544983

research-article2014
CRE0010.1177/0269215514544983Clinical RehabilitationHalvarsson et al.

CLINICAL
Article REHABILITATION

Clinical Rehabilitation

Balance training with multi-task 2015, Vol. 29(4) 365­–375

© The Author(s) 2014
Reprints and permissions:
exercises improves fall-related self- sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/0269215514544983

efficacy, gait, balance performance cre.sagepub.com

and physical function in older adults

with osteoporosis: a randomized
controlled trial

Alexandra Halvarsson1,2, Erika Franzén1,2 and Agneta Ståhle1,2

Abstract
Objective: To evaluate the effects of a balance training program including dual- and multi-task exercises
on fall-related self-efficacy, fear of falling, gait and balance performance, and physical function in older
adults with osteoporosis with an increased risk of falling and to evaluate whether additional physical
activity would further improve the effects.
Design: Randomized controlled trial, including three groups: two intervention groups (Training, or
Training+Physical activity) and one Control group, with a 12-week follow-up.
Setting: Stockholm County, Sweden.
Participants: Ninety-six older adults, aged 66-87, with verified osteoporosis.
Interventions: A specific and progressive balance training program including dual- and multi-task three
times/week for 12 weeks, and physical activity for 30 minutes, three times/week.
Main measures: Fall-related self-efficacy (Falls Efficacy Scale-International), fear of falling (single-item
question – ‘In general, are you afraid of falling?’), gait speed with and without a cognitive dual-task at
preferred pace and fast walking (GAITRite®), balance performance tests (one-leg stance, and modified
figure of eight), and physical function (Late-Life Function and Disability Instrument).
Results: Both intervention groups significantly improved their fall-related self-efficacy as compared to the
controls (p ⩽ 0.034, 4 points) and improved their balance performance. Significant differences over time
and between groups in favour of the intervention groups were found for walking speed with a dual-task
(p=0.003), at fast walking speed (p=0.008), and for advanced lower extremity physical function (p=0.034).
Conclusions: This balance training program, including dual- and multi-task, improves fall-related self-
efficacy, gait speed, balance performance, and physical function in older adults with osteoporosis.

Keywords
Balance, elderly, osteoporosis, gait, falling
Received: 31 January 2014; accepted: 5 July 2014

1Karolinska Institutet, Department of Neurobiology, Care Corresponding author:

Sciences and Society, Division of Physiotherapy, Stockholm, Alexandra Halvarsson, Karolinska Institutet, Department of
Sweden Neurobiology, Care Sciences and Society (NVS), Division of
2Karolinska University Hospital, Department of Physical Physiotherapy, 23100, Huddinge, SE-141 83, Sweden.
Therapy, Stockholm, Sweden Email: alexandra.halvarsson@ki.se

