The manufacture of prescription drugs, a multibillion-dollar industry, has provided many of the 20th *1119century’s greatest success stories and some of its worst tragedies. Because they cure disease, alleviate pain, and prolong life, prescription drugs have been a great benefit to society. But prescription drugs sometimes cause severe complications and side effects, inflicting great anguish as well as temporary and even permanent disability on some individuals.
This court’s task in the present case is to set rules defining prescription drug manufacturers’ tort liability for personal injuries caused by their products’ side effects. More particularly, the task is to determine under what circumstances a drug manufacturer should be held liable in tort for personal injury damages proximately caused by a failure to warn about the possibility of a particular drug complication.
The majority holds that a prescription drug manufacturer is liable in tort for all drug-related injuries about which it did not warn, provided only that the risk of injury was either actually known or in some manner scientifically ascertainable by the manufacturer when it distributed the drug. In his dissent, by contrast, Justice Baxter proposes that a prescription drug manufacturer be held liable for failure to warn of possible drug-related injuries only if the person injured by the drug succeeds in proving that the manufacturer’s decision not to warn was unreasonable.
I find neither approach entirely satisfactory. Combining what I view as the best features of both, I would hold that initially, to establish a prima facie case, a person seeking damages for drug-related injures on a failure-to-wam theory need prove only that at the time of distribution the manufacturer either knew or should have learned, through the application of commonly accepted scientific methods and reasonably available technologies, of the particular risk of harm; but I would hold also that after the party seeking recovery makes this showing, the manufacturer may defend on the basis that it acted reasonably, in light of all relevant considerations, in not warning of the particular risk.
I.
What standard of liability should govern when a consumer is injured by a properly manufactured dmg, but the manufacturer has not warned of the dmg’s association with the type of injury suffered? That is the issue here. Because failure-to-wam liability for injuries caused by a dmg is part of the larger field of products liability law, I begin with a discussion of the development and current state of that body of law.
Under the common law as it existed before the development of products liability law, an injured consumer could recover only under contractual *1120warranty law, which limited recovery to those who “stood in privity of contract”—that is, the immediate purchasers of the defective product. (Prosser & Keeton on Torts (5th ed. 1984) § 96, p. 681.) All too often, however, consumers lacked such privity because they had bought the product not from the manufacturer directly but from intermediaries such as retailers. Disclaimers of any warranty by the manufacturer erected yet another obstacle to recovery. (Ibid.)
The inadequacy of the protection afforded injured consumers under contract law eventually led the courts to extend the existing fault-based tort concept of negligence to the area of products liability. (MacPherson v. Buick Motor Co. (1916) 217 N.Y. 382 [111 N.E. 1050]; see generally, Prosser & Keeton on Torts, supra, § 96, pp. 682-683; Prosser, The Assault Upon the Citadel (Strict Liability to the Consumer) (1960) 69 Yale L.J. 1099, 1100-1103.) Serious impediments to recovery continued to exist, however, even under negligence law. For instance, the injured consumer, because of lack of knowledge about the manufacturing process, was not ordinarily in a position to identify the cause of a defect or to show that the defect had resulted from a manufacturer’s failure to act reasonably. (Prosser, The Assault Upon the Citadel (Strict Liability to the Consumer), supra, 69 Yale L.J. at p. 1116.) To ensure adequate protection to the user of a manufactured product, several legal scholars and jurists, most notably Justice Traynor of this court, adopted the view that a manufacturer should be held liable for defective product injuries regardless of its negligence. (Escola v. Coca Cola Bottling Co. (1944) 24 Cal.2d 453, 462-463, 467 [150 P.2d 436] (conc. opn. of Traynor, J.).)
This concept, referred to as products liability or strict liability in tort and extending not only to the manufacturer but to all others in the chain of distribution, ultimately became the law. (Prosser, The Fall of the Citadel (Strict Liability to the Consumer) (1966) 50 Minn. L.Rev. 791, 791-805.) California was the first to embrace this concept when, in 1963, in the landmark case of Greenman v. Yuba Power Products, Inc. (1963) 59 Cal.2d 57 [27 Cal.Rptr. 697, 377 P.2d 897, 13 A.L.R.3d 1049], this court held: “A manufacturer is strictly liable in tort when an article he places on the market, knowing that it is to be used without inspection for defects, proves to have a defect that causes injury to a human being.” (Id. at p. 62; see generally, Prosser, The Fall of the Citadel (Strict Liability to the Consumer), supra, 50 Minn. L.Rev. at pp. 803-805.)
As this court has acknowledged, “the very purpose of our pioneering efforts in this field was to relieve the plaintiff from problems of proof inherent in pursuing negligence (Escola v. Coca Cola Bottling Co., supra, 24 *1121Cal.2d 453, 461-462 (Traynor, J. concurring)) and warranty (Greenman v. Yuba Power Products, Inc., supra, 59 Cal.2d 57, 63) remedies, and thereby ‘to insure that the costs of injuries resulting from defective products are borne by the manufacturers . . . (Id; see Price v. Shell Oil Co. [1970] 2 Cal.3d 245, 251 [85 Cal.Rptr. 178, 466 P.2d 722].)” (Cronin v. J.B.E. Olson Corp. (1972) 8 Cal.3d 121, 133 [104 Cal.Rptr. 433, 501 P.2d 1153], ellipses in Cronin.)
