Gait & Posture 85 (2021) 65–70

Available online 13 January 20210966-6362/© 2021 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

Full length article

Exercise effects on backward walking speed in people with dementia: A randomized controlled trial

Annika Toots a,*, Lillemor Lundin-Olsson a, Peter Nordstrom b, Yngve Gustafson b, Erik Rosendahl a

a Department of Community Medicine and Rehabilitation, Physiotherapy, Umeå University, Umeå University, 901 87, Umeå, Sweden b Department of Community Medicine and Rehabilitation, Geriatric Medicine, Umeå University, Umeå, 901 87, Sweden

A R T I C L E I N F O

Keywords: Backward walking speed Dementia Exercise Residential facilities Mobility limitations

A B S T R A C T

Background: Multidirectional walking, including backward walking, is integral to daily activities, and seems particularly challenging in older age, and in people with pathology affecting postural control such as dementia. Research Question: Does exercise influence backward walking speed in people with dementia, when tested using habitual walking aids and without, and do effects differ according to walking aid use? Methods: This study included 141 women and 45 men (mean age 85 years) with dementia from the Umeå De-mentia and Exercise (UMDEX), a cluster-randomized controlled trial study set in 16 nursing homes in Umeå, Sweden. Participants were randomized to a High-Intensity Functional Exercise (HIFE) program targeting lower limb strength-, balance and mobility exercise or to a seated attention control activity. Blinded assessors measured 2.4-meter usual backward walking speed, at baseline, 4 - (intervention completion) and 7-month follow-up; tested 1) with habitual walking aids allowed, and 2) without walking aids. Results: Linear mixed models showed no between-group effect in either backward walking speed test at 4 or 7 months; test 1) 0.005 m/s, P = .788 and –0.006 m/s, P = .754 and test 2) 0.030 m/s, P = .231 and 0.015 m/s, P =.569, respectively. In interaction analyses, exercise effects differed significantly between participants who habitually walked unaided compared with those that used a walking aid at 7 months (0.094 m/s, P = .027). Significance: In this study of older people with dementia living in nursing homes, the effects of exercise had no overall effects on backwards walking speed. Nevertheless, some benefit was indicated in participants who habitually walked unaided, which is promising and merits further investigation in future studies.

1. Introduction

The ability to walk throughout older age is important since it is related to greater levels of physical activity, independence in activities of daily life (ADLs) and survival [1]. In addition, lower forward walking speed is an early predictor of cognitive impairment and dementia [2], and deteriorates concomitantly with disease severity [3]. Even walking forward on level surfaces, a seemingly simple and automatic motor task, may be considered complex and reliant on higher-level cognitive func-tion in older adults [4,5]. Walking and balance has been linked to several cognitive domains [6], of which executive function appears to be particularly important [5]. Walking is also an inherently unstable ac-tivity, during which many falls occur. Falls are the most common cause of injury in older age, can be devastating to physical function and in-dependence, and a frequent cause of death [7]. In people with dementia,

who alongside impaired cognition often have walking dysfunction [8], the risk of falls and fall-related injuries are even higher than their cognitively healthy counterparts [7,9].

Multidirectional walking, such as walking forward and backward or to the sides, is an integral part of many daily activities, e.g. in prepa-ration for sitting, or for opening and closing door. Compared to forward walking, backward walking appears to be particularly sensitive to age- related changes [10,11], as well as disease specific pathology affecting postural stability such as Parkinson’s disease [12]. In adults with de-mentia, a comparison of forward and backward walking speed has also indicated considerable detriment to the latter [4]. Furthermore, there is evidence to suggest that a deterioration in backward walking speed is related to increased risk of falls [11].

Physical exercise, in the form of moderate to high intensity strength- and balance training, has in meta-analyses shown to be the most

* Corresponding author at: Department of Community Medicine and Rehabilitation, Physiotherapy, Umeå University, 901 87, Umeå, Sweden. E-mail address: annika.toots@umu.se (A. Toots).