Downloaded from cre.sagepub.com at LSU Libraries on November 13, 2016

366 Clinical Rehabilitation 29(4)
Introduction
Falls and fall-related injuries are one of the most
common causes of ill health and morbidity in older
adults and lead to functional impairment, disabil-
ity, lower quality of life, and fractures.1 Fear of
falling increases with age, is more common among
women, and has been associated with an increased
risk of falling1 and fall-related injuries,2 as are
decreased walking speed,3,4 balance performance,4
and physical function.5
Walking speed is a strong independent predictor
of self-perceived function in older adults,6,7 and
older people with the ability to walk faster than 1
m/s have generally good functional status, lower
risk of health events, and better survival.8
The ability to divide attention between a motor
and a cognitive task, also known as dual-tasking, is
a natural component of our daily activities and
known to affect balance performance, gait, and the
risk of falling in the elderly.9 When performing two
or more task simultaneously that are attention
demanding, at least one of the task may deterio-
rate;10 for example, an added cognitive task will
influence gait pattern with lower walking speed as
a result.11
Osteoporosis is common among older people
worldwide and older adults with osteoporosis have
reduced balance control, gait performance, and an
increased fear of falling.12,13 Taken together, this
may lead to activity avoidance, decreased physical
function and quality of life, as well as an increased
risk of falling.2 Falls by older people with osteopo-
rosis can lead to consequences that may be both
detrimental and of great cost to society.14
A Cochrane review regarding exercises for
improving balance control in older people con-
cluded that there is weak evidence for several exer-
cises such as gait, balance, coordination, and
functional tasks, to be effective in improving bal-
ance in older adults, and that further high-quality
methodological research is required.15,16
Our research group has developed a balance
training program with functional and progressive
tasks for older adults with balance problems. The
program has been evaluated successfully in both
short- and long-term perspectives, resulting in a
reduced fear of falling and improved gait in healthy
Downloaded from cre.sagepub.com at LSU Libraries on November 13, 2016
older adults who have a fear of falling and balance
deficits.17,18 This intervention program now has
been developed further for other groups of the
elderly population, such as those with osteoporosis
and the fear of falling, because this group has been
shown to have an increased risk of falling and fall-
related injuries and fractures.2
The aim of the present study was to evaluate the
effects of a balance training program including
dual- and multi-task exercises on fall-related self-
efficacy, fear of falling, gait and balance perfor-
mance, and physical function in older adults with
osteoporosis with an increased risk of falling, as
well as to evaluate whether additional physical
activity would further improve the effects.
Materials and methods
The study was a randomized controlled trial
(BETA-study; NCT01417598, ClinicalTrials.gov)
with a 12-week follow-up. Data were collected
from spring 2010 to autumn 2011. Community-
dwelling older adults with osteoporosis and an
increased risk of falling were recruited by adver-
tisement in local newspapers in Stockholm County,
the Swedish Osteoporosis Society, and Karolinska
University Hospital. Osteoporosis was defined as a
bone mineral density value of ≥ 2.5 standard devia-
tions below the mean seen in young female sub-
jects.25 Inclusion criteria included: age ≥65 years
with diagnosed osteoporosis, being afraid of falling
and/or having experienced at least one fall in the
last 12 months, and independence in ambulation.
Participants were excluded if they had experienced
fractures during the last year, had a Mini-Mental
State Examination (MMSE) score < 24,19,20
severely decreased vision, or other diseases or con-
straints that might interfere with participation in
the exercise program. The study was approved by
the local ethics committee in Stockholm, Sweden
(2009/819-32, 2012/1829-32), and all participants
gave their written informed consent.
Participants were assessed by experienced
physiotherapists at baseline and at the 12-week
follow-up under the same conditions. Baseline
testing included a learning session where the par-
ticipants performed all the physical measurements
once. Subsequently, the sessions started randomly
Halvarsson et al. 367
with either questionnaires or physical assessments.
After the baseline testing, the subjects were rand-
omized in blocks of nine into three different
groups: two intervention groups (Training, and
Training+Physical activity), or a Control group,
using Web-based software. The test leaders were
blinded to group allocation at baseline; however, it
was no longer possible after baseline testing,
because some of the test leaders were also involved
in the balance training. Sample size was calculated
on the primary outcome measure Falls Efficacy
Scale-International on data from a previous study
by Halvarsson and colleagues,17 resulting in 21
subjects in each group.
The balance training was performed in groups
of six to10 participants, 45 minutes/session, and
three times/week for 12 weeks. Two to three physi-
otherapists were present at each session to ensure
the participants’ safety and to make progression of
the balance tasks possible. The training comprised
exercises targeting various systems for postural
control, such as stability limits (changes in base of
support during sitting and standing, reaching/lean-
ing), sensory orientation (walking/standing/sitting
on uneven surfaces, eyes open/closed), gait (walk-
ing at different pace and/or performing dual- and
multi-task – i.e., by adding cognitive or/and motor
tasks to an exercise), and postural responses (reac-
tions to balance loss that will occur when the bal-
ance is highly challenged). The exercises were
performed on three different levels (basic, moder-
ate, and advanced), enhancing progression, and
making it challenging for each individual. Exercises
were individually adjusted to be challenging for
each participant; for example, during sitting, stand-
ing, or walking on a compliant surface, the surfaces
had different density and size depending on the
individual’s level. 21 Participants needed to attend
at least 24 of 36 possible training sessions (66%) to
be included in the follow-up assessments.
In addition to the training program, participants
in the Training+Physical activity group were
instructed to walk for at least 30 minutes, three
times/week (preferably with poles, i.e., Nordic
walking) during the intervention period and keep
