H. An , H. Han , K. Shin , K. Lim , I. Jang
Department of Clinical Pharmacology & Therapeutics, Seoul National University College of Medicine & Hospital, Seoul, Korea
Objectives: This study aimed to modify the Bland-Altman method which is used for evaluating agreement between two measurement techniques with repeated measures data through the random effects model.
Methods and Results: The Bland-Altman analysis using a graphical method to plot the difference of two measurement techniques against the mean for each subject is most frequently used to assess agreement between two techniques. However, with repeated measures data, the standard Bland-Altman method ignores the time profiles of repeated measurements by summarizing them as a mean on each subject [1]. Random effects model using SAS (v.9.2) which considers correlations between repeated measurements in each subject reduced information loss of the standard Bland-Altman method which sums up repeated measurements as a mean.
Conclusions: we analyzed agreement of two measurement techniques using the modified Bland-Altman method which utilized all the data and exposed the time profiles of differences between two measurement techniques. In addition, we outlined how our random effects model could account for the dependent nature of the repeated measures data, and additional explanatory variables, to provide reliable estimates of agreement in this setting.
References:
[1] Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res 1999; 8: 135-60
Reference: PAGE 21 () Abstr 2397 [www.page-meeting.org/?abstract=2397]
Poster: Other Modelling Applications