Search on website
Show more
chevron-left-black Summaries

Biostats: Bland Altman plot

The Bland Altman Plot (also known as the Mean-difference plot) is a statistical method of comparing two measurements of the same variable (for example, comparing invasive blood pressure measurement and non-invasive blood pressure measurement). The y-axis plots the difference between the two values (V1-V2) and the x-axis is the mean of the two measurements ([V1+V2]/2, which is the best estimate of the “true value”). Additionally, the plot has three lines: mean, mean + 2SD, and mean – 2SD. The mean +/- 2SD are known as the “limits of agreement.” If these limits fall within the maximum allowed difference between the two methods (meaning the differences within mean +/- 2SD are not clinically important), the methods are considered to be in agreement and may be used interchangeably.

Updated definition 2020.

A useful statistical tool for comparing the means between two measurements of the same variable is the Bland Altman plot. For example, if you were comparing the CO2 measurements of a new capnography device to the current capnography device used in the OR, a Bland Altman plot would let you compare the measurements obtain from each device to determine how similar or different they are. The x axis is the average of the measurements at a point in time from each device. The y-axis is the difference between that pair of measurements. A line is then constructed to travel parallel to the x-axis of this scatter plot and represents the amount of bias or inaccuracy of the new measurement device as compared to the current standard device. A bias line at zero means no difference between the new and old measurement techniques. If the bias line represents the accuracy of the new measurement technique, then the lines of agreement represent the precision. There are lines that travel again parallel to the x axis that contain 95% of the scatter plot points, the upper and lower lines of agreement. If the lines of agreement are narrow together, this suggests a precise measurement tool. If they are wide, the contrary is suggested. To determine if the precision of the tool is of clinical relevance, the lines of agreement can be divided by the mean of the measurements to obtain the percentage error, which can useful in comparing precisions of investigational devices as compared to the current gold standard.


  1. Odor PM, Bampoe S, Cecconi M. Cardiac Output Monitoring: Validation Studies–how Results Should be Presented. Curr Anesthesiol Rep. 2017; 7:410–5. PubMed Link