570377
research-article2015
CRE0010.1177/0269215515570377Clinical RehabilitationLanders et al.
CLINICAL REHABILITATION
Article
Does attentional focus during balance training in people with Parkinson’s disease affect outcome? A randomised controlled clinical trial
Clinical Rehabilitation 2016, Vol. 30(1) 53–63 © The Author(s) 2015 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0269215515570377 cre.sagepub.com
Merrill R Landers, Rebecca M Hatlevig, Alyssa D Davis, Amanda R Richards and Leslee E Rosenlof
Abstract Objective: To compare the effects of attentional focus to augment balance outcomes in individuals with Parkinson’s disease. Design: Randomised controlled clinical trial. Setting: University gait and balance research laboratory. Participants: Forty-nine individuals with idiopathic Parkinson’s disease. Interventions: Participants were randomly assigned into one of four groups (three balance intervention groups and one control). The three intervention groups all received the same 4-week balance training program augmented with either external, internal, or no focus instructions. The control group did not receive any balance training. Main measures: Outcomes were measured at baseline, post intervention, 2-weeks post intervention, and 8-weeks post intervention and included: Sensory Organization Test, Berg Balance Scale, Self-Selected Gait Velocity, Dynamic Gait Index, Activities-Specific Balance Confidence Scale, and obstacle course completion time. Results: There were no differences among the groups in trajectory over the course of the trial for all outcomes (ps ⩾ .135). All groups improved from baseline to post intervention and from baseline to 2-weeks post intervention for all outcomes (ps ⩽ .003), except Self-Selected Gait Velocity, which did not change over the course of the trial (P = .121). Conclusions: Attentional focus instructions to augment a 4-week balance training program did not result in any change over and above a control group in measures of gait and balance in individuals with Parkinson’s disease. Additionally, while all four groups improved, there was no difference among the groups, including the control, suggesting that the 4-week balance training program in this trial was not effective.
Department of Physical Therapy, University of Nevada Las Vegas, USA
Corresponding author: Merrill R Landers, Department of Physical Therapy, University of Nevada, Las Vegas, NV, 4505 Maryland Parkway Box 453029, Las Vegas, Nevada, 89154-3029, USA. Email: [email protected]
Downloaded from cre.sagepub.com at University Library Utrecht on March 23, 2016
54
Clinical Rehabilitation 30(1)
Keywords Parkinson disease, attentional focus, postural instability, motor learning, balance Received: 16 March 2014; accepted: 28 December 2014
Introduction There are few treatment options for postural instability in Parkinson’s disease. Pharmacotherapy and deep brain stimulation have been shown to provide only modest benefit.1 However, exercise and physiotherapy have been shown in systematic reviews to provide some benefit in improving postural stability.2,3 A recent meta-analysis concludes that physiotherapy, specifically highly challenging balance training, may improve balance performance;4 however, many balance treatment approaches and parameters (e.g. intensity, frequency, duration) have not been sufficiently vetted. One such training strategy that has shown some potential is attentional focus instruction to improve motor performance. Because individuals with Parkinson’s disease are so amenable to cueing,5 an attentional focus strategy during balance training may improve balance training in those with Parkinson’s disease. Studies have shown that an external focus of attention during a motor task (i.e. attending ones attention to the effects of a motor task) is more effective than either no attentional focus or an internal focus of attention during the task (i.e. focusing attention to one’s body during a motor task). It is theorized that an external focus of attention reduces cognitive demands during an evolving movement and results in the utilization of more automatic motor processing.6 Adoption of an external focus of attention during a motor task would allow those with cognitive impairment, which is a prominent non-motor feature of Parkinson’s disease,7,8 to utilize automatic processing without placing high demands on already heavily taxed cognitive pathways.6 The bulk of the research on external focus of attention has mostly focused on sports performance; however, several studies have shown that attentional focus can positively impact balance performance.9–12 Chiviacowsky et al. also found that balance learning could be enhanced in older adults by encouraging them to adopt an external focus while balancing on a stabilometer.13
Adopting an external focus of attention during balance tasks has also been shown to improve postural stability in individuals with Parkinson’s disease.14,15 In these two studies, this benefit in balance performance was observed in close temporal sequence from the external focus instructions. However, these studies only reported the immediate impact of attentional focus instructions; neither study investigated balance learning or lasting benefit from attentional focus instruction. Therefore, the purpose of this trial was to see if external focus instructions during a balance training program would produce additional benefit to balance and gait outcomes over and above other attentional focus strategies (internal and no focus) and a control group. A secondary purpose of the trial was to evaluate if the 4-week balance training program utilized in this trial was sufficient to drive improvements in outcomes regardless of attentional focus relative to a control group that did not participate in a balance training program.
Methods The overall design was a block-randomised, controlled, factorial designed trial with four different experimental treatment groups (three treatment groups with different attentional focus strategies and one control) being tested before and after a 4-week balance training program. Measurements were taken immediately prior to beginning the 4-week trial, immediately after (4 weeks), and then 2 (6 weeks) and 8 weeks (12 weeks) later. The secondary purpose was to combine the three treatment groups and then compare them to a control. All training and testing took place at the Gait and Balance Laboratory at the University of Nevada, Las Vegas.
Participants Forty-nine community-dwelling individuals with idiopathic Parkinson’s disease (neurologist diagnosed) were recruited under Institutional Review Board
Downloaded from cre.sagepub.com at University Library Utrecht on March 23, 2016
55
Landers et al.
Assessed for eligibility N=96
Analysis
Follow-up
Alloca on
Randomised N = 49
Excluded (n = 47) • Did not meet inclusion or exclusion criteria (n = 43) • No medical clearance (n = 1) • Other reason (n = 3)
Allocated to Group A N = 12
Allocated to Group B N = 13
Allocated to Group C N = 12
Allocated to Group D N = 12
Follow-up at 4 weeks N = 10 1 drop-out 1 medical
Follow-up at 4 weeks N = 12 1 drop-out
Follow-up at 4 weeks N = 11 1 medical
Follow-up at 4 weeks N = 11 1 drop-out
Follow-up at 6 weeks N = 10
Follow-up at 6 weeks N = 12
Follow-up at 6 weeks N = 11
Follow-up at 6 weeks N = 11
Follow-up at 12 weeks N = 10
Follow-up at 12 weeks N = 11 1 drop-out
Follow-up at 12 weeks N = 11 1 medical
Follow-up at 12 weeks N = 10 1 drop-in
Intent-to-treat analysis N = 12 Per protocol analysis N = 10
Intent-to-treat analysis N = 13 Per protocol analysis N = 11
Intent-to-treat analysis N = 12 Per protocol analysis N = 10
Intent-to-treat analysis N = 12 Per protocol analysis N = 10
Figure 1. Flow diagram of participant recruitment, allocation and analysis.
(University of Nevada, Las Vegas) approval using advertisement through local Parkinson’s disease support groups and print media. Participants were excluded from the study if they were non-ambulatory or if significant comorbidities were present (e.g. stroke, total hip/knee replacement). They were also excluded if they had a history of surgical intervention for Parkinson’s disease. Participants were instructed to maintain their routine medication schedule and participated in the testing and interventions during peak “ON” periods of the medication regimen if receiving dopaminergic therapy. Figure 1 illustrates the subject recruitment, allocation, and analysis.
Procedures Participants were screened for eligibility and then randomly assigned by a research assistant into one of four groups (A=balance training + external focus instructions, B=balance training + internal focus instructions, C=balance training + no attentional focus instructions, or D=control) using block randomization with a block size of four and a 1:1:1:1 allocation ratio. A random numbers table was used to generate pre-filled envelopes specifying the group. The three treatment groups (A, B, and C) participated in a 4-week balance treatment
Downloaded from cre.sagepub.com at University Library Utrecht on March 23, 2016
56
Clinical Rehabilitation 30(1)
program (various static and dynamic balance tasks), and the control group (D) received no balance training during this period and were simply instructed to maintain their normal daily regimen for the 4 week trial (Appendix A). Participants in Groups A-C trained three times per week, about 45 minutes per day, for 4 weeks. Each treatment session for groups A, B and C consisted of the following three elements, all of which were based on evidence-based principles (Appendix A): 10 minutes of treadmill training without holding onto the railing so as to challenge balance,16 10 minutes of obstacle course negotiation,17,18 and 10 minutes of balance training on a compliant surface in a harness (tandem stance, narrow support stance, single leg stance, eyes closed and external perturbations).4 External perturbations consisted of up to five kilograms of expected and unexpected nudging. The obstacle course consisted of stepping over three obstacles that were 12.7 centimeters tall, walking on a balance beam forward, backward, and side-stepping, weaving through five cones, and finishing by turning 180 degrees to start the course again. This course was repeated three times, which took approximately five minutes to complete; a 30 second rest period was given, and the five minute routine was repeated, for a total of 10 minutes spent on the obstacle course. For safety reasons, all participants performed each of these training tasks in an overhead, non-deweighting harness that would prevent a fall to the floor. This harness had considerable slack in the system to allow the participant to experience loss of balance and subsequent automatic postural responses with only minimal feedback or interference from the harness. Each of the training tasks were manipulated for the varying balance capabilities of the participants. While the balance tasks were the same for all participants, the challenge of each task was adjusted for each participant. The overarching goal of this approach was to ensure that each participant was being challenged;4 that is, the tasks were individually tailored to be neither too easy nor too hard. For example, we had several different rocker boards that allowed different levels of lateral
excursion. Once the easiest rocker board (least excursion) was mastered, the subject was moved to the next most challenging rocker board. Due to the nature of the instructions in the trial, the research assistants who supervised the training and outcomes assessment were not blind to the treatment conditions. Likewise, participants were not fully blinded to condition either. Since three groups were getting exercise and one was not, it was self-evident in a general sense (treatment versus control) to which group the participants had been allocated. However, participants were not informed of the specific attentional focus group assignment for groups A, B, and C. Because the participants were not informed of the specific hypotheses of the study, they were blinded to that allocation. Additionally, since the balance training was the same for groups A, B, and C, participants had no way of knowing to which of the three treatment groups they were assigned.
Outcome measures Outcomes were measured at baseline, immediately post intervention (4 weeks), 2-weeks post intervention (6 weeks) and 8-weeks post intervention (12 weeks). At each measurement session, the following were evaluated (see Appendix B for evidence of reliability and validity): Sensory Organization Test,19 Berg Balance Scale,20 Self-Selected Gait Velocity,21 Dynamic Gait Index,22 and Activitiesspecific Balance Confidence Scale.23 The Sensory Organization Test was measured using the NeuroCom Smart® Balance Master system (Balance Master) (NeuroCom®, a division of Natus®, Clackamas, Oregon, USA). The Sensory Organization Test was completed using computerized dynamic posturography (NeuroCom SMART EquiTest®). The Sensory Organization Test is a score of the participant’s ability to perform different static balance conditions on a force plate by selectively taxing the three main components of balance (vision, vestibular information, and proprioception). The Sensory Organization Test score is a composite of the balance conditions and is expressed in terms of percentage ranging from zero (large sway) to 100 (small sway). The Berg Balance
Downloaded from cre.sagepub.com at University Library Utrecht on March 23, 2016
57
Landers et al. Scale is a performance-based assessment of static and dynamic balance consisting of 14 tasks used to determine the participant’s ability to balance in different functional situations. Scores range from zero to 56 with higher scores indicative of better balance performance. The Self-Selected Gait Velocity is the velocity at which the participant comfortably walks 10 meters. The Dynamic Gait Index is a performance-based assessment of eight dynamic walking tasks. The participant is graded on a scale of zero to 24, with higher scores indicating higher function. The Activities-specific Balance Confidence Scale is a self-reported measure of confidence with various daily functional tasks that require balance and is scored from zero to 100 with a higher score suggesting more confidence. Obstacle course negotiation, a novel assessment created for this trial, was also timed at each measurement point. The research assistant who measured the outcomes was not blinded to group assignment.
analysis. The intent-to-treat approach included all participants as originally randomised even if they had dropped out of the study. For those who had dropped out of the study, their missing data points were imputed using the last observation carried forward method. The per protocol analysis consisted only of data from the participants who completed the trial, as they were originally randomised. Based on an a priori analysis using 80% power and a moderate effect size (estimated from Landers et al. and Wulf et al. studies), it was estimated that between 17 and 19 participants per group would be necessary to detect an interaction effect. Based on a 10–15% anticipated drop-out rate, 22 participants per group would be allocated. A planned interim futility analysis was conducted at the midway point of this trial to determine if the trial should be continued or not. This analysis was done after the first of two phases of data collection when 10–12 participants per group had completed the trial.
Statistics
Results
All data were analyzed using PASW Statistics Version 18.0 (SPSS Inc, Chicago, Illinois, USA). Level of significance was set at α=0.05. In the primary analysis, the data were analyzed using a 4 (group: A, B, C, and D) X 4 (time: baseline, post intervention, 2-week post intervention, 8-week post intervention) mixed factorial analysis of variance (ANOVA) for each of the 6 outcome measures. A secondary analysis was used to compare all three treatment groups to the control group over time. In this secondary analysis, the three balance training groups (A, B, and C) were combined into one group (Intervention) and compared to the group D (Control). This analysis was a 2 (group: intervention and control) X 4 (time: baseline, post intervention, 2-week post intervention, 8-week post intervention) mixed factorial ANOVA for each of the 6 outcomes measures. All outcome measures (Sensory Organization Test, Berg Balance Scale, Self-Selected Gait Velocity, Dynamic Gait Index, Activities-specific Balance Confidence Scale, obstacle course) were analyzed twice using both an intent-to-treat analysis and a per protocol
Of the 49 participants with Parkinson’s disease that were enrolled in the study, 12 were allocated to Group A, 13 to Group B, 12 to Group C, and 12 to Group D (see Figure 1 for a flow diagram of the enrollment). There were no significant differences among the groups for age, gender, Parkinson’s disease severity (modified Hoehn & Yahr scale),24 cognitive status (Mini-Mental State Exam),25 Unified Parkinson’s Disease Rating Scale,26 and fall history (Table 1). Based on the findings of the pre-planned interim futility analysis at the midtrial point, it was determined to halt the trial as the treatment effect was not sufficiently strong to warrant continued allocation of resources to recruitment and treatment of the second half of participants. The following results are from that interim futility analysis.
Primary analysis: Comparing the four groups over time For means and standard deviations of the four groups over the trial see Table 2. In the intent-to-treat
Downloaded from cre.sagepub.com at University Library Utrecht on March 23, 2016
58
Clinical Rehabilitation 30(1)
Table 1. Participant characteristics of those who completed the trial. Group A (n=10) Age (years) Mean ± SD 72.2 ± 4.4 Gender Male 4 Female 6 Modified Hoehn & Yahr score 1.5 4 2.0 2 2.5 1 3.0 2 4.0 1
Group B (n=11)
Group C (n=10)
Group D (n=10)
Statistical comparison
70.2 ± 4.4
70.1 ± 9.5
74.3 ± 8.8
ANOVA P = 0.413
8 3
7 3
6 4
Chi square P = 0.220
2 0 1 6 2
0 4 3 3 0
1 0 4 4 1
Chi square P = 0.057
Mini-Mental State Exam Mean ± SD 27.6 ± 1.1 26.6 ± 3.9 Unified Parkinson’s Disease Rating Scale Mean ± SD 26.7 ± 13.6 39.4 ± 10.9 Fall status Faller 5 6 Non-faller 7 7
29.6 ± 0.5
28.5 ± 2.1
ANOVA P = 0.056
37.3 ± 8.3
33.3 ± 10.7
ANOVA P = 0.122
analysis of these data, there were no difference among the four groups over the course of the trial for the Sensory Organization Test (P = .135, ηp2 = .094, power = .659), Berg Balance Scale (P = .527, ηp2 = .057, power = .433), Self-Selected Gait Velocity (P = .624, ηp2 = .050, power = .380), Dynamic Gait Index (P = .402, ηp2 = .066, power = .485), Activities-specific Balance Confidence Scale (P = .249, ηp2 = .080, power = .578), and obstacle course (P = .654, ηp2 = .048, power = .330). However, all of the participants when combined together did change over the trial for the Sensory Organization Test (P = .003, ηp2 = .101), Berg Balance Scale (P = .003, ηp2 = .097), Dynamic Gait Index (P = .006, ηp2 = .091), Activities-specific Balance Confidence Scale (P = .002, ηp2 = .114), and obstacle course (P < .001, ηp2 = .189). Post hoc analyses revealed a trend of significant improvement in several outcomes from baseline to post intervention and from baseline to 2-weeks post intervention for all groups regardless of treatment group; overall, no other changes among the rest of the measurement times were observed (Table 3). There was no change over time for all
8 4
6 6
Chi square P = 0.632
participants combined for Self-Selected Gait Velocity (P = .121, ηp2 = .042, power = .500). The results of the per protocol analyses paralleled the intent-to-treat analyses.