Contents lists available at ScienceDirect

Gait & Posture

journal homepage: www.elsevier.com/locate/gaitpost

https://doi.org/10.1016/j.gaitpost.2020.12.028 Received 10 March 2020; Received in revised form 22 December 2020; Accepted 24 December 2020

Gait & Posture 85 (2021) 65–70

66

efficacious falls preventative intervention for older people (approxi-mately 21 % less falls) [13]. While earlier studies of the effects of ex-ercise in people with dementia indicate that balance and forward walking speed, as well as dependence in ADLs can be improved [14], research on backward walking is a new and unexplored area. Few studies have investigated the effects of exercise on backward walking and, and albeit encouraging, the results remain preliminary [15]. No study has investigated the effects in people with dementia.

We have previously reported findings suggesting beneficial effects from exercise on balance and forward walking speed in people with dementia living in nursing homes [16,17]. The aim of this study was to investigate the effects of a high-intensity functional exercise program on backward walking speed in the same population. Furthermore, walking aid use is common among older people. Using a walking aid provides

additional support when balance or muscle strength is poor. In addition, given that walking requires a degree of cognitive input, the walking aid may further reduce the cognitive challenge of forward walking [4,18]. As a result, walking aids may limit detection of walking deficits and reduce responsiveness of the walking speed test [17,19]. A further aim therefore, was to evaluate whether effects on backward walking speed differed according to walking aid use.

2. Methods

This study was part of the Umeå Dementia and Exercise (UMDEX) study [16,17] in which backwards walking speed was a prespecified secondary outcome measure. The UMDEX study was designed as a parallel-group, cluster-randomized controlled trial in older people with

Fig. 1. Flow of participants throughout the study. ADLs, Activities of Daily Living; BWS, Backward walking speed with walking aids allowed, MMSE, Mini-Mental State Examination.

A. Toots et al.

Gait & Posture 85 (2021) 65–70

67

dementia living in nursing homes. The study compared a high-intensity functional exercise program with a seated control activity. The length and number of sessions were equal between the activities to control for attention. Attention can have important effects on cognition in this population, which generally has limited social interactions. The study protocol is published on the ISRCTN registry (ISRCTN31767087). All individuals included in the study gave informed oral consent to partic-ipation confirmed by their next of kin. The study complies with princi-ples outlined in the Declaration of Helsinki, and was approved by the Regional Ethics Review Board in Umeå (2011-205-31 M).

2.1. Participants

Participants were recruited from 16 nursing homes in Umeå, Swe-den. Inclusion criteria were Mini-Mental State Examination (MMSE) score ≥10 [20], dementia diagnosis [21], age ≥65 years, dependent on assistance in ≥1 personal ADL [22], ability to stand up from a chair with armrests with assistance of ≤1 person, physician’s approval, and ability to hear and understand spoken Swedish sufficiently to participate in assessments. A total of 864 nursing home residents were screened. Included participants did not differ in age or MMSE score (P = 0.189 and P = 0.713, respectively) to those who declined participation (n = 55, Fig. 1). A larger proportion of men than women declined participation (34 % versus 18 %; P= 0.008). Baseline characteristics of the 186 par-ticipants are presented in Table 1. The sample size was calculated for the main outcome in the UMDEX study, the Barthel ADL Index [16]; 183 participants were required to verify significant intervention effects at a statistical power of 80 % at the 4-month follow-up, a significance level of 0.05, two-sided, and a presumed dropout rate of 10 %.

2.2. Randomization

Participants were randomized after completion of enrolment process and baseline assessment to ensure concealed allocation. Participants who lived in the same wing, unit, or floor were formed into 36 clusters (n = 3–8 each) to reduce contamination. The randomization was stratified in all nursing homes except one that had only a single cluster; the object being to have participants in both exercise and attention control activ-ities at each nursing home to reduce the risk of site-specific factors influencing the outcome. Two researchers not involved in the study performed randomization by drawing lots using sealed opaque envelopes.

2.3. Outcome measures

Physiotherapists (PTs) and physicians blinded to activity allocation and previous test results performed all measurements at baseline and follow-ups.

