BMJ Open Sport Exerc Med: first published as 10.1136/bmjsem-2017-000256 on 21 September 2017. Downloaded from http://bmjopensem.bmj.

com/ on March 28, 2021 by guest. Protected

Open Access Review

Verbal augmented feedback in the

rehabilitation of lower extremity
musculoskeletal dysfunctions: a
systematic review
Marianne Storberget,1 Linn Helen J Grødahl,1 Suzanne Snodgrass,2
Paulette van Vliet,2 Nicola Heneghan3

To cite: Storberget M, ABSTRACT

Grødahl LHJ, Snodgrass S, What is already known?
Background Verbal augmented feedback (VAF) is
et al. Verbal augmented
commonly used in physiotherapy rehabilitation of
feedback in the rehabilitation " Verbal augmented feedback is commonly used
individuals with lower extremity musculoskeletal
of lower extremity for exercise prescription; however, its effective-
musculoskeletal dysfunction or to induce motor learning for injury ness is unknown.
dysfunctions: a systematic prevention. Its effectiveness for acquisition, retention " Exercise with an external focus is more effective
review. BMJ Open Sport and transfer of learning of new skills in this population for motor learning in musculoskeletal conditions,
Exerc Med 2017;3:e000256. is unknown. but it is unclear which modes are most beneficial
doi:10.1136/bmjsem-2017- Objectives First, to investigate the effect of VAF for
000256 for motor learning.
rehabilitation and prevention of lower extremity
musculoskeletal dysfunction. Second, to determine its
Accepted 17 July 2017
effect on motor learning and the stages of acquisition,
retention and transfer in this population.
Design Systematic review designed in accordance What are the new findings?

by copyright.
with the Centre for Reviews and Dissemination and
reported in line with Preferred Reporting Items for " There is moderate evidence that verbal
Systematic Review and Meta-analysis. augmented feedback is effective in the rehabilita-
Method MEDLINE, Embase, PubMed and five tion of musculoskeletal lower limb injuries.
" Future studies should evaluate outcomes relating
additional databases were searched to identify primary
studies with a focus on VAF for prevention and to retention and transfer to evaluate achievement
rehabilitation of lower extremity musculoskeletal of motor learning.
dysfunction. One reviewer screened the titles and
abstracts. Two reviewers retrieved full text articles for
final inclusion. The first reviewer extracted data, extremity biomechanics and postural control have been
whereas the second reviewer audited. Two reviewers identified. This suggests that further research into this
independently assessed risk of bias and quality of topic is warranted where an investigation of long-term
evidence using Cochrane Collaboration’s tool and effects of interventions is required. All stages
Grading of Recommendations Assessment, (acquisition, retention and transfer) should be
1
Development and Evaluation, respectively. evaluated.
School of Sport, Exercise
Results Six studies were included, with a total sample
and Rehabilitation Sciences,
University of Birmingham,
of 304 participants. Participants included patients with
Birmingham, UK lateral ankle sprain (n=76), postoperative ACL INTRODUCTION
2
School of Health Sciences reconstruction (n=16) and healthy individuals in injury It is estimated that 22 million sports-related
(Physiotherapy), University prevention (n=212). All six studies included injuries occur annually in the UK alone,
of Newcastle, Callaghan, New acquisition, whereas retention was found in five
South Wales, Australia with the knee and ankle being common
studies. Only one study examined transfer of the
3
Centre of Precision injury sites.1 According to Murphy,
achieved motor learning (n=36). VAF was found to be
Rehabilitation for Spinal Connolly and Beynnon (2003), sports-
Pain, School of Sport,
effective for improving lower extremity biomechanics
and postural control with moderate evidence from five
related injuries are globally estimated to
Exercise and Rehabilitation
studies. account for $1 billion per year in medical,
Sciences, University of
Birmingham, Birmingham, Conclusion VAF should be considered in the sick leave and management costs. Rehabili-
UK rehabilitation of lower extremity musculoskeletal tation and prevention of these
dysfunctions. However, it cannot be unequivocally musculoskeletal injuries constitute a signifi-
Correspondence to confirmed that VAF is effective in this population, cant part of physiotherapy workload, and
Marianne Storberget; owing to study heterogeneity and a lack of high-quality societal and economic costs are considerable
mstorberget@hotmail.com evidence. Nevertheless, positive effects on lower given many are of a working age.2

Storberget M, et al. BMJ Open Sport Exerc Med 2017;3:e000256. doi:10.1136/bmjsem-2017-000256 1

 BMJ Open Sport Exerc Med: first published as 10.1136/bmjsem-2017-000256 on 21 September 2017. Downloaded from http://bmjopensem.bmj.com/ on March 28, 2021 by guest. Protected
Open Access

Figure 1 Flow chart illustrating search process and identification of studies.

by copyright.
Exercise prescription is integral to physiotherapy instruments, a balance board or external verbal instruc-
rehabilitation and prevention of musculoskeletal inju- tions or cues.11 12
ries. As well as changes in skeletal muscle structure,3 Evidence suggests that AF is effective for motor
exercise induces motor learning if done with sufficient learning achievement12 13; however, high-quality
repetition.4 5 Motor learning is defined as ‘a set of research is lacking. AF comprises a range of modes, in
processes associated with practice or experience general content, timing and focus of attention. In
leading to relatively permanent changes in the capa- terms of focus of attention, verbal augmented
bility for producing skilled action’ (p22). The process feedback (VAF) can either be delivered where the
of motor learning is broken down into three key feedback is focused towards the body or the body part
stages: (1) acquisition: the initial stage of learning a (internal) or where the movement’s effect on the envi-
new skill, for example, performance of exercise, ronment is the focus (external).9 11 14 A recent
(2) retention: evidence of skill achievement after cessa- systematic review concluded that exercise with an
tion of exercise and (3) transfer: the ability to perform external focus is more effective for motor learning in
the attained skill in a different motor task, including musculoskeletal conditions9; however, it is unclear
activities of daily life.5 which modes of VAF may be most beneficial to induce
An essential part of motor learning is neuroplasticity, motor learning and in turn enhance the effectiveness
which is the potential for the nervous system to change of injury rehabilitation and prevention.
in response to sensory information.6 While much work VAF is already widely used in practice during
has been done on motor learning in neurological retraining of individuals with lower limb injuries
conditions such as strokes,7–9 there has been less of a having the advantage of not requiring costly equip-
focus on the unimpaired brain, despite the healthy ment. However, in the absence of robust evidence, an
brain having a greater potential for change.9 10 Clini- investigation of verbal feedback as a form of AF on
cally, to achieve the desired outcomes such as motor motor learning is required to underpin clinical prac-
learning, feedback on performance of exercises is tice. Additionally, a closer look at retention and
required. To enhance motor learning, intrinsic and transfer tests of newly gained motor skills would be of
extrinsic approaches are advocated,11 where the high interest. Until now, no systematic reviews have
former is mediated through an individual’s sensory investigated the effect of VAF on musculoskeletal
system and the latter, also termed augmented feedback dysfunctions. Therefore, the main objective of this
(AF), involves an external source, such as biofeedback study was to determine the effect of VAF in the

