Gait Retraining for the Reduction of Injury

Occurrence in Novice Distance Runners

1-Year Follow-up of a Randomized Controlled Trial
Zoe Y.S. Chan,*y BEng, Janet H. Zhang,y MBBS, Ivan P.H. Au,y BSc, Winko W. An,z MEng,
Gary L.K. Shum,§ PT, PhD, Gabriel Y.F. Ng,y PT, PhD, and Roy T.H. Cheung,y PT, PhD
Investigation performed at the Gait & Motion Analysis Laboratory,
Department of Rehabilitation Sciences, The Hong Kong Polytechnic University, Hong Kong

Background: The increasing popularity of distance running has been accompanied by an increase in running-related injuries,
such that up to 85% of novice runners incur an injury in a given year. Previous studies have used a gait retraining program to
successfully lower impact loading, which has been associated with many running ailments. However, softer footfalls may not nec-
essarily prevent running injury.
Purpose: To examine vertical loading rates before and after a gait retraining program and assess the effectiveness of the program
in reducing the occurrence of running-related injury across a 12-month observation period.
Study Design: Randomized controlled trial; Level of evidence, 1.
Methods: A total of 320 novice runners from the local running club completed this study. All the participants underwent a baseline
running biomechanics evaluation on an instrumented treadmill with their usual running shoes at 8 and 12 km/h. Participants were
then randomly assigned to either the gait retraining group or the control group. In the gait retraining group (n = 166), participants
received 2 weeks of gait retraining with real-time visual feedback. In the control group (n = 154), participants received treadmill
running exercise but without visual feedback on their performance. The training time was identical between the 2 groups. Partic-
ipants’ running mechanics were reassessed after the training, and their 12-month posttraining injury profiles were tracked by use
of an online surveillance platform.
Results: A significant reduction was found in the vertical loading rates at both testing speeds in the gait retraining group (P \
.001, Cohen’s d . 0.99), whereas the loading rates were either similar or slightly increased in the control group after training
(P = .001 to 0.461, Cohen’s d = 0.03 to 20.14). At 12-month follow-up, the occurrence of running-related musculoskeletal injury
was 16% and 38% in the gait retraining and control groups, respectively. The hazard ratio between gait retraining and control
groups was 0.38 (95% CI, 0.25-0.59), indicating a 62% lower injury risk in gait-retrained runners compared with controls.
Conclusion: A 2-week gait retraining program is effective in lowering impact loading in novice runners. More important, the
occurrence of injury is 62% lower after 2 weeks of running gait modification.
Registration: HKUCTR-1996 (University of Hong Kong Clinical Trials Registry).
Keywords: running; kinetics; biofeedback; injury prevention

Running is a popular sport globally. The rapid growth of with 37% to 79% of runners sustaining an injury in a given
the running population can be partially reflected by the year.4,14 This means that 3 of 4 regular runners will incur
number of participants in many distance running events an injury within 3 years. Compared with elite runners, nov-
worldwide. In 2015, 17.1 million finishers participated in ice runners are more vulnerable,11 partially because they are
more than 30,000 races held in the United States.35 This less physically prepared for distance running.7 In view of this
increase in the popularity of running can be explained by situation, studies of physical training programs to prevent
its positive effect on the cardiovascular and mental health running-related injuries have been undertaken, and the
of runners.44 However, due to the repetitive nature of run- effectiveness of such programs has been questioned.6,7,30,43
ning, running-related musculoskeletal injuries are common, The findings of previous studies clearly indicated that
a physically conditioned runner under a structured training
protocol may still be at risk of injury if the biomechanical
The American Journal of Sports Medicine risk factor is not addressed.
1–8
DOI: 10.1177/0363546517736277
Investigators have studied the relationship between bio-
Ó 2017 The Author(s) mechanics and running-related injury. Among different

1
2 Chan et al The American Journal of Sports Medicine

biomechanical risk factors, such as the magnitude of METHODS

ground-reaction force peaks,41 a high level of vertical load-
ing has been reported to be associated with many injury Study Design and Participants
conditions in runners, such as patellofemoral pain,10,14 tib-
ial stress fractures,3,28 and plantar fasciitis.28 Greater ver- This laboratory-based study was a single-blinded, random-
tical average loading rate (VALR) or vertical instantaneous ized controlled trial. The experimental procedure was
loading rate (VILR) is caused by an increased vertical body reviewed and approved by the Departmental Research Com-
stiffness during landing.18,20 It has been suggested that an mittee of the Department of Rehabilitation Sciences, The
increased vertical stiffness is associated with injury Hong Kong Polytechnic University, and the trial was regis-
because a greater force acts on the body over a smaller tered at HKU Clinical Trials Registry (Ref: HKUCTR-1996).
joint excursion, which causes poor shock attenuation. A total of 412 novice runners (\2 years of running experi-
Many running techniques, such as Chi running and Pose ence) who regularly run more than 8 km/wk and were
running, target the modification of running gait for a softer aged 18 to 50 years were recruited in this study. Partici-
landing.15,33 However, information about the ability of pants were free from any active injury for at least 6 months
these methods to modify running gait is mainly anecdotal. before the study. To avoid a floor effect, all the participants
Previous studies have used a gait retraining program of underwent an initial running screening and those with
8 sessions in 2 weeks using real-time visual feedback to VALR less than 70 body weight (BW) per second during
control impact loading.21,29 In this training protocol, par- usual speed running were excluded.
ticipants ran on a treadmill and the training time in each
session was gradually increased from 15 to 30 minutes Baseline Measurements
over the 8 sessions, while the real-time visual feedback
was progressively removed in the last 4 sessions. Partici- All participants who met the study criteria and provided
pants’ impact loading was reduced 18% to 20% after the written consent underwent a baseline running biomechan-
training, and this reduction was maintained at the 1- ics assessment. They were asked to run on an instru-
month follow-up in a feedback-free state.24 Other biofeed- mented treadmill (AMTI force-sensing tandem treadmill)
back gait retraining programs using the same training at 8 km/h (slow pace) and 12 km/h (fast pace) for 5 minutes
and feedback-weaning protocol have been applied to other with their usual running shoes. The test sequence was ran-
cohorts and were shown to cause a favorable running gait domized by use of an online program (www.random.org),
pattern transition.13 Although the running biomechanics and a 5-minute rest period was provided between the 2
were not exactly identical between treadmill and over- running trials.
ground conditions, translation of the training effect from Ground-reaction force data were sampled at 1000 Hz for
treadmill-based training to overground running has been the last minute of the run. Data were then filtered using
observed in previous gait retraining studies.34 One plausi- a second-order, recursive Butterworth low-pass filter at
ble explanation was the comparable neuromuscular con- 50 Hz. A threshold of 10 N in the vertical ground-reaction
trol27 and kinetics32 between the 2 conditions, favoring force was used to determine foot-strike and toe-off. The
the translation of the training effect to the alternative run- VALR and VILR were obtained by the method described
ning environments. in a previous study.12 In brief, VALR and VILR were the
However, favorable running biomechanics may not average and maximum slopes of the line through the 20%
equate to injury-free running. Hitherto, no published stud- point and the 80% point of the vertical impact peak, respec-
ies have examined the effect of a gait retraining program tively. In case of an undetectable or absent vertical impact
on injury prevention in novice runners. Therefore, this peak within 1 stance phase, the vertical impact peak value
randomized controlled trial sought to evaluate the effec- would be taken as the force at 13% stance phase.4 Both
tiveness of a gait retraining program on modulation of VALR and VILR were normalized by body weight and
impact loading and to determine whether it can prevent averaged across all footfalls within the 1-minute trial.
running-related injury in a group of novice runners. We
hypothesized that participants receiving gait retraining Sample Size
would present lower VALR and VILR during running. In
contrast, the vertical loading rates would remain similar The required sample size was calculated for the primary
in the control group. It was also hypothesized that gait outcome variable, the annual occurrence of running-
retraining would lower the occurrence of running-related related musculoskeletal injury. According to previous stud-
injury when compared with the controls. ies, the occurrence varied between 37% and 79% in a given

*Address correspondence to Zoe Y.S. Chan, BEng, Gait & Motion Analysis Laboratory, ST004, G/F, Core S, The Hong Kong Polytechnic University,
Hung Hom, Hong Kong (email: [email protected]).
y
Department of Rehabilitation Sciences, The Hong Kong Polytechnic University, Hong Kong.
z
Department of Biomedical Engineering, Boston University, Boston, Massachusetts, USA.
§
Faculty of Sport & Health Sciences, University of St Mark & St John, Plymouth, UK.
One or more of the authors has declared the following potential conflict of interest or source of funding: This study was supported by the Health & Med-
ical Research Fund (Project No. 12131621), Food & Health Bureau, the Government of the HKSAR.
AJSM Vol. XX, No. X, XXXX 1-Year Follow-up After Gait Retraining in Novice Runners 3

Figure 1. Runners receiving visual biofeedback during gait retraining. They were asked to reduce the vertical impact peak by
softening their footfalls.

year.4,14 A reduction of 25% of the occurrence in the gait

retraining group compared with the control group was con-
sidered clinically significant and relevant.6 A logistic rank
surviving power analysis was performed with a hypothe-
sized 25% reduction of the annual occurrence, an attrition
rate of 5%, a power of 80%, and an alpha level of 5%; it was
determined that 380 runners (190 in each group) were
needed to detect an effect of the 2-week gait retraining
program.

Randomization
After the baseline measurement, all participants were
assigned to either the gait retraining group or control
group. To ensure that the participants were matched
between the 2 groups, a stratified randomization was per-
formed. Participants were stratified for current running
Figure 2. Training time and biofeedback time arrangement
mileage (8-12 km/wk; 12-16 km/wk; .16 km/wk) and sex.
in the gait retraining group.
A block size of 4 was used in the randomization sequence.
For each stratum, participants were allocated by drawing
a sealed opaque envelope. Participants were asked to ‘‘run softer’’ so that the amplitude
of vertical impact peak would be reduced or even diminished
(Figure 1). The training time was gradually increased from
Gait Retraining Group
15 minutes to 30 minutes over the 8 sessions, and visual feed-
Participants in the gait retraining group underwent a 2-week back was progressively removed in the last 4 sessions (Figure
gait retraining for landing stiffness modulation according to 2). The participants were then advised to maintain their new
the protocol established in a previous study.10 In brief, they gait pattern during their daily living or regular running prac-
participated in 8 sessions of gait modification over 2 weeks tice after the training.
(4 sessions per week). During the training, participants
were asked to run at a self-selected speed on an instrumented Control Group
treadmill (AMTI force-sensing tandem treadmill). Visual bio-
feedback in the form of a vertical ground-reaction force signal Similar to the gait retraining group, participants in the
from the treadmill was displayed on the monitor in front. control group were invited to the laboratory for 8 times
4 Chan et al The American Journal of Sports Medicine

TABLE 1
Baseline Characteristics of Participants in the Gait Retraining and Control Groupsa

Characteristics Gait Retraining (n = 166) Control (n = 154) P

Sex 82 males, 84 females 76 males, 78 females .993

Age, y 33.6 6 9.5 34.2 6 9.5 .559
Weight, kg 60.0 6 12.6 61.6 6 12.0 .235
Height, m 1.66 6 0.09 1.65 6 0.09 .843
Running experience, mo 16.8 6 5.2 16.6 6 5.0 .720
Weekly mileage, km 19.5 6 7.0 18.5 6 6.1 .172
VALR at 8 km/h, BW/s 65.95 6 9.90 67.81 6 9.97 .094
VALR at 12 km/h, BW/s 81.28 6 13.59 83.51 6 11.41 .115
VILR at 8 km/h, BW/s 90.69 6 13.90 92.32 6 10.81 .245
VILR at 12 km/h, BW/s 111.87 6 14.51 114.32 6 16.42 .160

a
Data are reported as mean [plus or minus] SD, unless otherwise indicated. BW, body weight; VALR, vertical average loading rate; VILR,
vertical instantaneous loading rate.

in 2 weeks. They were asked to run on an instrumented conducted if necessary. In addition, to avoid overreliance on
treadmill at a self-paced speed, but no feedback on their statistical tests,26 the effect size, in terms of Cohen’s d, was
running biomechanics was provided. The running time used to quantify the strength of comparisons. Cohen’s d val-
was identical to the protocol in the gait retraining group. ues around 0.2, 0.5, and 0.8 are considered as small, medium,
and large effect sizes, respectively.39 Since the current study
was not designed to investigate the effects of gait retraining
Reassessment on any particular injury type, the injury patterns in the 2
study groups were compared descriptively. Mantel-Cox test
All participants were reassessed 2 weeks after the first was used to compare the survival curves of the participants
evaluation. The testing procedure was identical to the with an injury in the gait retraining group and the control
baseline assessment. group. A Cox proportional hazards regression was conducted
to assess the difference in the occurrence of injury during the
12-month follow-up period after training. All analyses were
Tracking of Injury Occurrence performed following the ‘‘intention to treat’’ principle. All sta-
tistical tests were performed by SPSS software (version 23;
After the training program was completed, all participants
SPSS Inc), with level of significance set as .05.
were asked to log into an online running injury surveil-
lance platform, which was designed based on a previous
study.1 At the first login, they were required to report their
injury history and average weekly mileage over the past 6 RESULTS
months. At each of the 12 subsequent logins for each
month, they were asked to report their weekly mileage, Initially 412 participants volunteered for this study; 22 of
other training program involvement, and injuries (if any) them were excluded due to the preset criteria (Figure 3).
over the past month. They were required to specify the per- After stratified randomization, 195 runners were allocated
son who made the diagnosis for their injuries. An injury to the gait retraining group and another 195 runners were
was operationally defined as any running-related musculo- assigned to the control group. Of these 390 participants,
skeletal complaint40 that was diagnosed by a medical pro- 320 completed all follow-up assessments; 70 dropped out
fessional, such as a physician, physical therapist, or at different stages due to scheduling conflicts or personal
orthopaedic surgeon, and that caused the participant to reasons. No between-group differences in any demographic
miss at least 2 days of training. To ensure validity of the or baseline outcomes were found (P values . .094; Table 1).
injury data, those who had reported an injury were con- Participants in both groups reported no adverse effects.
tacted by a researcher to authenticate the injury incident. The 2 3 2 mixed-design ANOVA revealed significant inter-
action effects between training and time for both VALR (P \
.001, h2p = 0.344-0.367) and VILR (P \ .001, h2p = 0.353-
Statistics 0.541) at both testing speeds. Pairwise comparisons
revealed a significant reduction in VALR (P \ .001, Cohen’s
Baseline characteristics of participants in the gait retraining d = 1.06-1.12) and VILR (P \ .001, Cohen’s d = 0.99-1.01)
and control groups were compared by use of 2-tailed t tests after gait modification (Figure 4). In the control group, no
and chi-square statistics for continuous and discrete varia- significant difference was found in VALR at 8 km/h after
bles, respectively. A 2 3 2 mixed-design analysis of variance training (P = .461), but VALR at 12 km/h and VILR at
(ANOVA) was used to compare the interaction effect of train- both testing speeds were increased (P \ .029; Cohen’s
ing (gait retraining vs control) and time (before and after d = 20.09 to 20.14; Figure 4). For between-groups compar-
training) on VALR and VILR. Pairwise comparisons were isons, VALR and VILR in the gait retraining group were
AJSM Vol. XX, No. X, XXXX 1-Year Follow-up After Gait Retraining in Novice Runners 5

TABLE 2
No. (%) of Running-Related Injuries
in Gait Retraining and Control Groups

Condition Gait Retraining Group Control Group

Patellofemoral pain 4 (14) 18 (29)

Plantar fasciitis 2 (7) 23 (38)
Iliotibial band syndrome 3 (11) 8 (13)
Hamstrings strain 3 (11) 8 (13)
Achilles tendinitis 5 (18) 0 (0)
Calf strain 5 (18) 0 (0)
Shin splints 3 (11) 1 (2)
Patellar tendinitis 2 (7) 0 (0)
Meniscal injury 1 (3) 3 (5)

study (Cohen’s d = 0.99-1.12). Such discrepancy can be

explained by the instruction and feedback provided to par-
ticipants. Most of the previous studies used an explicit and
visible biomechanical parameter as a marker for the bio-
feedback training, such as foot-strike pattern,10,38 stride
frequency,17 or lower limb alignment.24 These modifica-
Figure 3. Consort diagram. tions could be observed and measured without the use of
sophisticated laboratory equipment, and thus runners
could attempt or practice modifications outside the train-
significantly lower than in the control group at both testing ing sessions, possibly enhancing the effect of the retrain-
speeds after training (P \ .001, Cohen’s d = 1.16-1.52). ing. This speculation is supported by the fact that
At the 12-month follow-up, 16% and 38% runners another study using an implicit parameter, tibial shock,
reported running-related musculoskeletal injury in the reported a smaller reduction of VALR and VILR (Cohen’s
gait retraining group and control group, respectively. The d = 1.3-1.7) after gait retraining.12 Even so, studies relat-
types of injuries reported differed between gait retraining ing attentional focus and motor learning suggested that
and control groups (Table 2). We observed more Achilles feedback promoting an external focus was more effective
tendinitis (18%) and calf strain (18%) in the gait retraining than feedback promoting an internal focus on both the
group, while no such injuries were observed in the control learning outcome and the retention.45,46 In the present
group. The most common injuries in the control group were study, participants were provided with real-time exter-
plantar fasciitis (38%) and patellofemoral pain (29%), nally focused feedback (ie, vertical ground-reaction force)
while only 7% and 14% of participants in the gait retrain- without instructions on the detailed movements required
ing group had these conditions. Mantel-Cox test indicated to achieve a reduced impact peak. This arrangement was
a significant difference in the survival curves between considered to be optimal for gait retraining and to favor
the 2 groups (Figure 5). The hazard ratio between the retention during the follow-up period.
gait retraining and control groups was 0.38 (95% CI, In the present study, unlike previous studies in which
0.25-0.59), indicating a 62% lower injury occurrence in the assessment and training speeds were set by research-
gait-retrained runners compared with controls. ers, participants completed the gait retraining at their
own training pace. They also wore their own usual running
shoes, such that the training was performed in a setting
DISCUSSION that best imitated their natural training conditions. This
design was intended to minimize the effect of speed and
This single-blinded, randomized controlled trial sought to footwear change on loading rates9,22 and ensure sustain-
evaluate the effectiveness of a laboratory-based gait ability of the modified gait when participants returned to
retraining program on the impact of loading control and their regular training.
running-related musculoskeletal injury prevention in nov- Lower VALR or VILR after gait retraining is achieved
ice runners. In accordance with our original hypotheses, by a reduction in the vertical body stiffness during
gait retraining was a safe and effective intervention to impact.18,20 The relationship between stiffness and run-
lower VALR and VILR during running. More crucially, ning injury is well established in animal models but not
the laboratory-based gait retraining program significantly in humans. A rate-dependent relationship between loading
reduced the occurrence of running-related musculoskeletal and bone injury has been demonstrated in rabbits,31,37
injuries by 62% during a 12-month follow-up period. dogs,8 and bovine.2 It has been suggested that increased
Previous gait retraining studies reported large reduc- strain rate is typically associated with greater risk of
tions of VALR (Cohen’s d up to 3.32) and VILR (Cohen’s bony injuries in animals. In human studies, higher VALR
d up to 3.74),23 which are greater than found in the present and VILR have been reported in a group of injured athletes
6 Chan et al The American Journal of Sports Medicine

Figure 4. Vertical average loading rate (VALR) and vertical instantaneous loading rate (VILR) at 8 km/h and 12 km/h before and
after training. BW, body weight.

incidence of calf injury (ie, calf strain and Achilles tendini-

tis) than the control group. This pattern can be explained
by a greater strain on the ankle plantar flexors when the
participants attempted to soften their footfalls by switch-
ing their foot-strike pattern,25 which has been shown to
be effective in lowering vertical loading rates.19
The findings of this study support the use of visual bio-
feedback in reducing impact loading and preventing
injury, which could have a direct effect on reducing health
care costs. A recent study reported that the economic bur-
den of a single running-related injury is approximately
$90.16 Given that more than 60 million people currently
engage in running, whether for recreational or competitive
reasons,36 and up to 79% of runners incur an injury in
a given year,5,42 the total cost of running-related injuries
is estimated at $4 billion annually. Further study could
investigate the cost-effectiveness and economic effect of
the visual biofeedback gait retraining program.
Figure 5. Kaplan-Meier plot of running-related injury survival
Several limitations should be considered in light of the
between participants from the gait retraining group and the
findings presented in this study. First, the current gait
control group.
retraining program can be delivered only in a biomechanics
laboratory, which is not commonly accessible to most run-
with patellofemoral pain10 and plantar fasciitis28 than in ners. Since impact loading is an invisible biomechanical
their healthy counterparts. Such observations are in line marker, future research should explore the potential for
with the injury pattern in our control group participants. wearable sensor technology to allow for VALR and VILR
In contrast, the gait retraining group had a higher measurement in an outdoor environment. Second, we did
AJSM Vol. XX, No. X, XXXX 1-Year Follow-up After Gait Retraining in Novice Runners 7

not measure running mechanics outside the laboratory; 13. Davis IS, Futrell E. Gait retraining: altering the fingerprint of gait. Phys
thus, sustainability of the modified gait biomechanics in Med Rehabil Clin N Am. 2016;27(1):339-355.
14. Davis IS, Powers CM. Patellofemoral pain syndrome: proximal,
the actual environment remains unexamined. Third, simi-
distal, and local factors, an international retreat, April 30-May 2,
lar to a previous study,1 we used an online platform to 2009, Fells Point, Baltimore, MD. J Orthop Sports Phys Ther.
monitor injury patterns of the participants for 12 months. 2010;40(3):A1-16.
Although we contacted every participant who had reported 15. Dreyer D, Dreyer K. ChiRunning: A Revolutionary Approach to Effort-
an injury in order to maximize data validity, we did not less, Injury-Free Running. New York, NY: Simon & Schuster; 2009.
contact uninjured participants to verify that they had not 16. Hespanhol Junior LC, Huisstede BMA, Smits D-W, et al. The
experienced an injury, and therefore injury occurrence NLstart2run study: economic burden of running-related injuries in
novice runners participating in a novice running program. J Sci
may have been underestimated in both groups. Finally, Med Sport. 2016;19(10):800-804.
the exclusion of experienced runners may have affected 17. Hobara H, Sato T, Sakaguchi M, Sato T, Nakazawa K. Step fre-
the generalizability of our findings. quency and lower extremity loading during running. Int J Sports
Med. 2012;33(4):310-313.
18. Hunter I. A new approach to modeling vertical stiffness in heel-toe
distance runners. J Sports Sci Med. 2003;2(4):139-143.
CONCLUSION 19. Lieberman DE, Venkadesan M, Werbel WA, et al. Foot strike patterns
and collision forces in habitually barefoot versus shod runners.
A 2-week gait retraining program using visual biofeedback Nature. 2010;463(7280):531-535.
is effective in lowering impact loading in novice runners. 20. McMahon TA, Cheng GC. The mechanics of running: how does stiff-
More important, the occurrence of running-related muscu- ness couple with speed? J Biomech. 1990;23(suppl 1):65-78.
21. Milner CE, Ferber R, Pollard CD, Hamill J, Davis IS. Biomechanical
loskeletal injury was 62% lower after 2 weeks of gait mod-
factors associated with tibial stress fracture in female runners. Med
ification over a 12-month follow-up period. Sci Sports Exerc. 2006;38(2):323-328.
22. Munro CF, Miller DI, Fuglevand AJ. Ground reaction forces in run-
ning: a reexamination. J Biomech. 1987;20(2):147-155.
23. Napier C, Cochrane CK, Taunton JE, Hunt MA. Gait modifications to
REFERENCES
change lower extremity gait biomechanics in runners: a systematic
review. Br J Sports Med. 2015;49(21):1382-1388.
1. Altman AR, Davis IS. Prospective comparison of running injuries
24. Noehren B, Scholz J, Davis I. The effect of real-time gait retraining on
between shod and barefoot runners. Br J Sports Med. 2016;50(8):
hip kinematics, pain and function in subjects with patellofemoral pain
476-480.
syndrome. Br J Sports Med. 2011;45(9):691-696.
2. Archdeacon MT, Jepsen KJ, Davy DT. The effects of torsional loading
25. Nunns M, House C, Fallowfield J, Allsopp A, Dixon S. Biomechanical
conditions and damage on bovine cortical bone strength. In: Presented
characteristics of barefoot footstrike modalities. J Biomech. 2013;46(15):
at Proceedings of the 1996 Fifteenth Southern Biomedical Engineering
2603-2610.
Conference. IEEE; 1996:445-448. doi:10.1109/SBEC.1996.493271.
26. Nuzzo R. Scientific method: statistical errors. Nature. 2014;506(7487):
3. Bennell K, Crossley K, Jayarajan J, et al. Ground reaction forces and
150-152.
bone parameters in females with tibial stress fracture. Med Sci Sports
27. Oliveira AS, Gizzi L, Ketabi S, Farina D, Kersting UG. Modular control
Exerc. 2004;36(3):397-404.
of treadmill vs overground running. PLoS One. 2016;11(4):e0153307.
4. Blackmore T, Willy RW, Creaby MW. The high frequency component
28. Pohl MB, Hamill J, Davis IS. Biomechanical and anatomic factors
of the vertical ground reaction force is a valid surrogate measure of
associated with a history of plantar fasciitis in female runners. Clin
the impact peak. J Biomech. 2016;49(3):479-483.
J Sport Med. 2009;19(5):372-376.
5. Bovens AM, Janssen GM, Vermeer HG, Hoeberigs JH, Janssen MP,
29. Pohl MB, Mullineaux DR, Milner CE, Hamill J, Davis IS. Biomechan-
Verstappen FT. Occurrence of running injuries in adults following
ical predictors of retrospective tibial stress fractures in runners. J
a supervised training program. Int J Sports Med. 1989;10(suppl
Biomech. 2008;41(6):1160-1165.
3):S186-S190.
30. Pope RP, Herbert RD, Kirwan JD, Graham BJ. A randomized trial of
6. Bredeweg SW, Zijlstra S, Bessem B, Buist I. The effectiveness of
preexercise stretching for prevention of lower-limb injury. Med Sci
a preconditioning programme on preventing running-related injuries
Sports Exerc. 2000;32(2):271-277.
in novice runners: a randomised controlled trial. Br J Sports Med.
31. Radin EL, Ehrlich MG, Chernack R, Abernethy P, Paul IL, Rose RM.
2012;46(12):865-870.
Effect of repetitive impulsive loading on the knee joints of rabbits. Clin
7. Buist I, Bredeweg SW, Bessem B, van Mechelen W, Lemmink KAPM,
Orthop. 1978;131:288-293.
Diercks RL. Incidence and risk factors of running-related injuries dur-
32. Riley PO, Dicharry J, Franz J, Croce UD, Wilder RP, Kerrigan DC. A
ing preparation for a 4-mile recreational running event. Br J Sports
kinematics and kinetic comparison of overground and treadmill run-
Med. 2010;44(8):598-604.
ning. Med Sci Sports Exerc. 2008;40(6):1093-1100.
8. Burr DB, Martin RB, Schaffler MB, Radin EL. Bone remodeling in
33. Romanov NS, Robson J. Dr. Nicholas Romanov’s Pose Method of Run-
response to in vivo fatigue microdamage. J Biomech. 1985;18(3):189-200.
ning: A New Paradigm of Running. Coral Gables, FL: PoseTech; 2004.
9. Chambon N, Delattre N, Guéguen N, Berton E, Rao G. Is midsole
34. Roper JL, Harding EM, Doerfler D, et al. The effects of gait retraining
thickness a key parameter for the running pattern? Gait Posture.
in runners with patellofemoral pain: a randomized trial. Clin Biomech.
2014;40(1):58-63.
2016;35:14-22.
10. Cheung RTH, Davis IS. Landing pattern modification to improve
35. Running USA. 2016 State of the sport—U.S. road race trends. http://
patellofemoral pain in runners: a case series. J Orthop Sports Phys
www.runningusa.org/state-of-sport-us-trends-2015?returnTo=main.
Ther. 2011;41(12):914-919.
Accessed July 18, 2017.
11. Cook SD, Brinker MR, Poche M. Running shoes: their relationship to
36. Scarborough N. Number of people who went jogging or running
running injuries. Sports Med Auckl NZ. 1990;10(1):1-8.
within the last 12 months in the United States from spring 2008 to
12. Crowell HP, Davis IS. Gait retraining to reduce lower extremity load-
spring 2017 (in millions). Statista - The Statistics Portal. https://
ing in runners. Clin Biomech Bristol Avon. 2011;26(1):78-83.
8 Chan et al The American Journal of Sports Medicine

www.statista.com/statistics/227423/number-of-joggers-and-runners- 42. van Gent RN, Siem D, van Middelkoop M, van Os AG, Bierma-
usa/. Accessed October 11, 2017. Zeinstra SMA, Koes BW. Incidence and determinants of lower
37. Serink MT, Nachemson A, Hansson G. The effect of impact loading extremity running injuries in long distance runners: a systematic
on rabbit knee joints. Acta Orthop Scand. 1977;48(3):250-262. review. Br J Sports Med. 2007;41(8):469-480; discussion 480.
38. Shih Y, Lin K-L, Shiang T-Y. Is the foot striking pattern more impor- 43. van Mechelen W, Hlobil H, Kemper HC, Voorn WJ, de Jongh HR.
tant than barefoot or shod conditions in running? Gait Posture. Prevention of running injuries by warm-up, cool-down, and stretching
2013;38(3):490-494. exercises. Am J Sports Med. 1993;21(5):711-719.
39. Sullivan GM, Feinn R. Using effect size—or why the P value is not 44. Williams PT. Reduction in incident stroke risk with vigorous physical
enough. J Grad Med Educ. 2012;4(3):279-282. activity: evidence from 7.7-year follow-up of the national runners’
40. Taunton J, Ryan M, Clement D, McKenzie D, Lloyd-Smith D, Zumbo health study. Stroke J Cereb Circ. 2009;40(5):1921-1923.
B. A retrospective case-control analysis of 2002 running injuries. Br J 45. Wulf G. Attention and Motor Skill Learning. Champaign, IL: Human
Sports Med. 2002;36(2):95-101. Kinetics; 2007.
41. van der Worp H, Vrielink JW, Bredeweg SW. Do runners who suffer 46. Wulf G, Dufek JS. Increased jump height with an external focus due
injuries have higher vertical ground reaction forces than those who to enhanced lower extremity joint kinetics. J Mot Behav. 2009;41(5):
remain injury-free? A systematic review and meta-analysis. Br J 401-409.
Sports Med. 2016;50(8):450-457.

For reprints and permission queries, please visit SAGE’s Web site at http://www.sagepub.com/journalsPermissions.nav.