Who is responsible for reference ranges?

Abstract

Whyare reference ranges so important? All laboratory professionals are aware that they are only a guide to the interpretation of laboratory tests since they are so profoundly in£uenced by physiology, pathology and analytical methods. Yet despite this, reference ranges are given an unwarranted level of absolute truth by our clinical colleagues. Is it merely a matter of education for clinicians? Or are the ranges between laboratories sodiierent thatpatientcare isadverselyaiected? Education is certainly an important factor.We have all seen inappropriate patient management when a speci¢c test result has trivially crossed a demarcation line. Is patient care adversely aiected? This is di⁄cult to answer and anecdotes, however compelling the tale they tell, cannot be used as evidence. In the current climate of evidence-based medicine, reference ranges might be considered the laboratory expression of evidence. If this latter point is correct, who is responsible for ensuring that reference ranges are accurate: the immediate deliverer of test results (i.e. the laboratory) or the developer/manufacturer/supplier of the laboratory tests? Laboratories are all responsible for the tests they choose to provide and should therefore be accountable for providing appropriate guidance on the interpretation of these tests. Some of the guidance will be mandated by external bodies, for example glucose concentrations that de¢ne diabetes or cholesterol concentrations that are used to guide medication. Interpretation of other analytes is more problematic due to a lack of generally accepted evidence and because local variations in population mix aiect test results. This has unfortunately led to the use of published reference ranges which have inadequate reference to either the populations studied or the methods employed. The next level of responsibility lies in the protocols used to de¢ne reference ranges. The International Federation of Clinical Chemistry (IFCC) have issued a series of papers on the development of reference ranges. They describe the selection of subjects and how they and their samples should be prepared and collected.1 Moreover, they recommend that 120 subjects should be used for each range, but this may need to be supplemented to allow for physiological and ethnic factors. Studies of this magnitude are outside the range for a routine clinical laboratory. In the UK alone there are over 250NHS laboratories and for each to develop its own ranges would require enormous resources to be spent in the process. A collaborative approach using common protocols for selection of subjects and sample collection would allow laboratories to share this development. This has stimulated the Nordic Reference Interval Project (NORIP) in Scandinavia, where reference ranges are being developed which encompass both population mix and local analytical factors.2 Are manufacturers responsible for the provision of reference ranges? Kits are currently provided with a reference range, often with vague provenance, and which usually do not give information regarding their adherence to the criteria set by the IFCC. Furthermore, most kits carryan opt-out clause advising laboratories to de¢ne their own ranges. Given the current ¢nancial climate in laboratory medicine, it is clearly unreasonable to expect much more. However, a degree of collaboration is required unlike that illustrated in this issue of theAnnals, where a manufacturer altered an assay architecture without warning users.3 This change in the assay has the potential for serious implications on interpretation and therefore patient management and outcome. Should the regulatory authorities demand appropriately formulated reference ranges as part of the submission or reformulation of tests, as implied by EuropeanUnion InVitro Diagnostics Directive 98/794, extra resources will be required by the diagnostics manufacturers, whowill pass those costs on to already cash-strapped laboratories. Any project to develop reference ranges that will bridge analytical platforms and diierent populations will result in a widening of the ranges. More importantly, it will require the collection of a considerable data-set of pre-analytical factors. This will necessitate full ethical approval and consent. It is possible to use accumulated laboratory data,4 but this approach only oiers data on age and gender, which may be inadequate for adults. As clinical care moves into a more distributed model, patients will receive their care through networks of medical services. In order to ensure a transparent continuity of care, it will be essential to have rationalized reference ranges. A project to standardize reference ranges would be an ambitious exercise but one which will bring bene¢ts for those patients and doctorswhomove fromonehospital toanother, and for the translation of research in clinical diagnostics into practice.This is anexample of the current challenges to which our professional bodies should respond.