I have created B&A plots to test the reproducibility and validity of a novel FFQ. Attached is one of the plots that i have created, I am finding it difficult to interpret them and to estimate the percentage of people that fell out of the limits of agreements. would appreciate your insight.
thank you
Bland and Altman introduced the Bland-Altman (B&A) plot to describe agreement between two quantitative measurements by constructing limits of agreement. These statistical limits are calculated by using the mean and the standard deviation (s) of the differences between two measurements. To check the assumptions of normality of differences and other characteristics, they used a graphical approach.
The resulting graph is a scatter plot XY, in which the Y axis shows the difference between the two paired measurements (A-B) and the X axis represents the average of these measures ((A+B)/2). In other words, the difference of the two paired measurements is plotted against the mean of the two measurements. B&A recommended that 95% of the data points should lie within ± 2s of the mean difference. This is the most common way to plot the B&A method, but it is also possible to plot the differences as percentages or ratios, and one can use the first method or the second one, instead of the mean of both methods.
The bias is computed as the value determined by one method minus the value determined by the other method. If one method is sometimes higher, and sometimes the other method is higher, the average of the differences will be close to zero. If it is not close to zero, this indicates that the two assay methods are systematically producing different results.
Interpreting the Bland-Altman results
Bland-Altman plots are generally interpreted informally, without further analyses. Ask yourself these questions:
(1) How big is the average discrepancy between methods (the bias)? You must interpret this clinically. Is the discrepancy large enough to be important? This is a clinical question, not a statistical one.
(2) How wide are the limits of agreement? If it is wide (as defined clinically), the results are ambiguous. If the limits are narrow (and the bias is tiny), then the two methods are essentially equivalent.
(3) Is there a trend? Does the difference between methods tend to get larger (or smaller) as the average increases?
(4) Is the variability consistent across the graph? Does the scatter around the bias line get larger as the average gets higher?
If there is a proportional bias, then the difference in energy (y-axis) should increase with the mean energy intake (x-axis).
As seen from the plot, the energy difference does not appear to increase with the energy intake. But you may conduct a linear regression analysis to test the null hypothesis.
You may refer to the video below for the details:
https://www.youtube.com/watch?v=vo7oQ_8OOhk
Bland and Altman introduced the Bland-Altman (B&A) plot to describe agreement between two quantitative measurements by constructing limits of agreement. These statistical limits are calculated by using the mean and the standard deviation (s) of the differences between two measurements. To check the assumptions of normality of differences and other characteristics, they used a graphical approach.
The resulting graph is a scatter plot XY, in which the Y axis shows the difference between the two paired measurements (A-B) and the X axis represents the average of these measures ((A+B)/2). In other words, the difference of the two paired measurements is plotted against the mean of the two measurements. B&A recommended that 95% of the data points should lie within ± 2s of the mean difference. This is the most common way to plot the B&A method, but it is also possible to plot the differences as percentages or ratios, and one can use the first method or the second one, instead of the mean of both methods.
The bias is computed as the value determined by one method minus the value determined by the other method. If one method is sometimes higher, and sometimes the other method is higher, the average of the differences will be close to zero. If it is not close to zero, this indicates that the two assay methods are systematically producing different results.
Interpreting the Bland-Altman results
Bland-Altman plots are generally interpreted informally, without further analyses. Ask yourself these questions:
(1) How big is the average discrepancy between methods (the bias)? You must interpret this clinically. Is the discrepancy large enough to be important? This is a clinical question, not a statistical one.
(2) How wide are the limits of agreement? If it is wide (as defined clinically), the results are ambiguous. If the limits are narrow (and the bias is tiny), then the two methods are essentially equivalent.
(3) Is there a trend? Does the difference between methods tend to get larger (or smaller) as the average increases?
(4) Is the variability consistent across the graph? Does the scatter around the bias line get larger as the average gets higher?
Salam
If you normalize your data by one of the statistical method the percentage of people that fell out of the limits will be estimated and linear regression is appropriate method to test hypo.
متشکرم
محسن میری
You can use linear regression analysis using the difference b/w measurements as the dependant variable, and the average of the measurements as the independent variable. If this test is statistically significant, you can demonstrate that there is, or that there is not, proportional bias. This is an additional statistical test that you can use to further discuss your Bland Altman plot.
@stephaniebernard is the citation for that regression method the Bland Altman paper? They talk about using linear regression, but the language is a little unclear. I have seen people suggest this on YouTube, but can't seem to find anyone making a direct reference to a paper.
Martin De Vita Hi Martin,
These two citations might be helpful, especially the latter one:
Bland , J. M. , & Altman , D. G. ( 1999 ). Measuring agreement
in method comparison studies . Statistical Methods in Medical
Research , 8 , 135 – 160 .
Bland , J. M. , & Altman , D. G. ( 2007 ). Agreement between methods
of measurement with multiple observations per individual .
Journal of Biopharmaceutical Statistics , 17 , 571 – 582 .
Best, Philip
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics