Wednesday, November 10, 2010

Head to head: literal falsity helmet claims rejected

Riddell, Inc. v. Schutt Sports, Inc., --- F.Supp.2d ----, 2010 WL 2803461 (W.D.Wis.)

Background: According to the New York Times, “[r]ecent engineering advances made by Riddell, Schutt, Adams and other manufacturers have undoubtedly improved the performance of the football helmet, which from its leather roots has always symbolized football’s duality of valor and violence. But helmets communicate a level of protection that they do not provide, experts said, in part because of lax industry standards and practices.”

Riddell sued Schutt for patent infringement, false advertising, and trade libel/product disparagement related to football helmets and faceguards. Together the two make up 90% of the football helmet marketplace, with Riddell having slightly more in dollar sales. Schutt brought the “usual counterclaims”: noninfringement, invalidity and inequitable conduct, as well as counterclaims for false advertising. Riddell moved for partial summary judgment on Schutt’s false advertising/deceptive trade practices claims.

Beginning in 2002, the University of Pittsburgh Medical Center conducted a study to compare the concussion rates and recovery times for athletes wearing Riddell’s Revolution helmet compared to those wearing traditional helmets. Riddell underwrote the study, providing salary support for two lead authors of the study, Micky Collins and Mark R. Lovell. A third author, Mark Ide, was a Riddell employee. Three other authors are co-owners of ImPACT, “a company that manufactures and distributes computerized neurocognitive testing software.” In 2003, ImPACT and Riddell agreed that ImPACT would not sell “in conflict” with Riddell and would pay Riddell for any ImPACT sale completed through a Riddell-initiated contact. The study used ImPACT concussion management software.

The research proposal for the study included directional hypotheses, which means that the researchers had some expectations about what would happen. Here, they hypothesized that athletes wearing the Riddell helmet would have significantly fewer incidences of cerebral concussion and fewer and shorter markers of concussion severity and dysfunction relative to the controls. The study was not randomized—the high school students studied were allowed to choose Riddell’s Revolution or a traditional helmet.

From 2002-2004, the study tracked more than 2000 high school players, slightly more than half wearing the Revolution. The Revolution helmets were reused after the first year. The traditional helmets were from school inventories and not necessarily new. All non-new helmets were refurbished and recertified each year to industry standards.

In a 2002 analysis of pilot data, the authors found nearly identical concussion rates. The 2003 data showed that the difference “approached” statistical significance. The 2004 data showed 2207 participants, 1173 with the Revolution helmet. This report showed 5.29% wearing the Revolution helmet with diagnoses of cerebral concussions, while 7.16% wearing traditional helmets sustained concussions. According to the authors, this difference “approached but did not reach statistical significance.” The final three-year study, however, considered only 2141 participants, 1173 fitted with the Revolution. Excluding 66 participants produced 5.3% (that is, rounding up from 5.29%) and 7.6% concussion rates respectively, which the authors described as “statistically significant,” and demonstrating “a trend toward a lowered incidence of concussion,” but cautioned that the “limited size sample precludes a more conclusive statement of findings at this time.”

The researchers submitted the study for publication in Neurosurgery, a neurology journal. It was subject to substantial criticism during the peer review process for cherry-picking data and for the “substantial conflict of interest” faced by the researchers. Among other things, reviewers pointed out that helmet selection was neither randomized or controlled, and that younger people tended to use the older helmet, which is significant because younger people may be more susceptible to concussion. Likewise, another reviewer found that disparities in the age of the helmets used “invalidate[] any comparison,” since the Riddell helmets were new, which is well recognized to affect performance, but the other helmets were of unknown age.

Regardless, the results were published in Neurosurgery in 2006. The article “compared the concussion rates of the 1,173 participants fitted with the Revolution helmet (5.3%), with the rates of the 968 participants fitted with traditional helmets (7.6%), concluding that there was a 2.3% decreased absolute risk for sustaining a concussion and a 31% decreased relative risk for athletes wearing the Revolution helmet. The study also subtracted “actual and estimated” (?) athletes with previous concussions, producing a 3.7% concussion rate for 1109 players wearing Revolution helmets and 6.2% for 916 players wearing traditional helmets, about 41% decreased relative risk. Comments were appended, including one opining that the study “suffers from a serious, if not fatal methodological flaw” raising doubt about the “significance of the data” because the age of the traditional helmets was not known. Another commenter stated that the study “has several limitations,” including the fact that “[h]elmet selection was neither randomized, nor controlled” and that “each of the authors has a business relationship with either” ImPACT or Riddell.

