A Systematic Literature Review

[

research report

]

CRISTINE AGRESTA, PT, PhD1 • ALLISON BROWN, PT, PhD2

Journal of Orthopaedic & Sports Physical Therapy® Downloaded from www.jospt.org at on August 1, 2015. For personal use only. No other uses without permission. Copyright © 2015 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

Gait Retraining for Injured and Healthy Runners Using Augmented Feedback: A Systematic Literature Review

R

egular cardiovascular exercise, such as running, has many known physical and psychological benefits, including weight control, decreased hypertension, decreased blood glucose in type 2 diabetes, and improved mood state.29,44 However, running does not come without risk. Lower extremity injury rates have been reported to range from 19% to 79% in runners.34 Risk factors are multifaceted and include

both external factors, such as high or a sudden increase in weekly mileage,33 and internal factors, such as insufTTSTUDY DESIGN: Systematic literature review.

TTOBJECTIVES: This review sought to determine the efficacy of real-time visual and/or auditory feedback for modifying kinematics and kinetics during running gait.

TTBACKGROUND: Real-time visual and auditory

feedback has gained popularity in the clinical and research settings. Rehabilitation time and injury prevention may be improved when clinicians are able to modify running mechanics in a patient population.

TTMETHODS: A thorough search of PubMed,

CINAHL, and Web of Science from 1989 to January 2015 was performed. The search sought articles that examined real-time visual or auditory feedback for the purposes of modifying kinematics or kinetics in injured or healthy runners. Study design and methodological quality were rated using a 20-point scale.

TTRESULTS: Ten studies were identified for inclusion in the review, 2 of high and 8 of moderate

ficient muscle strength, insufficient muscle length,15,33 or abnormal running mechanics.26,40 methodological quality. There was a consensus in the literature that the use of real-time feedback is effective in reducing variables related to ground reaction forces, as well as in positively modifying previously identified risky lower extremity kinematic movement patterns in healthy runners and those with patellofemoral pain and chronic exertional compartment syndrome. No one method of feedback was identified as being superior. Mirror and 2-dimensional video feedback were identified as potential methods for running-gait modification in a clinical setting.

TTCONCLUSION: In conjunction with traditional therapeutic interventions, real-time auditory and visual feedback should be considered for treating injured runners or addressing potentially injurious running mechanics in a healthy population. J Orthop Sports Phys Ther 2015;45(8):576-584. doi:10.2519/jospt.2015.5823 TTKEY WORDS: auditory, biomechanics, injury, mirror, patellofemoral pain, visual

Multiple prospective studies have documented the presence of altered running mechanics in runners who develop running injuries. 6,24,25,35,37 These findings suggest a causal relationship between abnormal running mechanics and subsequent injury. While there is some evidence to support the impairment-based treatment model in treating running-related injuries, a growing body of literature suggests that addressing these altered running mechanics may be of good long-term benefit. 17,20,26 One method of modifying a runner’s mechanics is through visual or auditory feedback. A few studies have investigated the effect of providing such feedback to runners with currently symptomatic patellofemoral pain (PFP). 4,27,41 While these studies have documented improved running mechanics following feedback sessions, more importantly, they have also shown decreased running-related pain and symptoms. These findings suggest that feedback training may be an effective approach to treat running injuries. The aim of this review was to evaluate the efficacy of real-time visual and/or auditory feedback to modify kinematic and kinetic gait patterns that have been associated with running injury.

1 Human Performance Innovation Lab, School of Kinesiology, University of Michigan, Ann Arbor, MI. 2School of Health Related Professions, Rutgers, The State University of New Jersey, Newark, NJ. The authors certify that they have no affiliations with or financial involvement in any organization or entity with a direct financial interest in the subject matter or materials discussed in the article. Both authors contributed equally to this work. Address correspondence to Dr Cristine Agresta, Central Campus Recreation Building, 401 Washtenaw Avenue, Ann Arbor, MI 48109. E-mail: [email protected] t Copyright ©2015 Journal of Orthopaedic & Sports Physical Therapy®

576 | august 2015 | volume 45 | number 8 | journal of orthopaedic & sports physical therapy

