576 So.2d 728 (1991)
Scott and Denise ADAMS, Appellants,
v.
G.D. SEARLE & CO., INC., Appellee.
No. 89-02843.
District Court of Appeal of Florida, Second District.
January 18, 1991.
Rehearing Denied April 4, 1991.
Michael D. Eriksen of Wagner, Nugent, Johnson, Roth, Romano, Eriksen & Kupfer, P.A., West Palm Beach, for appellants.
A. Broaddus Livingston and Alan C. Sundberg of Carlton, Fields, Ward, Emmanuel, Smith & Cutler, P.A., Tampa, for appellee.
PATTERSON, Judge.
Scott and Denise Adams brought a product liability action against G.D. Searle & Co., Inc., the manufacturer of Denise's intrauterine contraceptive device (IUD), after Denise became pregnant and suffered a septic abortion. The Adamses alleged in four counts that the IUD was defective and Searle was strictly liable for damages; that Searle failed to test the IUD adequately; that Searle's product labeling inadequately warned physicians and patients of the dangers associated with the IUD; and that Searle fraudulently misrepresented the IUD as safe. The trial court granted summary judgment as to the entire complaint in favor of Searle. We reverse.
Because of unrelated medical difficulties, Denise Adams cannot use oral contraceptives. She selected an IUD as her method of contraception based on the recommendation of her physician, Dr. Helwig. On January 2, 1980, Dr. Helwig prescribed for her and inserted the "Copper 7" (Cu-7), a copper-containing IUD manufactured by Searle.[1]
When Denise decided to try the Cu-7, Dr. Helwig gave her a patient brochure which Searle provided in accordance with FDA regulations. This brochure warned of specific dangers associated with IUD's in general and the Cu-7 in particular. It also identified adverse reactions that have been reported with the use of IUD's. Denise read the brochure and signed a card stating, "I received and have read the patient booklet, Cu-7."
Almost a year after Dr. Helwig inserted the Cu-7, Denise became pregnant. Dr. Helwig's normal procedure in such cases is to remove the IUD, because this reduces the risk that the patient will miscarry later in the pregnancy. However, in Denise's case, he found that the IUD's polypropylene removal string had retracted into her uterus. This limited Denise to three options: she could elect to have an abortion; Dr. Helwig could probe for the IUD, which in itself could cause an abortion; or Denise could risk trying to carry the infant to term with the IUD in place. Denise chose the latter, and the pregnancy proceeded.
When she was in the middle trimester, about twenty-four weeks along, Denise began to suffer flu-like symptoms, fever, and premature rupture of membranes. To safeguard her life, physicians induced labor and on April 14, 1981, she delivered a premature infant son who died within hours. The child's death certificate listed the cause of death as premature rupture of the membranes and uterine infection caused by the Cu-7.
On January 21, 1983, the Adamses filed this action against Searle. They sought damages for Denise's permanent injuries, wrongful death of the child, loss of consortium for Scott, mental anguish, loss of support from the child, and medical and burial expenses. In Count I, the strict liability count, they alleged the existence of specific defects in the Cu-7. In Count II, the "negligent failure to test" count, the Adamses alleged that Searle failed to perform tests necessary to discover the Cu-7's defects and dangers. In Count III, the failure-to-warn count, they alleged that Searle knowingly failed to warn either physicians or patients about the defects and dangers. And, in Count IV, the fraud count, they alleged that to induce physicians to prescribe the Cu-7, Searle knowingly withheld facts it knew were material.
Early in the litigation, the trial court dismissed the fraud claim for failure to state a cause of action. However, within the four corners of the complaint, see Geer v. Bennett, 237 So.2d 311 (Fla. 5th DCA 1970), the Adamses alleged that Searle misrepresented and omitted specific material facts about the Cu-7 to Dr. Helwig and Denise with the intention to induce them to rely on the misrepresentations and omissions; and that Dr. Helwig and Denise did so to Denise's detriment. These allegations state a cause of action for fraud, and we therefore reverse the dismissal of Count IV.
The remaining counts were disposed of by the trial court's order granting Searle's motions for summary judgment. In discussing whether summary judgment was appropriate as to each count, we address first the claim for negligent failure to test, next the question of failure to warn, and last the matter of strict liability for design defects.