The concept of products liability then spread nationwide. In 1965, products liability was embraced by the American Law Institute (hereafter ALI) when it adopted section 402A of the Restatement Second of Torts. That section states that a seller of a “product in a defective condition [that is] unreasonably dangerous to the user or consumer” is liable for physical harm caused by the product. (Rest.2d Torts, § 402A, pp. 347-348.) Soon most jurisdictions recognized some form of strict liability for defective products. (Prosser & Keeton on Torts, supra, § 99, p. 694.)
Products liability is not absolute liability, however; a manufacturer is not the insurer of the product. (See, e.g., Anderson v. Owens-Corning Fiberglas Corp. (1991) 53 Cal.3d 987, 1003-1004 [281. Cal.Rptr. 528, 810 P.2d 549]; Daly v. General Motors Corp. (1978) 20 Cal.3d 725, 733 [144 Cal.Rptr. 380, 575 P.2d 1162].) A plaintiff must still prove that the product was defective.
“Strict liability has been invoked for three types of defects—manufacturing defects, design defects, and ‘warning defects,’ i.e., inadequate warnings or failures to warn.” (Anderson v. Owens-Corning Fiberglas Corp., supra, 53 Cal.3d at p. 995; see Barker v. Lull Engineering Co. (1978) 20 Cal.3d 413 [143 Cal.Rptr. 225, 573 P.2d 443, 96 A.L.R.3d 1].) A manufacturing defect occurs when a product does not conform to the manufacturer’s intended design (Brown v. Superior Court (1988) 44 Cal.3d 1049,1057 [245 Cal.Rptr. 412, 751 P.2d 470]); a design defect occurs either when the design of the product fails to meet consumer expectations as to safety or when, on balance, the risk of danger inherent in the challenged design outweighs the benefits of the design (Soule v. General Motors Corp. (1994) 8 Cal.4th 548, 566-567 [34 Cal.Rptr.2d 607, 882 P.2d 298]); a “warning defect” occurs when a manufacturer has issued no warnings or has failed to adequately warn of dangers posed by the product (see Anderson v. Owens-Corning Fiberglas Corp., supra, 53 Cal.3d at p. 995). The issue here pertains to the third of these categories—liability for a prescription drug manufacturer’s failure to warn.
Courts and commentators have long noted that prescription drugs pose unique products liability issues, and that expansive liability for drug-related *1122injuries could deter manufacturers from developing and marketing medical drugs of benefit to society. During the discussion of a draft of what later became section 402A of the Restatement Second of Torts, a member of the ALI proposed that drugs be expressly exempted from section 402A because imposing products liability on drugs might stifle medical research and testing. (Brown v. Superior Court, supra, 44 Cal.3d at pp. 1057-1058, quoting 38 ALI Proc. 19, 90-98 (1961).) In response, Dean Prosser, who was the reporter for the Restatement Second of Torts, suggested that the issue be dealt with in a comment to section 402A. (44 Cal.3d at p. 1058.) This led to the drafting, and approval, of comment k. (See Comment, Comment K Immunity to Strict Liability: Should all Prescription Drugs be Protected? (1989) 26 Hous. L.Rev. 707, 736 [“the beneficial nature of prescription drugs provides the very reason for excepting them from standard rules of strict liability”].)
Comment k to section 402A of the Restatement Second of Torts recognizes the significant risk to the public interest presented by imposing excessive liability on the manufacturer of prescription drugs. Under the heading “Unavoidably unsafe products,” comment k observes that commonly used drugs, while highly beneficial, are often incapable of being made entirely safe. It then notes: “It is also true in particular of many new or experimental drugs as to which, because of lack of 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.” (Rest.2d Torts, § 402A, com. k, p. 354.) It concludes that a manufacturer of drugs that are sold with a proper warning “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.” (Ibid.)
Courts and commentators have debated the exact scope and meaning of comment k to section 402A of the Restatement Second of Torts. (See, e.g., Schwartz, Unavoidably Unsafe Products: Clarifying the Meaning and Policy Behind Comment K (1985) 42 Wash. & Lee L.Rev. 1139 [discussing various possible interpretations of comment k]; Henderson & Twerski, A Proposed Revision of Section 402A of the Restatement (Second) of Torts (1992) 77 Cornell L.Rev. 1512, 1537, 1542 [criticizing comment k for lack of clarity].) This court has concluded that “in fact the principle it [comment k] states is based on negligence.” (Brown v. Superior Court, supra, 44 Cal.3d at p. 1059.)