45-08 Agresta.indd 576

7/15/2015 3:31:40 PM

TABLE 1

Results of Electronic Database Search PubMed Yield

CINAHL Yield

Web of Science Yield

(gait OR locomotion OR running) AND (feedback OR retraining) AND runners

33

4

56

(feedback OR gait retraining) AND (load OR force OR impact) AND (gait OR running OR locomotion)

171

58

(feedback OR gait retraining) AND (step OR stride OR spatiotemporal) AND (gait OR running OR locomotion)

178

42

(running OR gait) AND (retraining OR feedback) AND (visual OR step rate)

147

51

(feedback OR gait retraining) AND (kinematic OR kinetic) AND (gait OR running OR locomotion)

204

30

Total yield

733

185


Key Word Search

56

*Values are n. Total yield across databases, n = 974.

METHODS Search Strategy

P

ubMed, CINAHL, and Web of Science were searched from 1989 to January 2015 for peer-reviewed publications that investigated the effect of feedback training during running. Searches were limited to articles in English with human subjects 18 years and older. Key words and search terms included the following: feedback, gait retraining, kinematic, kinetic, spatiotemporal, stride, step, gait, running, locomotion, load, force, and impact. The sequence with search retrievals is summarized in TABLE 1. For the purpose of this review, “feedback” refers to augmented or extrinsic feedback. Augmented feedback is defined as information from an external source, provided to the runner regarding his or her individual performance relative to the desired performance.31 The runner intentionally uses this information (feedback) to make corrections to his or her running gait to achieve a targeted performance parameter. Two reviewers (C.A. and A.B.) independently screened articles based on titles and abstracts. Studies were accepted or excluded based on inclusion and exclusion criteria. Following this first screening, each reviewer

independently retrieved appropriate articles and a full-text screen was performed. Following the full-text screen, there was discussion by the reviewers until an agreement was reached on the final list of included articles. The FIGURE shows a flow diagram summarizing the selection process. In January 2015, a second search was conducted to confirm that no additional articles had been published that met the inclusion/exclusion criteria.

Inclusion and Exclusion Criteria To be considered for inclusion, studies had to include interventions that utilized feedback training during running; to report on kinematic, kinetic, muscle electromyography, or spatiotemporal variables; and to focus on the lower extremity. Studies were excluded if they reported on interventions that included running-gait retraining without feedback on task performance; participants with prosthetic limbs, neurological impairments, or congenital impairments; feedback during tasks other than running; and children or participants under 18 years of age.

Assessment of Methodological Quality The methodological quality of each study was independently rated by the 2 review-

ers using a 20-point scale proposed by Burns and Miller.3 This appraisal tool consists of 16 items focused on study design, experimental control, study participants, methodology, and outcomes. The first item, addressing study design, is worth 5 points. Each additional item is worth 1 point, for a maximum score of 20. Once scored, studies were assigned a quality rating of high (score range, 1420), medium (score range, 7-13), or low (score range, 1-6), based on their final scores. Articles with disagreement in quality scores between raters were reread and discussed to reach consensus.

Data Extraction and Analysis Each author independently reviewed the articles and extracted the following data: study population, number of participants in each group, participant demographics, intervention protocol, method of feedback delivery, follow-up, outcomes, and statistical significance. Due to heterogeneity of study design and outcome measures, a meta-analysis could not be performed. Data analysis focused on the gait parameters targeted for modification by the feedback training. These parameters included hip and pelvis kinematics, joint moments, spatiotemporal parameters, and ground reaction forces.

RESULTS Study Selection

T

he initial literature search yielded a total of 974 potential articles. Nine hundred fifty-five articles were removed after screening titles and abstracts and removing duplicates. Three articles were added from hand searches, which left a total of 22 articles that were read independently by the 2 reviewers for content and quality. After reading the full-text articles, 9 studies were excluded for lack of focus on feedback training as an intervention or its influence on running characteristics.1,2,5,16,18,19,30,42,43 Another paper was excluded because it provided data on the same population as the original article.39 Finally, 2 pa-

journal of orthopaedic & sports physical therapy | volume 45 | number 8 | august 2015 | 577


7/15/2015 3:31:41 PM


[ pers were excluded for poorly defined participant characteristics and protocol.10,21 Thus, a total of 10 articles were considered appropriate and included for review.4,7-9,11-13,22,27,41 A summary of study characteristics and outcomes is provided in TABLE 2. Of the 10 studies included in this review, only 2 were classified as being of high quality, whereas the remaining 8 were of medium quality. The individual item scores and total quality assessment score for each study are summarized in TABLE 3. All studies lost quality points for a lack of assessor or participant blinding. With only 2 randomized controlled trials in the group, many studies lost quality points for study design as well as for lack of an experimental control group.

