Importance Running-related injuries are highly prevalent.
Objective Synthesise published evidence with international expert opinion on the use of running retraining when treating lower limb injuries.
Design Mixed methods.
Methods A systematic review of clinical and biomechanical findings related to running retraining interventions were synthesised and combined with semistructured interviews with 16 international experts covering clinical reasoning related to the implementation of running retraining.
Results Limited evidence supports the effectiveness of transition from rearfoot to forefoot or midfoot strike and increase step rate or altering proximal mechanics in individuals with anterior exertional lower leg pain; and visual and verbal feedback to reduce hip adduction in females with patellofemoral pain. Despite the paucity of clinical evidence, experts recommended running retraining for: iliotibial band syndrome; plantar fasciopathy (fasciitis); Achilles, patellar, proximal hamstring and gluteal tendinopathy; calf pain; and medial tibial stress syndrome. Tailoring approaches to each injury and individual was recommended to optimise outcomes. Substantial evidence exists for the immediate biomechanical effects of running retraining interventions (46 studies), including evaluation of step rate and strike pattern manipulation, strategies to alter proximal kinematics and cues to reduce impact loading variables.
Summary and relevance Our synthesis of published evidence related to clinical outcomes and biomechanical effects with expert opinion indicates running retraining warrants consideration in the treatment of lower limb injuries in clinical practice.
Statistics from Altmetric.com
Running as a means of exercise is popular for both recreation, and as a sport in its own right. However, running is not risk free. Injury incidence ranges from 19% to 78% among studies with follow-up periods between 1 week and 18 months.1–2 Common running injuries include exertional lower leg pain, plantar fasciopathy, Achilles tendinopathy, calf pain, medial tibial stress syndrome, patellofemoral pain (PFP), iliotibial band syndrome (ITBS), patellar tendinopathy, hamstring injury including proximal tendinopathy and gluteal tendinopathy.1 Proposed intrinsic factors for running-related injury include older age,3 higher body mass index,4 history of prior injury,1 limb length discrepancy,5 abnormal anatomical alignment5 and foot posture,6 ,7 and faulty foot loading patterns.8 ,9 Proposed extrinsic factors include level of competition,10 training accumulation,11 shoe type5 and training surface.12 Multiple interventions have been developed in an attempt to mitigate these risks and treat injuries that develop, including exercise programmes to improve strength, neuromuscular control and flexibility,13 ,14 footwear modification,13 ,15 foot orthoses,13 ,16 ,17 and taping techniques.13 ,18 Despite extensive research in all areas, a lack of effective long-term treatment strategies remains a source of frustration for many runners and clinicians.
Alteration to running technique may help to treat running injuries by reducing load in certain muscle groups and joints.19 ,20 This theory was tested by Davis21 introducing the coaching concept of running retraining to treat lower limb injuries in a case series of five patients who reported pain reduction and maintained a changed running technique. The process involved (1) identifying any theoretical abnormal running mechanics which may be contributing to tissue overload; (2) establishing if running mechanics could be altered; and (3) facilitating the desired running mechanics changes and encouraging motor learning to ensure maintenance of any change. Despite nearly 10 years since the publication of this paper, only four small case series evaluating the effectiveness of running retraining interventions in reducing pain have been published.22–25
Biomechanical studies in asymptomatic populations have examined retraining strategies extensively (including altering step rate, strike pattern, hip and knee motion, trunk position, step width, and impact loading variables), and reported to change running kinematics, kinetics and electromyography. A recent systematic review by Napier et al26 concluded a number of biomechanical variables can be altered by running retraining interventions. Additionally, they concluded the most effective strategy for reducing impact loading variables including vertical impact peak, and average and instantaneous loading rate was provision of real-time feedback related to kinetics and/or kinematics. However, the importance of these impact loading variables to injury are not supported by strong evidence, and the review by Napier et al26 did not provide a synthesis of current published clinical outcomes, despite emerging evidence.22–25
Many interventions studied in previous running retraining research may not translate to clinical practice well, due to issues with practicality (eg, real-time three-dimensional (3D) motion analysis feedback), and the ‘one size fits all’ approach of associated methodology (eg, all participants transition from a rearfoot to forefoot strike). It also remains unclear how running retraining may interact with other interventions with an established evidence base such as exercise rehabilitation.27–29 Given this lack of evidence to guide clinicians and researchers in the practical implementation of running retraining, consultation with those already prescribing running retraining in clinical practice may provide valuable insight. Such consultation is part of evidence-based practice, particularly where clinical trials are lacking.30
Our mixed-methods study had three objectives: (1) to systematically review and summarise the clinical and biomechanical evidence for implementing running retraining to treat lower limb injuries; (2) explore the clinical reasoning related to the use of running retraining by international experts; and (3) synthesise these elements to provide guidance for clinicians and researchers who seek to implement and evaluate running retraining interventions in the future.