The negligence-based standard of comment k to section 402A of the Restatement Second of Torts reflects a recognition that imposing strict *1123liability on manufacturers of prescription drugs would harm the public interest by discouraging the development and availability of medications that cure diseases, alleviate pain and suffering, or sustain life. In the words of Dean Prosser: “The argument that industries producing potentially dangerous products should make good the harm, distribute it by liability insurance, and add the cost to the price of the product, encounters reason for pause, when we consider that two of the greatest medical boons to the human race, penicillin and cortisone, both have their dangerous side effects, and that drug companies might well have been deterred from producing and selling them.” (Prosser, Torts (4th ed. 1971) § 99, p. 661, fns. omitted.)
This concern—not to impose an excessively broad standard of liability on prescription drug manufacturers—continues today. Thus, the current tentative draft of the Restatement Third of Torts contains a special section governing products liability for prescription drugs and medical devices. With regard to warning defects, section 8 provides: “(d) A prescription drug or medical device is not reasonably safe because of inadequate instructions or warnings when [^D (1) reasonable instructions or warnings regarding foreseeable risks of harm posed by the drug or medical device are not provided to prescribing and other health care providers . . . .” (Rest.3d Torts: Products Liability (Tent. Draft No. 2, Mar. 13, 1995) § 8, p. 210.)1
This court has in recent years addressed liability for “warning defects” for products other than prescription drugs (Anderson v. Owens-Corning Fiberglas Corp., supra, 53 Cal.3d 987 [evidence of whether risk known or knowable admissible in failure-to-wam case not involving prescription drugs]) as well as liability for prescription drug “design defects” (Brown v. Superior Court, supra, 44 Cal.3d 1049 [prescription drug manufacturers exempt from strict liability for design defect claims]). But this court has not previously defined the elements necessary to establish a drug manufacturer’s liability for failure to warn of a risk presented by a prescription drug.
To summarize, products or strict liability law was developed to compensate consumers injured by defective products without so expanding liability as to discourage manufacturers from developing and marketing products beneficial to the public. General principles of products liability law appropriately accommodate these competing interests for most types of products. *1124As will be seen, prescription drugs require a standard of liability and a burden-of-proof allocation somewhat different from those used for other products. In this case, neither the majority nor the dissent has adequately accommodated two different consumer interests at stake.
II.
At first glance, the majority’s holding—that a prescription drug manufacturer need warn only of those risks that are “known or reasonably scientifically knowable” (maj. opn., ante, at p. 1109)—appears quite reasonable. A close look, however, reveals that the majority imposes on a manufacturer a duty to warn of any risk that arguably may exist. The sheer breadth of that duty threatens two fundamental public interests. First, the majority’s holding will result in the problems of overwaming this court has previously recognized, thereby undermining the public interest in consumer protection. Second, by subjecting prescription drug manufacturers to excessive liability, the majority’s standard jeopardizes the important public interest of encouraging the development, availability, and affordability of beneficial prescription drugs.
To discern what the majority means by the phrase “known or reasonably scientifically knowable” is no easy task. Throughout its opinion, the majority itself gives different descriptions of the meaning of that phrase (see maj. opn., ante, at p. 1108 [“known to the scientific community”]; id. at p. 1112 [“ ‘generally recognized and prevailing best scientific and medical knowledge available’ ”]; id. at p. 1113, fn. 3 [“knowledge obtainable ‘ “by the application of reasonable, developed human skill and foresight” ’ ”]; ibid. [“ ‘true choice judgment’ ”]; id. at p. 1115 [“reasonable scientist operating in good faith should have known under the circumstances of the evidence”]; id. at p. 1116 [“evidence established a casual link between an alleged side effect and a prescription drug”]; ibid, [“knowable by the application of scientific and medical knowledge available”]).
The majority’s various characterizations of “knowability” lead me to conclude that the majority considers a risk to be “knowable,” even though not actually known to the manufacturer, if the risk at the time of distribution had been or reasonably could have been scientifically identified. The word “risk” simply means the possibility of a loss. (Webster’s New Collegiate Dict. (9th ed. 1988) p. 1018, col. 1.) Because the majority does not suggest that the degree of probability of loss is of any significance, the majority seemingly equates the identification of a bare possibility of harm with “knowability” and hence a duty to warn. Thus, under the majority’s holding, once a prescription drug plaintiff proves that a risk has been or reasonably *1125could have been identified by the scientific community, the manufacturer has a duty to warn regardless of whether warning is otherwise appropriate or reasonable.2
The majority insists that its conclusion is dictated by our decisions in Brown v. Superior Court, supra, 44 Cal.3d 1049, and Anderson v. Owens-Corning Fiberglas Corp., supra, 53 Cal.3d 987. (Maj. opn., ante, at pp. 1108-1109,1116-1117.) But, as Justice Baxter convincingly demonstrates in his dissent, the majority is wrong. (Dis. opn. of Baxter, J., post, at pp. 1153-1158.) This court has not previously decided this issue.
In my view, the majority’s holding goes too far in imposing liability, because it fails to recognize the complexity involved in scientifically identifying a risk as meaningful and in determining whether a warning is appropriate. As this court has noted, the quality of scientific evidence “may range from extremely vague to highly certain.” (Finn v. G. D. Searle & Co. (1984) 35 Cal.3d 691, 701 [200 Cal.Rptr. 870, 677 P.2d 1147].) Scientific studies suggesting associations between products and injuries may themselves be subjected to legitimate question as to the validity of their methods and the soundness of their conclusions. (See Ramirez v. Plough, Inc. (1993) 6 Cal.4th 539, 556 [25 Cal.Rptr.2d 97, 863 P.2d 167, 27 A.L.R.5th 899] [explaining that in 1986, methodology of scientific studies showing an association between aspirin and Reye’s syndrome had been questioned, and the Food and Drug Administration had determined that further studies to confirm or disprove association were necessary]; see generally, Foster et al., Phantom Risk: Scientific Inference and the Law (1993).) The majority’s apparent standard that a prescription drug manufacturer is strictly liable for failure to warn of any “knowable” risk fails to recognize, much less deal with, the complexity of scientific evaluations.
A possible result of such a standard is the destruction of the viability of any warnings. “[B]oth common sense and experience suggest that if every *1126report of a possible risk, no matter how speculative, conjectural, or tentative, imposed an affirmative duty to give some warning, a manufacturer would be required to inundate physicians indiscriminately with notice of any and every hint of danger, thereby inevitably diluting the force of any specific warning given. [Citations.] The strength of the causal link thus is relevant both to the issue of whether a warning should be given at all, and, if one is required, what form it should take.” (Finn v. G. D. Searle & Co., supra, 35 Cal.3d at p. 701.) Too broad a standard may prove ineffective and even counterproductive. (See Ramirez v. Plough, Inc., supra, 6 Cal.4th at p. 553; Twerski et al., The Use and Abuse of Warnings in Products Liability—Design Defect Litigation Comes of Age (1976) 61 Cornell L.Rev. 495, 513-517; Prosser & Keeton on Torts, supra, § 96, p. 686 [“Too much detail can be counterproductive. A warning to be effective must be read and understood.”].)
The problems of overwaming are exacerbated if warnings must be given even as to very remote risks, which drug manufacturers may find necessary under the vague standard that the majority adopts today. “Not only would such remote risk warnings crowd out potentially useful warnings but they would also focus consumer attention on the fairy tale bogeyman. One cannot cry wolf without paying the price over the long term.” (Henderson & Twerski, Doctrinal Collapse in Products Liability: The Empty Shell of Failure to Warn (1990) 65 N.Y.U. L.Rev. 265, 318.)
According to the majority, its standard of liability would not result in inundating consumers with too many warnings, because the manufacturer’s duty to warn runs to the physician rather than the patient, and because there is no duty to warn of obvious dangers. (Maj. opn., ante, at p. 1116.)
It is true, as the majority states, that a prescription drug manufacturer has no duty to ensure that a warning given a physician reaches the patient. (Stevens v. Parke, Davis & Co. (1973) 9 Cal.3d 51, 65 [107 Cal.Rptr. 45, 507 P.2d 653, 94 A.L.R.3d 1059].) But even physicians, the “learned intermediaries,” may be overwhelmed by excessive warnings. As this court has recognized, both common sense and experience suggest that to require a manufacturer “to inundate physicians indiscriminately with notice of any and every hint of danger” inevitably dilutes “the force of any specific warning given.” (Finn v. G. D. Searle & Co., supra, 35 Cal.3d at p. 701; contra, maj. opn., ante, at p. 1116, fn. 6.) And, contrary to the majority’s assertion, the absence of a duty to warn of known or obvious dangers has little import in the context of prescription drugs, because few of the sophisticated potential risks of prescription drugs could be considered obvious.
The majority’s holding exposes prescription drug manufacturers to such broad liability that they may restrict or cease the development and distribution of life-sustaining and lifesaving drugs, thereby defeating a strong public *1127interest. As this court has observed: “[TJhere is an important distinction between prescription drugs and other products such as construction machinery [citations], a lawnmower [citation], or perfume [citation], the producers of which were held strictly liable. In the latter cases, the product is used to make work easier or to provide pleasure, while in the former it may be necessary to alleviate pain and suffering or to sustain life. Moreover, unlike other important medical products (wheelchairs, for example), harm to some users from prescription drugs is unavoidable. Because of these distinctions, the broader public interest in the availability of drugs at an affordable price must be considered in deciding the appropriate standard of liability for injuries resulting from their use.” (Brown v. Superior Court, supra, 44 Cal.3d at p. 1063, italics added.) This distinction between prescription drugs and other products, the legal significance of which the majority denies (see maj. opn., ante, at p. 1116, fn. 5), is firmly rooted in products liability law and was the basis for this court’s unanimous decision in Brown v. Superior Court, supra, at pages 1063-1065.3
The concern repeatedly voiced by the courts and legal commentators that the imposition of excessive liability on prescription drug manufacturers may discourage the development and availability of life-sustaining and lifesaving drugs is well founded, as the following discussion demonstrates.