2 Storberget M, et al. BMJ Open Sport Exerc Med 2017;3:e000256. doi:10.1136/bmjsem-2017-000256

 BMJ Open Sport Exerc Med: first published as 10.1136/bmjsem-2017-000256 on 21 September 2017. Downloaded from http://bmjopensem.bmj.com/ on March 28, 2021 by guest. Protected
Open Access

METHODS
Table 1 MEDLINE search strategy: Ovid MEDLINE(R) Protocol and registration
1946 to December week 4 2016 A protocol was developed in accordance with Preferred
Search term Reporting Items for Systematic Review and Meta-anal-
ysis protocols (PRISMA-P) 2015 statement and
1 Feedback/
registered with International Prospective Register of
2 Motor learning.mp Systematic Reviews (CRD42016035349). The actual
3 Augmented feedback review is reported in accordance with the PRISMA
statement, and the Cochrane Handbook for Systematic
4 attentional focus.mp.
Reviews of Interventions was used to inform the
5 focus of attention.mp. conduct of the study.
6 injury.mp.
7 jump.mp.
Eligibility criteria
The search strategy was informed through consultation
8 landing.mp. with individuals with subject and methodological
9 biomechanics.mp. expertise (NH, SS, PvV) and following a scoping
10 Ankle/ or ankle.mp.
search. It was devised to answer the question and
framed in accordance with patient,intervention,
11 sprain.mp. comparison, outcome, study design (PICOS).
12 exp Anterior Cruciate Ligament/
13 kinematics.mp. Participants
Studies containing participants with a musculoskeletal
14 Transfer.mp.
dysfunction/injury or healthy subjects at risk of devel-
15 Acquisition.mp. oping a lower extremity musculoskeletal injury (injury
16 retention.mp. prevention) were included.
17 exp Learning/
Intervention

by copyright.
18 extrinsic feedback.mp.
Studies with the aim of assessing the effect of VAF and
19 verbal feedback.mp. investigating the effect of focus of attention (internal or
20 Instruction.mp. external) as way of providing VAF were included.11
However, articles focusing on video feedback and
21 Ground reaction force.mp.
general instructions were excluded. Instructions were
22 exp Kinetics/ considered as general if they were not focused on
23 exp Lower Extremity/ biomechanics of the lower limbs, for example, instruc-
tions that were not intended as the intervention.
24 external focus of attention.mp.
Results for VAF had to be presented separately to that
25 exp Motor Skills/ of other interventions if present. Someone other than
26 injury prevention.mp. the participant, such as the therapist, had to provide
the feedback (self-talk as feedback was excluded).
27 exp Rehabilitation/
Finally, the VAF had to be delivered verbally by the
28 3 and 26 therapist prior to or during the performance of the
29 21 and 24 task.
30 2 and 3 and 6
Comparison
31 3 and 8 and 9 Studies had to compare VAF with either different types
32 2 and 3 of AF, no AF or a control condition.
33 2 and 4
Outcome
34 3 and 26
Studies needed to include an outcome measure related
to motor learning such as improvement (or loss) of
lower extremity biomechanics/postural control in
different motor learning stages (acquisition, retention
rehabilitation and prevention of lower extremity and transfer).
musculoskeletal dysfunctions. A secondary objective If the outcome measure was not focused towards
was to evaluate the effect of VAF on motor learning rehabilitation/prevention of lower limb injuries,
with respect to the key stages of acquisition, retention for example, lower extremity biomechanics, the articles
and transfer. were excluded (figure 1).

Storberget M, et al. BMJ Open Sport Exerc Med 2017;3:e000256. doi:10.1136/bmjsem-2017-000256 3

 4
Table 2 Study characteristics and outcomes
Study:
name of Injury/injury
authors, Setting/ prevention
year country and exercise Study design Population Type of feedback Outcome measures
Open Access

1 Benjaminse Controlled ACL injury RCT n=90 healthy EFA by a visual stimuli compared with IFA by a verbal stimulus Vertical ground reaction force, knee kinetics and knee
et al, 201529 laboratory prevention. recreational Control: no feedback. General instructions prior to task. kinematics (sagittal angles and moments of the trunk,
setting/The Sidestep basketball VAF: “Bend your trunk forward, bend your knee and keep your hip, knee, ankle and range of motion. Frontal knee plane
Netherlands cutting. players knee straight above your foot”. (p2) moment) for acquisition and retention.
Female: 45 Enough time to warm up and familiarisation. Two retention sessions without feedback: after 1 week
Male: 45 Three trials and 4 weeks.
Mean age:
Female: 22.3
Male: 24.9

2 Gokeler Outpatient MSK injury: Between-group n=16 patients EFA versus IFA. Jump distance, knee valgus angle at initial contact,
et al, 201528 physical postop ACL experimental Female: 7 VAF: peak knee valgus angle, knee flexion angle at initial
therapy reconstruction. design Male: 9 EFA: “. . . I want you to think about pushing yourself off as hard contact, peak knee flexion ankle, total ROM and time to
facility/The (>4 months (randomisation) Mean age: as possible from the floor”. peak angles for acquisition.
Netherlands post-ACLR). Internal focus: IFA: “. . . I want you to think about extending your knees as (Successful jump if patient could keep balance for at
Single leg hop 23.75 rapidly as possible”. (p116) least 2 s postlanding).
jump. External focus: 5 min warm-up.
22.63 Three times practice.
Five trials.
30 s recovery time trial/jump.

