Is comparison of changes in cardiac output, assessed by different methods, better than only comparing cardiac output to the reference method?