A systematic review of statistical methods used in constructing pediatric reference intervals

Objectives: This study aims to investigate current medical literature with focus on statistical methods used to construct pediatric reference intervals and identify potential gaps within the process of reference interval estimation.

Design and methods: A systematic review of methods was performed. Extensive search criteria were developed and search was conducted on Embase, Medline, and PubMed databases to identify relevant articles. The articles were further screened using predefined inclusion and exclusion criteria. The selected articles were then included in our final systematic review.

Results: Our review reveals that there are gaps within current methodology and reporting of pediatric reference intervals. Not all publications followed the Clinical and Laboratory Standards Institute (CLSI) guidelines, and there is a large variation in the methods used. Discrepancies particularly arose when reference intervals were calculated for partitions with small sample sizes. In addition, the dynamic nature of pediatric data was not mostly captured when certain partitioning techniques were used.

Conclusions: There are areas within the pediatric reference interval development process that need attention. Partitioning methods particularly need to be explored with the goals of reducing subjectivity and enabling researchers to capture the best representative partitions possible. Moreover, the complicated nature of pediatric data often limits the sample size available for each partition and appropriate methods need to be considered in such cases. Researchers are also strongly encouraged to accompany their reference limits with confidence intervals to show sampling variability and demonstrate precision of their limits. These issues exemplify the need for a pediatric specific guideline that outlines a standardized way of establishing reference intervals.