notes of how many minutes and times/week they
walked. Participants in the Control group were
Downloaded from cre.sagepub.com at LSU Libraries on November 13, 2016
encouraged to live their regular lives and were
offered the same balance training at the end of the
study.
The primary outcome was fall-related self-effi-
cacy as measured by the Falls Efficacy Scale –
International (FES-I).22-24
Secondary outcomes were effects for fear of
falling, gait, balance, physical function, and physi-
cal activity level. Fear of falling was measured
with a single-item question – ‘In general, are you
afraid of falling?’ – with possible responses being
‘not at all’, ‘a little’, ‘quite a bit’, and ‘very much’.25
To measure gait speed, the GAITRite® system
(CIR Systems, Inc., Haverton, PA, USA),26,27 was
used. Participants walked at their preferred speed
with and without a cognitive dual-task (reciting
every second letter of the Swedish alphabet), as
well as at their maximal speed. Six trials for each
condition were collected, and a mean of each con-
dition was used for analysis. The test leaders regis-
tered the total number of letters recited and the
total number of errors.
Balance performance was assessed with the
one-leg stance and the modified figure-of-eight
test. For the one-leg stance, subjects stood alter-
nately on the right or the left leg for as long a time
as possible (maximum 30 seconds) with their hip
and knee slightly flexed, arms hanging down and
eyes open, 28-30 three times/leg, and a mean of all
six measures was used for analysis. The modified
figure-of-eight test consisted of two circles (ø163
cm) forming a figure of eight. The test leader meas-
ured the time taken to complete two figures of
eight and noted the number of oversteps (i.e., occa-
sions when no part of the shoe touched the taped
line).29 The modified figure-of-eight was per-
formed three times and a mean was used for
analysis.
To assess self-reported physical function, the
function component of the Late-Life Function and
Disability Instrument (LLFDI) was used.31,32 The
function component evaluates self-reported diffi-
culties in performing 32 physical activities com-
prised of three subscales: upper extremity, basic
lower extremity, and advanced lower extremity.
Physical activity level was assessed with the
Frändin-Grimby activity scale,33 which includes
368 Clinical Rehabilitation 29(4)
physical activity, exercise, and household activi-
ties, ranging from 1 (hardly any physical activity)
to 6 (hard or very hard exercise regularly, and sev-
eral times a week).
For statistical calculations, PASW Statistics ver-
sion 20.0 (SPSS Inc., Chicago, IL, USA) was used.
Data are presented as means, ± standard deviations
(SD), medians, minimum (min) maximum (max),
number of (n), and percentage (%). To analyse differ-
ences for ordinal variables or not-normally distrib-
uted variables (Falls Efficacy Scale – International,
one-leg stance, modified figure-of-eight, and physi-
cal activity level), the difference between the two test
occasions was calculated, and the Kruskall-Wallis
test and the Mann-Whitney U-test were used for
comparisons between groups. For within-group
comparisons, the Wilcoxon signed-rank test was
used. Normally distributed quantitative data (gait
speed and Late-Life Function and Disability
Instrument) were investigated with two-factor,
repeated measures analysis of variance (ANOVA)
as a mixed design (General Linear Model), with a
main effect of factor 1 (time) and factor 2 (group),
and interaction effects of factors 1 and 2. P-value
was set to ⩽ 0.05.
Results
A total of 351 persons reported their interest to take
part in the study. Ninety-six (94 women and 2 men)
met the inclusion criteria and were randomly allo-
cated to: Training group (n=34), Training+Physical
activity (n=31), or a Control group (n=31). The
mean age was 76 years (range: 66-87), and 78%
had experienced a fall during the previous year.
During the follow-up period, 27 dropped out result-
ing in just 69 participants fulfilling the entire study
period (Training group: n=25, Training+Physical
activity group: n=18, and Control group: n=26);
see Figure 1. There were no significant differences
(p ⩽ 0.05) at baseline between the three groups,
except for physical activity level (p = 0.045), with
a significant difference between Training group
and Training+Physical activity group (p = 0.008).
There were no significant differences between the
drop-outs and those fulfilling the study. For base-
line characteristics of the 69 participants who
Downloaded from cre.sagepub.com at LSU Libraries on November 13, 2016
fulfilled the study period, see Table 1. Adherence
rates to the training sessions were 89%, with a
range of 66-100%. All except one participant in the
Training+Physical activity group fulfilled the
added physical activity intervention. According to
the notes the participants in the Training group
were regularly physically active as compared to the
Training+Physical activity group, and there were
no differences in which activities they performed.
Consequently, there were no differences between
these two groups regarding the intervention.
All the three groups had significantly improved
their fall-related self-efficacy, measured with the
Falls Efficacy Scale – International, from baseline
to follow-up. Further analysis also revealed that
both training groups had a significantly larger
improvement on the Falls Efficacy Scale –
International score as compared to the Control
group.
No significant change from baseline to follow-
up was found for fear of falling between the three
groups. However, both intervention groups had a
significant decrease in fear of falling from baseline
to follow-up; see Table 2.
No significant interaction was shown for pre-
ferred walking speed. However, both training
groups increased their preferred walking speed
from baseline to follow-up while the Control group
remained unchanged.
At fast speed, a significant interaction was
shown, revealing that Training group (0.13 m/s)
and Training+Physical activity group (0.10 m/s)
significantly increased their fast walking speed
from baseline to follow-up. During walking at pre-
ferred speed with a cognitive dual-task, a signifi-
cant change was shown over time and between
groups, revealing that only the Training+Physical
activity group significantly increased their walking
speed during dual-task condition (0.14 m/s). See
Table 2.
There were no significant differences from
baseline to follow-up for both one-leg stance and
modified figure-of-eight when comparing the three
groups (Table 2). However, within-group analysis
revealed that both training groups increased their
time in one-leg stance (5.5s -11.9s) from baseline
to follow-up, and the Training+Physical activity
Halvarsson et al. 369

