Data Mining in Establishing the Indirect Reference Intervals of Biochemical and Haematological Assays in the Paediatric Population: A Review

Reference intervals (RIs) are fundamental values accompanying medical laboratory results that allow interpretation by medical practitioners, thus influencing patient management. Traditionally, RIs are established by recruiting 120 healthy reference individuals and applying statistical analysis to the results. This method is challenging due to the technical and ethical issues involved. Therefore, many laboratories either adapt RIs provided by the manufacturers of their analytical platforms or the results of RI studies done in other countries. The advent of data mining technology has allowed an alternative method, the indirect RIs (IRIs) approach, which applies appropriate statistical techniques to patient data stored in the laboratory electronic medical records to establish the IRIs. This review briefly highlights the historical aspect of IRI determination, provides a general outline of the steps involved and reviews publications that have used data mining to establish the paediatric IRI over the past ten years.