2.3.1. Backward walking speed Usual backward walking speed was measured over 2.4 m (8 ft), a

distance frequently used in tests of forward walking speed. The distance was marked on the wall, and a visible target (a chair) placed 6 m from start. From a standing still position, participants were asked to walk backward in their usual pace to the target or until told to stop. Time was measured with a digital stopwatch, from when instructed to start until crossing the 2.4 m marking, thus included acceleration but not decel-eration. Habitual walking aids were allowed, the procedure was repeated once, and walking speed calculated (m/s). To reduce balance support, a second test without walking aids was conducted as above. Participants who habitually used a walking aid performed the second test without it. Participants who walked unsupported performed the first test only, their time was carried over to the second test. Gait belts with handles were used to reduce the risk of falls in case of unsteadiness. Reasons for not being able to perform the backward walking speed test were registered and classified (physical impairment, motivation, or

other causes e.g. pain, dizziness, absence from ward). Only the walking aid used at baseline was allowed at subsequent follow ups.

2.3.2. Descriptive measurements Usual forward walking speed (2.4 m) was measured using the same

procedure as above, and the mean of two trials calculated (m/s) [23]. Dependence in ADLs was measured using Barthel ADL Index (0–20) [24]. Cognition was measured using the MMSE (0–30) [20] and the Verbal fluency test (executive function). Behavioral and psychological symptoms in dementia were measured using the Neuropsychiatric In-ventory (0–144). Symptoms of depression were assessed using the 15-item Geriatric Depression Scale (0–15). Balance was measured using the Berg Balance Scale (BBS, 0–56). Dementia type, depressive

Table 1 Baseline characteristics.

Total Exercise Control n = 186 n = 93 n = 93

Age, yrs, mean ± SD 85.1 ±7.1

84.4 ±6.2

85.9 ±7.8

Female, n (%) 141 (75.8)

70 (75.3) 71 (76.3)

Dementia type, n(%) Vascular 77

(41.4) 36 (38.7) 41 (44.1)

Alzheimer’s 67 (36.0)

34 (36.6) 33 (35.5)

Other 27 (14.5)

15 (16.1) 12 (12.9)

Mixed Alzheimer’s/Vascular 15 (8.1) 8 (8.6) 7 (7.5) Diagnoses and medical conditions, n (%) Depressive disorders 107

(57.5) 53 (57.0) 54 (58.1)

Delirium (within previous week) 102 (54.8)

48 (51.6) 54 (58.1)

Stroke 57 (30.6)

33 (35.5) 24 (25.8)

Heart failure 56 (30.1)

24 (25.8) 32 (34.4)

Prescription medication, n (%) Analgesics 112

(60.2) 55 (59.1) 57 (61.3)

Antidepressants 102 (54.8)

58 (62.4) 44 (47.3)

Neuroleptics 31 (16.7)

11 (11.8) 20 (21.5)

Number of drugs, mean ± SD 8.3 ± 3.8 8.4 ± 4.0 8.2 ± 3.7 Assessments, mean ± SD Usual mobility aid, n(%)

Wheelchair 24 (12.9)

11 (11.8) 13 (14.0)

Rollator 117 (62.9)

64 (68.8) 53 (57.0)

Stick or crutch 4 (2.2) 1 (1.1) 3 (3.3) Unsupported 41 (22) 17 (18.3) 24 (25.8)

Mini-Mental State Examination (0− 30)a 14.9 ±3.5

15.4 ±3.4

14.4 ±3.5

Verbal Fluency, n = 182a 6.4 ± 3.8 6.8 ± 4.1 6.0 ± 3.5 Barthel ADL Index (0− 20)a 10.9 ±

4.4 10.7 ±4.5

11.0 ±4.4

Neuropsychiatric Inventory (0− 144)b 14.8 ±14.2

15.2 ±15.8

14.4 ±12.6

Forward Walking speed (with or without walking aid) m/s, n=177

0.47 ±0.21

0.47 ±0.20

0.47 ±0.22

Forward Walking speed (without walking aids) m/s, n=99

0.55 ±0.21

0.54 ±0.22

0.56 ±0.21

Backward Walking speed (with or without walking aids) m/s, n=161

0.23 ±0.14

0.22 ±0.14

0.23 ±0.14

Backward Walking speed (without walking aids) m/s, n=86

0.24 ±0.15

0.23 ±0.14

0.24 ±0.16

Numbers reported after covariates indicate number of measurements available when values were missing.

a Higher score indicates better status. b Lower score indicates better status.