Figure 1.  Flow chart of individuals reporting interest in study, number of: excluded, allocated by random to a
Training Group, Training+Physical activity Group, or Control Group at baseline.

group had significantly reduced their over steps in
the modified figure of eight.
A significant improvement was shown over time
and between groups, in the advanced lower extrem-
ity function, measured with the Late-Life Function
and Disability Instrument. Both training groups sig-
nificantly increased their advanced lower extremity
scores from baseline to follow-up, while the Control
group scores remained unchanged. No significant
interactions were seen in the other domains of the
Late-Life Function and Disability Instrument.
Within-group analysis, from baseline to follow-up,
revealed significant improvement for the two train-
ing groups in the total score and in the subscale basic
lower extremity scores; additionally, the Training
group also increased their upper extremity function.
No significant differences between groups were
shown at follow-up for physical activity level;
however, within-group analysis revealed that the
Training+Physical activity group significantly
increased their physical activity level from base-
line to follow-up. See Table 2.
Since there were no differences when compar-
ing the two groups for physical activity level and
which activity performed, in other words, the
added physical activity in the Training+Physical
activity group, the two training groups were merged
into one Intervention group. This resulted in

Downloaded from cre.sagepub.com at LSU Libraries on November 13, 2016

370 Clinical Rehabilitation 29(4)

Table 1.  Baseline characteristics for the 69 participants that completed the study period divided into the three
randomized groups: Training, Training+Physical activity, and Control groups.

