Professional Documents
Culture Documents
Review
A R T I C L E I N F O A B S T R A C T
Article history: This systematic review and critical evaluation of the literature investigates whether advanced age
Received 15 September 2008 compromises obstacle crossing for unconstrained and time-constrained conditions. Eight electronic
Received in revised form 25 May 2009 databases were searched for articles with terms relating to obstacle crossing during walking in their title,
Accepted 30 May 2009
abstract or keywords. 15 articles were reviewed from an initial yield of 727 articles. The methodological
quality of each article was critiqued and data extracted by two reviewers. Young and older adults were
Keywords: shown to contact obstacles infrequently when adequate time was available to adapt foot placement in
Ageing
relation to the obstacle. When less time was available to adjust the foot trajectory, older adults contacted
Obstacle crossing
Accidental falls
obstacles more often than younger people. Older adults adopted a slower, more conservative obstacle
Gait crossing strategy. They demonstrated greater hip flexion during the swing phase of gait for the lead and
trail limbs as well as greater hip flexion, hip adduction and ankle dorsiflexion during the stance phase for
the lead and trail limbs. There was also evidence of reduced internal moments across the hip and ankle
during key events in the obstacle crossing gait cycle in older adults. Despite using a more conservative
obstacle crossing strategy, older adults are at greater risk of contacting obstacles for time-constrained
conditions.
ß 2009 Elsevier B.V. All rights reserved.
2. Methods
0966-6362/$ – see front matter ß 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.gaitpost.2009.05.022
B. Galna et al. / Gait & Posture 30 (2009) 270–275 271
databases included CINAHL (from 1982), Cochrane library (from 1800), Embase 2.4. Reliability of the reviewing process
(from 1980), Inspec (from 1898), Medline – ISI (from 1950), PsycInfo (from 1806),
PubMed (from 1950) and Web of Science – ISI (from 1900). Targeted searching of To ensure the reliability of the quality assessment and data extraction tools, each
frequently cited journals, authors and article reference lists ensured that all tool was piloted with a test article by two reviewers (BG and AP). A set of guidelines
relevant articles were located. and examples for completing article reviews were also provided to both reviewers
To be deemed relevant for review, an article had to contain words referring to to improve inter-reviewer reliability. A third reviewer (MM) independently
walking, an obstacle and crossing in the title, abstract and keywords. As differences identified discrepancies between the completed quality appraisal and data
in terminology and writing style exist between researchers, a list of synonyms was extraction forms of the two reviewers (BG and AP). The consensus method
used for each key term to identify all relevant articles. For example, the words ‘gait’, suggested in the Cochrane Systematic Review Handbook [9,p. 103] was used by the
‘locomotion’ and ‘ambulation’ were searched in conjunction with the word ‘walk’. two reviewers (BG and AP) to resolve discrepancies.
Key terms within the search strategy were also matched and exploded using
medical subject headings (MeSH) in databases such as Medline. Articles that are 3. Results
related to animal studies, robotics, pulmonary obstruction, children, osteoarthritis
and amputees were excluded using another set of key terms. As an example, the
following search strategy was used with the Web of Science (‘‘*’’ indicates a 3.1. Yield
wildcard and ‘‘TS’’ indicates a search of titles, abstracts and keywords):
The electronic database search resulted in an initial yield of 727
TS: (walk* or gait or locomot*or ambulat*) articles (Fig. 1), including 232 articles relating to obstacle crossing
AND TS: (obstacle* or obstruct*) in humans. Two original articles were sourced by hand searching
AND TS: (step* or cross* or negotiate* or adapt* or avoid*)
relevant journal titles, reference lists and frequently cited authors.
NOT TS: (child* or pulmonary* or obstructive* or osteoarthritis or amput* or cat or
animal or robot*) The final yield included 15 articles related to obstacle crossing in
older adults and young adults.
The initial yield was obtained by combining all original articles from electronic 3.2. Quality assessment
databases and targeted searches. Articles not relating to obstacle crossing were
excluded during an initial screening of titles and abstracts by the reviewer (BG). The
remaining abstracts were checked against the inclusion and exclusion criteria. The Table 1 summarises the quality assessment for each article. The
full text of an article was examined if there was insufficient information contained articles stated the aims sufficiently, gave an appropriate descrip-
in the abstract to decide whether to include the article. There was one case where tion of the participants, stated the inclusion and exclusion criteria,
there was insufficient detail in the full text article and the authors were contacted detailed the key outcomes clearly, employed a suitable methodol-
for further details.
