Should you run naked? Au naturale? Do barefoot runners get fewer injuries? The first real science milestone in the story of natural running & running injury prevention (finally) arrived in 2016 … about 12 years since the fad began, when Lieberman and Bramble famously concluded that humans are “born to run” and (news flash!) did not evolve in running shoes. This gave some scientific justification for a popular new running style, and helped to inspire a new minimalist shoe industry with amazing profits — mostly based on a health claim.
So, does running barefoot, or in barely-there shoes, actually prevent injury? If you haven’t heard already, place your bets!
(And don’t misss the brilliant sequel, Shit Runners Say To Barefoot Runners 2:48)
There have been many faddish running styles and shoes that might, perhaps, maybe, help prevent or treat overuse injuries of the knee, legs, and feet — but no one really knows, and never has. The value of these things has always been based only on speculation.1
I write a lot about repetitive strain injuries in runners — many articles, and four books, about iliotibial band syndrome, shin splints, plantar fasciitis, and patellofemoral syndrome — and for years I have mostly dismissed barefoot/ minimalist/natural running as a “fad,” the biggest one since the 70s. The nature of overuse injuries means that running biomechanics and technique can’t be anywhere near as important as the sheer amount of activity. Training volume and frequency probably move the needle more than any other factors by quite a bit. It’s unlikely that any running style or footwear can change the forces on your joints so much that it really matters very much one way or the other. If you’re doing so much running that you’re going to injure your legs or feet sooner or later, no biomechanical advantage is going to do any more than slightly delay your fate.
It might do that. It might matter, a little. Of course it might. And I realize that many badass distance runners are pushing the envelope and interested in the tiniest sliver of an edge, but there’s still a problem: tiny slivers of edges are super hard to prove. And even for regular folks, the benefit could be subtle. So I’ve been reserving judgement, saying wake me up when we’ve got some real data. For years, there was little else to say about barefoot running or any other running fad.
Things started to get interesting in 2012, but it wasn’t until 2016 we finally — more or less — got a clear answer about the effect of barefoot running on injury rates.
Dr. Daniel Leiberman’s 2012 paper on injury rates, published in Medicine & Science in Sports & Exercise in 2012,2 reported “a 2.6-times greater chance of having a severe injury” from running in shoes than running barefoot or minimally shod. “It’s a huge difference,” he says. “As far as I know, it’s the biggest effect ever shown on running injury.”3
Large injury prevention effects in sports medicine are indeed virtually unheard of. If it is true, it’s a big deal, a huge deal. But is it true? Almost certainly not. The study had critical limitations. (Show of hands: who saw that coming? Everyone? Good.) By design, a study like this one cannot in principle actually demonstrate the cause of running injuries, let alone the “biggest effect ever.” Before I explain the problem, let’s review the story of barefoot running so far…
The first running fad was running itself: jogging or “yogging” boomed in the 1970s in the United States. About 25 million Americans started running regularly.
A new industry was born, and was soon pumping out “advanced” running shoes with many absurd features and claims, but above all the promise to protect runners from injuries. In spite of this, injury rates remained more or less stable,4 and there’s still no evidence that any popular running shoe design can prevent injury, and some direct evidence that they do not.5 The runners of the world (well, many of them) turned to the idea of getting rid of their shoes …
prehistory — Everyone was barefoot. For a long time.
1960 — Abebe Bikila ran barefoot to win the first of his several consecutive marathon gold medals.
1955–1967 — Bruce Tulloh broke European records, usually barefoot. Later, Tulloh ran 2,876 miles across America … but shod.
1999 — Invention of the first “barefoot” shoe. Designer Robert Fliri proposed minimalist shoes to Vibram CEO Tony Post, who became a true believer in “natural running,” convinced that it was a “solution to the knee pain and soreness he was experiencing when running.” Lieberman’s 2004 paper was important, but people were thinking about this before he was.
2004 — Bramble and Lieberman publish the paper in Nature that injected a huge dose of scientific credibility into natural running: “Endurance running and the evolution of Homo”.6 They famously concluded, “The fossil evidence of these features suggests that endurance running is a derived capability of the genus Homo, originating about 2 million years ago, and may have been instrumental in the evolution of the human body form.”
2005 — Vibram FiverFingers are unleashed on the world, soon after Bramble & Lieberman’s paper. Coincidental timing? Maybe. The product’s commercial success is notable and distinctly faddish — worth about a half million dollars in 2006, with sales destined to triple annually since.
2007 — Bramble and Lieberman elaborate on the “persistence hunting” idea in a follow-up paper for Sports Medicine, arguing that humans are not just good endurance runners, but fantastic ones, an unusual athletic superpower that evolved to “help meat-eating hominids compete with other carnivores.”7 Dr. Lieberman did a terrific interview about this for ABC Radio’s The Science Show.8
2009 — Christopher McDougall publishes a successful book, Born to Run.9 The book is an instant hit, the barefooting bible, but (naturally) it oversells the value of running naturally:10 there is still zero data on injury prevention.
2010 — Lieberman publishes again, but only to confirm that barefoot running involves less impact — no better than “circumstantial evidence” in a court room, but it certainly fanned some flames of hope.11
2011 — The natural running trend soars to new heights. FiveFingers sales hit about 200 million, and new products flood into a market now worth as much as $2 billion, “accounting for virtually all the growth in the $6 billion running market” (businessweek.com). With legions of true believers wearing silly shoes, mostly because of one book, parody is inevitable (e.g. Shit Barefoot Runners Say 2:17, or the tweet at the end of this section).
Spring 2012 — In a paper for Exercise & Sport Sciences Reviews, summarizing his ideas about barefoot running, Lieberman admirably and cleary concludes that “we simply do not yet know” if natural running prevents injuries. Nobody pays any attention to this.
Summer 2012 — Lieberman publishes the first almost kinda sorta direct evidence that natural running is more than just a beautiful idea: “habitually rearfoot strike have significantly higher rates of repetitive stress injury than those who mostly forefoot strike.” This is quoted everywhere, of course, and Lieberman himself (in Brains Plus Brawn for Edge.org) seems to have made up his own mind that barefoot running does indeed prevent injury, judging from how he ends that talk. But it’s not really the last word. It can’t be. It was the wrong kind of research — which I will explain below.
Spring 2014 — Vibram settles a class-action lawsuit, pays $3.75 million to customers, and removes all claims that their FiveFingers product “strengthens” or “reduces injury” until they can be proved. It’s good news, but companies will keep making premature health claims as long as the potential profits trump the risk of getting spanked for it (so, forever).
Spring 2014 — And then, finally, in early 2016, the British Journal of Sports Medicine published the first genuinely relevant evidence — the right kind of evidence — on injury rates.
So now what?
— Rob Delaney, comedian
With the publication of Lieberman’s 2012 paper on injury rates, it seemed like natural running might be on the verge of an impressive vindication. Or maybe the data truly needed for vindication would fail to appear, in which case the fad deserves to fizzle? No, that won’t happen, because too many people have bought their minimalist shoes and “believe” in them. The topic is more likely to be stuck in the middle for a long time. In 2010, Alex Hutchinson of Sweat Science wrote:
I have a sinking feeling that this is just the beginning of the storm — we’re going to see a whole bunch of studies coming out, accompanied by press releases and news stories, that capitalize on this interest without really telling us what we want to know. Hopefully there are also people doing the long, painstaking, prospective research that would really shed new light on this question.
Was Lieberman’s 2012 paper “painstaking, prospective research”? No: it’s actually retrospective research, which has strictly limited power to answer the question. It is well-established that you can only learn so much in science by looking up bad things that have already happened and trying to match them up with characteristics in the present. There are just too many ways it can be misleading. Here’s a little research methods joke (there are so few!):
How many researchers does it take to change a light-bulb? None if it’s a retrospective study, because the light-bulb has already changed itself.
To nail a conclusion, you really have to do prospective research: you have to set up an experiment with conditions that are as ideal as possible, and ask very carefully designed questions about the results before you collect data. This is quite different than checking to see if the answer you want seems to exist in data previously collected for some other purpose. As a general rule, researchers can find almost anything they want to find in complex data, even when it’s not there at all.As a general rule, researchers can find almost anything they want to find in complex data, even when it’s not there at all, because it’s polluted with all kinds of confounding factors. This is the flaw baked into retrospective research that always limits its power, especially in smaller studies like this one, and especially when there’s a high risk of bias.
Randomized controlled trials, the gold standard of study design, are prospective studies. Only prospective RCTs can logically connect cause and effect. Was forefoot running really the cause of lower injury rates? Unknown! And indeed the abstract of the paper wisely declares that it “does not test the causal basis” for these results. Hurray for honest abstracts.
Common sense screams back: Of course it was the cause! What else? That big a difference? Come on! Gotta be! Emotionally, it’s an almost irresistible connection. And that is the connection Lieberman makes in “Brains Plus Brawn.”
Nevertheless, we should resist. But first, another digression.
Let’s take a break from the dry stuff and consider some pictures. Lots of pictures. What do scads of foot strike photos from 10k olympic trials say about barefoot running? This is smoking gun evidence… of exactly nothing. Briana Rognlin:
The foot strikes, paired next to each athlete’s place and time in the trials, vary widely. But they all belong to elite athletes who, whether they made it to the Olympics or not, are amongst the world’s most talented runners. And yet, some clearly drive their heels into the ground, some land squarely on their forefoot, and others seem to practically land on their pinkie toes.
So that clears that up. Back to resisting the siren call of correlations.
This period was one of those “moments” in science that separates the sober from the drunk-on-hope.
It is shockingly easy to be suckered by an appealing correlation like Lieberman reported in 2012. Many alleged causal relationships in the history of science, just as “obvious” as this one or even more so, have turned out quite wrong. In fact, this is one of the most common ways in which science ends up looking indecisive to the public: because so much restrospective research misleads everyone (even scientists, sometimes especially scientists) to premature conclusions that have to be corrected. Awkward!
Retrospective research particularly suffers from “confounding factors” — unknown factors that may be the real explanation for the results. There are other possible reasons, quite a lot of them, why forefoot-striking Harvard runners might have lower injury rates. It is not necessarily running barefoot (or in barely-there running shoes) that did the trick. For instance:
Forefoot strikers may do other things differently as well. They may have adopted that running style out of an general zeal for injury prevention, and may therefore have other injury-reducing habits. We literally cannot confirm that smashing the ground harder with your legs actually increases the risk of injuring them, and it might actually be exactly the other way around.Perhaps those runners are also twice as likely as their heel-striking peers to reduce loading when they detect early warning signs of injury — but the conditions of the experiment were not controlled, so we can’t know.
Almost nothing about this topic is as clear cut as anyone would hope.
The whole forefoot-versus-rearfoot thing is just astonishingly murky. For instance, rather incredibly, it’s not even safe to assume that higher impact is a risk factor for stress fractures.12 We literally cannot confirm that smashing the ground harder with your legs actually increases the risk of injuring them, and — this is really weird — it might even be exactly the other way around. It’s not clear.
Nor can you count on forefoot and rearoot striking to line up tidily with less and more impact — oddly enough, some runners manage to rearfoot strike rather lightly, and some forefoot strikers are surprisingly lead footed.13 And, from the same source, we learn that runners cannot even accurately report their own style: “lots of people are fooling themselves about how they’re running” (Hutchinson). (I find that last bit fascinating.)
Biomechanics expert Benno Nigg on forefoot running: “It increases some forces and decreases others, mainly shifting the load to different structures,” from his book Biomechanics of Sport Shoes, well-described by Amby Burfoot for RunnersWorld.com. (That’s a great little read, and an excellent companion to this rant.)
If you want to get picky and technical — and I really do — retrospective research is actually useless for doing anything except generating hypotheses.14
Why does calf muscle size in humans vary so much? The bulkiness of the calf is an amplified sign of what kind of heel bone you have.
Anna Ahn, assistant professor of biology at Harvey Mudd College, showed that calf size is probably driven by how far the heel bone projects backwards.15 The calcaneus is the lever that the calf muscles pull on, via the Achilles tendon, to flex the ankle (point the toes). The shorter that bony lever, the more muscle is required to control it and the foot.
Elite sprinters, by the way, have heel lengths way higher than average for homo sapiens. A long foot lever is a speed booster!
Although Ahn’s paper didn’t get into this, I will: it’s reasonable to extrapolate from her results that these anatomical differences probably produce quite different injury risks, and mostly uncontrollable ones at that. That makes for a fine example of why it would be unwise to run around declaring that barefoot running is “better,” because it almost certainly depends on sneaky, idiosyncratic biomechanical variables like this.
Barefoot running is bouncier and probably relieves some impact stresses for some people, but it also relies much more heavily on the strength and elasticity of the calf and Achilles tendon. That might well work well — i.e. less risk of Achilles tendinitis,16 perhaps — for people with longer heels and better leverage on their feet. Or … perhaps it works better for the folks with the stronger calves that they’ve developed because of their short heels? Only more research is going to answer that one, I think.
Meanwhile, you can see the danger of making any assumptions one way or the other.
