Wheelchair Propulsion Training
Principal Investigator: Michael L. Boninger, M.D.
Funding Source: National Institutes of Health
2006-2008
Manual wheelchair
users are at high risk of developing upper limb pain. Recent research, which
includes work complete at our center and ergonomics literature point to specific
propulsion techniques that may prevent injury. The specific aims of this
proposal were to 1) Develop wheelchair propulsion-training programs that
minimized potential injurious biomechanics; 2) Test if the training programs can
cause lasting changes in propulsion biomechanics; and 3) To determine if the
addition of real time feedback to a multimedia program is better than a
multimedia program alone at improving propulsion mechanics and maintaining
improvements over time. The study design was a longitudinal controlled trial
with a control group, a multimedia group, and a multimedia and real-time
feedback group. Training occurred over three weeks and subjects were followed
for three months.
A total of 28 long
time manual wheelchair users (4 female and 24 male) were enrolled in this study.
Specific aim 1 was achieved by developing a wheelchair propulsion training
program based on biomechanical literature, clinical practice guidelines,
ergonomic principles and motor learning theory. The program was designed to
teach efficient propulsion technique which may help minimize the development of
upper extremity pain and injury. The two interventions created as a part of the
training study were a real time biomechanical feedback (RTF) software display
and a multimedia instructional program (MMP). The MMP was an automated
instructional video and slide presentation highlighting common injuries that
occur as a result of propulsion and how to use specific propulsion techniques to
minimize their development. The RTF component of the study, incorporated
principles of motor learning theory translated into visual feedback-learning
software presenting the same key spatio-temporal and kinetic variables as
explained by the MMP. This included minimizing the number of real time feedback
variables presented simultaneously while displaying them discontinuously and in
random order between training and rest periods. Contact angle (CA)
(degrees pushed during each propulsive phase), velocity (m/s), and cadence
(strokes/sec) were the feedback variables selected because they have been shown
to have a strong association with the development of upper limb impairments. The
presentation of cadence and push angle feedback ultimately encouraged subjects
to take longer, less frequent strokes, to decrease force exerted at a given
velocity, without causing unintended changes in force direction.
To investigate
specific aim 2, data was collected for all subjects at a self selected and pre
determined velocity during propulsion on a dynamometer, carpet, ramp and tile.
Preliminary analysis of data from 22 individuals revealed that as hypothesized,
the treatment groups (FB and IO) displayed statistically significant favorable
biomechanical changes pre to post training (baseline to 3 months) compared to
the CG. Although analysis for all conditions is not complete, the FB group did
not consistently display statistically significantly differences from the IO
group.
For specific aim
3, full statistical analysis for long term change (baseline-3months post
training) has been performed for the self selected speeds conditions on the
dynamometer and ramp. Main effect statistical findings indicate that while
propelling on a dynamometer 3 months after training, both FB and IO treatment
groups displayed a statistically significant change (increase) in contact angle
where control group CG subjects remained nearly static (p =.001). The
interaction contrast effects showed that both FB and IO groups each displayed
significant increases in contact angle compared to the CG (p =.001) however
significant difference were not found between FB and IO groups (p =.216).
Statistical significance was also found in cadence (decrease) pre to post
training both in main effect (p =.014) and between the IO and CG groups (p
=.007). The change in cadence of the FB group compared to both IO and CG were
not found to be significant (p =.24, p =.11). In addition, subject age was found
to be a significant predictor of contact angle size on the dynamometer at
baseline and 3 months post training for all groups(p =.047). Sex, level of
injury, and weight were not found to be predictors of contact angle or
frequency. For the ramp condition, significance was found in the main effect for
change in contact angle (increase) (p=.021) and with the FB group compared to
the CG (p=.006). Comparisons of IO to CG and FB to IO were not found to be
statistically significant (p =.17, p =.09). Main effect changes in cadence were
not found to be statistically significant (p=.2).
We are no longer enrolling participants in this study. Data analysis is ongoing. A manuscript is in review related to manual wheelchair propulsion training using biomechanical real time visual feedback. It is our hope that the design of a training program can help to minimize the upper extremity pain and injury often associated with manual wheelchair propulsion. In the future, these experiments may assist clinicians in training patients in manual wheelchair propulsion.