Evidence based practice and nursing research

Age and Ageing 2015; 44: 205–212 doi: 10.1093/ageing/afu175

© The Author 2014. Published by Oxford University Press on behalf of the British Geriatrics Society. All rights reserved. For Permissions, please email: [email protected]

Published electronically 5 November 2014

Cognitive motor interference for preventing falls in older adults: a systematic review and meta-analysis of randomised controlled trials

XUEQIANG WANG1,2,†, YANLING PI3,†, PEIJIE CHEN1, YU LIU4, RU WANG4, CHETWYN CHAN5

1 Sport Medicine and Rehabilitation Center, Shanghai University of Sport, Shanghai, China 2 Department of Rehabilitation Medicine, Shanghai Shangti Orthopaedic Hospital, Shanghai, China 3 Department of Rehabilitation Medicine, Shanghai Punan Hospital, Shanghai, China 4 Key Laboratory of Exercise and Health Sciences of Ministry of Education, Shanghai University of Sport, Shanghai, China 5 Department of Rehabilitation Sciences, The Hong Kong Polytechnic University, Hong Kong, China

Address correspondence to: P. Chen. Tel: (+86) 02151253003; Fax: (+86) 02151253426. Email: [email protected]

Abstract

Objective: we conducted a systematic review to determine the effect of cognitive motor interference (CMI) for the prevention of falls in older adults. Methods: we searched studies through Medline, Embase, the Cochrane Library, Web of Science, CINAHL, PEDro and the China Biology Medicine disc. Only randomised controlled trials examining the effects of CMI for older people were included. The primary outcome measure was falls; the secondary outcome measures included gait, balance function and re- action time. Results: a total of 30 studies of 1,206 participants met the inclusion criteria, and 27 studies of 1,165 participants were used as data sources for the meta-analyses. The pooling revealed that CMI was superior to control group for fall rate [standard mean difference (SMD) (95% CI) = −3.03 (−4.33, −1.73), P < 0.0001], gait speed [SMD (95% CI) = 0.36 (0.07, 0.66), P = 0.01], step length [SMD (95% CI) = 0.48 (0.16, 0.80), P = 0.003], cadence [SMD (95% CI) = 0.19 (0.01, 0.36), P = 0.03], timed up and go test [SMD (95% CI) = −0.22 (−0.38, −0.06), P = 0.007], centre of pressure displacement [SMD (95% CI) = −0.32 (−1.06, 0.43), P = 0.04] and reaction time [SMD (95% CI) = −0.47 (−0.86, −0.08), P = 0.02]. Conclusion: the systematic review demonstrates that CMI is effective for preventing falls in older adults in the short term. However, there is, as yet, little evidence to support claims regarding long-term benefits. Hence, future studies should investi- gate the long-term effectiveness of CMI in terms of fall prevention in older adults.

Keywords: cognitive motor interference, falls, gait, balance, older people, systematic review

Introduction

Falls are a major cause of injury in older people (aged over 65) [1, 2]. More than one in three older people fall at least once a year, and 5–10% of older people with a history of falls will have a head injury, fracture or serious laceration [3, 4]. The severity of fall-related complications also increases with age [5]. Falls are the result of an interaction between physio- logical risk factors and environmental factors. For example, muscle weakness, impaired vision and slow reaction time are

common symptoms among individuals who are likely to fall [6, 7]. Moreover, older adults are more likely to fall when per- forming concurrent tasks, such as walking while performing other cognitive or motor tasks [8].

A cognitive motor task is, in effect, two tasks—a cogni- tive task and a motor task—that are carried out simultan- eously [9, 10, 11]. Cognitive motor interference (CMI) may lead to a decline in the performance of one or both tasks when these are performed simultaneously. Most daily activ- ities require the ability to maintain balance while performing various tasks simultaneously [10, 11]. Therefore, training for the performance of two tasks simultaneously may help to prevent falls [11]. Although a few studies [12, 13] have†X.W. and Y.P. contributed equally to this study.

205

suggested that CMI could be more effective than single-task exercise for improving balance function in older people, no consensus currently exists regarding the matter.

