On December 15, 1998, Mai Tran drove home from work in Orlando, Florida, in her 1983 Toyota Cressida. Her car crossed the center line and collided head-on with another vehicle. Tran's Cressida was equipped with a restraint system consisting of a manual lap belt and an automatic shoulder belt. The shoulder belt was a "passive" restraint. When the driver's door was opened, the belt slid along a motorized track towards the front of the car, allowing the driver to enter or exit. When the door closed, the belt slid back on its track into place, restraining the driver.

Tran was not wearing the manual lap belt during the accident. As a result of the collision, Tran suffered a spinal cord injury that rendered her quadriplegic. Tran and her husband1 sued Toyota Motor Corporation, the manufacturer of her vehicle, alleging negligence and strict liability in the manufacture, design, and testing of the Cressida, and that these defects were the cause of her injury. Specifically, she contended that the Cressida's automatic shoulder belt improperly fit shorter passengers like Tran. Tran was between 5'2" and 5'4" at the time of the accident. Tran asserted that the shoulder belt rode across her neck at the point of her injury. Tran claimed the belt instead should have been positioned to ride across her shoulder and sternum. Toyota's defense was that the passive restraint system was not defectively designed, that the shoulder belt did not cause Tran's spinal cord injury, that the belt could not have been across Tran's neck given the details of her injury, and that the cause of the injury was the inertial forces of the collision.

At the conclusion of an eight-day trial, the jury, finding that the vehicle's passive restraint system was not defective and that Toyota was not negligent, returned a verdict for Toyota. The district court entered a final judgment in accordance with the verdict, and Tran timely appealed.2 Tran presents three claims on appeal, and we address them in turn.

Tran contends that the court's instruction to the jury on strict liability design defect misstated the law. In a diversity case, the jury charge must accurately state the substantive law of the forum state. Wilson v. Bicycle South, 915 F.2d 1503, 1510 (11th Cir.1990). "[T]he manner of giving jury instructions is procedural rather than substantive," and thus our review is governed by federal law. Id. at 1511. "We review jury instructions de novo to determine whether they misstate the law or mislead the jury to the prejudice of the objecting party." Conroy v. Abraham Chevrolet-Tampa, Inc., 375 F.3d 1228, 1233 (11th Cir.2004) (quoting Palmer v. Bd. of Regents of the Univ. Sys. of Ga., 208 F.3d 969, 973 (11th Cir.2000)).

"Under Florida law, a strict product liability action requires the plaintiff to prove that (1) a product (2) produced by a manufacturer (3) was defective or created an unreasonably dangerous condition (4) that proximately caused (5) injury." McCorvey v. Baxter Healthcare Corp., 298 F.3d 1253, 1257 (11th Cir.2002) (citing Edward M. Chadbourne, Inc. v. Vaughn, 491 So. 2d 551, 553 (Fla.1986)).

Tran requested a jury instruction on design defect drawn from the Florida Standard Jury Instruction PL 5, which provides in relevant part that:

A product is unreasonably dangerous because of its design if the product fails to perform as safely as an ordinary consumer would expect when used as intended or in a manner reasonably foreseeable by the manufacturer or the risk of danger in the design outweighs the benefits.

Standard Jury Instructions Civil Cases, 778 So. 2d 264, 271 (Fla.2000). The court ruled that this instruction was "inappropriate" and declined to issue Tran's requested instruction. The court instead issued a jury instruction crafted from the Restatement (Third) of Torts: Product Liability § 2. The relevant portion read as follows:

A product is defective in design when the foreseeable risk of harm posed by the product could have been reduced or avoided by the adoption of a reasonable alternative design by the seller and the omission of the alternative design renders the product not reasonably safe to the user. This standard for judging whether a product is defective in design incorporates a reasonableness ("risk utility balancing") test. More specifically, the test is whether a reasonable alternative design would, at reasonable cost, have reduced the foreseeable risks of harm posed by the product and, if so, whether the omission of the alternative design by the seller rendered the product not reasonably safe. The balancing of risks and benefits in judging product design and marketing must be done in light of the knowledge of risks and risk-avoidance techniques reasonably attainable at the time of distribution.

