Article Text

Download PDFPDF
Optimising the efficacy of gait retraining
  1. Irene Davis
  1. Correspondence to Professor Irene Davis, Department of Physical Medicine and Rehabilitation, Harvard Medical School, Boston, Massachusetts 02138, USA; isdavis{at}

Statistics from

A recent randomised control trial compared the effects of three rehabilitation programmes for patellofemoral pain: education, education and exercises and education and gait retraining.1 I was excited to read this, as I consider gait retraining an important component of patellofemoral pain treatment. Physical therapists are movement experts and in the best position to address faulty movement patterns. Many running-related injuries, especially those more chronic and requiring medical attention, are associated with faulty running mechanics.2 Muscle strengthening alone may not address faulty mechanics.3 While muscle strength is important and can be considered the ‘hardware’ of movementit is the motor pattern, or the ‘software’, that dictates the movement quality. When movement patterns become habitual, altering them during high demand, functional activities such as running takes time and practice and can be challenging.

The goal of gait retraining is to adopt and retain a new motor pattern. Research suggests that this requires an acquisition phase, where extrinsic feedback is provided on a predetermined schedule so that the learner can associated the feedback with their proprioceptive input.4 This is followed by a transfer phase, where the feedback is gradually faded to enable the learner to execute the motor pattern independent of feedback and rely on internal cues. Additionally, to discern whether a motor skill has been learned, it should persist under distraction, transfer to another environment (eg, treadmill versus overground) or to another skill.5 This indicates that the motor pattern no longer requires conscious effort to execute. For example, in Willy et al, the reduced hip adduction following the running retraining was also observed in a single leg squat.6

Several aspects of the running training in the current study may have improved the success of this intervention. Runners were only provided five 10 min treadmill sessions with ‘physiotherapist feedback’, either to increase their cadence or run softer with a forefoot strike pattern. It is advisable to specify the feedback time, whether the feedback was verbally scripted, controlled between patients or faded over time is needed. There are numerous published of this structured feedback approach using established motor control principles,6–8 including a progressive increase in runtime from 10 to 30 min over eight sessions, with faded feedback in the last four sessions. Finally, it is advisable to provide evidence regarding whether the gait pattern had been ‘retrained’ or whether the runner was simply able to reproduce the cadence when tested.

Different retraining techniques (ie, changing footstrike pattern vs increasing cadence) may affect the retraining effects. Adopting a forefoot strike (FFS) pattern results in very different force profile than increasing cadence. A FFS pattern also increases foot and ankle load. Therefore, this change is best accompanied by a foot and ankle strengthening programme to minimise injury risk. This is especially important in the current trial, where runners were maintaining their normal mileage and not given time for muscle adaptation. This lack of preparation may have contributed to the reported calf injury. Additionally, one-third of the patients were non-rearfoot strikers at baseline, but it is unclear how many were FFS versus midfoot strikers. A FFS pattern is typically associated with shorter strides and higher step rates. The ability to increase cadence in these runners is likely limited, and it is possible that an alternate approach was needed. Finally, non-rearfoot strikers (especially FFS) naturally have low vertical loadrates9 creating a floor effect at baseline, thus reducing potential for further reduction. The study would have had fewer confounding issues if only rearfoot strike (RFS) runners were included and if either increasing cadence or transitioning to a FFS was used.

The authors suggest that the success of previously published retraining studies using a structured eight session, faded feedback designs were partly a result of reducing overall load. However, Roper et al 8 noted a significantly greater reduction in pain in the structured retraining group compared the group where only load was controlled. Furthermore, other studies reported that all runners returned to their pretraining running volume following the intervention and with pain levels ranging from 0 to 0.5/10 Visual Analogue Scale (VAS) post-training, including 1 and 3 month follow-ups.6 8 Willy et al reported that most participants were running even higher volumes at the 3 month follow-up.6 These results challenge the notion that the results of these studies were a function of load reduction. Additionally, while pain was reduced in this current randomised controlled trial, patients were still running with an average pain of 3.5/10 VAS following their retraining compared with nearly 0/10 in previously published structured programmes. The average pain value exceeded the level (VAS 2/10) that runners were instructed to maintain as part of their education. This raises questions on adherence to the education component of this intervention, which is a fundamental question of all trials.

In summary, I commend the authors for conducting this comparative study of several interventions for runners with patellofemoral pain. These types of intervention studies with long-term follow-up are challenging to conduct, but greatly needed to guide physical therapy practice. However, the study could have been improved with inclusion of only RFS runners and one gait retraining intervention. Additionally, a more structured gait retraining approach using motor control principles may have resulted in greater pain resolution, as shown in prior studies. These issues should be considered when designing future gait retraining studies to provide a valid assessment of their efficacy.



  • Competing interests None declared.

  • Provenance and peer review Commissioned; internally peer reviewed.

Request Permissions

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.

Linked Articles