Dissent by Judge FERGUSON.
OPINION
KLEINFELD, Circuit Judge.This is a products liability case involving no questions of federal law, just a question of state law.2
Facts
Carol Adams had persistent severe back and neck pain, radiating down into her fingers. The pain was caused by spurring on her vertebrae and deterioration of the disks in her neck. It did not respond to steroids, traction, or other conservative treatment, so she elected to have surgery to remove the spurs and the bad disks. Her neurosurgeon removed three disks and parts of three vertebrae. To maintain the stability of the spine, he removed a piece of bone from her hip and inserted it in the gap in her neck vertebrae. This is a standard way to do the cervical fusion operation he was performing.
The neck has to be held still during the three months or so until the piece of hip bone fuses with the neck bones to form a solid and stable substitute for the removed pieces of vertebrae. This used to be done, in an operation of this magnitude, by requiring the patient to wear a halo for four months. Holes would be drilled into the skull of the patient in order to screw the halo to the skull, and the halo would be held in place by a kind of vest. Obviously, this was hard for the patient to bear.
In 1991, Synthes Spine came out with a device to avoid the halo. Instead of drilling holes in the patient’s head and requiring the patient to wear the affixed halo, the surgeon would screw a metal plate to the bones of the spine during the surgery. This would hold the bones in place while they fused. Ms. Adams’ doctor, Leslie Bornfleth, M.D., testified that while patients were routinely advised that they *1116could have the halo instead, none wanted it once they could opt for the Synthes Spine plate. Ms. Adams made that election.
The operation went fine. The bones in Ms. Adams’ neck properly fused. But three years later, in 1998, Ms. Adams had neck pain and difficulty swallowing, and consulted Allan J. Drapkin, M.D. (Dr. Bornfleth had by then retired from practice). Dr. Drapkin’s x-rays showed that the plate had broken and one of the screws had come partly out. The fusion was solid, and the plate was no longer necessary to hold her spine together, so Dr. Drapkin recommended surgery to “remove the plate, and particularly remove that screw” in order to relieve “mechanical compression on the esophagus.” He performed the removal surgery, and it succeeded without complication.
A person can’t buy one of these Synthes Spine plates at the drugstore. They are sold only to physicians. Dr. Bornfleth, a neurosurgeon- who installed quite a few, did not even have them in his office. They were kept at the hospital. This was a medical device, fairly new when Dr. Born-fleth installed it in 1995 (four years after its introduction), and was sold to and used by physicians such as Dr. Bornfleth.
The directions that come with the device say to remove it once the bones have fused. In the “Precautions” section of the “SUGGESTIONS CONCERNING ORTHOPAEDIC METALLIC INTERNAL FIXATION DEVICES,” the instructions say that these implants can break. So. “[w]hile the surgeon must make the final decision,” the instructions recommend that the implants be removed once the bones, have fused:
Removal after fracture healing. Metallic implants can loosen, fracture, corrode, migrate, .cause pain, or stress shield bone even after a fracture has healed, particularly in young, active patients. While the surgeon must make the final decision on implant removal, we recommend that whenever possible and practical for the individual patient, fixation devices should be removed once their service as an aid to healing is accomplished. Implant removal should be followed by adequate postoperative management to avoid refracture.
Despite this manufacturer’s recommendation, a lot of surgeons don’t remove the implants. Dr. Bornfleth went to meetings where surgeons discussed how to use the implants, and the consensus was “there is no reason to take the plates out because if the fusion is solid and the plate is there, it’s going to cause no problem.” As Dr. Drapkin, who removed the broken plate, testified, the risk of a plate breaking is “very small,” so ordinarily “there is no reason to expose the patient to a risk of a second surgery on a routine basis” to remove it. Dr. Drapkin said he would still use a Synthes Spine plate if he were performing this sort of surgery, even after Ms. Adams’ plate broke. She was the unfortunate person for whom the “very small” risk materialized, necessitating removal.
Ms. Adams sued Synthes Spine for selling a defective product: Her damages were for the pain in her throat after the plate broke, the anxiety of the several days between Dr. Drapkin’s telling her it was broken and performing the surgery to remove it, and associated pain and suffering and medical expenses. The district court granted summary judgment against Ms. Adams, and she appeals on the theory that the product was defectively designed because it could break.
Analysis
We review summary judgment de novo, to determine whether, viewing the evidence in the light most favorable to the nonmoving party, there are any genuine *1117issues of material fact.3 Having so reviewed the case, we have concluded that the district judge’s carefully analyzed decision was correct, and we affirm.
The theory of the plaintiffs case, as briefed on appeal, is that the plate didn’t conform to the reasonable expectations of the consumer, Dr. Bornfleth, because it broke, and he wasn’t adequately warned that it could break.
The applicable Washington statute on products liability says that in determining whether a product is “reasonably safe,” the trier of fact “shall consider whether the product was unsafe to an extent beyond that which would be contemplated by the ordinary consumer.”4 Under Washington law, the “consumer” of a prescription-only medical device such as this is the physician, not the patient in whom it is installed,5 a critical point which plaintiff concedes.
The Washington Supreme Court held in Terhune v. AH. Robins Co.6 that the manufacturer of the Daikon Shield was not strictly liable to a woman whose uterus was perforated by one, because medical products “available only on prescription or through the services of a physician”7 are, in Washington, treated as falling under comment k of the Restatement (Second) of Torts § 402A.8 As the Restatement puts it, some products are “incapable of being made safe” but, because of their benefit, are not “unreasonably dangerous,” such as “many ... drugs, vaccines, and the like, many of which for this very reason cannot legally be sold except to physicians, or under the prescription of a physician.”9 Where “proper warning is given,” the seller “is not to be held to strict liability for unfortunate consequences attending their use.”10
Washington applies this rule not only to such medical products as' vaccines for deadly diseases, but, as the Daikon Shield case illustrates, much more broadly, to medical products where the physician acts as a “learned intermediary between the manufacturer or seller and the patient.”11 In Terhune, the issue was whether to apply the rule to a product, the Daikon Shield, that was not essential to; “cure a malady” but merely “requirefd] a physician’s services, his knowledge and his skill.” 12 The court held that the manufacturer’s duty to warn ran only to the physician, not to the patient, and “the manufacturer should not be held to , account if it has done its duty in this regard” and provided appropriate information to the physician, “even though the physician decides, in the exercise of his own judgment, to withhold the information from his patient.” 13
Had there been any doubt about the vigor and breadth of Terhune, it would have been eliminated by the more recent case of Young v. Key Pharmaceuticals, Inc.14 In Young, an asthma treatment *1118caused severe, permanent brain damage in a small child, a known risk that occasionally materialized.15 The Washington Supreme Court held that comment k applied (which is to say that strict liability did not apply) to “prescription medical products” without any separate determination on a case-by-case basis “because of the character of the medical profession and the active, intermediate involvement of a physician.” 16
This is a strict liability case. The plaintiff offered no evidence in district court and no argument here for a negligence theory. And Washington law rules out strict liability for prescription medical products such as the Synthes Spine plate, provided that proper warning is given to the physician.
As for whether the given warning was adequate, appellant’s argument is basically that the warning wasn’t clear enough for a doctor to notice or understand, and “most convincingly, it was obvious that the ‘warning’ was inadequate because NO ONE FOLLOWS IT.”17 As for whether it was clear enough for a physician to read and understand it, there was no evidence that it wasn’t. Dr. Bornfleth testified that he probably read it. The words in the warning are perfectly clear. What’s more, this isn’t something where physicians just mail away for it, read the directions, and start screwing these things onto patients’ spines. Dr. Bornfleth testified that he went to professional meetings where use of these devices was discussed by professionals in the field, and his understanding was that the device shouldn’t be removed, even though the manufacturer said it should be.
That the surgeons usually did not remove the device does not show that they didn’t understand the warning, just that they generally didn’t agree with the recommendation. Dr. Drapkin explained why- — -the risk of the surgery to remove the plate exceeded the risk of leaving it in, because in most, cases it wouldn’t break. A second surgery could safely be left for the rare cases when it broke. Even the rarest events happen sometimes, to someone. It is extremely unfortunate that Ms. Adams was that someone, that she suffered pain, and anxiety from a broken Synthes Spine plate. (Because her argument implies that, had Synthes Spine given an adequate warning, Dr. Bornfleth would have opened up her neck and performed a second surgery to remove the device after her bones fused, she can’t really be complaining about getting a second surgery, just that the breaking of the plate caused a great deal of mental distress and pain leading up to the second surgery.)
There isn’t any evidence in the record from which reasonable jurors could conclude that the warning was inadequate. It plainly said that the plate could break and that the manufacturer recommended removal. That physicians didn’t follow the recommendation 'doesn’t show that they couldn’t or didn’t read it and understand it, just that in their medical judgment, it wasn’t wise to follow it. Dr. Bornfleth testified that he was “surprised” that the plate broke in Ms. Adams’ case. That doesn’t show anything inadequate about the, warning, just that breakage was rare. Whenever a rare event occurs, the occurrence is surprising. People often choose to bear the risk of a rare event in order to avoid a certain and unattractive alternative, which in this case would be a second surgery, unnecessary in most cases.
*1119The dissent seems to be based on the notion that, had the warning been adequate, all physicians would know that they ought to remove the Synthes Spine plate after the bones have fused. That is simply a substitution of our dissenting colleague’s medical judgment for the medical judgment of the neurosurgeons and ortho-paedic surgeons who actually have the experience and knowledge needed to make competent medical judgments about whether and when such a device ought to be removed. Their judgment differs from his. It is not reasonable to suppose that that is because the words “precaution,” “the plate may fracture,” and “the plate should be removed” are too difficult and obscure for neurosurgeons and orthopaedic surgeons to read and understand. They’ve just decided that one surgery is better than two, except in the occasional instances where the device breaks. As Dr. Bornfleth testified, the surgery to remove the device entails the usual surgical risks of anaesthesia, infection, etc., and in his judgment, that risk wasn’t worth running.
AFFIRMED.
. This opinion is published pursuant to Ninth Circuit Rule 36-2(g).
. See United States v. Shumway, 199 F.3d 1093, 1102-03 (9th Cir.1999); Fed.R.Civ.P. 56(c).
. Wash. Rev.Code § 7.72.030(3).
. See Terhune v. A.H. Robins Co., 90 Wash.2d 9, 577 P.2d 975, 978 (1978).
. Id.
. Id. at 978.
.Id. at 977.
. Restátement (Second) of Torts § 402A cmt. k (1965) (emphasis in original).
. Id.
. Terhune, 577 P.2d at 978.
. Id.'
. Id. at 979.
. 130 Wash.2d 160, 922 P.2d 59 (1996) (en banc).
. Id. at 61.
. Id. at 64 (internal citations and quotation marks omitted).
.Brief for Appellant at 9, Adams v. Synthes Spine Co., No. 00-35094 (emphasis in original).