Bland-Altman plots when you have collected multiple observations per individual

 The method that Prism uses to create Bland-Altman plots implicitly assumes that each pair of values entered is independent. This assumption is violated when you collect data from multiple individuals and measure each individual multiple times. If you have ten individuals, and use both methods three times on each, you do not have thirty independent pairs of measurements. If you enter all the data into Prism, the resulting Bland-Altman plot will be misleading. 

Bland and Altman have published extensions of their method to deal with this situation, but Prism does not (yet) implement this method.

 Bland and Altman. Agreement between methods of measurement with multiple observations per individual. Journal of Biopharmaceutical Statistics (2007) vol. 17 pp. 571-582