Agreement on article inclusion was reached by two reviewers (BG and MM).
ogy for answering their research questions and discussed the
Articles were included in the systematic review if they investigated a sample of results of their study appropriately. Details provided were
healthy older adults and younger adults walking over an obstacle. Only English adequate to replicate the research although methods for partici-
language articles were considered for analysis. It was beyond the scope of the pant recruitment and sampling were not reported often. The
review to investigate other forms of adaptive gait such as turning, slips avoidance,
reliability of key outcome measures was only addressed in two of
running or stair climbing. If a study investigated both obstacle crossing and another
form of locomotion, only details pertaining to obstacle crossing were considered. the 15 studies and the internal validity of the key outcome
Treadmill studies were excluded on the basis that gait adaptations for an obstacle measures was only reported once. The clinical implications of the
on a treadmill do not allow for natural spatiotemporal modifications to gait in research was also lacking in detail or clarity in seven studies and
relation to the obstacle. Tripping studies were excluded because simulated trips in not addressed in seven studies. Most investigations controlled for
previous studies were induced by the researchers rather than occurring naturally
[6].
height, age and gender but typically did not control for limb
asymmetries and strength differences between groups. No article
2.2. Quality had particularly poor scientific quality so data concerning the
results were extracted from all articles.
The quality of each article was assessed independently by two researchers (BG
and AP) on the premise that ‘‘. . .it may limit bias, minimise errors and improve
reliability of findings. . .’’ [9,p. 93]. The quality of a study was defined as its capacity
3.3. Sample characteristics
to avoid potential bias and to generate results that were able to be generalised. This
definition encompasses both internal and external validity as dimensions of quality. The mean age for healthy older adults in each of the studies
Internal validity refers to the accuracy of the measurements. External validity refers ranged from 68 to 76 years in all except one study which recruited
to whether the results can be generalised to the population of interest [10,p 172].
participants between the ages of 84 and 93 years [13]. The mean
It was predicted that many of the articles would have an ex-post facto design
[10,p 154]. This type of design compares different levels of an independent variable age of the control groups of young adults ranged from 22 to 28
that the researcher cannot actively manipulate, such as age. No valid and reliable years. Both male and female participants were recruited into each
tools had been developed to assess the quality of such studies. A quality appraisal investigation, except for one study that included only males [14]
tool was developed by the authors for this systematic review (Table 1). An accepted and two that did not report the sex of participants [15,16].
generic tool for appraising the quality of a range of quantitative studies designed by
Law [11 p.305-308] was used in the development of the current tool. Additional
questions regarding whether potential confounding factors, such as strength and 3.4. Obstacle contacts and foot clearance
walking speed, were controlled for when comparing age groups were included.
Questions focussed on internal validity, external validity and the ability for the The number of obstacle contacts, time constraint and foot
methods to be replicated. A scoring system was developed to quantify the quality of
clearance results are summarised in Table 2. 10 studies reported
each study and to assess the methodological strengths and weaknesses of the
reviewed articles. A similar quality appraisal scoring system has previously been whether participants contacted the obstacle. Of these, four
published [12]. Each question on the quality appraisal tool was scored out of one. A included time constrained conditions, where participants had to
score of one indicated high quality research and zero indicated lower quality. avoid an obstacle that appeared suddenly during the walking trial
[8,17–19]. Obstacle contacts were more frequent in older adults
2.3. Data extraction compared to younger adults under time-constrained conditions in
Data concerning the sample and key results of each article were obtained each of these studies. Obstacle contacts were reported in six
independently by two reviewers (BG and AP) using a data extraction form. Data studies that did not constrain the amount of time participants had
extracted from each article included the description of the sample and the key to adapt to an obstacle. No obstacle contacts occurred for either age
results concerning obstacle crossing success, foot clearance, spatiotemporal, and group in four of these studies. In the two investigations that
kinematic and kinetic measures of obstacle crossing in older adults. Obstacle
crossing success is reported as the ability of a participant to avoid any part of their
obstacle contacts occurred, Chen et al. [20] reported four older
foot contacting an obstacle when attempting to walk over it, including stepping on adults hit the obstacle and both an older and young adult contacted
an obstacle. the second of two obstacles in a study by Lowrey et al. [21].
