Journal of Gerontology: MEDICAL SCIENCES
2005, Vol. 60A, No. 12, 1553–1557
Copyright 2005 by The Gerontological Society of America
Time Requirement for Young and Elderly Women
to Move Into a Position for Breaking a Fall
With Outstretched Hands
Stephen N. Robinovitch, Sarah C. Normandin, Paula Stotz, and Jessica D. Maurer
Injury Prevention and Mobility Laboratory, School of Kinesiology, Simon Fraser University,
Burnaby, British Columbia, Canada.
Background. Risk for hip fracture during a fall is reduced by contacting the ground first with the outstretched hands.
However, it is unclear whether the time required for young and elderly individuals to move the hands into a protective
position exceeds that available during a typical fall.
Methods. We tested whether young (n¼30; aged 18–35 years) and elderly women (n¼30; aged 70–88 years) differed in
the time required to move their hands into a protective position for breaking a fall. Participants stood either facing or sideways
to shoulder-height targets (simulating forward and sideways falls, respectively), which they were instructed to contact as
quickly as possible after hearing an aural go cue. Total contact time was partitioned into reaction time and movement time.
Results. Young women contacted the targets faster than elderly women in both forward (530 6 60 vs 615 6 88 ms;
p , .001) and sideways trials (658 6 80 vs 799 6 145 ms; p , .001). This difference was due to faster movement times
for young participants. There was no difference between groups in reaction time.
Conclusions. Previous studies have shown that during actual falls from standing, wrist and pelvis contact occur at
680 6 116 and 715 6 160 ms, respectively. Comparing these values to our results suggests that the typical elderly
woman should be able to move her hands quickly enough to break a forward fall, but not a sideways fall.
F
ALLS are a major cause of injury in elderly persons and
the underlying reason for 90% of hip fractures (1). In
addition to bone strength, risk for hip fracture during a fall
depends on the mechanics of the fall, and in particular,
on the configuration of the body segments at impact. For
example, fracture risk is increased 30-fold if impact occurs
to the hip region, and decreased 3-fold if impact occurs to
one or both outstretched hands (2–4).
The task of breaking a fall with the outstretched hands can
be divided into three stages: detection of imbalance and
initiation of hand movement, movement of the hands into
a final landing position, and energy absorption during
impact. Studies have shown that during unexpected falls in
young individuals, the total time between fall initiation and
impact to the wrist and pelvis (total contact time) is 680 6
116 ms and 715 6 160 ms, respectively (5). Impairments in
the initiation of hand movement following imbalance
(reaction time) and transport of the hands into a protective
position (movement time) could prevent elderly individuals
from meeting these time demands.
Previous studies provide inconclusive evidence on
whether differences exist between young and elderly adults
in the time required to move the hands into a position for
breaking a fall. For example, DeGoede and colleagues (6)
found that older women were 20% slower than younger
women in the movement time required to lift their hands
from thigh level to arrest a pendulum swinging toward their
face. However, because reaction time and total contact time
were not reported, their results did not provide enough
information to predict whether the women tested could react
and move their hands quickly enough to break a real-life fall.
Other studies have presented opposing results. Kim and
Ashton-Miller (7) recently reported that, when falling forward onto an inclined surface, older men had larger angular
velocities at the elbow and shoulder, and faster hand contact
times than did younger men. This result was interpreted to
reflect increased urgency or fear, or reduced ability to control
and soften impact severity, among elderly men. The
researchers reported total contact times, but did not show
how reaction time and arm movement time contributed to the
total contact time. Furthermore, these previous studies
focused only on simulations of forward falls.
In the current study, we investigated whether young and
elderly women differed in the time required to move their
hands to contact targets that were either in front of them or
at their side, simulating the task of breaking a forward or
sideways fall with the outstretched hands. We also compared our measured contact times to the range of contact
times previously reported for falls from standing.
