This products liability case presented questions of state law which we certified to the Idaho Supreme Court. Toner v. Lederle Laboratories, 779 F.2d 1429 (9th Cir.1986). That court having issued a full opinion in response to the certification, Toner v. Lederle Laboratories, 112 Idaho 328, 732 P.2d 297 (1987), we now complete our disposition of the appeal, and we affirm the judgment for the plaintiff based on the jury’s finding of negligence.
The facts are set forth in detail in our previous opinion. 779 F.2d at 1430-31. Kevin Toner, the plaintiff in the trial court, was paralyzed after vaccination with TriImmunol, a vaccine manufactured by the defendant, Lederle Laboratories. A state court action commenced by the plaintiff’s parents was removed to federal court based on diversity of citizenship. 28 U.S.C. § 1441 (1982). The plaintiff’s principal contention was that Lederle had failed to develop and market Tri-Solgen, an alternative vaccine that would have been safer than the one given to him. In a special verdict, the jury found that Lederle was negligent in connection with the product, and that the negligence was the proximate cause of the plaintiff’s injuries. In further special verdicts, the jury rejected the plaintiff's alternative theories of strict liability and breach of warranty. The jury awarded $1,131,200 in damages on the negligence claim. See 779 F.2d at 1433.
On appeal Lederle argues that the jury instructions on negligence were inadequate for failure to state that certain drugs are unavoidably unsafe and that it is permissible to market such products despite risks inherent in their use. See Restatement (Second) of Torts § 402A comment k (1965). Lederle also contends that the jury’s rejection of Toner’s strict liability claim necessarily is a conclusion that any defect in the product was unavoidable, which, at least on the facts of this case, is fatally inconsistent with a finding that Lederle was negligent in its manufacture or distribution. We examine these arguments in light of Idaho’s response to our certified questions.
The Idaho Supreme Court reformulated our certified questions as follows:
(1) Under Idaho law, do the principles set forth in Restatement (Second) of Torts § 402A comment k (1965) apply to strict liability claims, and in particular to the claim in this suit?
(2) (a) Under Idaho law, do the principles set forth in Restatement (Second) of Torts § 402A comment k apply to negligence claims, and in particular to the claim in this suit?
(b) If the above question no. (2)(a) is answered affirmatively, did the trial court’s instructions on negligence sufficiently incorporate those principles?
732 P.2d at 303. The answers to these questions suffice to dispose of the first issue raised by Lederle, the sufficiency of the negligence instructions; and guide us further in the determination whether the jury verdicts are inconsistent. We conclude the jury instructions on negligence were sufficient under Idaho law and that the jury’s special verdicts were not inconsistent.
The opinion of the Idaho Supreme Court discloses that Idaho adheres to the “unavoidably unsafe product” defense, as set forth in comment k, in strict liability cases based on defective design. 732 P.2d at 308. The doctrine recognizes that some products *512pose inherent risks, but that the benefits outweigh the risks inherent in their intended use, so that strict liability will not be imposed. Under Idaho law, as in most states, however, the plaintiff may proceed under a negligence cause of action even if comment k provides a defense to strict liability. Id. at 303 n. 5, 309-10. Quoting a decision of the Washington Supreme Court, the Idaho court in the case before us states that “[n]egligence and strict liability are not mutually exclusive because they differ in focus: negligence focuses upon the conduct of the manufacturer while strict liability focuses upon the product and the consumer’s expectation.” Id. at 303 n. 5 (quoting David v. Globe Mach. Mfg. Co. 102 Wash.2d 68, 684 P.2d 692, 696 (1984)).
The court also tells us that analysis under comment k is similar to negligence analysis in a general sense:
under negligence analysis, the utility of the act depends upon the value of the interest advanced, the extent to which it is advanced, and the opportunity for a less dangerous course of conduct, Restatement (Second) of Torts § 292 (1965), just as comment k’s application depends on the value of the product’s benefit, the extent to which the benefit accrues, and the availability of a feasible alternative design. The risks too are similarly considered by comment k and negligence law. See Restatement (Second) of Torts § 293 (1965). Such a weighing is implicit in the duty to use due care to avoid injuries while rendering services. Stephens v. Stearns, 106 Idaho 249, 257, 678 P.2d 41, 49 (1984). In sum, the determination under comment k that the design of a product is unavoidably unsafe and yet affords benefits outweighing its risks varies little from the determination under negligence law that the designing and marketing of the product was reasonably done.
