Thread by @KarlZelik

You go for a run down the street.
You feel the ground force on your feet.
You may think these reveal
The bone loads that you’ll feel,
But this thinking is just incomplete.

Read more in our #biomechanics pub. #running #stressfracture #wearables
tinyurl.com/y7qgjtkc
1/n

This paper reflects 2 years of fun, thought-provoking collaboration w/ @leonscottmd @EmilyMatijevich & Lauren B. The paper also benefited greatly from feedback & conversations on Twitter -- we even formally thanked the Twitterverse in Acknowledgements section -- so thank you! 2/n

@leonscottmd approached us a couple yrs ago asking if #wearables could be used to monitor & prevent bone stress fractures -- painful microcracks due to repeated bone loading -- in hopes of reducing the number of injured #runners showing up to his clinic each week. 3/n

After a yr of perusing literature, we performed a study, led by @EmilyMatijevich, in which 10 recreational runners ran over various speeds & slopes. We measured ground reaction force (GRF) & kinematics, & used #biomechanical algorithms to estimate force on the tibia leg bone. 4/n

Since stress fractures result from repeated bone loading, we simply wanted to know whether any of the commonly-used #GRF metrics (impact peak, #loadingrate, active peak, impulse) were strongly correlated with tibial load metrics (peak force, impulse). 5/n

What we found was pretty stunning... not a single GRF metric was strongly correlated to tibial force! 76 of the 80 subject-specific correlation coefficients we computed indicated that higher GRF metrics were not strongly correlated with higher tibial forces. 6/n

We observed high inter-subject variability in correlation coefficients, though most were negligible, weak or moderate. Correlating GRF impact peak & loading rate with peak tibial load resulted in r=-0.29 +/- 0.37 & r=-0.20 +/- 0.35 (i.e., weak/negligible). Example plot (N=1). 7/n

This means that just b/c GRFs (impacts, loading rates, etc.) increase, we cannot assume bone loading (or stress fracture risks due to this loading) also increases! Increases in GRFs should NOT be assumed to be a surrogate for, or indicator of, increases in tibial loading. 8/n

What is particularly interesting (& troubling) is that we did a quick search of the sport science literature between 2015-2017. We discovered that >50 peer-reviewed publications PER YEAR assume, report or interpret GRF metrics to signify increased bone loading or #injuryrisk. 9/n

In our paper we provide an extensive discussion of other evidence -- related to biomechanics, epidemiology, bone mechanics, etc. -- which further calls into question the way GRF metrics are commonly interpreted as indicators of musculoskeletal loading and injury risk. 10/n

Next in the paper we take a hard look at what these findings mean for #wearables -- there are a growing number of accelerometer & pressure insole devices that claim to give feedback on musculoskeletal loading or overuse injury (e.g. stress fracture) risk to runners/athletes 11/n

The problem is that these wearables use GRF-correlated signals to assess musculoskeletal loading or injury risk. But as noted earlier, GRFs (& correlates from accelerometers, etc) do not signify, and cannot be used as a surrogate for, forces felt by bones like the tibia. 12/n

I'm not sure how to say this other than... the scientific evidence strongly suggests that these wearables are measuring the wrong thing; if the goal is to monitor overuse injury risks due to repeated loading on bones and other tissues inside the body. 13/n

Our paper contains extended discussion on all these topics. We published the work Open Access & tried to write the paper -- especially Intro & Discussion -- in a manner that would be accessible/understandable to scientists and non-scientists alike. 14/n
tinyurl.com/y7qgjtkc

Collectively, the evidence summarized in the paper suggests critical flaws in how GRFs are commonly used to assess musculoskeletal loading & injury risk, which then misguides how GRF metrics are being applied to sport training, product development and wearable devices. 15/n

I hope this paper can help raise awareness & transparency of these correctable issues. I sincerely encourage the sport science & wearables communities to revisit/reconsider their use & interpretation of GRF metrics (& signals from accelerometers & pressure insoles). 16/n

@EmilyMatijevich created free interactive software that allows you to load our study data & explore it yourself. For instance you can select a subset of running speeds & slopes, & look at how GRF metrics do or do not correlate with tibia bone forces.
tinyurl.com/y76zxzyf
17/n

Here's a link to youtube videos that show examples of how you can use this interactive graphical user interface (GUI) to explore the #biomechanics data and automatically compute correlations between GRFs & bone loading.
tinyurl.com/y7av67s8
18/n

What's next for us? Well, we are working on a new way to fuse data from multiple wearable sensors to monitor the total force on the tibia bone (from both muscle and the GRF). Early results look promising & we may be looking for commercial partners to explore translation.
19/n

If you made it this far into thread, thanks! Reward: a final Limerick.

The force due to ground reaction
May be a stress fracture distraction.
Don’t assume force on shoe
To mean tibia load too
Since bone load’s mostly from muscle contraction.

20/20

tinyurl.com/y7qgjtkc

Also here's the two-minute video summary created by @VanderbiltU. Thanks @eagles_zack & @HeidiHallTN