Shoes and inserts: how the broken model works

One of the first things we blame whenever we get injured is our shoes.  Most runners are keenly aware of all the minute details of their footwear, and they are easy to pin the blame on.  Whether it's the shoddy upper of the updated model, a too-stiff midsole, shoes that are too run-down, or an abrupt change in brands, we are quick to link our shoes with our maladies.  And for good reason, many podiatrists and doctors would state--shoes can be the cause of a new injury, especially if it occurs within a week or two after a change in footwear.  Problem is, no one can agree on exactly what about running shoes makes them good or bad.

If you try to buy shoes at a chain store, you're probably on your own, unless the salespeople are trying to earn a few extra bucks.  At a specialty running store, you'll probably be exposed to some arcane witchcraft to divine the perfect shoe--the wet foot test, meant to measure arch height, and the standing pronation test, intended to measure rearfoot motion, are the two most common.  Inspecting wear patterns on an old shoe and a pronation examination while running are sometimes also done.  The general idea is as follows: people who underpronate/supinate (the terms are used interchangably) and have high arches can't absorb shock effectively, so they need a "cushioning" shoe.  People who have mild pronation and medium arches need support from a "supportive" shoe.  Finally, people with low arches who pronate severely need very supportive "motion control" shoes.  The driving principle is that pronation is bad and arches need support.

There are several problems with this approach.  The biggest one is that it doesn't work.  As we'll soon see, pronation does not reliably predict injuries, and attempting to correct it does not avoid them either.  Furthermore, the standard tools for correcting pronation, namely, shoe changes and custom orthotics, don't reliably reduce pronation! But perplexingly, switching to a more supportive shoe or getting a custom orthotic from a podiatrist works a good amount of the time! 

Before we tackle that, I'll briefly go over how pronation works, since there's a lot of misinformation about exactly what it is.  Biomechanicists (or is it "biomechanics"?) and podiatrists scoff at the word "pronation"--they prefer "rearfoot eversion," since this more accurately describes what is going on.  Here is an excerpt from a RunnersWorld video that illustrates a runner pronating:

I've skipped ahead to the interesting part. Also, stop watching after about 1:00, since pretty much nothing the narrator says is true.

In principle, pronation during running is fairly simple: after the foot strikes the ground and flattens (whether it is flattening from a forefoot strike or a heelstrike), the foot rotates inward about the subtalar joint.  In my previous post, I mentioned that studies have shown that no matter the stiffness of the surface, peak impact forces are the same.  In a 1998 paper, Wright et al. showed that this phenomena can be explained by a passive mechanism.  That is, the muscles do not interfere in a dynamic way to absorb impact shock.  Rather, the muscles, bones, and joints act like a system of springs, rods, and hinges during the impact phase. 

Wright's model of the leg was significantly more sophisticated than the simple "three hinge leg" model used in some other studies.

The role of pronation in this process is to translate vertical impact force into rotational force, which attenuates the impact by spreading it out over a longer period of time.  Everybody pronates; some do it more than others. Pronating was probably originally suspected of causing injury because it looks odd and out of place.

Running shoes and inserts attempt to correct pronation by supporting the medial arch of the foot.  Remember, pronation is when the foot rolls inwards, so putting a wedge on the inside of the foot should (theoretically) prevent that inward rolling.  This correction is easily visible in a static position; images like this one are common on shoe and insert websites:

 The difference is fairly obvious.  However standing is not running, and as we'll soon see, the difference in degree of pronation while actually running is not nearly as drastic or straightforward as the image above.

So why doesn't the "pronation paradigm" work? First, there is little evidence that the degree of rearfoot eversion/pronation is related to injury.  The biggest study I'm aware of that has found excessive pronation to predict injury is this one by Williams et al.  The study examined 400 subjects over three years.  Some 46 developed "exercise related lower leg pain" (shin splints), 29 of which developed it in both legs.  This gave the researchers 75 injured legs, which they compared with 167 uninjured legs from the students who stayed healthy.  Why they chose 167 is beyond me, but their statistics showed a significant difference in several parameters related to pronation.  The injured group displayed pronation that was faster, greater in magnitude, and occurred later in the stance phase than the uninjured group.  Unfortunately, some experimental oversights make these results untrustworthy.  First, the subjects were recreational athletes who competed in a variety of sports; the study did not use runners exclusively, even though the biomechanical evaluation was a running test.  Second, this biomechanical evaluation was done barefoot.  Running barefoot has a variable effect on rearfoot kinematics (including pronation), so there is no way to know whether someone who pronates barefoot pronates to the same degree in a shoe!

To be sure, there are some other studies that have found a weak relationship between pronation and injury.  But there have been just as many that have found no relationship (like this one, which concluded that "lower-extremity alignment is not a major risk factor for running injuries in [a] relatively low mileage cohort").

Additionally, the two tools used to alter pronation--shoe support and orthotics--don't reliably alter pronation! Benno Nigg, whose work I've cited several times in this post and the previous one, concisely summarizes in a seminal 2000 paper:

"Results from studies with bone pins in the calcaneus, the tibia, and the femur showed only small, nonsystematic effects of shoes or inserts on the kinematics of these bones during running. Even more surprising, the differences in the skeletal movement between barefoot, shoes, and shoes with inserts were small and nonsystematic.The results of this study suggest that the locomotor system does not react to interventions with shoes, inserts, or orthotics by changing the skeletal movement pattern. These experimental results do not provide any evidence for the claim that shoes, inserts, or orthotics align the skeleton."

