Paul R. and Cynthia J. Schilf sued Eli Lilly & Company and Quintiles Transna*948tional Corporation (“Lilly”). They alleged Lilly’s failure to warn and deceit caused the death of their son, Peter Raymond Schilf. Lilly moved for summary judgment, which the district court granted. Jurisdiction being proper under 28 U.S.C. § 1291, this court reverses and remands.
I.
This court states the facts most favorably to the Schilfs.
On November 26, 2004, Cynthia Schilf accompanied her sixteen-year-old son Peter to an appointment to discuss his depression with their family practitioner Dr. Richard G. Briggs. Peter complained of having various symptoms of depression since at least the prior summer. Dr. Briggs diagnosed Peter with depression and gave him samples of the antidepressant medication Cymbalta. These samples had been removed from the packaging and thus had no warning information. Dr. Briggs spoke with Cynthia and Peter Schilf about the risks of antidepressant treatment. Dr. Briggs recalls telling them that while there “may be an increased association with antidepressants and suicidal ideations and gestures,” “[n]o completed suicides occurred during the clinical trials,” and “Cymbalta was not specifically studied.” Dr. Briggs was referencing an FDA study and chose to prescribe Cymbalta in part because he believed it was less linked to suicide than another antidepressant evaluated in that study, Prozac.
In fact, there were five completed suicides in Lilly-sponsored clinical trials of Cymbalta, which was studied separately from the drugs Dr. Briggs referenced. Just over a month before Peter’s appointment, the FDA issued a Public Health Advisory telling the public that it directed manufacturers of antidepressants to include in their packaging a “black box” warning: “Antidepressants increase the risk of suicidal thinking and behavior (suicidality) in children and adolescents with major depressive disorder (MDD) and other psychiatric disorders.” A black box warning describes special problems, particularly those that may lead to death or serious injury, in a prominently-displayed box so that it is readily apparent. On the same day, the FDA issued a press release entitled, “FDA Launches a Multi-Pronged Strategy to Strengthen Safeguards for Children Treated with Antidepressant Medications.” In a separate letter to manufacturers, the FDA said, “A causal role for antidepressants in inducing suicidality has been established in pediatric patients.”
Within a day of receiving the medication samples, Peter, with the oversight of his father, searched the internet for Cymbalta and found Lilly’s website for it. Peter’s father testified that if he had noticed a warning about suicidality, he would not have allowed Peter to take the medication.
On December 24, 2004, Peter committed suicide. One month later, Lilly revised the Cymbalta literature to include the FDA-approved black box warning.
II.
This court reviews de novo a grant of summary judgment. Mason v. Corr. Med. Servs., Inc., 559 F.3d 880, 884 (8th Cir.2009). Summary judgment should be granted when — viewing the facts most favorably to the nonmoving party and giving that party the benefit of all reasonable inferences — the record shows that there is no genuine issue of material fact. See Fed.R.Civ.P. 56(c); Torgerson v. City of Rochester, 643 F.3d 1031, 1042 (8th Cir. 2011) (en banc). An issue is “genuine” if the evidence is sufficient to persuade a reasonable jury to return a verdict for the nonmoving party. Anderson v. Liberty Lobby, Inc., 477 U.S. 242, 248, 106 S.Ct. *9492505, 91 L.Ed.2d 202 (1986). “As to materiality, the substantive law will identify which facts are material.” Id. At summary judgment, the court’s function is not to weigh the evidence and determine the truth of the matter itself, but to determine whether there is a genuine issue for trial. Id. at 249, 106 S.Ct. 2505.
To survive summary judgment, the Schilfs must establish a genuine issue of material fact whether an adequate warning would have altered Dr. Briggs’ decision to prescribe Cymbalta. See In re Prempro Prods. Liab. Litig., 586 F.3d 547, 569 (8th Cir.2009); see also Ehlis v. Shire Richwood, Inc., 367 F.3d 1013, 1016 (8th Cir. 2004) (“The learned intermediary doctrine states that adequate warnings to prescribing physicians obviate the need for manufacturers of prescription products to warn ultimate consumers directly.”). The Schilfs rely on the heeding presumption— the presumption that a reasonable person (here, Dr. Briggs) would act according to an adequate warning. See Restatement (Second) of Torts § 402a cmt. j. It is likely that South Dakota would adopt this presumption. See McElhaney v. Eli Lilly, 739 F.2d 340, 340 (8th Cir.1984). The district court, relying on Thom v. Bristol-Myers Squibb Co., 353 F.3d 848, 856 (10th Cir.2003), ruled that Dr. Briggs’ behavior rebutted the presumption because he prescribed Cymbalta with knowledge of its risks.
