Sykes, Circuit Judge.

Richard Leonard Carlson, Randal Wayne LeNeave, Attorneys, HUNEGS, LENEAVE & KVAS, P.A., Wayzata, MN, for Plaintiff-Appellant.

Chloe G. Pedersen, James D. Helenhouse, Attorney, FLETCHER & SIPPEL, Chicago, IL, for Defendant-Appellee.

Lawrence M. Mann, Attorney, ALPER & MANN, Bethesda, MD, Amicus Curiae for ACADEMY OF RAIL LABOR ATTORNEYS.

Sean Malone Sullivan, Attorney, DALEY MOHAN GROBLE, P.C., Chicago, IL, Amicus Curiae for ASSOCIATION OF AMERICAN RAILROADS.

Before Sykes, Barrett, and St. Eve, Circuit Judges.

Sykes, Circuit Judge.

Jeffery Kopplin brought two claims against the Wisconsin Central railroad under the Federal Employers' Liability Act ("FELA"), 45 U.S.C. §§ 51 et seq. Both rest on the same allegation: that Kopplin injured his elbow in an effort to operate a broken railroad switch while employed by Wisconsin Central. The district court entered summary judgment for the railroad in part because Kopplin could not prove that the broken switch caused his injury. While the parties raise several other questions, that alone is sufficient to affirm.

I. Background

Prior to his injury, Kopplin worked for Wisconsin Central as a train conductor. On January 24, 2014, he pulled a train into the Fond du Lac yard. To bring the train onto the correct track, Kopplin had to get out and "throw" a switch, which involves pulling a handle to correctly align the tracks. The weather that morning was severe, with below-freezing temperatures and 20- to 30-mile-per-hour winds. As a result ice and snow had built up inside the switch's mechanisms. Kopplin tried to remove the ice and snow with a simple broom—the only tool Wisconsin Central had provided—but after straining himself for several minutes, the switch would not budge.

Kopplin claims that this effort was the initial cause of a long-term elbow disability, though the evidence is less than clear. A video of the incident shows no immediate signs of injury. And Kopplin never mentioned any pain symptoms to his coworkers until two hours later—time in which he continued to perform other physical tasks.

After his physician diagnosed him with medial and lateral epicondylitis, Kopplin took time off work to receive treatment. Among other things, he received an effective pain-relief injection in February. By April the injury had fully healed. But in August the pain suddenly reemerged when Kopplin tried to drive a riding lawnmower one-handed while holding his son. After that his career as a conductor was effectively over.

Kopplin then brought two related FELA claims against Wisconsin Central, both alleging that the railroad was responsible for the broken switch and the injury it allegedly caused. The first is a run-of-the-mill negligence claim. The second is a negligence per se claim premised on Wisconsin Central's alleged failure to comply with 49 C.F.R. § 213.135, the regulation that sets national standards for switches. Kopplin's sole causation expert was Dr. Etienne Mejia, who testified by deposition that the pain-relief injection Kopplin received often provides only temporary relief, which could explain the pain's reemergence. However, Dr. Mejia conceded that he never investigated whether something other than the January 24 incident could have caused the initial injury. In fact, he testified that he knew so little about Kopplin's job that it would be mere speculation to say throwing a switch even could cause the elbow injury. Moreover, he admitted that he did not investigate whether Kopplin's other physical activities—say, riding a lawnmower in a dangerous fashion—could have caused the renewed elbow problems in August.

For two months after the deposition, Kopplin made no attempt to supplement Dr. Mejia's testimony. But after Wisconsin Central moved for summary judgment, Kopplin attached to his response a new affidavit by Dr. Mejia. The contents of that affidavit were markedly different than the deposition testimony. Dr. Mejia definitively stated that the January 24 incident caused the elbow injury, explaining that the nature of the injury was so clear that there was no need to even consider other potential causes. In the end, Kopplin's effort to bolster his causation evidence was in vain. The judge refused to consider the affidavit because it contradicted sworn deposition testimony. And without the affidavit, she found Dr. Mejia's testimony unreliable under Daubert v. Merrell Dow Pharmaceuticals, Inc. , 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993). As a result, Kopplin had no causation evidence at all.

The judge addressed several other questions, including the extent to which regulations promulgated under the Federal Railroad Safety Act define the standard of care for FELA actions and the extent to which 49 C.F.R. § 213.5(a) imposes a notice requirement for negligence per se claims. Because the failure to prove causation is fatal to both FELA claims, see Walden v. Ill. Cent. Gulf R.R. , 975 F.2d 361, 364 (7th Cir. 1992), we need not reach those issues here.

II. Discussion

We review a summary judgment de novo, asking whether the movant has shown "that there is no genuine dispute as to any material fact." Hansen v. Fincantieri Marine Grp., LLC , 763 F.3d 832, 836 (7th Cir. 2014) (quotation marks omitted). We review the exclusion of the affidavit "for abuse of discretion, giving the trial judge much deference." Buckner v. Sam's Club, Inc. , 75 F.3d 290, 292 (7th Cir. 1996). Finally, "we review de novo a district court's application of the Daubert framework. If the district court properly adhered to the Daubert framework, then we review its decision to exclude (or not to exclude) expert testimony for abuse of discretion." C.W. ex rel. Wood v. Textron, Inc. , 807 F.3d 827, 835 (7th Cir. 2015) (citations omitted).

We start with the admissibility of Dr. Mejia's affidavit. As the judge explained, a party may not "create an issue of fact by submitting an affidavit whose conclusions contradict prior deposition or other sworn testimony." Buckner , 75 F.3d at 292. The affidavit here contradicts Dr. Mejia's testimony in at least two ways. First, Dr. Mejia was asked at his deposition whether "there could be other various causes of this type of condition" besides the January 24 incident. He answered unequivocally, "Yes." But then in his affidavit, Dr. Mejia wrote that there was no need to consider other causes because "[t]he etiology and diagnosis [were] clear" that "the patient suffered from left traumatic medial epicondylitis as a result of the injury of January 24, 2014." That clearly contradicts his original statement that other causes could be at play.

Second, Dr. Mejia was asked at his deposition whether throwing a switch "seem[ed] like the kind of activity that could lead to the tendinosis," and he answered, "It would be speculation on my part...." That admission is squarely at odds with his affidavit's definitive conclusion that Kopplin injured his elbow throwing the switch. See id. at 293 (excluding a supplemental affidavit's detailed description of a fact when the affiant had disclaimed knowledge of that same fact at her deposition).

To be sure, we have carved out several exceptions to the general rule barring contradictory supplemental affidavits. None apply here. For instance, we've said that a party may offer an affidavit in response to a summary-judgment motion "to clarify ambiguous or confusing testimony." Bank of Ill. v. Allied Signal Safety Restraint Sys. , 75 F.3d 1162, 1171 (7th Cir. 1996). Yet nothing about Dr. Mejia's deposition testimony was ambiguous or confusing: without qualification, he said that other factors could have caused this condition. Similarly, while we have held that an affidavit may contradict sworn deposition testimony if "it is based on newly discovered evidence," id. at 1172, even Kopplin concedes that Dr. Mejia received all of the materials supporting his affidavit before his deposition. Finally, a new affidavit may be appropriate if the earlier testimony was "the result of a memory lapse." Cook v. O'Neill , 803 F.3d 296, 298 (7th Cir. 2015). Kopplin argues that this exception applies because Dr. Mejia did not have the full medical record at his fingertips during the deposition itself. But nothing in Dr. Mejia's responses indicates that he was struggling to recall what those records said. To the contrary, his responses were direct and honest admissions that he never considered certain issues at all.