Inclusion and exclusion criteria
Running retraining was defined as ‘the implementation of any cue or strategy to alter an individual's running technique’.21 Studies evaluating other interventions in conjunction with running retraining were considered if the effects of retraining could be clearly delineated (eg, altering footwear combined with instruction of strike pattern, compared with altering footwear alone). Two streams of studies were sought:
Studies evaluating clinical outcomes (ie, changes to pain and/or function) following running retraining interventions in symptomatic running populations;
Studies evaluating changes to lower limb biomechanics (kinetic, kinematic or neuromotor) during running in symptomatic and asymptomatic populations.
Studies with less than 10 participants in total or in the running retraining intervention group were excluded. This criteria were applied to minimise the risk of potentially false-positive or false-negative findings influencing the evidence synthesis.
MEDLINE, EMBASE, CINAHL and Current Contents were searched in June 2015. The search strategy and results for each data base can be found in online supplementary file 1.
Titles and abstracts found during the initial electronic search were uploaded into Endnote X6 (Thomson Reuters, New York, USA), duplicates removed, and each screened for inclusion by two independent reviewers (DRB and BSN). To resolve disagreement about exclusion, a third reviewer (CJB) was available. Where necessary, the full text was retrieved.
Two independent reviewers (CJB and DRB) evaluated the methodological quality of each included study using the Downs and Black quality index which consists of 27 items (maximum score of 28).31 A third reviewer (BSN) was available to resolve any disagreements. Based on quality assessment scores, studies were categorised as high quality (≥20), moderate quality (17–19) or low quality (≤16). Owing to the lack of randomised controlled trials and high-quality studies identified, meta-analysis was not performed. However, where possible, similar findings were combined in results tables for various retraining interventions, and the quality of these associated studies subsequently used to determine the level of evidence for each finding based on a modified version of the van Tulder et al32 criteria:
Strong=consistent findings among multiple studies including at least three high-quality studies;
Moderate=consistent findings among multiple trials, including at least three moderate-quality/high-quality studies or two high-quality studies;
Limited=consistent findings among multiple low-quality/moderate-quality studies, or one high-quality study;
Very limited=findings from one low-quality/moderate-quality study.
Prospective interview participants were initially identified from author lists of running retraining-related literature. Experts were required to be actively participating in running retraining research, have at least 5 years clinical experience, and be prescribing running retraining regularly to treat patients with running injuries. It was felt that experts with a good blend of clinical experience and research knowledge would be able to provide the best information on the perceptions of current evidence (published and unpublished), and its external applicability for clinical practice. Sixteen international experts from the UK, the USA, Canada and Australia were included and interviewed. Among them were 11 physiotherapists, 2 physical therapists, 1 sports physician, 1 medical doctor and 1 running coach. Further details related to participant characteristics are detailed in online supplementary file 2. Ethical approval was granted by La Trobe University's Faculty of Human Ethics Committee (FHEC13/151). Each participant provided informed consent.
One interviewer (JC) completed and recorded all interviews via Skype or in person where possible. Each was then transcribed for further analysis. The interviewer was a physiotherapist with 9 years clinical experience, and used running retraining interventions as part of their clinical practice. To facilitate discussion, a topic guide (see online supplementary file 3) was presented to each participant during the interview process. Content of the topic guide was based on a preliminary review of the literature and discussion between the research team.
Qualitative data were evaluated using a ‘Framework’ approach33 by a physiotherapist (CJB) with experience in conducting interviews and evaluating data related to qualitative research. Each transcript was read to gain familiarity, and then a thematic framework was formed by mapping the ideas and opinions stated by the interviewees and combining these to generate themes and subthemes, subsequently tabulated with each interviewee being coded to enable anonymous quote attribution. Additional interviews were performed until data saturation, whereby no new themes were identified. An additional physiotherapist (PM) and sports physician (AF-M) also read through each interview transcript to reinforce the analysis. Of particular interest was information related to the current evidence base, appropriateness of running retraining to treat lower limb injuries, specific lower limb conditions and associated retraining strategies which may be effective in clinical practice, and the practical application of specific running retraining strategies and their interaction with other interventions. Triangulation was by means of respondent validation and performed by presenting each interviewee with the final themes, subthemes and accompanying findings, and requesting any free comment. Any new comment was added to the framework analysis.
Synthesis of review findings with expert opinion
For the purpose of synthesising evidence with expert opinion, retraining interventions discussed and suggested for various lower limb injuries were tabulated along with illustrative quotes and potential biomechanical rationale identified in the systematic review.
