In February 2012, a safety engineer at Evenflo, one of the biggest sellers of children’s booster seats, wanted the company to make a major change to its instructions for parents. He recommended Evenflo stop selling booster seats for children who weigh less than 40 pounds.

Citing government research, the engineer, Eric Dahle, emailed high-ranking executives to tell them that children lighter than 40 pounds would be safer in car seats that use harnesses to hold their small bodies in place. Making the change would match Canadian regulations and better align with recommendations from the American Academy of Pediatrics.

A marketing executive “vetoed” Dahle’s safety recommendation, an internal Evenflo record shows. Later that year, the subject came up again. The same executive, who had been promoted to vice president of marketing and product development, expressed his exasperation. “Why are we even talking about this?” he wrote in an email, adding, “I have looked at 40 lbs for the US numerous times and will not approve this.”

Evenflo’s decision to keep the weight recommendation for its Big Kid booster low in the U.S. was emblematic of how the company — locked in a marketing battle with its biggest competitor — has repeatedly made decisions that resulted in putting children at risk. Not only did it sell its seats for children under 40 pounds, but Evenflo touted its Big Kid boosters as “SIDE IMPACT TESTED” without revealing that its own tests showed a child seated in its booster could be in grave danger in such a crash.

On its website, Evenflo told parents those tests were rigorous, simulating realistic side-impact crashes, which were responsible for more than a quarter of deaths of children under 15 killed in vehicle collisions in 2018. While less common than head-on crashes, side impacts are more likely to result in serious injuries in part because there’s only a door separating the passenger from the intruding vehicle.

In reality, Evenflo tests were anything but stringent, internal company documents show. The company’s tests show that when child-sized crash dummies seated in Big Kid boosters were subjected to the forces of a T-bone collision, they were thrown far out of their shoulder belts. Evenflo’s top booster seat engineer would later admit in a deposition if real children moved that way, they could suffer catastrophic head, neck and spinal injuries — or die.

Yet Evenflo gave its seats passing grades. Indeed, the company’s test bar was so low, the only way to fail was if the child-sized dummy ended up on the floor or the booster itself broke into pieces.

ProPublica obtained multiple years of Evenflo’s side-impact test videos, thousands of pages of sworn depositions of company employees and marketing materials that laid out the business objectives for the Big Kid that, until now, had mostly been shielded by secrecy orders in court cases around the country. These records provide a rare window into one company's marketing and safety decisions.

Evenflo could make up its own test, and pass itself, because of a regulatory failure. Nearly 20 years ago, Congress enacted a law requiring the National Highway Traffic Safety Administration to implement rules to improve the safety of car seats and boosters and to minimize children’s head injuries in side-impact collisions. But the regulators never enacted standards for side-impact tests for boosters and other car seats in such crashes. Today, those seats need only pass a crash test that simulates a head-on collision.

It’s not clear whether children are at risk in other boosters in the same type of side-impact crashes. There is no good federal data on how often children in booster seats sustain injuries or if one company’s booster seats are safer or more dangerous than another’s. Since all the companies operate in the same lax regulatory environment, their side-impact tests, too, are allowed to remain secret.

Evenflo’s chief competitors, Graco and Dorel, have also faced lawsuits brought by the families of children injured or killed in accidents, though those suits involved head-on collisions. The families alleged the companies marketed boosters for kids too young or too small to be protected. Graco and Dorel also advertise seats that they say provide side-impact protection. When asked by ProPublica, Dorel and Graco declined to provide test videos or even the basic details of how their side-impact tests are performed. A Graco spokeswoman said the company is “committed to child safety.”

Unlike Evenflo, Dorel told ProPublica it stopped selling boosters for children under 40 pounds in 2016. But at least six other companies in addition to Evenflo continue to market boosters for American children who weigh as little as 30 pounds. Graco is among them, though some Graco TurboBoosters sold on the websites of Walmart, Amazon and Target now include this statement: “To continue to meet industry standards, we have increased the weight minimum from 30 to 40 lb.”

In a response to detailed written questions about this story, Evenflo General Counsel Amy Blankenship wrote that her company has been a pioneer in side-impact testing of boosters and other child car seats, and that it provides safe, effective and affordable products, including the Big Kid. Evenflo’s Big Kid has always complied with federal regulations, she wrote, and rules written by the National Highway Traffic Safety Administration allow sale of boosters for children who weigh as little as 30 pounds.

Blankenship said Evenflo has records of hundreds of accidents in which children seated in Big Kid boosters were unscathed or received only minor injuries “unrelated to the car seat.” She said some parents credited the Big Kid with helping to save their children’s lives.

“Unquestionably, our company records on Big Kid alone are anecdotal proof of the improved level of safety,” Blankenship said. She added, “No child restraint or booster seat can ensure that a child will not be injured in a car crash, especially a severe one; just like no vehicle manufacturer can ensure that all occupants will escape injury in every crash.” In severe accidents in which children are properly positioned in their booster seat, she contended, the injuries of a child weighing less than 40 pounds would be no different from a child who weighs more.

Referring to the federal regulators, Blankenship wrote in response to ProPublica that “none of the questions or criticisms you lodge have been raised by NHTSA at any time. In addition, many of the issues you raise would be applicable to the millions of BPB [belt-positioning booster] seats sold by other manufacturers. Your article therefore seeks to condemn an entire category of car seats successfully used by millions of consumers and which helped save thousands of lives in a myriad of crash scenarios.”

Evenflo’s marketing has helped make the Big Kid line a top-selling booster seat in the United States. The company, which is a subsidiary of China-based Goodbaby International Holdings Ltd., has sold more than 18 million Big Kid boosters. One customer was Jillian Brown’s father, who said in his deposition that he picked out his daughter’s Big Kid in a BJ’s Wholesale on Long Island in 2014 because of the number of times it said “safety” on the box.

Two years later, Jillian’s mother was driving her and her sister to day care on their first day back after a vacation to Disney World. Jillian, then 5, was properly belted into her Big Kid booster in the back seat on the passenger side of their Suzuki Forenza. She weighed just under 37 pounds.

As Lindsey Brown entered an intersection to make a left turn, she saw “the image of a black car” just before it slammed into the driver’s side of her car, she later recalled. The impact propelled Jillian violently to the left. Her shoulder slipped out of the seat belt. As her chest and stomach jackknifed over the lap portion of her seat belt, her head rotated downward and stretched her neck.

After the car shuddered to a stop, Lindsey Brown looked frantically back at her daughters. The door near Jillian’s sister, Samantha, bent inward like a crumpled piece of aluminum foil. Next to her, Jillian was slumped all the way forward over the lap belt. When first responders arrived, they found Jillian unresponsive. Her lips were blue.

Jillian had suffered what medical journals call “internal decapitation.”

Samantha healed from the injuries she sustained in the crash. But Jillian is paralyzed from the neck down. Now 8, with flowing brown hair, Jillian steers her electric wheelchair with her tongue and communicates by mouthing words barely above a whisper. She plays games and does homework by focusing her eyes on letters, words and icons on a computer device attached to her chair. A ventilator keeps her alive and she has round-the-clock nursing care. Jillian’s father is suing Evenflo and others in Suffolk County Supreme Court in New York.

Evenflo continued to sell booster seats for very small children even after it was put on notice about the possible jeopardy they faced. Years before Jillian’s accident, the families of two other children of similar weights had sued Evenflo when their Big Kid boosters failed to protect them in side-impact accidents. One suffered a traumatic brain injury; the other, internal decapitation.

Asked about the three cases, two of which Evenflo has confidentially settled, Blankenship blamed bad driving. “In the lawsuits you mention, it is undisputed that significant driver error was the primary causative factor of the children’s injuries,” she said. She added that Jillian’s seat “performed as it was designed to do and did not cause” her injury.

Despite what the marketing materials say, parents shouldn’t draw the wrong idea about the safety of Evenflo seats, David Sander, then an Evenflo engineer, said in a 2016 deposition in one of the previous Big Kid negligence cases. “We side-impact test our seats,” Sander said, “but I don't think we say that we offer any type of side-impact protection.”

The Launch

In the early 2000s, many states began requiring that school-aged children use car seats or boosters, recognizing that they didn’t fit in regular seat belts. Graco, Evenflo’s rival, launched the TurboBooster, which lifted children up so that their seat belts fit properly. It quickly became popular.

Evenflo began to lose out in this booming new category. It designed the Big Kid with the goal of “regaining control in the market” from Graco, according to internal Evenflo records. Evenflo’s team proposed creating a booster seat with similar features to Graco’s but priced to sell for about $10 less. Graco, a unit of Newell Brands Inc., and Evenflo aren’t making the Teslas and Porsches of car seats like those sold at high-end retailers for as much as $200. Instead, parents can find Graco and Evenflo products at Walmart, Target, Amazon and other large retailers. The plan was for Evenflo’s line to cost $39.99 to $49.99 initially, anticipating the booster’s annual sales could reach $11.4 million by its third year on the market.

Evenflo’s business plan for the Big Kid, which focused on mothers as the key purchasers, determined they did little research on booster seats. “Because all seats have to meet Federal safety standards, most moms feel that all seats have relatively the same level of safety,” it said. “Consequently, price and aesthetics drive the purchase.”

Evenflo didn’t just undercut Graco on price. When Evenflo launched the Big Kid in the early 2000s, Graco labeled its booster as safe for children at least 3 years old with a minimum weight of 30 pounds and minimum height of 38 inches.

Evenflo told parents the Big Kid was safe for babies as young as 1 as long as they weighed 30 pounds or more. There was no minimum height.

Dorel also sold boosters for children as young as 1. Indeed, the company has twice been sued in federal court by the families of 2-year-old children using its boosters, one who was paralyzed in a 2014 Iowa crash and another who was killed in a 2013 Alabama accident. Dorel settled both suits confidentially.

Evenflo’s engineers would later concede that 1-year olds don’t belong in Big Kids. Joshua Donay, an Evenflo project engineer on the Big Kid, said in a 2016 deposition in a case in Duval County Circuit Court in Florida that he would “not put a 1-year-old in any belt-positioning booster, Big Kid, Graco, you name it.” “I would keep them in an infant seat,” added Donay, who left Evenflo in 2005. Dahle, the top Evenflo booster seat engineer, acknowledged in a deposition in a different case that not only should a 1-year-old never use the Big Kid, a 2-year-old shouldn’t either.

The engineers were reflecting the scientific consensus that booster seats don’t adequately protect toddlers. To get the full safety benefit in crashes, the adult seat belt has to remain on the strong parts of a child’s body: across the middle of the shoulder and across the upper thighs. Even if toddlers are tall enough for the belt to reach the shoulders, children that young rarely sit upright for long and often wriggle out of position.

Watch: Video reveals how Evenflo put children at risk

In contrast, a tightly adjusted five-point harness secures shoulders and hips, and goes between the legs. Harnesses secure children’s bodies so that they are less likely to be ejected and disperse the crash forces over a wider area. There’s a reason NASCAR drivers wear harnesses.

While car seat recommendations have changed over the years, the American Academy of Pediatrics has been consistent about one key safety principle: Parents should not move children to a booster seat until they reach the maximum weight or height of their harnessed seat. Back in the early 2000s — before Evenflo, Graco, Dorel and others told parents boosters were safe for children weighing as little as 30 pounds — the American Academy of Pediatrics advised that kids who weigh 40 pounds or less were best protected in a seat with its own internal harness. That was the limit of most harnessed seats back then. Today, the overwhelming majority of harnessed seats on the market can accommodate children who weigh up to 65 pounds and who are as tall as 4 feet, 1 inch.

In Canada, the government does not allow the sale of boosters to children under 40 pounds. That’s been the case since 1987.

Evenflo has run afoul of Canada’s tighter rules. Its booster seats have been recalled three times there, most recently in 2012, for labeling that said a child could use the seat at 30 pounds or less. Canadian regulators also recalled a Dorel seat, the Safety 1st booster, for labeling that said it was safe for children as small as 30 pounds in 2002.

In pitching the Big Kid for small children, Evenflo capitalized on an important aspect of booster seats’ commercial appeal: Parents often face a barrage of whining from children who want out of harnessed seats they see as babyish. In addition, installing harnessed car seats can be a confusing, sweat-inducing chore, and studies show many aren’t installed correctly. Boosters are less complicated to use and can easily be moved from one vehicle to another. Parents often see switching from harnessed seats to boosters as an exciting milestone, but in fact, that transition reduces the protection a child would receive in a crash, said Dr. Ben Hoffman, an Oregon pediatrician and a lead author of the American Academy of Pediatrics’ policy statement on car seats.

Evenflo’s Blankenship disagreed, contending that boosters and harnessed seats “provide comparable overall safety.” A child’s head in a frontal collision can move farther forward in a harnessed seat if it’s installed without a tether, a strap that secures the seat to the car, compared with a booster, Blankenship said. She said federal data shows that only about half of child safety seats are installed using a tether.

She said Evenflo does not believe its recommendations are inconsistent with those of the American Academy of Pediatrics or the National Highway Traffic Safety Administration. As an illustration, she provided a page from an Evenflo manual that tells parents when to transition their children from a harnessed seat to a booster. But Evenflo’s advice runs counter to the American Academy of Pediatrics’ best practices. Referring to a harnessed seat as a “toddler seat,” the Evenflo manual said, “Even children who have not outgrown their toddler seat can benefit from the use of a booster seat, if it is used properly.”

Evenflo’s Big Kids sold briskly for several years before the company changed the marketing materials to say boosters were unsafe for babies and toddlers. In 2007, Evenflo increased the minimum age on the Big Kid to 3 and added a minimum height of 38 inches. Failure to follow those instructions, Evenflo warned, “can result in your child striking the vehicle’s interior during a sudden stop or crash, potentially resulting in serious injury or death.”

Federal law doesn’t specify a minimum age for boosters. Evenflo was not required to recall the Big Kid boosters it had labeled as safe for 1-year-olds. The company did not tell parents who had already bought the seats that they should stop putting toddlers in them.

In its first few years on the market, the Big Kid had become what Evenflo in an internal design review called “the reliable workhorse in the Evenflo platform stable.”

“Perceived Side Protection”

Despite the Big Kid’s success, by 2008 Graco was still outselling Evenflo. The marketing department wanted to make the Big Kid look more like the TurboBooster on the shelves of big box retailers, Dahle, the Evenflo engineer, said in a deposition in Jillian’s case last year. The company felt the Big Kid’s “on-shelf perception” was poor compared with the TurboBooster because Graco’s seat looked like it had more side support, he said in a deposition in another case.

To make its seat look more like Graco’s, Evenflo added side wings — curved extensions that protrude from the backrest of the Big Kid. One Evenflo document describing the strategy behind the product launch said the consumer benefits of these new side wings included “increased perceived side protection.”

Then Evenflo began to test its Big Kid with side wings. While Congress had directed the National Highway Traffic Safety Administration to develop side-impact testing rules for child safety seats in 2000, by 2008, it had not completed the task. (The national test, finally proposed in 2014, now languishes as federal regulators and the car seat industry wrangle over what makes a good test. NHTSA did not respond to repeated requests for comment.)

Absent a federal standard, Evenflo made up its own rules.

Videos from the company’s tests show child-sized dummies careening far outside the boundaries of the booster seat, where a child’s head, neck and spine would be vulnerable. While the purpose of a seat belt is to distribute the crash forces over the strong bones of the body — the shoulders and hips — the Evenflo test instead showed the belt slipped off the shoulder and wound up taut around the soft abdomen and ribs. In real life, that could cause internal organ damage.

Evenflo’s testing records showed the new Big Kid with side wings did no better than the old version.

The dummies in the seat with wings were just as prone to lurching far outside the seat, with their seat belts slipping off their shoulders and straining at their abdomens. In its written response, Evenflo’s Blankenship said the side wings do have “safety advantages in some side impact crashes,” declining to elaborate.

One video, taken from a deposition of Evenflo engineer Dahle in Jillian’s case, compares an Evenflo side-impact test of an old Big Kid on the left to a test of the Big Kid with side wings on the right. Both tests used a dummy modeled on a 3-year-old.

Both show the head, neck and spine vulnerable to serious injuries.

ProPublica showed the test videos to Hoffman, the pediatrician who was a lead author of the American Academy of Pediatrics car seat recommendations. In an interview, Hoffman said such violent movement at high speed of the dummy in the booster could lead to abdominal, brain and spinal injuries in a real child, including paralysis or death.

“This looks horrific, and I can’t imagine it being in any way shape or form better under real life circumstances,” Hoffman said.

The Big Kids had slight differences in fabric, base or model number, but no matter which Big Kid Evenflo tested, Dahle said in a 2019 deposition, the dummies’ bodies all moved in the same way.

After each test, a technician filled out a form that asked whether the test showed “dummy retention.” Jeremy Belzyt, an Evenflo senior test technician whose job involved crash testing boosters, explained in a deposition in another lawsuit that he would check “yes” or “no.”

Jeff Rowe, an attorney for Jillian and another girl who also suffered internal decapitation in a crash while seated in a Big Kid booster, asked Belzyt to explain what “dummy retention” meant.

“It's just did it stay in the seat or did it fall out of the seat and end up on the floor,” Belzyt said.

Using photos from Evenflo’s internal side-impact tests, Rowe asked whether Belzyt thought the Big Kid retained the dummies.

---

Rowe: I'm going to hand you — or I'm going to show you what has been marked Plaintiff's Exhibit 104A.

Rowe: Would you have marked dummy retention, yes?

Belzyt: Yes.

Later, Rowe handed Belzyt a photo of another Evenflo test using a dummy modeled after a 6-year-old.

Rowe: Here’s Plaintiff's Exhibit 122A.

Rowe: Is that a dummy retention, yes?

Belzyt: Yes.

Rowe later handed him a photo of another Evenflo test.

Rowe: All right. And then Plaintiff's Exhibit 137A, is that a dummy retention, yes?

Belzyt: Yes.

Rowe: And Plaintiff's Exhibit 138A, what would you answer for dummy retention?

Belzyt: Yes.

---

Belzyt said in his deposition that he sent his test reports to an engineer who would decide whether the Big Kid passed or failed.

These all passed.

Indeed, Belzyt, who worked at Evenflo for 13 years, said in an interview with ProPublica that he never performed a side-impact test on a booster that was deemed a failure.

To safety-minded parents, Evenflo presented its tests as tough, going above and beyond what the government required. (The federal government, after all, still does not require a side-impact test.) Evenflo once boasted on a corporate blog called “The Safety Net” that its “rigorous test simulates the government side-impact tests conducted for automobiles.”

That was misleading. One of the federal government’s tests simulates a side-impact crash by having a 3,015-pound barrier moving at 38.5 miles per hour smash into a vehicle. Another federal test involves a car pulled sideways at 20 miles per hour into a stationary pole. The Evenflo side-impact tests, by contrast, were conducted on a bench resembling a vehicle’s seat that moves at 20 miles per hour and is suddenly decelerated. It has no barrier smashing into the bench and no pole.

Sarah Haverstick, the Evenflo safety advocate who wrote the blog post, later said in a deposition in one of the Big Kid cases that she had never witnessed any of Evenflo’s side-impact tests. “Not being an engineer, I do not know specifics about it,” she said.

Asked about Haverstick’s claim that the company’s tests were comparable to the government’s, Evenflo’s Blankenship declined to specify which government test Haverstick was referring to, instead writing that “the forces” in Evenflo’s side-impact test are similar to those in a federal test for cars. She declined to make Haverstick available but said that Haverstick consulted with others who were well-versed in the company’s testing methods before writing her blog post.

Had Evenflo chosen not to develop a side-impact test, the company would have been criticized for “doing only the minimum required by law,” Blankenship wrote. She chastised ProPublica for attempting to “poke holes” in the company’s testing methods, writing that the questioning put the company in the position “of being damned if you do and damned if you don’t.”

The public has remarkably little access to information about booster seat makers’ safety test standards or results. In lawsuits, such materials are often sealed.

Jillian’s attorneys, Rowe and Doug Gentile, are also suing Graco in the District Court of Shelby County, Texas, on behalf of the parents of a boy who weighed 32 pounds and suffered a traumatic brain injury in a head-on crash while seated in a Graco TurboBooster. They allege that Graco knew its booster wasn’t safe for children under 40 pounds but sold it for them anyway. In a court filing, Graco denies the allegation.

Last year, the boy’s attorneys had a side-impact test performed on a Graco TurboBooster. The child-sized dummy in the Graco booster hurtled out of the shoulder belt in the same way it did in the Evenflo Big Kid. ProPublica shared with Graco the data and video from that test, which was performed in a lab at the Medical College of Wisconsin that also performs tests for federal researchers; Graco did not respond to questions about the test.

In 2008, Evenflo launched the new Big Kid with side wings, touting its testing. Subsequent marketing materials Evenflo sent Walmart, Target and Babies R Us emphasized in large bold letters that the top feature of the new Big Kid was that it was “Side Impact Tested.”

“Knowing that one in four automobile accidents are side impact collisions, we believe it’s important to go beyond the current government standards when designing the next generation of Evenflo car seats, including the Big Kid LX,” one marketing pitch read.

Children Are Injured

In November 2009, the family of a 4-year-old boy sued Evenflo in Onslow County Superior Court in North Carolina, alleging his Big Kid booster failed to protect him in a side-impact crash. The boy suffered a traumatic brain injury in a December 2006 crash that occurred as his mother was taking her two sons to day care. The family’s suit said Evenflo shouldn’t have marketed the seat for a child his size — about 36 pounds at the time of the accident — when the company knew he would have been safer in a seat with an internal harness.

The family and Evenflo agreed that the boy was properly strapped into his Big Kid booster, the earlier version without the side wings or side-impact testing label. The company later reached a confidential settlement with the family.

In the 35-page complaint, the family’s attorneys outlined more than a decade of recommendations from safety advocates, pediatricians and researchers that children under 40 pounds were far safer in harnessed seats.

As the North Carolina boy’s case was being litigated, the American Academy of Pediatrics in 2011 made a widely publicized safety announcement. Parents should keep their children in rear-facing child safety seats for as long as possible before moving them to forward-facing harnessed seats, the group said. And switching children from forward-facing harnessed seats to boosters when they reached 40 pounds was no longer ideal. By then, maximum weights on many harnessed seats had reached 65 pounds or 80 pounds, so the pediatrician group emphasized that children should remain in those seats and only switch to boosters when they outgrew their harnessed seats. “It is important to note that every transition is associated with some decrease in protection; therefore, parents should be encouraged to delay these transitions for as long as possible,” the pediatrician group wrote at that time.

That same day, NHTSA released new guidelines that echoed the pediatrician group’s recommendations.

The actions of the pediatrician group and federal regulators did not prompt Evenflo to change its marketing materials and manuals. The company still told parents the Big Kid was safe for children who weighed as little as 30 pounds and were as young as 3.

Across the border in Canada, Evenflo told parents in bold capital letters that a child less than 40 pounds risked “SERIOUS INJURY or DEATH” using the same model of Big Kid. In 2012, Evenflo considered making the instructions for seats in the U.S. match those in Canada, where by regulation they can only be sold for use by children 40 pounds and up.

That February, Dahle, the company’s top engineer for boosters, urged his colleagues to increase the minimum weight in the U.S. to 40 pounds and the minimum age to 4 for Evenflo boosters. In a PowerPoint presentation, Dahle wrote that there is an “increased risk of injury” for children between 3 and 4 riding in boosters instead of seats with harnesses because they are immature and may not sit properly. “Keeping the seat at 30 lbs encourages parents to transition them earlier because they can, and the booster is a less expensive option,” he wrote.

Evenflo should discourage early transitions, Dahle wrote. A harnessed seat, he said, “is the better option. We should encourage that behavior by modifying the weight rating to 40 lbs.” He added, “To overcome the misuse, we should follow the NHTSA Guidelines and increase the age rating to 4 yrs old also.”

To bolster his suggestions, Dahle sent his colleagues a 2010 federal report on booster seat effectiveness. Dahle highlighted the most important takeaway of this study, which looked at a decade of crash data: that 3- and 4-year-old children had a reduced risk of injury in crashes when they were using harnessed seats rather than boosters and that early graduation to boosters may “present safety risks.” Another section Dahle highlighted said that children should remain in harnessed seats until they are 4 or weigh 40 pounds and that harnessed seats may offer more side support and “better containment” for smaller children in crashes.

In a meeting days after Dahle sent his report, McKay Featherstone, then a senior marketing director at Evenflo, shot down Dahle’s weight recommendation, records show. In what Dahle described in a deposition as a “corporate decision,” the company raised the minimum age to 4 but still said the seat was safe for kids weighing as little as 30 pounds.

Dahle reflected on that decision in an email to a colleague when the notion of raising the weight to 40 pounds came up again six months later. “McKay vetoed that earlier this year,” Dahle wrote. “After much discussion he was not persuaded.”

Indeed, Featherstone was the Evenflo marketing executive who that July had chided a colleague for suggesting such a change and let him know that he had “looked at 40 lbs for the US numerous times and will not approve this.”

In a 2019 deposition, Dahle tried to walk back how he had described Featherstone’s veto in his email. Dahle said the marketing department did not have the power to veto safety engineers and that he ultimately agreed with the decision. He referred a reporter’s call to Blankenship, who declined to make him available for an interview. Blankenship said the consensus at the company was to keep the minimum weight at 30 pounds to accommodate tall, thin children who had outgrown their harnessed seats. Featherstone, who no longer works for Evenflo, forwarded ProPublica’s questions to the company and declined to comment.

By the time of that 2012 company decision, Evenflo had decided the Big Kid model that Jillian’s family would later buy looked “dated” compared with the competition. Jon Conaway, the senior product manager for car seats, said in a 2019 deposition that the company was phasing out that model, which the company sometimes called the “kill process.” Evenflo needed to “put lipstick on a pig” and jazz up new Big Kid models with lights and speakers, he explained. But it continued selling the old model with manuals dated 2008 that told parents the seat was safe for 3-year-olds.

In a 2019 deposition in Jillian’s case, Conaway displayed a notable incuriosity about one of his biggest product lines. He said he wasn’t aware of the broader debate within the company about when a child was heavy enough or old enough to move to a booster safely. He also said he didn’t know that the National Highway Traffic Safety Administration recommended delaying transitions to boosters until a child outgrew harnessed seats.

Gentile, one of Jillian’s attorneys, asked Conaway in a deposition last fall, “When did you first become aware of that NHTSA guideline, if ever?”

Conaway answered, “Right now, when you read it.”

Gentile pressed Conaway on whether he had read the federal study that said moving younger children to boosters too early was dangerous. Conaway said he didn’t get past Page 2 of the 13-page report. “I would have got to this math equation and immediately closed it,” he said.

Blankenship, declining to make Conaway available for an interview, said he worked with others who were “extremely familiar with the federal regulations and industry practices.”

Even for the seats Evenflo decided would get the increased age recommendation, the company did not always include this information on the box, where parents could see it as they were deciding what to buy, court records show.

Amy Arias had no way of knowing that her daughter, Rhea, was too young for the Big Kid when she bought the booster seat at a San Antonio Walmart in early 2013.

Although Evenflo had increased the seat’s minimum age to 4, the box Rhea’s seat came in only provided weight and height guidelines. Arias purchased the seat thinking it was safe for Rhea, then 3.

Rhea was always big for her age. She was on a cheerleading team and, according to her mom, whenever she had to be one of the kids lifted for a cheerleading stunt, “little girls would just cringe at the thought of having to hold her up.”

On Friday, April 26, 2013, Arias was driving with her mother and Rhea to Arias’ sister’s house for a girls’ weekend at the beach. Rhea fell asleep and her grandmother noticed that the child had crumbs all over her from eating the sandwiches she had prepared. She suggested they get off the highway to clean her up.

As Arias exited the highway, she remembers her mom saying, “No, Amy, no.” A tractor-trailer smashed into their car. The grandmother, who was seated on the side of the SUV that the truck hit, said in a deposition that she fractured her pelvis, broke some ribs and had hairline skull fractures, but made a full recovery.

Rhea, belted into her Big Kid, was on the side of the SUV opposite the collision. Like Jillian, though, Rhea suffered internal decapitation. She, too, is paralyzed from the neck down and must use a ventilator to breathe.

An emergency room record estimated her weight at 37 pounds. One of Evenflo’s experts in the family’s litigation against the company asserted that Rhea weighed 42 pounds based on other medical records. Even if that were true, she still would have been too young for the seat.

Evenflo’s Blankenship said in severe crashes, children will not remain within the confines of their car seats, no matter whether they’re in harnessed seats or boosters. Evenflo contends that Rhea’s crash was more severe than 99% of all side-impact crashes. But the expert hired by Rhea’s attorneys told ProPublica she wouldn’t have sustained such serious injuries had she been in a harnessed seat.

Though he was not speaking about any specific child, Evenflo’s Dahle acknowledged in a deposition last year that once the head of a child leaves the protective confines of the seat, the head can strike something or the neck can stretch too far and damage the spinal cord.

In that same deposition, Dahle and Jillian’s attorney Rowe watched a video that showed a dummy belted into a Big Kid move in just such a fashion during an Evenflo side-impact test. The seat belt slipped off the dummy’s shoulder and the dummy’s head and torso flailed far outside the seat. By Evenflo’s standard, this passed the test, Dahle said.

Rowe asked Dahle to compare that test with one Evenflo performed on the SecureKid, an Evenflo seat that can accommodate a child up to 65 pounds with an internal harness.

In that test, the head and torso of the dummy remained entirely within the confines of the harnessed seat. Rowe and Dahle looked at the test results side by side: the harnessed seat on the left, the Big Kid on the right.

Rowe asked Dahle which of the two seats put the dummy’s neck in severe extension. Dahle said that the Big Kid did.

“And which child is more at risk for injurious head contact?” Rowe asked.

“The one on the right,” Dahle said, referring to the dummy in the Big Kid.

Rowe later asked him: “As Evenflo's director of manufacturing engineering, do you think any reasonable parent concerned about his little girl's safety would move his child out of a five-point harness seat she still fits in and into a Big Kid if Evenflo showed him what a ‘pass’ was on these side-impact tests?”

“I don’t know how to answer that question,” Dahle replied.

Hoffman, a lead author of the American Academy of Pediatrics’ car seat recommendations, said that if he could show the testing videos to other pediatricians, they would “spread our gospel” that parents shouldn’t transition their children to boosters until they had outgrown their harnessed seats.

“What would you rather sit in?” he asked. “I mean, there’s no universe in which I would look at that booster and say, ‘Wow, I really want my head to snap like that.’”

The version of the Big Kid Jillian’s father, Jay Brown, bought had the “Side Impact Tested” label stitched into the seat back.

At the time of the crash, Jillian weighed just under 37 pounds. She could have stayed in her harnessed car seat for a long time, but when her father read the Big Kid safety guidelines, he thought she was ready for the move.

Brave

Before the crash, Jillian liked to be in motion.

The operator of her Long Island day care had horses, and at 5 years old, Jillian loved to ride a pony named Missy. She mastered the balance bike before her older sister did. The girls took gymnastics classes together.

Jillian, who goes by the nickname J.J., had an air of independence about her. In preschool she made friends easily, and her sister always wanted to be near her. But Jillian also liked to slip away and play on her own. At a trampoline birthday party, she jumped with the other kids but then left the group to jump by herself. Her favorite movie was “Brave.”