Variable Training Training+Physical Control

group activity group group

  n=25 n=18 n=26

Gender female/male 25/0 18/0 25/1
Age mean (min-max) 76 (66-87) 77 (69-87) 76 (68-85)
Body mass index (kg/m2) median (min- 25 (19-29) 25.5 (16-35) 26 (17-33)
max)
Height mean (min-max) 162 (153-171) 159 (141-171) 163 (149-189)
Mini Mental State Examination median 29 (24-30) 28 (25-29) 28 (25-30)
(min-max)
Civil status living alone/together 60%/40% 67%/33% 65%/35%
Living house/apartment 24%/76% 28%/72% 19%/81%
Education level 16%/36%/48% 33%/33%/34% 15%/42%/43%
Work blue/white collar 60%/40% 56%/44% 58%/42%
Number of prescribed medications 3 (1-15) 6 (1-12) 4 (1-13)
median (min-max)
Diabetes yes 8% 6% 0%
Hypertension yes 48% 44% 39%
Myocardial infarction yes 4% 0% 8%
Heart failure yes 4% 17% 4%
Angina yes 2% 0% 0%
Stroke yes 8% 6% 4%
Lung disease yes 0% 28% 19%
Cancer yes 12% 39% 27%
Experienced a fall during the last 12 72% 72% 89%
months Yes
Fear of falling 0%/60%/28%/12% 0%/44%/28%/28% 0%/31%/42%/27%
Not at all/a little/quite a bit/very much

Variables are presented as percentages, mean, median, and min-max.

significant differences between the groups over
time; in other words, the differences were more
pronounced when analysing the two groups now
called Intervention compared to the Control group.
Further analysis revealed that the Intervention
group had significantly improved in all variables
measured from baseline to follow-up, except for time
in modified figure of eight (p=0.056). See Table 3.
Discussion
This balance training program focusing on dual-
and multi-task exercises improves fall-related self-
efficacy in older adults with osteoporosis, as well
as preferred walking speed with and without a cog-
nitive dual-task, fast walking, and self-reported
physical function. However, it is unclear if added
physical activity gives additional benefits to the
balance training program, as all participants already
were regularly physically active as recommended.
After the intervention, the fall-related self-
efficacy as measured with the Falls Efficacy Scale
– International, decreased in all groups, a similar
finding to our previous study on healthy older peo-
ple at risk of falling.23 The Falls Efficacy Scale –
International score has to change more than 2.9
points to ensure a clinically relevant change
between groups.24 In the present study, both training

Downloaded from cre.sagepub.com at LSU Libraries on November 13, 2016

 Table 2.  Median, min-max, mean, SD for FES-I, OLS, MFE, gait speed during preferred speed with and without a cognitive dual-task, fast speed, physical function
assessed with LLFDI at baseline and at 12-weeks’ follow-up for the three groups (Training, Training+Physical activity, and Control).
Variables Training group (TG) n=25 Training + Physical activity group (TPAG) Control group n=26 Effect size  
n=18

  Baseline Follow-up p-value Baseline Follow-up p-value Baseline Follow-up p-value p-value TG-CG TPAG-CG TG-TPAG  
baseline baseline baseline
follow-up follow-up follow-up

Kruskall Non-parametric  
Wallis
FES-I median (min- 26 (20-47) 22 (17-32) <0.001 26 (20-46) 22 (17-38) 0.001 27.5 (21-39) 25 (17-41) 0.015 0.044 0.3 0.0 0.0  
max)
OLS median (min- 6.2 (1.8-30.0) 18.1 (1.7-30.0) 0.019 4.7 (0.5-30.0) 10.2 (1.9-30.0) 0.020 8.9 (1.1-30.0) 9.0 (0.6-30.0) 0.977 0.099 0.3 0.0 0.0  
max), s
MFE time median 24.5 (18.0-59.6) 26.4 (16.4-58.0) 0.053 26.0 (17.9-62.2) 28.9 (20.0-56.0) 0.500 27.8 (15.3-89-2) 25.8 (18.6-55.5) 0.990 0.444 0.2 0.2 0.0  
(min-max), s
MFE oversteps   3 (0-9)   2 (0-13) 0.567   4 (0-24)   3 (0-8) 0.035   2 (0-50)   3 (0-46) 0.445 0.063 0.2 0.1 0.2  
median (min-max), n
Physical activity   4 (2-4)   4 (3-5) 0.102   3 (2-4)   3 (2-4) 0.025 3.5 (2-5)   4 (2-4) 0.813 0.220 0.2 0.0 0.1  
median (min-max)
Fear of falling   1 (1-3)   1 (0-3) 0.044   2 (1-3)   1 (0-3) 0.046   2 (1-3)   2 (0-3) 0.771 0.243 0.2 0.2 0.0  
median (min-max)
No percent 0 40 0 22 0 8  
A little percent 60 32 44 50 31 23  
Quite a bit percent 28 16 28 17 42 42  
Very much percent 12 12 28 11 27 27  
Gait speed (m/s) Time Time*group Cohen′s d independent
mean (SD) samples
Preferred speed 1.21 (0.21) 1.28 (0.22) 0.025 1.09 (0.21) 1.18 (0.19) 0.010 1.17 (0.21) 1.16 (0.21) 0.929 0.005 0.086 0.5 0.7 0.2
single-task
Preferred speed 1.15 (0.24) 1.19 (0.24) 0.131 1.00 (0.36) 1.14 (0.31) <0.001 1.09 (0.24) 1.06 (0.29) 0.368 0.005 0.003 0.5 1.1 0.4
dual-task