A systematic review [14] that included 28 articles has con- cluded that the effectiveness of CMI in improving physical functioning in older adults is limited. However, this review only focused on a description of the studies and on qualitative synthesis rather than on meta-analysis. Another systematic review [15] reported that the gait speed of healthy participants was slower under dual-task conditions than single-task ones. However, the systematic review was not conducted to deter- mine the effect of CMI comparing with single-task exercise or no intervention.

At present, it remains unclear whether CMI is more bene- ficial in terms of preventing falls in older people than single- task exercise or no intervention. In preventing falls among the elderly, designing the most effective exercise plan based on scientific evidence is crucial. Doing so ensures that valu- able staff time and resources are not wasted. Thus, we per- formed a systematic review and meta-analysis of all available randomised controlled trials (RCTs) to estimate the effective- ness of CMI for preventing falls in older adults compared with single-task exercise or no intervention.

Methods

Search strategy

Relevant studies were identified by electronically searching the following data sources (January 1974–August 2013): Medline, Embase, the Cochrane Library, Web of Science, CINAHL, PEDro and China Biology Medicine disc. The search was limited to RCTs but had no language restrictions. We also sup- plemented the electronic searches by manual searches. The full electronic search strategies for all databases are provided in Supplementary data, Appendix e-1 available in Age and Ageing online. We registered the Systematic review in PROSPERO (Registration number: CRD42012002606, http://www.crd. york.ac.uk/PROSPERO).

Inclusion criteria

• Types of studies: we included published articles with com- pleted RCTs. No restrictions were made regarding language or the date of the trial.

• Types of participants: we only selected trials with both male and female participants aged over 60. Any articles that included participants with any pathology affecting the balance function, such as Parkinson’s disease, stroke or serious arthritis, were excluded.

• Types of interventions: we only considered those trials that compared a treatment group, which performed CMI, and a control group, which performed single-task exercise (e.g. walking or strength and balance exercises) or no interven- tion. In a typical CMI, subjects needed to answer a series of simple addition/subtraction questions (e.g. 5 + 3 = 8, 100– 7 = 93) while they were performing a motor task [16].

• Types of outcome measures: these included the following: (i) gait variables such as gait speed, stride length, step length and cadence; (ii) balance function, gauged via the timed up and go test, activities-specific balance confidence scale, functional reach test, one leg test, chair stand test, berg balance scale, Tinetti test and centre of pressure displace- ment; (iii) measures related to falls, including the falls effi- cacy scale, the number of falls and the number of people who fall; and (iv) reaction time.

Selection of studies

Two authors independently used the same selection criteria to screen titles, abstracts and full papers of the relevant articles. Studies were excluded if they did not meet the se- lection criteria. Any disagreement was resolved through dis- cussion; if disagreement persisted, then a third author was asked to make final selection decisions regarding inclusion or exclusion.

Data extraction and management

The following data were extracted: study characteristics (e.g. author and year), participant characteristics (e.g. age and number of subjects), description of interventions, duration of trial period, types of outcomes assessed and time point/s. The same two authors who performed the selection of studies extracted the data from the included articles. Any dis- agreement was resolved through discussion, and if disagree- ment persisted, a third author was consulted.

Assessment of risk of bias in the included studies

The Physiotherapy Evidence Database scale [17, 18] was used to evaluate the risk of bias for inclusion in the systemat- ic review. The following information were assessed: random allocation, blind subjects, blind therapists, blind assessors, concealed allocation, baseline comparability, intention-to- treat analysis, adequate follow-up, between-group compari- sons, point estimates and variability. Two review authors independently used a pre-determined 10-item scoring system to assess the methodological quality of each study. Review authors did not evaluate their own studies. Again, a third author was consulted if there was any disagreement.

Statistical analysis

For the meta-analysis, we required means and standard deviations (SD) for outcome measures. If continuous data were reported as median and inter-quartile range (IQR), the median would be assumed to be equivalent to the mean, and the relationship of IQR and the standard deviation is roughly as follows [19]: SD = IQR/1.35. It was necessary to contact the author if standard deviations were not shown or deriv- able in the original publication.

