Does how “hard” your feet land when you run affect how likely you are to get injured?
One way to find out is to assemble a few hundred runners, take detailed measurements of their landing forces and other biomechanical parameters, then monitor them for a couple of years to see who gets injured. That’s what Irene Davis of Harvard did, along with colleagues from South Dakota State and the University of Lincoln, US, in a paper recently published in the British Journal of Sports Medicine.
So what was the verdict? Well, that depends on how you slice it.
The most straightforward analysis is to simply compare the runners who got injured with those who didn’t. Of the 249 female recreational runners in the study, 144 suffered injuries during the two-year follow-up period. On average, there were no differences between the two groups: they were pretty much the same average age, ran similar mileage, and had no differences in their vertical impact parameters.
But that’s not where the paper stops.
Instead, the analysis subdivides the groups further, focusing on the injured runners who consulted a health professional about their injury, and comparing them to the small group (21 women) of uninjured runners who reported that they’d never experienced a running injury in their entire previous running history. When you compare these two subgroups, the diagnosed-injury group did have higher “impact-related” numbers than the never-injured group.
This is an interesting result, and pretty much lines up with what you’d expect—but there are a few points worth mentioning.
In a perfect world, I would have loved to see this study registered, so that I could go back and see what the research listed as the primary outcome variables when they started the study. Maybe they always intended to focus on those two subgroups (diagnosed-injury and never-injured), but you always wonder when you see the focus on a sub-group analysis instead of the seemingly obvious main comparison.
It’s also worth pointing out that the diagnosed-injury group ran 192 kilometres per month, compared to 150 kilometres per month in the never-injured group. While the statistical analysis tries to account for this difference, you can’t get around the fact that the injured runners were running significantly more, on top of any biomechanical differences.
Finally, I think the study could have been framed differently. The very first paragraph jumps in with statements about how humans “evolved to run barefoot” with a mid-foot or forefoot landing, whereas “modern cushioned running shoes” lead people to land on their heels. The conclusion section uses the results to put a big emphasis on adopting a mid-foot or forefoot strike.
The problem with all this is that the study explicitly focused only on runners who land on their heels. They invited all prospective subjects into the lab, analysed their gait, and excluded anyone who didn’t land on their heels with both feet while running. So the strongest conclusion you can draw is that, among runners who land on their heels, those who land most softly are most likely to avoid injury.
Is the same true for mid-foot or forefoot strikers? That’s an entirely different question, because their injury patterns are different, for example with load on the Achilles tendon a comparatively bigger issue. The impact forces on landing are also qualitatively different for heel strikers (there’s an “impact transient” when the heel hits that is entirely absent in mid-foot and forefoot strikers). You can’t compare a bunch of people from Group A and use the results to conclude that everyone should switch to Group B.
To be fair, I think Gretchen Reynolds’ article did a great job at keeping the focus on the benefits of landing lightly rather than switching footstrike. Davis had some nice quotes in it about how the soft-landing heel strikers looked—“like an insect running across water,” in one case—and offered advice about imagining that you’re running on eggshells, or perhaps increasing cadence. I think that’s an appropriate message to draw from the study.
Am I nitpicking here? Yes. The truth is that I find the overall conclusion of the paper (higher vertical impact loading is associated with greater chance of injury) entirely plausible. I think it’s probably true, and it’s great that Davis and her colleagues are collecting much-needed prospective data on questions like this.
But the bigger message I draw from this paper is that the differences are subtle, and are probably not a dominant factor in driving the supposed “epidemic” of modern running injuries. The fact that the main comparison between injured and uninjured didn’t show any impact differences is one clue that this isn’t the “smoking gun” we’re waiting for.
My sense is still that “training load errors” are the driving factor in most running injuries: too much, too fast, too soon. It’s not that things like landing forces don’t matter at all; they do, and it’s particularly worth thinking about them if you repeatedly get injured.
It’s just a question of emphasis. The idea that figuring out the “right” way to run will make you injury-proof is understandably and enduringly popular. But it’s nowhere near as important as learning to be patient.