3 Laufer et al, Military MSK injury: RCT n=40 volunteers. EFA versus IFA. Main outcome measure: overall stability for acquisition
200723 outpatient <4 months Female: 4 VAF: and retention (48 hours).
physical postgrade 1 or Male: 36 EFA: “Keep your balance by stabilising the platform”. APSI, OSI and MLSI.
therapy/Israel 2 ankle sprain. Mean age: IFA: “Keep your balance by stabilising your body”. (p106)
Balance IFA: 21.1 Two practice trials (30 s each)
training on a EFA: 20.5 Two tests (20 s each)
dynamic 30 s rest between trials
stabiliometer.

4 Prapavessis University of MSK injury RCT n=91 volunteers VAF versus IF Reduced GRF from vertical jump for short-term effect
and McNair, Auckland/ risk: Female: 35 VAF: (focus on the control of specific joint movement during for acquisition.
199924 New Zealand prevention of Male: 56 landing); “When you do your next jump, position yourself on the
injury of lower Augmented: 41 balls of your feet with bent knees just prior to landing. On
limb. (26 males and landing, lower the heels slowly to the ground and bend the
Vertical jump. 15 females) knees until well after the landing”.24
Sensory: 50 (30 (p354)
males and 20 IF: use the experience of their first jump to land in a way that
females) would minimise the stress of their next landing.
Mean age: No warm-up, break and practice reported.
16.07

5 Rotem- Military MSK injury: RCT n=36 male EFA versus IFA. Acquisition and transfer of a learnt balance capability/
Lehrer and outpatient <4 months volunteers EFA: “Keep your balance by stabilising the platform”. postural control to a more advanced task (48 hours).
Laufer, physical postgrade 1 or Mean age: 20.9 IFA: “Keep your balance by stabilising your body”. (p566) APSI, OSI and MLSI
200725 therapy 2 ankle sprain. 20 practice trials (20 s each).
department/ Balance 30 s rest between trials.
Israel training on
stabiliometer.

Continued

Storberget M, et al. BMJ Open Sport Exerc Med 2017;3:e000256. doi:10.1136/bmjsem-2017-000256

by copyright.
BMJ Open Sport Exerc Med: first published as 10.1136/bmjsem-2017-000256 on 21 September 2017. Downloaded from http://bmjopensem.bmj.com/ on March 28, 2021 by guest. Protected
 BMJ Open Sport Exerc Med: first published as 10.1136/bmjsem-2017-000256 on 21 September 2017. Downloaded from http://bmjopensem.bmj.com/ on March 28, 2021 by guest. Protected
Open Access

Study design

ACLR, anterior cruciate ligament reconstruction; APSI, Anterior/Posterior Stability Index; EFA, external focus of attention; GRF, ground reaction force; IF, intrinsic feedback; IFA, internal focus
Randomised controlled trials (RCTs), quasi-rando-

Change of landing biomechanics for prevention of ACL

Successful trial if participants landed with both feet at

of attention; MSK, musculoskeletal; MLSI, Medial/Lateral Stability Index; OSI, Overall Stability Index; postop ACL, postoperative anterior cruciate ligament; RCT, randomised controlled trial;
mised controlled studies, cross-over designs, single
case experimental, pre–post studies using primary data
were included. Case studies were excluded owing to

injury (acquisition and retention, 48 hours).

their low ranking on the research pyramid. Studies

the same time and in the right position.

where the full text could not be retrieved (ie, confer-
ence abstracts) and non-human studies were also
excluded.
Outcome measures

Information sources and search

Bibliographic databases were searched from 28
September 2015 to 26 December 2016. MEDLINE
(1946– to January week 2 2016), Embase (1974 to
January 2016), Physiotherapy Evidence Database,
Cinahl Plus, ProQuest, Web of Science, PubMed and
VAF: (One-time dose of external feedback) “Focus on landing as

Cochrane Library were used to find eligible studies.

EFA versus control group. Both groups were told to ‘jump as

Grey literature, relevant reviews or books about motor

Warm-up, three practice trials, 30 s rest between the five

learning and reference lists of primary studies were

searched to find supplementary papers and informa-
tion. Search strategies were developed in
Feedback group received additional instruction:

MEDLINE using Medical Subject Headings (MESH)

and free text (table 1). Search strategies for the other
databases were based on the MEDLINE search and
developed in consultation with a biomedical librarian.
light as a feather”. (p20)

Study selection

by copyright.
One reviewer screened the title and abstract of studies
Type of feedback

high as possible’.

(MS), in line with the inclusion criteria. Full text

training trials.

articles were retrieved and screened by two reviewers

(MS and LHJG). In any case of disagreements,
consensus was reached by discussion, or a third
reviewer (NH) was consulted if needed. The manage-
n=31 volunteers

Mean age: 21.3

Mean age: 21.0

ment of the included papers and removal of duplicates

females and 7

females and 8
with no injury

Feedback: 8
Population

were supported by the Reference Manager Software

Female: 16

Control: 8
Male: 15

RefWorks.
males

males

Data collection process

(counterbalance).

Study data were extracted by the first reviewer (MS)

Thesis project:

randomisation
Study design

Unpublished.

and audited for accuracy by the second reviewer

ROM, range of motion; VAF, verbal augmented feedback.
blocked

(LHJG). A data extraction form was created prior to

design

the collection and was piloted to avoid any discrepan-

cies of interpretation. Details of the extracted
information are reported in table 2 (study character-
and exercise

prevention of
Injury/injury
prevention

ACL injury.
MSK injury

Double leg

istic and outcome) and table 5 (results). Consensus

jump task.

between reviewers (MS and LHJG) was attained by

risk:

dialogue, and an opinion from the third reviewer (NH)

was sought if necessary.
Oregon State
University/
Setting/

Risk of bias in individual studies

country
Continued

USA

The two reviewers (MS and LHJG) independently

assessed risk of bias by using the Cochrane Collabora-
tion’s tool (see table 3). Each study was rated against
Weilbrenner,

the defined types of bias. One of the excluded studies

authors,
name of

201426
Study:
Table 2

was used as a pilot. The inter-rater agreement for

year

study bias ratings between the reviewers was measured

using the Cohen’s kappa statistics. Any disagreements
6

Storberget M, et al. BMJ Open Sport Exerc Med 2017;3:e000256. doi:10.1136/bmjsem-2017-000256 5

 BMJ Open Sport Exerc Med: first published as 10.1136/bmjsem-2017-000256 on 21 September 2017. Downloaded from http://bmjopensem.bmj.com/ on March 28, 2021 by guest. Protected
Open Access

Table 3 Summary assessment of the overall risk of bias—Cochrane Collaboration’s tool