A broad range of factors may be considered in determining whether an alternative design is reasonable and whether its omission renders a product not reasonably safe. The factors include, among others, the magnitude and probability of the foreseeable risks of harm, the instructions and warnings accompanying the product, and the nature and strength of consumer expectations regarding the product, including expectations arising from product portrayal and marketing. The relative advantages and disadvantages of the product as designed and as it alternatively could have been designed may also be considered. Thus, the likely effects of the alternative design on product costs; the effects of the alternative design on product longevity, maintenance, repair, and esthetics; and the range of consumer choice among products are factors that may be taken into account. The relevance of these factors and other factors vary, depending on the facts as you find them. Moreover, the factors interact with one another. For example, evidence of the magnitude and probability of foreseeable harm may be offset by evidence that the proposed alternative design would reduce the efficiency and the utility of the product. On the other hand, evidence that a proposed alternative design would increase production costs may be offset by evidence that the product portrayal and marketing created substantial expectations of performance or safety, thus increasing the probability of foreseeable harm. Depending on the mix of these factors, a number of variations in the design of a given product may be relevant to determining whether a product is defective. In sum, the rule that a product is defective in design if the foreseeable risks of harm could have been reduced by a reasonable alternative design is based on the commonsense notion that liability for harm caused by product designs should attach only when harm is reasonably preventable.

R. 187 at 9-10.

While the court's instruction did mention "the nature and strength of consumer expectations" as one factor in the risk-utility test it directed the jury to apply, it did not, as Tran requested, provide for a consumer expectation test as an independent basis for liability. The court's instruction was an erroneous statement of Florida law.

A few months after Tran's trial, the Florida Fifth District Court of Appeal decided Force v. Ford Motor Co., 879 So. 2d 103 (Fla.Dist.Ct.App.2004). In that case, plaintiff Force alleged that he was injured in an automobile collision when his seatbelt failed to restrain him. He sought a jury instruction, drawn from the standard Florida jury instruction, that provided both the consumer expectations test and the risk-utility test. The trial court agreed with the defendants that only the risk-utility test applied, and instructed the jury accordingly. Id. at 105

The District Court of Appeal reversed. First, the court held that every case to have addressed the issue confirmed the applicability of the consumer expectations test under Florida products liability law, "at least for some products." Id. at 108. Then, the court addressed the defendants' contention that the consumer expectations test was inappropriate in complex product cases, where the jury "simply has no idea how [the product] should perform." Id. at 109 (internal quotation omitted). Surveying cases, the court ultimately concluded that seatbelts were not such a product, and that consumers were capable of forming expectations about their performance. Id. at 109-10.

Force controls our decision on this issue. Toyota attempts to distinguish Force by noting that here the district court included consumer expectations as a factor in the risk-utility analysis, whereas the trial court in Force did not mention consumer expectations at all. However, Florida law recognizes consumer expectations as "one of the independent standards to be applied in at least some Florida products liability cases." Id. at 108 (emphasis added).

We emphasize that we do not hold that the consumer expectations test jury instruction is required in all product liability cases. We merely hold, like the court in Force, that the instruction is proper as an independent basis for liability under Florida law when the product in question is one about which an ordinary consumer could form expectations. Under Florida law, seatbelts are such a product. The district court did not have the benefit of the Force court's analysis, but in light of that case we must conclude that the court erred in not instructing the jury that it could find for Tran under a consumer expectations theory.

