This is an interlocutory appeal from a district court order under Fed.R.Civ.P. 23 conditionally certifying a nationwide plaintiff class and subclass in a products liability case against the manufacturer of a drug used for the treatment of epilepsy. The jurisdiction of the district court was grounded on diversity, and our jurisdiction is pursuant to certification under 28 U.S.C. § 1292(b).

The drug in question, known as Felbatol, is manufactured by defendants Carter-Wallace, Inc. and Wallace Laboratories (Carter-Wallace). Carter-Wallace began marketing the drug in August 1993 without giving any special warning of serious side effects. Between January 1994 and July 1994, Carter-Wallace received reports that some patients had developed aplastic anemia following use of the drug.^[1] In August 1994, Carter-Wallace mailed letters to the physician community warning them of this risk. By September 1994, Carter-Wallace had also received reports of liver failure in connection with use of the drug. Again, Carter-Wallace mailed letters to the physician community warning them of this risk.

The district court determined that the prerequisites of Fed.R.Civ.P. 23(a) had been met.^[2] The district court conditionally certified a plaintiff class consisting of "all persons who began using Felbatol prior to August 1, 1994." The district court also certified a "serious injury" subclass, defined as "all persons within the Felbatol user class who have developed or will develop aplastic anemia or liver failure, as a result of using Felbatol."

Pursuant to Fed.R.Civ.P. 23(c)(4)(A), the district court limited class certification to the issues of strict liability, negligence, failure to warn, breach of implied and express warranty, causation in fact, and liability for punitive damages. The district court stated that "[w]ith respect to these particular issues, common questions of law and/or fact predominate over any questions affecting only individual members and a class action is superior to other available methods for adjudication of the controversy." The court's order thus echoed the preponderance and superiority requirements of Fed.R.Civ.P. 23(b)(3).^[3] The court specifically excluded the individual issues of proximate causation, compensatory damages, and the amount of punitive damages from certification.

In its certification order, the court did not discuss whether the adjudication of the certified issues would significantly advance the resolution of the underlying case, thereby achieving judicial economy and efficiency. Nor did the court discuss any alternative methods for adjudicating these claims.

According to the named plaintiffs, during the brief period involved in this litigation the drug was prescribed to over 100,000 patients, who were told that the drug was unlike other anti-epilepsy drugs in that this one had few adverse side effects. Plaintiffs claim that over 3,000 people have reported some adverse reactions from the drug to the United States Food & Drug Administration, and there have been over seventy reported cases of aplastic anemia or liver damage, including nearly twenty reported deaths. Withdrawal from the drug has also been difficult for many patients.

Plaintiffs contend, with considerable justification, that because the case involves only one manufacturer, only one product, only one marketing program, and a relatively short period of time, the case is more manageable for class action purposes than cases that involve multiple manufacturers, multiple products, multiple marketing programs, and a long period of time. It appears undisputed that the claims of all members of the class will raise some common issues concerning the knowledge and conduct of Carter-Wallace. Apparently, in recognition of these common issues, the Judicial Panel on Multidistrict Litigation (JPML) has consolidated pretrial proceedings in all federal Felbatol cases and transferred them to the Northern District of California.

Carter-Wallace argues, with at least equal justification, that the existence of common issues of law or fact is a necessary but not the sole requirement for class certification, and that the class certified here does not meet other Rule 23 requirements. Carter-Wallace places particular stress on the Rule 23(b)(3) requirements that the common issues of fact predominate over individual issues and that the class action be superior to other methods of adjudicating the claims. Specifically, Carter-Wallace contends that the numerous adverse reactions of each plaintiff are intertwined with the certified liability issues, and that the law on each liability theory varies widely from state to state. Additionally, Carter-Wallace notes that the problems with the numerous adverse reactions affect the Rule 23(a) prerequisites of typicality and adequacy of representation in that the drug has had a variety of different effects on different people and further, that the class does not contain any representative who has allegedly developed aplastic anemia from taking the drug. Carter-Wallace also contends that class adjudication will be unmanageable and inefficient and that alternative, superior methods of adjudication exist.

Carter-Wallace's threshold contention in this appeal is, however, even more sweeping. It is that, regardless of any specific problems with this particular certification, class certification is never appropriate for multi-state plaintiffs asserting personal injury claims against manufacturers of drugs and medical devices. Carter-Wallace cites this circuit's opinion in In re Northern Dist. of California, Dalkon Shield IUD Prods. Liab. Litig. (Dalkon Shield ), 693 F.2d 847, 854-55 (9th Cir.1982), cert. denied, 459 U.S. 1171, 103 S.Ct. 817, 74 L.Ed.2d 1015 (1983), and recent cases from other circuits to support its broadside attack. See, e.g., Castano v. American Tobacco Co., 84 F.3d 734 (5th Cir.1996); In re American Medical Sys., 75 F.3d 1069 (6th Cir.1996); In re Rhone-Poulenc Rorer, Inc., 51 F.3d 1293 (7th Cir.), cert. denied, --- U.S. ----, 116 S.Ct. 184, 133 L.Ed.2d 122 (1995). Our review of the record suggests that a principal reason why the district court entered twin certifications, first to create class litigation, and then to secure appellate review of that creation, was to obtain a ruling from this court on whether the law of this circuit supports Carter-Wallace's threshold position.

We hold that the law of this circuit, and more specifically our leading decision in Dalkon Shield, does not create any absolute bar to the certification of a multi-state plaintiff class action in the medical products liability context. We decline to hold, at least at this early stage of the litigation, that there can never be a plaintiff class certification in this particular case. We do hold, however, on the basis of the record before us, that we must vacate this class certification order, because there has been no demonstration of how this class satisfies important Rule 23 requirements, including the predominance of common issues over individual issues and the superiority of class adjudication over other litigation alternatives.

I. Class Actions in Products Liability Litigation

The history of class action certifications and products liability cases in this circuit and elsewhere has not been luminous. Indeed the Advisory Committee on Civil Rules for the 1966 revision to Rule 23 cast doubt on the availability of class actions in mass tort cases. See Fed.R.Civ.P. 23, advisory committee's notes to 1966 amendment, Subdivision (b)(3) ("[a] 'mass accident' resulting in injuries to numerous persons is ordinarily not appropriate for a class action"). Nevertheless, courts have generally proceeded on a case-by-case basis and considered the appropriateness of class action treatment under the particular circumstances presented. See 7B Charles Alan Wright, Arthur R. Miller & Mary Kay Kane, Federal Practice and Procedure: Civil 2d, § 1783 at 74-75 (2d ed.1986); see also 3 Herbert B. Newberg & Alba Conte, Newberg on Class Actions, § 17.05 (3d ed.1992) (noting modern trend has been to expand use of class action litigation in mass tort context). The lead decision in this circuit was handed down in 1982 and vacated a nationwide punitive damages class and a statewide compensatory liability class of persons who had used allegedly defective intrauterine contraceptive devices. In re Northern Dist. of California, Dalkon Shield IUD Prods. Liab. Litig., 693 F.2d 847 (9th Cir.1982), cert. denied, 459 U.S. 1171, 103 S.Ct. 817, 74 L.Ed.2d 1015 (1983).

In rejecting the nationwide class certification under Rule 23(b)(1)(B), we were clearly troubled in Dalkon Shield by the problems that would arise in endeavoring to apply the varying punitive damage standards of fifty different jurisdictions. We did not, however, hold this commonality obstacle fatal. Id. at 850. There was in Dalkon Shield the added problem that no plaintiff, and no plaintiff's lawyer, had agreed to represent the class so that the requirements of typicality and adequacy of representation could not be satisfied. Id. at 850-51.

In considering the certification of the California liability class under Rule 23(b)(3), we commented in Dalkon Shield on the problems presented by products liability actions where, unlike the mass tort involving a single catastrophic event such as an airplane crash or cruise ship food poisoning, "[n]o single happening or accident occurs to cause similar types of physical harm or property damage." Id. at 853. We also discussed the inherent difficulties of proving proximate cause and a breach of a duty of care under a negligence theory, where there are different types of injuries and multiple defendants. Id. at 854-55. We were further troubled by the requirement that common issues predominate over individual issues in a certification of an entire case for class treatment; it appeared that only the underlying facts raised a common nucleus of issues, while the liability questions included highly individualized issues of damages and proximate cause. Id. at 856. Finally, we held that class adjudication would not be superior to individualized litigation given: first, the lack of any showing that class adjudication would save time or expense, and second, the management difficulties caused by the complexity and multiplicity of issues as well as the plaintiffs' hostility to the class action. Id.

We were careful in Dalkon Shield, however, not to preclude the future certification of more limited classes or subclasses pursuant to Rule 23(b)(3), or to rule out the possibility of broader class action certification in other products liability cases. See id. at 852-54, 856. Although Dalkon Shield pointed out many of the problems common to products liability litigation in meeting Rule 23's class certification requirements, we cannot conclude that Dalkon Shield creates an absolute bar to such certification in this circuit. As leading commentators have pointed out, the case was unusual in that there was simply no plaintiff or plaintiff's counsel ready, willing, and able to represent the class. See, e.g., 3 Newberg & Conte, supra, § 17.12 at 17-31. In addition, Dalkon Shield involved multiple defendants and multiple marketing schemes, unlike the present case where a single manufacturer marketed one drug over a limited period of time. Compare Dalkon Shield, 693 F.2d at 856 (holding district court erroneously certified class where manufacturer advertised in various medical journals and trade-show advertisements to different doctors), with In re Copley Pharmaceutical, 158 F.R.D. 485, 487, 491-93 (D.Wyo.1994) (certifying class where one manufacturer marketed four contaminated batches of one prescription drug).

The leading cases in other circuits in which class certifications have been approved are the "Agent Orange" litigation in the Second Circuit and the "School Asbestos" litigation in the Third Circuit. See In re Agent Orange Prod. Liab. Litig., 818 F.2d 145 (2d Cir.1987), cert. denied, 484 U.S. 1004, 108 S.Ct. 695, 98 L.Ed.2d 648 (1988); In re School Asbestos Litig., 789 F.2d 996 (3d Cir.), cert. denied, 479 U.S. 852, 107 S.Ct. 182, 93 L.Ed.2d 117, and cert. denied, 479 U.S. 915, 107 S.Ct. 318, 93 L.Ed.2d 291 (1986). Those cases also had some unique features.

In Agent Orange, the Second Circuit made it quite clear that the common issue in that case that caused class litigation to be both appropriate and superior to other forms of litigation was the common existence of a government contractor defense.

In our view, class certification was justified under Rule 23(b)(3) due to the centrality of the military contractor defense. First, this defense is common to all of the plaintiffs' cases, and thus satisfies the commonality requirement of Rule 23(a)(2). Second, because the military contractor defense is of central importance ... this issue is governed by federal law, and a class trial in a federal court is a method of adjudication superior to the alternatives. If the defense succeeds, the entire litigation is disposed of. If it fails, it will not be an issue in the subsequent individual trials. In that event, moreover, the ground for its rejection, such as a failure to warn the government of a known hazard, might well be dispositive of relevant factual issues in those trials.

Agent Orange, 818 F.2d at 166-67 (citations omitted).

In School Asbestos, the plaintiffs were school districts seeking compensation for property damages, not for personal injuries. The Third Circuit viewed that class action as much more manageable than a personal injury case would have been because, in essence, the effect of asbestos in different buildings is the same and the effect of asbestos on different people is not. See School Asbestos, 789 F.2d at 1010-11.

A leading decision in the Seventh Circuit has recently cast a pall on the future of class action certifications in products liability cases in that circuit. See In re Rhone-Poulenc Rorer, Inc., 51 F.3d 1293 (7th Cir.), cert. denied, --- U.S. ----, 116 S.Ct. 184, 133 L.Ed.2d 122 (1995); see also Castano v. American Tobacco Co., 84 F.3d 734 (5th Cir.1996) (decertifying national class of all nicotine-dependent persons, and expressing approval of Rhone-Poulenc ). The Seventh Circuit in Rhone-Poulenc issued a writ of mandamus ordering the district court to decertify a class of plaintiff hemophiliacs who were allegedly infected by the human immunodeficiency virus (HIV) as a result of using blood solids manufactured by the defendants. The Seventh Circuit majority was heavily influenced by at least three factors.

First, the majority expressed a general distaste for requiring defendants to place high economic stakes in the hands of a single jury. See Rhone-Poulenc, 51 F.3d at 1299. The majority also noted that there was a great likelihood that plaintiffs' legal claims lacked merit, given that twelve of thirteen individual suits had resulted in verdicts favorable to the defendants. See id. at 1299-1300. This concern does not appear to be in line with the law of this circuit that has not looked favorably upon granting extraordinary relief to vacate a class certification. See, e.g., Arthur Young & Co. v. United States Dist. Court, 549 F.2d 686, 698 (9th Cir.), cert. denied, 434 U.S. 829, 98 S.Ct. 109, 54 L.Ed.2d 88 (1977). There is also authority disapproving a separate hearing to consider the merits of the plaintiffs' claims when determining class certification. See Eisen v. Carlisle & Jacquelin, 417 U.S. 156, 177-78, 94 S.Ct. 2140, 2152-53, 40 L.Ed.2d 732 (1974); Blackie v. Barrack, 524 F.2d 891, 901 (9th Cir.1975), cert. denied, 429 U.S. 816, 97 S.Ct. 57, 50 L.Ed.2d 75 (1976); see also 7B Charles Alan Wright, Arthur R. Miller & Mary Kay Kane, supra, § 1785 at 125 (discussing Eisen and the Court's express rejection of a preliminary hearing to determine the merits of the litigation).

Second, the Rhone-Poulenc majority found that the class action would require a jury to determine "the negligence of the defendants under a legal standard that does not actually exist anywhere in the world." Id. at 1300. The court expressed concern with the ability of the district court to condense the law of the fifty states and the District of Columbia into a single jury instruction on negligence. See id. at 1300-02. The court thus focused on the district court's decision to create a hypothetical negligence standard. The district court in this case did not create such a hypothetical standard.

Third, the Rhone-Poulenc court perceived Seventh Amendment problems in the district court's bifurcation of class issues from individual issues, such as comparative negligence and proximate causation. See id. at 1302-03. The court determined that the district court's plan was inconsistent with the principle that the findings of one jury are not to be reexamined by a different jury. See id. at 1303. This constitutional concern of the Rhone-Poulenc court may not be fully in line with the law of this circuit, and constitutional issues were never squarely presented to the district court. See Arthur Young, 549 F.2d at 696.

We therefore do not accept Carter-Wallace's invitation in this case to adopt the principles of Rhone-Poulenc as the law of this circuit.

We are more sympathetic to the approach taken by the Sixth Circuit in In re American Medical Sys., 75 F.3d 1069 (6th Cir.1996). American Medical rejected class certification involving ten different models of penile implants that were implanted over a twenty-two year period. The court granted mandamus to decertify a nationwide class where the district court failed to identify common issues, explain why common issues predominate over individual issues, or make a finding of superiority. The court held that district courts must conduct a "rigorous analysis" into whether the prerequisites of Rule 23 are met before certifying a class. See id. at 1078-79. The Sixth Circuit has also recognized, however, that in the mass tort context, class adjudication of certain issues may be more efficient and expeditious than individualized litigation. See Sterling v. Velsicol Chem. Corp., 855 F.2d 1188 (6th Cir.1988).

Our reluctance to close the door on class action litigation in products liability cases is reinforced by current legal developments that could make class litigation more manageable. There has, for example, been discussion of federal class action legislation. See, e.g., Thomas D. Rowe, Jr., Beyond the Class Action Rule: An Inventory of Statutory Possibilities to Improve the Federal Class Action, 71 N.Y.U.L.Rev. 186 (1996) (discussing several areas in which legislation might enhance federal class actions); William W. Schwarzer et al., Judicial Federalism: A Proposal to Amend the Multidistrict Litigation Statute to Permit Discovery Coordination of Large-Scale Litigation Pending in State and Federal Courts, 73 Tex.L.Rev. 1529 (1995) (proposing amendments to the multidistrict litigation statute, 28 U.S.C. § 1407(a), to include state court cases). Further, the American Law Institute is now concluding its work on products liability in the Restatement of the Law of Torts. See Restatement (Third) of Torts: Products Liability (Tent. Draft No. 3, 1996); see also James A. Henderson, Jr. et al., Optimal Issue Separation in Modern Products Liability Litigation, 73 Tex.L.Rev. 1653, 1661-67 (1995) (discussing new Restatement as a reflection of current state of products liability law).

In addition, the Advisory Committee on Civil Rules is in the process of modifying Rule 23, and has proposed authorizing the possible certification of settlement classes that need not meet the requirements of Rule 23(b)(3). See Fed.R.Civ.P. 23(b)(4) (Draft Aug. 15, 1996); see also Samuel Estreicher, Foreword, Federal Class Actions After 30 Years, 71 N.Y.U.L.Rev. 1, 6 & n.26 (1996) (noting that proposed (b)(4) category would allow trial courts to certify class actions for purposes of settlement, even though the requirements of subdivision (b)(3) might not be met for trial); Edward H. Cooper, Rule 23: Challenges to the Rulemaking Process, 71 N.Y.U.L.Rev. 13 (1996) (discussing proposed changes to Rule 23). We observe that this idea has met with substantial opposition from a number of quarters. See, e.g., In re General Motors Corp. Pick-Up Truck Fuel Tank Prods. Liab. Litig., 55 F.3d 768, 786-94 (3d Cir.) (holding that under present rule settlement class must meet all Rule 23 requirements and expressing concern about dangers of overrewarding attorneys and undercompensating class members), cert. denied, --- U.S. ----, 116 S.Ct. 88, 133 L.Ed.2d 45 (1995); see also Georgine v. Amchem Prods., 83 F.3d 610, 624-25 (3d Cir.1996), cert. granted, --- U.S. ----, 117 S.Ct. 379, --- L.Ed.2d ----(1996); John C. Coffee, Jr., Class Wars: The Dilemma of the Mass Tort Class Action, 95 Colum.L.Rev. 1343 (1995); John C. Coffee, Jr., Understanding the Plaintiff's Attorney: The Implications of Economic Theory for Private Enforcement of Law Through Class and Derivative Actions, 86 Colum.L.Rev. 669 (1986). It is to be hoped that the debate on the proposed rule modification will add to our understanding of the appropriate role of class litigation in tort litigation.

For these reasons, we reject Carter-Wallace's position that the law of this circuit should prohibit any class certifications in products liability litigation. We therefore turn to the appropriateness of this particular certification order.

II. The Class Certification Order in This Case

This court reviews a district court's decision to grant class certification for abuse of discretion. See Six (6) Mexican Workers v. Arizona Citrus Growers, 904 F.2d 1301, 1304 (9th Cir.1990). In order for a class action to be certified, the plaintiffs must establish the four prerequisites of Fed.R.Civ.P. 23(a) and at least one of the alternative requirements of Fed.R.Civ.P. 23(b). See Fed.R.Civ.P. 23(b). An action may be maintained as a class action if the court finds that: (1) common questions of law and fact predominate over questions affecting individual members, and (2) a class action is superior to other available methods for the fair and efficient adjudication of the controversy. Fed.R.Civ.P. 23(b)(3); Dalkon Shield, 693 F.2d at 855-56.

The certification order which we review is brief and conclusory. The record reflects that it was entered with the express hope on the part of the district judge of encouraging settlement, and to trigger a ruling from this court on the more general issue of the viability of class certification in this circuit. The order is provisional and contemplates the possibility of future modifications, additions, or refinements of subclasses. The order was entered at an early stage in the proceedings, and the record simply does not reflect any basis for us to conclude that some key requirements of Rule 23 have been satisfied.

It is not clear that Plaintiffs have met either the typicality or adequacy of representation requirement. See Fed.R.Civ.P. 23(a)(3) and (4). The plaintiff-class representatives include two individuals who have had difficulty withdrawing from Felbatol and returning to prior medications, one alleging liver failure and one some unspecified type of liver damage. No named plaintiff has experienced aplastic anemia as a result of taking the drug, even though this condition is one of the most serious of the alleged adverse consequences. The named plaintiffs thus may not be able to provide adequate representation for those who have suffered different injuries. See Dalkon Shield, 693 F.2d at 854-55.

Additionally, notice may be problematic. The number of known users who have reportedly suffered actual injuries from the drug is relatively small in comparison with all the users of the drug, so that many potential members of the classes cannot yet know if they are part of the class. We therefore have serious due process concerns about whether adequate notice under Rule 23(c)(2) can be given to all class members to enable them to make an intelligent choice as to whether to opt out. See 7B Charles Alan Wright, Arthur R. Miller & Mary Kay Kane, supra, § 1786 at 197-98.

The first requirement of Rule 23(b)(3) is predominance of common questions over individual ones. Implicit in the satisfaction of the predominance test is the notion that the adjudication of common issues will help achieve judicial economy. See 1 Newberg & Conte, supra, § 4.25 at 4-86. Even if the common questions do not predominate over the individual questions so that class certification of the entire action is warranted, Rule 23 authorizes the district court in appropriate cases to isolate the common issues under Rule 23(c)(4)(A) and proceed with class treatment of these particular issues. See Dalkon Shield, 693 F.2d at 856; see also Copley, 158 F.R.D. at 491, 7B Charles Alan Wright, Arthur R. Miller & Mary Kay Kane, supra, § 1790 at 276; 1 Newberg & Conte, supra, § 4.25 at 4-81.

Here, the certification order merely reiterates Rule 23(b)(3)'s predominance requirement and is otherwise silent as to any reason why common issues predominate over individual issues certified under Rule 23(c)(4)(A). There has been no showing by Plaintiffs of how the class trial could be conducted. See e.g., Castano, 84 F.3d at 741-44. The district court abused its discretion by not adequately considering the predominance requirement before certifying the class. See Dalkon Shield, 693 F.2d at 856; cf. Agent Orange, 818 F.2d at 163-67; School Asbestos, 789 F.2d at 1010-11.

Last, but certainly not least, the district court must find that a class action is superior to other methods of adjudication. Fed.R.Civ.P. 23(b). Where classwide litigation of common issues will reduce litigation costs and promote greater efficiency, a class action may be superior to other methods of litigation. See Dalkon Shield, 693 F.2d at 856. A class action is the superior method for managing litigation if no realistic alternative exists. See Fed.R.Civ.P. 23(b)(3); 7A Charles Alan Wright, Arthur R. Miller & Mary Kay Kane, supra, § 1779 at 552. But here, as in Dalkon Shield, there has been no showing why the class mechanism is superior to alternative methods of adjudication, particularly when coupled with the discovery coordination that is made possible by the JPML consolidation. See Dalkon Shield, 693 F.2d at 856. Again, the certification order merely reiterates Rule 23(b)(3)'s superiority requirement but contains no discussion of alternatives or why class adjudication is superior.

The deficiencies in this certification are quite like those that caused the Sixth Circuit to reject the certification in American Medical, 75 F.3d at 1080-86. We similarly conclude that the district court abused its discretion by certifying particular issues for class adjudication. The district court's order is VACATED and the case is REMANDED for further proceedings.

VACATED AND REMANDED.