NPR and many other news outlets are reporting that brain injury rates have increased in cities with bike-share programs. The implication is that these bikeshare riders are not riding with helmets, and getting brained. That is the conclusion of a paper published by Dr. Janessa Graves of Washington State University.
And yet the data in the paper showed the exact opposite. Total number of head injuries in cities with bikeshare declined significantly — despite an increasing number of bicyclists. All that changed was that there was a marginal increase in the proportion of injuries that were head related. However, the paper can’t ascertain whether those head injuries had anything to do with bikeshare, or even whether those injured were wearing helmets.
This is the worst kind of junk science. While it is understandable that journalists could be fooled by this bullshit, how did this paper ever pass peer review?
Systemic Failure 
You could ask on Twitter:
https://twitter.com/janessagraves
Wouldn’t the results be consistent with this model:
More bikes on the road makes it safer to bike for everyone. So much safer, in fact, that total injuries went down. But, with a lower proportion of the bike riding population wearing helmets, a larger fraction of the reported injuries were head injuries.
I think you could argue that requiring helmets (as we’re about to do with our bike share in Seattle) would push the usage down far enough to get rid of the first positive effect, while also getting rid of the second negative effect. But, it’s definitely a more subtle argument than simply calling the study bullshit.
While your model is plausible, I don’t see anything about the results in this study that supports it. Their data lacks information on whether cyclists who suffered head injuries were wearing helmets, or using a bike rental at the time.
You also have to consider the source. One of the authors, Dr. Rivara, has a history of making bold claims about helmet effectiveness, that upon closer examination prove to be completely wrong.
Also consistent with the addition of rental riders having fewer accidents (slower and more cautious) but because they’re not given helmets, but when they do have an accident they’re more vulnerable to head injury.
So head injuries drop while other injuries drop more therefore … bikeshare is good at preventing all injuries, but particularly good at preventing some unmeasured class or classes of the “everything but head injury” group?
[…] « Another Junk Science Helmet Study […]
[…] another strange example, a helmet fanatic claimed that bike share increased brain injuries when the data in the study showed a significant decrease […]
[…] Another Junk Science Helmet Study […]
[…] riders generally do not wear helmets. He predicted mass carnage as a result, and published a paper that purported to show a 14% increased risk of head injuries as a result of […]