A. Toots et al.

Gait & Posture 85 (2021) 65–70

68

disorders, and delirium diagnoses were recorded through review of electronic records of past medical history, which in most cases included brain imaging. In addition, a specialist in geriatric medicine further reviewed and confirmed the diagnoses according to DSM-IV-TR criteria [21], taking into account current pharmaceutical treatment, and assessment results.

2.4. Intervention

PT’s led the exercise activities, and occupational therapists (OT) or an OT assistant the attention control activities; conducted at nursing homes in small groups (n = 3–8). The intervention consisted of 5 ses-sions per fortnight for the duration of 4 months (40 sessions in total), with each session lasting approximately 45 min. Participants unable to attend group sessions were offered supervised individual sessions when possible. No activities were offered after completion of the 4-month intervention. Participation in activities other than provided by the study were not restricted at any time. At the end of each session, leaders completed a structured protocol for each participant pertaining to adverse events, and strength- and balance exercise intensity in the ex-ercise group; where exercise intensity was estimated individually as high, moderate, or low according to a predefined scale [25]. Over the 4-month intervention period, adherence in the exercise group was 73 % and 70 % in the attention control group. In the exercise group, partici-pants performed strength exercises with moderate intensity (40 %) and at high intensity (49 %) of attended sessions, and balance exercises with moderate intensity (26 %) and high intensity (68 %) of attended sessions [26]. All adverse events recorded during exercise sessions were minor or temporary [26].

2.4.1. Exercise activity The exercise intervention was based on the high-intensity functional

exercise program (HIFE), which includes a model for exercise selection and a definition of exercise intensity (available online at https://www. hifeprogram.se/en). In brief, the HIFE program comprises 39 func-tional exercises for improved lower limb strength, balance, and mobility, including multidirectional walking and stepping, and stepping over obstacles, to be performed with high intensity and in weightbearing positions similar to daily activities. High-intensity in strength exercise was defined as 8-12 repetition maximum and in balance exercise when postural stability was fully challenged. Exercises were initially selected and tailored based on participants’ functional deficits e.g. amount of assistance needed to ambulate. Participants were supervised individu-ally to promote the highest possible exercise intensity, and adapted accordingly through progressive adjustment of load and base of support, while also taking into account participants’ symptoms and changes in health and functional status. For safety, participants wore gait belts with handles for PTs to provide support if postural stability was threatened, thereby preventing falls. Unnecessary support, however light, was avoided as it is known to affect postural control [27].

2.4.2. Attention control activity The attention control group participated in structured activities that

were developed by the OTs/OT assistant that led the activities. The activities were structured around topics believed to be interesting for older people, including local wild life, seasons, and holidays. While seated in a group, participants conversed, sang, listened to music or readings, and/or looked at pictures and objects.

2.5. Statistics

An a priory strategy for selection of possible confounders was formed. Comparisons between exercise and attention control group and associ-ations (r ≥ 0.3) with changes in outcome measures at 4 and 7 months were estimated for all variables in Table 1, pre-selected as possible confounders, using Student t-test, the Chi2 test, and Pearson correlation

coefficients. No variable was found to associate with change in outcome measures above predefined levels. The variable antidepressant use differed between groups (P = .04) and was adjusted for in all analyses. In agreement with the intention-to-treat principle, available data for each participant was analyzed according to original allocation and regardless of level of attendance. Longitudinal changes in backward walking speed from baseline to 4 and 7 months were analyzed in linear mixed effects models, using interaction terms for activity and time point and adjust-ment for age, sex, and antidepressant use as fixed effects, and individual and cluster allocation as random effects. Baseline values for outcome measures were included in the dependent variable, thus using all available data. The adjusted within-group differences were compared using estimated marginal means from the models. Subgroup analyses were conducted according to walking aids used in the test by adding interaction terms to adjusted models. Use of walking aids were dichot-omized as with or without walking aid. All analyses were performed using IBM SPSS statistics for Macintosh, version 25.0 (IBM Corp. Armonk, NY: IBM Corp.), two tailed, and with P < 0.05 considered statistically significant.