In April 2015, while Jillian was still in preschool, Rhea’s father sued Evenflo in U.S. District Court for the Southern District of Texas, alleging the Big Kid was “unreasonably dangerous” in side-impact collisions for children who weighed less than 40 pounds. He reached a confidential settlement with Evenflo.

In May 2016, as Jillian was finishing her final year of preschool, Evenflo came out with another Big Kid booster manual. The company had finally raised the weight minimum to 40 pounds but not for all of its Big Kids. This new set of instructions made it clear that Jillian was too small for the seat. But Evenflo never told its customers whose kids were already traveling in Big Kid boosters, so the Browns had no idea Jillian was in a seat that wasn’t safe for a child her size.

Jillian’s father, who is also represented by Aileen Kavanagh, sued Evenflo in July 2017. The case is ongoing.

Evenflo General Counsel Blankenship said the company chose to keep the minimum weight recommendation at 30 pounds to “provide options for parents whose children were too tall” for their harnessed seats but had not yet reached 40 pounds. The company changed some U.S. seats to 40 pounds so that it could sell them in Canada, she said.

She said Jillian’s crash was more severe than 98% of all side-impact crashes. Just as in the Arias case, the expert hired by Jillian’s attorneys told ProPublica she wouldn’t have sustained such serious injuries had she been in a harnessed seat.

At the time of the 2016 accident, Jillian was getting ready to start kindergarten. After the crash, she spent 10 weeks in the intensive care unit and another four months at a rehab hospital two hours from the family home. Her mom stayed with her for half the week, her dad the other half.

Jillian no longer had the ability to cough or clear secretions from her lungs on her own. Pneumonia, collapsed lungs and other ailments sent her back to the acute hospital four times during her stay in rehab.

After six months in different hospitals, the Browns wanted to go home. The discharge planner talked about two options: bringing Jillian home or admitting her to a long-term care facility.

“Well, not a question for me,” Lindsey Brown recalled.

While at the rehab center, the Browns could summon a medical team in emergencies with a push of a button. At home they had to learn to save Jillian’s life on their own. Twice, when new nurses didn’t recognize signs of trouble, the Browns knew how to resuscitate her.

Since the bathroom is on the second floor of their house in Ridge, New York, they have to unhook Jillian from her ventilator and carry her upstairs to bathe her. It’s a two-person job: one washes Jillian while the other pumps the Ambu bag, a football-shaped air bladder connected to a mask that helps her breathe. Eventually, Jillian will be too heavy to carry, and they’ll have to build a full bathroom on the first floor.

The Browns’ house already reflects Jillian’s transformed life. They moved her bedroom from the second floor to the downstairs dining room and installed a generator so if the power goes out, Jillian’s ventilator will keep running. The once pristine walls are marked with dents and scuffs from when Jillian learned to steer her power wheelchair with her tongue.

Longwood Central School District sent a teacher to their home until Jillian was stable enough to attend. She’s now in a mainstream class and, in third grade, is a great speller and reads books above her grade level. The district installed a wheelchair swing for her to use and encouraged her family to enter her in the wheelchair race at the annual Nassau County Empire State Games for the Physically Challenged. Jillian won.

Despite her resilience, her vulnerabilities weigh on the Browns. There is always a risk of what they call “popping off” — the ventilator tubing becoming disconnected — or getting clogged with Jillian’s secretions. With her voice as faint as it is, Jillian can’t cry out for help.

“You can’t even put the garbage out because she’s going to pop off,” Jay Brown said. “We can’t get the mail because she’s going to die while you’re getting the mail.”

Jillian needs 24-hour care, but nurses call in sick or need time off. The need to fill in for them, Lindsey said, led her to quit her job as a health insurance implementation manager last year.

Jillian’s care takes up so much time and energy that Jay and Lindsey Brown struggle to give Samantha the attention she needs.

They resist thinking too far in the future. Where will Jillian go when they’re too old and frail to care for her? They don’t want that responsibility to fall on Samantha. At this point, they don’t have a plan.

So far, Jillian hasn’t asked her parents about the accident, about the Big Kid, about her life. They know those questions will eventually come.

“We're not looking forward to that day,” Jay Brown said.

It is still possible to purchase Evenflo Big Kid boosters with the out-of-date, unsafe parameters. Last month, ProPublica ordered two Big Kid boosters directly from Evenflo’s website. The boxes did not specify an age for use. The labels said they were safe for children weighing as little as 30 pounds.

After ProPublica’s inquiries, the website changed: Now it says the minimum weight is 40 pounds. So ProPublica ordered another Big Kid from Evenflo’s website. The booster’s box, manual and label still say it is safe for children as small as 30 pounds.