OBJECTIVE

To determine the test-retest reliability and the minimal clinical change determination for accelerometric parameters, estimated by a new accelerometric-based method that estimates 3-dimensional (3D) linear accelerations of the tibia and femur during comfortable and fast walking speeds.

DESIGN

Test-retest study.

SETTING

Gait laboratory in a research center.

PARTICIPANTS

Patients (N=25; 6 men, 19 women) with symptomatic knee osteoarthritis (OA).

INTERVENTIONS

Not applicable.

MAIN OUTCOMES MEASURES

Subjects attended 2 walking sessions in which accelerometers were rigidly fixed by means of an exoskeleton to the femoral and tibial segments. In both sessions, 3D accelerations were collected for 25 seconds for each of the walking speeds. Mean accelerometric pattern was calculated using 15 gait cycles. From each mean pattern, maximal, minimal, and range values were extracted from the loading phase period. The root mean square (RMS) value was also calculated for every pattern. Relative and absolute reliability were determined using intraclass correlation (ICC) and standard error (SE) of measurement, respectively. Minimal detectable change was calculated for each parameter as the least significant difference.

RESULTS

Tibial and femoral accelerations showed reliable values across sessions 1 and 2 with ICCs greater than or equal to .75 for 96% and 88% of the parameters at comfortable and fast speeds, respectively. The SE of measurement ranged from .01 to .05g for the RMS value and from .05 to .35g for maximal, minimum, and range point parameters.

CONCLUSIONS

The proposed method is the first to have determined the reliability and the minimal detectable change for tibial and femoral acceleration parameters in knee OA patients during a treadmill walking evaluation. The minimal detectable change determined in this study will be used to determine improvement or deterioration of knee OA patients after rehabilitation.