Even if the affidavit were perfectly consistent with Dr. Mejia's prior statements, a larger problem remains. In essence the affidavit sets forth a brand new expert opinion on a topic beyond the scope of anything in Dr. Mejia's prior disclosures. In his original expert report, Dr. Mejia discussed Kopplin's treatment history and prognosis but never explained how the switch actually caused the disability. The issue surfaced for the first time—at least to any meaningful degree—in the affidavit itself. By then, the time had long passed to disclose a new report on a previously unexplored topic: Kopplin attached it as an exhibit to his summary-judgment response on June 27, 2017, months after the district court's December 30, 2016 deadline for Kopplin's expert reports. See FED. R. CIV. P. 26(a)(2)(D) ("A party must make [expert] disclosures at the time and in the sequence that the court orders.").

Without the affidavit the Daubert analysis is relatively straightforward. Under Daubert the court considers "whether the expert is proposing to testify to (1) scientific knowledge that (2) will assist the trier of fact to understand or determine a fact in issue." 509 U.S. at 592, 113 S.Ct. 2786. The ultimate question is whether the expert's approach is scientifically valid, which requires a careful examination of its "evidentiary relevance and reliability." Id. at 594–95, 113 S.Ct. 2786. The focus is on the expert's methodology, not his ultimate conclusions. See id. at 595, 113 S.Ct. 2786.

Both relevance and reliability are problems here. As to reliability, the judge identified a number of causation questions that Dr. Mejia conceded he never considered. Each concession significantly undermined the validity of his methods. The most troubling were his admissions that he never considered whether factors other than the switch could have caused the initial injury in January, nor whether other factors could have caused the renewed symptoms in August. The judge found this unacceptable, and that was not an abuse of discretion. See Brown v. Burlington N. Santa Fe Ry. Co. , 765 F.3d 765, 773–74 (7th Cir. 2014) (faulting an expert's differential etiology not just for failing to "rule in" the alleged cause but also for failing to "rule out" other potential causes).

As to relevance, only one of the opinions Dr. Mejia gave at his deposition is even probative of causation: his testimony that the pain may have resurfaced in August because the pain-relief injection Kopplin received often wears off. That is, Dr. Mejia had one theory for how the January injury could have had long-term effects. Even that is a partial theory because he admitted that he did not know whether throwing the switch could have caused the January injury in the first place. He testified that it would be "speculation" to say one way or another. Because Dr. Mejia's opinion is only marginally relevant, there is little reason to think that his testimony would be helpful to the trier of fact.

Kopplin has two final objections. First, he argues that even without the affidavit and despite all the problems with Dr. Mejia's deposition testimony, he should prevail because the injury's origin is obvious. It is true that we do not require expert testimony when causation is so clear that "a layperson can understand what caused the injury." Myers v. Ill. Cent. R.R. Co ., 629 F.3d 639, 643 (7th Cir. 2010). For example, a pedestrian hit by a truck would generally not need an expert to prove the cause of his broken leg. See id. But this case is much different. There are several steps between Kopplin's effort to fix the switch and his long-term disability, and none is clear. For instance, take the fact that the injury resurfaced when Kopplin attempted to hold his son while riding a lawnmower. To put it mildly, we are skeptical that the average layperson knows whether operating heavy machinery one-handed can contribute to medial and lateral epicondylitis. And because it would not be obvious to a layperson, expert testimony was indeed necessary.

Second, Kopplin insists that his claims should survive because Wisconsin Central's expert Dr. Jan Bax noted in a report that "Mr. Kopplin sustained a work-related strain to his left-elbow on January 14." But that one stray line does very little work. To start, the report never says that the broken switch caused the injury. It says only that the injury was "work-related," which could refer to a number of different things. The report also oddly says the injury began on January 14, ten days before Kopplin operated the broken switch. Moreover, Dr. Bax faces many of the same problems as Dr. Mejia—namely, that there is no evidence he considered whether other factors may have caused the injury. In fact, there is no evidence at all that Dr. Bax's testimony would have been admissible under Daubert . Perhaps the greatest flaw is that he never said a word about the reemergence of the injury in August. One way or another, Kopplin still needs admissible expert testimony that the January 24 incident caused a long-term disability. That report is not it.

As mentioned, causation is a necessary element of every FELA claim. See Walden , 975 F.2d at 364. So Kopplin's failure to present reliable expert testimony on that issue is fatal.

AFFIRMED .