biasplot: A package to effective plots to assess bias and precision in method comparison studies
Patrick Taffé
Institute for Social and Preventive Medicine
University of Lausanne
Lausanne, Switzerland
[email protected]
|
Mingkai Peng
Department of Community Health Sciences
University of Calgary
Calgary, Canada
[email protected]
|
Vicki Stagg
Calgary Statistical Support
Calgary, Canada
[email protected]
|
Tyler Williamson
Department of Community Health Sciences
University of Calgary
Calgary, Canada
[email protected]
|
Abstract. Bland and Altman’s (1986, Lancet 327: 307–310) limits of agreement
have been used in many clinical research settings to assess agreement between
two methods of measuring a quantitative characteristic. However, when the
variances of the measurement errors of the two methods differ, limits of
agreement can be misleading. biasplot implements a new statistical
methodology that Taffé (Forthcoming, Statistical Methods in Medical
Research) recently developed to circumvent this issue and assess bias and
precision of the two measurement methods (one is the reference standard, and
the other is the new measurement method to be evaluated). biasplot
produces three new plots introduced by Taffé: the “bias plot”,
“precision plot”, and “comparison plot”. These help the investigator visually
evaluate the performance of the new measurement method. In this article, we
introduce the user-written command biasplot and present worked examples
using simulated data included with the package. Note that the Taffé
method assumes there are several measurements from the reference standard and
possibly as few as one measurement from the new method for each individual.
View all articles by these authors:
Patrick Taffé, Mingkai Peng, Vicki Stagg, Tyler Williamson
View all articles with these keywords:
biasplot, limits of agreement, differential bias, proportional bias, Bland–Altman's plot, method comparison, measurement, empirical Bayes, BLUP
Download citation: BibTeX RIS
Download citation and abstract: BibTeX RIS
|