The trial court correctly disposed of the claim for negligent failure to test. That count alleged that Searle negligently failed to perform tests upon the Cu-7 to discover its defects and dangers. We agree with Kociemba v. G.D. Searle & Co., Inc., 707 F. Supp. 1517 (D.Minn. 1989), that a manufacturer's duty to inspect and test is not a separate cause of action. The duty to test, as that court concluded, is a subpart of a manufacturer's duty to design a product with reasonable care, and thus is subsumed in the plaintiffs' claims for defective design and failure to warn. 707 F. Supp. at 1527. We therefore affirm the summary judgment as to this claim.
The trial court, in granting final summary judgment for Searle, found as a matter of law that the Cu-7's product labeling was adequate. Although the adequacy of prescription drug warnings can become a question of law where the warning is accurate, clear, and unambiguous, it otherwise remains a question of fact. See Upjohn Co. v. MacMurdo, 562 So.2d 680, 681-2 (Fla. 1990); Felix v. Hoffmann-LaRoche, Inc., 540 So.2d 102, 105 (Fla. 1989). In this case, the Adamses produced evidence tending to show that the warnings were not accurate, clear, and unambiguous; therefore, the adequacy of the warnings was a question of fact for the jury, and the trial court erred in deciding it as a matter of law.[2]
As to the defective design claim, we find that the summary judgment must be reversed. To grant the motion for summary judgment as to this issue, the trial court necessarily must have found that Searle could not be held responsible under strict liability for defects in the design of the Cu-7. It could have arrived at this conclusion only by applying comment k of the Restatement (Second) of Torts (1959) section 402A, to the Cu-7.[3]
Comment k protects the manufacturers of certain products from strict liability for design defects. It states:
UNAVOIDABLY UNSAFE PRODUCTS.
There are some products which, in the present state of human knowledge, are quite incapable of being made safe for their intended and ordinary use. These are especially common in the field of drugs. An outstanding example is the vaccine for the Pasteur treatment of rabies which not uncommonly leads to very serious and damaging consequences when it is injected. Since the disease itself invariably leads to a dreadful death, both the marketing and the use of the vaccine are fully justified, notwithstanding the unavoidable high degree of risk which they involve. Such a product, properly prepared, and accompanied by proper directions and warnings, is not defective nor is it unreasonably dangerous. The same is true of many other 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. It is also true in particular of many new or experimental drugs as to which, because of the time and opportunity for sufficient medical experience, there can be no assurance of safety, or perhaps even of purity of ingredients, but such experience as there is justifies the marketing and use of the drug, notwithstanding a medically recognizable risk. The seller of such products, again with the qualification that they are properly prepared and marketed, and proper warning is given, where the situation calls for it, is not to be held to strict liability for unfortunate consequences attending their use, merely because he has undertaken to supply the public with an apparently useful and desirable product, attended with a known but apparently reasonable risk.
Restatement (Second) of Torts (1959) section 402A, comment k. Thus, by its terms comment k applies to products which current knowledge and technology cannot make safe for their ordinary use, but for which society has a need great enough to justify using the product despite its dangers. Comment k increases the standard a plaintiff must meet from that of strict liability to negligence.[4]See Toner v. Lederle Laboratories, 112 Idaho 328, 336, 732 P.2d 297, 305 (1987).
Some courts apply comment k uniformly to all prescription drugs. See, e.g., Brown v. Superior Court, 44 Cal.3d 1049, 1061, 751 P.2d 470, 477, 245 Cal. Rptr. 412, 418 (1988); McKee v. Moore, 648 P.2d 21 (Okla. 1982); Terhune v. A.H. Robins Co., 90 Wash.2d 9, 577 P.2d 975 (1978). While we recognize that policy reasons exist to support such an approach, we are disinclined to adopt it for several reasons.
First, the language of comment k indicates that it was not intended to apply to all prescription drugs.[5] It speaks of "some products," "especially common in the field of drugs," "many other drugs," and "many new and experimental drugs." By definition, "some" and "many" are less than all. See Toner, 112 Idaho at 339, 732 P.2d at 308; Kociemba v. G.D. Searle & Co., 680 F. Supp. 1293 at 1300 (D.Minn. 1988).
Second, the comment itself adopts a risk/benefit analysis in using the Pasteur rabies vaccine as an "outstanding example" of a product which is protected by comment k. It notes that injecting the vaccine may lead to very serious consequences, but because the disease "invariably leads to a dreadful death," the vaccine's marketing and use are justified. Applying comment k uniformly to all prescription drugs therefore rejects the comment's own approach to determining its scope.
Third, we believe that the policy reasons supporting a blanket approach are countervailed by those supporting a more selective application of the comment. For example, courts which apply comment k to all prescription drugs frequently state that the public's interest in the development of prescription drug products requires the user to bear all costs of injury unless the drug product was negligently manufactured, negligently designed, or accompanied by inadequate product warnings; otherwise, drug manufacturers might be discouraged from developing new products. See, e.g., Brown, 44 Cal.3d 1049 at 1062, 751 P.2d 470 at 478, 245 Cal. Rptr. 412 at 420. We do not believe this policy has a sound foundation.
We believe that a more selective application will encourage, rather than discourage, improvements in prescription products. Comment k was designed in part to protect new and experimental drugs. As the Toner court noted:
Comment k states: "There are some products which, in the present state of human knowledge, are quite incapable of being made safe for their intended and ordinary use." Obviously, for this to be true, the design must be as safe as the best available testing and research permits.
112 Idaho at 306, 732 P.2d at 306 (citations omitted). Thus, a product which is as safe as current testing and research permits should be protected. The reverse is also true; a product which is not as safe as current technology can make it should not be protected. See Toner, 112 Idaho at 337-8, 732 P.2d at 306-7; Belle Bonfils Memorial Blood Bank v. Hansen, 665 P.2d 118 at 123 (Colo. 1983); Castrignano v. E.R. Squibb & Sons, Inc., 546 A.2d 775 (R.I. 1988). In Castrignano, the Rhode Island Supreme Court stated that "[W]e believe the societal interest in the development and marketing of prescription drugs will be adequately served by extending comment k to prescription drugs on a case-by-case basis." 546 A.2d at 781.
The allegations in this case illustrate why we prefer this approach. The Adamses allege that the Cu-7's polypropylene withdrawal string had a "memory" which caused it to tend to retract into the uterus; that Searle knew of this problem; that by 1978 it had identified polyethylene as a safer alternative; that it used polyethylene successfully in another IUD; and that despite its knowledge, Searle continued to use polypropylene for the Cu-7's string. Applying comment k to all prescription drugs does nothing to discourage such practices.
Under a selective application theory, if comment k protected the Cu-7 originally,[6] then certainly that protection would have been withdrawn after Searle learned that polyethylene made the Cu-7 safer but did nothing. To keep the Cu-7 within the scope of comment k, Searle would have had to replace the polypropylene string as soon as possible. Thus, it appears that a selective application should encourage improvements in prescription products.
Another factor courts have cited as supporting the determination that comment k should apply to all prescription drugs is that the Federal Food and Drug Administration (FDA) approves such products and their warning label inserts. See, e.g., Leibowitz v. Ortho Pharmaceutical Corp., 224 Pa.Super. 418, 307 A.2d 449 (1973). However, it seems likely to us that a drug manufacturer is in a better position to monitor the current state of knowledge and technology, as applied to its products, than is the FDA. We hesitate to hold that a manufacturer is excused from making changes it knows will improve its product merely because an older, more dangerous version received FDA approval.
We therefore reject a blanket approach and decline to apply comment k to all prescription products. Instead, we follow those courts which hold that comment k is an affirmative defense to a strict liability claim. See, e.g., Ortho Pharmaceutical Corp. v. Heath, 722 P.2d 410 at 416 (Colo. 1986); Toner, 112 Idaho at 339, 732 P.2d at 308; Feldman v. Lederle Laboratories, 97 N.J. 429, 479 A.2d 374, 383 (1984); Belle Bonfils, 665 P.2d at 122-3, 125, n. 12; Senn v. Merrell-Dow Pharmaceuticals, Inc., 305 Or. 256, 751 P.2d 215 (1988); Castrignano, 546 A.2d at 782. These courts generally adopt a risk/benefit analysis in determining whether comment k applies to a given product. We believe such an approach better comports with the language of comment k and the policy underlying it.
We therefore hold that the seller has the burden to establish the application of comment k. We also hold that whether comment k applies is a mixed question of fact and law. If reasonable minds could reach only one conclusion, then the trial judge may rule; but if reasonable minds might differ, then the matter must be submitted to the jury. See Castrignano, 546 A.2d at 782.
