal For laboratories to validate common reference intervals for use in their laboratories a ‘toolkit’ of suitable methods should be established. This may include a number of the available techniques listed above depending on available resources and the analyte being validated. For example data mining for a hospital laboratory may be less valid than for a laboratory with outpatient work, and validation of skewed or age-related intervals with small numbers of normal subjects may be under-powered. Conclusion The final responsibility for adoption of a common reference interval remains with the individual laboratory. Simple tools are required to ensure population, pre-analytical or analytical differences do not lead to erroneous use of common intervals and to facilitate rapid adoption when suitable.