A.
The district court first found that “a warning from [Lilly] would not have informed Dr. Briggs of anything he did not already know” about the risks of Cymbal-ta. The court wrote: “Dr. Briggs was aware of the same warnings that [the Schilfs] now say [Lilly] should have given to prescribing physicians such as Dr. Briggs.” See Ehlis, 367 F.3d at 1016 (8th Cir.2004) (noting that a plaintiff cannot prevail on a failure-to-warn claim if the prescribing physician knew the information that would have been in an adequate warning).
The Schilfs’ desired warnings were that (1) five suicides occurred during Cymbalta clinical trials (including one during a trial for a condition other than depression), (2) there is a causal role for Cymbalta in suicidality, and (3) the suicide risk in taking antidepressants is increased in children and adolescents.1 Dr. Briggs did not recall any suicides occurring in the clinical trials of Cymbalta, and he believed that no causal connection had been established between Cymbalta and suicidality.
Dr. Briggs was not aware of the five suicides that occurred during clinical trials *950of Cymbalta. He testified he reviewed the Cymbalta package insert, which cautions that “completed suicide” and “suicide attempt” were “infrequent adverse events.” At his deposition, he did not recall this warning and was interested in “see[ing] the information.” The insert did not state the number of suicides that occurred during the clinical trials. It also did not indicate in any way that the “completed suicide” and “suicide attempt” adverse events were more significant than the other events detailed in fine-print surrounding them. Dr. Briggs testified that before prescribing Cymbalta, he would have wanted to know the details about any suicide that occurred during its clinical trials.
Dr. Briggs testified that while he was not aware of a causal link between Cymbalta and suicide, he was aware of an association. The package insert clearly said that “a causal role for antidepressants in inducing suicidality has not been established,” but did caution that patients should “[n]evertheless” be observed for suicidality. Dr. Briggs’ understanding of the FDA’s determination was that there was an association between antidepressants and suicide but that “[t]hey weren’t saying the risk was there.”2 Asked what he understood in 2004 about the increased risk in taking antidepressants and suicide, he answered: “I think it was unknown at that time based on the information available.” Asked whether “there’s an increase in suicide risk when taking antidepressants,” Dr. Briggs answered, “I think it remains to be seen, yeah.” A warning that an adverse effect is “associated” with a medication — like the one Dr. Briggs gave — is not a warning that a causal connection exists. See Thom, 353 F.3d at 853-54 (collecting cases to that effect).
Dr. Briggs’ testimony is unclear about what he knew about the increased risk of suicide among children and adolescents who take Cymbalta. He testified that he had not seen any reports on the effects of Cymbalta on people under the age of 18 but could “remember seeing Cymbalta mentioned with pediatrics use.” No one asked Dr. Briggs the source or the contents of the “mention.”
The district court incorrectly stated that Dr. Briggs testified that he “read” the 2004 FDA press release and that the Schilfs admit he did so. To the contrary, Dr. Briggs testified that he was “aware” of the press release because it was “in the media” — “on the news, radio, televisions, all those.” He also did not answer a question whether he had read the FDA study that was the basis of the press release. He answered yes to a question whether he had “seen and become aware of’ the press release. No one elicited testimony about what Dr. Briggs thought the press release said.
There are genuine issues of material fact whether Dr. Briggs knew the suicide-related information that an adequate warning would have contained.3 See Ehlis, 367 F.3d at 1016 (8th Cir.2004).
*951B.
The district court found that Dr. Briggs would have prescribed Cymbalta to Peter even if he knew of its actual risks. See In re Prempro Prods. Liab. Litig., 586 F.3d at 569 (noting that a plaintiff cannot prevail on a failure-to-warn claim if an adequate warning would not have changed the prescribing physician’s decision).
Dr. Briggs’ deposition is unclear whether he would have still prescribed Cymbalta if given information about the clinical trial suicides or any causal role for Cymbalta in inducing suicidality. No one asked Dr. Briggs whether he would have prescribed Cymbalta to Peter if Lilly — through a black box warning, a letter, or its sales representatives — had informed him of the suicide-related information. Dr. Briggs stated that he has not made a decision not to prescribe Cymbalta given the information about suicides, but he could not remember whether he had prescribed it to an adolescent since Peter’s suicide. The district court relied heavily on Dr. Briggs’ statement that he still believed his prescription decision was appropriate, finding that he testified that “he would prescribe Cymbalta for Peter Schilf given adequate warnings.” The question Dr. Briggs was asked could be asking if Dr. Briggs’ decision was appropriate at the time it was made, given the information he had then. Read this way, the question would be consistent with those asked earlier in the deposition. When asked if there were anything he would have done differently, Dr. Briggs answered: “Not at the time. I did — I did exactly what I would have done.” The district court’s conclusion that Dr. Briggs would prescribe Cymbalta to Peter again with an adequate warning also fails because, even at the time of his deposition, Dr. Briggs was not aware of the suicide-related information.
The district court’s conclusion is also inconsistent with Dr. Briggs’ behavior. Dr. Briggs testified that he does not tell a patient every warning and precaution for a drug he prescribes, but he does discuss the ones that are “most likely to apply.” He discussed suicide with the Schilfs before prescribing Cymbalta — noting his belief that no completed suicides occurred in the trial. He also specifically chose Cymbalta because it was not included in the pooled study on which the FDA based its suicide warning.
Alternatively, Lilly argues that the 2004 Cymbalta warnings were adequate because Peter’s father would not have allowed him to take the medication if he had read the suicide-related statements in the warnings. There is no testimony, however, about how Peter’s father would have evaluated those statements in the original context. Considering the suicide-related statements contextually is critical because the 2004 Cymbalta warning also stated that “a causal role for antidepressants in inducing suicidality has not been established.”
Dr. Briggs’ testimony and behavior indicate that knowledge of the five suicides during the Cymbalta trials or of any causal role for Cymbalta in inducing suicidality may have changed his prescribing behavior. There are genuine issues of material fact whether an adequate warning would have changed Dr. Briggs’ decision to prescribe Cymbalta to Peter. See In re *952Prempro Prods. Liab. Litig., 586 F.3d at 569.
* * * * # *
The judgment of the district court is reversed, and the case remanded for further proceedings consistent with this opinion.
. The dissent states that the Schilfs did not state a desire for a causal role warning "[u]n-til this appeal.” This is incomplete. In addition to twice referencing causality in the amended complaint, the Schilfs presented the argument several other times. In their opposition to Lilly’s motion for summary judgment, the Schilfs stated that summary judgment was inappropriate because “Dr. Briggs had no idea that causality had been established for antidepressant induced pediatric suicidality.” In the district court’s hearing on the matter, counsel for the Schilfs stated no less than three times that "causality has been established.” The Schilfs’ counsel stated once even more plainly: "This pill can cause you to take your own life.” The Schilfs' expert on this point was also the subject of several motions by Lilly. In the opinion on general causation, the expert wrote that Cymbalta "can cause younger patients to become suicidal.” The district court also understood causality to be one thrust of the Schilfs’ arguments. In its order granting summary judgment, the district court wrote that it would presume the warnings inadequate for the purposes of the summary judgment determination because "[ajlthough suicidality is mentioned, the warnings provided by Lilly prior to the black box warnings do not convey a causal connection between taking Cymbalta and suicidality.”
. The dissent’s argument that Dr. Briggs understood the contents of the FDA study misses the mark. Even if Dr. Briggs was familiar with this information, he did not understand that the study also represented the risks of talcing antidepressants other than those specifically studied, including Cymbalta, and based his prescription decision on that belief. There is a genuine issue of material fact whether he knew the suicide-related risks of Cymbalta.
. Lilly asserts that federal pre-emption justifies summary judgment. Lilly argues that because the FDA did not approve its language asserting a causal role for Cymbalta in suicidality, the Schilfs’ failure-to-warn claim on that basis is pre-empted. (Without mentioning pre-emption, the dissent also intimates that the FDA’s determination not to allow the "causal role” language in antidepressant labeling resolves the warning issue.) Lilly's argument — restricted to the one sentence the *951FDA rejected — would not resolve this case. Additionally, Lilly does not argue it was barred from disseminating information through its sales representatives. Last, the Schilfs’ claim is not limited to the time after the FDA issued its Public Health Advisory, press release, and new warnings; they point to those events as evidence of Cymbalta's alleged misbranding.