Downloaded from cre.sagepub.com at LSU Libraries on November 13, 2016

Error in the 21 (14) 17 (11) 0.032 19 (13) 16 (11) 0.146 23 (14) 17 (12) 0.005 <0.001 0.754 0.1 0.3 0.1
performance of
the dual-task in
percentage
Fast speed 1.53 (0.26) 1.66 (0.26) <0.001 1.42 (0.28) 1.52 (0.29) 0.003 1.52 (0.34) 1.52 (0.34) 0.880 <0.001 0.008 0.8 0.9 0.1
LLFDI mean (SD)  
Functional total 59.9 (8.2) 64.4 (6.8) <0.001 56.0 (6.1) 58.6 (8.6) 0.038 57.7 (6.6) 59.3 (8.9) 0.132 <0.001 0.157 0.6 0.2 0.3
Upper extremity 71.5 (7.8) 76.3 (9.1) 0.003 68.4 (9.7) 69.0 (9.4) 0.742 71.6 (7.5) 74.4 (9.6) 0.069 0.005 0.216 0.3 0.3 0.5
Basic lower 68.8 (11.9) 72.9 (9.7) 0.016 65.5 (8.9) 70.1 (13.5) 0.020 67.1 (10.3) 69.4 (12.2) 0.157 0.001 0.613 0.2 0.3 0.1
extremity
Advanced lower 53.4 (13.3) 61.0 (11.6) <0.001 46.3 (10.4) 50.4 (13.6) 0.034 48.3 (12.1) 49.9 (13.6) 0.313 <0.001 0.034 0.7 0.3 0.4
extremity

The Kruskall-Wallis test was used to analyse differences between groups over time for the FES-I, OLS, and MFE; a two-factor repeated-measures ANOVA was used for gait speed and physical function.
SD: standard deviation, FES-I: Falls Efficacy Scale International, OLS: one leg stance, MFE: modified figure-of-eight test, LLFDI: Late Life Disability and Function Instrument.
Significance level at ⩽ 0.05 is marked by bold type. Effect size for non-parametric data and Cohen′s d for independent samples.
 372

Table 3.  Median, min-max, mean, SD for FES-I, OLS, MFE, gait speed during preferred speed with and without a cognitive dual-task, fast speed, physical
function assessed with LLFDI at baseline and at 12-weeks’ follow-up for the two groups (Intervention (both training groups) and Control).
Variables Intervention group n=43 Control group n=26 p-value Effect size  

  Baseline Follow-up p-value Baseline Follow-up p-value  

baseline baseline
follow-up follow-up

Mann-Whitney U Non-parametric  
FES-I median (min-max) 26 (20-47) 22 (17-38) <0.001 27.5 (21-39) 25 (17-41) 0.015 0.012 0.3  
OLS median (min-max), s 6.0 (0.5-30.0) 12.8 (1.7-30.0) 0.001 5.0 (1.1-30.0) 5.9 (0.6-30.0) 0.977 0.032 0.3  
MFE time median (min-max), s 25.8 (17.9-62.2) 28.4 (16.4-58.0) 0.056 27.8 (15.3-89.2) 25.8 (18.6-55.5) 0.990 0.207 0.2  
MFE oversteps median (min-max), n 4 (0-24) 2 (0-13) 0.034 2 (0-50) 3 (0-46) 0.445 0.046 0.2  
Physical activity median (min-max) 3 (2-4) 4 (2-5) 0.007 3.5 (2-5) 4 (2-4) 0.813 0.108 0.2  
Fear of falling median (min-max) 1 (1-3) 1 (0-3) 0.004 2 (1-3) 2 (0-3) 0.771 0.099 0.2  
No percent    0   33 0 8  
A little percent   54   40 31 23  
Quite a bit percent   28   16 42 42  
Very much percent   19   12 27 27  
Gait speed (m/s) mean (SD) Time Time*group Cohen′s d
Independent
samples
Preferred speed single-task 1.16 (0.22) 1.24 (0.21) 0.001 1.17 (0.21) 1.16 (0.20) 0.929 0.043 0.031 0.6
Preferred speed dual-task 1.09 (0.30) 1.17 (0.27) 0.001 1.09 (0.24) 1.06 (0.29) 0.378 0.131 0.005 1.0
Error in the performance of the 20 (14) 17 (11) 0.010 23 (14) 17 (12) 0.879 <0.001 0.488 0.2
dual-task in percentage
Fast speed 1.49 (0.27) 1.60 (0.28) <0.001 1.52 (0.34) 1.52 (0.34) 0.005 0.001 0.002 0.5

