Assessing agreement between clinical measurement methods

Nursing practice and research often require the use of one measurement method in place of another, creating the need to know how well one clinical measurement method approximates another. The usual statistical analyses for assessing agreement between measurement methods are the correlation coefficient and the paired t test. This article suggests that these statistical methods do not address the question of agreement. It proposes the limits of agreement as a more informative analysis when a researcher is interested in replacing one measurement method with another. The application of limits of agreement analysis to assess repeatability of one measurement method is also introduced.