METHODS
Participants
A total of 30 young women (mean age ¼ 22 6 4 [standard
deviation, SD] years) and 30 elderly women (mean age¼77 6
5 years) participated in the study. Mean participant weight
was 55.6 6 9.3 kg (young) and 68.0 6 8.8 kg (elderly). Mean
participant height was 163 6 6 cm (young) and 160 6 7 cm
(elderly). We recruited young women through university
posting boards and elderly women through community
1553
1554
ROBINOVITCH ET AL.
newspapers. Prospective participants were screened initially
for eligibility with a telephone interview. Those women
who seemed to meet our inclusion criteria based on the telephone interview were further screened on physical examination by an experienced physiotherapist. Exclusion criteria
included: (a) regular exercise averaging once a week or more
for the past 3 months, (b) impairment of neuromuscular
function secondary to neurological disease (e.g., traumatic
brain injury, Parkinson’s disease, cerebral palsy, multiple
sclerosis, diabetic neuropathy), (c) amputation or other debilitating orthopedic conditions (e.g., joint replacements,
rheumatoid arthritis), (d) inability to raise the arms to shoulder
height, and (e) inability to follow simple instructions in
English. Each participant provided written informed consent,
and the experimental protocol was approved by the Research
Ethics Board of Simon Fraser University.
Figure 1. Side view of experimental setup for (A) forward trials and (B)
sideways trials. In forward trials, the participant stood facing shoulder-height
targets. In sideways trials, the participant stood sideways with her dominant arm
closest to the targets. The target distance, d, was adjusted so that impact occurred
with the elbows flexed at 458. The participant was instructed to contact the
targets as ‘‘quickly as possible’’ with both hands after hearing an aural go cue.
C, We identified the contact time by the sudden increase in compressive force
applied to a force plate mounted behind the dominant-hand target. This example
is from a forward trial with an elderly participant. The aural go cue was provided
at t ¼ 0, and the vertical line indicates the contact time at 617 ms.
Protocol
For all participants, we calculated maximum isometric
shoulder flexor and shoulder abductor torques (s). A
handheld dynamometer (Nicholas Manual Muscle Tester
[MMT] 01160; Lafayette Instrument Co., Lafayette, IN) was
used to record the peak force (F) generated by the participant
during an isometric 5-second resistance. The distance (d)
between the dynamometer (point of force application) and
the participant’s shoulder joint center was measured using
a standard measuring tape. We calculated shoulder torques
using the equation s ¼ Fd. Each participant’s shoulder
torques were normalized to body weight (N) 3 body height
(m), and then averaged over three trials.
To simulate breaking a fall with the outstretched hands,
we instructed participants to contact a pair of shoulderheight targets (one for each hand) that were mounted on
a standing frame (Figure 1). We conducted two sets of 10
trials in pseudorandom order, with a pseudorandom time
interval (approximately 10–30 seconds) between successive
trials. In the set of forward trials, the participant stood facing
the targets. In sideways trials, the participant stood with the
targets at her side such that her dominant hand was closest
to the targets. The participant’s distance from the targets was
adjusted so that her elbows would be flexed 458 at hand
impact. Each target was padded with 20 cm-thick foam to
minimize the resulting contact forces and reduce the
participant’s hesitation to strike the targets. The participant
began each trial with her hands relaxed at her sides and her
feet in a stationary position. We instructed her to contact
the targets as ‘‘quickly as possible’’ with the palms of both
hands, without moving her feet, after hearing an aural go
cue (a beep of 200 Hz frequency and 350 ms duration).
We used two force plates (MU2535; Bertec Corp.,
Columbus, OH), one mounted behind each target, to measure contact forces for the right and left hands at 960 Hz.
