American Honda Motor Co., Inc. v. Sarah Milburn

Supreme Court of Texas ══════════ No. 21-1097 ══════════

American Honda Motor Co., Inc., Petitioner,

v.

Sarah Milburn, Respondent

═══════════════════════════════════════ On Petition for Review from the Court of Appeals for the Fifth District of Texas ═══════════════════════════════════════

Argued September 13, 2023

JUSTICE LEHRMANN delivered the opinion of the Court, in which Chief Justice Hecht, Justice Blacklock, Justice Busby, Justice Bland, Justice Huddle, and Justice Young joined.

JUSTICE BLACKLOCK filed a concurring opinion, in which Justice Busby joined.

JUSTICE DEVINE filed a dissenting opinion, in which Justice Boyd joined.

In this products liability action involving an automobile manufacturer’s alleged negligent design of a seat-belt system, the primary issue presented concerns the statutory rebuttable presumption

of nonliability that attaches when a product’s design complies with applicable federal safety standards. The trial court rendered judgment on the jury’s verdict in the plaintiff’s favor, and the court of appeals affirmed. The court of appeals held, among other things, that legally sufficient evidence supports the jury’s findings that the presumption of nonliability applied and that the presumption was rebutted. We hold as a matter of law that the presumption both applied and was not rebutted. Because that holding is dispositive, we reverse the court of appeals’ judgment and render judgment for the manufacturer without addressing the remaining issues.

I. Background A. The Seat-Belt System

The standard seat-belt system in motor vehicles is a Type 2 system that integrates the shoulder belt and lap belt into one continuous piece of webbing with three points of attachment. Typically, the shoulder belt attaches to the car’s frame or seat, the lap belt attaches to the car’s floor or seat, and the passenger creates the third point of restraint by pulling the belt across her body and latching it into a buckle next to her hip. The seat-belt system at issue here is a ceiling-mounted detachable Type 2 anchor system, which is used in the third-row middle seat of the 2011 Honda Odyssey, among many other vehicles. The belt in the 2011 Odyssey is mounted to the ceiling on the right side of the middle seat and has a detachable anchor that latches into a minibuckle at the right hip. When the anchor is attached, the passenger creates the third point of restraint by buckling the belt at her left hip. But if the

passenger fastens the belt in that way when the anchor is not attached, and the passenger fails to reattach the anchor, the passenger’s lap will remain unbelted. A key-like object is required to detach the anchor. The reason for the detachability feature is to allow the seat belt to retract into the ceiling, which in turn facilitates the rear seat’s folding flat into the floor pan, significantly increasing the amount of cargo space without risking damage to the seat belt. Generally, the anchor is intended to remain connected when the seat is in the upright position. The Odyssey’s owner’s manual contains warnings about using the seat belt when the anchor is detached, and an additional warning label is located on the seat belt itself. It is undisputed that the detachable-anchor system used in the rear middle seat of the 2011 Odyssey complies with mandatory federal regulations, as the Federal Motor Vehicle Safety Standards expressly allow that system to be used in passenger cars manufactured after September 1, 2007, in the middle seating position of a seat for which the seat back can be folded nearly or fully flush with the floor. See 49 C.F.R. § 571.208, S4.2.7.4.

B. The Accident

On the evening of November 14, 2015, Sarah Milburn went out with five friends to a bar in Uptown Dallas. Shortly after midnight, the group decided to meet up with other friends at a different bar a short distance away. A member of the group called an Uber, and Uber driver Arian Yusufzai picked them up in a 2011 Odyssey. Milburn sat in the third-row middle seat. That seat’s detachable anchor was not latched, and the webbing was retracted into the ceiling-mounted retractor. Milburn fastened her seat belt by pulling the belt down from the ceiling

across her body and attaching it to the buckle at her left hip, leaving her lap unbelted. 1 As the Odyssey entered an intersection, a pickup truck hit the minivan on the passenger side, causing it to overturn and come to rest on its roof. 2 The impact caused Milburn to move forward and to the right until her neck was “clotheslined” by the shoulder-strap portion of the belt. While the other five passengers exited the van unassisted, paramedics extracted Milburn on a backboard and took her to the hospital. Milburn, who was twenty-three years old at the time of the accident, suffered severe injuries to her cervical vertebrae that rendered her a quadriplegic.

C. The Lawsuit

Milburn sued Honda, Uber Technologies and two of its subsidiaries, Yusufzai, and the Odyssey’s owner. 3 Before trial, Milburn settled with all defendants except Honda and amended her petition to assert only product liability claims against Honda. Milburn alleged that the Odyssey was “defective and unreasonably dangerous in that it was

not adequately designed, manufactured or marketed to minimize the risk of injury.” More specifically, she alleged that Honda was negligent in (1) designing the Odyssey “with a third-row center seatbelt system which an ordinary passenger would likely not be able [to] use as designed because the intended method of use was dangerously unclear, confusing, counter intuitive, and misleading”; (2) failing to adequately test and evaluate the usability and safety of the seat belt system; and (3) failing to provide adequate warnings and instructions. 4 Honda raised several affirmative defenses in its answer. Relevant here, Honda first pleaded that the Odyssey’s compliance with mandatory federal safety standards gave rise to a presumption of nonliability under Texas Civil Practice and Remedies Code Section 82.008. That statute entitles a manufacturer to a rebuttable presumption that it is not liable in a products liability action based on a product’s design if the manufacturer establishes that the design complied with applicable federal safety standards or regulations “that governed the product risk that allegedly caused harm.” TEX. CIV. PRAC. & REM. CODE § 82.008(a). The claimant may then rebut that presumption by establishing that the standards or regulations “were inadequate to protect the public from unreasonable risks of injury or damage.” Id. § 82.008(b)(1). As an additional affirmative defense, Honda pleaded that, under Civil Practice and Remedies Code Chapter 33, its liability should be reduced by the “percentage of responsibility” attributable to Milburn,