After screening title and abstract, full texts were obtained for 70 studies. The primary reason for exclusion on obtainment of full text was inadequate participant numbers (ie, less than 10 participants per group), and included six studies on step rate manipulation,34–38 five studies on strike pattern comparison36 ,39–42 including one in a PFP population,41 one study on step width manipulation,43 one study on cues to reduce hip adduction,44 two studies on reducing impact loading variables45 ,46 and three combination studies.40 ,47–49 Additional reasons for exclusion included one combination study for absence of clinical or biomechanical analysis,50 one combination study for including transition to minimalist footwear during gait retraining,51 two step rate manipulation studies52 ,53 for absence of statistical comparison of biomechanical differences and one step rate manipulation study54 which combined hill running. A total of 46 studies met the inclusion criteria. A flow chart of the search results can be found in online supplementary file 1.
Results of the quality assessment can be found in online supplementary file 4. Of the 46 included studies, 13 were high quality, 25 were moderate quality and 8 were low quality. Of particular note, no study attempted to blind the participants or assessors; only 6 studies22 ,55–59 reported whether adverse events were experienced by participants, and 21 of the 46 studies did not report actual probability values. Additionally, 27 of the 46 studies lacked adequate justification of the statistical tests used to assess the main outcome data, commonly using parametric tests without describing whether the distribution of data was screened for normality prior to analysis. Accordingly, the use of parametric tests was considered inappropriate for such studies.
Evidence for clinical outcomes following running retraining
Four studies22–25 investigating clinical outcomes were identified (table 1). Limited evidence indicates 6-weeks of visual (video) and verbal feedback to transition from rearfoot strike (RFS) to midfoot strike (MFS) or forefoot strike (FFS) and increased step rate can reduce pain24 ,25 and compartmental pressures24 in patients with running-related anterior exertional lower leg pain.24 Limited evidence indicates 2-weeks (eight sessions) of visual and verbal feedback to reduce peak hip adduction in female patients with PFP possessing more than 20° peak hip adduction can reduce pain at 1-month22 ,23 and 3-month follow-up.22
Biomechanical effects of running retraining interventions
Forty-six studies evaluating the biomechanical effects of running retraining interventions met the inclusion criteria,22–25 ,55–96 including evaluation of step rate manipulation (see online supplementary file 5); altering strike pattern (see online supplementary file 6); proximal retraining strategies (see online supplementary file 7); and modifying impact loading variables, contact time and stiffness (see online supplementary file 8). Nineteen studies59–75 ,96 evaluated the biomechanical effects of step rate manipulation in isolation, with one including a symptomatic population,75 and one also examining the effects of altering contact time.70 Fifteen studies56–58 ,76–82 ,90 ,92–94 ,96 evaluated the biomechanical effects of altering strike pattern, all in asymptomatic participants. Two studies evaluated the biomechanical effects of a retraining intervention combining transition from rearfoot to forefoot or midfoot strike pattern with other retraining interventions including an increase to step rate24 and cues to alter proximal mechanics.25 Three studies55 ,83 ,95 evaluated the biomechanical effects of cues to reduce impact loading variables at foot strike. Four studies84–87 evaluated the biomechanical effects of altering step width, but none of these included symptomatic participants. Three studies evaluated the biomechanical effects of cues to alter proximal mechanics,22 ,23 ,88 with two aiming to reduce hip adduction,22 ,23 and one to increase forward trunk lean.88 One additional study89 evaluated the biomechanical effects of a combination of retraining strategies on biomechanics, but this was excluded from further analysis due to poor transparency regarding the specific intervention used.
Interview transcript analysis identified 10 sections with 29 themes and 75 subthemes. Three sections included ‘current evidence base’ (see online supplementary file 9.1; 2 themes, 9 subthemes), ‘appropriateness of running retraining’ (see online supplementary file 9.2; 1 theme, 4 subthemes) and ‘specific conditions’ (tables 2 and 3 and online supplementary file 9.3; 4 themes, 11 subthemes). Four sections were related to the practical application of running retraining strategies, including a reduction in overstride (ie, horizontal distance from foot strike to centre of mass (COM)) and increase in step rate (table 4; two themes, four subthemes); alteration to strike pattern (table 5; one theme, four subthemes); proximal retraining strategies (table 6; four themes, five subthemes); and approaches to modify impact loading variables, contact time and stiffness (table 7; two themes, four subthemes). Additional themes and subthemes included barriers and facilitators to running retraining implementation (see online supplementary file 10.1; 5 themes, 15 subthemes), the influence and relative importance of adjunctive interventions (see online supplementary file 10.2; 7 themes, 16 subthemes), and the potential for running retraining to prevent injuries or improve performance (see online supplementary file 10.3; 1 theme, 3 subthemes).