3. Results

In total, 141 women and 45 men, with mean ± standard deviation (SD) age of 85.1 ± 7.1 years and a MMSE score of 14.9 ± 3.5 were included (Table 1). Seventy-seven (41.2 %) participants had vascular dementia and 67 (36.0 %) Alzheimer’s disease. Of participants who performed the backward and forward walking test, 121 (75 %) and 135 (76 %) used a walking aid, respectively. General reasons for missing assessments i.e. deceased, moved, or when consent was withdrawn, as well as, specific reasons for missing backward walking speed, are pre-sented in Fig. 1. When tested without walking aids (beside general reasons for missing assessments) 100, 88, and 90 participants had missing backward walking speed values at baseline and respective follow-ups, mainly because of physical impairment, but also motivation, or other reasons. At baseline, the mean (SD) backward walking speed in participants who used a walking aid (n = 121) was 0.20 ± 0.12 m/s, while in those who habitually walked without a walking aid (n = 40) it was 0.30 ± 0.18 m/s.

There were no differences between exercise and control groups in either test of backward walking speed, at 4-months (at intervention completion) or 7-months follow-up (Table 2). In interaction analyses, exercise effects on backward walking speed significantly differed ac-cording to walking aid use, with larger effects in participants who walked without a walking aid in the first test at 7 months (0.094, 95 %CI = 0.011, 0.177, P = 0.027, Table 3), while no significant difference was found at 4 months (0.049, 95 %CI=–0.031, 0.130, P = 0.229, Table 3).

4. Discussion

To the best of our knowledge, this is the first study to investigate exercise effects on backwards walking in people with dementia. The study found no overall effects from a high-intensity functional exercise program on backward walking speed at either 4- or 7-month follow-up. The effects differ according to walking aid use at the 7-month follow-up only, with positive effects indicated in participants who habitually walked without a walking aid.

Very few studies have investigated the effects of exercise on back-ward walking speed in older adults, with much variation in experimental designs and exercise regimes limiting comparisons and inferences [15]. One clinical controlled trial of individuals with Parkinson’s disease, found that neither tango nor forward walking treadmill exercise had effects on backward or forward walking speed compared to an active control activity (stretching) [28]. In contrast, a small pilot study in people early post stroke, found that backward walking training benefitted both forward and backwards walking speed compared to standing balance training [29].

A. Toots et al.

Gait & Posture 85 (2021) 65–70

69

Still, the findings correspond well with effects on forward walking speed in our previous study, which also found no overall effects when tested using habitual walking aids or without [17]. The non-significant results may be explained by the individual tailoring of the exercises to the participants’ functional deficits (lower proportion of dynamic bal-ance exercise in walking with lower functional ability). Thus, partici-pants with greater functional deficits at baseline, many of who used walking aids, may have had less exercises in walking, which could have impacted overall effects on backward walking speed.

The larger effect observed in participants who habitually walked without a walking aid is also in line with results from our previous study [17]. The result seemed to indicate that participants in the exercise group maintained their baseline walking speed to a greater extent than participants in the control group. Considering the limited number of participants in this subgroup however, this result needs validation in larger studies.

Many older people with dementia living in nursing homes use walking aids, which was also evident in our study. Although no exercise

effects on backward walking speed were observed in this subgroup, the intervention may still have had effects on balance and muscle strength, as indicated previously [16,17]. That an exercise intervention affects balance yet not walking speed may seem contradictory, but has been shown before in a study of people with dementia living in nursing homes, and where the proportion of walking aid use was akin to ours [30]. A higher exercise dose, for example longer duration, may be required for effects on muscle strength and balance to translate to im-provements in walking speed in people with higher levels of mobility disability.

