I have a problem with that 1% in the sense that it's correct for injuries that required a hospital visit where the helmet failed to protect enough... But what about all those other crashes that people walked away with no injuries or slight injuries because they were wearing a helmet and it saved them enough to not require a hospital visit, nobody knows those numbers but I suspect the potential is a lot higher than 1% chance of an incident with injuries requiring a hospital visit, if all those numbers were known, the incident level for possible injuries would rise, and if nobody wore a helmet because it's only a 1% chance... that number may skyrocket to 2%... 3%...4%...5%+...? JMO

I understand the objection and it's a valid one concerning many of the studies showing how effective helmets are. But here I was only addressing the risk of having the accident, regardless of helmets.

Yes and I agree the 1% is the risk of having an accident where the helmet failed to protect enough to avoid a hospital visit... But what the actual risk of having an accident where head hits the ground is unknown, totally unknown... That's my point.(trying to be)... So what if the actual chance is 5% but I decide that 1% is safe enough and start to not wear a helmet? A 5% compared to 1% is a big difference, we just don't know the actual % number... We don't have any % numbers where a helmet actually did it's job and shouldn't dismiss that greater possibility...

1% risk of an accident, with or without a helmet. The "actual" risk with a helmet will be lower, not higher. If the helmet had, say, a 50% chance of helping you in an accident, then it's 1% x 50%, or .5%.

That's the reason I insisted on separating the two probabilities with the other guy. They really are two independent probabilities, and one is a lot easier to put an accurate number to than the other. It's easier to know total accidents, and some estimate of cycling trips or cyclists, than it is to put a number on how much a helmet will help given an accident. Because we have better (but not perfect) data on those. But you have to multiply the two probabilities together in the end.

But my thinking is, let say 100 people visit the emergency. OK it works out to 1% of the people we know about, but what if 1,000 people actually hit their heads but the helmet did it's job for 900 people out of the 1,000 that bounced their head on the pavement but they didn't go to the emergency... Kind of skews the numbers if you don't include those 900 people... No?

I agree and it's one of the reasons that accident data is poor. Minor accidents will be under-reported in ER data.

I cited the US Department of Transportation data, which I presumed relied on accident reports rather than hospital data. But that might not be a valid assumption - I'll try to find out.

And that's what I'm trying to point out, that all these statistics show is when something went wrong and the helmet failed to protect enough... There is no statistic on when something went wrong and the helmet helped... Thus, by making the choice to not wear a helmet you could increase the chance of injury way more than some people would think just looking at those statistics/risk numbers where the helmet failed. That's the way I see it.

EDIT; What we need to know to make an intelligent/educated decision is the one thing we don't have any numbers for... How many people did helmets protect enough to keep from going to the ER... If we knew that it was only 1% I too would stop wearing a helmet as I can work with that number no problem, but what if, when in the group where head meets pavement (and it's probably not as small a number as most people think), it's 5%, 10%, 20%+ of that, that the helmet is helping enough to keep from going to ER? So with that lack of data I choose wearing a helmet as that's the prudent thing to do...

You suspect a significant number of non-injury traffic accidents, because of the helmet. Well, you can set a reasonable approximate upper bound to that. Assume ALL of the injury accidents were non-helmeted. Helmet use is approximately 19% among adults (includes very casual riders). So if helmets would have protected in every injury and those accidents were equally distributed among helmeted and non-helmeted (reasonable, isn't it?) then there were 5 times as many accidents, and only 1/5 of them recorded as injury. So the maximum injury accident rate would be 5% per year.

Well, I see we are still swagging away.

But helmeted riders have a serious trauma rate (medical attention sought for an injury) on the order of once every 20 years.
Unhelmeted riders have a serious trauma rate (medical attention sought for an injury) on the order of once every 8 years.

(Trauma, that is injury with self care, about once every 5 years.)

Your choice.

-mr. bill

And 5% or more would put it over the top for me as far as risk, 1% I could live with... Here there is a helmet law and I think 80% of the people are wearing helmets, so using that number = 20%.???

Source for this info? And by whose definition is Severe Trauma, any injury requiring medical attention?

(Sorry, FTFY, serious, their definition.) Hoffman, MR et al.

-mr. bill

A URL or at least journal name and article title would be nice - i.e. something one can actually look up.

I do see some positive progress here. Most everyone but the hard core anti helmet types admit that most all the reports and studies are flawed since they do not include all accidents. Especially the ones that helmets prevented injury. As I pointed out, no one reports that I wasnt injured.

The silliness of that definition of serious trauma is just as wacky. What is "Hoffman, MR et al."

Of course it fits right in this recent discussion of risk where any injury, no matter what the severity
is given equal value.

et al = Lambert WE, Peck EG, Mayberry JC.

-mr. bill

Then there are the hardcore types who have no idea how risk is properly evaluated and consider their own ignorance of the subject as a virtue.

Cute

Is it this one?
Bicycle commuter injury prevention: it is time to focus on the envi... - PubMed - NCBI
The abstract doesn't mention the once every 8 years figure but does have the once every 20 years one. Interestingly it states:
"Traumatic events were not related to rider demographics, safety practices, or experience levels." I'd think helmet usage would be one consideration under safety practices, so it sounds like they didn't see any significant effect on rate of traumatic events based on that. [Unfortunately I don't have access to the full article right now but could get it soon.]

Based on the previously stated every 8 years figure I was afraid that I was overdue for at least 7 serious traumatic events.

Also note that the referenced study is from a study of bicycle commuters. There are many more types of cyclists.

The summary is a bit inelegant - "safety practices" are lights, reflective clothing, and mirrors.

"Except for helmet use, there were no statistical differences between those commuters who experienced a serious traumatic event and those who did not."

....

"Thirteen percent (6) of commuters who did not wear a helmet experienced a serious traumatic event [in the year of the study], compared with 5% (43) of those who did (p 0.023). In an additional analysis comparing commuters who reported a traumatic event with those who reported a serious traumatic event, lack of helmet use was the only statistical difference between the two groups (p 0.013)."

-mr. bill

Of course the so-called Serious Trauma data includes every boo-boo or scratch that any accident "victim" thought worthy of medical attention.Presumably including accidents that produced no injury but the victim sought medical attention to be checked out, just in case (and normally assuming somebody's insurance would pay.)
It should also be pointed out that the so-called Serious Trauma data is not limited to portions of the body putatively "protected" by a helmet. Mr. Bill do you believe helmet wear also reduced the so called serious trauma rate to the entire body of those wearing a helmet?

IMO, only a moron about riskevaluation or a zealot with an agenda would give any credence to any conclusions from a report about accident risk that defines "serious trauma" as any injury that gets medical attention.

And unsurprisingly we are back to this. The only studies that some will accept are studies that can not exist.



-mr. bill

Then can I safely assume that you take seriously the Hoffman MR1, Lambert WE, Peck EG, Mayberry JC. report, its flaky definitions of injury severity, restricted sample selection, self reporting injury data collection method, and most telling give credence to the so-called conclusions about bicycling risk from this report? Yes? No?

Nuff said.

My apologies again for not currently having access to the full paper. But the wording "In an additional analysis" makes it sound like this was an analysis they decided to do only after they had gathered their data and not a hypothesis that they initially set out to test. As explained in wikipedia under 'Misuse of Statistics' such subsequent analysis may indeed be useful in generating new hypotheses to test in future studies with new data but should *not* be used to reach conclusions based on the data in the initial study. It's called "data dredging" - see
Misuse of statistics - Wikipedia, the free encyclopedia