Full opinion text
Opinion for the Court filed by BAZEL-ON, Circuit Judge. Opinion filed by McGOWAN, Circuit Judge, concurring in part and dissenting in part. BAZELON, Circuit Judge: Smith, Kline & French Laboratories (SKF), a Division of SmithKline Corporation, has since 1950 produced and marketed Dexamyl, a prescription drug used as an anorectic in the treatment of obesity. Dex-amyl is a combination drug containing Dexedrine (dextroamphetamine sulfate), an appetite suppressant, and amobarbital, a barbituate designed to reduce the possible adverse side effects of Dexedrine. On August 24, 1976, the Acting Commissioner of Food and Drugs published an order denying SKF a hearing and refusing to approve the pending new drug applications (NDAs) for Dexamyl. 41 Fed.Reg. 35741 (1976). SKF petitions this court to reverse the Acting Commissioner’s summary judgment order and to remand to the Food and Drug Administration (FDA) for an evidentiary hearing pursuant to 21 U.S.C. § 355(c). See id. at § 355(h). I. BACKGROUND A. The Statutory and Regulatory Framework The Federal Food, Drug, and Cosmetic Act (the Act) prohibits the marketing in interstate commerce of any “new drug” unless an NDA for the drug has been approved.' See '21 U.S.C. § 355(a). In 1962 the Act was amended to define “new drug” to mean any drug not generally recognized as safe and effective for its intended use. See 76 Stat. 781, 21 U.S.C. § 321(p). The Amendments provided that an NDA would not be approved if it were found, “after due notice to the applicant . . . and giving him an opportunity for a hearing,” 21 U.S.C. § 355(d), that there was a lack of “substantial evidence” that the drug was effective for its intended use. Id. Substantial evidence was defined to mean: evidence consisting of adequate and well-controlled investigations, including clinical investigations, by experts qualified by scientific training and experience to evaluate the effectiveness of the drug involved, on the basis of which it could fairly and responsibly be concluded by such experts that the drug will have the effect it purports or is represented to have under the conditions of use prescribed, recommended, or suggested in the labeling or proposed labeling thereof. Id. The Amendments exempted certain drugs from the requirement of providing substantial evidence as to their effectiveness. A “grandfather” clause provided: In the case of any drug which, on the first day immediately preceding the enactment date, (A) was commercially used or sold in the United States, (B) was not a new drug as defined by section 201(p) of the basic Act as then in force, and (C) was not covered by an effective application under section 505 of that Act [21 U.S.C. § 355], the amendments to section 201(p) [21 U.S.C. § 321(p)] made by this Act shall not apply to such drug when intended solely for use under conditions prescribed, recommended, or suggested in labeling with respect to such drug on that day. Section 107(c)(4), 76 Stat. 789 (1962). FDA has promulgated regulations implementing this statutory scheme. See 21 U.S.C. § 371(a). It has required that for a fixed combination prescription drug such as Dexamyl substantial evidence must be presented both that the drug is effective for its intended use and that each constituent component contributes to the claimed effects. 21 C.F.R. § 300.50(a) (1977). In addition, FDA has specified criteria for the “adequate and well-controlled investigations” acceptable as substantial evidence of effectiveness. Although the Act appears to contemplate a hearing if FDA does not approve an NDA, see 21 U.S.C. §§ 355(c), 355(d), supra note 2, FDA has developed summary judgment procedures in cases where an applicant has failed to submit substantial evidence of drug efficacy sufficient to meet regulatory standards. The Supreme Court has in principle approved such procedures, stating that, “[w]e cannot impute to Congress the design of requiring, nor does due process demand, a hearing when it appears conclusively from the applicant’s ‘pleadings’ that the application cannot succeed.” Weinberger v. Hynson, Wes-cott & Dunning, Inc., 412 U.S. 609, 621, 93 S.Ct. 2469, 2479, 37 L.Ed.2d 207 (1973). B. Proceedings Before FDA When SKF decided to market Dexamyl tablets in 1949, it advised FDA of its position that the drug was not a “new drug” under the Federal Food, Drug and Cosmetic Act. See Appendix B at lb-2b. FDA agreed and informed SKF of its conclusion that Dexamyl was not a new drug. Id. at 5b. Thus no NDA was ever filed for Dexa-myl. When the 1962 Amendments changed the statutory definition of a “new drug” to designate one not generally recognized by experts as safe and effective for its intended use and in addition required substantial evidence of efficacy, FDA was faced with the massive task of reevaluating almost all marketed drugs. For assistance in this task, FDA retained the National Academy of Sciences-National Research Council (NAS-NRC) to create expert panels to review the efficacy of drugs by class. One such expert panel studied certain amphetamine combination products. On August 18, 1970, the Commissioner of Food and Drugs announced his concurrence with the panel’s finding that these products were only “possibly effective” for their claimed anorectic effects. 35 Fed.Reg. 12678 (1970). The Commissioner invited those who held NDAs for these products and “persons marketing any of these drugs without approval” to “submit in a supplemental or original new-drug application data to provide substantial evidence of effectiveness . . . .” Id. at 12679. On August 6,1971, SKF submitted NDAs for Dexamyl, together with a letter stating its position that Dexamyl was covered by the “grandfather clause” of the 1962 Amendments and was thus not subject to the efficacy requirement. Joint Appendix (J.A.) at 3-6. FDA informed SKF by letter on January 15,1973, that these NDAs were not approvable because “[t]he studies submitted fail to demonstrate the contribution of the sedative-tranquilizer constituent to the total effect of the drug.” J.A. at 15. Under FDA regulations, SKF was given 30 days to make a “written request to file the application over protest.” 21 C.F.R. § 130.-5(d) (1973). Within the 30-day period, on February 12, 1973, SKF requested that its Dexamyl NDAs be filed over protest, again noting its position that Dexamyl was grandfathered under the 1962 Amendments. J.A. at 17-18. On the same day, February 12,1973, FDA published in the Federal Register an “Opportunity for Hearing on Proposal to Withdraw Approval of New Drug Applications,” in which FDA stated that information concerning the efficacy of combination anorectic drugs that had been submitted in response to its August 8, 1970 notice was “found not to provide substantial evidence that the drugs are effective as fixed combinations for their claimed uses.” 38 Fed. Reg. 4279 (1973). FDA proposed to withdraw approval of the listed NDAs and stated that “[a]ll identical, related, or similar products, not the subject of an approved new drug application, are covered by the new drug application(s) reviewed.” Id. An NDA holder “or any other interested person” could request an evidentiary hearing on the issues involved. Id. at 4280. SKF responded to the February 12, 1973 notice although the notice had not mentioned Dexamyl by name. J.A. at 21-26. SKF stated, first, that Dexamyl was not covered by the NDAs reviewed in the notice. SKF went on to argue, however, that, assuming that Dexamyl was so covered, SKF had presented sufficient evidence in its previous submissions to require a full evidentiary hearing. On March 30, 1973, FDA approved the continued marketing of Dexamyl pending a ruling on SKF’s request for a hearing. 38 Fed.Reg. 8290 (1973). On July 11, 1973, SKF submitted to FDA the results of five, new, double-blind, clinical trials testing the efficacy of Dexamyl (multi-investigator clinical trials). J.A. at 27-31. Two weeks later, on July 27, 1973, FDA informed SKF by letter that the record regarding Dexamyl was closed as of that date, and that no further data would be accepted from SKF. J.A. at 85. Three years later, on August 24, 1976, FDA published an order denying a hearing and refusing to approve the pending NDAs for Dexamyl. 41 Fed.Reg. 35741 (1976). The Acting Commissioner of Food and Drugs stated: that (1) there is a lack of substantial evidence that Dexamyl products have the effects they are represented to have under the conditions of use recommended, suggested or prescribed in their labeling and (2) there is a lack of substantial evidence that each component of the combination products contributes to the total effects claimed. Smith, Kline and French has failed to offer any data or legal reason to demonstrate the existence of a genuine and substantial issue of fact requiring a hearing. Id. at 35754. The Acting Commissioner based his holding on the fact that none of the evidence submitted by SKF met the regulatory standards for adequate and well-controlled clinical studies. The Acting Commissioner also concluded that Dexamyl was covered by the NDAs of the drugs named in the February 12,1973 order, and that therefore Dexamyl was not grandfathered under the 1962 Amendments. The order of August 24 provided that it was to become effective on September 3, but on August 26 SKF filed with FDA a petition for reconsideration and for an administrative stay pending judicial review. J.A. at 104-19. FDA denied the request for reconsideration, but granted the stay. Id. at 121. On October 15, 1976, SKF filed in this court a petition for the review of FDA’s August 24 order. Prior to August 24, SKF had received no criticism from FDA of SKF’s multi-investi-gator clinical trials. After reviewing FDA’s August 24 order, SKF asked Herbert Solomon, Ph.D., Professor of Statistics at Stanford University, and William M. War-dell, M.D., Ph.D., Associate Professor of Pharmacology and Toxicology at the University of Rochester Medical Center, to review these studies. These two experts prepared affidavits averring that the studies were adequate and well-controlled clinical investigations under applicable scientific and regulatory standards. Stating that the restricted time frame surrounding the issuance of the August 24 order “made it impossible for SK&F to present an evaluation of FDA’s factual contentions regarding the Dexamyl studies in its August 26 request,” SKF appended the Solomon and Wardell affidavits to a supplemental request for reconsideration on November 24, 1976. FDA denied this request on December 8, 1976. SKF subsequently moved this court to remand the case to FDA with directions that it reconsider its order of August 24 in light of SKF’s supplemental request for reconsideration and the attached affidavits. On January 19, 1977, this court denied SKF’s motion. II. DEXAMYL AND THE GRANDFATHER CLAUSE OF THE 1962 AMENDMENTS SKF contends that the grandfather clause of the 1962 Amendments exempts Dexamyl from the requirement of demonstrating its effectiveness. We need not linger long over this question, however, since it is evident that the effectiveness standards do apply to Dexamyl. To be exempted by the grandfather clause, a drug must meet at least three conditions. As of October 9, 1962, it must have been (1) commercially used or sold in the United States; (2) not a new drug as then defined by the Act; and (3) not covered by an effective NDA. In his August 24 order the Commissioner found that Dexa-myl was “covered” by the effective NDAs of the amphetamine combination products listed in the February 12, 1973 order. These NDAs were effective on October 9, 1962. 41 Fed.Reg. at 35753 (1976). SKF does not contest these facts, but urges instead that since the NDAs for the drugs listed in the February 12 order were filed subsequent to the marketing of Dexa-myl, Dexamyl should not be considered “covered” for purposes of the grandfather clause. Although the Supreme Court has held in USV Pharmaceutical Corp. v. Wein-berger, 412 U.S. 655, 93 S.Ct. 2498, 37 L.Ed.2d 244 (1973), that when a pioneer product is the subject of an approved NDA, subsequently marketed “me-too” drugs are covered by the NDA for purposes of the grandfather clause, SKF interprets this decision as resting primarily “on the perceived inequity that would otherwise result, putting the ‘me-too’ drug in a more favorable regulatory posture than the pioneer.” Brief for petitioner at 58. If Dexamyl were to be considered covered by the NDAs of subsequently marketed “me-too” drugs, SKF urges, similar inequities would occur. Unlike SKF, we do not read USV Pharmaceutical to rest on such equitable considerations. Instead we understand it to mandate a policy of uniform and comprehensive application of the 1962 Amendments so as to avoid “a competitive contest in the marketing of ineffective pre-1962 drugs.” USV Pharmaceutical Corp. v. Weinberger, 412 U.S. at 665, 93 S.Ct. at 2505. SKF would have us engage in an interpretation of the grandfather clause that “is discriminatory and needlessly so,” and that “would create a hiatus in the regulatory scheme for which there appears to be no cogent reason.” Id. at 664, 93 S.Ct. at 2505. The grandfather clause itself makes no distinction between pioneer and “me-too” drugs, but exempts only that generic class of drugs which, on October 9, 1962, were not covered by an effective NDA. Dexamyl is not such a drug, and it is consequently not exempt from the 1962 Amendments. III. THE APPROPRIATENESS OF SUMMARY JUDGMENT A. General Considerations In Weinberger v. Hynson, Wescott & Dunning, 412 U.S. 609, 93 S.Ct. 2469, 37 L.Ed.2d 207 (1973), the Supreme Court upheld FDA’s summary judgment procedures. The Court stressed the massive regulatory task facing FDA: If FDA were required automatically to hold a hearing for each product whose efficacy was questioned by the NAS-NRC study, even though many hearings would be an exercise in futility, we have no doubt that it could not fulfill its statutory mandate to remove from the market all those drugs which do not meet the effectiveness requirements of the Act. Id. at 621, 93 S.Ct. at 2479. After examining FDA regulations specifying standards for the “well-controlled investigations” required by the Act, the Court concluded that they provided “drug manufacturers . full and precise notice of the evidence they must present to sustain their NDA’s . . . .” Id. at 622, 93 S.Ct. at 2479. Given such specific regulations, FDA was not required to “provide a formal hearing where it is apparent at the threshold that the applicant has not tendered any evidence which on its face meets the statutory standards as particularized by the regulations.” Id. at 620, 93 S.Ct. at 2478. This reasoning, of course, only applied “to those regulations that are precise.” Id. at 621 n. 17, 93 S.Ct. at 2479. The Court noted that “[s]ome of the regulations, however, are not precise, as they call for the exercise of discretion or subjective judgment in determining whether a study is adequate and well controlled.” For example, [§ 314.111(a)(5)(ii)(a)(2)(i)] requires that the plan or protocol for the study include a method of selection of the subjects that provide [sic] “adequate assurance that they are suitable for the purposes of the study.” (Emphasis added.) The qualitative standards “adequate” and “suitable” do not lend themselves to clear-cut definition, and it may not be possible to tell from the face of a study whether the standards have been met. Thus, it might not be proper to deny a hearing on the ground that the study did not comply with this regulation. Id. We have noted elsewhere, however, that the FDA regulations defining “well-controlled investigations” cannot be categorically classified as precise or imprecise with any degree of certainty. Indeed their precision will “vary with the context in which [they are] invoked. ... [A] regulatory provision which seems vague in the abstract may nevertheless be conclusively at odds with a peculiarly deficient item of evidence.” Cooper Laboratories, Inc. v. FDA, 163 U.S.App.D.C. 212, 220, 501 F.2d 772, 780 (1974). This flexibility is necessary because the regulations for the most part express general norms of scientific research rather than exact rules of procedure. The task of a court of appeals reviewing an FDA grant of summary judgment thus becomes, at least in those cases which do not involve the violation of a manifestly “precise” regulation, the determination of whether the applicant’s submission on its face is so conclusively deficient in light of these norms that no issue of fact remains whether “experts qualified by scientific training and experience” could “fairly and responsibly” conclude, on the basis of the submission, that “the drug will have the effect it purports or is represented to have under the conditions of use prescribed, recommended, or suggested in the labeling or proposed labeling thereof.” See 21 U.S.C. § 355(d). The difficulties besetting this enterprise are formidable. The issues entail complicated questions of scientific methodology, an area in which courts have little institutional competence. These questions must be confronted, moreover, in the absence of an evidentiary record. FDA’s interpretation of its summary judgment regulations does not require it to provide an applicant the opportunity to respond to the agency’s determination that the applicant’s submission is conclusively inadequate in light of regulations defining “substantial evidence.” See 39 Fed.Reg. 9750, 9751 (1974). The record will thus normally contain only FDA’s unchallenged characterizations of an applicant’s submissions. Although courts will ordinarily exercise considerable deference to an “agency’s technical expertise and experience,” particularly with respect to questions involving “ ‘engineering and scientific’ considerations,” FPC v. Florida Power & Light Co., 404 U.S. 453, 463, 92 S.Ct. 637, 30 L.Ed.2d 600 (1972), it is not clear how far that deference should extend when an agency has deliberately prevented the creation of a record by which its determinations can be probed for their underlying “basis in fact.” Id. See USV Pharmaceutical Corp. v. HEW, 151 U.S. App.D.C. 184, 191, 466 F.2d 455, 462 (1972). Our function in reviewing agency action is normally only “to assure that the agency has given reasoned consideration to all the material facts and issues,” Greater Boston Television Corp. v. FCC, 143 U.S.App.D.C. 383, 394, 444 F.2d 841, 851 (1970), cert. denied, 403 U.S. 923, 91 S.Ct. 2233, 29 L.Ed.2d 701 (1971), but when a record is barren concerning an issue presented in a petition for review — in this case the issue of whether an applicant’s studies are methodologically adequate in light of the pertinent regulations — we are asked to perform even this limited function on the basis of faith alone. The instant case amply illustrates these difficulties. SKF strenuously urges that FDA’s criticisms of its multi-investigator clinical trials are based upon “imprecise” regulations and are scientifically fallacious. Since FDA refused to consider these contentions, however, the record before us consists chiefly of SKF’s NDA, together with its multi-investigator clinical trials, and FDA’s August 24, 1976 order. We are thus confronted on one side by the arguments of lawyers, and on the other by the untested conclusions of FDA. To decide the scientific merit of these disputes, on the basis of the record now before us, would certainly be to risk the dangerous unreliability likely to occur when “technically illiterate judges” attempt substantively to review mathematical and scientific questions. Ethyl Corp. v. EPA, 176 U.S.App.D.C. 373, 439, 541 F.2d 1, 67, cert. denied, 426 U.S. 941, 96 S.Ct. 2663, 49 L.Ed.2d 394 (1976) (Bazelon, C. J., concurring). That risk may, in the end, prove inescapable, but it can perhaps be minimized if it is kept firmly in mind that we need not resolve the scientific question of the methodological adequacy of SKF’s multi-investigator clinical trials, but the legal question of whether they are on their face conclusively inadequate in light of the pertinent regulations. The manner in which we resolve this latter question involves policy considerations: we must take account of both fairness to the petitioner and the public interest in effective drug regulation. Since FDA has precluded SKF from establishing on the record the factual predicates for its arguments that FDA’s criticism of the multi-investigator clinical trials are scientifically inaccurate, rudimentary fairness requires that we at least give the factual claims underlying SKF’s arguments the benefit of the doubt. See Gellhorn & Robinson, Summary Judgment in Administrative Adjudication, 84 Harv.L.Rev. 612 (1971). The Supreme Court has made clear, however, that, because these circumstances do not involve the Seventh Amendment right to a trial by jury, we need not engage in the sharp limitations on summary judgment required by Rule 56 of the Federal Rules of Civil Procedure. Weinberger v. Hynson, Westcott & Dunning, 412 U.S. 609, 622, 93 S.Ct. 2469, 37 L.Ed.2d 207 (1973). If we were to follow our usual practice under Rule 56 of accepting as true petitioner’s version of the facts, Bishop v. Wood, 426 U.S. 341, 347, 96 S.Ct. 2074, 48 L.Ed.2d 684 (1976), petitioner might be entitled, despite the present abortive state of the record, to a full adjudicatory hearing, even though such a hearing might well be “an exercise in futility.” Weinberger v. Hynson, West-cott & Dunning, 412 U.S. 609, 621, 93 S.Ct. 2469, 37 L.Ed.2d 207 (1973). Full adjudicatory hearings represent a serious drain on FDA’s limited resources, see Ames & McCracken, Framing Regulatory Standards to Avoid Formal Adjudication: The FDA As a Case Study, 64 Calif.L.Rev. 14, 19-20 (1976), and their profusion might well compromise FDA’s ability to “fulfill its statutory mandate to remove from the market all those drugs which do not meet the effectiveness requirements of the Act.” Wein-berger v. Hynson, Wescott & Dunning, 412 U.S. 609, 621, 93 S.Ct. 2469, 2479, 37 L.Ed.2d 207 (1973). The Supreme Court has stressed the role of FDA’s summary judgment procedures as a shield to protect the public interest against this eventuality. Id. As we said in Cooper Laboratories, Inc. v. FDA, “the court must vindicate the general public’s right to an efficient administration of the congressional mandate." 501 F.2d 772, 780, 163 U.S.App.D.C. 212, 220 (1974). Therefore we will, if possible, avoid ordering an adjudicatory hearing until we are assured, with some basis in a record, that a genuine issue of fact exists there to be resolved. B. The Multi-Investigator Clinical Trials SKF rests its claim to a hearing chiefly on its multi-investigator clinical trials. SKF argues that these trials establish both that Dexamyl is effective as an anorectic and that each of its components contributes to the effects claimed; i. e., that the Dexedrine in Dexamyl performs its weight-loss function undiminished by the presence of amobarbital, and that the amobarbital makes its own contribution by reducing adverse side effects experienced by patients who take Dexedrine alone. The trials tested Dexamyl tablets against Dexedrine. The studies were double-blind; neither investigators nor patients knew which of the two drugs had been prescribed to any particular patient. There were five investigators, scattered throughout the country; each was a physician who treated patients in his private medical practice for obesity. All followed the same protocol, under which 30 patients were selected for each study, 15 of whom were assigned randomly to one drug, and 15 to the other. The trials lasted eight weeks. In its August 24, 1976 order, FDA noted numerous deficiencies which, it claimed, rendered the trials conclusively inadequate in light of the pertinent regulations. Before this court, however, FDA has chosen to press only six of these deficiencies. Should even a single one of these six deficiencies prove valid, we must sustain FDA’s summary judgment order. 1. Suitability of Patient Population The subjects of the trials were selected from among patients between the ages of 18 and 60 who were at least 20 percent overweight, and who scored four or more on a prescribed Anxiety Manifestation Index. No patients were included if, among other things, they suffered from a number of specified diseases or had a history of drug abuse. The protocol also required that patients selected for a trial not have received any anorectic or tranquilizing medication for at least two weeks prior to their participation in the study, and tranquilizers, sedatives and other anorectic agents or measures for weight control were prohibited during the trial. Patients were placed on a 1200 calorie diet. Section 505(d) of the Act, 21 U.S.C. § 355(d), provides that for a study to constitute substantial evidence, experts must be able fairly and responsibly to conclude from it that the drug will have the effect it purports to have “under the conditions of use prescribed, recommended or suggested” in its labeling. FDA regulations provide that there must be a method of selection of subjects which provides adequate assurance that they are suitable for purposes of the study. See 21 C.F.R. § 314-lll(a)(5)(ii)(a)(2)(i), supra note 3. Dexa-myl’s present labeling indicates use with obese patients, but the subjects of the trials were anxious, obese patients. Since anxiety was one of the side effects of Dexedrine that amobarbital was designed to remedy, FDA argues that it is impossible to generalize the results obtained in the subpopulation of patients studied by the trials. Its point appears to be that the trials do not demonstrate whether Dexedrine produces in non-anxious obese patients side-effects that can be remedied by amobarbital. If no such side effects are produced, the prescription of amobarbital would be superfluous. FDA’s contention is well grounded in its regulations. The trials provide no “assurance” that Dexedrine, in the amounts contained in Dexamyl, produces in non-anxious obese patients side effects capable of being remedied by amobarbital. This conclusion is not fatal to SKF, however, since it implies only that Dexamyl’s labeling be altered to recommend Dexamyl for use with anxious, obese patients. SKF implies that it would accept such an alteration, reply brief for petitioner, at 10 n.7, and we must therefore examine the other deficiencies alleged by FDA to determine whether the trials are conclusively inadequate despite such an alteration in labeling. 2. Methods of Observation and Recording The trials were conducted by physicians whose private medical practices included obese patients, and each physician-investigator was aware that the trials were testing the side-effect potential of the two drugs involved. The physicians were expected to use their normal but trained powers of observation, supplemented by specific inquiry where they deemed it appropriate. They completed forms recording extensive data for all patients. Particular attention was paid to adverse effects, which were detailed as to dates (onset and termination), severity, and relationship to the drug prescribed. The supervising physicians recorded whatever action they took as the result of such an adverse effect, including any reduction of dosage or outright discontinuance of the drug involved. J.A. at 83. FDA regulations require a well-controlled study to explain its “methods of observation and recording of results, including the variables measured, quantitation, assessment of any subjects response, and steps taken to minimize bias on the part of the subject and observer.” 21 C.F.R. § 314.111(a)(5) (ii)(a)(3), supra note 3. FDA contends that the trials are conclusively inadequate in light of this regulation since they nowhere describe what procedure (observation, general questions or specific inquiries) was actually used by each investigator to elicit adverse reaction data, whether specific inquiry was resorted to under the same circumstances or whether each of the investigators asked their subjects the same questions. FDA argues that the absence of this information is particularly serious because the method of obtaining adverse reaction data may seriously affect the incidence of reported side effects, and because, without this information, it is impossible for an impartial observer to know whether a reported response is an adverse reaction to the drug under study or merely the persistence of a pre-existing condition. Both parties have framed their arguments as though the regulation in question were precise and merely formal in its requirements. SKF strenuously urges, for example, that it has provided an explanation of its methods of observation and recording of results in that the protocol of the trials called for data to be observed by physicians in the course and manner of their regular private practices and to be recorded on standardized patient report forms. FDA, on the other hand, urges with equal vehemence that “[n]o explanation [was] provided as to whether the investigators evaluated subjects solely on the basis of their own observations, or on the basis of questions asked the subjects, or both.” 41 Fed.Reg. at 35745. It is evident, however, that this framing of the question leads nowhere, since it fails to suggest criteria by which it might be determined whether the information provided by the trials’ protocol constitutes an “explanation” for purposes of the regulation. The question must therefore be reformulated so as to capture the basic purpose underlying FDA’s summary judgment regulations. We must ask, that is, whether the trials are so conclusively deficient in light of the norm of scientific research expressed by the regulation that no question of fact remains whether qualified experts could fairly and responsibly conclude from the trials whether Dexamyl was effective for its intended uses. SKF argues that such a question of fact exists. The trials were conducted in a double-blind fashion. Since neither investigators nor subjects knew which patients were receiving Dexamyl and which Dexedrine, SKF concludes that the possible irregularities in observation noted by FDA would not bias the trials’ results. Whether the level of “explanation” offered by SKF was sufficient under the regulation thus involves a complex question of scientific methodology. We, of course, have neither the knowledge nor the competence to ourselves resolve this question. Giving petitioner’s version of the facts the benefit of the doubt, however, we must agree that the question is presently unsettled. But we will not so abandon the public interest in the expeditious enforcement of the drug laws as to assume, in the absence of a more developed record, that SKF’s contentions have truly raised a genuine issue of fact requiring an adjudicatory hearing. 3. Comparability of Test and Control Groups The trials compared a “test” group of patients who received Dexamyl against a “control” group who received Dexedrine. FDA regulations require that a well-controlled study must include “[a] method of selection of the subjects that (iii) Assures comparability in test and control groups of pertinent variables . . .” 21 C.F.R. § 314.111(a)(5)(ii)(a)(2), supra note 3. FDA argues that the trials are conclusively inadequate in light of this regulation because they fail to demonstrate the comparability of test and control subjects regarding the incidence and degree of anxiety. SKF vigorously contests this characterization of the trials, contending that they did indeed include a method of selection assuring comparability. The subjects were assigned to test and control groups on a random basis, and a statistical analysis of the Anxiety Manifestation Index for both groups indicated a random distribution of anxiety. FDA discounts this analysis, however, stating that the regulations require a check on the results of randomization at the conclusion as well as at the beginning of a study. In this FDA appears to be improperly elaborating on its own regulations, which require only that a study provide a “method of selection of the subjects” that would ensure comparability. FDA’s apparent requirement that a study provide a check on the comparability of test and control groups at the conclusion of the study is thus not included in “the statutory standards as particularized by the regulations,” Weinberger v. Hynson, Wescott & Dunning, 412 U.S. 609, 620, 93 S.Ct. 2469, 2478, 37 L.Ed.2d 207 (1973), and cannot be used as a basis for summary judgment. FDA argues that the trials also failed to assure comparability between test and control groups with respect to the past use of amphetamines. This is important, FDA stresses, because “amphetamines as a class can produce both tolerance and dependence,” and because “a subject’s prior history of amphetamine use may affect . the incidence of various side effects.” 41 Fed.Reg. at 35746. SKF contends, however, that these factors are irrelevant because no subject was included in the trials who had used amphetamines during the two weeks immediately prior to the commencement of the studies and because this two week “washout” period was sufficient to eliminate any interference that might be caused by a subject’s prior history of amphetamine use. The issue of whether the two week washout period is sufficient to assure comparability in the selection of test and control groups appears to raise a question of fact. Whether this appearance is genuine, however, depends upon the state of opinion in the scientific community, a circumstance concerning which we are at present entirely ignorant. We are therefore unable to conclude from the record before us whether “a real, not formal, controversy exists.” Bruce Construction Corp. v. United States, 242 F.2d 873, 875 (5th Cir. 1957). 4. Statistical Pooling The trials analyzed data compiled by each of the five individual investigators. This data was then pooled or aggregated for further statistical analysis. FDA concluded that this pooling failed to provide the necessary assurance of comparability between test and control groups with respect to pertinent variables: In several respects, SKF failed to demonstrate that the data from the five investigators could be combined for statistical analysis because it was not shown that each block of patients, in any single study or in all five studies, was similar with respect to either the presence or the degree of anxiety (§ 314.111(a)(5)(ii)(a) (2)(iii)). Because each investigator may have used widely varying criteria to measure anxiety, and because the effect of different degrees of anxiety on side effects is unknown, it is impermissible to pool the results of these investigators. A further obstacle to pooling is the investigators’ failure to show that each block of subjects was comparable with respect to (1) prior history of amphetamine use, (2) the manner in which each investigator elicited adverse reaction data and/or (3) the method used to record adverse reactions. 41 Fed.Reg. at 35746. SKF argues that FDA’s conclusion overlooks the fact that all five investigators used the same experimental protocol, including the two week washout period for prior amphetamine use, the random assignment of subjects to test and control groups, and the double-blind administration of the trials. Thus, for example, if the washout period were effective in assuring comparability between test and control groups for a single investigator, it would also be so with respect to the aggregated data. Since we have already concluded that the effectiveness of the washout period is an open question, we must also consider this aspect of the issue of pooling as unsettled. FDA’s conclusion, however, may rest on the additional point that the protocol of the trials was not sufficiently detailed to sustain pooling of the data under 21 C.F.R. § 314.111(a)(5)(ii)(a)(3). See note 3 supra. The fact that different investigators used different techniques for soliciting or recording the data, for example, may present a bar to the statistical pooling of their results. But this is an intricate scientific question concerning which we cannot exercise independent judgment, and FDA has offered neither evidence nor explanation. Extending the benefit of the doubt to petitioner’s general claim that the protocol was in fact sufficiently detailed to permit pooling, we must conclude that the question of whether a genuine issue of fact here exists remains open. 5. Improper Control The trials compared Dexamyl to Dexedrine. FDA regulations state that a well-controlled study must provide “a comparison of the results of treatment or diagnosis with a control in such a fashion as to permit quantitative evaluation. ... An effective regime of therapy may be used for comparison . . . .” 21 C.F.R. § 314.-lll(a)(5)(ii)(a)(4)(iii), supra note 3. FDA concluded that the trials were conclusively inadequate in light of this regulation “since the active control used, dextroampheta-mine, does not permit a quantitative evaluation of the effects of Dexamyl.” Because the anorectic effect of dex-troamphetamine is only marginal to begin with, (i. e., its advantage over placebo is small), and because its anorectic effect is highly variable, its effectiveness is not demonstrable in every study. Therefore, before a finding of anorectic equivalence of dextroamphetamine and Dexamyl in a particular study can be considered as evidence of the effectiveness of Dexamyl, it must be demonstrated that dextroamphetamine was in fact effective under the precise conditions of the study, e. g., by including a placebo control in the study. This series failed to include a placebo control or otherwise show that dextroamphetamine was in fact effective under those circumstances. Accordingly, a finding of “no difference” between Dexamyl and dextroampheta-mine in a particular study can mean either that both were effective or that neither was effective in that study. 41 Fed.Reg. at 35750. SKF vigorously contests this characterization of Dexedrine (dextroamphetamine), pointing to FDA’s own conclusion in the Federal Register that: On the basis of the currently available evidence, the Commissioner concludes that oral dosage forms of amphetamine and/or dextroamphetamine are effective in the management of exogenous obesity as a short term (a few weeks) adjunct in a regimen of weight reduction based on caloric restriction for patients in whom obesity is refractory to other measures. 38 Fed.Reg. 4249 (1973). Surely, SKF argues, if FDA recognizes Dexedrine as effective, the trials were not on their face deficient to use the drug as an active control. We agree with SKF in this matter. FDA’s endorsement of Dexedrine as effective for the short term management of exogenous obesity provides at least prima facie support for SKF’s view of the appropriateness of Dexedrine as an active control. Since, in contrast to other issues we have heretofore considered, SKF’s position is supported by evidence, we conclude that there is a genuine issue of fact whether Dexedrine, although effective, is so unpredictable as to foreclose the possibility of quantitative evaluation. Such issues are not to be decided at summary judgment. 6. Analyst Bias FDA requires that a well-controlled investigation explain “the methods used to minimize bias on the part of the observers and the analysts of the data.” 21 C.F.R. § 314.lll(a)(5)(ii)(a)(4), supra note 3. FDA concluded that the trials were conclusively deficient in light of this regulation: The authors and SKF . . . failed to state what steps, if any, were taken to minimize bias on the part of the analysts of the data (§ 314.111(a)(5)(ii)(a)(4)). Steps must be taken to preclude the possibility of analyst bias (either conscious or unconscious), since it may cause them to make subjective decisions that vitally affect the results of the study, e. g., whether or not to include all patients in the analysis or only those who completed the entire 8 weeks. 41 Fed.Reg. at 35751. In its August 24 order, FDA appeared to overstep the bounds of its regulation. Although the latter requires only that there be “an explanation given of the methods used to minimize bias on the part of the . analysts of the data,” the order implies both that a well-controlled investigation must include such methods and that such methods be adequate. In its brief FDA is more circumspect, arguing only that SKF provided “[n]o such explanation.” Brief for respondent at 34. SKF rejects this argument, noting that the steps taken in the trials to minimize analyst bias are stated in the protocol and final report. These include a multiple covariance quantitative analysis, a Chi-Square analysis of qualitative results, a report of the details and results of these analyses, and an analysis both of all subjects and of only those subjects who completed the entire eight week program. Since it is impossible to determine, in the abstract, whether the information provided in the trials’ protocol constitutes a sufficient “explanation" for purposes of the regulation, we must ask, once again, whether the trials are so conclusively deficient with respect to the norm of scientific research expressed by the regulation that no issue of fact remains whether qualified experts could fairly and responsibly conclude from the trials whether Dexamyl was effective for its intended uses. We have no way of answering this question in the absence of a more developed record. Giving the benefit of the doubt to petitioner, we conclude that the question remains open. IV. THE QUESTION OF REMEDY If Dexamyl’s labeling were altered to recommend its use for anxious, obese patients, FDA’s summary judgment order cannot be sustained. It remains an open question, however, whether the trials are conclusively deficient in light of FDA regulations requiring comparability of test and control groups, an explanation of the methods used to observe and record results, and an explanation of the methods used to minimize analyst bias. These issues cannot be resolved in the absence of an evidentiary record. SKF has proffered evidence on these issues in the form of the Solomon and Warded affidavits. We are satisfied that these affidavits are material and that FDA’s rather abrupt procedures below provide reasonable grounds for SKF’s failure to adduce them before the agency. In such circumstances the Act specifically provides that the court may order such additional evidence to be taken before the [Commissioner] and to be adduced upon the hearing in such manner and upon such terms and conditions as to the court may seem proper. The [Commissioner] may modify his findings as to the facts by reason of the additional evidence so taken, and he shall file with the court such modified findings which, if supported by substantial evidence, shall be conclusive, and his recommendation, if any, for the setting aside of the original order. 21 U.S.C. § 355(h). The statute thus gives this court broad authority to fashion a remedy capable of balancing fairness to the petitioner against the public’s right to an expeditious enforcement of the Act. Since we cannot sustain FDA’s summary judgment order and since, at the same time, we have no assurance that there is presently a genuine issue of fact to be aired at an adjudicatory hearing, this record should be remanded to FDA for a proceeding to determine whether such a genuine issue of fact exists. The extent of this proceeding should be as limited as its circumscribed purpose. Were this proceeding to become too lengthy or elaborate, it would create precisely the drain on FDA’s resources FDA’s summary judgment regulations were designed to prevent. An appropriate adaptation of the procedures set forth in Part 15 of 21 C.F.R. might well prove sufficient, if it is understood that FDA itself must submit on the record and for the comments of petitioner evidence for FDA’s conclusions. Confident, however, that “court and agency [are] ‘collaborative instrumentalities of justice,’ ” S.S.W. Inc. v. Air Transport Ass'n, 89 U.S.App.D.C. 273, 279, 191 F.2d 658, 664 (1951), cert. denied, 343 U.S. 955, 72 S.Ct. 1049, 96 L.Ed. 1355 (1952), we decline to specify any particular procedures to be used on remand. “Our own responsibility as a court is as a partner in the overall administrative process — acting with restraint, but providing supervision.” Environmental Defense Fund, Inc. v. EPA, 150 U.S.App.D.C. 348, 360-61, 465 F.2d 528, 540-41 (1972). We require only that the proceeding on remand provide assurance that FDA has given reasoned consideration to all material facts and issues. One final question deserves attention. SKF submitted NDAs for Dexamyl in two forms: tablets and Spansule (sustained release) capsules. A Spansule capsule contains precisely two or precisely three times the amounts of the two active ingredients that are included in a single tablet. Tablets are prescribed for use two or three times daily, but patients take only one Spansule capsule each morning. See 41 Fed.Reg. at 35742. The multi-investigator clinical trials were conducted using Dexamyl tablets. SKF has submitted extensive data purporting to demonstrate that there is an equivalent “bioavailability” of Dexamyl’s active ingredients for both tablets and Spansules. It therefore claims that the results of the trials should be applied to the NDA for Dexamyl Spansules. FDA did not comment on these bioavaila-bility data in its August 24 order. We therefore have no way of knowing whether they constitute the adequate and well-controlled investigations required by statute. In the proceeding on remand, however, FDA should pass on this question. The record is remanded for further proceedings not inconsistent with this opinion. So ordered. . SKF produces Dexamyl in two forms at issue in these proceedings: tablets and sustained release capsules called Spansules. . 21 U.S.C. § 355(c) states: Within one hundred and eighty days after the filing of an application under this subsection, or such additional period as may be agreed upon by the Secretary and the applicant, the Secretary shall either— (1) approve the application if he then finds that none of the grounds for denying approval specified in subsection (d) of this section applies, or (2) give the applicant notice of an opportunity for a hearing before the Secretary under subsection (d) of this section on the question whether such application is approvable. If the applicant elects to accept the opportunity for hearing by written request within thirty days after such notice, such hearing shall commence not more than ninety days after the expiration of such thirty days unless the Secretary and the applicant otherwise agree. Any such hearing shall thereafter be conducted on an expedited basis and the Secretary’s order thereon shall be issued within ninety days after the date fixed by the Secretary for filing final briefs. Section 355(d) provides, in part: If the Secretary finds, after due notice to the applicant in accordance with subsection (c) of this section and giving him an opportunity for a hearing, in accordance with said subsection, that (1) the investigations, reports of which are required to be submitted to the Secretary pursuant to subsection (b) of this section, do not include adequate tests by all methods reasonably applicable to show whether or not such drug is safe for use under the conditions prescribed, recommended, or suggested in the proposed labeling thereof; (2) the results of such tests show that such drug is unsafe for use under such conditions or do not show that such drug is safe for use under such conditions; (3) the methods used in, and the facilities and controls used for, the manufacture, processing, and packing of such drug are inadequate to preserve its identity, strength, quality, and purity; (4) upon the basis of the information submitted to him as part of the application, or upon the basis of any other information before him with respect to such drug, he has insufficient information to determine whether such drug is safe for use under such conditions; or (5) evaluated on the basis of the information submitted to him as part of the application and any other information before him with respect to such drug, there is a lack of substantial evidence that the drug will have the effect it purports or is represented to have under the conditions of use prescribed, recommended, or suggested in the proposed labeling thereof; or (6) based on a fair evaluation of all material facts, such labeling is false or misleading in any particular; he shall issue an order refusing to approve the application. If, after such notice and opportunity for hearing, the Secretary finds that clauses (1) through (6) do not apply, he shall issue an order approving the application. The Secretary of Health, Education and Welfare (Secretary) has delegated all functions assigned him by the Act to the Commissioner of Food and Drugs. See 21 C.F.R. § 5.1(a)(1) (1977). . See, e. g„ 21 C.F.R. § 314.111(a)(5)(ii): (ii) The following principles have been developed over a period of years and are recognized by the scientific community as the essentials of adequate and well-controlled clinical investigations. They provide the basis for the determination whether there is “substantial evidence” to support the claims of effectiveness for “new drugs” and antibiotic drugs. (a) The plan or protocol for the study and the report of the results of the effectiveness study must include the following: (1) A clear statement of the objectives of the study, (2) A method of selection of the subjects that (/) Provides adequate assurance that they are suitable for the purposes of the study, diagnostic criteria of the condition to be treated or diagnosed, confirmatory laboratory tests where appropriate, and, in the case of prophylactic agents, evidence of susceptibility and exposure to the condition against which prophylaxix is desired. (ii) Assigns the subjects to test groups in such a way as to minimize bias. (iii) Assures comparability in test and control groups of pertinent variables, such as age, sex, severity, or duration of disease, and use of drugs other than the test drug. (3) Explains the methods of observation and recording of results, including the variables measured, quantitation, assessment of any subjects response, and steps taken to minimize bias on the part of the subject and observer. (4) Provides a comparison of the results of treatment or diagnosis with a control in such a fashion as to permit quantitative evaluation. The precise nature of the control must be stated and an explanation given of the methods used to minimize bias on the part of the observers and the analysts of the data. Level and methods of “blinding,” if used, are to be documented. Generally, four types of comparison are recognized: (/) No treatment: Where objective measurements of effectiveness are available and placebo effect is negligible, comparison of the objective results in comparable groups of treated and untreated patients. (ii) Placebo control: Comparison of the results of use of the new drug entity with an inactive preparation designed to resemble the test drug as far as possible. (iii) Active treatment control: An effective regimen of therapy may be used for comparison, e. g., where the condition treated is such that no treatment or administration of a placebo would be contrary to the interest of the patient. (iv) Historical control: In certain circumstances, such as those involving diseases with high and predictable mortality (acute leukemia of childhood), with signs and symptoms of predictable duration or severity (fever in certain infections), or in case of prophylaxis, where morbidity is predictable, the results of use of a new drug entity may be compared quantitatively with prior experience historically derived from the adequately documented natural history of the disease or condition in comparable patients or populations with no treatment or with a regimen (therapeutic, diagnostic, prophylactic) the effectiveness of which is established. (5) A summary of the methods of analysis and an evaluation of data derived from the study, including any appropriate statistical methods. . See 21 C.F.R. § 314.200 (1977): (a) The notice to the applicant, and to all other persons who manufacture or distribute identical, related, or similar drug products as defined in § 310.6 of this chapter, of an opportunity for a hearing on a proposal by the Director of the Bureau of Drugs to refuse to approve an application or to withdraw the approval of an application will state the reasons for his action and the grounds upon which he proposes to issue his order. (1) Such notice may be general (i. e., simply summarizing in a general way the information resulting in the notice) or specific (i. e., either referring to specific requirements in the statute and regulations with which there is a lack of compliance, or providing a detailed description and analysis of the specific facts resulting in the notice). ****** (g) A request for a hearing may not rest upon mere allegations or denials, but must set forth specific facts showing that there is a genuine and substantial issue of fact that requires a hearing with respect to the particular drug product(s) specified in the request for hearing. (1) Where a specific notice of opportunity for hearing (as defined in paragraph (a)(1) of this section) is used, it shall state that, if it conclusively appears from the face of the data, information, and factual analyses in the request for the hearing that there is no genuine and substantial issue of fact which precludes the refusal to approve the application or the withdrawal of approval of the application, e. g., no adequate and well-controlled clinical investigations meeting each of the precise elements of § 314.111(a)(5) and, for a combination drug product, § 300.50 of this chapter, showing effectiveness have been identified, or when a request for hearing is not made in the required format or with the required analyses, the Commissioner will enter summary judgment against the person(s) who requests the hearing, making findings and conclusions, denying a hearing. Any such order entering summary judgment shall set forth the Commissioner’s findings and conclusions in detail and shall specify why each study submitted fails to meet the requirements of the statute and regulations or why the request for hearing does not raise a genuine and substantial issue of fact or shall specify the requirements of this section with respect to format or analyses with which there is a lack of compliance. (2) Where a general notice of opportunity for hearing (as defined in paragraph (a)(1) of this section) is used and the Director of the Bureau of Drugs concludes that summary judgment against the person(s) requesting a hearing should be considered, he shall serve upon such person(s) by registered mail a proposed order denying a hearing. Such persons) shall have 60 days after receipt of such proposed order to respond with sufficient data, information, and analyses to demonstrate that there is a genuine and substantial issue of fact which justifies a hearing. (3) Where a general or specific notice of opportunity for hearing is used and the persons) requesting a hearing submits data or information of a type required by the statute and regulations, and the Director of the Bureau of Drugs concludes that summary judgment against such person(s) should be considered, he shall serve upon such person(s) by registered mail a proposed order denying a hearing. Such person(s) shall have 60 days after receipt of such proposed order to respond with sufficient data, information, and analyses to demonstrate that there is a genuine and substantial issue of fact which justifies a hearing. . At that time the Act defined a new drug as one not generally recognized among experts as safe for its intended use. . In 1949 Dexamyl was known as “Dexam,” and was marketed in tablet form. . In 1953, FDA informed SKF that Dexamyl Spansules were also not new drugs. Appendix B at 6b. In 1968 FDA promulgated regulations revoking “all opinions previously given by the Food and Drug Administration to the effect that an article is ‘not a new drug’ or is ‘no longer a new drug’ . . . .” 21 C.F.R. § 310.100(d) (1977). . The lack of an NDA was not all that unusual. There existed a large class of “me too” drugs which were also marketed without an approved NDA. “For every drug for which an NDA became effective between 1938 and 1962, as many as 13 other products similar in formulation and claimed therapeutic effect, were marketed without formal review.” J. Mashaw & R. Merrill, Introduction to the American Public Law System 498 (1975). The producers of these “me too” drugs claimed that the approved NDA of the pioneer drug was sufficient evidence of the general recognition of the safety of these “me too” drugs for their intended use. . Some 7000 drug formulations had been processed under the Act between 1938 and 1962, of which approximately 4000 were on the active market in 1962. Note, Drug Efficacy and the 1962 Drug Amendments, 60 Geo.L.J. 185, 207 n.141 (1971). In addition there were thousands of “me too” drugs. Weinberger v. Hynson, Wescott & Dunning, 412 U.S. 609, 614, 93 S.Ct. 2469, 37 L.Ed.2d 207 (1973). . The panels rated drug claims as “effective,” “effective but,” “probably effective,” “possibly effective,” or “ineffective.” Drug claims receiving ratings other then “effective” were not considered as having established their efficacy. . FDA regulations stated that upon filing such a request, “the application shall be reevaluated, and within 30 days of the date of receipt of such written request, the application shall be approved, or the applicant shall be given written notice of an opportunity for a hearing on the question whether the application is approvable.” 21 C.F.R. § 130.5(d) (1973), now codified at 21 C.F.R. § 314.110(d). . The necessity for such a broad notice was spelled out in FDA regulations: The Food and Drug Administration’s conclusions on the effectiveness of drugs are currently being published in the Federal Register as Drug Efficacy Study Implementation (DESI) Notices and as Notices of Opportunity for Hearing. The specific products listed in these notices include only those that were introduced into the market through the new-drug procedures from 1938-62 and were submitted for review by the National Academy of Sciences-National Research Council (NAS-NRC), Drug Efficacy Study Group. Many products which are identical to, related to, or similar to the products listed in these notices have been marketed under different names or by different firms during this same period or since 1962 without going through the new-drug procedures or the Academy review. Even though these products are not listed in the notices, they are covered by the new drug applications reviewed and thus are subject to these notices. All persons with an interest in a product that is identical, related, or similar to a drug listed in a drug efficacy notice or a notice of opportunity for a hearing will be given the same opportunity as the applicant to submit data and information, to request a hearing, and to participate in any hearing. It is not feasible for the Food and Drug Administration to list all products which are covered by an NDA and thus subject to each notice. However, it is essential that the efficacy conclusions be applied to all identical, related, and similar drug products to which those conclusions are reasonably applicable. Any product not in compliance with an applicable drug efficacy notice is in violation of section 505 (new drugs) and/or section 502 (misbranding) of the act. 21 C.F.R. § 130.40(a) (1973), now codified at 21 C.F.R. § 310.6(a) (1977). Cf. Weinberger v. Hynson,