The data were combined when similar contrasts were used in those studies with multiple comparisons. Three

206

X. Wang et al.

studies [12, 20, 21] reported different subgroups of CMI (such as dual-task fixed exercise, dual-task variable exercise), and in these cases, the means and standard deviations of these subgroups were pooled using the following formulas [19]:

We used Review Manager Software (RevMan5.2) to conduct the meta-analysis. We used the χ2 test and the I2 stat- istic to evaluate heterogeneity among the studies. When the heterogeneity test showed no significant difference (I2 < 50%; P > 0.1), we used the fixed-effects model; otherwise, the random-effects model was used. We considered P < 0.05 to be statistically significant. Funnel plot asymmetry was used to assess possible publication bias by Egger’s regression test.

Results

Search results

Among 3,571 identified records from Medline, Embase, the Cochrane Library, Web of Science, CINAHL, PEDro, the China Biology Medicine disc and manual search, according to their titles and abstracts, we identified 60 potentially eli- gible studies. After reviewing the full papers of the 60 poten- tial articles, 30 articles [8, 10, 12, 13, 20, 21, 22, 23, 24, 25, 26, 27, 28–32, 33, 34, 35, 36, 37–45] fulfilled the inclusion cri- teria. The remaining 30 articles were excluded, because their studies included participants with neurological illness (e.g. stroke, Parkinson’s disease and cognitive impairment) or par- ticipants were younger than 60 years. Other studies were excluded either because original data was no longer available from the authors or because they did not compare CMI with a control group. Supplementary data, Appendix 1 available in Age and Ageing online summarise the characteristics of each included study. The process of identifying eligible studies is outlined in Supplementary data, Appendix 1 available in Age and Ageing online.

Description of studies

The Physiotherapy Evidence Database scale (with scores from 1 to 10) was used for the methodological quality assess- ment and yielded a mean score of 5.43 (SD = 1.38). Briefly, every study was reported as using random allocation, but 22 of the included trials failed to adopt allocation concealment. A total of 16 studies tried to blind the assessors to the allo- cated treatment, and 7 trials performed intention-to-treat analysis. According to the total PEDro score, studies were

classified as high-, moderate- and low-quality RCTs. Only one article was low quality, other included articles were of moderate-to-high quality and 26.7% (8 out of 30) were high quality. Full details of the methodological quality of these trials are presented in Supplementary data, Appendix 2 avail- able in Age and Ageing online.

Effects of CMI on gait variables

Effects of CMI on gait speed: the results showed that CMI for gait speed was better than the control group in a random-effects model [SMD (95% CI) = 0.36 (0.07, 0.66), P = 0.01] (Figure 1; Table 1).

Effects of CMI on stride length: the results showed that CMI improved stride length compared with controls in a fixed-effects model [SMD (95% CI) = 0.26 (0.03, 0.48), P = 0.03] (Supplementary data, Appendix 2 available in Age and Ageing online; Table 1).

Effects of CMI on step length: CMI was better than controls in terms of improving step length in a fixed-effects model [SMD (95% CI) = 0.48 (0.16, 0.80), P = 0.003] (Supplementary data, Appendix 3 available in Age and Ageing online; Table 1).

Effects of CMI on cadence: CMI was better than the control group for improving cadence in a fixed-effects model [SMD (95% CI) = 0.19 (0.01, 0.36), P = 0.03] (Supplementary data, Appendix 4 available in Age and Ageing online; Table 1).

Effects of CMI on balance function

Effects of dual-task exercise on the timed up and go test: CMI was better than controls for improving the timed up and go test performance in a fixed-effects model [SMD (95% CI) = −0.22 (−0.38, −0.06), P = 0.007] (Figure 2; Table 1).

Effects of CMI on the functional reach test: the data did not reveal any difference between CMI and controls in a random-effects model [MD (95% CI) = 0.57 (−1.20, 2.33), P = 0.53] (Table 1).

Effects of CMI on performance in the activities-specific balance confidence scale: the data did not reveal any differ- ence between CMI and control group in a fixed-effects model [MD (95% CI) = 2.34 (−2.02, 6.70), P = 0.29] (Table 1).

Effects of CMI on the one leg test: the data did not reveal any difference between CMI and control group in a fixed- effects model [MD (95% CI) = 3.04 (−0.33, 6.40), P = 0.08] (Table 1).

