Bland–Altman plot

Bland–Altman plot

A Bland–Altman plot is a method of data plotting used in biostatistics to analyze the agreement between two different assays or measurements. It is named after J. Martin Bland and Douglas G. Altman, who described the approach in a 1986 paper.

Overview

The Bland–Altman plot is a graphical method to compare two measurements techniques. It is used to identify any systematic difference between the measurements (i.e., fixed bias) and to identify possible outliers. The plot is also known as a difference plot.

Construction

To construct a Bland–Altman plot, the following steps are typically followed: 1. Calculate the mean of the two measurements for each subject. 2. Calculate the difference between the two measurements for each subject. 3. Plot the differences against the means.

The plot usually includes:

• A horizontal line at the mean difference (which represents the average bias).
• Horizontal lines at the mean difference plus and minus 1.96 times the standard deviation of the differences (which represent the limits of agreement).

Interpretation

The Bland–Altman plot helps to visualize the agreement between two methods of measurement by showing the spread of differences and identifying any patterns. If the differences are randomly scattered around the mean difference, it suggests good agreement. If there is a trend or pattern, it may indicate a systematic bias or other issues with the measurements.

Applications

Bland–Altman plots are widely used in various fields, including clinical research, psychometrics, and method comparison studies. They are particularly useful when comparing a new measurement technique with a gold standard or when validating new instruments.

Advantages and Limitations

Advantages

• Simple to construct and interpret.
• Provides a clear visual representation of the agreement.
• Identifies both fixed and proportional bias.

Limitations

• Assumes that the differences are normally distributed.
• May not be suitable for small sample sizes.
• Does not provide a formal statistical test for agreement.

Related Concepts

• Correlation coefficient
• Regression analysis
• Measurement error
• Method comparison studies

See Also

• J. Martin Bland
• Douglas G. Altman
• Biostatistics
• Clinical research