The effect of learning on the 6MWD is large enough to be clinically important when the 6MWT is used to
evaluate response to treatment or change over time. In these situations, two 6MWTs should be performed
and the best 6MWD recorded. Use of two 6MWDs may also decrease the sample size requirements for
clinical trials, due to reduced variability in the pre- and post-intervention measures [11]. Where the 6MWD
is used as a one-off measure to stage disease or assess risk (e.g. likelihood of hospitalisation or mortality),
the magnitude of the learning effect may be less important and one test may be sufficient. However,
clinicians should be mindful of the learning effect if the 6MWD is approaching pre-defined thresholds on
which treatment decisions may be based [12, 13]; in this situation, repeat testing should be considered. One
test may also be sufficient for patients who have recently performed the test, where the learning effect is
smaller (e.g. end of pulmonary rehabilitation) [10