In 1976, because the threat of excessive tort liability was deterring drug manufacturers from developing a vaccine for swine flu needed to protect public health, Congress enacted the Swine Flu Act, under which the federal government assumed the risk of lawsuits arising from injuries associated with the vaccine. (Pub.L. No. 94-380 (Aug. 12, 1976) §2, 90 Stat. 1113; Franklin & Mais, Tort Law and Mass Immunization Programs: Lessons from the Polio and Flu Episodes (1977) 65 Cal.L.Rev. 754, 769; see Brown v. Superior Court, supra, 44 Cal.3d at p. 1064.)
In 1983, Benedictin, the only antinauseant drug available to pregnant women, was withdrawn from the market because the cost of insurance almost consumed the entire income from the sale of the drug and its price *1128had already increased by over 300 percent. (Brown v. Superior Court, supra, 44 Cal.3d at p. 1064.)
In the mid-1980’s, a producer of the diphtheria-pertussis-tetanus (DPT) vaccine withdrew its product from the market because of “extreme liability exposure, costs of litigation and the difficulty of continuing to obtain adequate insurance.” (Brown v. Superior Court, supra, 44 Cal.3d at p. 1064, quoting Hearing on Vaccine Injury Compensation Before Subcom. on Health and the Environment of House Com. on Energy and Commerce, 98th Cong., 2d Sess. (Sept. 10, 1984) p. 295.) By 1986, there were only two manufacturers of the DPT vaccine still marketing the drug and its costs had risen from 11 cents per dose in 1982 to $11.40 per dose in 1986, $8 of which was for insurance reserves. (Brown v. Superior Court, supra, 44 Cal.3d at p. 1064. ) Incidentally, that same year the unavailability of insurance prevented a manufacturer from marketing a new drug to treat vision problems. (Id. at p. 1065. )
In 1990, it was reported that the threat of products liability litigation had left the United States with just one manufacturer of the polio vaccine and one manufacturer of the combined measles, mumps and rubella vaccine. (Remarks of Sen. Heinz, 136 Cong. Rec. S16661, SI6662 (Oct. 25, 1990).)
More recently, the imposition of too harsh a standard for products liability on manufacturers of prescription drugs has been mentioned as a possible impediment to this country’s ability to combat the serious public health threat posed by Acquired Immune Deficiency Syndrome (AIDS). (See generally, Arnold, Developing, Testing, and Marketing an AIDS Vaccine: Legal Concerns for Manufacturers (1991) 139 U. Pa. L.Rev. 1077.)
To require prescription drug manufacturers to warn of all scientifically knowable risks, as the majority holds, exposes them to excessive liability. Such exposure creates a powerful disincentive on the part of prescription drug manufacturers to develop essential medications, especially when it is joined with the extensive liability exposure for alleged inadequacy of warnings given. (See, e.g., MacDonald v. Ortho Pharmaceutical Corp. (1985) 394 Mass. 131 [475 N.E.2d 65, 71] [adequacy of warning that use of oral contraceptive may cause abnormal blood clotting in such vital organs as the brain and that the abnormal clotting may be fatal presented question of fact because it did not use the word “stroke”]; see also Henderson & Twerski, Doctrinal Collapse in Products Liability: The Empty Shell of Failure to Warn, supra, 65 N.Y.U. L.Rev. at pp. 318-319.)
In short, the broad standard of liability to which the majority subjects manufacturers of prescription drugs may cause manufacturers to issue so *1129many warnings that the effectiveness of the most important warnings will be negated, and it jeopardizes the vital public interest in the development and availability of life- and health-sustaining drugs.
Although the latter interest is advanced by the dissent’s proposed rule, the dissent’s approach places too heavy a burden on the injured consumer.
III.
Justice Baxter’s dissent expresses the view that a prescription drug manufacturer’s liability for failure to warn should be based solely upon negligence principles. Under that approach, a manufacturer is liable for failure to warn only if the plaintiff establishes that a reasonable manufacturer in similar circumstances would have issued warnings. (Dis. opn. of Baxter, J., post, at pp. 1147, 1162; see 6 Witkin, Summary of Cal. Law (9th ed. 1988) Torts, § 729, pp. 56-58.)
The dissent agrees with the majority that the legal theories of negligent failure to warn and strict liability failure to warn are two different theories of liability: negligence looks to the reasonableness of a manufacturer’s conduct while strict liability considers it irrelevant. (Dis. opn. of Baxter, J., post, at pp. 1157-1158.) According to the dissent, the same policy considerations that led this court in Brown v. Superior Court, supra, 44 Cal.3d 1049, to adopt a negligence standard for prescription drug design defects apply with equal force when, as here, the alleged defect is a failure to warn of risks associated with the drug. (Dis. opn., post, at p. 1158.)
The dissent cites four reasons for its conclusion. First, it mentions that several commentators have concluded that a product’s warning may be characterized as a subset of the product’s design. (Dis. opn. of Baxter, J., post, at p. 1158.) Second, the dissent points out that products liability law has always recognized that prescription drugs may require different legal treatment than other products, and that the tentative draft of the Restatement Third of Torts incorporates negligence principles into its new standard for prescription drug manufacturer products liability. (Dis. opn., post, at pp. 1159- 1160.) Third, it notes that the prescription drug industry is heavily regulated by the federal Food and Drug Administration (FDA), and that one of the FDA’s primary functions is to evaluate drug warnings. (Id. at pp. 1160-1161.) Fourth, the dissent points out that the imposition of strict liability for failure to warn on prescription drug manufacturers would inhibit and delay the development and marketing of essential drugs. (Id. at p. 1162.)