Some methodological issues in particular merit further discussion. Even when walking aids were allowed, the relatively wide inclusion criteria resulted in floor effects on the backward walking speed test. Although the proportion of participants with missing values was considered too high for imputation to be valid, it seemed evenly dispersed between groups. Furthermore, the statistical method used in analyses (linear mixed effects models) allowed inclusion of all available measurements from all participants and time points, in keeping with the intention-to-treat principle. However, the losses to follow-up reduced the available number of participants by half, hence increasing the risk for type II error in models analyzing effects on walking speed without walking aids. Analyses were based on a single test of backward walking speed at baseline and subsequent follow-ups, and the reliability of the test has not been investigated previously. We used a digital stopwatch to measure walking speed. More advanced measurement methods, such as electronic walkway systems, may have provided more accurate or responsive estimates.

5. Conclusions

In this study of older people with dementia living in nursing homes the effects of exercise had no overall effects on backwards walking speed. Nevertheless, some benefit was indicated in participants who habitually walked unaided, which is promising and merits further investigation in future studies.

Author contributions

AT, LL-O, YG, PN and ER contributed to the study conception and design, acquisition of data, data analysis and interpretation. AT drafted the manuscript, and all authors revised the manuscript critically and approved the version to be submitted.

Funding

This work was supported by the Swedish Research Council (grant

Table 2 Within- and between-group differences from baseline in backward walking speed (BWS).

Test and time point

Exercise Control

Mean (SE)

Within group Mean (SE)a

Mean (SE)

Within- group Mean (SE)a

Between- group Mean (95 %CI)

P

BWS with or without WA, m/s Baseline 0.219

(0.016) – 0.235

(0.016) – –

4 months

0.224 (0.017)

0.005 (0.013)

0.235 (0.017)

0.000 (0.013)

0.005 (–0.031, 0.041)

.788

7 months

0.209 (0.017)

–0.010 (0.013)

0.230 (0.018)

–0.004 (0.013)

–0.006 (–0.043, 0.031)

.754

BWS without WA, m/s Baseline 0.216

(0.022) – 0.223

(0.023) – –

4 months

0.242 (0.023)

0.026 (0.018)

0.219 (0.023)

–0.004 (0.017)

0.030 (–0.019, 0.079)

.231

7 months

0.221 (0.024)

–0.005 (0.020)

0.212 (0.023)

–0.011 (0.018)

0.015 (–0.038, 0.068)

.569

CI, confidence interval; SE, standard error; WA, walking aids. All estimates from linear mixed-effects models adjusted for age, sex and anti-depressant use.

a From pair-wise comparison between follow-up and baseline.

Table 3 Within- and between-group differences from baseline in backward walking speed (BWS) according to walking aid use.

Time point and WA status Exercise Control

Mean (SE) Within-groupa Mean (SE) Mean (SE) Within-groupa Mean (SE) Interactionb Mean (95%CI) P

Baseline BWS without WA, m/s 0.318 (0.031) – 0.308 (0.028) – – BWS with WS, m/s 0.199 (0.017) – 0.207 (0.018) –

4 months BWS without WS, m/s 0.339 (0.033) 0.021 (0.026) 0.288 (0.028) − 0.020 (0.023) 0.049 (-0.031, 0.130) .229 BWS with WA, m/s 0.199 (0.018) 0.000 (0.015) 0.216 (0.019) 0.008 (0.015)

7 months BWS without WA, m/s 0.334 (0.034) 0.016 (0.027) 0.263 (0.023) − 0.045 (0.024) 0.094 (0.011, 0.177) .027 BWS with WA, m/s 0.180 (0.018) − 0.019 (0.015) 0.222 (0.020) 0.014 (0.016)

CI, confidence interval; SE, standard error; WA, walking aids. All estimates from linear mixed-effects models adjusted for age, sex and antidepressant use.

a Pair-wise comparisons between follow-up and baseline. b Difference in exercise effect between participants who walked without walking aids compared to those with walking aids; exercise effect being the difference from

baseline in the exercise group minus the difference in the control group. A positive interaction value indicates a greater exercise effect in favor of participants who walked without walking aids.