Downloaded from cre.sagepub.com at LSU Libraries on November 13, 2016

LLFDI mean (SD)  
Functional total 58.3 (7.6) 62.0 (8.1) <0.001 57.7 (6.6) 59.3 (8.9) 0.132 <0.001 0.114 0.4
Upper extremity 70.2 (8.7) 73.2 (9.8) 0.648 71.6 (7.5) 74.4 (9.6) 0.510 0.004 0.897 0.0
Basic lower extremity 67.4 (10.7) 71.7 (11.4) 0.001 67.1 (10.3) 69.4 (12.2) 0.154 0.002 0.333 0.2
Advanced lower extremity 50.4 (12.6) 56.5 (13.4) <0.001 48.3 (12.1) 49.9 (13.6) 0.316 <0.001 0.028 0.6

The Mann-Whitney U-test was used to analyse differences between groups over time for the FES-I, OLS, and MFE; a two-factor repeated-measures ANOVA was used for gait speed and physical
function.
SD: standard deviation, FES-I: Falls Efficacy Scale International, OLS: one leg stance, MFE: modified figure-of-eight test, LLFDI: Late Life Disability and Function Instrument.
Significance level at ⩽ 0.05 is marked by bold type. Effect size for non-parametric data and Cohen′s d for independent samples.
Clinical Rehabilitation 29(4)
Halvarsson et al. 373
groups changed their scores by 4 points, whereas
the Control group only showed a 1.5 point decrease,
resulting in clinically relevant changes in both train-
ing groups. Fear of falling, measured with the sin-
gle-item question – ‘In general, are you afraid of
falling?’ – also decreased in the two training groups.
Fear of falling leads to self-restricted physical
activity and avoidance of activities that may leads
to a fall,2 which may have a great impact on activi-
ties in daily living and lead to dependency. It is
important, especially among older people, to
reduce the level of the fear of falling and improve
fall-related self-efficacy.
During daily life situations, it is important to be
able to increase gait speed, especially as when cross-
ing the street or avoiding obstacles. Previous research
has shown that an improvement of 0.10 m/s in walk-
ing speed is considered a substantial change,34 and
impacts daily life in a positive way—that is, it
improves physical function and quality of life.7
After participating in the balance training pro-
gram, the participants were able to increase their
fast walking speed by more than 0.10 m/s. When
merging the two training groups together, we also
found improvement in preferred walking speed
(0.8 m/s).
Walking with a dual-task reflects activities in
daily life in which we perform tasks simultane-
ously, for example while walking and talking to
someone. The Training+Physical activity group sig-
nificantly increased their walking speed during the
dual-task condition by 0.14 m/s after the interven-
tion as compared to baseline. However, the
Training+Physical activity group had lower base-
line values and, therefore, may have had a greater
potential to improve during the study period. In line
with this supposition, both training groups walked
at similar speed during dual-tasking at follow-up.
The lack of improvement in balance perfor-
mance when comparing the three groups may indi-
cate a problem with power, as when merging the
two training groups to one intervention group, we
found significant improvements in one leg stance
and modified figure of eight.
We also observed a significant decrease of over
steps in the modified figure of eight, which may
indicate a better balance performance with the
Downloaded from cre.sagepub.com at LSU Libraries on November 13, 2016
ability to walk with less over steps during the same
measured time. Our results for time and over steps
correspond well with data from a cross-sectional
study on community-dwelling older adults that
consider themselves healthy.29
The balance training program in the present
study includes exercises that place a demand on the
lower extremities and particularly on more
advanced functions, such as walking on various
terrains, over obstacles, and on different surfaces
(slippery or uneven). The fact that participants
improved self-rated physical function, as measured
with the Late-Life Function and Disability
Instrument with respect to advanced lower extrem-
ity function, supports that the program was
successful.
That the sample was highly skewed in gender
distribution (two men) might be seen as a limita-
tion of the generalizability of the results, and the
results, therefore, may not be valid for men.
However, more women are diagnosed with osteo-
porosis,35 perceive a higher-level fear of falling,1
and may be more conscious of their balance defi-
cits than men and, therefore, may be more inter-
ested in participating in this type of study.
The lack of blinding of the assessors after base-
line testing could be considered as a bias. However,
in the present study, there were always two asses-
sors present during the physical tests in order to
ensure that the participants got the same standard-
ized instructions and that the tests were performed
in the same way.
In the present study, only the analyses of short-
term effects of the balance training program for
those who fulfilled the study period, per protocol
analysis, have been performed. Hence, the authors
could not interpret the long-term effects of this
study. It is therefore of interest to evaluate the long-
term effects and also include an intention-to-treat
analysis.
During the follow-up period, 27 participants
dropped out, mostly in the training groups. By
observing the descriptive data, the Training+Physical
activity group seems to have had more difficulties
with their balance performance and physical func-
tion in comparison with the other groups. This may
be one explanation for the dropout rate in the
374 Clinical Rehabilitation 29(4)
Training+Physical activity group. The analyses of
the dropouts revealed no differences regarding
causes to drop out.
Analysing the two training groups regarding
physical activity level and performed activity, no
differences were found between groups; in other
words, the addition of physical activity in one of
the groups did not lead to any further benefits, as
all participants in the training groups reached the
recommended level for added physical activity.
It would be of interest, and a complement to
these results, to know the participants’ perceptions
of the fear of falling, physical function, and bal-
ance performance after participation in this balance
training program. However, this would require a
qualitative research designed study. It would also
be of interest to know the long-term effects of this
intervention, and if the improvements we have
seen in the short-term persist over time.
For further analysis and evaluation of this bal-
ance training program, the authors will take the
group constellation into consideration, since there
were no differences in the performed intervention
between the two training groups. Having three
groups may have led to a lack of power. Therefore,
the authors of the present study suggest that further
analysis of this balance training program in this
randomized controlled trial should be performed
on two groups; in other words, on one intervention
group (were the two training groups are merged)
and one Control group.
Clinical messages
•• Balance training with dual- and multi-
task successfully improves fall-related
self-efficacy, gait speed, balance perfor-
mance and physical function in older
adults with osteoporosis.
•• It is unclear if added physical activity
gives additional benefits to the balance
training program.
Acknowledgements
The authors would like to thank all the physiotherapists,
test leaders, and participants in the study. A special thank
Downloaded from cre.sagepub.com at LSU Libraries on November 13, 2016
you goes to Lisbet Broman at Karolinska Institute for
guidance and help with data collection.
Conflict of interest
