Bland-Altman Plot

In recent years, the type of difference plot for evaluation of method comparison data suggested by Bland and Altman has become increasingly popular. The procedure was originally introduced for comparison of measurements in clinical medicine, but the procedure has been adopted also in clinical chemistry. The Bland-Altman plot is usually understood as a plot of the differences against the average results of the methods. Thus the difference plot in this version provides information on the relation between dif- 

Nonparametric limits may also be considered. The distribution of the differences as measured on the y-axis of the coordinate system corresponds to tlie relations outlined for the DoD plot, which represents a projection of the differences on the y-axis. A constant mean bias over the analytical measurement range changes the average concentration away from zero. The presence of random matrix-related interferences increases the width of the distribution. If the mean bias depends on the concentration or if the dispersion varies with the concentration or both, the relations become more complex, and the interval mean 2 SD of the differences may not fit very well as a 95% interval throughout the analytical measurement range. 

Although it is customary to display the estimated limits for the differences (often mean 2 SDdif), one may as an alternative display specification limits considered being reasonable as mentioned for the DoD plot. It may then be assessed whether the observed differences conform to these limits as discussed earlier (see Table 14-10). Application of the difference plot in various specific contexts has been considered. It has also been suggested to estimate a regression line for the differences as a function of the average measurement concentration.  

A Caution Against Incoerect Interpretation of Paired t-TESTS in Method Comparison Studies 

---

The Bland-Altman plot is superior to regression/correlation analysis when used to compare two methods of measurement. It is the method of choice when comparing one method to an agreed gold standard. 

Figure 21.7 An example of a Bland-Altman plot The bias is -1.5 and the limits of agreement-0.5 to -2.8. There is poor agreement in this case.

Figure 14-16 Bland-Altman plot of differences for the drug comparison example. The differences are plotted against the average concentration.The mean difference (42nmol/L) v ith 2 SD of differences is shown (dashed //nesj,...

Figure 14-18 Simulated example with positive and negative differences in the low and high range, respectively A, Bland-Altman plot. B, An x-y plot with diagonal (dotted straight line) and estimated Deming regression line fso//d //ne) with 95% confidence curves (dashed lines).

Continuous Bland-Altman plot Discrete percentage agreement... 

This graph shows that the measured fluid balance is strongly correlated with the calculated body water (r = 0.9324). When taking this correlation into account, both measurement methods have been compared using a Bland-Altman plot (see fig. 8 (b)). It shows that the mean error is low and all measured values lie within the limits of agreement (1.96 a), where cr is the standard deviation of the mean value /t. 

Figure 8. (a) Measured fluid balance vs. calculated body water (b) Bland-Altman plot for fluid balance. 5 CONCLUSION... 

Fig. 2 Comparison of PEP values obtained with the impedance (PEPz) and catheter (PEPa.) methods, a - trends of calculated PEPz and PEPa. values the arrow heads 1, 2, and 3 indicate the start of Dobutamine infusion with the rate of 20, 15 and 10 pg/kg/min b - regression plot and c - Bland-Altman plot for these values.

Fig. 2 Agreement between the Cobas Integra 700 and Kone Microlyte 6 methods as an Altman-Bland (AB) plot. The 95% limits of agreement are illustrated as dashed lines...

Method comparison between SIDA, radiobinding assay (A) or microbiological assay (B) for serum folates. The results are presented as Bland-Altman difference plots solid line represents the mean difference between the two methods. The SIDA method is used as the comparison method. Dashed lines represent the limits of agreement or 0.95 central intervals of the differences between the two methods (mean difference = 2 SD). A total of 46 serum samples were used in the analysis in A and 49 for the analysis in B. 

Fig. 2 Bland and Altman plot. Segmental sum mean resistance (fi) between methods 3 and 4...

Figure 1 shows the Bland and Altman plots of total body water (TBW) by deuterium oxide dilution (D2O) against TBW by anthropometry (ANT) and TBW by bioeleetrieal impedanee analysis (BIA). 

Bland JM, Altman DG. Comparing methods of measurement why plotting difference against standard method is misleading. Lancet 1995 346 1085-7. 

Bland M. J. and D.G. Altman, 1995. Comparing methods of measurements Why plotting difference against a standard method is misleading. Lancet 346, 1085-1087. 

---