272
Table 1
Summary of quality appraisal for individual studies.
Question Scoring criteria Brown Brown Chapman and Chen Chen Chen Chen Di Fabio Draganich Hahn and Hahn Lowrey Lu McFayden McKenzie Average
et al. et al. Hollands et al. et al. et al. et al. et al. and Kuo Chou et al. et al. et al. and Prince and Brown
[23] [17] [15] [20] [18] [8] [19] [13] [22] [24] [25] [21] [16] [14] [26]
3. Recruitment and 1 – Yes; 0.5 – yes, lacking 0 0 0.5 0.5 0.5 0.5 0.5 0.5 0 0.5 0.5 1 0 0.5 0 0.37
sampling methods detail or clarity; 0 – No
described
6. Key outcome variables 1 – Yes; 0.5 – only some defined; 1 1 1 1 1 0.5 1 0.5 0.5 1 0.5 1 1 1 1 0.87
clearly described 0.5 – yes, lacking detail or clarity;
0 – no
14. Clinical implications 1 – Yes; 0.5 – yes, lacking 0.5 0 0.5 0.5 0 0 0 0.5 0 0.5 1 0.5 0.5 0 0 0.30
stated detail or clarity; 0 - No
B. Galna et al. / Gait & Posture 30 (2009) 270–275 273
Fig. 1. Flowchart of articles included for review. The number under the text indicates the number of original articles (i.e. not duplicates) at each stage of the search.
Several obstacle crossing studies that were reviewed did not future obstacle crossing studies report the frequency and nature of
state whether any obstacle contact events occurred during testing. any obstacle contacts that occur.
Given the infrequent nature of obstacle contact events that occur There were conflicting results for the effects of age on foot
when conducting obstacle crossing studies, it is recommended that clearance. Both Chen et al. [20] and Lowrey et al. [21] reported that
Table 2
Summary of obstacle crossing success, time constraint and foot clearance.
Author Obstacle contacts Time-constraint Lead limb clearance Trial limb clearance
lead limb toe clearance was unaffected by age. Lu et al. [16] internal hip extensor, hip abductor and knee extensor moments
however, reported an interaction of lead limb toe clearance with during the stance phase of both lead limb crossing as well as
age and obstacle height, whereby older adults increased their toe reduced ankle evertor moments for the mid stance phase of the
clearance with increasing obstacle height and younger adults did trail limb. Older adults also had reduced peak ankle power during
not. Older adults had a higher lead limb toe clearance than younger late stance and reduced eccentric hip abductor activity. The power
adults for higher obstacles. In contrast, another study demon- generated to raise the stance hip vertically at the end of stance
strated that lead limb toe clearance was lower in older adults [14]. phase for the trail limb was reduced in older adults [14].
There was consistently no effect of age on trail limb toe clearance in
three studies [14,16,22]. Given the disparate characteristics of
experimental methodology and measurement of obstacle contacts 4. Discussion
and foot clearance, a meta-analysis was not performed.
Obstacle contacts were more frequent in healthy older adults
3.5. Spatiotemporal adaptations during obstacle crossing than young adults when there was limited time to avoid the
obstacle [8,17–19]. Treadmill studies have also shown that older
The effect of ageing on spatiotemporal variables of obstacle adults are more likely to contact obstacles when walking under
crossing was not consistent. Of the 13 studies that reported time-constrained conditions [27]. Older adults adopted a short
spatiotemporal data, five reported no difference in spatiotemporal step strategy to avoid obstacles under time-constrained condi-
variables between young and older adults [17,22–25]. In the eight tions, which could increase the risk of stepping on an obstacle.
that reported differences, older adults walked slower during Using a short step strategy might also indicate an inability or
obstacle crossing [13,14,15,20,21,26], with smaller steps reluctance of older adults to walk with long steps [18].