Using a 60 Hz, seven-camera motion measurement system
(ProReflex; Qualisys, Inc., Glastonbury, CT), we obtained
the 3-dimensional positions of 12 retroreflective skin surface
markers placed bilaterally at the acromion process, distal
third metacarpal, anterior superior iliac spine (ASIS), greater
trochanter, lateral malleolus, and fifth metatarsal head. The
aural go cue, video data, and force plate data were triggered
TIME REQUIREMENT FOR BREAKING A FALL
1555
Table 1. Outcome Variables From Hand Contact Trials in Young and Elderly Women
Forward Trials
Contact time, msz
Reaction time, msz
Movement time, msz
Sideways Trials
n Y, n E *
Young
Elderly
p Valuey
Young
Elderly
p Valuey
30, 30
30, 30
30, 30
530 6 60
212 6 35
318 6 40
615 6 88
231 6 42
384 6 75
,.001
.065
,.001
658 6 80
221 6 35
437 6 62
799 6 145
252 6 60
547 6 128
,.001
.017
,.001
Note: Data are means 6 standard deviation.
*nY is the sample size of young participants; nE is the sample size of elderly participants.
y
Significance of difference between young and elderly columns, based on independent samples t test. Values less than or equal to .005 are considered statistically
significant.
z
Data are for the dominant hand.
and acquired simultaneously using computer-based data
acquisition software (Qtrac; Qualisys).
Data Analysis
For each trial, we calculated reaction time as the time
interval between the aural go cue and the onset of hand
movement, and contact time as the time interval between the
go cue and contact of the dominant-hand target. Movement
time was determined as the difference between contact time
and reaction time (i.e., the interval between the onset of
hand movement and contact of the dominant hand. The
onset of hand movement was defined as the instant when
the hand surface marker began moving in a steady upward
motion, as determined from visual inspection (by a single
blinded investigator) of data acquired by the motion
measurement system. Contact time was identified by a sharp
increase in the normal force applied to the force plate
(Figure 1C), again from visual inspection of force plate data.
Statistics
Statistical analysis was performed using a repeatedmeasures analysis of variance, with movement direction
(forward vs sideways) as the repeating factor and age
category (young vs elderly) as the grouping factor. Dependent variables were reaction time, movement time, and total
contact time. Where the analysis of variance revealed
a significant effect of movement direction or age, post hoc
t tests were conducted to identify the source of these differences. Pearson correlations were used to examine associations between continuous variables. Statistical tests were
run using SPSS statistical analysis software (SPSS, Inc.,
Chicago, IL). After applying a Bonferroni correction for
the set of comparisons, p values less than or equal to .005
were considered statistically significant.
RESULTS
Effects of Age Differences and Movement
Direction on Total Contact Time
In both sets of trials, young participants were able to
contact the hand targets faster than were elderly participants
(Table 1). Mean contact time was 16% longer for elderly
participants than for young participants in forward trials
(615 6 88 vs 530 6 60 ms, mean difference ¼ 85 ms, 95%
CI, 46–124 ms; t ¼ 4.4, df ¼ 58, p , .001) and 21% longer
for elderly participants in sideways trials (799 6 145 vs
658 6 80 ms, mean difference ¼ 141 ms, 95% CI, 80–201 ms;
t ¼ 4.7, df ¼ 58, p , .001).
All participants were able to contact the hand targets faster
in forward trials than in sideways trials. For all participants
combined, mean contact time was 27% longer in sideways
trials than in forward trials (728 6 136 vs 572 6 86 ms,
mean difference ¼ 156 ms, 95% CI, 139–173 ms, t ¼ 18.0,
df ¼ 59, p , .001). For young participants, mean contact
time was 24% longer in sideways trials compared to forward
trials (658 6 80 vs 530 6 60 ms, mean difference ¼ 128 ms,
95% CI, 116–140 ms, t ¼ 22.3, df ¼ 29, p , .001); for elderly
participants, mean contact time was 30% longer in sideways
trials (799 6 145 vs 615 6 88 ms, mean difference ¼ 184 ms,
95% CI, 154–214 ms, t ¼ 12.4, df ¼ 29, p , .001).
Effect of Age Differences on Reaction
and Movement Times
We observed age-related slowing in movement time but
not in reaction time (Table 1). Although reaction times were
not significantly different between young and elderly participants, reaction time in the elderly women tended to be
slightly slower than in the young women, by 9% in forward
trials (231 6 42 vs 212 6 35 ms, mean difference ¼ 19 ms,
95% CI, 1–39 ms, t ¼ 1.9, df ¼ 58, p ¼ .065) and 14% in
sideways trials (252 6 60 vs 221 6 35 ms, mean difference ¼
31 ms, 95% CI, 6–57 ms, t ¼ 2.5, df ¼ 58, p ¼ .017); however,
mean movement time for elderly participants was 21%
slower than for young participants in forward trials (384 6
75 vs 318 6 40 ms, mean difference ¼ 66 ms, 95% CI, 35–
97 ms, t¼4.3, df¼ 58, p , .001) and 25% slower in sideways
trials (547 6 128 vs 437 6 62 ms, mean difference ¼ 109 ms,
95% CI, 57–161 ms, t ¼ 4.2, df ¼ 58, p , .001).
Total Contact Time, Reaction Time,
and Movement Time Relationships
Differences in contact time related more to differences
in movement time than reaction time for both forward
and sideways trials (Figure 2). For forward trials, contact
time increased as both reaction time and movement
time increased (r ¼ 0.631, p , .001; r ¼ 0.894, p , .001,
respectively). Similarly, for sideways trials, contact time
increased as both reaction time and movement time
increased (r ¼ 0.581, p , .001; r ¼ 0.931, p , .001,
respectively). Although there was not a significant association between reaction time and movement time in either
forward ( p ¼ .097) or sideways ( p ¼ .062) trials, it was
common to see slowing in both variables among the slowest
1556
ROBINOVITCH ET AL.
Figure 2. Measured values of contact time, reaction time, and movement time
in (A) forward fall and (B) sideways fall simulations. In each graph, two data
points are displayed for each participant: one showing contact time plotted
against reaction time (circles), and the other showing contact time plotted against
movement time (triangles). Data from young women are shown as filled
symbols; data from elderly women are shown as unfilled symbols. The vertical
dashed lines drawn at 680 ms and 715 ms indicate previous measures of the
average ‘‘threshold’’ time for contact to the wrist and pelvis, respectively, during
actual unexpected falls (6). Note that, on average, contact times were longer in
elderly women than in young women. In forward fall simulations, none of the
young participants and only 4 of the 30 elderly participants (13%) had contact
times that were longer than the threshold time for pelvis impact during an actual
fall. In sideways fall simulations, 7 of the 30 young participants (23%) and 23 of
the 30 elderly participants (77%) had contact times that were longer than the
threshold time for pelvis impact during an actual fall.
elderly participants (Figure 3). Furthermore, there was
a relationship between increased forward reaction time and
increased sideways reaction time (r ¼ 0.748, p , .001), and
between increased forward movement time and increased
sideways movement time (r ¼ 0.900, p , .001).
Effect of Shoulder Strength on Movement Time
We found a significant, but weak relationship between
shoulder strength and movement time. Normalized shoulder
flexor torque increased as movement time decreased for both
forward (r ¼0.487, p , .001) and sideways directions (r ¼
0.466, p , .001). Normalized shoulder abductor torque also
increased as movement time decreased for forward (r ¼
0.488, p , .001) and sideways directions (r ¼0.448, p ,
Figure 3. Traces of the position in the x-direction of the dominant-hand marker,
comparing an elderly participant with fast contact times (participant E1, shown
with solid lines) versus an elderly participant with slow contact times (participant
E2; shown with dashed lines) in forward fall (A) and sideways fall (B) simulations.