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Group 1 (e.g. males) Group 2 (e.g. females) Combined groups

Sample size N1 N2 N1 þ N2 Mean M1 M2 ðN1M1 þ N2M2Þ=N1N2

SD SD1 SD2

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ðN1 � 1ÞSD12 þ ðN2 � 1ÞSD22 þ ðN1N2=ðN1þN2ÞÞðM12 þ M22 � 2M1M2Þ

N1 þ N2 � 1

s

207

Cognitive motor interference for older adults

Effects of CMI on the berg balance scale: the data did not reveal differences between CMI and control group in a random-effects model [MD (95% CI) = 1.39 (−1.18, 3.96), P = 0.290] (Table 1).

Effects of CMI on the chair stand test: the data did not reveal any difference between CMI and control group in a fixed-effects model [MD (95% CI) = 1.0 (−0.32, 2.31), P = 0.14] (Table 1).

Effects of CMI on Tinetti test: the results showed that CMI was better than control group for Tinetti test in a random-effects model [SMD (95% CI) = −0.64 (−1.24, −0.05), P = 0.03] (Table 1).

Effects of CMI on centre of pressure displacement: the results showed that CMI was better than control group for centre of pressure displacement in a fixed-effects model [SMD (95% CI) = −0.4 (−0.77, −0.02), P = 0.04] (Supplementary data, Appendix 5 available in Age and Ageing online; Table 1).

Effects of CMI on fall

Effects of CMI on fall rate: the results showed that CMI was better than control group for fall rate in a random-effects model [SMD (95% CI) = −3.03 (−4.33, −1.73), P < 0.0001] (Table 1).

Effects of CMI on falls efficacy scale: the data did not reveal any difference between CMI and control group in a fixed-effects model [MD (95% CI) = −0.66 (−3.23, 1.91), P = 0.61] (Table 1).

Effects of CMI on reaction time

After the results were combined in a fixed-effects model, the data indicated that CMI was better than control group in terms of reaction time [SMD (95% CI) = −0.47 (−0.86, −0.08), P = 0.02] (Supplementary data, Appendix 6 available in Age and Ageing online; Table 1).

Publication bias

Egger’s regression test did not show any publication bias for gait speed (asymmetry test, P = 0.879) and time up and go test (asymmetry test, P = 0.971) (Figure 3).

Discussion

Many different types of physical activity interventions are used to try to prevent falls among older adults. Previous sys- tematic reviews have concentrated on single-task exercise (e.g. strength training and balance exercise), but this current review and meta-analysis brings together evidence from rele- vant articles evaluating CMI (i.e. simultaneous cognitive and motor tasks) in the context of preventing falls in older adults.

Our systematic review and meta-analysis of trials from 30 RCTs comprising 1,206 subjects provided evidence supporting the efficacy of CMI for improving gait and balance, and pre- venting falls in older adults. The results of our systematic

Figure 1. Meta-analyses of cognitive motor interference on gait speed. SD, standard deviation; 95% CI, 95% confidence intervals; IV, inverse variance.

208

X. Wang et al.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Table 1. Summary of results

Outcome Trials Participants Statistical method Effect estimate Heterogeneity P value

Gait Gait speed 12 [8, 12, 23, 24, 25, 26, 28,

31, 32, 33, 34, 40] 760 Standard mean difference

(IV, random, 95% CI) 0.38 [0.23, 0.52] 0.04 0.01

Stride length 8 [8, 20] 310 Standard mean difference (IV, fixed, 95% CI)

0.26 [0.03, 0.48] 0.23 0.03

Step length 3 [26, 31, 33] 165 Standard mean difference (IV, fixed, 95% CI)

0.48 [0.16, 0.80] 0.11 0.003

Cadence 5 [8, 26, 28, 31, 34] 533 Standard mean difference (IV, fixed, 95% CI)

0.19 [0.01, 0.36] 0.1 0.03

Balance function Timed up and go test 12 [8, 13, 21, 22, 25, 27, 28,

32, 34, 36, 39, 45] 629 Standard mean difference

(IV, fixed, 95% CI) −0.22 [−0.38, −0.06] 0.11 0.0007

Functional reach test 4 [13, 25, 28, 36] 140 Mean difference (IV, fixed, 95% CI)

