What’s in a number: Falling on the sword of cut-off points and reference limits?

Abstract

The marked increase in laboratory test volumes and costs internationally emphasises the need for demand management. One way that this can be implemented is by reducing unnecessary repeat testing with the provision of appropriate decision cut-off points (clinical decision limits [CDLs]) or reference intervals (RIs) with subsequent correct interpretation of laboratory results. The derivation of RIs and CDLs are fraught with technical and biological challenges. There is difficulty in conducting labour-intensive, costly, long, and complex studies, which require healthy volunteers that represent things such as different age groups, genders, races, and alternate states of health (e.g. pregnancy) within the population. It is also inappropriate to apply RIs or cut-off points from other populations, which is often what occurs when manufacturer-expected values are used. Lack of standardisation of international guidelines for CDLs and analytical methods poses a further problem. The effect of analytical and biological variation on results is also essential to consider when interpreting results. These make ideal RIs and CDLs difficult to attain and implement despite their critical need.