Secondary analysis: Comparing intervention to control over time Table 4 illustrates the means and standard deviations for the secondary analysis. In the intentto-treat analysis, there were no differences over the course of the trial when comparing the combined three balance treatment groups (A, B and C) to the control group for the Sensory Organization Test (P =.193, ηp2 = .033, power = .399), Berg Balance Scale (P =.207, ηp2 = .032, power = .392), SelfSelected Gait Velocity (P =.356, ηp2 = .023, power = .290), Dynamic Gait Index (P =.605, ηp2 = .012, power = .161), Activities-specific Balance Confidence Scale (P =.918, ηp2 = .003, power = 074), and obstacle course (P =.675, ηp2 = .010, power = .133). However, all of the participants when combined together did improve over the course of the
Downloaded from cre.sagepub.com at University Library Utrecht on March 23, 2016
59
Landers et al. Table 2. Means and standard deviations for the intent-to-treat analysis of all of the outcome variables for the primary analysis.
Sensory Organization Testa Berg Balance Scale Self-Selected Gait Velocityb Dynamic Gait Index Activity Balance Confidence Scale Obstacle Coursec
Group
Baseline
4 weeks
6 weeks
12 weeks
A B C D A B C D A B C D A B C D A B C D A B C D
66.7 ± 13.4 69.3 ± 12.1 57.9 ± 17.9 68.1 ± 9.3 45.0 ± 8.8 44.7 ± 7.6 45.5 ± 6.6 44.8 ± 4.6 1.22 ± 0.46 1.26 ± 0.40 1.26 ± 0.41 1.18 ± 0.30 17.8 ± 4.6 17.2 ± 3.3 18.0 ± 3.0 19.3 ± 2.3 67.1 ± 29.0 73.1 ± 20.8 71.2 ± 18.5 65.6 ± 18.3 72.6 ± 54.4 61.0 ± 32.1 83.9 ± 59.4 78.4 ± 69.4
68.4 ± 15.1 72.7 ± 14.6 68.5 ± 11.3 65.5 ± 18.9 47.9 ± 6.8 48.0 ± 7.0 47.3 ± 7.3 44.5 ± 7.5 1.27 ± 0.48 1.32 ± 0.36 1.36 ± 0.52 1.19 ± 0.27 19.2 ± 4.8 19.9 ± 3.1 18.6 ± 3.8 19.9 ± 3.4 68.2 ± 26.4 82.8 ± 15.0 79.1 ± 17.5 73.3 ± 11.5 62.3 ± 57.2 43.1 ± 26.8 63.4 ± 52.8 65.3 ± 58.5
67.4 ± 12.8 74.2 ± 13.2 71.1 ± 13.0 71.9 ±10.9 46.8 ± 6.8 47.9 ± 9.4 49.4 ± 5.7 46.3 ± 4.9 1.19 ± 0.41 1.35 ± 0.33 1.35 ± 0.46 1.32 ± 0.32 19.3 ± 5.2 19.7 ± 3.4 19.6 ± 3.3 19.9 ± 2.8 72.6 ± 25.0 77.5 ± 17.9 77.9 ± 20.3 70.1 ± 15.9 59.6 ± 58.2 46.0 ± 37.7 51.8 ± 33.8 55.6 ± 39.4
68.5 ± 13.7 75.1 ± 12.9 67.7 ± 14.0 69.5 ± 14.4 47.4 ± 8.7 46.4 ± 10.0 48.4 ± 8.0 47.6 ± 4.9 1.21 ± 0.43 1.34 ± 0.41 1.35 ± 0.46 1.30 ± 0.37 17.8 ± 7.0 17.9 ± 5.3 19.8 ± 4.9 19.8 ± 2.6 63.8 ± 28.8 81.3 ± 16.0 77.4 ± 20.8 68.6 ± 13.6 60.7 ± 60.2 53.0 ± 59.4 59.1 ± 54.0 56.6 ± 41.1
aSway
on a Scale of 0-100 with a score of 100 representing no sway. per second. cSeconds to complete the course. bMeters
trial for the Sensory Organization Test (P = .034, ηp2 = .061, power = .682), Berg Balance Scale (P = .022, ηp2 = .067, power = .741), Activitiesspecific Balance Confidence Scale (P = .007, ηp2 = .086, power = .835), and obstacle course (P < .001, ηp2 = .156, power = .983); however, Self-Selected Gait Velocity (P = .085, ηp2 = .046, power = .559) and Dynamic Gait Index (P = .105, ηp2 = .044, power = .502) did not change over the trial. Again, the results of the per protocol analyses paralleled the intent-to-treat analyses.
Discussion Our results suggest that attentional focus instructions during a standardized balance training program did not improve balance impairment and
balance activity outcomes in participants with Parkinson’s disease. These results suggest that the immediate attentional focus-driven improvements in balance that were observed in the Landers et al. and Wulf et al. studies did not translate in the present study into any lasting benefit in dynamic balance task performance.15,27 In addition, the standardized balance training intervention, regardless of attentional focus strategy, was not sufficient to drive improvements in balance compared to the control group. In general, all participants, including the control group, demonstrated improved balance performance across most measurements over the course of the trial. The simplest reason why external attentional focus instructions did not cause any improvement in the present trial is that attentional focus instructions
Downloaded from cre.sagepub.com at University Library Utrecht on March 23, 2016
60
Clinical Rehabilitation 30(1)
Table 3. P-values for the Bonferroni pairwise comparisons of only the outcome measures that significantly changed over the course of the trial. Baseline to 4 weeks
Baseline to 6 weeks
Baseline to 12 weeks
Intent to Treat Data (all participants as randomised) Sensory Organization Test .407 .006 .034 Berg Balance Scale .171 .005 .032 Dynamic Gait Index .100 .002 1.00 Activities-specific Balance .002 .082 .537 Confidence Scale Obstacle .001
research-article2015
CRE0010.1177/0269215515570377Clinical RehabilitationLanders et al.