Summary
Different types of within
Study bias study Overall risk
1 2 3 4 5 6

1 Benjaminse et al, U U H U L L H=1 Enrolment, allocation and testing done by the same person.
201529 L=2 Lack of information in terms of selection and attrition
U=3
2 Gokeler et al, L L U L L L H=0 Lack of information in terms of blinding
201528 L=5
U=1
3 Laufer et al, U U U L L L H=0 Lack of information in terms of selection bias and blinding
200723 L=3
U=3
4 Prapavessis and L U U L L L H=0 Allocation concealment not reported. Lack of information in
McNair, 199924 L=4 terms of blinding
U=2
5 Rotem-Lehrer U U U L L H=0 Lack of information in terms of selection bias and blinding
and Laufer, L=3
200725 U=3
6 Weilbrenner, L U U L L L H=0 Allocation concealment and blinding not reported
201426 L=4
U=2

Risk of bias criteria: 1, selection bias=random sequence generation; 2, selection bias=allocation concealment; 3, performance bias/detection
bias=blinding of personnel and blinding of participants/blinding of outcome assessors; 4, attrition bias=incomplete outcome data; 5,

by copyright.
reporting bias=short-term selective outcome reporting; 6, other bias=potential threats to validity, for example, consideration of a protocol.
Levels of risk of bias: H, high risk of bias; L, low risk of bias and U, unclear risk of bias.

were resolved through discussion. A third reviewer Study characteristics

(NH) was consulted if needed. Table 2 summarises the characteristics of the included
studies. There were five RCTs1 22–25 and one blocked
randomised design.26 Three of the studies enrolled
Quality of evidence participants with a musculoskeletal injury, of which two
Two reviewers (MS and LHJG) assessed the studies’ studies had participants with lateral ankle sprain23 25,
quality of evidence for each main outcome using the and the third investigated individuals with postopera-
Grading of Recommendations Assessment, Develop- tive anterior cruciate ligament reconstruction.22 The
ment and Evaluation (GRADE)15 16 system (table 4). remaining three studies enrolled healthy participants.1
24 26
The total sample size of all six studies was 304; 92
injured participants and 212 healthy. Three studies
Synthesis of results
compared external focus of attention (EFA) verbal
Final results derived from the risk of bias analysis and
feedback to internal focus of attention (IFA) verbal
quality assessment were included in the synthesis and
feedback.22 23 25 One study compared EFA verbal feed-
analysis of data. Owing to the heterogeneity across the
back to a control group. One study compared VAF with
studies, a meta-analysis was not appropriate. Narrative
no AF (they relied on their own intrinsic sensory
reporting was therefore used to synthesise results.
systems), and one study compared VAF only to a
control group.26 The outcome measures of the
RESULTS included studies were jump distance,22 stability/
Study selection balance/postural control,22 23 25 ground reaction force1
22 24
The search yielded 292 studies after duplicates were and knee kinematics.1 26
removed. Screening of titles and abstracts resulted in
24 studies being retrieved. Eleven of these were Risk of bias within the studies
included for further eligibility check. Five studies17–21 The percentage agreement between the two reviewers’
were excluded leaving six studies to be included in the risk of bias was 90.5% with kappa=0.806 (CI 0.626 to
analysis1 22–26 as agreed by the reviewers (MS, 0.987). Five of the studies presented unclear risk of
LHJG and NH). bias regarding allocation concealment.1 23–26 One

6 Storberget M, et al. BMJ Open Sport Exerc Med 2017;3:e000256. doi:10.1136/bmjsem-2017-000256

 BMJ Open Sport Exerc Med: first published as 10.1136/bmjsem-2017-000256 on 21 September 2017. Downloaded from http://bmjopensem.bmj.com/ on March 28, 2021 by guest. Protected
Open Access

Table 4 Quality of body of evidence based on the GRADE approach

Summary/
Number of Publication quality of
Outcome studies Limitation Inconsistency Indirectness Imprecision bias Upgrade evidence

Jump 1 RCT No NA No serious 1 None +1 High

distance serious indirectness 
limitation
Stability/ 2 RCTs23 25 1 None No serious 1 None +1 Moderate
postural indirectness 
control/
balance
GRF 2 RCTs1 24 1 None No serious 1 None +1 Moderate
indirectness 
Knee 2 RCTs1 22 1 None No serious 1 None +1 Moderate
kinematics 1 blocked indirectness 
randomised
design

All RCTs start as high quality. Assessment criteria: limitation: based on Cochrane risk of bias assessment. Downgraded by one level if more
than one unclear. Inconsistency: unexplained heterogeneity across studies. indirectness: heterogeneity for participants, intervention or
outcome measure in individual studies. imprecision: if no sample size justification and calculation: downgraded by one level. Publication bias.
Upgrade: if statistically significant effect: upgraded by one level.16
GRADE, Grading of Recommendations Assessment, Development and Evaluation; GRF, ground reaction force; RCT, randomised controlled
trial.

study had low risk for allocation concealment.22 Three Effect of VAF (EFA vs IFA) for musculoskeletal injury

by copyright.
studies had an unclear risk of bias owing to poor Three RCTs, with moderate to high strength of
reporting on randomisation sequence,23 25 the evidence, looked at EFA versus IFA in participants with
remaining were low risk or not applicable. Five studies a musculoskeletal injury.22 23 25 The study by Gokeler
showed poor reporting in terms of blinding, and one et al28 looked at the stage of acquisition and provided
study had high risk. With regard to reporting and VAF prior to testing. The study found no statistical
other biases, all studies had low risk of bias. One study differences between the EFA and IFA groups in terms
had unknown risk for attrition bias, the rest were rated of jump distance. For knee kinematics, the IFA group
as low (figure 2 and table 3). All studies had an overall had significantly lower knee flexion compared with the
unclear risk of bias, in accordance with the descriptions EFA group. Laufer et al23 assessed acquisition and
of overall risk of bias within the Cochrane Handbook retention (48 hours post-test) and reported an effect
for Systematic Reviews of Interventions.27 primarily for stance phase after three sessions of
training balance. Compared with IFA, EFA was supe-
rior on the effect of balance in the simpler stance
Quality of body of evidence position, especially for the acquisition phase. Retention
Using GRADE (table 4), all bodies of evidence were tests showed maintenance of newly gained skills.
downgraded owing to imprecision. All outcomes were Rotem-Lehrer and Laufer25 tested transfer of a
downgraded owing to the lack of sample size justifica- postural control task (48 hours post-test) and showed
tion and calculation (imprecision). Out of the quality significant differences in all stability measures of
classification, three studies were of moderate and one pretraining and post-training for EFA rather than IFA.
high quality. In total, four outcomes (jump distance, For both studies,23 25 no VAF was given during the
stability/postural control/balance, ground reaction force assessment, only for training.
and knee kinematics) were assessed.