Any weighing of a product's risk against its benefit should consider the value of the benefit, the seriousness of the risk, and the likelihood of both. For comment k to apply, the product's benefits must outweigh its known risks as of the date the product is distributed. Toner, 112 Idaho at 337, 732 P.2d at 306; Belle Bonfils, 665 P.2d at 122-3. In Toner, the court stated:
Comment k does not require sellers to be clairvoyant. When a product is "apparently useful and desirable," considering its benefits and risks, then it ought to be immune from strict liability. Such products can include new or experimental drugs for which "there can be no assurance of safety." Comment k only requires that the balance "apparently" tip toward the benefit of a product at the time of distribution.
... .
If new information later tips the balance toward the risk of a product, or if new developments make possible a safer design, at that point further distributions of the product are not protected by comment k.
Toner, 112 Idaho at 338, 732 P.2d at 306-7 (citations omitted). Thus, it appears that extending comment k's protection to prescription drug products on a case-by-case basis will adequately serve the public interest.
In this case, the trial court could not have granted summary judgment for Searle without finding as a matter of law that comment k protected Denise Adams's Cu-7. After reviewing the evidence, we conclude that reasonable minds might differ on this question. We therefore hold that the trial court erred in granting Searle's motions for summary judgment.
In conclusion, we reverse the order dismissing the Adamses' fraud claim for failure to state a cause of action. We also reverse the final summary judgment as to their claims for failure to warn and strict liability for design defects. We affirm the summary judgment as to the claim for "negligent failure to test," and remand the case for further proceedings consistent with this opinion.
FRANK, A.C.J., concurs.
PARKER, Judge, dissenting in part and specially concurring in part with opinion.
PARKER, Judge, dissenting and concurring.
I dissent from the majority's decision that summary judgment was improperly granted for Searle on the issue of the adequacy of the warning and specially concur on the majority's application of comment k of the Restatement of Torts (Second) as a defense to the strict liability claim. While the majority acknowledges that there are instances where the adequacy of a warning becomes a question of law that may be determined on summary judgment, it concludes, without further explanation, that Searle's Cu-7 warning creates a question of fact concerning its adequacy because it is not accurate, clear, and unambiguous citing the federal district court's statement in Zanzuri, 748 F. Supp. at 1516-17, denouncing the Cu-7 warning's lack of clarity. However, that statement in Zanzuri was concerned with symptoms other than complained of by the plaintiff in this case.
The plaintiff in Zanzuri suffered two ectopic pregnancies caused by pelvic inflammatory disease, leading to the eventual removal of both fallopian tubes and infertility. Thus, the opinion in Zanzuri does not control or prevent a decision in this case that the Cu-7 warning was adequate as a matter of law in advising the physician, as the learned intermediary, of the specific result which was obtained here septic abortion or infected miscarriage.
Our supreme court in MacMurdo, 562 So.2d at 680 and Felix, 540 So.2d at 102 has held the warnings applicable in those cases adequate as a matter of law upon finding that they were sufficiently clear to inform physicians of the precise symptoms or conditions suffered by the plaintiffs in those cases. Under MacMurdo, the fact that the warning placed doctors on notice that the symptoms complained of there continuous bleeding for three months or dysfunctional bleeding was a possible result of the use of the drug Depo-Provera was adequate even though the warning did not specify dysfunctional bleeding but only referred to breakthrough bleeding and change in menstrual flow as associated adverse reactions to the use of the drug.
Specifically, the package insert provided in this instance by Searle to physicians with the Cu-7 stated as follows:
Warnings
1. Pregnancy: ...
(b) Septic Abortion. Reports have indicated an increased incidence of septic abortion associated in some instances with septicemia, septic shock and death in patients becoming pregnant with an IUD in place. Most of these reports have been associated with the mid-trimester of pregnancy... . If pregnancy should occur with a Cu-7 in situ, the Cu-7 should be removed if the thread is visible or, if removal proves to be or would be difficult, interruption of the pregnancy should be considered and offered to the patient as an option... .
(c) Continuation of pregnancy. If the patient chooses to continue the pregnancy and the Cu-7 remains in situ, she must be warned of the increased risk of spontaneous abortion and the increased risk of sepsis, including death. The patient must be closely observed and she must be advised to report immediately all abnormal symptoms, such as flu-like syndrome, fever, abdominal cramping and pain, bleeding, or vaginal discharge, because generalized symptoms of septicemia may be insidious.
