Login / Signup

A review of Bland-Altman difference plot analysis in the veterinary clinical pathology laboratory.

A Russell Moore
Published in: Veterinary clinical pathology (2023)
Current guidelines recommend using Bland-Altman plots (BA-plots), also called Difference plots, as part of method comparison evaluation in the veterinary clinical pathology laboratory. Analysis of differences can meaningfully augment linear regression techniques and allows fuller summarization of the performance of two methods relative to each other. This work summarizes the current literature on BA-plot composition and evaluation. Model data is used to demonstrate data evaluation approaches based on the observed differences, the combined inherent imprecision of the methods, and clinically relevant performance goals. Common limitations of the approaches, including points of frequent misinterpretation, are presented. BA-plot analysis can be part of an intentionally crafted method comparison study that provides analytically and clinically relevant data.
Keyphrases
  • electronic health record
  • big data
  • systematic review
  • machine learning
  • public health
  • clinical practice
  • artificial intelligence
  • global health