Between-subjects differences of within-subject variability in repeated balance measures: consequences on the minimum detectable change.

When repeating tests on a subject it is important to know if changes of the results are significant. Researchers have used measures of reliability, e.g., intraclass correlation coefficient (ICC) and minimum detectable change (MDC), to assess reliability and validity of balance measures and equipment, and to determine what constitutes a significant change (SC). Using ICC and MDC assumes measurement differences are random in nature, i.e., that within-subject variability is similar between subjects. We investigated if this assumption holds true for center of pressure based balance measures. 20 repetitions of the tests comprising the modified Clinical Test of Sensory Integration in Balance (mCTSIB) protocol were performed by two 17-subject groups simulating laboratory and clinical conditions. Force platform based computerized dynamic posturography Vestibular Technologies CAPS(®) systems exceeding ISPGR recommended accuracy, precision and resolution were used to perform the testing and collect the data. The MDC was compared with the SC computed from individual subjects' within-subject standard deviation, with and without averaging 3 repetitions. We found within-subjects variability was not similar between subjects, rendering the applicability of ICC and MDC questionable. The MDC greatly underestimated the SC for some subjects and overestimated it for others, therefore it should be used with extreme caution, if at all. It seems that for balance measures the SC is more a subject's rather than an instrument's or test's characteristic and should be evaluated on an individual basis. We suggest doing so might provide useful additional clinical information about a subject.