All outcome measures were collected by an investigator who was blind
to the group assignment. Baseline demographic data including name,
age, gender, date of stroke onset and medical history were collected.
All subjects received an initial and a final assessment session. The
initial assessment was performed before the first treatment session of
the training programme and the final assessment was performed on
the day after the last treatment session. Outcome measures included
a 10MW test and Berg’s Balance Scale (BBS).
Gait performance was measured by a 10MW test. Subjects were
instructed to walk at their fastest speed on a 10-m walkway (14), and
were allowed to use their walking aids or assistive devices. Videorecording
of the tests allowed off-line calculation of gait speed, cadence
and stride length (15). One practice trial was given followed by 3 test
trials, and the trial with the fastest speed was used for further analysis.
The formulae for the calculation are as follows:
• gait speed (cm/s) = 1,000 cm/time (s) to walk the 10-m walkway,
• cadence (steps/min) = number of steps to walk the 10-m walkway
× 60/time (s) to walk the 10-m walkway,
• stride length (cm) = 1,000 cm/number of steps to walk the 10-m
walkway.
Balance performance was determined by BBS. The scale consists
of 14 items, including the ability to maintain sitting balance, static
and dynamic standing balance, and stability during functional transfer
tasks (16). Each item is scored from 0 (cannot perform the task) to
4 (the best performance), with a total score range of 0–56. BBS is a
well-established balance assessment tool that has been found to have
excellent inter- and intra-rater reliability (r > 0.98) (17). Before the
main study, the test-retest reliability of the procedures for testing gait
parameters and BBS was examined in 5 patients with chronic stroke.
The results indicated excellent reliability, as shown by the intra-class
correlation coefficients (gait speed = 0.97, cadence = 0.98, stride
length = 0.89, BBS = 1.00) (18).
Statistical analysis
SPSS 16 software was used for data analysis (SPSS Inc., Chicago, IL,
USA). Descriptive analysis was used to analyse the demographic data
in each subject group. A 2-way repeated measures analysis of variance
(ANOVA) was used to analyse gait and balance performance with time
(week 0 and week 2) as the within-group factor and gait (experimental
and control) as the between-group factor. When an interaction between
group and time was found, post-hoc tests were used to determine the
real within- and between-group differences. We further computed the
changes in each outcome measure (gait speed, cadence, stride length,
BBS score) between the pre- and post-treatment assessment sessions.
The mean differences and 95% confidence intervals (95% CI) were
calculated for the experimental and control groups. The level of significance
was set at 0.05.