If shoes don't have the ability to reliably alter pronation, we would expect large-scale studies using familiar tests like the pronator/supinator or wet foot tests to show little or no difference between people assigned the "right" shoe for their foot type and people assigned the "wrong" one.  In fact, this is exactly what we see.  This study by Ryan et al. divided up 81 women training for a half marathon into three groups: neutral, pronators, and severe pronators.  Each group was divided into three subgroups which were each assigned either a cushioning, stability, or motion control shoe to wear for all of their training.  So, for the 30 women who were deemed to be "pronators," 10 received a stability shoe (the "right" one), 10 received a cushioning shoe, and 10 received a motion control shoe. Surprisingly, runners who got the "wrong" shoe got injured slightly less than runners who got the "right" shoe, and everyone who got a motion control shoe was more likely to become injured than those who did not.  A similar study by Knapik et al. with over 1300 Air Force recruits which assigned shoes based on the "arch height" strategy saw no difference between groups.

All I've told you up until now is somewhat old-hat.  The failure of pronation control in preventing injuries has been covered in more detail by others, and the results of the aforementioned studies have been declared as proof of an enormous conspiracy wrought by an unholy alliance of money-hungry shoe companies and doctors.  The problem is that the conclusion is not as simple as it seems, and few people seem to bother with exploring why.  The evidence in the Ryan and Knapik studies, combined with Benno Nigg's review article, seems to lead to a simple conclusion: running shoes and running inserts can't prevent injuries because they don't change pronation, and pronation doesn't cause injuries (which is quickly misconstrued into "running shoes cause injury").  However, this is not the conclusion the data support!  As usual, the details are more nuanced. 

Why?  Because shoe inserts (and probably the right shoe) can prevent injury.  Several studies have found that inserts can relieve or prevent injury; here is one example by Mündermann et al. So, even though the inserts are not significantly altering pronation, they are preventing injury.  This points to a different mechanism for injury, which brings us back to Benno Nigg's article. Nigg spends the first half describing why impact forces and pronation have little or no relation to injury risk.  The second half proposes a new model for skeletal movement and its relation to injury.

According to Nigg, the body has a "preferred movement path."  Regardless of the footwear condition or presence or absence of inserts, the body activates the muscles to stay as close as possible to the preferred path.  This is why adding a thick medial wedge on the inside of the foot to oppose pronation does not significantly change the amount of pronation while running.  To overcome a medial wedge attempting to divert the foot from its preferred path, the body simply activates the eversion muscles of the leg more strongly.  Apparently, the muscles controlling the foot are easily strong enough to overcome a small foam wedge.  In contrast, if a shoe or shoe insert encourages the foot to move along its preferred path, the foot control muscles will not have to be as highly activated.  In his own words:

" If an intervention counteracts the preferred movement path, muscle activity must be increased. An optimal shoe, insert, or orthotic reduces muscle activity. Thus, shoes, inserts, and orthotics affect general muscle activity and, therefore, fatigue, comfort, work, and performance."

There is concern that "interventions" (shoes and inserts) that increase muscle activity may cause injury, but no evidence of this yet.  There is evidence that shoes which decrease muscle activity decrease the energetic cost of running, increasing efficiency.  Muscle activity is connected very closely with muscle vibration on impact--muscles act like springs on impact, and just like a real spring, they vibrate.  Much of Nigg's current work is looking at the effects of these muscular vibrations and whether they are related to injuries. 

Benno Nigg uses electromyography machines to prove the changes in muscle activity.  Fortunately, you don't need an EMG machine to tell whether shoes or inserts are encouraging or discouraging your preferred path of motion.  Your body communicates this in a fairly clear way: comfort. Nigg (and other researchers) have published studies indicating that the shoe and insert choices which prevent injury and reduce muscle activity are also the ones which subjects report to be the most comfortable.  In the Mündermann study above, the military recruits in the experimental group (which recieved shoe inserts and experienced a lower injury rate) rated their inserts as more comfortable than the control group, which had a simple flat insole.

The factors that affect comfort are not the same from person to person; this is why stability shoes are the right choice for some, and the wrong one for others (and this seems to have nothing to do with degree of pronation).  So, in a long and circuitous way, we have shown that in this case, common sense is right: wear the shoe that feels most comfortable!

Things are, from a scientific perspective, a bit more complicated than that, and I've been itching to do some more reading on how changes in muscle activity and vibration affect injury, as well as how various other changes (texturing for example) affects the way the foot interacts with the shoe and the ground.  Further, there are several factors that aren't taken into account in many of these studies that do come into play in real life.  Most studies use small variations of the same shoe, testing different arch heights, medial wedges, midsole hardness, etc.  But in the real world, much more changes from shoe to shoe.  Flexibility, the fit of the upper, the elevation of the heel, and the lacing system are just a few examples.  So even though one of the new "minimalist" shoes with a low-profile, flexible midsole may feel very comfortable, you may not be ready to run in it.  For exaomple, abruptly switching to a shoe with a lower heel height could induce Achilles tendonitis, or wearing a more or less-flexible shoe could cause foot problems.  Of course, it could prevent them as well! These are all topics for another time, though.  In the next week or two, look for another post on the anatomy of a running shoe and some analysis on the usefulness of some of these things.