The updated analysis follows an expression of concern the journal Injury Prevention added to the paper in June 2017, after the authors and an outside expert could not replicate the results.
The 2007 paper made a big claim: Children ages one to two years old are five times more likely to sustain serious injuries in a crash when restrained in a forward-facing car seat than a rear-facing seat.
Benjamin Hoffman, a professor of pediatrics at Oregon Health & Science University in Portland who was not involved in the 2007 research, told us:
The 2007 paper was a very significant component of current recommendations and policies.
In 2011, the American Academy of Pediatrics (AAP) changed its recommendation and began advising parents to keep children rear-facing until age two (instead of age one, as the previous policy advised). The 2011 recommendations cited the 2007 paper.
But the new findings, “Rear-facing versus forward-facing child restraints: an updated assessment,” published online Nov. 25, 2017, are much less clear-cut. According to Marilyn Bull, the last author on the 2007 paper and a middle author on the recent paper, the re-analysis “still reaffirms rear-facing is better than forward-facing,” but not to the extent the initial analysis reported. In the new analysis, Bull explained:
We cannot say how much better rear-facing car seats are because of a lack of statistical power.
Brian Johnston, the editor-in-chief of Injury Prevention, told us on November 10 that the journal plans to “retract the original paper,” though has not yet done so. Johnston also explained that the updated paper includes “statistical code that would allow an interested reader to replicate the analysis:”
This level of transparency is vital, especially in observational epidemiology, to maintain confidence in reported results and to allow for robust post publication peer review.
Replication issue comes to light
Questions about the paper arose when Jeya Padmanaban—a biostatistician hired by the car seat manufacturer, Dorel, as an expert witness in a lawsuit involving its forward-facing car seats—could not reproduce the 2007 findings. In February 2016, she sent a letter to the journal and authors, explaining that her “conclusions directly contradict the conclusions” presented in the 2007 paper.
When the study authors tried to replicate the results, they also failed.
In response, the journal issued an expression of concern to alert readers to concerns with the original paper, “Car safety seats for children: rear facing for best protection.” In it, the authors say they believe that the “survey weights were improperly handled in the initial analysis, which caused the apparent sample size to be larger than the actual sample size,” resulting in an inflated statistical significance.
The paper, published in December 2007, has been cited 46 times, according to Clarivate Analytics’ Web of Science.
Basem Henary, the first author on the 2007 paper, told us he stands by his initial analysis. Henary, who left academia several years ago and did not participate in the revised analysis, explained that in the 2007 analysis he used software that “handles the weights of the study designs VERY accurately.”
The authors’ 2017 re-analysis—which examines the data from the same years in the 2007 paper (1988 to 2003) and extends it through 2015—suggests “that a rear-facing seating position is safer” in both cases, says Johnston.
Padmanaban, however, reported the complete opposite: She said her analysis showed children between one and two years old are five times safer in forward-facing seats. Even when she extended the analysis through 2014, “I still found forward-facing to be about three to four times better than rear-facing.” Padmanaban was hired by Dorel to validate and update the Injury Prevention study for a lawsuit in which the company was sued by a family whose child was injured in a front-facing seat; the company tried to defend the safety of the seat in that position.
We asked Padmanaban for her take on the updated paper. She told us she still stands by her analysis and is working on a response to the new paper, which will be published soon.
Time to revisit recommendations?
The 2007 paper is cited by both the AAP’s policy statement behind the change in recommendations to keep kids’ rear-facing until 2, and the technical report behind that policy change, published in Pediatrics. Several states, including Oregon, Rhode Island and South Carolina, recently passed laws requiring parents to keep children rear-facing until two.
Will a retraction of the 2007 paper and the 2017 re-analysis potentially weaken or change current recommendations?
That’s not likely, according to Bull:
I don’t think our new paper will change recommendations or policy. It’s always possible some new research will change things … but not in the foreseeable future.
The authors of the updated analysis reaffirm the current recommendations “to keep children rear-facing as long as possible.”
Hoffman, a pediatrician who led Oregon’s legislation effort, agreed and noted that the 2007 paper was not the only data that informed the current AAP policy. “Compelling” findings from Sweden (1, 2) as well as basic physics played a role and “point to rear-facing being safer for young children,” he said.
But since Padmanaban’s analysis, which she has tried to publish, Dorel has walked back its previous rear-facing recommendation. According to a statement on Dorel’s website:
we continue to advise consumers of the guidelines provided by the AAP, but we do not require a minimum of age 2 in our instructions.
The lawsuit against Dorel, which prompted the company to request a re-analysis of the 2007 paper, ultimately found Dorel “negligent in warning or instructing” the safe use of one of its forward-facing car seats, after a child was left partially paralyzed from a crash. In June 2016, a jury ruled that Dorel pay the family $34 million in damages.
Hoffman noted that the benefit of the Dorel analysis is that “it restarted the conversation and hopefully will spur us to get closer to the truth:”
There’s not a lot of margin for error here. This is about saving kids’ lives from the leading cause of death in young children in the U.S.
Update, 1700 UTC, 12/22/17: The 2007 paper was retracted on December 20.
Like Retraction Watch? Consider making a tax-deductible contribution to support our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up on our homepage for an email every time there’s a new post, or subscribe to our daily digest. Click here to review our Comments Policy. For a sneak peek at what we’re working on, click here. If you have comments or feedback, you can reach us at email@example.com.