The Chicago Marathon was one of the headline marathon majors for 2018 thanks to some big names that went head to head in the international elite field. We were very fortunate at Runfisix to be able to collaborate with a runner competing in the second tier elite section sometimes referred to as the 'American Development' group. Our runner, Eric was aiming to run under 2 hours 30 minutes and gain valuable experience in marathon preparation and race-day execution.
Conditions were typical of fall season with rain and some wind but temperatures that weren't overly cold. At the sound of the gun many of the elite athletes pushed hard to establish very aggressive pacing plans and form strategic groups. In windy conditions elite runners know that staying within protective groups is an advantageous strategy that can conserve large amounts of mental and physical energy. This was demonstrated excellently by the eventual men's winner Mo Farah who sat with the lead group until well into the race before changing gears and driving on to victory in the last part of the course.
In keeping with our style in these big race case studies we will focus on one aspect of the runner's gait data in the post. For this data example we will look into a key metric (or metrics) that can highlight something valuable and insightful about the runner's quality of movement during the event. The collection of a number of different metrics increases the chance of observing something diagnostic and helps refine a more precise understanding. In this race Eric wore a pair of version 3 Runscribe footpods that simultaneously collected data on 8 different bilateral metrics.
The plot below shows Eric's basic background data, his pace curve and step rate curve with the course elevation profile in the background. What can see immediately is that Eric had a tough race that didn't go to his plan. His pace dropped progressively throughout the race until it stabilized at 2 hours and 5 minutes. His step rate response was not the same. Step rate appears to drop progressively until 52 minutes at which point it holds steady until the end of the race. Eric's overall finish time was 2 hours 34 minutes and hence the end result was not as quick as he wanted.
Pace in min/km in grey and step rate in spm in pink against time. Elevation profile in black.
Clearly then there was a change in his gait at 52 minutes where he found a turn over that was sustainable for him until the finish line. Eric's report back after the race confirms this adjustment at 52 minutes. Up until this point he was looking to stay with other runners in groups, but the pace was very variable and likely a little too fast. At 52 minutes he found himself on his own and made the conscious decision to find a sustainable effort level and style that he felt he could stick with.
So how was Eric moving during the race? Are there metrics within the data file that can tell us anything about his mechanics and injury exposure?
When we look at his contact times for his left and right sides we can see that they both indicate increased time on the ground as the race progresses and he gets more fatigued. However aside from performance analysis, from an injury prevention perspective his left and right sides remain extremely balanced all the way to the line. Plot 2 (below) shows his left and right contact times with near perfect overlap and symmetry. The overall average asymmetry for the race is just 1%.
Pace in min/km in grey and contact times in ms in blue and purple against time. Elevation profile in black.
With inspection of his foot-ankle area motion data we see yet more evidence of quite symmetrical movement. The 2 plots below show (firstly) maximum pronation excursion and (secondly) maximum pronation velocity. In both plots the curves are tracking closely to one another and indicative of good symmetry. The measured overall average asymmetry for the race was 11% for pronation excursion and 11% for pronation velocity. So clearly from the ankles down Eric was working well with no clear signs of significant issues. Interestingly for both parameters his right ankle area had the higher values and inferred greater instability.
Pace in min/km in grey and pronation excursions in degrees in light and bright green against time. Elevation profile in black.
Pace in min/km in grey, pronation velocities in degrees/s in blue and purple against time. Elevation profile in black.
The data file also includes a series of metrics that are based on recorded acceleration values when the feet strike the ground. Accelerations that will then be felt by body and require absorption by key leg muscles. The key metrics in this group are the impact G's and the braking G's. These come from the same raw data but are essentially the vertical (impact) and horizontal (braking) components. When we look at the braking G's in the plot below we see again very good symmetry between the left and right side with the absolute numbers being within expected range. We can also see that the values on both sides are increasing throughout the race. This is most likely due to the nature of the footstrike and which part of Eric's foot is striking the ground first. The overall average asymmetry in braking G's for the race is 6%.
Pace in min/km in grey, braking in G in light and bright turquoise against time. Elevation profile in black.
The next plot shows the impact G's for the left and right sides. This is the first metric that indicates markedly different values and notable asymmetry. The left side average value was 13.1G compared with the right side of 9.7G. This gives an overall average asymmetry for the race of 26%. At Runfisix we say that anything over 20% asymmetry for impact numbers is highly significant as a rule of thumb and could be a key indicator for injury exposure. Notable here is that the actual impact values themselves stay relatively constant during the race with just a minor change on the right side between 22 minutes and 1 hour 5 minutes.
Pace in min/km in grey, impact in G in light and bright purple against time. Elevation profile in black.
So Eric's impact numbers clearly hint at something and would be a great trigger to go into a more detailed assessment with him to understand why this is happening. Fortunately as the data collection was a blind test with no prior information we were able to talk to Eric about his impact asymmetry after the race and ask if there was anything that would be able to explain it. Eric said that he had indeed had an issue in the week prior to the race and that he over-stressed his right calf and had been sore in that muscle group in the week proceeding the race day.
It is a tenuous interpretation with just this information alone but we have seen relatively acute calf issues generating asymmetric impact values in cases within our database. In these cases the side with the lower impact numbers, was the injured side and the lower impact reflected a reluctance of the body to accept impact and shock in order to protect this side. We have seen numerous cases where the side with the injured calf also displayed higher pronation and pronation velocity numbers. With the inference being that increased instability in the foot - ankle area on that side meant it was more vulnerable to suffering lower leg injury. Of course it must be stressed that these are very much tenuous connections unless more complimentary data is collected.
Thanks for reading our Chicago Marathon case study, we hope that you gained a deeper understanding for how wearable gait data can be extremely valuable when particular metrics provide insightful asymmetries. In the future we will pursuing more race event case study examples. If you have ideas for races where we should collect data just let us know in the comments below.