Riddell has, since then, relied heavily on the study in its ads, using the following or similar language: “Research shows a 31% reduction in concussions in players wearing Riddell Revolution Helmets.” Some ads made an explicit comparison to traditional football helmets; others claimed reduced risk “up to 41%,” and others added that this was only for players who hadn’t previously suffered a concussion. Although the study tested only the Revolution helmets, and although there are a number of design and material differences between the Revolution and other models in the Revolution family, Riddell used the phrase in many advertisements for other helmets in the Revolution family. (Differences include: different face guard mechanisms, materials used in the outer shell and liner, locking rear pads, mandible designs, and padding structure. These differences could affect the helmet’s ability to prevent concussions.) One letter falsely attributed the 31% finding to Riddell Revolution Youth helmets, a “mistake” identified and changed after the mailing went out. PowerPoints for sales reps included similar misstatements about which helmets had been studied, such as “IQ HITS, Speed, Revolution IQ, Revolution, Speed Youth, IQ Youth, Revolution Youth and Revolution Little Pro helmets all offered as ‘exclusive benefit from Riddell’ that ‘reduce[ ] chances of a concussion by 31%/41%.’”

Riddell has had great market success since the concussion study was published. Riddell gets a $50 premium for the Revolution helmet, which Riddell attributes to the helmet’s technology and the concussion study.

The challenged ads all refer to the results of the concussion study, also known as establishment claims. Establishment claims can be shown to be false by showing that the cited test or study doesn’t prove the claim. Some courts have also stated that an establishment claim can be literally false if the test or study does prove the proposition, but wasn’t sufficiently reliable to permit one to conclude with reasonable certainty that it established the proposition for which it was cited. The court here expressed doubt about that, and held that it was not bound to follow the cases explaining that standard. Initially, I think that the court is a bit confused about what “prove” means, since it appears to hold that “prove” means something other than “establish the proposition for which it was cited to a reasonable degree of certainty.” I really have no idea what that other meaning would be.

The court, however, found that this standard leads to a strained reading of “literally false” (a term, I might point out, which nowhere appears in the Lanham Act; judicial glosses are purely a matter of judicial determination): “a court's determination that a test is ‘unreliable’ leads to a conclusion that a statement in the form ‘test shows x’ is literally false even if the test really does show x.” The court speculated that this might be an application of falsity by necessary implication, though the Seventh Circuit has also not adopted that doctrine. But necessary implication exists only if a consumer will “unavoidably receive a false message" from a statement. “The statement that a ‘test shows x’ (or even that it ‘proves’ x) merely suggests that the test is reliable; it does not send an ‘unavoidable’ message that it is.”

Comment: This is ridiculous. There is no reason other than wanting you to rely on the test that an advertiser would present to you the results of that test; the necessary implication is that the test is reliable. The court’s interpretation of “unavoidably” would strip the concept of any meaning, since we could always posit a consumer who would read an ad in an extreme and unlikely way. In addition, the court is missing the point: establishment claims have special force to consumers, which is why advertisers like to say “tests prove X” instead of “X.” If the challenger proves that the test isn’t sufficiently reliable to establish X, then the test doesn’t prove or show X, and the claim that it does is literally false. That is, the court treats the issue as if it were “is X true?” when there is also a claim that “tests prove X” in the ad, and that claim is, independently, falsifiable even without falsifying X.