Feedback Retraining Protocol Feedback schedule, running duration, and the duration of training varied across studies. Four of the 10 studies used the same feedback schedule for training,4,8,27,41 with feedback given consistently for 4 sessions, then gradually withdrawn over 4 sessions. In these 4 studies, participants were not permitted to run outside of gait-retraining sessions. Clansey et al7 provided 20 minutes of continuous feedback during all six 35-minute running sessions. Two studies11,12 used videotape feedback, a digital metronome to increase step rate, and verbal feedback to reduce the tendency to heel strike upon ground contact. Verbal feedback was offered during the 25- to 30-minute running sessions following a series of drills aimed at improving running mechanics (see Diebal et al12 for details on running drills). Messier and Cirillo22 provided concurrent visual and verbal feedback during minutes 1 and 10 of a 20-minute training session. The remaining 2 studies investigated the immediate effects of feedback using a single training session.9,13 Crowell et al9 used a single 30-minute running session divided into warm-up, feedback, no-feedback, and cool-down periods. Eriksson et al13 tested participants across 11 trials, while providing nonsimultaneous auditory or

research report

]

Records identified through database search, n = 974 • PubMed, n = 733 • Web of Science, n = 56 • CINAHL, n = 185

Records identified through hand search, n = 3

Records excluded after screening titles, abstracts, and duplicates, n = 955 Full-text articles assessed for eligibility, n = 22

Full-text articles excluded, n = 12 • Feedback training was not primary intervention, n = 91,2,5,16,18,19,30,42,43 • Sample population duplicated in another study, n = 139 • Poorly defined subject characteristics and protocols, n = 210,21

Studies included in the review, n = 104,7-9,11-13,22,27,41 FIGURE. Study selection flow diagram.

visual feedback. Selected gait parameters, in isolation or in combination, were targeted during each trial. All but 17 of the 10 studies included instructions from the experimenters for a “strategy or way to run” to achieve the desired outcome, and all but 14 of the 10 studies incorporated both visual and auditory feedback information during the training sessions. Nine of the 10 studies utilized visual feedback, 5 in the form of a visual representation of the metric to change,7-9,13,27 1 using a fulllength mirror,41 and 3 using video of the participant.11,12,22

Outcome Variables Noehren et al27 demonstrated a significant overall reduction in peak hip adduction angle (HADD) and peak contralateral pelvic drop following feedback training in runners with PFP. Although it was not statistically significant, peak hip internal rotation was also reduced. Modifications for HADD and contralateral pelvic drop persisted at the 1-month follow-up. Also, in runners with PFP, Willy et al41 found significant reductions in stance-phase peak HADD, peak contra-

lateral pelvic drop, and peak hip internal rotation after mirror feedback training. Only changes in HADD remained significant at the 1-month follow-up. Following 6 weeks of training to increase step rate and change to a forefoot strike pattern, runners with chronic exertional compartment syndrome (CECS) significantly increased step frequency, decreased step length, and reduced ground reaction forces (GRFs).11,12 After 8 feedback training sessions, Cheung and Davis4 demonstrated reductions in GRF that were maintained at the 3-month follow-up. During a single feedback session, the 5 participants in the Crowell et al9 study were able to make considerable reductions in the magnitude of the GRFs. In a similar study by Crowell and Davis,8 GRFs were reduced after feedback training, and these changes persisted at the 1-month follow-up. Similarly, Clansey et al7 also reported a reduction in the magnitude of the GRFs after training; however, all measures returned to baseline values at the 1-month follow-up. In addition, Clansey et al7 found significant differences in ankle and foot-strike angle after runners received visual and auditory