[13,14,18,21]. Older adults landed closer to the obstacle with There was little evidence to show that older adults contact
their lead limb after crossing in three studies [14,21] however, only ground-based obstacles more often than younger adults when
a significant difference between older and younger females was ample time is given to adjust foot placement. In most of the
reported in the study by Chen et al. [20]. In contrast, Lu et al. [16] reviewed articles, older adults adopted a slower, more con-
reported that older adults placed their foot closer to the obstacle servative strategy than young people when stepping over
prior to crossing and landed further from the obstacle after obstacles when time was not constrained [13,14,15,20,21,26].
crossing. Older adults were more likely to take more steps to reach Adopting a conservative strategy with advancing age helps to
an obstacle [13] and use a short step strategy compared to younger explain why so few obstacle contacts occurred during uncon-
adults [18]. A short step strategy was defined as the reduction of strained locomotion. Walking slower could also provide older
step length in the step prior to crossing the obstacle [18]. adults with more time to adjust their foot placement in relation
to the obstacle in order to reduce the risk of tripping
3.6. Kinematics and kinetics [13,14,15,20,21,26].
In previous studies that reported the kinematics of obstacle
Six studies reported the kinematics of obstacle crossing. Brown crossing, older adults were shown to be more flexed and adducted
et al. [17] did not find any effects of ageing on kinematics. Chen at the hip during stance [14,16]. An analysis of joint powers
et al. [20] tested hip, knee and ankle range of motion during indicated that older adults demonstrated reduced eccentric
obstacle crossing and when standing, holding onto a rail. The contraction of the stance limb hip abductors and vertical hip
standing condition was used to measure maximum range of power (defined as the power used to raise the stance hip) at toe-off
motion (RoM). The authors found lower limb RoM during obstacle were both reduced in older adults [14]. Reduced hip abductor
crossing was less than the RoM when standing and holding onto a strength may lead to pelvic drop during stance, effectively
rail. They concluded that although older adults had a reduced step lowering the height of the crossing limb. Age-related changes in
length during obstacle crossing, it was unlikely that it was due to proprioception, vision and cognitive function are also potential
their reduced lower limb RoM, because they did not use the same factors explaining deficits in obstacle avoidance [5,6,25], yet
lower limb RoM during obstacle crossing as they did during the further investigation of obstacle crossing under time-constrained
standing condition. Lowrey et al. [21] found that sagittal plane conditions is required.
trunk RoM was greater in older adults in the step prior to crossing Despite including only articles that tested healthy older adults,
than younger adults. it is unlikely that all of the participants were all free from age-
Three articles reported age-related kinematic differences related conditions that influence the ability to cross obstacles. It is
during lead or trail limb clearance. When the lead limb toe was possible that underlying deterioration in visual, orthopaedic and
over the obstacle, older adults had greater lead hip flexion and less cognitive systems associated with ageing might contribute to a
ankle eversion, whilst their trail hip was more flexed and adducted, decline in obstacle crossing performance.
and the ankle in greater dorsiflexion [14,16]. When the trail toe
was above the obstacle, older adults had slower trail limb hip
angular velocities [22], greater lead limb hip flexion and adduction, 5. Conclusion
and greater lead ankle dorsiflexion [16]. Older adults also
demonstrated greater pelvic drop of the trail limb at trail limb Healthy older adults contact obstacles more often than younger
toe off and when the trail foot was over the obstacle, whereby the adults when limited time is available to respond. It is possible that
pelvis was closer to the ground on the side of the trail limb [14]. more frequent contacts with ground-based obstacles during
Only two articles documented the kinetics of obstacle crossing. walking may contribute to an increased risk of injurious falls in
Draganich and Kuo [22] reported that when older adults walked at older adults, although this requires empirical confirmation.
similar speeds to their younger counterparts, they had greater peak
trail limb hip abduction internal moments in early trail limb
stance. Peak trail limb hip abduction, knee abduction and ankle Conflicts of Interest
abduction moments were also greater in older adults in late trail
limb stance. When walking speed was not constrained, McFayden There are no conflicts of interest to disclose for any of the
and Prince [14] showed that older adults tended to have reduced authors.
B. Galna et al. / Gait & Posture 30 (2009) 270–275 275