CLINICAL REHABILITATION
Article
Does attentional focus during balance training in people with Parkinson’s disease affect outcome? A randomised controlled clinical trial
Clinical Rehabilitation 2016, Vol. 30(1) 53–63 © The Author(s) 2015 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0269215515570377 cre.sagepub.com
Merrill R Landers, Rebecca M Hatlevig, Alyssa D Davis, Amanda R Richards and Leslee E Rosenlof
Abstract Objective: To compare the effects of attentional focus to augment balance outcomes in individuals with Parkinson’s disease. Design: Randomised controlled clinical trial. Setting: University gait and balance research laboratory. Participants: Forty-nine individuals with idiopathic Parkinson’s disease. Interventions: Participants were randomly assigned into one of four groups (three balance intervention groups and one control). The three intervention groups all received the same 4-week balance training program augmented with either external, internal, or no focus instructions. The control group did not receive any balance training. Main measures: Outcomes were measured at baseline, post intervention, 2-weeks post intervention, and 8-weeks post intervention and included: Sensory Organization Test, Berg Balance Scale, Self-Selected Gait Velocity, Dynamic Gait Index, Activities-Specific Balance Confidence Scale, and obstacle course completion time. Results: There were no differences among the groups in trajectory over the course of the trial for all outcomes (ps ⩾ .135). All groups improved from baseline to post intervention and from baseline to 2-weeks post intervention for all outcomes (ps ⩽ .003), except Self-Selected Gait Velocity, which did not change over the course of the trial (P = .121). Conclusions: Attentional focus instructions to augment a 4-week balance training program did not result in any change over and above a control group in measures of gait and balance in individuals with Parkinson’s disease. Additionally, while all four groups improved, there was no difference among the groups, including the control, suggesting that the 4-week balance training program in this trial was not effective.
Department of Physical Therapy, University of Nevada Las Vegas, USA
Corresponding author: Merrill R Landers, Department of Physical Therapy, University of Nevada, Las Vegas, NV, 4505 Maryland Parkway Box 453029, Las Vegas, Nevada, 89154-3029, USA. Email: [email protected]
Downloaded from cre.sagepub.com at University Library Utrecht on March 23, 2016
54
Clinical Rehabilitation 30(1)
Keywords Parkinson disease, attentional focus, postural instability, motor learning, balance Received: 16 March 2014; accepted: 28 December 2014
Introduction There are few treatment options for postural instability in Parkinson’s disease. Pharmacotherapy and deep brain stimulation have been shown to provide only modest benefit.1 However, exercise and physiotherapy have been shown in systematic reviews to provide some benefit in improving postural stability.2,3 A recent meta-analysis concludes that physiotherapy, specifically highly challenging balance training, may improve balance performance;4 however, many balance treatment approaches and parameters (e.g. intensity, frequency, duration) have not been sufficiently vetted. One such training strategy that has shown some potential is attentional focus instruction to improve motor performance. Because individuals with Parkinson’s disease are so amenable to cueing,5 an attentional focus strategy during balance training may improve balance training in those with Parkinson’s disease. Studies have shown that an external focus of attention during a motor task (i.e. attending ones attention to the effects of a motor task) is more effective than either no attentional focus or an internal focus of attention during the task (i.e. focusing attention to one’s body during a motor task). It is theorized that an external focus of attention reduces cognitive demands during an evolving movement and results in the utilization of more automatic motor processing.6 Adoption of an external focus of attention during a motor task would allow those with cognitive impairment, which is a prominent non-motor feature of Parkinson’s disease,7,8 to utilize automatic processing without placing high demands on already heavily taxed cognitive pathways.6 The bulk of the research on external focus of attention has mostly focused on sports performance; however, several studies have shown that attentional focus can positively impact balance performance.9–12 Chiviacowsky et al. also found that balance learning could be enhanced in older adults by encouraging them to adopt an external focus while balancing on a stabilometer.13
Adopting an external focus of attention during balance tasks has also been shown to improve postural stability in individuals with Parkinson’s disease.14,15 In these two studies, this benefit in balance performance was observed in close temporal sequence from the external focus instructions. However, these studies only reported the immediate impact of attentional focus instructions; neither study investigated balance learning or lasting benefit from attentional focus instruction. Therefore, the purpose of this trial was to see if external focus instructions during a balance training program would produce additional benefit to balance and gait outcomes over and above other attentional focus strategies (internal and no focus) and a control group. A secondary purpose of the trial was to evaluate if the 4-week balance training program utilized in this trial was sufficient to drive improvements in outcomes regardless of attentional focus relative to a control group that did not participate in a balance training program.
Methods The overall design was a block-randomised, controlled, factorial designed trial with four different experimental treatment groups (three treatment groups with different attentional focus strategies and one control) being tested before and after a 4-week balance training program. Measurements were taken immediately prior to beginning the 4-week trial, immediately after (4 weeks), and then 2 (6 weeks) and 8 weeks (12 weeks) later. The secondary purpose was to combine the three treatment groups and then compare them to a control. All training and testing took place at the Gait and Balance Laboratory at the University of Nevada, Las Vegas.
Participants Forty-nine community-dwelling individuals with idiopathic Parkinson’s disease (neurologist diagnosed) were recruited under Institutional Review Board
Downloaded from cre.sagepub.com at University Library Utrecht on March 23, 2016
55
Landers et al.