Slower contact times tended to be associated with delays in movement initiation
(i.e., slower reaction times [RT]), and with slower hand velocities and longer
deceleration periods before contact (leading to slower movement times [MT]). In
the forward condition, these two participants’ reaction times differed by 106 ms
(RTE1 ¼ 205 ms; RTE2 ¼ 311 ms), and their movement times differed by 324 ms
(MTE1 ¼ 236 ms; MTE2 ¼ 560 ms). In the sideways condition, these participants’
reaction times differed by 151 ms (RTE1 ¼ 167 ms; RTE2 ¼ 318 ms), and their
movement times differed by 425 ms (MTE1 ¼ 367 ms; MTE2 ¼ 792 ms).
.001). However, after adjusting for age, these correlations
did not reach statistical significance (r ¼ .24, p ¼ .06 and
r ¼ .22, p ¼ .094 for flexor torque and movement time in
the forward and sideways directions, respectively; r ¼.23,
p ¼ .08 and r ¼ .18, p ¼ .18 for abductor torque and
movement time in the forward and sideways directions).
DISCUSSION
We found that elderly women are slower than young
women in moving their hands into a protective position for
breaking a forward or sideways fall. Furthermore, this was
due primarily to slower movement times, as opposed to
differences in reaction times. To place our results in context,
consider that previous studies have shown that, during actual
falls from standing, the interval between fall initiation and
contact to the wrist averages 680 6 116 ms, and the interval
between fall initiation and impact to the pelvis averages
715 6 160 ms (5). In our forward trials, 100% of young
participants and 83% of elderly participants had total contact
TIME REQUIREMENT FOR BREAKING A FALL
times less than 680 ms. During our sideways trials, 63% of
young and only 13% of elderly women had total contact
times less than 680 ms. These results suggest that the typical
young woman can meet the time requirements for breaking
either a forward or sideways fall (as one would expect from
common experience). They also suggest that the typical
elderly woman should be able to react and move her hands
quickly enough to break a forward fall, but not a sideways fall.
Our results are in agreement with previous reports of agerelated declines in reaction time and movement time (8–10).
Furthermore, although the comparison is complicated by
the different methods used in the two studies, our movement times are similar to those reported by DeGoede and
colleagues (6) for simulated falls on the outstretched hands.
In DeGoede’s study, participants were asked to contact
shoulder-height targets as quickly as possible with the hands,
from a seated position. Trials were conducted under three
conditions: no-threat (participants contacted stationary targets), low-threat (participants contacted targets on a swinging
pendulum, which was released just as hand movement was
initiated), and high-threat (participants contacted targets on
a swinging pendulum, and were instructed to wait as long as
possible before initiating hand movement). Movement time
for the older women was 20% slower than that for the younger
women in the no- and low-threat conditions (335 6 39 [SD]
vs 281 6 30 ms for the no-threat condition; 338 6 39 vs
277 6 49 ms for the low-threat condition). Furthermore,
movement time in women did not differ significantly between
the no-threat, low-threat, and high-threat trials. This finding
supports the notion that performance in a reduced threat
situation (such as our experiment) may be a reasonable
indicator of protective abilities during a real-life fall.
There are important limitations to this study. Due to safety
concerns, our experiments did not involve actual falls, and
there are several reasons why response times may be different
in an actual fall. For example, the fear and urgency associated
with falling may affect response times in a complex manner.
Furthermore, falls produce a range of impact configurations
(5), and we cannot rule out the possibility of different trends
had we used a different impact configuration. Moreover,
detecting the onset of imbalance is the first stage in initiating
a protective response during an actual fall, and older adults
take longer than do younger adults to respond to a postural
perturbation. However, these delays are relatively small when
compared to age-related changes in movement time (11). For
example, in examining single-step recoveries from a forward
fall initiated by tether release, Wojcik and colleagues (12)
found that, although total step times averaged 60 ms faster in
young than in elderly women, the difference in average
reaction times was only 13 ms—similar to the differences
observed in our study. A further limitation of our study is
that, despite our efforts to screen prospective participants
for obvious neurological disease or musculoskeletal impairment, our findings may not be related to the effect of
age per se, but rather to subclinical neurological or musculoskeletal dysfunction. Finally, our between-group differences in reaction time (which averaged 19 ms [or 9%] in
forward trials, and 31 ms [or 14%] in sideways trials) bordered
on statistical significance. We expect that a larger sample
size would have revealed a statically significant difference.
