The appropriateness of Bland-Altman's approximate confidence intervals for limits of agreement

Gwowen Shieh*

*Corresponding author for this work

Research output: Contribution to journalArticle

6 Scopus citations

Abstract

Background: Percentiles are widely used as reference limits for determining the relative magnitude and substantial importance of quantitative measurements. An important application is the advocated Bland-Altman limits of agreement. Methods: To contribute to the data analysis and design planning of reference limit or percentile research, the purpose of this paper is twofold. The first is to clarify the statistical features of interval estimation procedures for normal percentiles. The second goal is to provide sample size procedures for precise interval estimation of normal percentiles. Results: The delineation demonstrates the theoretical connections between different pivotal quantities for obtaining exact confidence intervals. Moreover, the seemingly accurate approximate methods with equidistant from the principal estimators are shown to have undesirable confidence limits. It is found that the optimal sample size has a minimum for median or mean, and increases as the percentile approaches the extremes. Conclusions: The exact interval procedure should be used in preference to the approximate methods. Computer algorithms are presented to implement the suggested interval precision and sample size calculations for planning percentile research.

Original languageEnglish
Article number45
JournalBMC Medical Research Methodology
Volume18
Issue number1
DOIs
StatePublished - 22 May 2018

Keywords

  • Assurance probability
  • Expected width
  • Precision
  • Quantile
  • Sample size

Fingerprint Dive into the research topics of 'The appropriateness of Bland-Altman's approximate confidence intervals for limits of agreement'. Together they form a unique fingerprint.

  • Cite this