Our review of a district court's jury instruction is deferential, but we will reverse a district court because of an erroneous instruction if we are "left with a substantial and ineradicable doubt as to whether the jury was properly guided in its deliberations." Carter v. DecisionOne Corp., 122 F.3d 997, 1005 (11th Cir.1997) (quoting Johnson v. Bryant, 671 F.2d 1276, 1280 (11th Cir.1982)). Tran was prejudiced by the erroneous instruction because the jury was not aware that consumer expectations was an adequate and independent basis for liability, rather than merely one factor among many in the risk-utility balance. Toyota's contention that Tran could have argued her consumer expectations theory to the jury is misplaced because, under the instruction the court issued, she was unable to argue that unmet consumer expectations were an independently sufficient basis for liability. Indeed, the jury could not have found Toyota liable even if consumer expectations were unmet, if it determined that other factors in the risk-utility balancing test outweighed that factor. In sum, we cannot say that the jury instruction "sufficiently instructed the jury so that the jurors understood the issues and were not misled." Carter, 122 F.3d at 1005 (quoting Wilkinson v. Carnival Cruise Lines, Inc., 920 F.2d 1560, 1569 (11th Cir.1991)). Accordingly, we conclude that the district court erred in instructing the jury and that the error requires a remand for a new trial.

Tran next argues that the district court erred in excluding the testimony of Dr. Charles Clark. Tran proposed to present Dr. Clark as an expert witness to testify about Tran's neck injury from a "micro perspective." Appellants' Brief at 18. Tran was able to offer the testimony of the treating physician, Dr. Michael Cheatham. In addition, Tran presented the expert testimony of Dr. Joseph Burton, whose testimony Tran characterizes as encompassing a "macro perspective" on the collision and Tran's injury. Id. at 17.

After Dr. Burton testified, Toyota objected to Dr. Clark's testimony as cumulative. See Fed.R.Evid. 403. The court examined Dr. Clark's deposition and expert witness report, and extensively examined Dr. Clark's qualifications. The court concluded that Dr. Clark's opinions, and the bases for these opinions, were the same as those of Dr. Burton. The court sustained Toyota's objection and excluded Dr. Clark from testifying.

"The district court has broad discretion to determine the admissibility of evidence, and we will not disturb the court's judgment absent a clear abuse of discretion." United States v. McLean, 138 F.3d 1398, 1403 (11th Cir.1998). "An abuse of discretion can occur where the district court applies the wrong law, follows the wrong procedure, bases its decision on clearly erroneous facts, or commits a clear error in judgment." United States v. Brown, 415 F.3d 1257, 1266 (11th Cir.2005) (discussing admissibility of expert testimony) (citing McClain v. Metabolife Intern., Inc., 401 F.3d 1233, 1238 (11th Cir.2005)).

Tran relies on Johnson v. United States, 780 F.2d 902 (11th Cir.1986). In that case, we held that the exclusion of a third expert witness as cumulative was an abuse of the district court's discretion. Id. at 906. We noted that the excluded expert's "analysis was somewhat different," his testimony was "more comprehensive," and the witness "had different, and arguably better qualifications than the other experts." Id. The same is not true here.

Drs. Burton and Clark relied on the same medical evidence in forming their opinions. In addition to testimony about the collision and inertial forces, Dr. Burton testified about Tran's neck injury and the impact of the seat belt. These "micro" issues are the same as those about which Dr. Clark would have testified. Moreover, the treating physician, Dr. Cheatham, testified about Tran's injury as well. In sum, Tran presented extensive testimony to the jury suggesting that the seat belt caused her injury, and it is not at all clear that Dr. Clark would have added any different information that Tran could not have presented through Drs. Burton and Cheatham. Additionally, unlike in Johnson, Dr. Clark's qualifications are not significantly greater than the other doctors'. Finally, Tran could have called Dr. Clark when given an opportunity for rebuttal, but did not call him at that time. While we note that in Johnson we held that a third expert witness was not cumulative, whereas Dr. Clark was excluded from testifying as a second expert witness, the mere number of witnesses is not conclusive when these other factors support the district court's decision.

District courts have broad authority over the management of trials. Id. at 905. Part of this authority is the power to exclude cumulative testimony. Fed.R.Evid. 403; Johnson, 780 F.2d at 905. "Inherent in this [abuse of discretion] standard is the firm recognition that there are difficult evidentiary rulings that turn on matters uniquely within the purview of the district court, which has first-hand access to documentary evidence and is physically proximate to testifying witnesses and the jury." United States v. Jernigan, 341 F.3d 1273, 1285 (11th Cir.2003).

"[U]nder the abuse of discretion standard of review there will be occasions in which we affirm the district court even though we would have gone the other way had it been our call." Rasbury v. Internal Revenue Serv. (In re Rasbury), 24 F.3d 159, 168 (11th Cir.1994). On this record, we cannot say that the court would have abused its discretion had it allowed Dr. Clark to testify. The testimony likely would not have unduly prolonged the trial, Dr. Clark's practice and experience was somewhat different from that of the other doctors, and Tran might have presented her evidence differently had she known earlier that Dr. Clark would be excluded. Given our deferential standard of review, however, we cannot say that the district court's decision fell outside its permissible "range of choice." United States v. Kelly, 888 F.2d 732, 745 (11th Cir.1989). Therefore, we affirm the court's order excluding Dr. Clark's testimony as cumulative.

Finally, Tran argues that the district court erred in admitting into evidence a study of Cressida accidents performed by Dr. Donald Huelke in the 1980s ("the Toyota study" or "the study"). "[T]his court will afford great deference to the decisions of the district court with regard to evidentiary matters. We will only reverse a district court's ruling concerning the admissibility of evidence where the appellant can show that the judge abused his broad discretion and that the decision affected the substantial rights of the complaining party." Heath v. Suzuki Motor Corp., 126 F.3d 1391, 1395 (11th Cir.1997) (quoting Wood v. Morbark Indus., Inc., 70 F.3d 1201, 1206 (11th Cir.1995)).

The Toyota study was an examination of other accidents involving the Cressida's restraint system. Toyota introduced the study to demonstrate the system's overall effectiveness in a wide array of accidents. Tran asserts that the Toyota study should not have been admitted because Toyota did not prove that the accidents in the study were substantially similar to hers.

The doctrine of substantial similarity applies when one party seeks to admit prior accidents or occurrences involving the opposing party, in order to show, for example notice, magnitude of the danger involved, the [party's] ability to correct a known defect, the lack of safety for intended uses, strength of a product, the standard of care, and causation. In order to limit the substantial prejudice that might inure to a party should these past occurrences or accidents be admitted into evidence, courts have developed limitations governing the admissibility of such evidence, including the "substantial similarity doctrine." This doctrine applies to protect parties against the admission of unfairly prejudicial evidence, evidence which, because it is not substantially similar to the accident or incident at issue, is apt to confuse or mislead the jury.

Heath, 126 F.3d at 1396 (quoting Jones v. Otis Elevator Co., 861 F.2d 655, 661 (11th Cir.1988)) (internal citation and footnotes omitted; alteration in original).

The substantial similarity doctrine does not apply to situations, like this one, where the evidence is "pointedly dissimilar" and "not offered to reenact the accident." Heath, 126 F.3d at 1396-97. The evidence may have had some prejudicial effect on Tran's case by showing that the Cressida's restraint system generally performed well in a variety of accidents (a point that Tran's expert conceded). But the district court did not abuse its broad discretion in concluding that this prejudice did not outweigh the probative value of the study as part of Toyota's case that its restraint system was not defectively designed. Id. Therefore, we affirm the district court's ruling on Tran's objection to the admission of the Toyota study.

For the reasons stated above, we affirm the district court's decision to exclude the testimony of Dr. Clark, as well as its decision to admit the Toyota study into evidence. However, we conclude that the jury instruction regarding design defect products liability was erroneous. Accordingly, we vacate the district court's order and remand the proceeding for a new trial consistent with this opinion.

AFFIRMED IN PART; REVERSED AND REMANDED IN PART.