Synthesis of results Effect of VAF versus intrinsic feedback for musculoskeletal

The results are summarised in table 5 in a narrative injury prevention (healthy participants only)
form as the studies demonstrated heterogeneity with Moderate evidence from one RCT demonstrated a
respect to participants, sample, sample size, protocol, significantly lower ground reaction force for VAF
interventions and outcome measures. The selection of compared with intrinsic feedback in musculoskeletal
VAF phrases used during the intervention in the injury prevention. VAF was provided during testing in
studies can be found in table 2. the acquisition stage only.

Storberget M, et al. BMJ Open Sport Exerc Med 2017;3:e000256. doi:10.1136/bmjsem-2017-000256 7

 BMJ Open Sport Exerc Med: first published as 10.1136/bmjsem-2017-000256 on 21 September 2017. Downloaded from http://bmjopensem.bmj.com/ on March 28, 2021 by guest. Protected
Open Access

Figure 2 Risk of bias (high, unknown and low) within studies in terms of different categories.

Effect of VAF versus control for musculoskeletal injury overall body of evidence was deemed of a moderate
prevention (healthy participants only) quality (table 4), which means that further research can
Two studies,1 26 with moderate evidence, compared alter the estimated effect and beliefs about the strength
VAF with a control group in musculoskeletal injury of the evidence.15
prevention during the stages of acquisition and reten-
tion. The former study also included a visual feedback Effect of VAF (EFA vs IFA) for MSK injury
(video) group. In this study, no feedback was given in In the study by Gokeler et al,28 both landing strategy
the retention test. In the VAF group, significant effects and jump distance were measured. In terms of the

by copyright.
for all sessions were found for knee flexion angles in former, the IFA group’s landing strategy was assessed
females only, compared with video and control group. as stiffer compared with the EFA group. Theoretically,
Males in the video group had larger ground reactions this can lead to the risk of developing an ACL injury.22
(all sessions), greater knee flexion (regardless of Therefore, in this case, VAF with an EFA may be bene-
sessions) and reduced knee valgus moment (over time) ficial for anterior cruciate ligament injury prevention
compared with the VAF group and the control group.1 and motor learning achievement. For jump distance,
Weilbrenner26 did not find any significant changes in there were no statistical effects. Gokeler et al28 suggest
landing biomechanics for VAF during the task versus that an extra stimulus could be necessary to achieve a
the control group. significant effect for jump distance. Additionally, a
different wording of VAF, for example, instructing the
DISCUSSION participant to reach as well as jump, can improve the
The purpose of this review was to determine the effect effect of motor learning.22
of VAF in the rehabilitation and prevention of muscu- Laufer et al23 and Rotem-Lehrer and Laufer25 looked
loskeletal dysfunctions in the lower limbs, as well as to at postural control. They both found optimistic results
determine its effect on the different motor learning for EFA as way of providing VAF in the rehabilitation
stages; acquisition, retention and transfer. Five out of of musculoskeletal dysfunctions, as postural control
six studies reported statistically significant effects for enhancement is crucial for secondary prevention of
VAF. With evidence of moderate quality from five out lateral ankle sprains.23 25
of six studies, VAF was found to be effective for
improving ground reaction force and lower extremity Effect of VAF versus IF for MSK injury prevention (healthy
biomechanics (acquisition and retention) and postural participants only)
control (transfer), which are all crucial factors in reha- Healthy participants in the study by Prapavessis and
bilitation and prevention of injuries. Four out of the McNair24 demonstrated motor learning achievement
six studies did not include a control group22–25 and in terms of reduced ground reaction force after
cannot support VAF over any other intervention or no receiving VAF. A lower ground reaction force improves
intervention. landing biomechanics and can in turn prevent injuries
Caution should be made in interpreting this evidence if kept doing the same way in practice. However, the
given all studies were classed as unclear risk of bias investigators only looked at the stage of acquisition,
(figure 1 and table 3) and the evidence indicates a low and a relatively permanent change (ie, motor learning)
statistical credibility (table 5). Notwithstanding this, the was not confirmed. Additionally, an investigation of

8 Storberget M, et al. BMJ Open Sport Exerc Med 2017;3:e000256. doi:10.1136/bmjsem-2017-000256

 Table 5 Results
Motor control stages and Additional comments/overall
Study Effect authors conclusion risk of bias within the study

1 Benjaminse et al, " Females in VER group had greater knee flexion angles compared with females in CTRL and VIS group in each of the Acquisition & retention " No testing of transfer
201529 sessions (p<0.05). Male participants benefit from " Unclear risk of bias
" The males in VIS group had significantly larger vGRF in all sessions compared with males in the VER and CTRL group visual feedback. Visual
(p<0.05) feedback reduces knee joint
" The males in the VIS group had greater knee flexion moments compared with males in the VER and CTRL groups loading in males with high
(p<0.05) regardless of sessions. retention
" Males in the VIS group reduced knee valgus moment over time. Males in VER and CTRL group did not.
" Females in the VER group reduced their knee varus moment (non-significant change) over time.

2 Gokeler et al, " Mean jump distance, knee valgus angle at IC, peak knee valgus angle, time to peak knee valgus ankle and valgus ROM: Acquisition " No control group.
201528 no significant difference between IFA and EFA (p>0.05). EFA is more beneficial than IFA " Only short-term effect (no
" Knee flexion at IC: no significant group difference for non-injured legs (p=0.82). For injured legs it was a significantly for motor skill performance follow-up).
smaller knee flexion in IFA group (p=0.04). " Low statistical credibility:
" Peak knee flexion: significantly lower in IFA group for non-injured legs (p=0.01) and for injured legs: (p=0.01) compared small sample size (n=16).
with EFA. " No testing of transfer.
" IFA time to peak knee flexion for non-injured leg was significantly shorter (p=0.01) and for injured leg (p=0.0.02) " Unclear risk of bias.
compared with EFA.