0.57 [−1.20, 2.33] 0.98 0.53

Activities-specific balance confidence 3 [12, 21, 32] 76 Mean difference (IV, fixed, 95% CI)

0.26 [−7.11, 7.63] 0.31 0.94

One leg balance 2 [22, 28] 162 Mean difference (IV, fixed, 95% CI)

3.04 [−0.33, 6.40] 0.3 0.08

Berg balance scale 2 [12, 21] 59 Mean difference (IV, random, 95% CI)

0.18 [−2.73, 3.10] 0.1 0.9

Chair Stand Test 2 [13, 27] 120 Mean difference (IV, fixed, 95% CI)

0.59 [−1.67, 2.86] 0.48 0.61

Centre of pressure displacement 7 [8, 13, 23, 24, 33, 38, 44] 371 Standard mean difference (IV, random, 95% CI)

−0.32 [−0.80, 0.15] 0.01 0.04

Falls Falls rate 2 [8, 38] 204 Standard mean difference

(IV, random, 95% CI) −3.03 [−4.33, −1.73] 0.004 <0.0001

Falls efficacy scale 5 [25, 26, 31, 34, 39] 168 Mean difference (IV, random, 95% CI)

−0.66 [−3.23, 1.91] 0.04 0.61

Reaction time Reaction time 3 [33, 34, 36] 107 Standard mean difference

(IV, fixed, 95% CI) −0.47 [−0.86, −0.08] 0.44 0.02

Figure 2. Meta-analyses of cognitive motor interference on time up and go test. TUG, time up and go; SD, standard deviation; 95% CI, 95% confidence intervals; IV, inverse variance.

209

Cognitive motor interference for older adults

review showed significant differences in comparing CMI with single-task exercise or no intervention for eight outcomes, in- cluding gait speed, stride length, step length, cadence, the timed up and go test, centre of pressure displacement, fall rates and reaction time. Most of the observed differences among the groups were not large, but the improvements seen for gait speed, the timed up and go test, Tinetti test, centre of pressure displacement, fall rates and reaction time were at levels that may signify clinical importance. Moreover, no serious adverse events were reported in the 30 RCTs included in our meta-analysis.

In contrast to the above-mentioned outcome measures, there were no significant difference in other balance function and fall outcome measures (e.g. activities-specific balance confidence scale, functional reach test, one leg test, berg balance scale, chair stand test and falls efficacy scale) when comparing CMI with single-task exercise or no intervention.

However, it is possible that the number of included participants and studies were not sufficient to detect the effects of CMI.

Strengths and limitations

This is the first systematic review and meta-analysis to evalu- ate the effects of CMI in terms of balance, gait and falls in older people by comparing with single-task exercise or no intervention. Previous systematic reviews [14, 15, 46, 47, 48] either focused on qualitative synthesis rather than meta- analysis or did not compare CMI with single-task exercise or no intervention. Compared with prior systematic reviews, most of the included studies were new in the current study, and we only considered older people (aged over 60) for all of the articles in the current analysis.

Furthermore, a wide range of electronic databases were searched in our systematic review covering all relevant articles [49]. In addition, there were no restrictions on the basis of publication date or language, and most of the articles were published within the last 5 years. To decrease bias and tran- scription errors, two reviewers independently selected studies, extracted the data and evaluated the quality of the article. The results of our systematic review are considered to be very robust due to these characteristics.

However, there are some limitations in our systematic review. First, high-quality articles were still deficient in our systematic review. Based on PEDro score, only 26.7% (8 out of 30) RCTs were high quality in the systematic review. Most studies had inadequate reporting or poor methodological quality. Only 8 of the 30 studies showed how allocation of participants was concealed, and 23 of 30 studies did not perform intention-to-treat analyses. None of the 30 studies blinded participants and therapists, and only 16 studies blinded assessors.

Second, 22% of older people aged over 75 have mild cog- nitive impairment [50] and cognitive impairment in the elderly increases risk of falls, and hence, the improvement of cognitive function may be a reduction in fall risk [51]. Thus, we would have liked to have looked for differences in cognitive function measures between subjects undergoing CMI and single-task exercise or no intervention in our review, but few studies reported these cognitive function outcome measures.

Third, the follow-up durations were <1 year in all of the studies, and hence, the long-term effect of CMI could not be estimated.