However, the court took the position that, “if a cited test is unreliable, statements that the ‘test proves x’ are merely deceptive or misleading, not ‘literally false.’” But because this was at odds with, well, all the other cases, and because the parties didn’t dispute the standard, the court said that it would assume that the “not sufficiently reliable” test applied. The court’s reservations about the standard “inform[ed]” its reading, though. What counts as sufficient reliability? The court thought that it “appears to be tied to whether the methods and findings of the cited study are acceptable to the relevant scientific community.” (So basically, I wonder how this court decides Daubert motions; it did ultimately drop a reference to Daubert.) In the cases the court examined that found unreliability, “members of the scientific community, a government agency responsible for monitoring the field or the authors themselves identified a flaw in the study that undermined the study's conclusion.”

The challenged establishment claim was Riddell’s ads touting the Neurosurgery study's findings that the risk of sustaining a concussion for Revolution helmet wearers was between 31% and 41% (for those without a previous concussion) less than the risk for traditional helmet wearers.

Schutt identified the following concerns: (1) conflicts of interest based on Riddell's funding and participation in the study and Riddell's relationship with ImPACT (query whether Riddell complies with the new FTC Guides on endorsements); (2) the non-random sampling method; (3) the lack of information about the age and condition of traditional helmets used; (4) discrepancies between the number of participants in earlier reports and in the final report; (5) the "preliminary" nature of the study; and (6) the fact that the concussion data gathered in the years following the study were not included as part of the study.

These issues “give reasons to doubt the results of the study, but they do not show that the study was unreliable” to the relevant scientific community. Schutt relied too heavily on criticisms during the peer review process, including the criticism that the three-year cutoff of the study may have been simply an attempt to reach statistical significance. The study was published over these objections. Even a study that passes peer review may be unreliable, but approval for publication is some evidence of reliability. “[A] party seeking to attack the reliability of a peer-reviewed article should do more than recite objections made by reviewers.” Schutt added nothing to suggest that the reviewers were not in a position to assess the reliability of the study. (Which they did!  Why can't the court weigh their objections itself?) The fact that more traditional helmet wearers were included in earlier stages, and that additional concussion data were gathered in subsequent years, was also not important, because Schutt didn’t offer any evidence that the decision to drop those subjects/years was scientifically inappropriate or that the extra data would have been less favorable. Thus, Schutt failed to show that the results or methods of the concussion study were unreliable. (Compare the recent Kaiser v. Pfizer case, where the court found that data manipulation by a study sponsor was intentionally fraudulent, because even if you find positive results once you massage the data, it's the massaging that's scientifically inappropriate.)

Schutt also argued that there were significant differences between what the study showed and what the ads claimed it showed: (1) some ads touted helmets other than the Revolution helmets tested in the study; (2) certain ads made claims directed to players other than high school players, the age group tested in the study; and (3) certain ads compare Riddell's helmets to Schutt's high-end helmets, but the study compared Riddell's helmets only with "traditional helmets."

First, the court found that Schutt glossed over the specific language in the ads, focusing instead on context. The specific language "research shows a 31-41% reduction in concussions in players wearing Riddell Revolution helmets" accurately reflected the study, even when “nestled into” ads for helmets other than the Revolution. This might be misleading or deceptive, but there was no consumer survey evidence.

Riddell’s statement that the study showed concussion reduction in “Riddell Revolution Youth” helmets were, however, literally false. Riddell corrected this statement in its files. Apparently literally false statements in PowerPoint presentations to Riddell’s sales force weren’t ads, and Schutt lacked evidence that the sales force made those statements to the public.

Statements that the "technology" used in Riddell's Revolution line of helmets has been "shown to reduce the incidence of concussion" were trickier. It was fair to say that the study showed that the technology reduced concussion, since it identified the padding structure and shell shape as what distinguished it from traditional helmets. But it was uncertain whether the technology used in the family or line of helmets has been shown to reduce concussion. The different helmets have different design features, including differences in shell, face guard, padding, lining and other features, and even the current Revolution helmet is different from the original production.

But this uncertainty was Schutt’s problem: it had the burden to show literal falsity by proving relevant design differences between the tested helmets and other helmets in the family. Schutt’s evidence of difference didn’t show that the differences were relevant to concussion testing. It had general evidence that design differences can change concussion results, but not reason to think that the actual design differences in this case would be expected to do so.