The negligence standard advocated by the dissent preserves the public interest in the development and availability of life-sustaining and lifesaving *1130drugs because prescription drug manufacturers would be liable only if their decision not to warn was unreasonable. But the dissent’s approach would unnecessarily weaken the legal protection afforded injured consumers under products liability law.
As mentioned at the outset, a principal objective of products liability law is to relieve injured consumers of the problems of proof inherent in negligence and contractual warranty, common law concepts that were initially applied to consumers injured by defective products. (Ante, at pp. 11191120.) The dissent’s approach runs counter to this important goal.
To prove a prescription drug manufacturer was negligent in failing to warn, an injured consumer would have to establish not only that the risk was scientifically appreciable and valid, but also, for example, that the risk was of sufficient magnitude in relation to other risks as to render it unreasonable not to warn, that the risk was not too remote, and that under all of the circumstances a prudent manufacturer would have issued a warning. Thus, the consumer would have to show not only that the risk at issue was scientifically credible, but also that its relative severity and probability in relation to other scientifically credible risks made the decision not to warn unreasonable.
Having developed and manufactured the product that injured the consumer, the manufacturer rather than the consumer is the one with superior access to the pertinent scientific data. Yet, the negligence standard proposed by the dissent would saddle the consumer with the burden of obtaining and evaluating the relevant scientific information.
In addition, under the dissent’s approach, the injured consumer would have to prove that at the time of distribution there were no other factors that would make the manufacturer’s decision not to warn reasonable. The other factors would involve “a balancing of the seriousness of the harm, the probability that the harm will result, and the burden involved in taking the necessary precautions.” (Note, Formulating the Strict Liability Failure to Warn (1992) 49 Wash. & Lee L.Rev. 1509, 1512, fn. omitted.) A manufacturer that is charged with possessing the requisite scientific information, that is in the business of developing and marketing prescription drugs, and that deals directly and on an ongoing basis with the FDA in evaluating the propriety and extent of warnings (see, e.g., 21 C.F.R. §§ 201.57 (1996), 310 et seq. (1995), 312 et seq. (1995), 314 et seq. (1995)), is far better situated to address the issues relating to the reasonableness of a decision whether to warn of a particular risk. An injured consumer, by comparison, does not have direct access to the relevant information and cannot reasonably be *1131expected to have the resources necessary to satisfy the burden of proving negligence.
Accordingly, in my view, adoption of a negligence standard, as the dissent advocates, undermines the consumer protection goals underlying products liability law.
IV.
As explained above, there are serious drawbacks to the majority’s strict liability standard and the dissent’s negligence approach. Imposition of strict liability on prescription drug manufacturers may curtail the development of drugs beneficial to society, substantially increase the cost to the manufacturer and consequently the price charged to the consumer, and lead to the issuance of so many warnings as to negate their effectiveness. The dissent, on the other hand, by requiring an injured consumer to prove that a manufacturer’s failure to warn was negligent or unreasonable, would saddle the consumer with an impossibly high burden of proof. In short, the majority’s strict liability standard is too harsh towards the manufacturer of prescription drugs, while the dissent is too harsh towards the injured consumer.
I propose an intermediate approach, one that fairly accommodates the competing public policies. (See Schwartz, Prescription Products and the Proposed Restatement (Third) 61 Tenn. L.Rev. 1357, 1359 [pharmaceutical product’s ability to reduce pain, save lives, and restore health should be balanced with its potential for producing serious injuries].) Unlike the majority, I would not impose liability on prescription drug manufacturers for their failure to warn of every arguable risk, but only of those risks supported by credible scientific evidence or that upon reasonable inquiry would be supported by credible scientific evidence. Unlike the dissent, I would not require the plaintiff consumer to show as part of the plaintiff’s case that the manufacturer acted unreasonably; instead, I would allocate to the manufacturer the burden of proving that its decision not to warn was reasonable.
As I discussed earlier, there are two competing public policies at stake, each advancing a distinct and important consumer interest. On the one side, there are the interests that gave rise to the development of products liability law: to compensate consumers injured by deficient products and to encourage manufacturers to increase product safety. On the other side, there is society’s interest in affordable prescription drugs, many of which are life sustaining and lifesaving. The majority’s holding partially serves the first of these public policies by imposing a broad standard of liability, but it does so by jeopardizing the effectiveness of the warnings given and by undermining *1132the public interest in ensuring the development and availability of beneficial drugs. Although the dissent advances the latter policy, it weakens the protection of injured consumers by imposing on them too onerous a burden of proof. Thus, both the majority and the dissent sacrifice an important policy to the detriment of another equally important policy.