3 Laufer et al, Stability level 6: Acquisition & retention " No control group.
200723 " Main effect of time for APSI (p<0.001). No such effect for OSI and MLSI. VAF (EFA) is advantageous for " Low statistical credibility:
" Significant interaction between group and time for OSI (p=0.030) and APSI (p=0.019). learning postural control task small sample size (n=40)
" Between post-training and pretraining: only EFA group had a significant decrease in OSI (p=0.030) and in APSI (especially for acquisition " Gender difference:
(p<0.001). phase) over three sessions males>females.
" Stability indices: no significant difference between post-training and retention. " No testing of transfer.
Stability level 4: it looks as if something is wrong with the lay out of this in the PDF-file. Can you change that please? " Unclear risk of bias.
" Main effect of time for OSI (p=0.010) and for APSI (p<0.001) (significant improvement). No such effect for MLSI
" Acquisition phase: Improvements for OSI (p=0.011) and APSI (p<0.001). Retention phase: no significant difference
" Stability indices: no effect
" APSI EFA group: trend toward significant interaction between time and group (p=0.078).

Storberget M, et al. BMJ Open Sport Exerc Med 2017;3:e000256. doi:10.1136/bmjsem-2017-000256

4 Prapavessis and " No significant difference between Trial 1 (prior to feedback) GRF of the sensory and AF group. Acquisition " No control group.
McNair, 199924 " A significant difference was present between Trial 2 (postfeedback). AF group: lower GRF compared with sensory Adolescents need more " Only short-term effect (no
feedback group. AF group: a significantly lower GRF was also observed in Trial 2 compared with Trial 1. information than that provided follow-up).
by their prior experiences in " Gender: males>females.
jumping—to lower GRF in " No testing of transfer.
landing. It is evident that those " Unclear risk of bias.
receiving VAF can pick up
instructions related to
improvements of lower limb
kinematics
5 Rotem-Lehrer and " Statistically significant group-by-time interaction OSI: (p=0.001), APSI: (p=0.03) and MLSI: (p=0.01). Acquisition and transfer. VAF " No control group.
Laufer, 200725 " Significant difference of pretraining and post-training for EFA in all stability measures: (p<0.05). Change over time for IFA (EFA) is advantageous for " Low statistical credibility:
was not significant. transfer of postural control task small sample size (n=36).
over a 3-day period " Gender: only male
participants.
" Good: testing transfer.
" Unclear risk of bias

Continued
Open Access

9
by copyright.
BMJ Open Sport Exerc Med: first published as 10.1136/bmjsem-2017-000256 on 21 September 2017. Downloaded from http://bmjopensem.bmj.com/ on March 28, 2021 by guest. Protected
 BMJ Open Sport Exerc Med: first published as 10.1136/bmjsem-2017-000256 on 21 September 2017. Downloaded from http://bmjopensem.bmj.com/ on March 28, 2021 by guest. Protected
Open Access

injured participants is required to be able to confirm

the effect for (secondary) injury prevention.
Additional comments/overall

attention; MLSI, Medial/Lateral Stability Index; OSI, Overall Stability Index; p, p-value; PATSF, peak anterior tibial shear force; VAF, verbal augmented feedback; VER, verbal; vGRF, vertical
risk of bias within the study

small sample size (n=31).

Low statistical credibility:

No testing of transfer.
Unclear risk of bias.

AF, augmented feedback; APSI, Anterior/Posterior Stability Index; CTRL, control; EFA, external focus of attention; GRF, ground reaction force; IC, initial contact; IFA, internal focus of
Effect of VAF versus control for MSK injury prevention
(healthy participants only)
Benjaminse et al29 did not find positive effects in the
VAF group (with IFA) regarding lower extremity
biomechanics, compared with the visual feedback
"

"
"

group (with EFA) and control group. This may suggest

that EFA is better than IFA and supports previous find-
ings stated in the Introduction.9 Weilbrenner26 cannot
Motor control stages and

The VAF did not change

support the use of VAF with external focus of attention

Acquisition & retention

landing biomechanics
authors conclusion

for primary ACL injury prevention in individuals who

have not sustained a previous anterior cruciate liga-
ment injury. This study did however differ from all the
other studies in that the VAF did not include sugges-
tions as to how to use the body during the test (instead
used a metaphor—‘as light as a feather’). This may
" Peak knee varus moment: no significant main for effects for time (p=0.792) or group (p=0.752), and no significant group-

suggest that a mental imagery could prove less effective

" Peak anterior tibial shear force: significant main effect for time (p=0.017), but no significant effect for group (p=0.329) or
" Peak knee valgus angle: no significant main effects for time (p=0.223) and group (p=0.844) or group-by-time interaction
" Knee flexion angle at initial contact: significant main effect (p=0.026), but no significant time (p=0.433) or group-by-time

Kinetics: " Peak knee extension moment: no significant main effect for time (p=0.588) and group (p=0.747) or group-by-

as feedback during exercise compared with actual

group-by-time interaction (p=0.029). Subjects across groups: significantly greater PATSF at retention than baseline.
" Frontal plane knee angle at initial contact: no significant main effects for time (p=0.469) or group (p=0.752) and no

verbal feedback.
interaction effects (p=0.523). Feedback group had significantly greater knee flexion compared with control.

Motor learning: stages and achievement

Acquisition is key to motor learning; however, reten-
tion and transfer are even more important in order to

by copyright.
prevent injuries. The ability to show satisfactory results
in retention and transfer tests is important considering
the new tasks and challenges concerned with returning
to play.29 Retention is the patient’s ability to show skill
achievement or improvement of the same task some
time after the acquisition phase, without having prac-
tised it.30 Transfer, on the other hand, requires
additional skills where the patient has to demonstrate
motor learning in a different, yet similar task.5 30 A
significant group-by-time interaction effect (p=0.288).

skill is therefore not considered as fully learnt before

the patient can show successful results in retention and
transfer tests. It is important to bear in mind though
that these tests do not always give us straightforward
by-time (p=0.801) interaction effect.

conclusions owing to factors such as temporary fatigue

time interaction (p=0.908).

or anxiety.30
Only one study tested transfer, the rest of the studies
assessed solely acquisition and retention (three) or
acquisition alone (two). In view of the two studies
effects (p=0.775).

testing short-term effects (acquisition and retention),3

24
an answer to whether a motor learning achievement
ground reactions force; VIS, visual.
Kinematics:

was present cannot truly be obtained without a transfer

test, as learning only occurs if the participants can
Effect

show relatively permanent changes.5 31 However, the

Continued

question then becomes, how long does it take before a

change can be reasonably considered as long-term?
The longest follow-up was 4 weeks. Prapavessis and
Weilbrenner,

McNair24 suggest that a longer follow-up, such as a

201426
Table 5

Study

year, could provide more realistic results. It might

however depend on the intensity and frequency of the
exercise, as to how long an effect is expected to last.
6

10 Storberget M, et al. BMJ Open Sport Exerc Med 2017;3:e000256. doi:10.1136/bmjsem-2017-000256

 BMJ Open Sport Exerc Med: first published as 10.1136/bmjsem-2017-000256 on 21 September 2017. Downloaded from http://bmjopensem.bmj.com/ on March 28, 2021 by guest. Protected
Open Access

Analysis in relation to existing evidence clinical relevance, future studies are recommended to
Regarding the first objective, the present review test individuals suffering from a musculoskeletal injury.
supports previous findings stating that there is a lack Further use of reporting guidelines for research publi-
of quantity and quality of current evidence for VAF cations may enhance the quality of the evidence base
and musculoskeletal dysfunctions,12 13 and it is there- by ensuring a robust methodological process is used
fore not possible to determine whether VAF is with transparent designs and methods.
effective. In terms of focus of attention, the results To determine best practices, it would be relevant to
from three of the included papers22 23 25 support look at other aspects of VAF delivery, such as timing
previous evidence that VAF with EFA is more effective and frequency of all three fundamental stages of motor
than VAF with IFA.9 12 32 They all confirmed statisti- learning: acquisition, retention and transfer. Looking
cally significant results regarding EFA and VAF, but at the current systematic review, transfer was assessed
conclusions should be taken with caution owing to an 48 hours after the acquisition phase, and retention was
overall moderate quality of evidence. tested 4 weeks (no feedback provided) or 48 hours23 26
postacquisition stage. (Feedback was provided in the
latter study). In studies looking at the stroke popula-
Strengths and limitations tion and the group of healthy participants, i there is a
The review used a thorough literature search in eight wide variation in terms of timing of retention and
databases, inclusion and exclusion criteria were devel- transfer tests following acquisition phase: 1 day, 2 days,
oped a priori and the protocol was registered. The 3 days, 1 week, 4 weeks and 7 weeks.32–36 One study
review is written in line with PRISMA, and GRADE was defined two types of retention tests: immediate reten-
used to determine the overall quality of the synthesised tion (5 min) or delayed retention (next day).37
results. In light of the heterogeneity of evidence, recommen-
Only six studies were identified for inclusion and in dations cannot be made regarding timing of retention
order to involve a sufficient number of studies, studies and transfer tests postacquisition stage. It does
of injury prevention were included. Owing to the however seem like a minimum of 24 hours postacquisi-
heterogeneity of the studies, a meta-analysis was not tion stage should be a requirement for retention and
appropriate. transfer tests. In addition, several studies agree on the
Moreover, it should be noted that the protocols

by copyright.
fact that no feedback in these tests should be provided.
differed between the studies and showed a wide varia- One thing is clear; there is a need for the development
tion with respect to sample, intervention, outcome of a standard by which these tests must be conducted.
measure, gender, time for practice, warm-up and rest. In terms of the interventions chosen, they should be in
Rest might influence performance in terms of line with the Medical Research Council Framework for
preparing the body for exercise and to prevent Complex Interventions.38 Ultimately, we need the
fatigue.26 Time postmusculoskeletal injury differed interventions to bring out meaningful long-term
among studies (table 2), and both novice and experi- outcomes such as return to play or reduced prevalence
enced sporting participants were included. This means of injury—to provide physiotherapists with confidence
that pain scores, balance and skills will vary between within the evidence based clinical practice.
the subjects—factors that may influence the level of
motor learning achievement. Another factor to
consider is the participants’ age and gender. All six CONCLUSION
studies included relatively young subjects, usually The results from this systematic review suggest that
adolescents and both males and/or females. It is said there is moderate evidence that VAF is effective in the
that females have a higher risk of developing injury in rehabilitation and prevention of lower extremity
puberty, and testing females at this age is important musculoskeletal dysfunctions. From this review and
for injury prevention. However, transferring these notwithstanding the lack of high-quality evidence,
results to the management of males or adults/elderly improvements in terms of lower extremity biome-
may not be possible. chanics in a jumping task or enhanced postural control
while balancing were found following VAF. Future
high-quality studies are required to specifically evaluate
Clinical and research implications VAF, including different parameters associated with
Based on the above findings, it is still unclear whether feedback and long-term effects of interventions, where
physiotherapists can fully trust current evidence in acquisition, retention and transfer are evaluated.
terms of providing VAF in a clinical context with
respect to musculoskeletal injuries in the lower limbs.
The systematic review has detected inconsistencies with Contributors MS was the first reviewer. LHJG was the second reviewer. SS,
the use of VAF in published studies. Furthermore, PvV and NH were coauthors. All authors read, edited and approved the final
examination of healthy participants is not sufficient to manuscript.
demonstrate whether VAF is effective in the rehabilita- Funding The funding of the publication will be covered by the University of
tion of musculoskeletal dysfunctions. To provide more Birmingham.

Storberget M, et al. BMJ Open Sport Exerc Med 2017;3:e000256. doi:10.1136/bmjsem-2017-000256 11

 BMJ Open Sport Exerc Med: first published as 10.1136/bmjsem-2017-000256 on 21 September 2017. Downloaded from http://bmjopensem.bmj.com/ on March 28, 2021 by guest. Protected
Open Access