Fourth, the incidence and prevalence of chronic diseases increase with age stroke and serious arthritis. However, due to heterogeneity, older people with pathological causes of poor balance are excluded in this article. Hence, it is neces- sary to perform a meta-analysis to determine the effect of CMI on falls for different patients with pathological causes of poor balance.

Implications for research

In brief, articles should adhere to generally accepted stan- dards of reporting clinical trials (e.g. the Consolidated

Figure 3. (A) Funnel plot of 11 studies included in meta- analysis of studies regarding the publication bias for gait speed (asymmetry test P = 0.879); (B) Funnel plot of 10 studies included in meta-analysis of studies regarding the publication bias for time up and go test (asymmetry test, P = 0.971). SE, standard error; SMD, standardised mean difference.

210

X. Wang et al.

Standards of Reporting Trials [CONSORT] statement) [52]. As previously mentioned, most of the studies in our system- atic review had short follow-up periods, and the sample sizes in individual studies were modest. Consequently, there is a need for large-scale RCTs for CMI. At present, there are many types of CMI that could have different results in older adults. Therefore, it is necessary to perform a systematic review of different CMI for determining the optimal inter- vention programme in older people. Eventually, theories regarding the mechanisms by which CMI may prevent the likelihood of falls in older people should be explored further.

Conclusions and implications for practice

In our systematic review, we found that CMI is more effective than single-task exercise or no intervention in older adults for reducing falls and improving gait and balance. This pro- vides useful information for older people (not include older people with any pathology affecting the balance function) as well as for medical staff and health-care decision-makers in coming up with effective exercise regimes for this age group.

Key points

• This is the first systematic review to objectively estimate the effects of CMI for older people.

• A wide range of electronic databases were searched in our systematic review covering all relevant articles.

• There were no restrictions on the basis of publication date or language.

Acknowledgements

We thank KGSupport for English language editing and Alexandra Halvarsson for providing detail data for our analyses.

Conflicts of interest

None declared.

Funding

This study was supported by the Key Laboratory of Exercise and Health Sciences (Shanghai University of Sport); Ministry of Education, the First-class Disciplines of Shanghai Colleges and Universities (Psychology); Science and Technology Commission of Shanghai Municipality (No. 13490503500).

Supplementary data

Supplementary data mentioned in the text are available to subscribers in Age and Ageing online.

References

The very long list of references supporting this review has meant that only the most important are listed here and are represented by bold type throughout the text. The full list of references is available on Supplementary data, Appendix e-2 available in Age and Ageing online.

1. Kannus P, Sievänen H, Palvanen M, Järvinen T, Parkkari J. Prevention of falls and consequent injuries in elderly people. Lancet 2005; 366: 1885–93.

2. Tinetti ME, Baker DI, King M et al. Effect of dissemination of evidence in reducing injuries from falls. N Engl J Med 2008; 359: 252–61.

3. Michael YL, Whitlock EP, Lin JS et al. Primary care–relevant interventions to prevent falling in older adults: a systematic evi- dence review for the U.S. Preventive Services Task Force. Ann Intern Med 2010; 153: 815–25.

4. Moyer VA, U.S. Preventive Services Task Force. Prevention of falls in community-dwelling older adults: U.S. Preventive Services Task Force recommendation statement. Ann Intern Med 2012; 157: 197–204.

5. Chang JT, Morton SC, Rubenstein LZ et al. Interventions for the prevention of falls in older adults: systematic review and meta-analysis of randomised clinical trials. BMJ 2004; 328: 680.

6. Tinetti ME. Clinical practice. Preventing falls in elderly persons. N Engl J Med 2003; 348: 42–9.

8. Trombetti A, Hars M, Herrmann FR, Kressig RW, Ferrari S, Rizzoli R. Effect of music-based multitask training on gait, balance, and fall risk in elderly people: a randomized con- trolled trial. Arch Intern Med 2011; 171: 525–33.

9. Sigman M, Dehaene S. Dynamics of the central bottleneck: dual-task and task uncertainty. PLoS Biol 2006; 4: e220.

10. Silsupadol P, Siu KC, Shumway-Cook A, Woollacott MH. Training of balance under single- and dual-task conditions in older adults with balance impairment. Phys Ther 2006; 86: 269–81.