Assessed for eligibility N=96
Analysis
Follow-up
Alloca on
Randomised N = 49
Excluded (n = 47) • Did not meet inclusion or exclusion criteria (n = 43) • No medical clearance (n = 1) • Other reason (n = 3)
Allocated to Group A N = 12
Allocated to Group B N = 13
Allocated to Group C N = 12
Allocated to Group D N = 12
Follow-up at 4 weeks N = 10 1 drop-out 1 medical
Follow-up at 4 weeks N = 12 1 drop-out
Follow-up at 4 weeks N = 11 1 medical
Follow-up at 4 weeks N = 11 1 drop-out
Follow-up at 6 weeks N = 10
Follow-up at 6 weeks N = 12
Follow-up at 6 weeks N = 11
Follow-up at 6 weeks N = 11
Follow-up at 12 weeks N = 10
Follow-up at 12 weeks N = 11 1 drop-out
Follow-up at 12 weeks N = 11 1 medical
Follow-up at 12 weeks N = 10 1 drop-in
Intent-to-treat analysis N = 12 Per protocol analysis N = 10
Intent-to-treat analysis N = 13 Per protocol analysis N = 11
Intent-to-treat analysis N = 12 Per protocol analysis N = 10
Intent-to-treat analysis N = 12 Per protocol analysis N = 10
Figure 1. Flow diagram of participant recruitment, allocation and analysis.
(University of Nevada, Las Vegas) approval using advertisement through local Parkinson’s disease support groups and print media. Participants were excluded from the study if they were non-ambulatory or if significant comorbidities were present (e.g. stroke, total hip/knee replacement). They were also excluded if they had a history of surgical intervention for Parkinson’s disease. Participants were instructed to maintain their routine medication schedule and participated in the testing and interventions during peak “ON” periods of the medication regimen if receiving dopaminergic therapy. Figure 1 illustrates the subject recruitment, allocation, and analysis.
Procedures Participants were screened for eligibility and then randomly assigned by a research assistant into one of four groups (A=balance training + external focus instructions, B=balance training + internal focus instructions, C=balance training + no attentional focus instructions, or D=control) using block randomization with a block size of four and a 1:1:1:1 allocation ratio. A random numbers table was used to generate pre-filled envelopes specifying the group. The three treatment groups (A, B, and C) participated in a 4-week balance treatment
Downloaded from cre.sagepub.com at University Library Utrecht on March 23, 2016
56
Clinical Rehabilitation 30(1)
program (various static and dynamic balance tasks), and the control group (D) received no balance training during this period and were simply instructed to maintain their normal daily regimen for the 4 week trial (Appendix A). Participants in Groups A-C trained three times per week, about 45 minutes per day, for 4 weeks. Each treatment session for groups A, B and C consisted of the following three elements, all of which were based on evidence-based principles (Appendix A): 10 minutes of treadmill training without holding onto the railing so as to challenge balance,16 10 minutes of obstacle course negotiation,17,18 and 10 minutes of balance training on a compliant surface in a harness (tandem stance, narrow support stance, single leg stance, eyes closed and external perturbations).4 External perturbations consisted of up to five kilograms of expected and unexpected nudging. The obstacle course consisted of stepping over three obstacles that were 12.7 centimeters tall, walking on a balance beam forward, backward, and side-stepping, weaving through five cones, and finishing by turning 180 degrees to start the course again. This course was repeated three times, which took approximately five minutes to complete; a 30 second rest period was given, and the five minute routine was repeated, for a total of 10 minutes spent on the obstacle course. For safety reasons, all participants performed each of these training tasks in an overhead, non-deweighting harness that would prevent a fall to the floor. This harness had considerable slack in the system to allow the participant to experience loss of balance and subsequent automatic postural responses with only minimal feedback or interference from the harness. Each of the training tasks were manipulated for the varying balance capabilities of the participants. While the balance tasks were the same for all participants, the challenge of each task was adjusted for each participant. The overarching goal of this approach was to ensure that each participant was being challenged;4 that is, the tasks were individually tailored to be neither too easy nor too hard. For example, we had several different rocker boards that allowed different levels of lateral
excursion. Once the easiest rocker board (least excursion) was mastered, the subject was moved to the next most challenging rocker board. Due to the nature of the instructions in the trial, the research assistants who supervised the training and outcomes assessment were not blind to the treatment conditions. Likewise, participants were not fully blinded to condition either. Since three groups were getting exercise and one was not, it was self-evident in a general sense (treatment versus control) to which group the participants had been allocated. However, participants were not informed of the specific attentional focus group assignment for groups A, B, and C. Because the participants were not informed of the specific hypotheses of the study, they were blinded to that allocation. Additionally, since the balance training was the same for groups A, B, and C, participants had no way of knowing to which of the three treatment groups they were assigned.
Outcome measures Outcomes were measured at baseline, immediately post intervention (4 weeks), 2-weeks post intervention (6 weeks) and 8-weeks post intervention (12 weeks). At each measurement session, the following were evaluated (see Appendix B for evidence of reliability and validity): Sensory Organization Test,19 Berg Balance Scale,20 Self-Selected Gait Velocity,21 Dynamic Gait Index,22 and Activitiesspecific Balance Confidence Scale.23 The Sensory Organization Test was measured using the NeuroCom Smart® Balance Master system (Balance Master) (NeuroCom®, a division of Natus®, Clackamas, Oregon, USA). The Sensory Organization Test was completed using computerized dynamic posturography (NeuroCom SMART EquiTest®). The Sensory Organization Test is a score of the participant’s ability to perform different static balance conditions on a force plate by selectively taxing the three main components of balance (vision, vestibular information, and proprioception). The Sensory Organization Test score is a composite of the balance conditions and is expressed in terms of percentage ranging from zero (large sway) to 100 (small sway). The Berg Balance
Downloaded from cre.sagepub.com at University Library Utrecht on March 23, 2016
57
Landers et al. Scale is a performance-based assessment of static and dynamic balance consisting of 14 tasks used to determine the participant’s ability to balance in different functional situations. Scores range from zero to 56 with higher scores indicative of better balance performance. The Self-Selected Gait Velocity is the velocity at which the participant comfortably walks 10 meters. The Dynamic Gait Index is a performance-based assessment of eight dynamic walking tasks. The participant is graded on a scale of zero to 24, with higher scores indicating higher function. The Activities-specific Balance Confidence Scale is a self-reported measure of confidence with various daily functional tasks that require balance and is scored from zero to 100 with a higher score suggesting more confidence. Obstacle course negotiation, a novel assessment created for this trial, was also timed at each measurement point. The research assistant who measured the outcomes was not blinded to group assignment.
