Elsevier

Clinical Biochemistry

Volume 71, September 2019, Pages 14-16
Clinical Biochemistry

Choosing the best statistical method for reference interval estimation

https://doi.org/10.1016/j.clinbiochem.2019.06.006Get rights and content

Highlights

  • Parametric method provides accurate reference interval estimation for Gaussian data.

  • Harrell-Davis method is a good alternative to the traditional non-parametric method.

  • RIs should be calculated by multiple methods to choose the most precise estimate.

Section snippets

Conflicts of interest

The authors have no conflicts of interest to declare.

Acknowledgments

This work was supported by a Foundation Grant from the Canadian Institutes of Health Research (CIHR) to KA. VH was supported by a CIHR Doctoral award.

First page preview

First page preview
Click to open first page preview

References (18)

There are more references available in the full text version of this article.

Cited by (11)

  • Effect of sample size and the traditional parametric, nonparametric, and robust methods on the establishment of reference intervals: Evidence from real world data

    2021, Clinical Biochemistry
    Citation Excerpt :

    However, the CIs of the RIs for the three methods are relatively wide when the sample size is 120. The CI width of the upper limit of the RI is particularly larger than 0.2 times of the RI width, indicating low precision when the RI of the TSH is established for 120 samples [6,7]. Unsurprisingly, the upper and lower limits of the RIs established using the three methods were more consistent when using a larger sample size (greater than or equal to 2000).

  • Reference interval of monocyte distribution width (MDW) in healthy blood donors

    2020, Clinica Chimica Acta
    Citation Excerpt :

    Specifically, the current CLSI document recommends the use of the non-parametric method, as the standard method, and the robust method, as an alternative in case of small samples. The Harrell-Davis method represents a bootstrap version of the traditional non-parametric method, but it is more efficient than the simple non-parametric method [25]. However, although the Harrell-Davis method has an excellent ability to establish RI, it involves more complex computation and, consequently, requires more expertise.

  • Real-world big-data studies in laboratory medicine: Current status, application, and future considerations

    2020, Clinical Biochemistry
    Citation Excerpt :

    This requires the participation of programmers, data analysts, and statisticians. Based on published literature [18,23] and the experience of our research team [24–26], we present the general study process for establishing RIs using real-world data from clinical laboratories in Fig. 2. The research process includes data collection, data cleaning, data description, statistical analysis, and the publication of results.

  • Harmonizing the post-analytical phase: focus on the laboratory report

    2024, Clinical Chemistry and Laboratory Medicine
View all citing articles on Scopus
1

These authors contributed equally to this work.

View full text