1557
Conclusion
Our study provides important new information on agerelated differences in the ability to move the hands into
a position for breaking a fall—a protective response that has
been shown to substantially reduce hip fracture risk. We
found that, in simulated forward falls, both young and
elderly groups had average contact times that were faster
that the time typically available during a fall. However, in
simulated sideways falls (which are most likely to produce
hip fracture), most elderly participants were unable to
achieve contact times within the time available during a fall.
Further research is required to determine whether exercise
programs can improve the efficacy of upper extremity fallprotective responses.
ACKNOWLEDGMENTS
This study was supported by operating grants from the Centers for
Disease Control and Prevention (R49CCR019355), the National Institutes
of Health (R01AR46890), and the Canadian Institutes of Health Research
(MOP 64414). Stephen N. Robinovitch was supported by a New
Investigator Award from the Canadian Institutes of Health Research and
a Scholar Award from the Michael Smith Foundation for Health Research.
Address correspondence to Stephen N. Robinovitch, PhD, School of
Kinesiology, Simon Fraser University, Burnaby, BC, V5A 1S6, Canada.
E-mail: [email protected]
REFERENCES
1. Melton LJ, Chao EYS, Lane J. Biomechanical aspects of fractures.
In: Riggs BL, Melton LJ, eds. Osteoporosis: Etiology, Diagnosis, and
Management. New York: Raven Press; 1988:111–131.
2. Nevitt MC, Cummings SR. Type of fall and risk of hip and wrist
fractures: the study of osteoporotic fractures. The Study of Osteoporotic
Fractures Research Group. J Am Geriatr Soc. 1993;41:1226–1234.
3. Schwartz AV, Kelsey JL, Sidney S, Grisso JA. Characteristics of falls
and risk of hip fracture in elderly men. Osteoporos Int. 1998;8:
240–246.
4. Parkkari J, Kannus P, Palvanen M, et al. Majority of hip fractures occur
as a result of a fall and impact on the greater trochanter of the femur:
a prospective controlled hip fracture study with 206 consecutive
patients. Calcif Tissue Int. 1999;65:183–187.
5. Hsiao ET, Robinovitch SN. Common protective movements govern
unexpected falls from standing height. J Biomech. 1998;31:1–9.
6. DeGoede KM, Ashton-Miller JA, Liao JM, Alexander NB. How
quickly can healthy adults move their hands to intercept an approaching
object? Age and gender effects. J Gerontol Med Sci. 2001;56A:
M584–M588.
7. Kim KJ, Ashton-Miller JA. Biomechanics of fall arrest using the upper
extremity: age differences. Clin Biomech (Bristol, Avon). 2003;18:
311–318.
8. Ketcham CJ, Stelmach GE. Age-related declines in motor control.
In: Birren JE, Schaie KW, eds. Handbook of the Psychology of Aging.
San Diego, CA: Academic Press; 2001:313–348.
9. Stelmach GE, Amrhein PC, Goggin NL. Age differences in bimanual
coordination. J Gerontol Psychol Sci. 1988;43B:P18–P23.
10. Yan JH, Thomas JR, Stelmach GE. Aging and rapid aiming arm
movement control. Exp Aging Res. 1998;24:155–168.
11. Thelen DG, Wojcik LA, Schultz AB, Ashton-Miller JA, Alexander NB.
Age differences in using a rapid step to regain balance during a forward
fall. J Gerontol Med Sci. 1997;52A:M8–M13.
12. Wojcik LA, Thelen DG, Schultz AB, Ashton-Miller JA, Alexander NB.
Age and gender differences in single-step recovery from a forward fall.
J Gerontol Med Sci. 1999;54A:M44–M50.
Received August 9, 2004
Accepted March 3, 2005
Decision Editor: John E. Morley, MB, BCh