Exactly. Need to look at the entire population of injured riders, not just those injured badly enough to go to the hospital. Particularly you then need to look at ratio of helmet use between those with similar injuries to parts other than the head as well as similar types of crashes. Only when you can normalize for the factors other than the helmet can you then begin to look for any real information about the effects of helmet use.


The trouble is that's a difficult (expensive) proposition and unlikely to get funded well enough to do properly -- FAR easier to get funding for a couple grad students to go comb through the records at six hospitals and a morgue and write a paper based on the results.


This is an example of why it's important to look at any study's methodology before drawing conclusions from its results. Many times a study is looking at a sample that doesn't allow conclusions about the population as a whole.