There is limited evidence to support running retraining in the treatment of exertional lower leg pain24 ,25 and PFP.22 ,23 Additionally, the expert panel were advocates of running retraining for cases of chronic or recurrent injury, and where potential biomechanical deficiencies linked to injury can be established (see online supplementary file 9.2). Experts interviewed suggested lower limb injuries which may benefit from running retraining include (but are not limited to) exertional lower leg pain, plantar fasciopathy, Achilles tendinopathy, calf pain, medial tibial stress syndrome, PFP, ITBS, patellar tendinopathy, hamstring injury including proximal tendinopathy and gluteal tendinopathy (tables 2 and 3; and online supplementary file 9.3). Despite a paucity of current evidence in injured populations, there is substantial evidence for the immediate biomechanical effects of running retraining interventions in an uninjured population (see online supplementary files 5–8). Given the current state of evidence, this information is important to assist guidance of current clinical practice and future research, and is incorporated throughout the discussion.
Conditions with limited empirical evidence
Limited evidence indicates that transitioning from a rearfoot to forefoot or midfoot strike pattern combined with increasing step rate, or altering proximal mechanics to facilitate hip flexion is effective in managing anterior exertional lower leg pain over a 6-week period, with positive clinical outcomes sustained for up to a year following the intervention24 ,25 (table 1). This approach was strongly supported by experts interviewed, and further biomechanical support is provided by very limited to limited evidence demonstrating reduced tibialis anterior muscle activity81 ,90 and forces80 with a forefoot strike (see online supplementary file 6), and limited evidence indicating reduced tibialis anterior muscle activity60 and ankle dorsiflexion61 ,71 ,72 around the time of foot strike with increased step rate (see online supplementary file 5). Additional running retraining recommendations for anterior exertional lower leg pain included other strategies to reduce overstride (eg, altering proximal mechanics), with some experts suggesting this may be more important than changing the strike pattern, and cues to reduce impact loading variables which may reduce ankle dorsiflexion at foot strike83 (see online supplementary file 8). Further research is needed to identify which running retraining strategies have the greatest effect on biomechanics and symptoms in individuals with exertional lower leg pain.
Limited evidence indicates visual and verbal feedback to reduce peak hip adduction in females with PFP possessing greater than 20° peak hip adduction is effective in its treatment,22 ,23 with positive clinical outcomes sustained for up to 3 months post-intervention22 (see table 1). This approach is generally supported by expert opinion (table 3), and recent prospective findings implicating excessive hip adduction as a risk factor for PFP.97 Reducing overstride through various approaches including increasing step rate was also suggested as important in the treatment of PFP (table 3). Importantly, there is strong evidence indicating reduced patellofemoral joint (PFJ) stress/load,68 ,75 ,96 moderate evidence indicating peak knee flexion,66 ,68 ,75 and limited evidence indicating reduced knee power absorption,63 ,66 internal knee extensor moments63 ,71 ,73 ,75 and peak hip adduction59 ,66 with an increased step rate (see online supplementary file 5).
Other suggested running retraining interventions for PFP included increasing forward trunk lean, which limited evidence indicates will reduced PFJ stress;88 and cues to reduce impact loading variables (see online supplementary file 8). Interestingly, transitioning from a rearfoot to a forefoot strike in runners with PFP did not emerge as a strong theme among experts, despite limited evidence indicating reduced peak and accumulative PFJ reaction force and PFJ stress when transitioning to forefoot striking.94 ,96 Additionally, reductions in running-related pain have previously been reported in a small case series of three runners with PFP when transitioning to a forefoot strike.41 Further research is needed to establish the potential therapeutic value of transitioning from rearfoot to forefoot striking in runners with PFP. Considering PFP is a multifactorial condition,98 it is likely that a tailored retraining strategy is needed in a clinical setting, and further research is required to understand the interaction of running retraining strategies with other evidence-based interventions.99
Conditions with support for running retraining from expert opinion
Poor agreement exists among experts on the value of running retraining in plantar fasciopathy treatment (table 2). Running retraining suggestions included reducing overstride, increasing step rate and feedback to reduce impact loading variables, based on evidence linking high loading rates to plantar fasciopathy.100 Supporting this proposal is limited evidence for reduced loading rates with cues to reduce impact loading variables55 ,83 (see online supplementary file 8) and increase step rate59 ,67 (see online supplementary file 5). Some experts recommended that transitioning from a rearfoot to a forefoot or midfoot strike to reduce loading rates81 may assist in plantar fasciopathy treatment, while others suggested that this may be detrimental due to increased arch strain in forefoot compared with rearfoot strike running79 (see online supplementary file 6). Further research on the effects of strike pattern manipulation in plantar fasciopathy is needed to provide clarity on these differing opinions.
