St. Louis, MO
Posted - 03/02/2011 : 8:17 PM
The helmet story told by safety professionals also claim helmets change the equation so that those who would die without one are only injured. Injuries, then, should be important-and yet we hear little about them. So let's look at injuries in the Louisiana Experiment and see if the helmet story proves true:
We'll use the data from Louisiana's Highway Safety Research Group (LHSRG) through Louisiana State University that breaks injuries down into moderate, severe and fatal in terms of helmeted and unhelmeted riders.
What is a Severe or Moderate Injury?
LHSRG doesn't define what comprises severe or moderate injuries but we'll assume the KABCO injury coding scale was used[i]. If so, severe would translate to incapacitating injuries and moderate to non-incapacitating injuries. Incapacitating injuries would be those that are not fatal but prevent the victim from performing activities s/he was normally able to do before the crash-for example, fractures or concussions. Moderate injuries would then be those that are obvious at the scene but aren't either fatal or obviously incapacitating-for example, sprains, contusions or many (but not all) lacerations.
The following graph tracks each kind of injury for both helmeted and unhelmeted riders. It should be kept in mind that riders suffer and die from a variety of injuries that do not involve the head in any way including chest trauma, internal bleeding, ruptured organs. Wearing a helmet will not prevent those.
During the repeal years, both helmeted and unhelmeted injuries are closely clustered. After reinstatement there's a huge separation. However, during this time, helmet use never dropped below 42 percent while after the reinstatement helmet use rose to 98 percent and that could explain the clustering. Even so, unhelmeted injuries of all kinds outpaced helmeted ones.
In both conditions and as one would expect, theres more moderate injuries than severe ones and more severe injuries than fatalities.
Otoh, there's also an increase in injuries in the reinstatement years that's not explained by more helmeted riders. For example, in 1999 injuries totaled 512. In 2002-two full years into the repeal-the total injuries for all three kinds of crashes was 619-a 21 percent increase. In 2006-two years after the reinstatement-the total was 861. In four years, then, injury crashes had gone up 39 percent or almost double the percentage increase from the midst of the repeal years.
During the repeal years, there's also more fluctuation between the various kinds of injuries for both helmeted and unhelmeted riders. And during the reinstatement years, the kind of injuries are more closely clustered according to helmet use/non-use. This is particularly evident for helmeted injuries. There is no apparent reason for this.
Let's look more closely at each condition in terms of the actual numbers of injured riders.
Under the repeal years, unhelmeted fatalities rose for the first four years. This what the helmet story would tell us to expect as more riders chose to ride without a helmet. The uptick in severe and fatal injuries and vast increase in moderate ones could simply be the result of a huge influx of unhelmeted riders.
Severe injuries rose as well, however, in 60% of the years, there's almost no difference between severe and fatal injury numbers. This relationship between severe and fatal injuries is much tighter after the reinstatement than before and there's no obvious reason why that should be.
The bulk of injuries are moderate, which would be expected but there appears to be no correlation between moderate and severe injuries as there is between severe and fatal injuries.
The helmet story implies that helmets prevent fatalities and turn them into moderate or severe injuries and reduce severe injuries and turn them into moderate ones. The behavior of the three kinds of unhelmeted injuries, though, doesn't support that even though fatalities did rise as predicted.
After the reinstatement, however, a closer relationship between moderate and severe/fatal crashes appears among the unhelmeted and the moderate injuries plummet. Why would this happen?
So let's examine that by looking at the relationship between injury severity and helmet use:
In some ways it's almost the reverse image to unhelmeted fatalities: overall, there's a closer relationship between moderate injuries and severe ones-and a closer relationship between severe and fatal injuries-during the repeal years and a looser one once the universal helmet law was reinstated. But it is basically a mirror image.
There are some differences: while moderate injuries zoom up under reinstatement, there's no wild fluctuation from year to year. And, from 2007-2009, moderate and severe injuries appear to correlate very well however, this is not seen in fatalities. Three years, though, may represent a blip rather than a trend.
Moderate injuries are, by far, the preferred outcome-the increase in moderate injuries in the reinstatement years would be a positive sign if the severe and fatal injury rate was depressed as a consequence as it suggests that helmets are effective in changing outcomes in the same kind of crashes.
But we saw that moderate injuries zoomed up under the unhelmeted condition as well.
Moderate injuries are the normative outcomes of certain kinds of crashes-such as low-sides where riders don't impact a solid, fixed object. Severe and fatal injuries are the common result of crash configurations-such as frontal impacts.
The helmet story hangs on whether helmets really do turn fatalities into serious injuries and serious injuries into moderate ones so, let's compare apples to apples by the percentage of each kind of injury:
Overall, there's an extremely stable relationship between all kinds of injuries: moderate ones are the overwhelming majority for both conditions followed by severe then fatal ones.
Moderate injuries under both conditions over the entire time span averaged between 69 percent (unhelmeted) and 76 percent (helmeted)-but in both conditions, the average percentage dropped slightly after reinstatement.
However, helmeted moderate injuries averaged out, over the eleven years to be 7.76 percent lower than unhelmeted ones. Averaged helmeted severe injuries were 4.76 lower and fatalities were 4.75 lower than unhelmeted averages.
The helmet story, then, held up in that regard: if all things were equal and helmets were the only variable-which they may not be-then helmets appear to have made a small difference when it came to severe and fatal injuries. Otoh, less than a 5 percent difference is not seen to be statistically significant. But, as one reader points out, if you're the one it made a difference for, it matters a lot. Even so, this presumes that the injuries that killed or wounded the extra five percent were head injuries-and that may or may not be true.
However, there's less difference between the percentage of severe and fatal crashes than we may have expected. The following chart presents the difference between fatal and severe injuries for each condition:
The difference between severe injuries and fatalities in Louisiana
In eight out of eleven years (72.7%) there was a greater difference between unhelmeted severe and fatalities and helmeted severe and fatalities. Which is something the helmet story wouldn't have predicted.
The helmet story in Louisiana appears to be like rider training and licensing: it should be true that helmets save lives-they just don't save them in statistically provable ways.
However, if we break down the averages into repeal and reinstatement years, helmeted fatalities went up almost 2 percentage points under reinstatement while severe injuries remained almost the same and moderate injuries went down. In terms of helmeted injuries, the increase was entirely in fatalities.
Otoh, the percentages of unhelmeted injuries remained almost identical during repeal and reinstatement years. Whatever is driving the difference in helmeted deaths, either it's not having the same effect on those who do not wear helmets or it's negating the helmet benefit in some ways.
Although 2 percent is tiny-it's still an alarming development simply because, over several years, that increase was solely in helmeted fatalities and not in severe injuries.
Since we see the same pattern with moderate-fatal injuries (though more exaggerated under the unhelmeted condition) it raises the obvious possibility that the differences are more attributable to the number of different crashes that varied from year to year that drove injury rates rather than helmet use.
In addition, we see that while helmeted statistics performed slightly better overall but worse in reinstatement years while unhelmeted statistics were the same it also points to some other factor that's operating. It could be that a certain number of crashes themselves are becoming more severe and negate the helmet's safety benefit to the same state as riding helmetless.
Otoh, we could be seeing off-setting or risk-compensation or risk homeostasis occurring or adverse recruitment among helmeted but not unhelmeted riders. More on that in the future.
In the next entry, we'll briefly compare injuries to registrations.
[i] The KABCO coding scale: K=Killed; A=Incapacitating Injury; B=Non-Incapacitating Injury; C=Possible Injury; O=No Injury; and U=Injured, severity unknown.