03 January 2006

Rock and Roll

SUVs kill babies.

That may be a bit of an exaggeration, or at least a decontextualized distortion, of a new study published this week in the journal of the American Academy of Pediatrics. But compared with today's confusing media coverage of the study, it's at least straight to the point.

The truth, of course, is that all sort of vehicles, and all sort of other things, kill babies. What the new Pediatrics paper (not available online yet) actually says that is that you are not reducing the chance that you will injure or kill your children if you drive an SUV instead of a ordinary car. And because we know that SUVs threaten the lives of everyone in anything smaller on the roads, the net societal impact of the 250 percent increase in SUV sales between 1995 and 2002 is negative.

The child-injury stats will surprise many parents, as they probably assumed that bigger is better when it comes to physical protection on the road. It's not really news as far as adult injuries go, though. Statistics showing a slight increase in injury rates for SUV riders comprise a central portion of High and Mighty: SUVs--The World's Most Dangerous Vehicles and How They Got That Way, a book-length exposé by auto journalist Keith Bradsher that came out three years ago.

But because parents tend to react more strongly to emotional arguments rather than statistics, and they're too busy hauling their kids to soccer games to read a 464-page piece of nonfiction, they will be surprised, shocked even, to learn that instead of improving the safety of their offspring's world, they have actually put them at increased risk for some kinds of injury.

This is a perfect example of why we should all have more respect for the tools of science. The way the universe works is often counter-intuitive. Common sense, particularly sense derived from an emotionally charged, advertising-and-propaganda-filled environment, can get you killed.

The coverage the story received this morning certainly doesn't help. The most alarming (almost as alarming as my deliberately misleading opening) comes from a brief UPI item. under the headline "Children's risk triple in SUV rollovers":
PHILADELPHIA, Jan. 3 (UPI) -- Children riding in SUVs involved in rollover crashes are three times as likely to be injured, research released Tuesday concludes.
That's technically true, but the triple risk actually refers to children involved in rollovers compared with non-rollover crashes, not rollovers in cars. Big difference.

Most other news sources did a more responsible, but still confusing job. The Mississippi Sun Herald's headline was "SUVs no safer for children than cars." Most cherry-picked statistics like the Sun Herald's observation that "A child who was properly restrained in a seatbelt or child seat during an SUV rollover had twice the chance of serious injury." Many included a paraphrase of the study's finding that "Rollover crashes increased the risk of injury in both vehicle types."

Furthermore, there's a "doubled risk of rollovers in SUVs" (Chicago Sun-Times) and "though rollovers represent only 3 percent of accidents, they account for more than a third of annual highway deaths" (the Sun Herald again). The Bloomberg business wire offers this take: "Child injuries from rollover accidents were more common in SUVs and outweighed the safety benefits of larger, heavier vehicle frames."

Confused? Depending on what story, and how closely one reads the stories, it may seem like SUVs are three times, two times, a bit more, or just as likely to get your child injured or killed.

The bottom line is that the heft of SUVs does offer some protection, but only enough to compensate for the increased likelihood of rolling over. What the study actually says is that
"... despite the larger size of SUVs and the consequent perception of improved safety, children riding in SUVs have a similar risk of injury, compared with children riding in passenger cars. The protective effect of increased vehicle weight offered by SUVs is tempered by their higher risk of rollover crashes "


"After adjustment for all of the aforementioned factors, the risk of injury was not significantly different for children in SUVs versus passenger cars."
This is not something that lay people can easily figure out by themselves. They have to rely on trained statisticians to tease conclusions out of the enormous accident databases of the federal government.

In addition, responsible and science-literate reporters must help them make sense of the numbers. None of the stories I found mentioned the "odds ratio" the researchers used to describe the comparative risk. The final OR for kids in SUVs was 1.5 times that of kids in other cars, with a 95-percent confidence interval of 0.88-2.57. That means the real risk could be slightly lower or as high as two-and-a-half times. I would expect that after being exposed to those numbers, a parent would be even less comfortable with the idea of driving their kids to soccer practice in an SUV.

Also overlooked in my review of the coverage so far was the bigger picture. Any responsible analysis would come to the conclusion that SUVs are not good for society or the planet. Marketers tell customers they can drive off-road in comfort, although hardly anyone actually does that -- and good thing, too, since it's illegal just about everywhere as well as being bad for wilderness and wildlife. They use too much gasoline, pose a threat to anyone not ensconced in the extra 1,317 pounds that it takes for an average SUV to qualify for an exemption to the national fuel-efficiency standards, and make life miserable for small towns trying to keep their roads free of potholes.

And now it turns out they're not any safer.

The problem is, trading in your SUV for a smaller vehicle only transfers the risk to someone else. In the used-vehicle market, the buyer will likely be younger than most parents, meaning a less-skilled driver will now be driving a machine that's more difficult to drive in the first place. Meaning more accidents and rollovers...

Oh well.


Post a Comment

Links to this post:

Create a Link

<< Home