Competing interests None declared. ligament and patellofemoral joint injury prevention. Knee
2014;21:891–5.
Provenance and peer review Not commissioned; externally peer reviewed. 18. Myer GD, Stroube BW, DiCesare CA, et al. Augmented feedback
Data sharing statement No additional data are available. supports skill transfer and reduces high-risk injury landing
mechanics. Am J Sports Med 2013;41:669–77.
Open Access This is an Open Access article distributed in accordance with 19. Onate JA, Guskiewicz KM, Sullivan RJ. Augmented feedback
the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, reduces jump landing forces. J Orthop Sports Phys Ther
which permits others to distribute, remix, adapt, build upon this work non- 2001;31:511–7.
commercially, and license their derivative works on different terms, provided 20. Parsons JL, Alexander MJ. Modifying spike jump landing
biomechanics in female adolescent volleyball athletes using video
the original work is properly cited and the use is non-commercial. See: http:// and verbal feedback. J Strength Cond Res 2012;26:1076–84.
creativecommons.org/licenses/by-nc/4.0/ 21. Stroube BW, Myer GD, Brent JL, et al. Effects of task-specific
© Article author(s) (or their employer(s) unless otherwise stated in the text of augmented feedback on deficit modification during performance of
the tuck-jump exercise. J Sport Rehabil 2013;22:7–18.
the article) 2017. All rights reserved. No commercial use is permitted unless
22. Gokeler A, Benjaminse A, Hewett TE, et al. Feedback techniques to
otherwise expressly granted. target functional deficits following anterior cruciate ligament
reconstruction: implications for motor control and reduction of
REFERENCES second injury risk. Sports Med 2013;43:1065–74.
23. Laufer Y, Rotem-Lehrer N, Ronen Z, et al. Effect of attention focus
1. Benjaminse A, Otten B, Gokeler A, et al. Motor learning strategies in
on acquisition and retention of postural control following ankle
basketball players and its implications for ACL injury prevention: a
sprain. Arch Phys Med Rehabil 2007;88:105–8.
randomized controlled trial. Knee Surg Sports Traumatol Arthrosc
24. Prapavessis H, McNair PJ. Effects of instruction in jumping
2017;25:2365–76.
technique and experience jumping on ground reaction forces. J
2. Woolf AD, Pfleger B. Burden of major musculoskeletal conditions.
Orthop Sports Phys Ther 1999;29:352–6.
Bull World Health Organ 2003;81:646–56.
€ck M. Functional, structural and molecular plasticity of
3. Flu 25. Rotem-Lehrer N, Laufer Y. Effect of focus of attention on transfer of
mammalian skeletal muscle in response to exercise stimuli. J Exp a postural control task following an ankle sprain. J Orthop Sports
Biol 2006;209:2239–48. Phys Ther 2007;37:564–9.
4. Magill R. Motor learning and control. 1st ed. New York: McGraw-Hill, 26. Weilbrenner J. The influence of external focus of attention feedback
2011. on ACL injury related landing biomechanics. Honors baccaluareate
5. Muratori LM, Lamberg EM, Quinn L, et al. Applying principles of of exercise and sports sciences. Oregon State University
motor learning and control to upper extremity rehabilitation. J Hand 2014.
Ther 2013;26:94–103. 27. Higgins J, Green S, The Cochrane Collaboration. Chapter 8:
6. Snodgrass SJ, Heneghan NR, Tsao H, et al. Recognising Assessing risk of bias in included studies. In: Higgins J, Green S,
neuroplasticity in musculoskeletal rehabilitation: a basis for greater eds. Cochrane handbook for systematic reviews of interventions. 1st
collaboration between musculoskeletal and neurological ed, Version 5.1.0 [Updated March 2011]; 2011.
physiotherapists. Man Ther 2014;19:614–7. 28. Gokeler A, Benjaminse A, Welling W, et al. The effects of attentional
7. Buchbinder R, Maher C, Harris IA. Setting the research agenda for focus on jump performance and knee joint kinematics in patients
improving health care in musculoskeletal disorders. Nat Rev after ACL reconstruction. Phys Ther Sport 2015;16:114–20.

by copyright.
Rheumatol 2015;11:597–605. 29. Benjaminse A, Gokeler A, Dowling AV, et al. Optimization of the
8. Durham KF, Sackley CM, Wright CC, et al. Attentional focus of anterior cruciate ligament injury prevention paradigm: novel
feedback for improving performance of reach-to-grasp after stroke: feedback techniques to enhance motor learning and reduce injury
a randomised crossover study. Physiotherapy 2014;100:108–15. risk. J Orthop Sports Phys Ther 2015;45:170–82.
9. Sturmberg C, Marquez J, Heneghan N, et al. Attentional focus of 30. Schmidt R, Lee T. Motor control and learning. 1st ed. Champaign, IL:
feedback and instructions in the treatment of musculoskeletal Human Kinetics, 2011.
dysfunction: a systematic review. Man Ther 2013;18:458–67. 31. Shumway-Cook A, Woollacott M. Motor control. 4th ed. Philadelphia:
10. van Vliet PM, Heneghan NR. Motor control and the management of Wolters Kluwer Health/Lippincott Williams & Wilkins, 2012.
musculoskeletal dysfunction. Man Ther 2006;11:208–13. 32. Wulf G. Attentional focus and motor learning: a review of 15 years.
11. Ribeiro DC, Sole G, Abbott JH, et al. A rationale for the provision of Int Rev Sport Exerc Psychol 2013;6:77–104.
extrinsic feedback towards management of low back pain. Man Ther 33. van Vliet PM, Wulf G. Extrinsic feedback for motor learning after
2011;16:301–5. stroke: what is the evidence? Disabil Rehabil 2006;28:831–40.
12. Sigrist R, Rauter G, Riener R, et al. Augmented visual, auditory, 34. Wulf G, McConnel N, Ga €rtner M, et al. Enhancing the learning of
haptic, and multimodal feedback in motor learning: a review. sport skills through external-focus feedback. J Mot Behav
Psychon Bull Rev 2013;20:21–53. 2002;34:171–82.
13. Hartveld A, Hegarty J. Augmented feedback and physiotherapy 35. Wulf G, Prinz W. Directing attention to movement effects
practice. Physiotherapy 1996;82:480–90. enhances learning: a review. Psychon Bull Rev 2001;
14. Wulf G, Schmidt RA. Feedback-induced variability and the learning 8:648–60.
of generalized motor programs. J Mot Behav 1994;26:348–61. 36. Wulf G. Attention and motor skill learning. 1st ed. Champaign, IL:
15. Guyatt GH, Oxman AD, Vist GE, et al. GRADE: an emerging Human Kinetics, 2007.
consensus on rating quality of evidence and strength of 37. Wulf G, Chiviacowsky S, Schiller E, et al. Frequent external-focus
recommendations. BMJ 2008;336:924–6. feedback enhances motor learning. Front Psychol 2010;1:1.
16. Schu€nemann H, Brozek J, Guyatt G, et al. GRADE Handbook, 2013. 38. Craig P, Dieppe P, Macintyre S, et al. Developing and evaluating
17. Munro A, Herrington L. The effect of videotape augmented feedback complex interventions: the new Medical Research Council guidance.
on drop jump landing strategy: Implications for anterior cruciate BMJ 2008;337:a1655.

12 Storberget M, et al. BMJ Open Sport Exerc Med 2017;3:e000256. doi:10.1136/bmjsem-2017-000256