12. Silsupadol P, Shumway-Cook A, Lugade V et al. Effects of single-task versus dual-task training on balance performance in older adults: a double-blind, randomized controlled trial. Arch Phys Med Rehabil 2009; 90: 381–7.

14. Pichierri G, Wolf P, Murer K, de Bruin ED. Cognitive and cognitive-motor interventions affecting physical functioning: a systematic review. BMC Geriatr 2011; 11: 29.

15. Al-Yahya E, Dawes H, Smith L, Dennis A, Howells K, Cockburn J. Cognitive motor interference while walking: a sys- tematic review and meta-analysis. Neurosci Biobehav Rev 2011; 35: 715–28.

17. Coombes BK, Bisset L, Vicenzino B. Efficacy and safety of corticosteroid injections and other injections for management of tendinopathy: a systematic review of randomised controlled trials. Lancet 2010; 376: 1751–67.

18. Pinto RZ, Maher CG, Ferreira ML et al. Drugs for relief of pain in patients with sciatica: systematic review and meta-ana- lysis. BMJ 2012; 344: e497.

22. Vaillant J, Vuillerme N, Martigné P et al. Balance, aging, and osteoporosis: effects of cognitive exercises combined with physiotherapy. Joint Bone Spine 2006; 73: 414–8.

23. You JH, Shetty A, Jones T, Shields K, Belay Y, Brown D. Effects of dual-task cognitive-gait intervention on memory and gait dynamics in older adults with a history of falls: a preliminary investigation. Neuro Rehabilitation 2009; 24: 193–8.

211

Cognitive motor interference for older adults

25. Takai I. Influence of balance exercises during an attention task on fear of falling among frail elderly fallen. Nihon Ronen Igakkai Zasshi 2010; 47: 220–5.

26. Halvarsson A, Oddsson L, Olsson E, Farén E, Pettersson A, Ståhle 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. Clin Rehabil 2011; 25: 1021–31.

28. Yamada M, Aoyama T, Tanaka B, Nagai K, Ichihashi N. Seated stepping exercise in a dual-task condition im- proves ambulatory function with a secondary task: a ran- domized controlled trial. Aging Clin Exp Res 2011; 23: 386–92.

29. Lussier M, Gagnon C, Bherer L. An investigation of response and stimulus modality transfer effects after dual-task training in younger and older. Front Hum Neurosci 2012; 6: 129.

30. Pichierri G, Coppe A, Lorenzetti S, Murer K, de Bruin ED. The effect of a cognitive-motor intervention on voluntary step execution under single and dual task conditions in older adults: a randomized controlled pilot study. Clin Interv Aging 2012; 7: 175–84.

31. Pichierri G, Murer K, de Bruin ED. A cognitive-motor inter- vention using a dance video game to enhance foot placement

accuracy and gait under dual task conditions in older adults: a randomized controlled trial. BMC Geriatr 2012; 12: 74.

32. Plummer-D’Amato P, Cohen Z, Daee NA, Lawson SE, Lizotte MR, Padilla A. Effects of once weekly dual-task train- ing in older adults: a pilot randomized controlled trial. Geriatr Gerontol Int 2012; 12: 622–9.

35. Melzer I, Oddsson L. Improving balance control and self- reported lower extremity function in community-dwelling older adults: a randomized control trial. Clin Rehabil 2013; 27: 195–206.

36. Marmeleira JF, Godinho MB, Fernandes OM. The effects of an exercise program on several abilities associated with driving performance in older adults. Accid Anal Prev 2009; 41: 90–7.

46. Lee H, Sullivan SJ, Schneiders AG. The use of the dual-task paradigm in detecting gait performance deficits following a sports-related concussion: a systematic review and meta- analysis. J Sci Med Sport 2013; 16: 2–7.

47. Verhaeghen P, Steitz DW, Sliwinski MJ, Cerella J. Aging and dual-task performance: a meta-analysis. Psychol Aging 2003; 18: 443–60.

Received 16 February 2014; accepted in revised form 7 May 2014

212

X. Wang et al.

Copyright of Age & Ageing is the property of Oxford University Press / USA and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.