578 | august 2015 | volume 45 | number 8 | journal of orthopaedic & sports physical therapy


7/15/2015 3:31:42 PM

TABLE 2 Participant Characteristics

Control Group

Gait Parameter to Change

Outcome Measures

Cheung and Davis4

3 females with PFP; age range, 26-32 y; 10-30 km/wk

No

Dosage: 8 sessions over 2 wk, 15to 30-min sessions, SSS Instruction/strategy: verbal (“shorten stride length, avoid RFS landing”) Feedback: auditory sensor in heel of Pedar insole

Foot strike

VIP, VALR, VILR

VIP preintervention: 1.3  0.1 BW, 1.7  0.2 BW, 1.7  0.1 BW; postintervention: 1.4  0.1 BW, 1.5  0.1 BW, 1.1  0.1 BW VALR preintervention: 48.2  6.9 BW/s, 55.8  7.6 BW/s, 41.4  8.5 BW/s; postintervention: 37.7  7.7 BW/s, 38.7  8.3 BW/s, 35.2  8.6 BW/s VILR preintervention: 69.6  7.0 BW/s, 72.3  7.9 BW/s, 41.4  8.5 BW/s; postintervention: 37.7  7.7 BW/s, 38.7  8.3 BW/s, 35.2  8.6 BW/s

Clansey et al7

3 healthy male RFS; mean age, 33 y; 30 km/wk

10 matched controls

Dosage: 6 sessions over 3 wk, 20 min, 3.7 m/s Feedback: visual (traffic symbol), auditory (pitch)

GRF

VIP, VALR, VILR

PTA preintervention, 10.67  1.85 g; postintervention, 7.39  1.48 g† VALR preintervention, 66.54  17.45 BW/s; postintervention, 54.62  14.41 BW/s† VILR preintervention, 113.87  33.01 BW/s; postintervention, 92.10  27.06 BW/s† VIP preintervention, 2.72  0.17 BW; postintervention, 2.62  0.16 BW

Crowell et al9

5 healthy females; mean age, 26 y; 32 km/wk

No

Dosage: 1 session, 30 min, SSS Instruction/strategy: verbal (“run softer”) at beginning of first retraining session Feedback: visual (graph of PTA)

GRF

VIP, VALR, VILR

PTA preintervention, +6% g; postintervention, –60% g† VIP preintervention, –6% BW; postintervention, –30% BW† VALR preintervention, –19% BW/s; postintervention, –39% BW/s† VILR preintervention, –15% BW/s; postintervention, –39% BW/s

Crowell and Davis8

10 healthy RFS (6 male, 4 female); 16 km/wk

No

Dosage: 8 sessions over 2 wk, 15to 30-min sessions, SSS Instruction/strategy: verbal (“run softer”) at beginning of first retraining session Feedback: visual (graph of PTA)

GRF

VIP, VALR, VILR

PTA preintervention to postintervention difference, –48% g† VIP preintervention to postintervention difference, –19% BW VALR preintervention to postintervention difference, –32% BW/s† VILR preintervention to postintervention difference, –34% BW/s†

Diebal et al11

2 RFS; CECS; age, 21 y

No

Dosage: three 30-min sessions per week for 6 wk Feedback: visual (videotape of runner), verbal (“run quietly”; explanation of proper running technique)

Foot strike, step rate

Step length, step rate, peak VGRF, impulse

Step frequency: preintervention, 2.74 and 2.66 steps/s; postintervention, 2.99 and 3.37 steps/s Step length: preintervention, 1.15 and 1.21 m; postintervention, 1.05 and 1.02 m Peak VGRF: preintervention, 2.68 and 2.42 BW; postintervention, 2.44 and 2.21 BW Impulse: preintervention, 193.64 and 382.10 ns; postintervention, 183.56 and 314.19 ns

Study


Study Characteristics and Outcomes

Protocol

Outcomes*

Table continues on page 580.

feedback focusing on peak tibial acceleration, with the runners demonstrating a more plantar-flexed ankle position during early stance. After 5 weeks of feedback training targeting running mechanics, Messier and Cirillo22 found that runners responded favorably to the desired modifications set by the investigators. These modifica-

tions included a significantly more dorsiflexed ankle position during early stance (P