Through most of the story so far, obviously no one can possibly have any real clue if we’re better off running paleo-style. Which is completely unsurprising, really.17 Certainly the folks selling minimalist shoes don’t actually know.18 Without good evidence of injury prevention, natural running and minimalist shoes are not much more than a fashion statement and a gamble, but that clearly isn’t stopping many people from buying it, literally and figuratively. Although humans may be born to run, we sure as hell aren’t born to reserve judgement. We love wishful thinking and wishful believing. In fact, just like endurance running, we’re freakishly good at it. That’s why Lieberman’s caution in 2012 was so noteworthy and admirable:
Barefoot running raises more questions about injury than we have answers at the moment … We simply do not yet know if experienced barefoot runners have fewer injuries than habitually shod runners.
That was an important statement: the “father” of the natural running fad, making one of the only cautious expert declarations about this ever made. (It was echoed by Rixe et al, practically at the same time.19)
And yet it was met with a huge round of indifference. Everyone should have paid attention — remember, it was the conclusion to a major summary of the topic! — but almost literally no one did. If you Google the exact quote, my own citations dominate a single page of results. Amby Burfoot, who is awesome, also blogged it (Peak Performance, Runner’s World). There are no other prominent references to the quote whatsoever: just Amby and me. Meanwhile, articles about the potential of natural running have been in many mainstream publications, and of course Dr. Lieberman’s new study is being shouted from the roof tops.
Lieberman’s (retrospective!) 2012 results are interesting, and at the time I admitted that they could be promising. But kept on reserving judgement and refused to “recommend” natural running to injured runners until the power of natural running to stop (or treat, a different thing) repetitive strain injuries is confirmed by prospective research.
Without that, it’s obviously un-guessable. And yet that’s exactly what everyone seems to be happy to do: guess guess guess.
There are probably dozens of well-known running techniques and hundreds or even thousands of groovy products that have never been properly tested and would quite likely fail if they were. Somewhere in there, there’s probably stuff that helps, at least a little bit, but it is literally impossible to know without testing.
So I don’t waste a heck of a lot of energy on trying to figure out if these things work when I know that the only way that it possibly can be figured out is not going to happen. To conclude, more from Benno Nigg:
Nigg has noted that running injuries have not changed over the years despite the massive development of the running-shoe industry. Unlike others, he hasn't jumped to the conclusion that shoes are bad, or that barefoot or minimalist-running or forefoot-striking is the answer. Instead, looking at the same data, Nigg concludes: Okay, apparently shoes aren't a big part of the equation.
So get over it.
In 2016, Allison Altman and Irene Davis published the first prospective comparison of injury rates in shod versus barefoot running, in British Journal of Sports Medicine.20 And “prospective” is what we needed the whole time: following a bunch of initially uninjured runners of both types to see what happens to them. Only this kind of study is qualified to answer the question. It wasn’t a perfect test, but it was a lot better than anything else so far.The results are clear & unsurprising: there was no important difference in injury rates, just the types of injuries.
For this test, 200 experienced runners were studied over the course of a year. The results are clear and unsurprising: there was no important difference in injury rates, just the types of injuries. Each was better in some ways, worse in others. Although the paper emphasizes “fewer overall injuries” for barefoot runners, injury rates are what matters — the number of injuries per 1,000 kilometres, say — and they were “not statistically different between groups due to significantly less mileage run in the barefoot group.”
But there’s an important caveat: The barefoot runners put in just 24km/week, while runners in shoes ran 41km/week without an increase in injury rate. As Alex Hutchinson put it for Runner’s World, “The only way the comparison has any relevance is if they’re arguing that barefoot running reduces injuries by preventing you from running as much as you’d like.”
Naturally, a larger, longer study may have different results. But this is an excellent start, and we can now say with high confidence that barefoot running is not a panacea for running injuries — and more and better data will probably, if anything, be even more embarrassing for the natural running fad’s claims of injury prevention.
Fuller et al21 — A small (61 runners) randomized controlled trial comparing running pain and injury rates in regular running shoes versus minimalist shoes. Despite a nice gradual adaptation period of 26 weeks, these experienced runners developed more pain and injuries in the minimalist shoes: 16 of them got injured, compared to 11 runners in the regular shoe group. “Runners should limit weekly training distance in minimalist shoes to avoid running-related pain.” The study isn’t big enough for firm conclusions, but it adds to the mounting evidence that running in minimalist shoes may actually cause more injuries, rather than preventing them.
I am a science writer, former massage therapist, and I was the assistant editor at ScienceBasedMedicine.org for several years. I have had my share of injuries and pain challenges as a runner and ultimate player. My wife and I live in downtown Vancouver, Canada. See my full bio and qualifications, or my blog, Writerly. You might run into me on Facebook or Twitter.
Special thanks to Dr. David Gorski of Science-Based Medicine, and Dr. Rob Tarzwell, for helping me to wrap my head around the limitations of restrospective research.