A. Toots et al.

Gait & Posture 85 (2021) 65–70

70

numbers K2009-69P-21298-01-4, K2009-69X-21299-01-1, K2009-69P- 21298-04-4, K2014-99X-22610-01-6); Forte – Swedish Research Council for Health, Working Life and Welfare (formerly FAS – Swedish Council for Working Life and Social Research); the Vårdal Foundation; the Swedish Dementia Association; the Promobilia Foundation; the Swedish Society of Medicine; the Swedish Alzheimer Foundation; the King Gus-tav V and Queen Victoria’s Foundation of Freemasons; the European Union Bothnia-Atlantica Program; the County Council of Vasterbotten, the Umeå University Foundation for Medical Research; the Ragnhild and Einar Lundstrom’s Memorial Foundation; and the Erik and Anne-Marie Detlof‘s Foundation. The study sponsors had no role in the design, methods, subject recruitment, data collections, analysis or preparation of paper.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

We would like to acknowledge all who contributed to the data collection and implementation of the UMDEX study, as well as, the So-cial Authorities of the municipality of Umeå, care staff, and participants.

References

[1] S. Studenski, S. Perera, K. Patel, et al., Gait speed and survival in older adults, JAMA 305 (2011) 50–58.

[2] O. Beauchet, C. Annweiler, M.L. Callisaya, et al., Poor gait performance and prediction of dementia: results from a meta-analysis, J. Am. Med. Dir. Assoc. 17 (2016) 482–490.

[3] G. Allali, C. Annweiler, H.M. Blumen, et al., Gait phenotype from mild cognitive impairment to moderate dementia: results from the GOOD initiative, Eur. J. Neurol. 23 (2016) 527–541.

[4] H. Johansson, L. Lundin-Olsson, H. Littbrand, Y. Gustafson, E. Rosendahl, A. Toots, Cognitive function and walking velocity in people with dementia; a comparison of backward and forward walking, Gait Posture 58 (2017) 481–486.

[5] J.M. Hausdorff, G. Yogev, S. Springer, E.S. Simon, N. Giladi, Walking is more like catching than tapping: gait in the elderly as a complex cognitive task, Exp. Brain Res. 164 (2005) 541–548.

[6] N. Demnitz, P. Esser, H. Dawes, et al., A systematic review and meta-analysis of cross-sectional studies examining the relationship between mobility and cognition in healthy older adults, Gait Posture 50 (2016) 164–174.

[7] N.L. Baker, M.N. Cook, H.M. Arrighi, R. Bullock, Hip fracture risk and subsequent mortality among Alzheimer’s disease patients in the United Kingdom, 1988–2007, Age Ageing 40 (2011) 49–54.

[8] L.M. Allan, C.G. Ballard, D.J. Burn, R.A. Kenny, Prevalence and severity of gait disorders in Alzheimer’s and non-Alzheimer’s dementias, J. Am. Geriatr. Soc. 53 (2005) 1681–1687.

[9] L.M. Allan, C.G. Ballard, E.N. Rowan, R.A. Kenny, Incidence and prediction of falls in dementia: a prospective study in older people, PLoS One 4 (2009) e5521.

[10] Y. Laufer, Effect of age on characteristics of forward and backward gait at preferred and accelerated walking speed, J. Gerontol. A Biol. Sci. Med. Sci. 60 (2005) 627–632.

[11] N.E. Fritz, A.M. Worstell, A.D. Kloos, A.B. Siles, S.E. White, D.A. Kegelmeyer, Backward walking measures are sensitive to age-related changes in mobility and balance, Gait Posture 37 (2013) 593–597.

[12] M.E. Hackney, G.M. Earhart, Backward walking in Parkinson’s disease, Mov. Disord. 24 (2009) 218–223.