Likewise, Schutt’s argument that ads directed at groups outside high school students were literally false because the study wasn’t designed to apply to those age groups failed. Just being overly broad isn’t literal falsity:
In this case, the challenged advertisements state simply that the concussion study showed decreased concussion rates, without providing the limitation that the study applied only to high school students. … The absence of limiting language does not require the audience to conclude that there must have been no limitations in the study. (There are always limitations.) … The context does not require a conclusion that the study must have tested the segment of the population that forms the audience. At most, such a context suggests that the study had a broad study group or that the results can be applied to the audience. This means the advertisements may be misleading, not that they are literally false.

Schutt also argued that, despite the study’s failure to test Schutt’s high-end helmets, Riddell used the study to compare those helmets. This is mostly based on PowerPoint presentations to sales reps, which aren’t ads and don’t support a claim that representatives were making false statements. The PowerPoints may have guided sales reps, but they weren’t “scripts.” (Is it really plausible that sales reps were downplaying claims they heard internally? Still, I can see where getting some customer/sales rep testimony would have helped.)

Anyway, the presentations weren’t literally false. They included statements such as: "With RCRT, Speed reduces the chances of concussion by 31%/41%. XP doesn't stack up." The court (in conflict with FDA/FTC standards for making comparative statements of this sort, which prefer or require head-to-head testing, no pun intended) found that literal falsity would have required language to the effect that the study actually tested Speed v. XP or otherwise tested high-end Schutt helmets. “Although the presentations and internal discussions may suggest that sales representatives were trained to mislead, they fail to suggest that representatives were trained to make literally false statements.”

Schutt also challenged comparative ads that didn’t provide the point of comparison, traditional helmets: general superiority claims of “reducing concussion” by 31%, along with “The Riddell Revolution helmet is the standard against which all football helmets are measured--shown in published research to reduce the risk of concussion by nearly a third."

Assuming that falsity by necessary implication is a valid claim in the Seventh Circuit, Schutt still lost. Schutt argued that, because Schutt and Riddell are the two principal competitors in the market, failure to disclose the point of comparison necessarily implied testing against all product offerings, including Schutt’s high-end helmets. The court disagreed. Case law supports the proposition that a superiority claim may necessarily implicate a principal competitor, even without naming the competitor. However, that doesn’t necessarily imply comparison with a particular product of the competitor’s—Schutt didn’t contest Riddell’s claims of superiority over Schutt’s standard helmets. The lack of a qualifer could mislead, but is not necessarily false.

The “standard against which all football helmets are measured” ad required separate analysis, because it did suggest that the comparison was all helmets. But “the language is odd enough to weigh against reading the advertisement that way.” The ad didn’t say that all helmets had been tested against the Revolution, and this “subtle difference” in language made it ambiguous. “In other words, what makes it such a "leader," under one reading, is the simple fact that it has been tested and other high-end helmets have not.” Thus, there was no falsity by necessary implication.

Schutt only showed one instance of literal falsity, and it couldn’t show injury from those ads. Its evidence of injury was “scant,” coming down to the fact that Riddell charges a $50 premium for helmets using the Revolution technology and that it has “converted” high school and college players. But Schutt didn’t submit evidence of lost sales or market share, and there was no evidence that players switched from Schutt helmets to the falsely advertised Riddell Youth helmets. Moreover, the falsity was merely technical—had the ad simply asserted that the results of the study could be applied to the Youth helmet as part of the same family, “the reader would have gotten the same message and there would have been no false advertisement problem.” Under these circumstances, it would be unjust to award damages, and there was no reason to grant an injunction, since Riddell removed the reference and there was no suggestion that they intend to reuse that language. (Different courts take different approaches to voluntary discontinuance; perceived good faith affects whether injunctive relief is nonetheless deemed appropriate.)

The court concluded with a caution: Schutt lost “not because Riddell's advertisements were particularly open and honest, but rather because Schutt tried to take the easiest evidentiary path to success: literal falsity.”

Schutt's state-law Deceptive Trade Practices claim failed because it required Schutt to show "pecuniary loss, which it could not do, and also because the law does not provide a cause of action for misrepresentations made to non-parties. It was “not designed to protect product manufacturers from the deceptive acts of their competitors.”

No comments: