BLACK, Circuit Judge:
This appeal arises from a wrongful death action brought by the natural mother and father of Jason Christopher. Jason suffered from classic hemophilia, a congenital blood clotting disorder. As part of his treatment, Jason used a number of different blood products, including one known generically as Factor VIII concentrate and sold by Appellant Armour Pharmaceutical (Armour) under the trade name Faetorate. Jason died of complications from AIDS in February of 1992. Plaintiffs claimed that Jason was infected with the HIV virus by an infusion of Armour’s concentrate in late 1983, as a result of Armour’s negligent failure to warn Jason’s treating physician that its product might transmit AIDS. A jury found in favor of plaintiffs and awarded them damages of $2,007,256.13.
We address whether plaintiffs’ evidence established sufficient causation for the jury to find in their favor and whether the district court erroneously instructed the jury on Armour’s learned intermediary defense.1 We affirm in part and reverse and remand in part, holding that plaintiffs’ evidence established sufficient causation to render a verdict against Armour, but that the district court erred in its jury instructions on the learned intermediary rule by constructing an improper legal standard not supported by the record.
I. PROCEDURAL BACKGROUND
Jason Christopher’s parents brought a wrongful death action against Armour, claiming that Armour negligently failed to warn Jason’s prescribing physician that Faetorate might transmit the HIV virus. Faetorate is a biologic, licensed by the Federal Drug Administration (FDA), that contains Factor VIII clotting protein necessary for the treatment of hemophilia. The FDA must approve this biologic for sale. 21 C.F.R. § 601.2(a) (1983). It is available only by prescription through a physician. The written prescribing information packaged with each vial of Faetorate must include FDA-approved warnings with respect to potential hazards of use. 21 C.F.R. § 201.57(e) (1983); see 21 C.F.R. § 601.2(a) (1983).2 Armour cannot change the warnings without the express approval of the FDA. 21 C.F.R. § 601.12(b) (1983).
Prior to January 1984, Armour’s package labeling did not contain a warning with respect to AIDS. On September 30, 1983, [1187]*1187Armour requested a labeling change for Fac-torate to include the following AIDS warning:
The possibility exists that Acquired Immune Deficiency Syndrome, (AIDS), an immunologic disorder with extremely severe consequences, may be transmitted by blood, blood products and blood derivatives, including clotting factors. However, the causative agent has neither been isolated nor identified. This information should be considered in determining patient care and treatment.
The FDA approved this proposed change on January 25, 1984.
Although plaintiffs did not contest the adequacy of this warning, they claimed that Armour was negligent in not seeking approval of the warning at an earlier time, alleging that Armour had reasonable evidence of a potential risk of transmission of HIV from their product before September 1983. They additionally claimed that Armour was negligent in failing to send an individual letter to physicians, known as a Dear Doctor letter, warning of the risk.3
In order to carry their burden of proof on causation, plaintiffs had to show that their son was infected with HIV by an infusion of Armour’s concentrate, Factorate, during a discrete period of time in which Armour negligently failed to warn the prescribing physician of the possibility of AIDS when it knew that reasonable evidence existed of that risk with its product. Armour asserted that plaintiffs did not prove causation because they did not establish with reasonable medical probability that Jason was infected by Factorate as opposed to the other blood treatments he had received. Armour also argued that Jason’s prescribing physician, Dr. Jerry Barbosa, knew of the possibility that AIDS might be transmitted by concentrate when he first treated Jason with Armour’s product. Given this alleged knowledge, Armour asserted, under the learned intermediary rule, that any failure to warn Dr. Bar-bosa of that possibility at an earlier time was not the proximate cause of Jason’s infection.
The district court denied Armour’s Rule 50 motion for judgment as a matter of law, finding sufficient causation to go to the jury. The court also did not give Armour’s requested jury instruction on the learned intermediary defense. Initially, the court charged the jury, consistent with Armour’s instruction, that Armour had to prove that Dr. Barbosa knew at the time of treatment of the “possibility” that AIDS might be transmitted by blood products, the language contained in the warning subsequently approved by the FDA and conceded as adequate by plaintiffs. The court continued, however, with the presently-challenged instruction that Armour had to prove that Dr. Barbosa knew of “reasonable evidence of an association of a serious hazard, that is, AIDS, with Factor VIII concentrate.”
After the jury verdict in favor of plaintiffs, Armour filed both a motion for new trial and an alternative motion for judgment notwithstanding the verdict, alleging insufficient causation and error in the jury instructions. The district court, in two lengthy orders, denied both post-trial motions. Armour then appealed.
II. FACTUAL BACKGROUND
Jason Christopher suffered from moderate hemophilia, a disorder in which the afflicted person is subject to spontaneous bleeds that can cause severe arthritis, crippling, or death. The severity of the clotting disorder depends upon the extent to which the hemophiliac lacks a clotting protein in plasma called Factor VIII. To limit or treat bleeding, hemophiliacs must use blood products, donated by others, which contain the missing clotting factor. There are several types of blood products used to treat hemophilia— whole blood, cryoprecipitate, and Factor VIII concentrate — which contain the Factor VIII [1188]*1188clotting protein.4 Over the course of his life, Jason Christopher used all three products.
The disease known as AIDS first came to the attention of the United States medical community in 1981.5 The first report of AIDS in hemophiliacs occurred in July of 1982 when the Center for Disease Control (CDC) reported that three hemophiliacs were diagnosed with AIDS. Although the cause of the new disease was unknown, the CDC report suggested that hemophilia patients might be at risk because of their treatment with blood products. Evidence of the possibility that AIDS might be transmitted through blood products continued to surface through July 19836, at which time Armour concluded that there was reasonable evidence of an association of AIDS with blood products and decided to seek FDA approval of a revision to its package labeling. The proposed warning was submitted in September 1983 and approved in January 1984.
Armour introduced a new heat-treated concentrate in February 1984. Studies had shown that this heat-treated Factor VIII concentrate might reduce the risk of transmission of non-A non-B Hepatitis. Although it was not known at that time, the heat-treated concentrate later was shown to protect against AIDS. The heat-treated concentrate contained the identical AIDS warning as approved for the non-heat-treated concentrate.
Jason used each type of blood product available for treatment of his hemophilia. Jason’s medical records established his use of blood products as follows. In March and April of 1982, he received several units of whole blood and 64 units of cryoprecipitate at military bases in Germany. He received 33 units of cryoprecipitate in Washington, D.C., between April and June of 1982, 6 units in October 1982, and 4 more in January 1983. In total, Jason received 107 units of cryopre-cipitate. On three occasions in July 1982, Jason received non-Armour Factor VIII concentrate. In August 1982, he was diagnosed with acute Hepatitis B, caused by one of his earlier treatments with blood products.
In February 1983, Jason began treatment with Dr. Barbosa, the chief pediatric hematologist at All Children’s Hospital in St. Pe-tersburg, Florida. He received his first Armour concentrate on July 15, 1983, after an injury requiring therapy. Dr. Barbosa treated Jason with two more infusions of Factorate in August and two in November— once on November 10 and once on November 30. On November 29, 1983, Jason developed [1189]*1189a flu-like illness, with symptoms including fever, enlarged lymph nodes, and sore throat. After November 30, 1983, Jason received no further concentrate until May of 1984, when he began receiving heat-treated concentrate. Jason tested positive for HIV in late May 1985, developed AIDS in early 1990 and died in February 1992.
It was undisputed at trial that Jason could have been infected with the HIV virus by any one of the units of blood products used to treat his condition from May 1982 until May 1984, when the heat-treated concentrate was introduced. In order to prove Jason’s infection was caused by Armour’s concentrate, plaintiffs introduced the testimony of Dr. Barbosa, Jason’s treating physician, and Dr. William Robinson, a professor of medicine with a specialty in infectious diseases, including AIDS and hepatitis.
Dr. Robinson’s opinion testimony was based upon epidemiological studies and clinical evidence of Jason’s symptoms on November 29, 1983. The epidemiological studies used by Dr. Robinson, analyzed HIV infection rates in hemophiliacs using concentrate or cryoprecipitate. There was no dispute that the studies demonstrated an association between AIDS and these products and that, in the populations surveyed, concentrate users had a higher proportion of infection than did cryoprecipitate users. In addition, the studies showed that hemophiliacs using a greater amount of Factor VIII, either concentrate or cryoprecipitate, had a higher incidence of infection than did hemophiliacs using a lesser amount of those products. Dr. Robinson concluded from these studies that “the rate of what patients are infected and the percentage of patients infected in any year is related to in some way to the quantity of Factor VIII that they received.”
In Dr. Robinson’s opinion, in comparison to the hemophiliacs in the studies, Jason received such small amounts of Factor VIII in concentrate or cryoprecipitate that his risk of infection was very low.7 Nonetheless, based upon the higher incidence of AIDS in concentrate users than in cryoprecipitate users, and based upon the very small amount of cryoprecipitate that Jason used, Dr. Robinson opined that if Jason were infected by either product, it was more likely to have been concentrate. He further testified that Jason was more likely to have been infected from one of his later injections of concentrate than from one of his earlier treatments, concluding that Jason’s risk of infection increased as the amount of Factor VIII concentrate he received increased. Consequently, he targeted late 1983 to early 1984 as the time most likely for Jason to have been infected.
In addition, Dr. Robinson testified, based on examining Jason’s medical records, that the clinical symptoms exhibited by Jason on November 29, 1983, were a likely manifestation of an acute primary HIV infection.8 Although he conceded that the symptoms were not absolutely diagnostic of such an infection and that they mirrored symptoms of a common cold, Dr. Robinson testified that, in his expert opinion, they were the only ones in Jason’s medical history perfectly consistent with a primary HIV infection. Dr. Robinson also opined that this illness was connected with Jason’s treatment in early November with Factorate. By November 1983, it had been more than ten months since Jason had received cryoprecipitate or any other blood product except Armour’s Factor VIII con-[1190]*1190céntrate. The November 10 Factorate injection was particularly suspect, because Jason’s illness occurred nineteen days after that treatment — a time frame fitting within the typical one- to six-week incubation period for HIV.
Dr. Robinson acknowledged that he could not determine where an individual patient such as Jason would fall in the epidemiological studies, and that he could not identify to a reasonable medical probability whether a concentrate user was infected by a particular unit of concentrate. He concluded, however, given Jason’s clinical symptoms and the incubation period of the virus, combined with the epidemiological studies showing a greater likelihood of infection among concentrate users than among cryopreeipitate users, that Jason more, likely than not was infected in November 1983 when he received Armour’s Factorate.
Dr. Barbosa, Jason’s treating physician, testified that, “statistically speaking,” Jason “probably” was infected with concentrate rather than cryopreeipitate because concentrate came from thousands of donors unlike cryopreeipitate, which comes from one donor. He further stated that the clinical symptoms that Jason exhibited in November 1983, though consistent with a common cold, were “very suggestive” of primary HIV infection. He did not remember ever seeing those cluster of symptoms in Jason’s medical records prior to November 1983.
In its defense, Armour presented evidence and testimony to show that even if the jury found it to have issued an untimely warning, its negligence was not the proximate cause of Jason’s injury because of the learned intermediary rule. Armour argued that if Dr. Barbosa, when administering Factorate to Jason, had the same knowledge as was issued later in the concededly adequate warning, the jury could find that there was no causal link between its alleged negligence and Jason’s infection. Armour introduced evidence that Dr. Barbosa, a board-certified pediatric hematologist/oncologist, kept informed about new developments in his specialty by reading medical journals, going to medical meetings, and conferring with colleagues. Dr. Barbosa admitted that it was “certainly possible” that he had read articles published in early 1983 that suggested the risk of AIDS transmission through blood products. Dr. Barbosa testified that he did not alter treatment of Jason because he was not sure that AIDS was transmitted by Factor VIII and because he had no hard data to make him change his mind.9
III. STANDARD OF REVIEW
The standard of review with respect to the district court’s denial of Armour’s motions for judgment as a matter of law is whether there was substantial evidence of causation, taken in the light and with all reasonable inferences most favorable to the plaintiffs. Lamb v. Sears, Roebuck & Co., 1 F.3d 1184, 1187 (11th Cir.1993). This is a question of law subject to de novo review. Elrod v. Sears, Roebuck & Co., 939 F.2d 1466, 1469 (11th Cir.1991).
A district court has broad discretion in formulating jury instructions. Georgetown Manor, Inc. v. Ethan Allen, Inc., 991 F.2d 1533, 1538 (11th Cir.1993). Motions for new trial on the basis of erroneous and prejudicial jury instructions are committed to the discretion of the trial court and reviewed to ascertain whether there has been a clear abuse of that discretion. Pate v. Seaboard R.R., Inc., 819 F.2d 1074, 1077 (11th Cir.1987).
This Court examines jury instructions as a whole to determine whether they fairly and adequately addressed the issue and correctly stated the law. Georgetown Manor, Inc., 991 F.2d at 1538. Jury instruc[1191]*1191tions must be put in context; we consider the allegations of the complaint, the evidence presented, and the arguments of counsel when determining whether the jury understood the issues or was misled. Hasenfus v. Secord, 962 F.2d 1556, 1562 (11th Cir.1992), cert. denied, - U.S. -, 113 S.Ct. 972, 122 L.Ed.2d 127 (1993).
IV. DISCUSSION
On this appeal we address whether plaintiffs’ evidence established sufficient causation for the jury to find in their favor and whether the. district court erroneously instructed the jury on Armour’s learned intermediary defense.
A. Causation
Armour contends that there was insufficient evidence to prove causation and that the district court accordingly was obligated to direct a verdict in its favor. In order to establish a prima facie case in this diversity action, Florida products liability law required plaintiffs to prove by a preponderance of the evidence, with “reasonable medical probability,” that Armour’s alleged negligence in mid to late 1983 was the proximate cause of Jason’s HIV infection. Reaves v. Armstrong World Industries, Inc., 569 So.2d 1307, 1309 (Fla.Dist.Ct.App.1990). In other words, plaintiffs must show that is “more likely than not” that the defendant’s act was a substantial factor in bringing about the injury. Id. at 1309 (quoting Gooding v. University Hospital Building, Inc., 445 So.2d 1015, 1018 (Fla.1984)). “A mere possibility of such causation is not enough; and when the matter remains one of pure speculation or conjecture, or the probabilities are at best evenly balanced, it becomes the duty of the court to direct a verdict for the defendant.” Id. In this case, plaintiffs presented sufficient evidence to prove by a preponderance of the evidence that Armour’s alleged negligence was the proximate cause of Jason’s infection.10 See Lamb, 1 F.3d at 1187 (citing Boeing Co. v. Shipman, 411 F.2d 365, 374-75 (5th Cir.1969) (en banc)).
The only portion of Armour’s insufficient causation argument that merits discussion is the allegation that the statistical probabilities derived from the epidemiological studies, as analyzed by Dr. Robinson, were not scientifically valid and thus could not be used to determine the time or source of Jason’s infection. Upon careful review of Dr. Robinson’s testimony, there is no question that several of his statements, viewed in isolation, were statistically invalid11 and, as such, that the district court should not have admitted them.12 See Daubert v. Merrell Dow Pharmaceuticals, Inc., — U.S. -, -, 113 S.Ct. 2786, 2796, 125 L.Ed.2d 469 (1993) (holding that the trial court must screen experts’ evidence to ensure its reliability by conducting a “preliminary assessment of whether the reasoning or methodology underlying the testimony is scientifically valid ... [and] properly can be applied to the facts in issue”). In many other instances, [1192]*1192however, Dr. Robinson made statements that were inconsistent with the allegedly invalid testimony. For example, although he opined that Jason was more likely to be infected in late 1983 because Jason had received increasing amounts of concentrate each year, he also stated — completely opposite to his prior conclusion — that he could not pinpoint which vial of concentrate infected Jason, and that there was not a statistically significant difference between receiving three or six vials of concentrate.
As to those instances when Dr. Robinson’s testimony could be read to be statistically inaccurate, what is very clear is that Armour made no objections at any time. Armour filed no motion in limine challenging Dr. Robinson’s proposed testimony, nor did it raise any contemporaneous objections. In order to preserve this issue on appeal, Armour must have objected to the challenged testimony. See Saunders v. Chatham County Board of Commissioners, 728 F.2d 1367, 1368 (11th Cir.1984).
Entering an objection to evidence is not primarily a matter of building a record for appeal or a tactical maneuver by counsel. Its principal purpose is that counsel bring to the attention of the trial judge evidence that counsel considers inadmissible or prejudicial so that, if there is an error involved, the court has a chance to correct it on the spot.
Id.
What occurred in this case provides a classic example of why contemporaneous objections need to be made. Some of Dr. Robinson’s testimony could have been read to be inaccurate and inconsistent, but he was afforded no opportunity to clarify the apparent inconsistencies. If Armour believed that Dr. Robinson’s testimony was statistically invalid, it should have objected to that testimony, giving him the chance to explain his answers. Objecting also would have provided the district court with the opportunity not only to make a ruling on the accuracy and admissibility of the challenged testimony but also to clarify that testimony.
Absent an objection, we can review the challenged evidence only for plain error. Id. “Plain errors are obvious and substantial errors which seriously affect the fairness, integrity or public reputation of judicial proceedings.” United States v. Hope, 901 F.2d 1013, 1020 (11th Cir.1990), cert. denied, 498 U.S. 1041, 111 S.Ct. 713, 112 L.Ed.2d 702 (1991). In reviewing all of the evidence and Dr. Robinson’s entire testimony, including those statements that clearly were inconsistent with the allegedly invalid testimony, we hold that the district court did not commit plain error.
B. Jury Instruction
Armour rested its primary defense on the learned intermediary rule. It requested a specific jury instruction, which was denied, and now challenges the instruction that was given. Armour contends that the district court utilized an incorrect standard and misled the jury as to the governing law, thus depriving Armour of its affirmative defense. Before addressing the challenged instruction, we first review the learned intermediary rule in order to clarify the defense upon which the instruction was based.
The learned intermediary rule is a corollary to the rule that a manufacturer of prescription drugs or products discharges its duty to warn by providing the physician with information about risks associated with those products. Felix v. Hoffmann-LaRoche, Inc., 540 So.2d 102, 104 (Fla.1989). The manufacturer’s duty to warn of drug hazards runs to the physician, not directly to the patient. Id. The learned intermediary rule provides that the failure of the manufacturer to provide the physician with an adequate warning of the risks associated with a prescription product is not the proximate cause of a patient’s injury if the prescribing physician had independent knowledge of the risk that the adequate warning should have communicated. Id. at 105; Zanzuri v. G.D. Searle & Co., 748 F.Supp. 1511, 1517 (S.D.Fla.1990). Thus, the causal link between a patient’s injury and the alleged failure to warn is broken when the prescribing physician had “substantially the same” knowledge as an adequate warning from the manufacturer should have communicated to him. Tatum v. Schering Corp., 795 F.2d 925, 928 (11th Cir.1986).
[1193]*1193Armour challenges the jury instruction, alleging that the charge imposed a burden of proof greater than that required to defeat proximate cause under the learned intermediary rule. According to the rule, if Dr. Barbosa, the prescribing physician, had substantially the same knowledge as Armour’s warning would have communicated to him, then there would be no proximate cause between Armour’s alleged negligence and Jason’s infection. See Tatum, 795 F.2d at 928. Armour’s warning stated that:
The 'possibility exists that Acquired Immune Deficiency Syndrome, AIDS, an immunologic disorder with extremely severe consequences, may be transmitted by blood, blood products and blood derivatives, including clotting factors. However, the causative agent has neither been isolated not identified. This information should be considered in determining patient care and treatment, (emphasis added).13
Armour contends that in order to sustain its burden on the learned intermediary defense,it needed only to prove that Dr. Barbosa had substantially the same knowledge of the AIDS risk as would have been disclosed in the information set forth in this warning. Armour therefore requested the district court to charge the jury on its affirmative defense with an instruction that would reflect the warning’s language.
In addition to the requested charge, the district court instructed the jury that its “verdict would be for Armour if [it] found by the greater weight of the evidence that Jason Christopher’s hemophilia treater, that is, Dr. Barbosa, knew himself at the time in question that there was reasonable evidence of an association of a serious hazard, that is, AIDS, with Factor VIII concentrate.” (emphasis added). Armour asserts that the “reasonable evidence” language did not reflect the information set forth in its warning but instead distorted the learned intermediary rule by forcing Armour to prove that Dr. Barbosa had a level of knowledge greater than that disclosed by the warning. According to Armour, its warning would have communicated to Dr. Barbosa exactly what it stated — that the “possibility exists” that AIDS could be transmitted through blood products; if it was proven that Dr. Barbosa had substantially the same knowledge as in that warning, then Armour could have prevailed. Armour argues that the warning would not have communicated the additional knowledge that there was “reasonable evidence” of an association of AIDS with Factor VIII concentrate and that by including this language in the jury charge, the court imposed a higher and erroneous layer of proof on its affirmative defense.
Plaintiffs’ position is that federal regulations required Armour to initiate a warning change upon finding reasonable evidence of an association of its drug with a serious hazard. 21 C.F.R. § 201.57(e). They contend that the issuance of a warning from a pharmaceutical company, separate from the language therein, communicates to the physician that there is reasonable evidence of an association of serious adverse reactions with the product at issue because the manufacturer would not be permitted to issue any warning unless it had such evidence. (Red Brief at 41). See 21 C.F.R. § 201.57(e). Plaintiffs reason that, in addition to the language of Armour’s warning, which would have conveyed to Dr. Barbosa that the “possibility existed” of a risk of AIDS transmission, the warning standing alone would have communicated to him that “reasonable evidence” of the risk existed. They thus conclude that the inclusion of the standard in the jury charge was proper.
If the treating physician understood that a warning is issued only after a manufacturer has reasonable evidence of a serious hazard, the existence of a warning, regardless of its language, would communicate that information to the physician. In that circumstance, [1194]*1194equating the physician’s knowledge of the risk with the manufacturer’s knowledge for the purposes of the learned intermediary rule, as the district court did here, would be appropriate. The manufacturer’s statutory duty based on its knowledge of reasonable evidence of a risk, however, is wholly unrelated to the physician’s knowledge of the risk if the physician did not demonstrate that the warning conveyed that information to him. Absent additional knowledge of the manufacturer’s obligation, the standard under the learned intermediary rule remains whether the physician had substantially the same knowledge as would have been communicated in the actual warning, and any jury instruction given should reflect that standard.
“The purpose of jury instructions is to give the jury a clear and concise statement of the law applicable to the facts of the case.” Pesaplastic, C.A. v. Cincinnati Milacron Co., 750 F.2d 1516, 1525 (11th Cir.1985) (quoting Hanover Fire Insurance Co. v. Sides, 320 F.2d 437, 444 (5th Cir.1963)). In order to assess the propriety of the jury instruction in light of the facts of this ease, we must determine if there was any evidence that Armour’s warning would have communicated to Dr. Barbosa that there was reasonable evidence of an association between AIDS and blood products. See Bank South Leasing, Inc. v. Williams, 778 F.2d 704, 706 (11th Cir.1985) (stating that there must be evidence adduced at trial to support the giving of an instruction). Here, the district court’s charge to the jury regarding Dr. Barbosa’s knowledge would have been appropriate only if there were a factual basis for the charge. Such a basis could have been testimony by Dr. Bar-bosa that he was aware of the reasonable evidence standard or that warnings from drug companies conveyed such information to him.
Not only was there no testimony that physicians in general understood the reasonable evidence regulatory standard, there was no testimony that Dr. Barbosa was aware of such a standard. Absent knowledge of the underlying standard, the only information that Armour’s warning possibly could have conveyed to Dr. Barbosa was the language contained therein — that the “possibility existed” of an association between AIDS and blood products. The reasonable evidence standard thus was irrelevant to Dr. Barbo-sa’s knowledge of the risks associated with Factor VIII and should not have been included in the jury instruction on the learned intermediary defense.