[13] C. Sherrington, Z.A. Michaleff, N. Fairhall, et al., Exercise to prevent falls in older adults: an updated systematic review and meta-analysis, Br. J. Sports Med. 51 (2017) 1750–1758.

[14] D. Forbes, S.C. Forbes, C.M. Blake, E.J. Thiessen, S. Forbes, Exercise programs for people with dementia, Cochrane Database Syst. Rev. 4 (2015), CD006489.

[15] J. Wang, J. Xu, R. An, Effectiveness of backward walking training on balance performance: A systematic review and meta-analysis, Gait Posture 68 (2019) 466–475.

[16] A. Toots, H. Littbrand, N. Lindelof, et al., Effects of a high-intensity functional exercise program on dependence in activities of daily living and balance in older adults with dementia, J. Am. Geriatr. Soc. 64 (2016) 55–64.

[17] A. Toots, H. Littbrand, H. Holmberg, et al., Walking aids moderate exercise effects on gait speed in people with dementia: a randomized controlled trial, J. Am. Med. Dir. Assoc. 18 (2017) 227–233.

[18] J. Ohlin, A. Ahlgren, R. Folkesson, et al., The association between cognition and gait in a representative sample of very old people - the influence of dementia and walking aid use, BMC Geriatr. 20 (2020) 34.

[19] M. Schwenk, M. Schmidt, M. Pfisterer, P. Oster, K. Hauer, Rollator use adversely impacts on assessment of gait and mobility during geriatric rehabilitation, J. Rehabil. Med. 43 (2011) 424–429.

[20] M.F. Folstein, S.E. Folstein, P.R. McHugh, “Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician, J. Psychiatr. Res. 12 (1975) 189–198.

[21] American Psychiatric Association, Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision, American Psychiatric Association, Washington, DC, 2000.

[22] S. Katz, A.B. Ford, R.W. Moskowitz, B.A. Jackson, M.W. Jaffe, Studies of illness in the aged. The index of Adl: a standardized measure of biological and psychosocial function, JAMA 185 (1963) 914–919.

[23] J.M. Guralnik, L. Ferrucci, C.F. Pieper, et al., Lower extremity function and subsequent disability: consistency across studies, predictive models, and value of gait speed alone compared with the short physical performance battery, J. Gerontol. A Biol. Sci. Med. Sci. 55 (2000) M221–231.

[24] C. Collin, D.T. Wade, S. Davies, V. Horne, The Barthel ADL Index: a reliability study, Int. Disabil. Stud. 10 (1988) 61–63.

[25] H. Littbrand, N. Lindelof, E. Rosendahl, The HIFE Program: The High-Intensity Functional Exercise Program, 2nd ed. Umeå, ISBN 978-91-7601-166-9, Umeå University, Department of Community Medicine and Rehabilitation, Geriatric Medicine, 2014.

[26] A. Sondell, E. Rosendahl, Y. Gustafson, N. Lindelof, H. Littbrand, The applicability of a high-intensity functional exercise program among older people with dementia living in nursing homes, J. Geriatr. Phys. Ther. (2018).

[27] A.M. Baldan, S.R. Alouche, I.M. Araujo, S.M. Freitas, Effect of light touch on postural sway in individuals with balance problems: a systematic review, Gait Posture 40 (2014) 1–10.

[28] K.S. Rawson, M.E. McNeely, R.P. Duncan, K.A. Pickett, J.S. Perlmutter, G. M. Earhart, Exercise and parkinson disease: comparing tango, treadmill, and stretching, J. Neurol. Phys. Ther. 43 (2019) 26–32.

[29] D.K. Rose, L. DeMark, E.J. Fox, D.J. Clark, P. Wludyka, A backward walking training program to improve balance and mobility in acute stroke: a pilot randomized controlled trial, J. Neurol. Phys. Ther. 42 (2018) 12–21.

[30] E.W. Telenius, K. Engedal, A. Bergland, Effect of a high-intensity exercise program on physical function and mental health in nursing home residents with dementia: an assessor blinded randomized controlled trial, PLoS One 10 (2015), e0126102.

A. Toots et al.