Running retraining for Achilles tendinopathy and calf pain was cautiously recommended by some but not all experts interviewed (table 2), and adequate rehabilitation exercise was generally considered more important. Recommended retraining strategies for Achilles tendinopathy included reducing overstride and increasing step rate, which limited evidence indicates will reduced stance phase soleus muscle forces68 and ankle dorsiflexion at midstance68 (see online supplementary file 5). Additional considerations lacking supporting biomechanical evidence included reducing impact loading variables, increasing lower limb stiffness to reduce midstance ankle dorsiflexion, and increasing hip extension to reduce power requirements of the gastrocnemius-soleus complex during propulsion (see online supplementary file 8). Transitioning individuals with a pronounced forefoot strike to a rearfoot or midfoot strike was another proposed strategy for Achilles tendinopathy and calf pain (table 2), supported by very limited to limited evidence for reductions in plantar flexor impulses and internal joint moments,79 ,92 and gastrocnemius-soleus muscle forces80 (see online supplementary file 8). Running retraining strategies including reducing overstride and increasing step rate, reducing impact loading variables and transitioning strike pattern towards a midfoot strike was advocated by most experts for medial tibial stress syndrome (table 2). Additional consideration to reducing hip adduction and increasing step width was also suggested, which limited to very limited evidence indicates may reduce peak rearfoot eversion;84 ,87 rearfoot eversion and forefoot dorsiflexion excursion;87 peak internal ankle inversion moment;84 and tension, compression and shear stresses on the tibia86 (see online supplementary file 7).
Similar to PFP, cues to reduce peak hip adduction and internal rotation were recommended by some experts in the treatment of ITBS (table 3), an approach supported by findings that greater peak adduction in female runners may be a risk factor for ITBS development.101 Additionally, increasing step width to reduce cross-over gait (hip adduction at foot strike) was suggested (table 3), and this is supported by very limited to limited evidence for reduced peak internal knee abduction moment and impulse,84 ITB strain and strain rate,85 and peak hip adduction84 ,85 (see online supplementary file 7). There were inconsistent beliefs regarding the potential for running retraining to assist in the treatment of patellar tendinopathy (table 3), with some cautious recommendations to consider reducing overstride, increasing step rate, reducing impact loading variables, and transitioning from a rearfoot to a midfoot or forefoot strike. Limited-to-moderate evidence indicating reduced patellar tendon forces in midstance,68 internal knee extensor moments,63 ,73 ,75 knee power absorption63 ,66 ,96 and peak knee flexion66 ,75 ,68 supports increasing step rate (see online supplementary file 5). Biomechanical rationale to support a transition from a rearfoot to a forefoot strike is less clear, with limited evidence indicating a reduction in knee power absorption,82 ,92 but conflicting evidence related to internal knee extensor moment80 ,92 ,93 (see online supplementary file 6). This, combined with potential increases to tissue stresses more distally with a transition to a forefoot strike (eg, increased ankle power absorption; see online supplementary file 6), indicate that caution may be needed when considering this strategy for patellar tendinopathy and other conditions. Considering comparable reductions in knee, loading can be achieved by increasing step rate by 10%59; this may be a safer approach.
Running retraining for hamstring injury including proximal hamstring tendinopathy was generally supported (table 3). Recommendations included cues to reduce impact loading variables and anterior pelvic tilt, along with various methods to reduce overstriding, including increased step rate and promoting more hip and knee flexion during swing. While there is a paucity of biomechanical evidence examining the effects of promoting increased hip and knee flexion, there are a number of considerations from the biomechanical literature relating to increased step rate. First, limited evidence indicates greater hamstring muscle forces68 and activity during late swing60 with increased step rate (see online supplementary file 5), which in theory may place greater load on the hamstring muscle or tendon. However, limited evidence also indicates increasing step rate may reduce hip energy absorption63 ,66 and peak internal hip extensor moment66 ,71 (see online supplementary file 5), and thus reduce hamstring muscle and tendon load during early stance. Further research evaluating the effects of increasing step rate on symptoms in individuals with hamstring injury is needed to clarify the importance of these various biomechanical changes.