If there is no basis in the record for the instruction given, such error may raise a “substantial and ineradicable doubt as to whether the jury was properly guided in its deliberations,” and reversal may be required. McElroy v. Firestone Tire & Rubber Co., 894 F.2d 1504, 1509 (11th Cir.1990) (quoting National Independent Theatre Exhibitors, Inc. v. Charter Financial Group, Inc., 747 F.2d 1396, 1402-03 (11th Cir.1984), cert. denied, 471 U.S. 1056, 105 S.Ct. 2120, 85 L.Ed.2d 484 (1985)). Given the complete lack of evidence in the record to support the inclusion of the reasonable evidence standard in the instruction, we hold that the district court erred by inaccurately charging the jury on Armour’s affirmative defense. Cf. Bank South Leasing, 778 F.2d at 706 (holding that the district court erred in failing to properly charge the jury on defendants’ affirmative defenses).
In some instances, even an inaccurate instruction may not mandate reversal. If the totality of the instructions “properly express the law applicable to the case, there is no error even though an isolated clause may be inaccurate, ambiguous, incomplete or otherwise subject to criticism.” Somer v. Johnson, 704 F.2d 1473, 1478 (11th Cir.1983) (quoting Johnson v. Bryant, 671 F.2d 1276, 1280 (11th Cir.1982)). In order to place this jury instruction in context and to determine whether it might have prejudiced Armour by confusing or misleading the jury as to the applicable principles of law, see Pate, 819 F.2d at 1077, we have carefully examined the record, including the allegations of the complaint and counsels’ arguments, Hasenfus, 962 F.2d at 1562. Although an inaccurate instruction may be harmless in the context of the whole trial, this was not such a case.
Because the learned intermediary rule was the heart of Armour’s defense, it was critical that the court properly guide the jury as to the standard to be applied. Mid-Texas Communications Systems, Inc. [1195]*1195v. American Tel. and Tel. Co., 615 F.2d 1372, 1386-87 (5th Cir.), cert. denied, 449 U.S. 912, 101 S.Ct. 286, 66 L.Ed.2d 140 (1980).14 “A defendant is entitled to have the court instruct the jury on the theory of the defense, as long as it has some basis in the evidence and has legal support.” United States v. Hedges, 912 F.2d 1397, 1405-06 (11th Cir. 1990) (quoting United States v. Orr, 825 F.2d 1537, 1542 (11th Cir.1987)). Failing to properly charge the jury on a defense theory may constitute reversible error. See Hedges, 912 F.2d at 1405-06. We review the trial in its entirety to determine the prejudicial effect of this instruction. See Hasenfus, 962 F.2d at 1562. In this case, not only did the court charge the jury with an instruction not supported by evidence adduced at trial, plaintiffs’ counsel reinforced the improper jury instruction by stressing the reasonable evidence standard in his closing argument.15 Given the importance of this instruction to Armour’s primary defense, the reiteration of an unsupported and irrelevant standard in closing argument strengthens our certainty that the jury was improperly guided in its deliberations. See Miller v. Universal City Studios, Inc., 650 F.2d 1365, 1372 (5th Cir. 1981).
Upon examining the totality of the instructions and after carefully reviewing the record, we are left with a substantial and ineradicable doubt that the jury was properly instructed and hold that the district court erroneously charged the jury on a standard not supported by evidence adduced at trial. As the district court erroneously instructed the jury on Armour’s defense, a new trial is required.
V. CONCLUSION
Because we find that plaintiffs presented sufficient evidence for the jury to find that Armour’s Factorate was the cause of Jason Christopher’s infection, we affirm the district court’s denial of Armour’s motion for judgment notwithstanding the verdict on the issue of causation. We also affirm the district court’s denial of Armour’s motion for a new trial based on allegedly erroneous missing witness arguments. The district court’s improper jury instruction on the learned inter-' mediary rule, however, constitutes reversible error and mandates a new trial.
AFFIRMED in part, REVERSED and REMANDED in part.