analysis. The intent-to-treat approach included all participants as originally randomised even if they had dropped out of the study. For those who had dropped out of the study, their missing data points were imputed using the last observation carried forward method. The per protocol analysis consisted only of data from the participants who completed the trial, as they were originally randomised. Based on an a priori analysis using 80% power and a moderate effect size (estimated from Landers et al. and Wulf et al. studies), it was estimated that between 17 and 19 participants per group would be necessary to detect an interaction effect. Based on a 10–15% anticipated drop-out rate, 22 participants per group would be allocated. A planned interim futility analysis was conducted at the midway point of this trial to determine if the trial should be continued or not. This analysis was done after the first of two phases of data collection when 10–12 participants per group had completed the trial.
Statistics
Results
All data were analyzed using PASW Statistics Version 18.0 (SPSS Inc, Chicago, Illinois, USA). Level of significance was set at α=0.05. In the primary analysis, the data were analyzed using a 4 (group: A, B, C, and D) X 4 (time: baseline, post intervention, 2-week post intervention, 8-week post intervention) mixed factorial analysis of variance (ANOVA) for each of the 6 outcome measures. A secondary analysis was used to compare all three treatment groups to the control group over time. In this secondary analysis, the three balance training groups (A, B, and C) were combined into one group (Intervention) and compared to the group D (Control). This analysis was a 2 (group: intervention and control) X 4 (time: baseline, post intervention, 2-week post intervention, 8-week post intervention) mixed factorial ANOVA for each of the 6 outcomes measures. All outcome measures (Sensory Organization Test, Berg Balance Scale, Self-Selected Gait Velocity, Dynamic Gait Index, Activities-specific Balance Confidence Scale, obstacle course) were analyzed twice using both an intent-to-treat analysis and a per protocol
Of the 49 participants with Parkinson’s disease that were enrolled in the study, 12 were allocated to Group A, 13 to Group B, 12 to Group C, and 12 to Group D (see Figure 1 for a flow diagram of the enrollment). There were no significant differences among the groups for age, gender, Parkinson’s disease severity (modified Hoehn & Yahr scale),24 cognitive status (Mini-Mental State Exam),25 Unified Parkinson’s Disease Rating Scale,26 and fall history (Table 1). Based on the findings of the pre-planned interim futility analysis at the midtrial point, it was determined to halt the trial as the treatment effect was not sufficiently strong to warrant continued allocation of resources to recruitment and treatment of the second half of participants. The following results are from that interim futility analysis.
Primary analysis: Comparing the four groups over time For means and standard deviations of the four groups over the trial see Table 2. In the intent-to-treat
Downloaded from cre.sagepub.com at University Library Utrecht on March 23, 2016
58
Clinical Rehabilitation 30(1)
Table 1. Participant characteristics of those who completed the trial. Group A (n=10) Age (years) Mean ± SD 72.2 ± 4.4 Gender Male 4 Female 6 Modified Hoehn & Yahr score 1.5 4 2.0 2 2.5 1 3.0 2 4.0 1
Group B (n=11)
Group C (n=10)
Group D (n=10)
Statistical comparison
70.2 ± 4.4
70.1 ± 9.5
74.3 ± 8.8
ANOVA P = 0.413
8 3
7 3
6 4
Chi square P = 0.220
2 0 1 6 2
0 4 3 3 0
1 0 4 4 1
Chi square P = 0.057
Mini-Mental State Exam Mean ± SD 27.6 ± 1.1 26.6 ± 3.9 Unified Parkinson’s Disease Rating Scale Mean ± SD 26.7 ± 13.6 39.4 ± 10.9 Fall status Faller 5 6 Non-faller 7 7
29.6 ± 0.5
28.5 ± 2.1
ANOVA P = 0.056
37.3 ± 8.3
33.3 ± 10.7
ANOVA P = 0.122
analysis of these data, there were no difference among the four groups over the course of the trial for the Sensory Organization Test (P = .135, ηp2 = .094, power = .659), Berg Balance Scale (P = .527, ηp2 = .057, power = .433), Self-Selected Gait Velocity (P = .624, ηp2 = .050, power = .380), Dynamic Gait Index (P = .402, ηp2 = .066, power = .485), Activities-specific Balance Confidence Scale (P = .249, ηp2 = .080, power = .578), and obstacle course (P = .654, ηp2 = .048, power = .330). However, all of the participants when combined together did change over the trial for the Sensory Organization Test (P = .003, ηp2 = .101), Berg Balance Scale (P = .003, ηp2 = .097), Dynamic Gait Index (P = .006, ηp2 = .091), Activities-specific Balance Confidence Scale (P = .002, ηp2 = .114), and obstacle course (P < .001, ηp2 = .189). Post hoc analyses revealed a trend of significant improvement in several outcomes from baseline to post intervention and from baseline to 2-weeks post intervention for all groups regardless of treatment group; overall, no other changes among the rest of the measurement times were observed (Table 3). There was no change over time for all
8 4
6 6
Chi square P = 0.632
participants combined for Self-Selected Gait Velocity (P = .121, ηp2 = .042, power = .500). The results of the per protocol analyses paralleled the intent-to-treat analyses.