the settling defendants, and the driver of the pickup truck. As to the Uber entities specifically, Honda alleged that their negligence, negligent undertaking, and fraudulent and negligent misrepresentations were a producing cause of Milburn’s injury. Milburn specially excepted to Honda’s pleadings relating to the Uber entities’ proportionate responsibility under Chapter 33, arguing that (1) submission of the Uber entities for comparative apportionment would be improper because Uber’s liability was derivative of another tortfeasor (Yusufzai) and thus (2) all evidence regarding the allegations against the Uber entities should be excluded as irrelevant. Milburn then moved for partial summary judgment, arguing no viable legal claim would permit the submission of the Uber entities in a proportionate-responsibility question. The trial court granted the motion. The parties proceeded to a three-week jury trial. Milburn presented two expert witnesses to testify about the seat-belt system’s design. Joellen Gill, a human factors engineering consultant, 5 testified that the seat-belt system’s design was defective from a human factors perspective because it was foreseeable that owners would not reliably maintain the belt in the anchored position and that passengers would neither fasten the unanchored belt correctly nor recognize that they were belted incorrectly. And she concluded that Honda did not adequately mitigate those foreseeable risks. In so opining, Gill

explained that her conclusions were supported by two “usability studies” conducted by Milburn’s counsel in which fifty out of fifty-three people asked to fasten an unanchored seat belt in a 2011 Odyssey’s rear center seat did so incorrectly. Honda sought to exclude Gill’s testimony on the grounds that she was not qualified and that her testimony was unreliable; Honda further sought to exclude the usability studies as unreliable and unscientific. The trial court denied those requests. Milburn also presented the expert testimony of Steven Meyer, a mechanical engineer specializing in automotive safety, who concluded that the likelihood of the seat belt’s incorrect usage, weighed against the utility of the additional cargo space, rendered the system unreasonably dangerous as designed. He further testified that a safer alternative design that is both economically and technologically feasible is an “all-belts-to-seat” (ABTS) system, meaning the belt is attached to the seat, with no detachability feature, and thus cannot be erroneously used as a two-point belt. 6 Honda presented its own engineering and human factors experts to testify about the product’s design. Michael Klima, a mechanical engineer, opined that the Odyssey’s seat-belt system was “well-designed, consistent with peers out in the motoring public today.” He further opined that the ABTS alternative design Meyer suggested would require additional seat height and bulk, which in turn would interfere with the ability to retract the seat into the floor. Nathan

Dorris, a human factors and safety consultant, opined that the design of the seat-belt system was “reasonably safe . . . from a human factors perspective,” taking into account expected knowledge and attitudes regarding seat belts, including the expectation of a lap belt and the ability to ask a driver for assistance. Honda’s corporate representative testified that no other complaints or lawsuits regarding the seat belt had been filed and that the system “is used widely throughout Honda and the rest of the industry and has been for almost 20 years.” When asked whether Honda conducted any usability studies with respect to the detachable seat belt in the 2011 Odyssey, he explained that Honda would not have done such a study “since it wasn’t a new technology” and was “used in the marketplace for almost over 10 years before it was applied to this vehicle.” Honda began using roof-mounted detachable seat belts in its vehicles in 2001 or 2002 and had used them in the Odyssey since 2005. In a single liability question, the jury was asked whether Honda was negligent in designing the 2011 Odyssey and whether Honda’s negligence, if any, proximately caused Milburn’s injury. The charge instructed the jury that “[f]or American Honda to have been negligent, there must have been a defect in the designing” of the Odyssey. The charge went on to define “design defect” as “a condition of the product that renders it unreasonably dangerous as designed, taking into consideration the utility of the product and the risk involved in its use. For a design defect to exist there must have been a safer alternative design.” In turn, the charge defined “safer alternative design” as a product design other than the one actually used that in reasonable probability (1) would have prevented or

significantly reduced the risk of the injury in question without substantially impairing the product’s utility and (2) was economically and technologically feasible at the time the product left the control of American Honda by the application of existing or reasonably achievable scientific knowledge. (Formatting altered.) The jury answered “yes” to the liability question. The charge then asked the jury two questions about the Odyssey’s compliance with federal safety regulations. The jury first found that the Odyssey’s design complied with mandatory federal safety standards or regulations “that were applicable to the product at the time of manufacture and that governed the product risk that allegedly caused harm,” meaning that the statutory presumption of nonliability applied. However, the jury also found that those standards or regulations were “inadequate to protect the public from unreasonable risks of injury or damage,” meaning Milburn rebutted that presumption. The jury went on to find that Milburn’s and Yusufzai’s negligence also proximately caused the injury in question and attributed 63% of the responsibility to Honda, 32% to Yusufzai, and 5% to Milburn. Finally, the jury found that approximately $37 million in damages would fairly and reasonably compensate Milburn for her injuries. The trial court rendered judgment on the verdict, reducing the amount awarded by 5% for Milburn’s proportionate responsibility and applying a credit based on Milburn’s settlements with the other defendants. Because the jury attributed more than 50% of the responsibility to Honda, the trial court deemed Honda jointly and severally liable for Milburn’s recoverable damages. See TEX. CIV. PRAC. & REM. CODE § 33.013(b)(1). Accordingly,

the trial court’s judgment ordered that Milburn recover from Honda just under $26 million in damages and prejudgment interest. Honda appealed, arguing that the evidence was legally and factually insufficient to support the jury’s finding that Honda negligently designed the Odyssey, proximately causing the injury in question, and the jury’s finding that the nonliability presumption was rebutted. Honda further argued that the trial court erred in excluding the settling Uber entities from the jury’s consideration of proportionate responsibility. The court of appeals affirmed. 668 S.W.3d 6 (Tex. App.—Dallas 2021). On the sufficiency issues, the court concluded that the case amounted to a “battle of the experts” and that it was within the jury’s province to resolve the conflicting evidence in Milburn’s favor. Id. at 20, 26. In so holding, the court rejected Honda’s challenges to Gill’s qualifications and the reliability of her testimony. Id. at 22–24. Regarding the evidence of a safer alternative design, the court concluded that Meyer sufficiently explained how the ABTS system could be implemented while maintaining the stowaway feature of the rear seat and that, even if his testimony on that point was conclusory, conflicting evidence was presented about whether the loss of cargo space associated with the stowaway seat would substantially impair the utility of the product. Id. at 25–26. On the proportionate-responsibility issue, the court of appeals held that (1) Honda’s negligence theories could not support submission of the Uber entities to the jury to apportion liability because those entities’ liability was derivative of Yusufzai’s and (2) no

evidence supported a fraud or negligent-misrepresentation claim. Id. at 31–34. In its petition for review, Honda renews its legal sufficiency challenges to the jury’s findings of regulatory inadequacy and negligence. 7 Honda alternatively argues that it is entitled to rendition of judgment under federal conflict-preemption principles. Finally, Honda asserts that, if rendition is inappropriate, it is nevertheless entitled to a new trial because the trial court erred in refusing to submit the Uber entities in the proportionate-responsibility question.

II. Discussion

A products liability claim involving a design defect requires proof that the product’s design rendered it unreasonably dangerous, taking into consideration the utility of the product and the risk involved in its use, and that a safer alternative design existed. Timpte Indus., Inc. v. Gish, 286 S.W.3d 306, 311 (Tex. 2009); TEX. CIV. PRAC. & REM. CODE § 82.005(a). 8 When the product at issue is a motor vehicle, the vehicle’s compliance with applicable federal motor vehicle safety standards is

often presented as evidence tending to show the design was not defective. A motor vehicle safety standard is “a minimum standard for motor vehicle or motor vehicle equipment performance,” 49 U.S.C. § 30102(a)(10), and a vehicle may not be manufactured for sale absent compliance with all applicable safety standards, id. § 30112(a)(1). Under federal law, such compliance does not exempt the manufacturer from liability at common law. Id. § 30103(e). But under Texas law, it does under certain circumstances. Specifically, in 2003, the Legislature enacted Civil Practice and Remedies Code Section 82.008 in response to “a finding that manufacturers and sellers were being held liable in products liability cases even though the products at issue complied with all applicable federal safety standards.” Kia Motors Corp. v. Ruiz, 432 S.W.3d 865, 869 (Tex. 2014). Section 82.008 entitles a product manufacturer to a presumption that it is not liable for injuries caused by its product’s design if the manufacturer establishes that (1) the design complied with mandatory federal safety standards or regulations, (2) the standards or regulations were applicable to the product at the time of manufacture, and (3) the standards or regulations governed the product risk that allegedly caused the harm. Id. (citing TEX. CIV. PRAC. & REM. CODE § 82.008(a)); 9 see also Wright v. Ford Motor Co., 508 F.3d 263, 271 (5th

Cir. 2007) (“Section 82.008(a) is not limited to preemptive regulations, and, in fact, appears to assume non-preemptive regulations . . . .”). 10 That presumption, however, is rebuttable. A claimant may rebut the presumption by establishing either that “the mandatory federal safety standards or regulations applicable to the product were inadequate to protect the public from unreasonable risks of injury or damage” or that “the manufacturer, before or after marketing the product, withheld or misrepresented information or material relevant to the federal government’s or agency’s determination of adequacy of the safety standards or regulations at issue in the action.” TEX. CIV. PRAC. & REM. CODE § 82.008(b). Only the rebuttal ground premised on inadequacy of the applicable federal safety standards is at issue. As in other cases addressing Section 82.008, the parties and the lower courts here treated the applicability of the presumption as an affirmative defense, on which Honda bore the burden of proof, and the presumption’s rebuttal as an exception to or matter in avoidance of that

defense, on which Milburn bore the burden. 11 See Kia Motors, 432 S.W.3d at 870 n.5 (noting that Kia raised the presumption as an affirmative defense in its pleadings and challenged the jury’s verdict in part because no evidence was presented rebutting the presumption); Wright, 508 F.3d at 274 (noting that if the rebuttal evidence relied on by the plaintiff “presents a fact question [with respect to regulatory inadequacy], then, in a jury tried case, it appears logical to conclude that the statute proceeds on the assumption that any such fact question as whether the presumption has been rebutted will be submitted to the jury”). The statutory language supports that treatment to some extent, as a defendant manufacturer or seller must “establish[]” that the presumption applies due to the product’s compliance with applicable safety standards, TEX. CIV. PRAC. & REM. CODE § 82.008(a), and if it does so, the claimant must then “establish[]” the inadequacy of those safety standards to rebut the presumption, id. § 82.008(b). As discussed below, however, the presumption’s applicability in the first instance will rarely lend itself to resolution by the factfinder. Further, considering the statutory context and structure of Section 82.008, 12 the presumption, when applicable, essentially

presents an independent hurdle that the claimant must clear to establish a manufacturer’s or seller’s liability on a defective-design claim. If the manufacturer establishes the presumption’s applicability and the claimant fails to rebut it, the manufacturer is “not liable” on the claim, and further inquiry into the claim’s elements is immaterial. Id. § 82.008(a). But if the claimant rebuts the presumption, the defendant’s compliance with applicable safety standards becomes just another piece of evidence relevant to whether the product was defectively designed.

A. Applicability of Presumption

We first address the presumption’s applicability: whether the 2011 Odyssey (1) complied with mandatory federal safety regulations that (2) were applicable to the Odyssey at the time of its manufacture and (3) “governed the product risk that allegedly caused the harm.” Id. The federal safety regulations at issue—the Federal Motor Vehicle Safety Standards—are promulgated by the National Highway Traffic Safety Administration (NHTSA). Although this issue was presented to the jury—who resolved it in Honda’s favor—and Milburn challenged the finding on appeal on no-evidence grounds, Milburn also favorably cites a recent Fifth Circuit decision for the proposition that the presumption’s applicability is a question of law for the court. Kim v. Am. Honda Motor Co., 86 F.4th 150, 169–70 (5th Cir. 2023). As we implied in Kia Motors, we agree that the presumption’s applicability, at least in most instances, will be a question of law rather than one of fact. See 432 S.W.3d at 874 (holding as a matter of law that the presumption did not apply because the pertinent federal safety standards did not govern the product risk that

allegedly caused the harm). Determining a product’s compliance with identified, mandatory safety regulations that were applicable to the product at the time of manufacture will typically be a straightforward affair with little room for disagreement. Indeed, there is no dispute here that the 2011 Odyssey complied with the mandatory motor vehicle safety standards that govern seat belts and were applicable to the Odyssey at the time of its manufacture. The only disputed issue is whether those standards “governed the product risk that allegedly caused the harm.” But that question too lends itself to resolution as a matter of law, see id., because it requires the interpretation of a government agency’s regulation in order to determine its relation to the relevant product risk—a quintessentially judicial task. See Wal-Mart Stores, Inc. v. Xerox State & Loc. Sols., Inc., 663 S.W.3d 569, 577 n.26 (Tex. 2023) (“As we do, federal courts interpret regulations by applying similar construction principles used to interpret statutes.”); State v. Shumake, 199 S.W.3d 279, 284 (Tex. 2006) (“Statutory construction is a question of law . . . .”). While we do not foreclose the possibility that there could be disputed relevant facts, whether the presumption applies in the first instance will generally be a question of law. With that backdrop in mind, we examine the pertinent safety standards and the product risk they allegedly govern. NHTSA promulgated the Federal Motor Vehicle Safety Standards in accordance with congressional direction to prescribe such standards for the purpose of “reduc[ing] traffic accidents and deaths and injuries resulting from traffic accidents.” 49 U.S.C. § 30101; see also 49 C.F.R. § 571.1. Most pertinent here is Safety Standard 208, which “specifies

performance requirements for the protection of vehicle occupants in crashes.” 49 C.F.R. § 571.208, S¹. The regulation’s stated purpose is “to reduce the number of deaths of vehicle occupants, and the severity of injuries, by specifying vehicle crashworthiness requirements in terms of forces and accelerations measured on anthropomorphic dummies in test crashes, and by specifying equipment requirements for active and passive restraint systems.” Id. § 571.208, S² (emphasis added). With narrow exceptions, Safety Standard 208 requires all passenger cars manufactured on or after September 1, 2007, to be equipped with a “Type 2 seat belt assembly that conforms to Standard No. 209 and to S7.1 and S7.2 of this standard at each rear designated seating position.” 13 Id. § 571.208, S⁴.1.5.5.1. The regulations go on to state: Any inboard designated seating position on a seat for which the entire seat back can be folded (including the head restraints and any other part of the vehicle attached to the seat back) such that no part of the seat back extends above a horizontal plane located 250 mm above the highest [seat reference point] located on the seat may meet the requirements of S4.1.5.5.1 [to equip the car with a Type 2 seat belt assembly that conforms to Standard 209 and to S7.1 and S7.2 of Standard 208] by use of a belt incorporating a release mechanism that detaches both the lap and shoulder portion at either the upper or lower anchorage point, but not both. The means of detachment shall be a key or key-like object.

The referenced standards govern requirements for seat-belt 13

assemblies such as hardware, durability, adjustment, and latch mechanisms. See 49 C.F.R. §§ 571.208, S7.1, S7.2; 571.209.

Id. § 571.208, S4.1.5.5.2. The third-row middle seat of the 2011 Odyssey qualifies as such an “inboard designated seating position on a seat for which the entire seat back can be folded . . . .” Id. Thus, Safety Standard 208 expressly authorizes “use of a belt incorporating a release mechanism that detaches both the lap and shoulder portion at either the upper or lower anchorage point, but not both,” as the means of equipping the vehicle with the required Type 2 seat-belt assembly for that seating position. Id. The Odyssey undisputedly complies with that standard by using a belt with a release mechanism that detaches both the lap and shoulder portion at the lower anchorage point, with a key-like object as the means of detachment. Relying largely on our decision in Kia Motors, Milburn argues that the Odyssey’s compliance with Safety Standard 208 is irrelevant because it does not govern the product risk at issue—specifically, “the risk that owners, drivers, and passengers will fail to reliably use the detachable seat belt system in a correct manner.” Honda responds that the product risk at issue is the dual detachability of the lap and shoulder belt, which the regulation expressly allows. We do not view those risks as mutually exclusive, as the risk of misuse accompanies a detachable seat-belt system. And for the reasons discussed below, we also conclude that Safety Standard 208 governs that risk. We begin with a discussion of Kia Motors, which guides our inquiry on this issue and is cited by both parties to support their opposing positions. That case involved a products liability suit against Kia for injuries suffered when a vehicle’s frontal air bag failed to deploy during a collision due to allegedly defective wiring connectors in the

air-bag system. Kia Motors, 432 S.W.3d at 868. The vehicle complied with Safety Standard 208’s requirement that it be equipped with front driver’s side and passenger’s side air bags that met certain protection criteria during a crash test. Id. at 870 (citing 49 C.F.R. § 571.208, S⁵.1, S⁶). Nevertheless, we held that Section 82.008’s presumption of nonliability did not apply because the regulation had no bearing on the risk associated with the alleged product defect: “the risk of occupant injury due to the failure of the air bag to reliably activate and deploy.” Id. at 874. We noted that the result would have been different, and the presumption would have applied, had the alleged defect been the lack of specific occupant-restraint equipment—such as a third frontal air bag— that the regulation did not require. Id. Milburn argues that, just as the regulation at issue in Kia Motors “presume[d] air bag deployment” and did “not measure or apply to air bag failure rates,” id. (emphasis omitted), the regulations applicable to seat belts “all presume that the seat belts are correctly being used in their intended manner” and do not contemplate the risk of misuse. That assertion ignores an important distinction between the air-bag regulations at issue in Kia Motors and the seat-belt regulations at issue here. It is true that the Safety Standards’ mandated crash tests measure the efficacy of seat belts when used in their intended manner, just as they measure the efficacy of air bags when they properly deploy. However, unlike the air-bag regulations reviewed in Kia Motors, the seat-belt regulations are not limited to how seat belts perform in a crash; they also contemplate equipment types, locations, and designs for various seating positions. And they expressly authorize the precise

design that Milburn alleges is defective: a system with a release mechanism that detaches both the lap and shoulder portion at the lower (or upper) anchorage point. In that respect, the hypothetical we posited in Kia Motors—that the presumption would apply to an alleged defect premised on the lack of a third frontal air bag when the regulation only required two—is instructive. Milburn essentially alleges that the Odyssey was defective because it used a dual-detachable seat-belt system in the third-row center seat instead of an ABTS system. But while Safety Standard 208 mandates a Type 2 assembly in that seating position, it does not mandate an ABTS system. Instead, NHTSA deliberately chose to allow the dual-detachable system. 14 Indeed, any detachable system necessarily involves a risk of misuse, and that risk was contemplated when NHTSA assessed whether to allow this type of system at all; that is, the risk of misuse was part of the cost–benefit analysis. NHTSA said as much in its published comments in 2004 to the final rule amending Safety Standard 208 to allow the dual-detachable system. See 69 Fed. Reg. 70904, 70908–09 (Dec. 2004). In those comments, NHTSA noted that the then-existing standard permitted detachability of the shoulder portion of the belt in certain seating positions but that many manufacturers were already using “a ‘minibuckle’ design that permits the entire belt to detach from the seat and retract into the upper shoulder anchorage,” like the one at issue here. Id. at 70908. That design was also being used in seats for

which manufacturers, at the time, were given the option of using either a Type 1 (lap-belt only) or Type 2 system, such as rear center seats and rear outboard seats located adjacent to an aisle. Id.; see 49 C.F.R. § 571.208, S4.1.4.2(a) (1989) (requiring most passenger cars manufactured on or after September 1, 1990, to be equipped with a Type 2 system “at every forward-facing rear outboard designated seating position”); id. § 571.208, S4.1.5.1(a)(2) (1993) (requiring most passenger cars manufactured on or after September 1, 1996, to be equipped with either a Type 1 or Type 2 seat belt assembly “at any rear designated seating positions that are not ‘rear outboard designated seating positions’”). “Given the advances in safety belt technology,” NHTSA was considering “whether it was appropriate to reconsider the detachability allowances for these seats.” 69 Fed. Reg. at 70908. NHTSA continued: In the [Notice of Proposed Rulemaking], the agency acknowledged that for certain seat designs Type 2 belts could not be installed without integrating the upper shoulder anchorage into the seat back or permitting designs that allow for detachability of the shoulder belt. Because detachable belts can be misused, we were particularly interested in exploring the possibility of integrated belts. Id. Ultimately, NHTSA concluded that “integrated [e.g., ABTS] belt designs are not an optimal design for all types of seats,” and it “decided to retain the existing detachability provisions with some revision” and “to expand the detachability provision to the inboard seating position of folding seats [like those in the Odyssey], bus seats, and outboard seats adjacent to an aisle.” Id. The limitation that prohibiting detachability would have on effective use of cargo space and the increased cost

associated with incorporating an integrated belt were both relevant to that expansion. Id. at 70908–09. In approving the dual-detachable minibuckle design, NHTSA also chose to expressly prohibit the use of a pushbutton mechanism to detach the belt, instead requiring use of a key-like object, to reduce the likelihood of inadvertently releasing the minibuckle. Id. Finally, NHTSA expressly “decided against permitting detachable belts at the outboard seating position of [folding] seats because . . . we believe manufacturers can use the roof or side pillars to attach the upper shoulder anchorage.” Id. at 70909 (emphases added). This discussion reflects a cost–benefit analysis involving the benefits of Type 2 detachable belts—including the kind of detachable belt used in the Odyssey—and the risks associated with their use. Indeed, the only discernible reason for limiting the use of detachable belts to certain seating positions is that the risk of misuse outweighs any potential benefits in other positions. NHTSA concluded that detachable belts were a sufficiently safe option for implementing a Type 2 seat-belt system at some seating positions considering the benefits associated with those systems, despite the recognized risk of misuse. Milburn argues that the regulation itself does not discuss the risk that people will not understand how to operate the detachable system. True, but it does “specify[] equipment requirements for active and passive restraint systems” for the purpose of “reduc[ing] the number of deaths of vehicle occupants[] and the severity of injuries,” 49 C.F.R. § 571.208, S2, and again, in promulgating those requirements NHTSA took into consideration the risk that detachable systems will be misused.

As noted, in Kia Motors, the regulation’s equipment requirements did not encompass (or even contemplate) the design aspects of the air-bag system intended to ensure that the air bag deployed when necessary. Here, the equipment requirements did contemplate and encompass certain design aspects of the seat-belt system intended to ensure adequate occupant restraint, including the allowance of Type 2 systems with a dual-detachable belt in certain seating positions. We hold that the 2011 Odyssey’s design complies with mandatory federal safety standards “that were applicable to the product at the time of manufacture and that governed the product risk that allegedly caused harm.” TEX. CIV. PRAC. & REM. CODE § 82.008(a). Accordingly, Honda was entitled to Section 82.008’s presumption of nonliability for injuries caused by the Odyssey’s design. 15 We therefore turn to whether the presumption was rebutted.

B. Rebuttal of Presumption: Evaluating Inadequacy of Safety Standard We have not had occasion to address what is necessary to establish that “the mandatory federal safety standards or regulations applicable to the product were inadequate to protect the public from unreasonable risks of injury or damage” under Section 82.008(b)(1),

thereby rebutting the presumption of nonliability. See Kia Motors, 432 S.W.3d at 874. However, the issue is squarely before us today. 16 The jury agreed with Milburn that the applicable safety standards were inadequate to protect the public, and Honda challenges the legal sufficiency of the evidence to support that finding. We review legal sufficiency challenges to a jury’s verdict by considering whether the evidence at trial, viewed in a light favorable to the verdict, “would enable reasonable and fair-minded people to reach the verdict under review.” Pediatrics Cool Care v. Thompson, 649 S.W.3d 152, 161 (Tex. 2022) (citations omitted). In arguing the insufficiency of the evidence, Honda asserts that a showing that a federal safety standard was “inadequate to protect the public from unreasonable risks of injury or damage” requires the testimony of a “qualified regulatory expert.” “Among other things,” Honda continues, that expert “would need to explain why, in the context of the entire regulatory history and the delicate balance between absolute safety and commercial feasibility, the agency’s determination

was an ‘inadequate’ solution to ‘protect’ against an ‘unreasonable’ risk faced by the entire public.” Here, Honda asserts, neither Gill nor Meyer evaluated or criticized the regulatory decision-making underlying NHTSA’s adoption of the pertinent standard. Milburn responds that the evidence demonstrates NHTSA “did not consider the risk that the detachable seat belt system would not be usable, and it did not require manufacturers to test for usability.” Further, she criticizes NHTSA’s decision to allow the detachable system as being driven by cost rather than safety. Finally, Milburn asserts that the absence of other customer complaints, lawsuits, and recalls cannot by itself overcome the jury’s finding of regulatory inadequacy. As an initial matter, we note that whether a product’s design is defective and whether the applicable safety standards with which the design complies are inadequate to protect the public necessarily constitute distinct inquiries. 17 If the standard for rebutting the presumption mirrored the standard for a product defect, then the presumption would serve no purpose at all. See TIC Energy & Chem., Inc. v. Martin, 498 S.W.3d 68, 74 (Tex. 2016) (“In [construing a statute], we consider the statute as a whole, giving effect to each provision so that none is rendered meaningless or mere surplusage.”). Accordingly, a determination that an applicable federal safety standard is inadequate to protect the public from unreasonable risks of injury requires

something other than proof of a product’s defective design—that is, something more than a conclusion that the risks outweighed the benefits with respect to the particular product design at issue. The court of appeals essentially ignored this distinction, summarizing its analysis of regulatory inadequacy as follows: Like many product liability cases, this case was a battle of the experts. The Milburns’ experts examined physical evidence, performed tests, reviewed data, performed calculations, criticized Honda’s experts, and concluded that the federal standards pertaining to the ceiling-mounted detachable anchor seat belt system for the third-row middle seat were inadequate to protect the public from unreasonable risks of injury or damage. Honda’s experts conducted the same examinations, tests, reviews, calculations, and critiques of the Milburns’ experts, but concluded that the standards were adequate to protect the public. The jury properly exercised its prerogative to resolve this conflicting evidence and believed the Milburns’ experts. 668 S.W.3d at 20. Assuming the admissibility of the evidence the court of appeals described, 18 the court of appeals made no effort to explain how

the showing required to prove regulatory inadequacy is in any way distinguishable from the showing required to demonstrate that the design of the Odyssey’s seat belt system for the rear center seat was defective—that is, unreasonably dangerous considering the product’s utility and the risks involved in its use. Timpte Indus., 286 S.W.3d at 311. Milburn similarly argues that “the fact that a product is defective [e.g., by virtue of its confusing design] is some evidence that the federal standards allowing [that] design are inadequate to protect the public.” But again, defective design and regulatory inadequacy are necessarily independent inquiries. And while evidence tending to show a product was defectively designed can certainly also be relevant to the adequacy of the regulation allowing that design, legally sufficient evidence of the former does not automatically amount to legally sufficient evidence of the latter. By conflating what is necessary to prove liability with what is necessary to rebut the presumption, both Milburn and the court of appeals render the presumption superfluous.

The dissent claims this conclusion rests on a faulty premise; the standards can mirror each other because the presumption highlights for the jury the significance of compliance with the federal standards, which benefits the manufacturer. Post at 6 (Devine, J., dissenting) (citing Egbert v. Nissan N. Am., Inc., 167 P.3d 1058, 1061 (Utah 2007)). We disagree. Had the Legislature merely wanted to highlight such compliance as part of the evaluation of whether a product is defectively designed, it could have done so. See, e.g., OHIO REV. CODE § 2307.75. Instead, the Legislature chose to create an independent presumption that can be rebutted only by establishing the inadequacy of the standard itself, allowing manufacturers to reasonably rely on those standards when designing their products. And the dissent ultimately accepts that the standards do not mirror each other because one focuses on the particular product and its intended user, while the other focuses on the applicable safety standard and the public as a whole. See post at 7 (Devine, J., dissenting). On this point, at least, we agree. At bottom, if Milburn’s interpretation of the statute is correct, then the analysis is as follows: a defendant is liable for a defective design if the plaintiff proves X, unless the product complies with an applicable federal safety standard, in which case the defendant is not liable unless the plaintiff proves X (which has already been proven). Such a reading would effectively delete the presumption of nonliability. We do not read Section 82.008 to provide such illusory protection to compliant manufacturers. Turning to the safety standard at issue, if evidence existed, for example, that NHTSA simply failed to consider the possibility of misuse

associated with detachable belts or based its decision to allow detachable belts solely on cost without any consideration for safety, we do not disagree that such evidence could overcome the presumption that the regulation adequately protected the public. See Nat’l Truck Equip. Ass’n v. NHTSA, 711 F.3d 662, 667 (6th Cir. 2013) (noting, in the context of a challenge to the validity of a federal agency rule, that federal courts will invalidate an agency action if, among other things, the agency has “entirely failed to consider an important aspect of the problem [or] offered an explanation for its decision that runs counter to the evidence before the agency” (quoting Motor Vehicle Mfrs. Ass’n of U.S. v. State Farm Mut. Auto. Ins. Co., 463 U.S. 29, 43 (1983))). But no such evidence was presented. Gill testified that she had not read the Safety Standards or the regulatory history, nor did she purport to evaluate NHTSA’s decision-making process. And to the extent Meyer opined that the standards were inadequate, his reasoning was that the permitted detachable design was unreasonably dangerous; in other words, he disagreed with NHTSA’s conclusion that the permitted design was not unreasonably dangerous. Meyer’s conclusory opinion that NHTSA got it wrong is not enough. Further, NHTSA’s thorough documentation of its decision to approve the dual-detachable seat belt for use in vehicles like the Odyssey provides no indication that NHTSA simply failed to consider the risks associated with a detachable belt. Indeed, the risk of misuse associated with detachable belts is the very reason NHTSA was reconsidering the allowance of detachable belts at all. 69 Fed. Reg. at

70908. 19 NHTSA weighed the benefits associated with requiring a Type 2 belt assembly (rather than allowing manufacturers the option of a Type 1 or Type 2 system), the risk of misuse associated with a Type 2 system that incorporates a detachable belt, the cost and feasibility of requiring a fully integrated seat-belt system in certain vehicles—such as those with removable or folding seats—and the risk of inadvertently detaching the minibuckle used in the authorized design. Nothing in the record suggests that the conclusions NHTSA reached ran counter to the evidence before it. Moreover, NHTSA recognized that many manufacturers were already using the kind of dual-detachable system that NHTSA ultimately approved in the rule. It was certainly within the agency’s province to consider the practical experience gleaned from the real-world use of the system, and there is no indication that NHTSA ignored any red flags related to that usage. As another example, to the extent Milburn contends that NHTSA simply “got it wrong”—that is, improperly weighed the myriad factors that go into a motor vehicle safety standard—we agree with Honda that considerably more evidence of the various considerations NHTSA must take into account in making regulatory determinations about vehicle safety is required than Milburn presented. For instance, in discussing an earlier safety standard that gave manufacturers the choice of installing either a Type 1 or a Type 2 belt in the rear center seating position, the Fifth Circuit noted NHTSA’s conclusion that requiring

Type 2 belts “would yield small safety benefits and substantially greater costs, given the lower center seat occupancy rate.” Carden v. Gen. Motors Corp., 509 F.3d 227, 231 (5th Cir. 2007), abrogated on other grounds by Williamson v. Mazda Motor of Am., Inc., 562 U.S. 323 (2011). The lower occupancy rate in a rear center seat is a consideration that would affect any safety regulation applicable to that seating position, but Milburn’s experts did not discuss it. Further, as discussed, Milburn presented no evidence of other incidents involving injuries from a dual-detachable seat belt, which was already being utilized before Safety Standard 208 was amended to expressly authorize it. When asked about the significance of the absence of reported injuries, Gill dismissed it: What we don’t know is if, indeed, there have been no complaints, how many people actually got into the car, the Honda Odyssey, sat in the second or third row center seat and put the seat belt on incorrectly with no consequences because there wasn’t a crash. While that’s true, Gill did not discuss how either of those considerations—the lack of reports and any statistics indicating that the seat belts are nevertheless being misused—would have affected NHTSA’s risk–benefit analysis. And although Gill opined that the driver of an Odyssey would be unlikely to reliably leave the seat belt in the anchored position when the seat is upright, 20 she did not discuss

expectations regarding whether a driver would reliably attach the belt before transporting a passenger in the rear middle seat or how those expectations would affect NHTSA’s analysis. In enacting Section 82.008, the Legislature made a policy decision that manufacturers at risk of liability for injuries caused by an allegedly defective design are entitled to rely on a federal agency’s cogent determination that the pertinent risks associated with that design are not unreasonable. NHTSA made such a determination with respect to the safety standard authorizing dual-detachable seat-belt systems in some vehicles and seating positions. Absent a comprehensive review of the various factors and tradeoffs NHTSA considered in adopting that

safety standard, 21 as a general matter neither we nor a jury can deem a particular regulation “inadequate” to prevent an unreasonable risk of harm to the public as a whole. 22 That said, even in the absence of evidence that the standard was inadequate based on the information available to NHTSA when it was adopted, we do not foreclose the possibility that subsequent developments could demonstrate that the standard was no longer adequate to protect the public from unreasonable risks of injury at the time of the compliant product’s manufacture. For example, a material change in technology or a proliferation of new studies or data about risks and injuries associated with a compliant product could demonstrate the

standard’s inadequacy. 23 We leave further exploration of the parameters of such a showing for another day, noting only that no evidence regarding such developments was presented here. As discussed, the record demonstrates a complete absence of prior complaints, lawsuits, or recalls associated with the Odyssey’s seat-belt system. Nor is there evidence of complaints, lawsuits, or injuries related to any of the many other vehicles that used a roof-mounted detachable minibuckle system, either before or after the applicable regulation was adopted. In sum, no evidence was presented that NHTSA engaged in an improper or erroneous decision-making process in approving the regulation that authorized the detachable seat-belt system used in the Odyssey’s rear center seat, and no evidence was presented of post-approval developments that call the regulation’s adequacy into question. Accordingly, legally insufficient evidence supports the jury’s finding of regulatory inadequacy, and the presumption that Honda was “not liable” for the injuries allegedly caused by the Odyssey’s design was not rebutted as a matter of law. 24

III. Conclusion

For the foregoing reasons, we hold that the 2011 Odyssey’s design complied with mandatory federal safety standards that were applicable to the Odyssey at the time of manufacture and governed the product risk that allegedly caused harm, entitling Honda to a presumption of nonliability. We further hold that the presumption was not rebutted, as no evidence supports the jury’s finding that the federal safety standards failed to adequately protect the public from unreasonable risks of injury. Accordingly, we reverse the court of appeals’ judgment and render a take-nothing judgment for Honda.

Debra H. Lehrmann Justice

OPINION DELIVERED: June 28, 2024

Footnotes

[1] As the court of appeals described, conflicting evidence was presented at trial regarding whether Milburn was wearing her seat belt at all. 668 S.W.3d 6, 27–28 (Tex. App.—Dallas 2021). It was within the jury’s province to resolve that factual dispute.

[2] The police report concluded that Yusufzai ran a red light. There was also testimony that he entered the intersection as the light changed from yellow to red.

[3] Milburn’s parents were also named plaintiffs, but they later nonsuited their claims. The named defendants were American Honda Motor Co., Inc.; Honda Motor Co., Ltd. (which was later nonsuited); Uber Technologies, Inc.; Rasier, LLC; Uber USA, LLC; Arian Yusufzai; and Dawood Kohistani.

[4] Although Milburn alleged design, manufacturing, and marketing defects in her petition, only a design-defect claim was submitted to the jury.

[5] Gill described human factors engineering, which began as a discipline in the 1950s, as combining traditional design engineering and cognitive psychology to consider how people interact with the products and systems that engineers design, with the goal of making those systems safer.

[6] Both the detachable system used in the Odyssey’s rear middle seat and the ABTS design that Meyer endorsed are “Type 2” systems in that they have three points of attachment. The difference is the detachability feature.

[7] In the court of appeals, Honda argued the evidence was insufficient as to both negligence and proximate cause. 668 S.W.3d at 12–14. Honda does not challenge causation in this Court.

[8] Under Texas law, a products liability action can be based on negligence or strict liability, among other theories. See TEX. CIV. PRAC. & REM. CODE § 82.001(2); Hyundai Motor Co. v. Rodriguez, 995 S.W.2d 661, 664 (Tex. 1999). Only a negligence question was submitted to the jury in this case; however, as noted, the jury was instructed that for Honda to have been negligent, “there must have been a defect in the designing of the 2011 Honda Odyssey.” “Defective design” was defined in accordance with the applicable common-law and statutory requirements.

[9] Section 82.008(a) states in full: In a products liability action brought against a product manufacturer or seller, there is a rebuttable presumption that the product manufacturer or seller is not liable for any injury to a claimant caused by some aspect of the formulation, labeling, or design of a product if the product manufacturer or seller establishes that the product’s formula, labeling, or design complied with mandatory safety standards or regulations adopted and promulgated by the federal government, or an agency of the federal government, that were applicable to the product at the time of manufacture and that governed the product risk that allegedly caused harm. TEX. CIV. PRAC. & REM. CODE § 82.008(a).

[10] Notwithstanding the federal statute’s express statement that compliance with federal law does not exempt a manufacturer from liability at common law, Honda maintains that federal law preempts Milburn’s claims. Because we hold that Honda is not liable under Texas law, we need not reach its preemption argument.

[11] See Zorilla v. Aypco Constr. II, LLC, 469 S.W.3d 143, 156 (Tex. 2015) (noting that the defendant bears the burden of proof “to establish [an affirmative] defense and obtain the requisite jury findings”); Walters v. Cleveland Reg’l Med. Ctr., 307 S.W.3d 292, 295 (Tex. 2010) (discussing the plaintiff’s burden with regard to the open-courts exception to the statute-of-limitations defense).

[12] In interpreting a statute’s plain language, we construe the words and phrases chosen by the Legislature in context. Aleman v. Tex. Med. Bd., 573 S.W.3d 796, 802 (Tex. 2019).

[14] We do not address whether the presumption would apply had the regulation neither expressly precluded nor expressly authorized use of the detachable system.

[15] Because the presumption applied as a matter of law, submission of the question to the jury was improper. However, because the jury found that the presumption applied, the error was harmless. See TEX. R. APP. P. 61.1(a).

[16] Several other states have statutes creating a similar rebuttable presumption that a product’s design is not defective if it complied with applicable government standards. See COLO. REV. STAT. § 13-21-403; FLA. STAT. § 768.1256; IND. CODE § 34-20-5-1; KAN. STAT. § 60-3304; KY. REV. STAT. § 411.310; MICH. COMP. LAWS § 600.2946(4); N.D. CENT. CODE § 28-01.3-09; OKLA. STAT. tit. 76, § 57.2(A)–(B); TENN. CODE § 29-28-104(a); UTAH CODE § 78B-6-703(2). Of those, only two contain any direction about how to rebut the presumption: Oklahoma’s statute is identical to our Section 82.008, OKLA. STAT. tit. 76, § 57.2(A)–(B), while Kansas’s provides for rebuttal when “the claimant proves by a preponderance of the evidence that a reasonably prudent product seller could and would have taken additional precautions,” KAN. STAT. § 60-3304(a).

[17] By contrast, in Ohio, “[t]he extent to which [a product’s] design or formulation conformed to any public or private product standard that was in effect when the product left the control of its manufacturer” is one of the statutory factors to consider in determining whether a product is defective in the first instance. OHIO REV. CODE § 2307.75(B)(4).

[18] As noted, Honda unsuccessfully sought to exclude both Gill’s testimony and the usability studies on which she partially relied, arguing that Gill, a human factors engineer, was not qualified and that neither her testimony nor the studies were reliable under E.I. du Pont de Nemours & Co. v. Robinson, 923 S.W.2d 549 (Tex. 1995). Honda does not dispute that human factors engineering is a recognized scientific discipline. See Mihailovich v. Laatsch, 359 F.3d 892, 915 (7th Cir. 2004) (explaining that, “[b]roadly speaking, a human factors analysis focuses on the interaction between human behavior and the design of a machine[ or] product” and that an expert engaging in that analysis “will consider whether, in light of predictable human behavior, the design or condition of the subject item poses a potential hazard”); see also Douglas R. Richmond, Human Factors Experts in Personal Injury Litigation, 46 ARK. L. REV. 333, 337 (1993) (noting that “numerous corporations, industries, and organizations employ human factors experts,” who “testify in all types of personal injury litigation”). Indeed, one of Honda’s own witnesses was presented as a human factors expert. Rather, Honda notes Gill’s lack of experience with the automotive industry in general and seat belts in particular, along with the fact that Gill was unfamiliar with the federal safety regulations governing seat belts. Milburn responds that “human factors experts need not have expertise on the specific product or industry at issue because their knowledge, skill, experience, training, and education apply universally across all products and industries in which humans are involved.” Because we may assume without deciding that the trial court did not abuse its discretion in admitting Gill’s testimony, we need not address the parties’ disagreement about Gill’s qualifications and the reliability of her testimony.

[19] Contrary to the dissent’s accusation, post at 14–15 (Devine, J., dissenting), we do not simply “surmise[]” that NHTSA considered the risk of misuse in promulgating the standard. See 69 Fed. Reg. at 70908.

[20] Gill gave two principal reasons for her conclusion that a driver is unlikely to leave the seat belt in the anchored position. First, she stated that when the seat belt is anchored, it hangs down in the center of the vehicle and creates a visual obstruction. Our review of the record does not support that assertion. Rather, photographs of a “surrogate” 2011 Odyssey show the anchored belt suspended in front of the headrest of the rear seat on the right side of the vehicle, not obstructing the middle of the rear window. Second, Gill testified that because a tool is necessary to release the anchor, “that means you have to walk around the vehicle [and] get in through the side door” to do so, making it “inconvenient for a user who maybe wants to stow [and] unstow the seat frequently.” While the need to use a key-like object makes detaching the anchor a more deliberate act, it is unclear why that requirement makes the process more complicated logistically. And while Gill also criticized the owner’s manual as failing to adequately communicate the importance of reattaching the anchor when the seat is not stowed, Milburn did not pursue a marketing-defect claim, which is distinct from a defective-design claim. Am. Tobacco Co. v. Grinnell, 951 S.W.2d 420, 426 (Tex. 1997) (stating that “[a] product may be unreasonably dangerous because of a defect in marketing, design, or manufacturing”); Caterpillar, Inc. v. Shears, 911 S.W.2d 379, 382 (Tex. 1995) (holding that liability for a marketing defect attaches “if the lack of adequate warnings or instructions renders an otherwise adequate product unreasonably dangerous”). In any event, while Gill testified that vehicle owners would not reliably leave the belt attached when the seat was not stowed, she said nothing about whether they would reliably attach the belt before transporting a passenger in the rear middle seat.

[21] Honda argues that a “regulatory expert” is always necessary to make this showing. While we do not foreclose the possibility that rebutting the presumption would not require an expert, certainly a witness familiar with the regulatory process for approving safety standards relating to the type of product at issue would be integral in most cases in demonstrating the inadequacy of a federal regulation. Cf. Robinson v. Ethicon Inc., No. H-20-3760, 2021 WL 5054648, at *7 (S.D. Tex. Nov. 1, 2021) (evaluating a proposed safer alternative design and holding that the plaintiff could have created a fact issue regarding the feasibility of that design by presenting expert testimony that the FDA likely would have approved it).

[22] The dissent asserts that in analyzing the evidence of a regulation’s inadequacy to protect the public, we “unduly restrict[] the focus to the federal government’s decision-making.” Post at 9 (Devine, J., dissenting). Relatedly, the dissent opines that the considerations underlying the agency’s adoption of a safety standard have no bearing on whether the standard is inadequate to protect the public. Id. at 21. We cannot agree. The agency’s decision-making is what led to the regulation whose adequacy is being challenged. We fail to see how a jury could second-guess that decision without evaluating the process by which it was reached.

[23] As the concurrence notes, post at 4 (Blacklock, J., concurring), and contrary to the dissent’s assertion, post at 3–4 (Devine, J., dissenting), nothing in our opinion should be read to limit the grounds on which the presumption may be rebutted. Indeed, we broadly recognize that, among other possible showings, a plaintiff could demonstrate to the jury that a federal agency simply got it wrong.

[24] In light of this holding, we need not address Honda’s challenge to the sufficiency of the evidence supporting the jury’s negligent-design finding or its challenge to the trial court’s refusal to submit the Uber entities in the proportionate-responsibility question.