Article

The equilibrium (EQ) score commonly used in computerized dynamic posturography is
normalized between 0 and 100, with falls assigned a score of 0. The resulting
mixed discrete-continuous distribution limits certain statistical analyses and
treats all trials with falls equally. We propose a simple modification of the
formula in which peak-to-peak sway data from trials with falls is scaled
according the percent of the trial completed to derive a continuous equilibrium
(cEQ) score. The cEQ scores for trials without falls remain unchanged from the
original methodology. The cEQ factors in the time before a fall and results in a
continuous variable retaining the central tendencies of the original EQ distribution. A random set of 5315 Sensory Organization Test trials were pooled
that included 81 falls. A comparison of the original and cEQ distributions and
their rank ordering demonstrated that trials with falls continue to constitute
the lower range of scores with the cEQ methodology. The area under the receiver
operating characteristic curve (0.997) demonstrates that the cEQ retained
near-perfect discrimination between trials with and without falls. We conclude
that the cEQ score provides the ability to discriminate between ballistic falls
from falls that occur later in the trial. This approach of incorporating time and
sway magnitude can be easily extended to enhance other balance tests that include
fall data or incomplete trials.
CI - Copyright (c) 2012 Elsevier B.V. All rights reserved.

Langue : ANGLAIS