Skip to main content


Fig. 4 | BMC Medicine

Fig. 4

From: A meta-analysis of the performance of the PimaTM CD4 for point of care testing

Fig. 4

Agreement analysis for 11,803 data paired observations between Pima CD4 and reference CD4 technology testing. Plot (a) is a modified Bland-Altman scatter plot with vertical axis as the mean bias (Pima CD4 - reference) and the horizontal axis the absolute CD4 count of the reference technology. The dotted lines illustrate the typical funnel shape of difference not being relative over the range in absolute CD4 counts. Plot (b) is the percentage similarity scatter plot with vertical axis the mean percentage similarity values and horizontal axis the absolute CD4 count of the reference. The dotted circle highlights observations pairs that are not clinically relevant outliers in this CD4 count range, but generate high percentage similarity values due to the nature of the method comparison formula. The vertical axis of the percentage similarity plot represents values <1,000. The line plot (c) represents the SD of the bias on the vertical axis and the percentage relative bias SD on the secondary vertical axis and the median CD4 count in four CD4 count categories (0–100 cells/μl; 100–350 cells/μl; 350–500 cells/μl; >500 cells/μl) on the horizontal axis. The legend shows the overlay of all three method comparison methods (Bland-Altman, percent similarity and percent bias) for specimens sorted by the specimen extraction method (venous and capillary)

Back to article page