I used the code outlined here:   https://onlinecourses.science.psu.edu/stat509/node/161/ https://onlinecourses.science.psu.edu/stat509/sites/onlinecourses.science.psu.edu.stat509/files/lesson18/19.2_agreement_concordanc/index.sas     However, one of my colleagues pointed out that whilst this would be OK if we had one measurement (Blood Pressure) per patient, the confidence intervals suggested by the above code would be incorrect of there was within subject correlation. Which is understandable as we are measuring BP's in different ways on the same patients. Similarly if we have repeated measurements, then bootstrapping with replacement would work. Going to take a look at Efron's 1994 book on  introduction to bootstrapping but is there a way to code for this or adjust the above code to make it work for repeated measures in the same individuals or different ways of measuring the same thing simultaneously in the same patients. Any suggestions? ... View more