Hello, I am comparing 2 measurement methods (automated versus manual blood pressure measurement) and considered using Bland-Altman plots in SPSS. My one-sample t test of differences was significant (p
Apart from summarising the characteristics (like: mean difference, overestimation etc.) of data for the two methods,
Bland-Altman plot shows good agreement between two methods and data is evenly distributed and concentrated around mean. On average, compared with manual BP measurement, automated systolic BP was 2.35 mm Hg higher [Note: if the difference u took is automated reading - manual reading or else it will become 2.35 mmHg lower]. Limits of agreement (-12.49 to 17.19 mmHg) is quiet close and most of the data points lie within the limits of agreement.
You mention linear regression, however, it is not valid in method comparison studies. This is because linear regression assumes that the independent variable is measured without error - not the situation here. Instead we use Deming or Passing-Bablok regression.
As Sir Adrian Exterman has rightly highlighted that liner regression may not be an appropriate tool in analyzing the automated Vrs manual blood pressure measurement, . Since your test result shows that p value is 0.05 which is a significant at 95 % confidence level, and based on the comparative Bland -Altman plots, which also shows that plots are correlated with each other. The test result shows that both manual and automatic test measure results are inline with each other...