Secondary analysis: Comparing intervention to control over time Table 4 illustrates the means and standard deviations for the secondary analysis. In the intentto-treat analysis, there were no differences over the course of the trial when comparing the combined three balance treatment groups (A, B and C) to the control group for the Sensory Organization Test (P =.193, ηp2 = .033, power = .399), Berg Balance Scale (P =.207, ηp2 = .032, power = .392), SelfSelected Gait Velocity (P =.356, ηp2 = .023, power = .290), Dynamic Gait Index (P =.605, ηp2 = .012, power = .161), Activities-specific Balance Confidence Scale (P =.918, ηp2 = .003, power = 074), and obstacle course (P =.675, ηp2 = .010, power = .133). However, all of the participants when combined together did improve over the course of the
Downloaded from cre.sagepub.com at University Library Utrecht on March 23, 2016
59
Landers et al. Table 2. Means and standard deviations for the intent-to-treat analysis of all of the outcome variables for the primary analysis.
Sensory Organization Testa Berg Balance Scale Self-Selected Gait Velocityb Dynamic Gait Index Activity Balance Confidence Scale Obstacle Coursec
Group
Baseline
4 weeks
6 weeks
12 weeks
A B C D A B C D A B C D A B C D A B C D A B C D
66.7 ± 13.4 69.3 ± 12.1 57.9 ± 17.9 68.1 ± 9.3 45.0 ± 8.8 44.7 ± 7.6 45.5 ± 6.6 44.8 ± 4.6 1.22 ± 0.46 1.26 ± 0.40 1.26 ± 0.41 1.18 ± 0.30 17.8 ± 4.6 17.2 ± 3.3 18.0 ± 3.0 19.3 ± 2.3 67.1 ± 29.0 73.1 ± 20.8 71.2 ± 18.5 65.6 ± 18.3 72.6 ± 54.4 61.0 ± 32.1 83.9 ± 59.4 78.4 ± 69.4
68.4 ± 15.1 72.7 ± 14.6 68.5 ± 11.3 65.5 ± 18.9 47.9 ± 6.8 48.0 ± 7.0 47.3 ± 7.3 44.5 ± 7.5 1.27 ± 0.48 1.32 ± 0.36 1.36 ± 0.52 1.19 ± 0.27 19.2 ± 4.8 19.9 ± 3.1 18.6 ± 3.8 19.9 ± 3.4 68.2 ± 26.4 82.8 ± 15.0 79.1 ± 17.5 73.3 ± 11.5 62.3 ± 57.2 43.1 ± 26.8 63.4 ± 52.8 65.3 ± 58.5
67.4 ± 12.8 74.2 ± 13.2 71.1 ± 13.0 71.9 ±10.9 46.8 ± 6.8 47.9 ± 9.4 49.4 ± 5.7 46.3 ± 4.9 1.19 ± 0.41 1.35 ± 0.33 1.35 ± 0.46 1.32 ± 0.32 19.3 ± 5.2 19.7 ± 3.4 19.6 ± 3.3 19.9 ± 2.8 72.6 ± 25.0 77.5 ± 17.9 77.9 ± 20.3 70.1 ± 15.9 59.6 ± 58.2 46.0 ± 37.7 51.8 ± 33.8 55.6 ± 39.4
68.5 ± 13.7 75.1 ± 12.9 67.7 ± 14.0 69.5 ± 14.4 47.4 ± 8.7 46.4 ± 10.0 48.4 ± 8.0 47.6 ± 4.9 1.21 ± 0.43 1.34 ± 0.41 1.35 ± 0.46 1.30 ± 0.37 17.8 ± 7.0 17.9 ± 5.3 19.8 ± 4.9 19.8 ± 2.6 63.8 ± 28.8 81.3 ± 16.0 77.4 ± 20.8 68.6 ± 13.6 60.7 ± 60.2 53.0 ± 59.4 59.1 ± 54.0 56.6 ± 41.1
aSway
on a Scale of 0-100 with a score of 100 representing no sway. per second. cSeconds to complete the course. bMeters
trial for the Sensory Organization Test (P = .034, ηp2 = .061, power = .682), Berg Balance Scale (P = .022, ηp2 = .067, power = .741), Activitiesspecific Balance Confidence Scale (P = .007, ηp2 = .086, power = .835), and obstacle course (P < .001, ηp2 = .156, power = .983); however, Self-Selected Gait Velocity (P = .085, ηp2 = .046, power = .559) and Dynamic Gait Index (P = .105, ηp2 = .044, power = .502) did not change over the trial. Again, the results of the per protocol analyses paralleled the intent-to-treat analyses.
Discussion Our results suggest that attentional focus instructions during a standardized balance training program did not improve balance impairment and
balance activity outcomes in participants with Parkinson’s disease. These results suggest that the immediate attentional focus-driven improvements in balance that were observed in the Landers et al. and Wulf et al. studies did not translate in the present study into any lasting benefit in dynamic balance task performance.15,27 In addition, the standardized balance training intervention, regardless of attentional focus strategy, was not sufficient to drive improvements in balance compared to the control group. In general, all participants, including the control group, demonstrated improved balance performance across most measurements over the course of the trial. The simplest reason why external attentional focus instructions did not cause any improvement in the present trial is that attentional focus instructions
Downloaded from cre.sagepub.com at University Library Utrecht on March 23, 2016
60
Clinical Rehabilitation 30(1)
Table 3. P-values for the Bonferroni pairwise comparisons of only the outcome measures that significantly changed over the course of the trial. Baseline to 4 weeks
Baseline to 6 weeks
Baseline to 12 weeks
Intent to Treat Data (all participants as randomised) Sensory Organization Test .407 .006 .034 Berg Balance Scale .171 .005 .032 Dynamic Gait Index .100 .002 1.00 Activities-specific Balance .002 .082 .537 Confidence Scale Obstacle .001
