Do novice runners have weak hips and bad running form?

The role of what I like to refer to as proximal control issues are taking on increasing importance in running injury management with a number of hip related issues identified as being associated with overuse injuries such as patellofemoral pain syndrome and medial tibial stress syndrome. This new study set out to look at some of those proximal issues around the hip and compare novice with experienced runners:

Do novice runners have weak hips and bad running form?
Anne Schmitz, Kelsey Russo, Lauren Nicholson, Brian Noehren
Gait & Posture; Available online 3 March 2014
First, we sought to better understand the predisposition of novice female runners to injury by identifying potential differences in running mechanics and strength between experienced female runners and active novice runners. Secondly, we aimed to assess the relationship between hip and trunk strength with non-sagittal hip kinematics during running. Two female populations were recruited: 19 healthy experienced runners and 19 healthy active novice runners. Strength measurements of the hip abductors and external rotators were measured using a hand held dynamometer while trunk endurance was assessed via a side-plank. Next, an instrumented gait analysis was performed while each participant ran at 3.3 m/s. Group comparisons were made using an independent t-test to identify differences in the impact peak, loading rate, peak non-sagittal hip joint angles, trunk endurance, and hip strength. Pearson’s correlation coefficients were calculated between hip kinematics and strength measurements. There were no statistically significant differences in impact peak, loading rate, peak non-sagittal hip kinematics, or strength. However, the novice runners did show a clinically meaningful trend towards increased peak hip internal rotation by 3.8 degrees (effect size 0.520). A decrease in trunk side-plank endurance was associated with an increased peak hip internal rotation angle (r = -.357, p = 0.03), whereas isometric strength was not related to kinematics. Programs aiming to prevent injuries in novice runners should target trunk performance and possibly hip neuromuscular control, rather than hip strength.

In this study the authors recruited two groups of female runners: experienced and novice (they were comparable on age, weight, height and the Tegner activity scale). The experienced group was defined as running more than 12 miles per week in the previous year. They measured a number of hip strength and endurance parameters as well as kinematic data around the hip while running on a treadmill.

They found no differences between the novice and experienced runners, but then seem to have gone on a fishing expedition to find something positive to say about the study. They claimed:

The effect size of hip internal rotation was clinically meaningful where novice runners tended towards more internal rotation. Trunk side-plank endurance and hip internal rotation angle were significantly correlated with each other where an increase in endurance was associated with a decreased internal rotation angle

The actual data for hip internal rotation was 11.2 (+7.4) degrees in the experienced group and 15.0 (+7.2) degrees in the novice group with the difference being 3.7 (+0.2) degrees. The p value for the difference was 0.12, not even close to the acceptable level of 0.05 for there to be a difference. So, there was no difference in hip internal rotation! They claimed that there was a trend to clinically meaningful difference because the effect size for the difference (using Cohen’s d) was 0.54 (which is normally a pretty good value). The problem is that when there is no statistical difference, effect sizes are meaningless. This just means that the big effect size could be due to chance (as the p=0.12), and there is probably no real effect or difference. So, their data does not support the conclusion that they made. They authors did sort of try to justify this:

The novice runners did show a clinically meaningful trend towards increased peak hip internal rotation by 3.7 degrees. However, this difference did not achieve statistical significance. Quantifying both the effect size, which shows the size of group separation, and the p-value, which indicates the chance that the difference is due to chance alone, allowed us to more deeply explore the data. In this case, we believe that we were able to observe a large effect that was not statistically significant because the t-test was more sensitive to within group variability (7 degrees) than the calculation of the effect size.

….but I don’t buy it.

The other finding they managed to dig up was:

A decrease in trunk side-plank endurance was associated with an increased peak hip internal rotation angle (r = -.357, p = 0.03)

While it is quite plausible that this is the case as this does make sense as if there is less endurance, the hip will probably rotate more internally. I just not convinced that their data supports that as they claim. They did multiple t-tests on the data and multiple Pearson’s correlations, so they need to do an adjustment of the acceptable p value of 0.05 as with multiple comparisons you will find something by chance eventually (ie type 1 error). In the methods section they stated, “P values less that 0.05 were defined as statistically significant“. They should have done a Bonferroni adjustment down of the acceptable p value for a statistical significance. So is the 0.03 low enough to call this correlation significantly different to zero? We do not even know how many correlations they did as this was the only one they reported in the results, but in the methods stated: “Pearson’s correlation coefficients were also calculated in SPSS to assess the relationship between hip kinematics and strength measurements“, so they obviously did more than the one they reported.The authors certainly fudged the x axis in the graph to exaggerate the visual appearance of the correlation.

As always, I go where the evidence takes me until convinced otherwise, and this study does not convince me that there is a difference in hip related parameters between experienced and novice female runners running on a treadmill.

Schmitz, A., Russo, K., Nicholson, L., & Noehren, B. (2014). Do novice runners have weak hips and bad running form? Gait & Posture DOI: 10.1016/j.gaitpost.2014.02.014

Last updated by .


No comments yet.

Leave a Reply