— Minor update: editing and clarification of the introduction; added a citation strengthening the case against barefoot running.
Can any running style or shoe type prevent running injuries? This is one of those reviews of a virtually non-existent literature. The authors speculate, as many others have, that styles and shoes probably involve risk trade-offs: less risk of one kind of injury, but more of another. But mostly the paper inevitably confirms that, as of 2012, there really is no hard data about this.BACK TO TEXT
Can a custom shoe prevent injuries by compensating for individual differences in running mechanics? Several hundred Marine Corps recruits were given motion control, stability, or cushioned shoes for their “low, medium, or high arches.” Those recruits got injured just as much as hundreds of other recruits who were given stability shoes, regardless of their arch shape. The authors conclused that prescribing shoes “based on the shape of the plantar foot surface had little influence on injuries.”
But what if the recruits didn’t actually wear the shoes much? One of the recruits contacted me and explained that “we spent most of our physical training in ill-fitting boots. I ran in those boots more than in my fitted running shoes. Throw in a few dozen officials and unofficial hikes and the study was probably useless. We mostly wore the shoes for recovery from the hikes, blisters, swollen feet and all.”
A report like this isn’t necessarily a data disaster — there are many possible mitigating factors — but it does reduce confidence in this otherwise persuasive evidence.BACK TO TEXT
Dan Lieberman gives a great interview. Audio no longer available, but the full transcript is.BACK TO TEXT
This paper shows that “habitually barefoot endurance runners often land on the fore-foot before bringing down the heel” and “generate smaller collision forces than shod rear-foot strikers,” which “may protect the feet and lower limbs from some of the impact-related injuries now experienced by a high percentage of runners.”
However, trying to infer injury prevention outcomes from these findings is pure guesswork. It’s so, so, so easy to see the words “smaller collision forces” and assume that this is a good, injury-preventing thing, but that is really just not a safe assumption. For all we know, that smaller collision force (assuming it’s real) is coming at the expense of, say, much greater stresses on your calves and achilles to control the sproingier ankle action … which might protect against shin splints, but send achilles tendinitis rates through the roof. We don’t know that stuff, and we can’t know until thousands of runners have been carefully studied.BACK TO TEXT
The typical human foot is an anatomical marvel of evolution with 26 bones, 33 joints, 20 muscles and hundreds of sensory receptors, tendons and ligaments. Like the rest of the body, to keep our feet healthy, they need to be stimulated and exercised. Vibram FiveFingers® footwear is different than any other footwear on the planet. Not only does it bring you closer to your environment, it also delivers a number of positive health benefits—by leveraging all of the body’s natural biomechanics, so you can move as nature intended.
Uh huh. So basically the same stuff — but now too vague to get busted for.BACK TO TEXT
A summary of the state of barefoot-minimalist running science, much shorter than Lieberman’s early 2012 summary, with the same conclusion: on the one hand, “A growing body of biomechanics research has emerged to support the advantages of ‘barefoot’ running,” but unfortunately, “No clinical studies have been published to substantiate the claims of injury reduction using a ‘minimalist’ style.”BACK TO TEXT
BACK TO TEXT
BACKGROUND: Advocates of barefoot running suggest that it is more natural and may be a way to minimise injury risk. In contrast, opponents believe shoes are needed to adequately cushion and support the foot. However, to date, there have been no prospective studies of injury patterns in barefoot and shod runners. The purpose of this study was to compare the incidence and rate of injuries between shod and barefoot runners.
METHODS: A prospective survey was conducted over the course of a year among 201 (107 barefoot and 94 shod) adult runners. Information regarding injuries and mileage was logged monthly using a custom, web-based database program. The number of injured runners, number of injuries per runner and injury rates were compared between habitual barefoot and habitual shod runners. Both musculoskeletal and plantar surface injuries were assessed.
RESULTS: Statistically fewer overall, diagnosed, musculoskeletal injuries/runner were noted in the barefoot group. However, injury rates were not statistically different between groups due to significantly less mileage run in the barefoot group. As expected, barefoot runners sustained a statistically greater number of injuries to the plantar surface of the foot. The descriptive analysis suggests a greater number of calf injuries, but lower number of knee and hip injuries in the barefoot group. Additionally barefoot runners reported less plantar fasciitis than the shod group.
CONCLUSIONS: Barefoot running is associated with fewer overall musculoskeletal injuries/runner, but similar injury rates. A larger scale cohort is needed to more accurately assess differences in individual injuries between these two groups.