When managing gluteal tendinopathy, running retraining strategies including reducing overstride, increasing step rate, and cues to reduce hip adduction/internal rotation and contralateral pelvic drop were generally advocated (table 3). We did not find any studies evaluating strategies specifically targeting contralateral pelvic drop, but very limited to limited evidence indicates visual and verbal cues to reduce hip adduction also reduces contralateral pelvic drop22 ,23 and peak internal hip abduction moments22 up to 3 months following a 2-week intervention (see online supplementary file 7). When increasing step rate, very limited to limited evidence indicates reduced gluteal muscle forces,68 and hip energy absorption during stance;63 ,66 peak internal hip extensor,66 ,71 abduction66 and external rotation66 moments; and peak hip adduction59 ,66 (see online supplementary file 7).
Practical application of running retraining
Addressing overstride and increasing step rate
Addressing the presence of overstriding was considered one of the most beneficial running retraining strategies (table 4), with experts frequently discussing the importance of the foot landing closer to a runner's COM. Despite strong recommendation from experts, evidence to support the importance of overstriding or horizontal distance between foot strike and COM to running injury is lacking. Defining what constitutes overstriding (eg, the distance which may lead to injury), how this relates to injury and how to address it should be priorities for future research. Increasing step rate will move foot strike closer to the COM,66 and was suggested to be beneficial for many injuries. One key recommendation is that increases to step rate should be gradual (between 5% and 10% at a time) to ensure manageable changes and avoid excessive fatigue.66 Experts proposed a range of cues to increase step rate, including metronomes, music and verbal cues (eg, ‘take shorter faster steps’), and all could be considered clinically.
Strike pattern alteration
Experts proposed a range of potential benefits of strike pattern alteration, including both transition from rearfoot towards forefoot or midfoot (eg, knee injuries), and from forefoot towards rearfoot or midfoot strike (eg, Achilles tendinopathy and calf pain). However, caution was recommended when considering changing a runner's strike pattern, with a substantial adaptation period suggested if choosing to make alterations, especially towards a forefoot strike pattern (table 5). Additionally, potentially greater importance of other retraining interventions was stressed, including reducing overstride, which transitioning from a rearfoot to forefoot strike may be detrimental to.81 Additionally, other interventions including proximal retraining strategies, increasing step rate,61 ,71 ,72 and use of minimalist footwear or barefoot running102 were proposed to potentially have desired effects on foot strike.
Proximal retraining (cues to modify hip, pelvic and trunk mechanics)
A number of strategies additional to visual (3D motion feedback or mirrors) and verbal (‘tighten your buttocks’ and ‘point your knees straight ahead’) cues published in the literature were proposed to reduce hip adduction and internal rotation (table 6). Considerations to both internal (eg, squeeze gluteals) and external (eg, point knees forward) cues were apparent, but consensus was lacking regarding the most appropriate approach, and it is possible the most effective cue may be specific to the individual. A number of retraining interventions may indirectly reduce peak hip adduction, including increasing step rate66 or step width.84 ,85 Possible cues to increase step width used in the literature include both visual (taped lines on the floor)84 ,87 and verbal instruction (eg, ‘widen your stance’).85 ,86 Although verbal instruction may be easier to implement in clinical practice, experts stressed the need to ensure individuals do not ‘overcorrect’ (table 6). Further research is needed to determine the most effective cues for reducing hip adduction and increasing step width and its effect of running-related pain.
Reducing impact loading variables
Some experts considered kinetic variables, including joint forces, tibial accelerations and loading rates, may be more important in relation to injury than kinematic variables (table 4), although evidence to support this premise is lacking. Additionally, most clinicians lack the required equipment to directly measure kinetics during running. It was suggested that impact noise during running may be closely related to actual impact loading variables such as vertical loading rates, and therefore may be used as a proxy measure. Additionally, moderate relationships between sagittal plane kinematics and subsequent loading (kinetic) variables during stance of running were recently reported, including increased step rate, decreased vertical COM excursion, decreased foot inclination angle at foot strike, and reduced horizontal distance between foot strike and COM (ie, decreased overstride).103 Considering these variables along with strategies to reduce impact noise may be a focus of retraining interventions where clinicians seek to reduce impact loading variables.
Barriers and facilitators
Pain and irritability may prevent the implementation of running retraining in the early stages of rehabilitation (see online supplementary file 10.1 which summarises expert discussion of barriers and facilitators to running retraining implementation). Inconsistent beliefs related to whether a runner should be allowed to run in pain or not were evident. Importantly, experts suggesting running-related pain could be allowed, believed so only if pain was minimal (eg, 3/10). This fits with previous research in Achilles tendinopathy, which indicates that continued sports participation with minimal pain did not impair rehabilitation outcomes if guided by a similar pain monitoring model.29 An additional consideration is that pain reduction may provide vital feedback to aid compliance to running retraining interventions.104 Other key potential barriers discussed included muscle function capacity, joint flexibility and skeletal structure. Further research is needed to determine how these variables impact on biomechanical and clinical outcomes with running retraining. Fatigue resulting from running retraining was suggested to limit how much could be changed at once (eg, magnitude of step rate increase), and the duration runners may be able to maintain these changes. However, it was suggested that fatigue associated with running retraining was likely to diminish with habituation.
Of the studies providing limited evidence for running retraining to reduce pain,22–25 two evaluated a programme including eight sessions of up to 30 min over a 2-week period,22 ,23 another used 18 sessions over 6-weeks24 and one used just 3 sessions over 6-weeks.25 Three22–24 of these four approaches are unlikely to reflect clinical practice due to associated costs they would entail, and are more intensive then those proposed by experts interviewed. The optimum number of sessions and time frame suggested to facilitate required changes varied among experts, and it was suggested this may be different for each runner. Some runners may be able to make clinically significant changes immediately, while others may take many months due to barriers such as pathology and associated pain, weakness, flexibility deficits, and intrinsic motivation. Importantly, experts highlighted these barriers may need to be addressed concurrently or prior to attempts to retraining running technique. Variable beliefs exist regarding the potential for changes as a result of running retraining to become habitual. Although few studies have explored this, limited evidence indicates a reduction in both peak hip adduction and running-related pain at 1–3 months follow-up can be achieved in females with PFP.22 ,23 Further longer term studies are needed to determine if habituation from running retraining interventions in injured runners can be facilitated.
Limiting cues to one or two at a time was recommended. Additionally, an individual response to cues is considered likely, meaning clinicians need to be able to adapt and tailor cues to ensure desired biomechanical changes. Video analysis and feedback may assist compliance, allowing visualisation of running technique before and after the running retraining. Use of mirrors was suggested by some experts to assist reductions in hip adduction and knee valgus, and changes to trunk position and strike pattern. However, this was not recommended by all experts interviewed. With just one study published supporting the use of mirrors in running retraining,22 further research is needed to establish their usefulness, along with other forms of clinically feasible biofeedback.
Undertaking running retraining on a treadmill was suggested to be practically easier than overground, although potential limitations in carry over to overground running gait were acknowledged. Vertical ground reaction forces have been reported to be similar between conditions.105 However, treadmill running has been reported to decrease peak and range of knee flexion,106–111 decrease ankle dorsiflexion range of motion, velocity and peak,111 ,112 and increase peak rearfoot/ankle eversion110 ,111 ,113 when compared with overground running. Further research is needed to establish to what extent clinical and biomechanical effects of running retraining on a treadmill translate to overground running.
Experts highlighted that running retraining is only part of the solution for the treatment of running injuries (see online supplementary file 10.2 which summarises expert discussion of adjunctive interventions during running retraining implementation). The importance of concurrently addressing muscle function and flexibility deficits was frequently discussed. References were made to both strength and motor control, with the latter being considered by some to be more important to an individual's ability to make changes to running technique. Manual therapy to address soft tissue flexibility and joint mobility restrictions was also considered as a potential adjunctive intervention by some experts. Further research to establish the impact of addressing muscle function and flexibility deficits through exercise and manual therapy on running retraining outcomes is needed.
The influence of footwear on injury and running retraining appears to be a divisive topic. While some experts believed footwear had minimal impact on outcomes, others considered footwear as vitally important. Those emphasising the importance of footwear generally promoted a more minimalistic approach, to assist in reducing impact loading variables and overstride, and transitioning towards a midfoot or forefoot strike. Importantly, caution and slow transition towards minimalist footwear was encouraged if choosing to implement this as an intervention, which is supported by recent evidence indicating an increased risk of pain and injury,114 and foot bone marrow oedema115 during 10–12-week transition periods. Barefoot running was suggested by some experts to potentially facilitate increased step rate, a more forefoot or midfoot strike, and reduced impact loading variables. Biomechanical rationale supports this premise,102 but importantly, some experts suggested barefoot running may be injurious if the transition is too fast or running retraining is not concurrently implemented. The interaction of barefoot running and footwear modification with running retraining interventions requires investigation.
Substantial supporting evidence for the effectiveness of foot orthoses to treat lower limb injuries exists.16 ,17 However, the perceived value of foot orthoses in running injury treatment varied among experts, with some considering them only as a ‘last resort’. Additionally, those advocating their use generally saw them as only a temporary solution or a tool to facilitate desired running retraining strategies. Taping was viewed in similar light, with its value perceived to be in reducing symptoms in the short term and facilitating desired running retraining changes.
Prevention, performance and ideal running pattern
Some experts suggested running retraining may play a role in injury prevention, although a lack of available evidence to guide implementation exists (see online supplementary file 10.3 which summarises expert discussion of prevention, performance and and ideal running pattern in relation to running retraining). One commonly discussed approach was reducing impact loading variables such as vertical loading rates, indicating research to investigate the merits of this may be warranted in the future. When discussing running performance and economy, experts suggested acute changes may reduce running economy. However, some, but not all experts, suggested that economy may then improve over the first couple of months, with possible improvements to performance in the longer term with interventions such as increasing step rate.
Beliefs related to the presence of an ideal running pattern in relation to both injury and performance were inconsistent, with some experts suggesting this to be individual to each patient. Others stated certain running characteristics could be considered better than others for all runners, including higher step rates and a forefoot/midfoot strike. The topic of strike pattern is controversial, with the recent growing interest of barefoot running leading to a strong promotion of rearfoot runners to transition to a forefoot strike to both treat injury and improve performance.116 Considering the lack of evidence to support this notion in relation to injury102 and performance,42 ,79 further research is needed on this topic.
Limitations and future research directions
This study provides what might be reported as a biased sample of ‘experts’ because we selected those who use and study running retraining. A sports biomechanics researcher or clinician who has moved away from this area of research because he/she feels it is not effective would not be included among our ‘experts’. The topic guide (see online supplementary file 3) used to facilitate discussion between the researcher and interviewee was based primarily on literature included in the systematic review, which may have biased discussions towards topics with current biomechanical evidence. To address this, the topic guide included scope for and encouraged discussion of clinical reasoning regarding interventions lacking evidence.
Initial evaluation and framework analysis was completed by an experienced physiotherapist with more than 5 years’ experience of providing running retraining as a specialist focus in clinical practice (CJB), which is another potential source of bias. To remove one level of bias, data accuracy and interpretation were checked by an additional researcher with qualitative research experience (PM), and two other researchers were also involved (AF-M and JC). Validity of qualitative findings was then further strengthened through triangulation of findings via respondent validation.
Only 5 of 46 studies included investigated an injured population, limiting the clinical applicability of findings from the evidence synthesis. We believe randomised controlled trials of running retraining efficacy in symptomatic populations are justified on the basis of current evidence. Such trials should also consider evaluating potential clinical predictors and biomechanical mechanisms, which may help tailor running retraining interventions to individuals and specific conditions. Other priorities highlighted by experts include development of technologies to provide more efficient feedback, developing options to allow mobile monitoring of compliance, and investigating the interaction of running retraining with other evidence-based interventions (see online supplementary file 9.1).
Expert clinical reasoning has been combined with a comprehensive evidence synthesis to guide clinicians and researchers who seek to implement and evaluate running retraining in the treatment of lower limb injury. Various options require consideration, including strategies to reduce overstride and increase step rate, altering strike pattern, reducing impact loading variables, increasing step width and altering proximal kinematics. Currently, limited evidence supports the effectiveness of transition from rearfoot to forefoot strike and increase step rate or altering proximal mechanics in individuals with anterior exertional lower leg pain, and visual and verbal feedback to reduce hip adduction in females with PFP. According to current expert opinion, other lower limb injuries which may benefit from running retraining include plantar fasciopathy, Achilles tendinopathy, calf pain, medial tibial stress syndrome, ITBS, patellar tendinopathy, proximal hamstring tendinopathy and gluteal tendinopathy. Tailoring approaches to each injury and individual is likely to be required to optimise outcomes, and running retraining should only be considered part of the solution when managing running injuries.
Tailoring running retraining strategies to each injury and individuals is needed to optimise outcomes.
What are the findings?
There is limited evidence for running retraining in the treatment of patellofemoral pain and anterior exertional lower leg pain.
Based on sound biomechanical rationale, running retraining may assist in the treatment of lower limb injuries including exertional lower leg pain, plantar fasciopathy, Achilles tendinopathy, calf pain, medial tibial stress syndrome, patellofemoral pain, iliotibial band syndrome, patellar tendinopathy, hamstring injury including proximal tendinopathy and gluteal tendinopathy.
The running retraining options that clinicians and patients might consider in clinical practice include strategies to reduce overstride and increase step rate, altering strike pattern, reducing impact loading, increasing step width and altering proximal kinematics.
Substantial evidence exists for the immediate biomechanical effects of running retraining interventions in uninjured populations.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Twitter Follow Christian Barton at @DrChrisBarton
Contributors CJB conceived the study concept and design. DRB, BSN and CJB completed literature searching. CJB and DRB completed quality assessment. CJB completed evidence synthesis. JC completed all interviews. CJB, PM and AF-M were responsible for qualitative data analysis and initial interpretation. All authors contributed to interpretation of findings, manuscript write up and final approval.
Funding Interview transcription was funded by La Trobe University's Lower Extremity Gait Studies (LEGS) programme.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.