The authors declare that there is no conflict of interest.
Funding
The study was supported by grants through the Regional
Agreement on Medical Training and Clinical Research
between Stockholm County Council and Karolinska
Institutet (ALF), as well as from the Swedish Research
Council (grant numbers 521-2010-2483 and 521-2013-
2525) and the Health Care Sciences Postgraduate School
at Karolinska Institutet.
References
1. Scheffer AC, Schuurmans MJ, van Dijk N, van der Hooft
T and de Rooij SE. Fear of falling: measurement strategy,
prevalence, risk factors and consequences among older
persons. Age and ageing 2008; 37: 19–24.
2. Delbaere K, Crombez G, Vanderstraeten G, Willems T
and Cambier D. Fear-related avoidance of activities, falls
and physical frailty. A prospective community-based
cohort study. Age and ageing 2004; 33: 368–373.
3. Chamberlin ME, Fulwider BD, Sanders SL and Medeiros
JM. Does fear of falling influence spatial and temporal
gait parameters in elderly persons beyond changes asso-
ciated with normal aging? The journals of gerontology
Series A, Biological sciences and medical sciences 2005;
60: 1163–1167.
4. Reelick MF, van Iersel MB, Kessels RP and Rikkert MG.
The influence of fear of falling on gait and balance in
older people. Age and ageing 2009; 38: 435–440.
5. Arfken CL, Lach HW, Birge SJ and Miller JP. The prevalence
and correlates of fear of falling in elderly persons living in
the community. Am J Public Health 1994; 84: 565–570.
6. Cress ME, Schechtman KB, Mulrow CD, Fiatarone MA,
Gerety MB and Buchner DM. Relationship between
physical performance and self-perceived physical func-
tion. Journal of the American Geriatrics Society 1995; 43:
93–101.
7. Studenski S, Perera S, Patel K, et al. Gait speed and sur-
vival in older adults. JAMA : the journal of the American
Medical Association 2011; 305: 50–58.
8. Cesari M, Kritchevsky SB, Penninx BW, et al. Prognostic
value of usual gait speed in well-functioning older peo-
ple–results from the Health, Aging and Body Composition
Study. Journal of the American Geriatrics Society 2005;
53: 1675–1680.
9. Segev-Jacubovski O, Herman T, Yogev-Seligmann G,
Mirelman A, Giladi N and Hausdorff JM. The interplay
between gait, falls and cognition: can cognitive therapy
reduce fall risk? Expert review of neurotherapeutics 2011;
11: 1057–1075.
Halvarsson et al. 375
10. Alexander NB and Hausdorff JM. Guest editorial: linking
thinking, walking, and falling. The journals of gerontol-
ogy Series A, Biological sciences and medical sciences
2008; 63: 1325–1328.
11. Verghese J, Kuslansky G, Holtzer R, et al. Walking while
talking: effect of task prioritization in the elderly. Archives
of physical medicine and rehabilitation 2007; 88: 50–53.
12. Lynn SG, Sinaki M and Westerlind KC. Balance charac-
teristics of persons with osteoporosis. Archives of physical
medicine and rehabilitation 1997; 78: 273–277.
13. Campbell AJ, Borrie MJ and Spears GF. Risk factors for
falls in a community-based prospective study of people 70
years and older. Journal of gerontology 1989; 44: M112–
M117.
14. Kanis JA. Diagnosis of osteoporosis and assessment of
fracture risk. Lancet 2002; 359: 1929–1936.
15. Howe TE, Rochester L, Jackson A, Banks PM and Blair
VA. Exercise for improving balance in older people.
Cochrane Database Syst Rev 2007: CD004963.
16. Howe TE, Rochester L, Neil F, Skelton DA and Ballinger
C. Exercise for improving balance in older people.
Cochrane Database Syst Rev 2011: CD004963.
17. Halvarsson A, Oddsson L, Olsson E, Faren E, Pettersson
A and Stahle A. Effects of new, individually adjusted,
progressive balance group training for elderly people with
fear of falling and tend to fall: a randomized controlled
trial. Clinical rehabilitation 2011; 25: 1021–1031.
18. Halvarsson A, Franzen E, Faren E, Olsson E, Oddsson L
and Stahle A. Long-term effects of new progressive group
balance training for elderly people with increased risk of
falling - a randomized controlled trial. Clinical rehabilita-
tion 2013; 27: 450–458.
19. Folstein MF, Folstein SE and McHugh PR. “Mini-mental
state”. A practical method for grading the cognitive state
of patients for the clinician. J Psychiatr Res 1975; 12:
189–198.
20. Folstein MF and Luria R. Reliability, validity, and clinical
application of the Visual Analogue Mood Scale. Psychol
Med 1973; 3: 479–486.
21. Halvarsson A, Dohrn I-M and Ståhle A. Taking balance
training for older adults one step further; the rationale for
and a description of a proven balance training programme.
2014: in press.
22. Yardley L, Beyer N, Hauer K, Kempen G, Piot-Ziegler
C and Todd C. Development and initial validation of the
Falls Efficacy Scale-International (FES-I). Age and age-
ing 2005; 34: 614–619.
23. Nordell E, Andreasson M, Gall K and Thorngren KG.
Evaluating the Swedish version of the Falls Efficacy
Scale-International (FES-I). Advances in Physiotherapy
2009; 11: 81–87.
24. Halvarsson A, Franzen E and Stahle A. Assessing the rela-
tive and absolute reliability of the Falls Efficacy Scale-
Downloaded from cre.sagepub.com at LSU Libraries on November 13, 2016
International questionnaire in elderly individuals with
increased fall risk and the questionnaire’s convergent
validity in elderly women with osteoporosis. Osteoporosis
international : a journal established as result of coopera-
tion between the European Foundation for Osteoporosis
and the National Osteoporosis Foundation of the USA
2013; 24: 1853–1858.
25. Kempen GI, Todd CJ, Van Haastregt JC, et al. Cross-
cultural validation of the Falls Efficacy Scale International
(FES-I) in older people: results from Germany, the
Netherlands and the UK were satisfactory. Disability and
rehabilitation 2007; 29: 155–162.
26. Menz HB, Latt MD, Tiedemann A, Mun San Kwan M
and Lord SR. Reliability of the GAITRite walkway sys-
tem for the quantification of temporo-spatial parameters
of gait in young and older people. Gait & posture 2004;
20: 20–25.
27. Webster KE, Wittwer JE and Feller JA. Validity of the
GAITRite walkway system for the measurement of aver-
aged and individual step parameters of gait. Gait & pos-
ture 2005; 22: 317–321.
28. Heitmann DK, Gossman MR, Shaddeau SA and Jackson
JR. Balance performance and step width in noninstitu-
tionalized, elderly, female fallers and nonfallers. Physical
therapy 1989; 69: 923–931.
29. Jarnlo GB and Nordell E. Reliability of the modified
figure of eight- a balance performance test for elderly
women. Physiotherapy Theory and Practice 2003; 19:
35–43.
30. Frandin K, Sonn U, Svantesson U and Grimby G.
Functional balance tests in 76-year-olds in relation to
performance, activities of daily living and platform tests.
Scand J Rehabil Med 1995; 27: 231–241.
31. Haley SM, Jette AM, Coster WJ, et al. Late Life Function
and Disability Instrument: II. Development and evalua-
tion of the function component. The journals of geron-
tology Series A, Biological sciences and medical sciences
2002; 57: M217–M222.
32. Roaldsen KS, Halvarsson A, Sarlija B, Franzen E and
Stahle A. Self-reported function and disability in late life
- cross-cultural adaptation and validation of the Swedish
version of the late-life function and disability instrument.
Disability and rehabilitation 2014; 36: 813–817.
33. Frändin K and Grimby G. Assessment of physicak activ-
ity, fitness and performance in 76-year-olds. Scand J Med
Sci Sports 1994; 4: 41–46.
34. Hardy SE, Perera S, Roumani YF, Chandler JM and
Studenski SA. Improvement in usual gait speed predicts
better survival in older adults. Journal of the American
Geriatrics Society 2007; 55: 1727–1734.
35. Kannus P, Uusi-Rasi K, Palvanen M and Parkkari J. Non-
pharmacological means to prevent fractures among older
adults. Ann Med 2005; 37: 303–310.