A balance of these important public policies may be achieved by a hybrid or intermediate approach that borrows the best features from strict liability as well as negligence. This approach has two steps: (1) The plaintiff would have the burden of establishing a prima facie case by proving that, at the time of distribution of the drug, scientifically credible evidence had identified the risk, or that upon reasonable inquiry the risk would have been identified by scientifically credible evidence; (2) then the burden of proof would shift to the manufacturer to show that its decision not to warn was reasonable.
As to the first step, I find guidance in previous decisions of this court. In determining the admissibility of new scientific techniques, this court has held that evidence of a technique is admissible only if it has gained acceptance in the particular scientific field to which it belongs. (People v. Kelly (1976) 17 Cal.3d 24, 30 [130 Cal.Rptr. 144, 549 P.2d 1240] [voiceprint analysis].) Acceptance of the technique for purposes of admissibility is established if the proponent of the evidence shows that a cross-section of the relevant scientific community accepts the technique as scientifically credible. (See People v. Leahy (1994) 8 Cal.4th 587, 612 [34 Cal.Rptr.2d 663, 882 P.2d 321] [horizontal gaze nystagmus]; People v. Shirley (1982) 31 Cal.3d 18, 56 [181 Cal.Rptr. 243, 723 P.2d 1354] [pretrial hypnosis of witnesses].) Acceptance does not require numerical majority support; in determining whether a technique has attained the requisite acceptance, both the quality and the quantity of the supporting and opposing evidence must be considered. (People v. Leahy, supra, at p. 612.)
Similarly, I would require the plaintiff in a prescription drug case to show that at the time of the manufacturer’s distribution of the drug evidence of a risk of harm from the drug was considered credible in the relevant scientific community or that a reasonable scientist would have investigated the possibility of a risk and the investigation would have resulted in scientifically credible evidence of the risk’s existence. Evidence of a risk would be scientifically credible if the data upon which it is based, the methodology employed, and its conclusions identifying the existence of a risk comply *1133with generally accepted scientific methodology and analysis.4 Scientific evidence that postulates the possibility of a risk or that is otherwise speculative or conjectural would be inadequate. Also, the relevant inquiry relates to the credibility of the scientific evidence in light of accepted scientific norms, not to the personal professional beliefs or preferences of an otherwise qualified expert.
The plaintiff’s identification of such a risk would not by itself subject the manufacturer to liability for failure to warn of the risk associated with the drug prescribed. Against the benefits that may be gained by a warning must be balanced the dangers of overwaming and of less meaningful warnings crowding out necessary warnings, the problems of remote risks, and the seriousness of the possible harm to the consumer. But requiring a plaintiff to prove that these factors supported the giving of a warning would force the plaintiff to independently obtain and evaluate all relevant scientific information bearing on these factors. Because such information is not readily accessible to plaintiffs, who may also lack the resources necessary to analyze and evaluate the information, it would impose too heavy a burden on plaintiffs to require them to prove that a warning should have been given because of factors such as the magnitude of the risk in relation to other risks, the seriousness of the harm, and the probability of the harm. Therefore, I would shift to the manufacturer the burden of proving justification for the failure to warn of the risk established by the consumer’s prima facie case. (See Evid. Code, §§ 605, 606.)
To satisfy its burden, the manufacturer would have to show that its failure to warn was reasonable in relation to the identified risk. The manufacturer, for instance, could introduce evidence that the risk did not pose a serious threat to health, that it was remote, that the number or relative severity of other risks justified a failure to warn, or that the scientific association between the drug and the risk was weak. The manufacturer’s compliance with product safety statutes or regulations such as those of the FDA would also be relevant, but not necessarily controlling.5 The same is true of industry standards and practices to the extent they relate to the reasonableness of the manufacturer’s decision not to warn. (See Prosser & Keeton on Torts, supra, § 33, pp. 193-196.)
*1134In my view, this allocation of the burden of proof is appropriate given the manufacturer’s superior access to and capability of evaluating the relevant scientific information; it also furthers a goal of products liability law of relieving injured consumers from evidentiary burdens that may be too onerous. (See Barker v. Lull Engineering Co., supra, 20 Cal.3d at p. 431 [holding that once a plaintiff makes a prima facie showing of injury caused by product’s design, the burden should shift to the defendant to prove that the product is not defective].)
Is this solution perfect? Perhaps not. But it does attempt to strike a fair balance between two distinct public interests: compensating consumers injured by defective products, and encouraging the development of prescription drugs, many of which are life sustaining and lifesaving.
I now turn to the facts of this case.
V.
Here, plaintiff Wilma Peggy Carlin sued defendant the Upjohn Company, among others, for injuries allegedly caused by taking Halcion, a drug manufactured by defendant and prescribed by plaintiff’s physician. Plaintiff alleges that Halcion is intended for the treatment of, among other things, insomnia, and that she ingested the drug from 1987 through November 1992. She asserts that Halcion caused her emotional, physical, and psychic injuries, and that defendant knew or should have known that the drug could cause these injuries.