732 P.2d at 310-11. The opinion states in summary that the jury instructions on negligence adequately enabled the jury to consider all the relevant factors. Id. at 311-12.
The decision of the Idaho Supreme Court allows us to hold with confidence that the jury’s determination of negligence was reached in accordance with Idaho law and that Lederle is not entitled to judgment on the negligence issue. The instruction on negligence permitted the jury to assess the reasonableness of Lederle’s conduct “in light of all the attendant circumstances,” Id. at 312, among which were the state of scientific knowledge and the utility of the Tri-Immunol vaccine in light of alternatives known by Lederle to be available. The concern we expressed in our prior opinion, that “the trial court may have omitted a material element of negligence in failing to instruct the jury to decide whether Tri-Immunol was an unavoidably unsafe product,” 779 F.2d at 1432, has been allayed.
We turn to the issue of whether the jury’s special verdicts are fatally inconsistent. We are bound to find the special verdicts consistent if we can do so under a fair reading of them. Gallick v. Baltimore & Ohio R. Co., 372 U.S. 108, 119, 83 S.Ct. 659, 666, 9 L.Ed.2d 618 (1963); Atlantic & Gulf Stevedores, Inc. v. Ellerman Lines, Ltd., 369 U.S. 355, 364, 82 S.Ct. 780, 786, 7 L.Ed.2d 798 (1962); Blanton v. Mobil Oil Corp., 721 F.2d 1207, 1213 (9th Cir.1983), cert. denied, 471 U.S. 1007, 105 S.Ct. 1874, 85 L.Ed.2d 166 (1985). When faced with a claim that verdicts are inconsistent, the court must search for a reasonable way to read the verdicts as expressing a coherent view of the case, and must exhaust this effort before it is free to disregard the jury’s verdict and remand the case for a new trial. Gallick, 372 U.S. at 119, 83 S.Ct. at 666. The consistency of the jury verdicts must be considered in light of the judge’s instructions to the jury. Id. at 120-21, 83 S.Ct. at 666-67; Bates v. Jean, 745 F.2d 1146, 1151 (7th Cir.1984).
The crux of the inconsistency issue is whether the jury could have found that Lederle was not strictly liable on the ground that Tri-Immunol was an unavoidably unsafe product, but also found that Lederle was negligent in failing to develop the Tri-Solgen vaccine. A similar inconsistency is alleged to exist between the negli*513gence verdict and the finding that Lederle did not breach the implied warranty of merchantability. Lederle contends that under Idaho law, the jury’s findings on the strict liability and breach of warranty issues necessarily constitute a finding that the Tri-Immunol vaccine was not defective. Lederle then asks how the jury could possibly have found Lederle negligent in marketing a nondefective product, especially when the balancing of risk versus utility that makes the product nondefective is similar to the analysis used to determine the reasonableness of the manufacturer’s conduct. See Toner, 732 P.2d at 310-11.
Although we recognize the abstract symmetry of Lederle’s argument, we must adhere to our duty to reconcile the jury verdicts if there is a reasonable way to do so. In light of the Idaho Supreme Court’s opinion and the instructions actually given to the jury, we think Lederle’s paradox can be resolved and the jury’s verdicts reconciled. As the Idaho Supreme Court points out, the focus in negligence is on the manufacturer’s conduct, while in strict liability it is on the product and the user’s expectations. Id. at 303 n. 5; Rojas v. Lindsay Mfg. Co., 108 Idaho 590, 592, 701 P.2d 210, 212 (1985). Comment k functions in the strict liability context to encourage production and marketing of useful products, but it does not remove the incentive for safe design by the manufacturer that is supplied by the threat of negligence liability. 732 P.2d at 310. The instruction given to the jury on the meaning of the strict liability term “defective condition, unreasonably dangerous” reflects this focus:
“Instruction No. 21: A product is in a defective condition, unreasonably dangerous to persons if it is more dangerous than would be expected by an ordinary person who may reasonably be expected to use it. The law does not say what would be expected by an ordinary person or who may reasonably be expected to use the product. Both of these are for you to decide.”
This focus is different from the emphasis in the instruction on negligence, which is phrased in terms of the manufacturer’s duty to “exercise ordinary and reasonable care not to expose the potential consumer to an unreasonable risk of harm.”
It is not enough for Lederle to argue that the jury’s finding of negligence concludes that Tri-Immunol was defective, while its finding on strict liability states a contrary view. The law and the instructions required the jury to examine the case from two different points of view. It is reasonable to read the special verdicts as saying that Lederle’s failure to develop the Tri-Solgen vaccine was unreasonable conduct, although the danger posed by the product itself was not greater than an ordinary consumer would reasonably expect.
Lederle contends that even accepting the differing focus of the jury instructions, Toner’s failure to pursue his strict liability claim of inadequate warning means that the ordinary user of the product must be assumed to have all the information available to Lederle. This being so, Lederle claims, the reasonableness determination in strict liability mirrors the reasonableness determination in negligence, and the verdicts are inconsistent. We disagree.
As an initial matter, we doubt that Lederle is entitled to treat the case as if the jury found adequate warning; the issue was not before it. We decline to treat Toner’s litigation decision not to pursue the warning theory as if it were a stipulation that Toner had adequate warning. Even if the jury had reached such a verdict, the vantage point of the ordinary consumer, in the strict liability instruction would not be identical to the vantage point of the manufacturer itself. Such a rule would do away with the distinction between strict liability and negligence set forth by the Idaho Supreme Court, and does not comport with the strict liability instruction that was given to the jury, which specifically stated that “the law does not say what would be expected by an ordinary person.”
Similarly, we do not think the jury’s verdict on the negligence issue is necessarily inconsistent with its finding that Lederle did not breach the implied warranty of merchantability. On that issue, the jury was instructed that “a breach of this war*514ranty occurs when the product is not fit for the ordinary purposes for which the product is to be used.” It is not beyond peradventure that the jury thought that Tri-Immunol was fit for the purposes for which it was to be used, but that Lederle was negligent in failing to replace it with a better product. This case is distinguishable from Witt v. Norfe, 725 F.2d 1277, 1279-80 (11th Cir.1984), in which the manufacturer of a glass shower door could not have been negligent in its production.if the product were fit for its intended purpose as found by the jury. In the case before us, the jury could have believed that the vaccine was not per se unfit for its intended use but that Lederle was negligent in failing to develop the alternative. The jury’s verdicts rendered pursuant to the instructions as given are amenable to an interpretation that makes them consistent.
Finally, we reject Lederle’s claim that the plaintiffs failed to show that Lederle’s negligence was an actual or proximate cause of the plaintiff’s injury. Much of the testimony at trial centered on whether the injury would have occurred but for Lederle’s failure to utilize the Tri-Solgen vaccine. There was expert testimony that it would not have. We reject Lederle’s argument that the plaintiff failed to prove causation because he failed to show that Tri-Solgen would have been approved by the FDA. The Third Circuit has determined that this issue is for the jury, keeping in mind the FDA’s statutory duty to monitor drugs for safety and effectiveness. Stanton v. Astra Pharmaceutical Prods., 718 F.2d 553, 569 (3rd Cir.1983). We agree. The testimony by witnesses for both sides regarding the length of time that the Tri-Solgen vaccine was available and the absence of evidence that the FDA would not have approved it had Lederle submitted the product was sufficient to permit the jury to infer that Toner’s injury could in fact have been avoided by the Tri-Solgen vaccine.
Were we charged with deciding whether jury verdicts are the most sensible way of allocating risks and costs regarding vaccines that have been proved beneficial, we might well design a different method. The question, however, is one of Idaho law. The Idaho Supreme Court enforces the traditional tort system, leaving for its legislature any comprehensive reform. We must defer to this state law determination.
AFFIRMED.