The uncertainty of a numerical laboratory result can be masked by the fact that the laboratory reports an absolute number, whereas users have limited knowledge of the confidence interval of the result. Interpretation of laboratory tests is in reality therefore an inexact science, a balance between clinical context and the likely relevance of a laboratory result.
This review considers the factors which contribute to result variability and examines the implications for interpreting differences between sequential laboratory results. It offers suggestions to deal with a problem which has not yet been much addressed in routine practice. The examples used are restricted to the discipline of clinical biochemistry, although the issues and principles apply to numerical (and indeed qualitative) results in other disciplines.
Laboratories could provide more guidance on the likelihood of a result being significant to assist users. There is a need for discussion about how this is best done, and compatible with electronic result delivery. Options for providing this information are considered.
Statistics from Altmetric.com
Competing interests: None declared.