The complaint attempts to state causes of action for strict liability for failure to properly prepare and warn of the dangerous propensities of Halcion, negligence, breach of warranty, failure to warn, and fraud. Citing Brown v. Superior Court, supra, 44 Cal.3d 1049, defendant manufacturer demurred to the causes of action for strict liability and breach of warranty, arguing that such causes of action cannot be maintained against a prescription drug manufacturer. The trial court sustained defendant’s demurrer without leave to amend to the causes of action for strict liability and breach of warranty. Thereafter, the Court of Appeal granted plaintiff’s petition for a writ of mandate and directed the trial court to overrule the demurrer, citing this court’s decision in Anderson v. Owens-Corning Fiberglas Corp., supra, 53 Cal.3d 987, as permitting a plaintiff in a prescription drug manufacturer case to maintain a cause of action in strict liability for a failure to warn of a known or knowable risk.
Although plaintiff’s cause of action is more properly termed “products liability for failure to warn,” her complaint contains allegations sufficient to *1135withstand a demurrer. Under applicable law, a demurrer may not be sustained if the complaint states facts sufficient to constitute a cause of action. (See, e.g., Blank v. Kirwan (1985) 39 Cal.3d 311, 318 [216 Cal.Rptr. 718, 703 P.2d 58].) Here, plaintiff alleged that defendant manufacturer knew or should have known of Halcion’s dangerous propensities at the time of its distribution. These allegations are sufficient to state a prima facie case.
Conclusion
The continuing debate over the proper standard of liability for prescription drug warning defects has largely been framed as though the only available alternatives were strict liability and negligence. Both the majority and the dissent in this case address the issue from within that limited framework. For the reasons stated above, however, I am convinced that the framework must be discarded in order to craft a standard of liability that achieves a just resolution of competing public policies.
My approach favors neither party at the expense of the other, as the majority and the dissent do. Instead, my approach seeks to strike a balance between two quite distinct consumer interests: providing compensation to the injured consumer, and encouraging the development and availability of affordable health- and life-sustaining prescription drugs.
Although I disagree with the majority’s standard of liability, I join the majority’s disposition affirming the judgment of the Court of Appeal, which directs the trial court to vacate its order sustaining the demurrer and to enter a new order overruling the demurrer.
The tentative draft of the Restatement Third of Torts may be part of a general retrenchment of products liability law. As one commentator observes: “In the past decade, however, further judicial expansion of tort and products liability law largely has come to an end. ‘A new tort era’ has been entered that has been marked by ‘the stabilization and contraction of [tort] doctrine,’ rather than its growth and development. Judicial opinions in the products liability field reflect growing concerns that the liability system is affecting the costs and availability of products adversely.” (Schwartz, The Impact of the New Products Liability Restatement on Prescription Products (1995) 50 Food & Drug L.J. 399, 400, brackets in Schwartz, fns. omitted.)
On page 1116 of its opinion, the majority states: “The application of the failure-to-wam theory to pharmaceuticals requires determinations whether available evidence established a causal link between an alleged side effect and a prescription drug, whether any warning should have been given, and, if so, whether the warning was adequate. These are issues of fact involving, inter alia, questions concerning the state of the art, i.e., what was known or reasonably knowable by the application of scientific and medical knowledge available at the time of manufacture and distribution of the prescription drug. They also necessarily involve questions concerning whether the risk, in light of accepted scientific norms, was more than merely speculative or conjectural, or so remote and insignificant as to be negligible.” (Italics added.) These statements only further obscure what the majority means by the phrase “known or reasonably scientifically knowable.” The italicized portions of the sentences quoted above are either yet another way of expressing the majority’s “knowability” standard that the duty to warn does not extend to unknowable risks or a concession that the standard of liability includes reasonableness of the manufacturer’s conduct. (But see maj. opn., ante, at p. 1112 [“ ‘the reasonableness of the defendant’s failure to warn is immaterial’ ”].)
In footnote 5 at page 1116 of its opinion, the majority characterizes my view as suggesting that the “costs” of prescription drug liability have been excessive and then asserts that it is unclear how my approach would reduce “those costs—unless the inevitable result would be to deprive an innocent injured person of any remedy.” It is true that under my approach not every person injured by the side effects of a prescription drug would recover. But limiting somewhat the scope of liability for failure to warn of risks associated with prescription drugs, as I would do, will best assure that millions of other innocent people—those suffering from debilitating or even fatal diseases—will have available to them prescription drugs that sustain life or health. Moreover, in requiring the plaintiff to prove that the risk was known or reasonably scientifically knowable, the majority’s holding too will “deprive an innocent injured person of any remedy.”
A manufacturer may have a duty to warn even in the absence of credible scientific evidence as defined here if the plaintiff can establish that a reasonable manufacturer would have warned.
I agree with the majority that a prescription drug manufacturer cannot be held liable for failure to warn if it was precluded from issuing a warning by the EDA. (Maj. opn., ante, at p. 1115 & fn. 4.) A contrary conclusion would be unreasonable because it would make manufacturer compliance with both federal and state laws impossible. I also agree with the majority that FDA regulations and action or inaction would be admissible as probative of whether a manufacturer should have warned of a risk. (Id. at p. 1114.)