... .
Precautions
... .
2. Patient evaluation and clinical considerations. .. .
... .
(1) The Cu-7 should be removed for the following medical reasons: ... pregnancy, if the thread is visible;
... .
Adverse Reactions
[P]regnancy has occurred with the Cu-7 in situ and when the Cu-7 has been partially or completely expelled.
The incidence of spontaneous abortion, when conception occurs with intrauterine devices in situ, appears to be increased over that in unprotected women... . The following complaints have also been reported with IUDs although their relation to the Cu-7 has not been established: ... septic abortion, ... .
In addition, the patient brochure Searle provided for physicians to give to their patients, which Adams acknowledged reading, conveys the following information to the patient:
Adverse Reactions
The following adverse reactions and side effects have been reported with IUDs and may occur after the Cu-7 is inserted:
Pregnancy with the Cu-7 in the uterus.
.....
Spontaneous abortion (miscarriage) if pregnancy occurs with the Cu-7 in the uterus. This possibility is greater if you have a Cu-7 than if you do not have an IUD.
... .
Difficult removal of the Cu-7.
The patient brochure further advised:
Special Warning about Pregnancy with an IUD in Place
... .
When a pregnancy continues with a Cu-7 in the uterus, there is an increased risk that serious complications may result such as sepsis (severe infection) and septic abortion (infected miscarriage) which may lead to death. Most of the cases of serious complications happen in the middle third of pregnancy and occur spontaneously.
Contrary to the majority's view, examination of Searle's Cu-7 warning in the light of MacMurdo and Felix can lead me to reach no other conclusion but that it was sufficient to advise Adams' physician that the precise condition which Adams suffered could result from her use of the Cu-7, and thus was adequate as a matter of law.
Furthermore, in this case as in MacMurdo there was no medical expert testimony that the Cu-7 warning was insufficient to notify physicians of the possibilities that a woman using a Cu-7 nevertheless could become pregnant and that that pregnancy could terminate in a septic abortion. The medical testimony which the plaintiff introduced in opposition to Searle's motion for summary judgment only was directed toward the quality or extent of the Cu-7 warning not to its sufficiency to convey the potential risk of a septic abortion. For instance, Adams' experts testified that the warning only informed physicians of associational or rare incidences of septic abortion, rather than stating a precise cause and effect relationship known by Searle to exist between the Cu-7 and septic abortion, and that the warning did not disclose information available to Searle with respect to the defraying qualities of the Cu-7 string that could cause it to retract into the uterus making removal difficult if not virtually impossible. While this testimony may be relevant in establishing misrepresentations or material omissions by Searle in the warning as alleged in the improvidently dismissed fraud count, it certainly does not defeat a finding as a matter of law that the warning was enough to make physicians aware of the possibility of a septic abortion from the use of the Cu-7.
The majority opinion further fails to address the deposition testimony of Adams' prescribing physician that he had knowledge independent of the Cu-7 warning that the complications sustained by Adams could occur. Dr. Helwig stated he was aware that a small percentage of women using the Cu-7 will become pregnant, and that there was a risk of spontaneous abortion or premature delivery associated with becoming pregnant with an IUD in place. According to Helwig, the risk was "well known" in the medical community.[7] Therefore, based upon the doctor's independent knowledge of those risks, any inadequacy of the warning could not have been the proximate cause of Adams suffering a septic abortion. See Felix, 540 So.2d at 105.
Turning now to the strict liability count and the application of comment k to that count, for reasons which I will explain below, I would adopt a different approach than that taken by the majority opinion. The majority rejects a blanket application of comment k to all prescription drugs and instead adopts a case-by-case risk/benefit analysis to comment k that would require a drug manufacturer to assert comment k as an affirmative defense, leaving it to the jury to make a factual determination whether the benefits of a particurlar drug under the circumstances of that case outweigh or justify the risks attendant with that drug. While I am persuaded in part by the majority's view, I decline to adopt its entire approach to the application of the comment.
To the extent that the majority refuses to adopt a blanket approach to the comment and instead holds that comment k must be raised as an affirmative defense to a strict liability claim to be determined on a case-by-case basis, I concur. Where the majority and I part company is with the portion of its holding that requires the court, and in certain instances the jury, to engage is a risk/benefit analysis to determine whether comment k should apply in a given case. Although there is some support for that view,[8] I believe the more balanced approach to comment k is that undertaken by the federal district courts in Zanzuri, 748 F. Supp. at 1520 (S.D.Fla. 1990) and Amore v. G.D. Searle & Co., 748 F. Supp. 845 (S.D.Fla. 1990), which cite to Florida law supportive of its conclusion.[9] The Zanzuri court held that:
[i]n order to avail itself of the affirmative defense to strict liability provided by Comment k, a defendant must satisfy the following three requisite elements: that the drug be (1) incapable of being made safe; (2) properly prepared and marketed; and (3) accompanied by a proper warning.
Id. at 1520. See also Amore, 748 F. Supp. at 852.
In my opinion, the above analysis is in accord with the express language of the comment and will likely achieve the same public policy results which the majority seeks to obtain through its risk/benefit analysis. For instance, the majority is concerned in this case that Searle may have had knowledge that there was an alternative safer material to the Cu-7 string that may have avoided the memory problem causing a retraction of the string into the uterus, thereby preventing its removal in the event of pregnancy, and that adopting a blanket application of comment k would in essence condone Searle's refusal to use the safer material for the string. That policy concern, however, is addressed by the requirements set out in the above test that the manufacturer, faced with a strict liability claim, must establish that the product is "incapable of being made safe" based upon the present state of knowledge and that the product was "properly prepared and marketed" before it receives the benefit of comment k.
Therefore, applying this three-part test to the instant case, I conclude for the reasons stated that the warning provided by Searle was adequate as a matter of law. However, because the record before us is insufficient to establish as a matter of law that the Cu-7 was incapable of being made safe and was properly prepared and marketed, in that a factual question is at least raised respecting the safeness of the material used for the Cu-7 string, I concur specially with the majority's decision that summary disposition of those issues was inappropriate at this time.
Accordingly, I would reverse in part the summary judgment for Searle and the court's dismissal of the fraud count and remand for the case to proceed on the strict liability and fraud counts consistent with the views I have expressed above.
NOTES
[1] Because the copper in the Cu-7 has effects on the user's body, the FDA classified this device as a prescription drug.
[2] One Florida federal district court, in reversing a summary judgment for Searle, has said: "Far from a model of clarity, the Cu-7 warning is the very example of a qualified warning, the adequacy of which must be resolved through a highly intensive factual inquiry." Zanzuri v. G.D. Searle & Co., 748 F. Supp. 1511, 1516, (S.D.Fla. 1990).
[3] Searle moved for summary judgment on this issue and noticed the motion for hearing, but it appears from the record that the parties ran out of time and were unable to argue the motion to the trial court.
[4] Comment k protects manufacturers from strict liability only for design defects. An injured party may seek strict liability for manufacturing defects or inadequate warnings even though comment k applies. See Toner, 112 Idaho at 336, 732 P.2d at 305.
[5] One court noted that, at the American Law Institute meeting where section 402A and comment k were adopted, the drafters of comment k entertained a proposal to include all prescription drugs within its scope. This proposal was defeated. Hill v. Searle Laboratories, 884 F.2d 1064, 1069 (8th Cir.1989).
[6] This court does not pass judgment upon whether the Cu-7 was entitled to the protection of comment k at any time.
[7] Dr. Helwig stated he informed his patients that the rate of miscarriage when a pregnancy occurred with an IUD in place was fifty to sixty percent if the device remained in situ, and forty percent if the device is removed.
[8] See, e.g., Hill v. Searle Laboratories, 884 F.2d 1064 (8th Cir.1989); Toner v. Lederle Laboratories, 112 Idaho 328, 732 P.2d 297 (1987); Belle Bonfils Memorial Blood Bank v. Hansen, 665 P.2d 118 (Colo. 1983).
[9] In Russell v. Community Blood Bank, Inc., 185 So.2d 749 (Fla. 2d DCA 1966), reversed on other grounds, 196 So.2d 115 (1967) and Mcleod v. W.S. Merrell Co., 174 So.2d 736 (Fla. 1965), it was stated that before applying comment k, it was necessary for the court to find that the product was "unavoidably unsafe."