Nehmer v. United States Veterans' Administration

712 F. Supp. 1404, 1989 WL 52821

• Court: District Court, N.D. California
• Decided: May 15, 1989
• Docket: Civ. A. C 86-6160 TEH
• Status: Published

Opinion

ORDER

THELTON E. HENDERSON, District Judge.

This matter comes before the Court on cross-motions for summary judgment. After careful and extensive consideration of the parties’ lengthy briefs and oral arguments of counsel, the Court hereby grants summary judgment in part to plaintiffs and summary judgment in part to defendants. The Court finds that defendants failed to comply with the Veterans’ Dioxin and Radiation Exposure Compensation Standards Act, 38 U.S.C. § 354 note (hereinafter “Dioxin Act” or “Act”) by erroneously interpreting two provisions of the Act. As a result, a portion of the regulation promulgated by the VA and codified at 38 C.F.R. § 3.311 is invalid, and the Court remands this case to the Veterans Administration for further proceedings not inconsistent with this opinion.

I. Factual Background.

This lawsuit is another round in the conflict between Vietnam Veterans and the United States Government over Agent Orange, a chemical defoliant used by the United States Armed Forces in Vietnam during the 1960’s to clear dense jungle land. Many veterans believe that their exposure to the chemical dioxin (contained in Agent Orange) has caused them to contract several debilitating diseases, particularly soft tissue sarcoma (“STS”) (malignant tumors that form in muscle, fat, or fibrous connective tissue) and porphyria cutanea tarda (“PCT”) (deficiencies in liver enzymes). In the late 1970’s, veterans filed a class action lawsuit against the manufacturers of Dioxin and the United States to obtain compensation for their alleged injuries. In 1984, the manufacturers agreed to establish a $180 million fund to compensate class members. In Re Agent Orange Product Liability Litigation, 597 F.Supp. 740 (E.D.N.Y.1984) aff'd 818 F.2d 145 (2d Cir.1987). However, actions against the United States were dismissed; the Second Circuit applied the doctrine developed in Feres v. United States, 340 U.S. 135, 71 S.Ct. 153, 95 L.Ed. 152 (1950) to bar recovery for injuries suffered through military service. In Re Agent Orange Product Liability Litigation, 818 F.2d 194 (2d Cir.1987).

Veterans also pursued their claims against the United States in another forum: they sought disability compensation from the Veterans’ Administration, claiming that the diseases they incurred were caused by exposure to Agent Orange during military service in Vietnam. However, the VA has consistently taken the position that only one disease — chloracne, a skin condition— arises from exposure to Agent Orange. Accordingly, the VA has routinely denied compensation for veterans who allege that exposure to Agent Orange has caused diseases other than chloracne. As of October 1, 1983, 9170 veterans filed claims with disabilities that they allege were caused by Agent Orange; 7709 were denied compensation because the VA found that the claimed diseases were not service connected. House Report No. 98-592, reprinted in U.S.Code Cong. & Adm.News, 98th Cong. 2d Sess., 1984, at 4452.

In 1984, Congress enacted the Veterans’ Dioxin and Radiation Exposure Compensation Standards Act (Dioxin Act), 38 U.S.C. § 354 note. The Act was passed amidst veterans’ “concern[s] about possible long-term health effects of exposure to herbicides containing dioxin,” section 2(1), as well as “scientific and medical uncertainty” regarding the long-term health effects of Agent Orange exposure. Section 2(l). 1 The Act’s purpose is to ensure that disability compensation is provided to veterans “for all disabilities arising after [service in Vietnam] that are connected, based on sound scientific and medical evidence, to such service.” Section 3. 2

The Act dramatically alters the process governing Agent Orange disability claims. Rather than have the VA determine in individual adjudicatory proceedings whether a particular veteran’s claimed disease was caused by Agent Orange exposure, the Act authorizes the Administrator of the VA (“Administrator”) to conduct rulemaking to determine which diseases will be deemed service connected for all diseases claimed to be caused by Agent Orange exposure. 3

To achieve these purposes, the Act requires the VA to appoint an advisory committee composed of experts in dioxin, experts in epidemiology, and interested members of the public to review the pertinent literature on dioxin and submit periodic recommendations and evaluations to the Administrator. Section 6. 4 The Act also compels the Administrator to adopt regulations governing the evaluation of the scientific evidence. Id. at Section 5(b)(1)(A). Finally, after receiving the recommendations of the Advisory Committee and other members of the public, the Act requires the Administrator to promulgate a regulation identifying diseases to be deemed service connected, based on “sound scientific and medical evidence”. Section 5(b)(2)(A)(i).

The Advisory Committee was established on March 4, 1985. The committee included four lay members with a demonstrated interest in dioxin issues, three recognized specialists on the health effects of dioxin, and five scientists with expertise in epidemiology, three of whom also had extensive experience studying the effects of dioxin.

On April 22,1985, the Administrator published a proposed rule. 50 Fed.Reg. 15848. The rule sets forth five factors to govern evaluation of the scientific evidence. Id. The rule also proposes service-connection for only one disease — chloracne manifesting itself within three months of a veteran’s service. The proposed rule states that “[s]ound scientific and medical evidence does not support a causal association between dioxin exposure” and any other diseases. Id. at 15849-15850.

Also on April 22, 1985, the Advisory Committee met for the first time. The Committee received background information on the Act and scientific studies on the effects of Agent Orange exposure. At the second meeting on June 24-25, 1985, the Committee reviewed a number of studies on dioxin’s effects on human populations and received written comments on the proposed regulation. The Committee recommended that the Administrator adopt the five factor guidelines in the proposed rule and agreed with the VA that only chloracne should be deemed service-connected. The Committee transmitted its findings to the Administrator.

On August 26, 1984, the VA published a final regulation identical to the proposed rule. 38 C.F.R. 3.311a(d) states that “sound scientific and medical evidence does not establish a cause and effect relationship between dioxin exposure” and any other disease but chloracne. The regulation does not preclude a veteran from proving in an individual case that a claimed disease was caused by Agent Orange exposure, but as of December 1987, over 31,000 veterans have been denied compensation under this regulation. Nehmer v. U.S. Veterans’Administration, 118 F.R.D. 113, 120 (N.D.Cal.1987) (order granting class certification).

Following adoption of the regulation, the Advisory Committee has continued to evaluate scientific studies, reviewing over seventy studies, reports, and articles on dioxin. However, the Committee has not recommended that the Administrator amend the regulation to grant service connection to any other disease.

II. Statement of the Case.

On February 2,1987, plaintiffs, (Vietnam veterans and survivors of veterans) filed this action. Plaintiffs allege that the final regulation is invalid because the Advisory Committee and the VA violated provisions of the Dioxin Act, the judicial review provisions of the Administrative Procedure Act, (APA) 5 U.S.C. §§ 701-706, the Fifth Amendment of the United States Constitution, and the Federal Advisory Commission Act (FACA) 5 U.S.CApp. §§ 1-15. On December 22,1987, this Court certified a class to litigate these claims. Nehmer v. Veterans’ Administration, 118 F.R.D. 113 (N.D.Cal.1987). The class consists of all current or former service members (or their survivors) who are eligible to apply for benefits based on dioxin exposure or who have already applied and been denied claims for benefits based on dioxin exposure. Both plaintiffs and defendants now move for summary judgment. Because this suit challenges a federal agency’s compliance with statutory provisions, and this court’s review of the agency’s action is limited for the most part to the administrative record, this case is well suited for summary judgment.

Plaintiffs challenge the Administrator’s and the Advisory Committee’s compliance with the statute on the following grounds:

1) The Advisory Committee failed to thoroughly review the pertinent scientific studies, and wrongfully chose not to review studies on animal populations, even though these studies allegedly demonstrate a link between dioxin exposure and diseases other than chloracne;

2) The VA’s five factor guidelines for evaluation of scientific evidence wrongfully precluded consideration of animal studies;

3) The VA wrongfully required that the scientific evidence demonstrate a “cause and effect” relationship between Agent Orange exposure and claimed diseases, instead of using the less demanding standard that there be a “statistical association” between Agent Orange exposure and claimed diseases;

4) The VA erroneously failed to give the benefit of the doubt to the veterans when confronted with conflicting evidence of approximately equal weight;

5) the Advisory Committee failed to produce and transmit accurate committee minutes to the Administrator, so that the Administrator’s review of the Committee’s conclusions was distorted.

The Court finds that the Advisory Committee’s review of the scientific literature as well as the VA’s five factor test was not arbitrary and capricious or otherwise contrary to law. The court also holds that the Advisory Committee’s failure to transmit perfectly accurate minutes to the VA does not constitute grounds for invalidating the Dioxin regulation. However, the Court agrees with plaintiffs that the Administrator’s adoption of the cause and effect test and failure to give the benefit of the doubt to veterans violated the Dioxin Act. The Court further determines that these errors sharply tipped the scales against the claims of veterans. Given the congressional finding of substantial scientific uncertainty regarding the effects of Agent Orange, we hold that these errors are not harmless; they may very well account for the conclusion that the Administrator reached in the Dioxin regulation.

Accordingly, we hereby void 38 C.F.R. § 311(d), the portion of the Dioxin regulation that denies service connection for all other diseases. We also void all benefit decisions made under the 38 C.F.R. § 311(d), and remand to the VA for further proceedings not inconsistent with this opinion.

III. Subject Matter Jurisdiction.

Defendants first argue that we are precluded from even reviewing the Dioxin regulation by virtue of 38 U.S.C. § 211(a). That statute precludes judicial review of “[t]he decisions of the Administrator [of the VA] on any question of law or fact under any law administered by the Veteran’s Administration providing benefits for veterans ...”

Despite the seemingly all inclusive language of the statute, courts have devised exceptions to allow judicial review of particular types of VA determinations. For example, the statute does not bar constitutional challenges to VA action. Johnson v. Robison, 415 U.S. 361, 94 S.Ct. 1160, 39 L.Ed.2d 389 (1974). Many courts, including the Ninth Circuit, have also held that the statute does not preclude challenges to the statutory validity of a particu lar regulation, finding that section 211(a) was designed “to prevent judicial review of decisions of the VA on individual cases,” but not to bar challenges to VA regulations on constitutional and statutory grounds. See Evergreen State College v. Cleland, 621 F.2d 1002, 1008 (9th Cir.1980), allowing educational institution and veterans to challenge educational benefit restrictions on statutory grounds. 5 But see Gott v. Walters, 756 F.2d 902, 909 (D.C.Cir.1985), vacated and rehearing granted 791 F.2d 172, (D.C.Cir.1985) remanded with instructions to dismiss as moot 791 F.2d 172 (D.C.Cir.1985).

Defendants contend that a recently decided Supreme Court case interprets section 211(a) to bar this challenge. In Traynor v. Tumage, 485 U.S. 535, 108 S.Ct. 1372, 99 L.Ed.2d 618 (1988), a statute permitted veterans to apply for an extension of time to use unused educational benefits as long as veterans were prevented from timely use of the benefits by a disability which in turn was not the result of willful misconduct. 38 U.S.C. § 1662(a)(1); 108 S.Ct. at 1386. The VA determined that alcoholism was willful misconduct, so that the extension of time would not be granted to alcoholics. Id. at 1376.

Plaintiffs alleged that this determination violated the Rehabilitation Act, 29 U.S.C. § 794, a statute that prohibits discrimination on the basis of handicap. Defendants countered that section 211(a) barred judicial review of plaintiffs’ claim. The Supreme Court found that section 211(a) does not preclude review of claims that a VA regulation violates a federal statute of general applicability, such as the Rehabilitation Act. Id. at 1380.

However, the court’s general discussion of the underlying policies of section 211(a) suggests that the instant claims may be precluded. The court identified two policies served by section 211(a): it prevents veterans from burdening courts with benefit denial claims, and it avoids enmeshing courts with complex technical decisions regarding VA policy. 108 S.Ct. at 1379, citing Robison, supra, 415 U.S. at 370, 94 S.Ct. at 1167.

The latter policy is implicated in this action. In contrast to the challenge in Traynor, the instant case involves a challenge to a statute that the VA has sole responsibility to administer. Moreover, the challenge advanced by plaintiffs does require the court to become enmeshed in highly technical, complex determinations to review the dioxin regulation and the procedures used to adopt the regulation. Thus, Traynor suggests, albeit in dicta, that section 211(a) bars the instant challenge.

However, Congress has rescued plaintiffs from judicial preclusion. Congress recently enacted S.ll, the Veteran’s Judicial Review Act, which then-President Reagan signed into law on November 18, 1988. The Act amends section 211(a) to explicitly authorize judicial review in the Federal Circuit Court for all statutory challenges to VA rulemaking filed after September 1, 1989. 38 U.S.C. § 223(c).

Because the Act only applies to actions filed after September 1989, the court must answer this question: what does the passage of the Act imply about Congressional intent for claims filed before that date? Plaintiffs argue, and the Court agrees, that this Act was intended to clarify pre-exist-ing Congressional intent to allow statutory challenges to VA rules, and not to reverse a prior policy embodied in section 211(a) to bar such challenges.

Senator Cranston, a sponsor of the Act, provides substantial support for this position. In a floor statement submitted on behalf of S.ll, Senator Cranston notes that section 223 was intended to “[cjlarify the state of the law” on section 211(a). 134 Cong.Record, S 16643 (October 18, 1988). Senator Cranston observed that prior to this legislation, the federal circuits disagreed as to whether section 211(a) barred review of VA regulations. Id. He then stated that the Supreme Court failed to resolve the matter in Traynor, supra, since the court only addressed challenges to VA regulations grounded in other federal statutes of general applicability. Id. Cranston then noted that the VA itself supported judicial review of regulations and supported the circuits which have interpreted section 211(a) to allow such review, but that the Justice Department continued to raise section 211(a) in litigation despite the VA’s position. Id. Thus, in the absence of a Supreme Court opinion resolving that judicial review was appropriate for challenges to regulations, Congress enacted this legislation to resolve the split in circuits and clarify pre-existing law. Id.

Senator Cranston also addressed the question of whether the Act would affect actions filed before that date. Senator Cranston stated that “no pending case would be adversely affected” by the legislation, because “once jurisdiction attaches in a case, a court continues to have jurisdiction over the matter until decision has been reached.” Id. at 16650. This statement presupposes that jurisdiction exists for challenges prior to enactment of this legislation, further confirming that the Act affirmed, rather than reversed, prior law on the question of judicial review.

These floor statements strongly suggest that judicial review has always been available for statutory challenges to VA action. Indeed, there is no legislative history suggesting the contrary — that Congress passed the Act to change existing law and introduce a new right to judicial review that had not been previously afforded. Therefore, the court finds that even if the Supreme Court’s discussion of section 211(a) in Traynor would preclude judicial review here, the enactment of S.ll resurrects the right to review that was arguably lost by the Traynor opinion.

We find that this action is not precluded by section 211(a), and we turn now to the merits of the action. 6

7F. Advisory Committee’s Alleged Noncompliance with the Act.

A. Pertinent Studies.

Plaintiffs first allege that the Advisory Committee violated the Dioxin Act by 1) failing to review a sufficient number of pertinent studies, and 2) excluding review of studies on animal populations. Plaintiffs contend that this alleged dereliction so tainted the Advisory Committee’s conclusions — and in turn, those of the Administrator — that the regulation should be invalidated.

As noted, the Act assigns a role of substantial importance to the Advisory Committee. The Committee is charged with the task of making “findings and evaluations regarding scientific studies” and submitting its findings to the Administrator on a periodic basis. 38 U.S.C. § 354 note section 6(d)(3). Statements by numerous senators and representatives reveal that Congress intended that the Committee would perform a thorough evaluation of the pertinent scientific studies, and that the Committee’s review would provide an important factual underpinning for the Administrator’s ultimate determination. See, e.g., Statement of Senator Pressler, 130 Cong. Rec. S 6163-6164 (May 22,1984), Statement of Rep. Montgomery, 130 Cong.Rec. H 11161 (Oct. 3, 1984).

Plaintiffs allege that the Committee failed to adhere to this mandate. Plaintiffs note that prior to the adoption of the regulation, the Committee held only one substantive meeting and reviewed only seven studies out of the hundreds conducted on the health effects of dioxin. Indeed, the Committee did not even review as many studies as the Administrator reviewed to reach its conclusion that the proposed rule was consistent with the scientific evidence. Furthermore, while the Committee has reviewed many more studies since promulgation of the rule, plaintiffs contend that the Committee has still only reviewed a small percentage of the total scientific literature. As a consequence, plaintiffs argue that the Administrator made his decision without the benefit of a thorough literature review conducted by a neutral body of experts.

Defendants argue that the Advisory Committee’s choice of which (and how many) studies to review is entitled to great deference. 7 We agree: a group of experts’ decisions regarding which scientific studies are the most pertinent to understanding the health effects of Agent Orange exposure is perhaps the epitome of a scientific determination which this court is ill-equipped to review. While plaintiffs attempt to characterize their claims as procedural challenges that only require the court to judge the Committee’s action against clear statutory standards, in reality plaintiffs are challenging the scientific wisdom of the Committee’s determinations. Indeed, plaintiffs rely heavily on expert declarations describing what they consider to be a thorough review of the scientific literature in an effort to persuade the court that the Committee failed to conduct this mandated thorough review.

In reviewing expert determinations such as this one, we follow the lead of numerous courts that have been asked to review the scientific validity of agency action. As one court stated in a challenge to the EPA’s regulation reducing the lead content of gasoline, the court must review the agency decision “not as the chemist, biologist or statistician that we are qualified neither by training nor experience to be, but as a reviewing court exercising our narrowly defined duty of holding agencies to certain minimal standards of rationality.” Ethly Corp. v. EPA, 541 F.2d 1, 36 (D.C.Cir.1976) (en banc). See also Nat. Cattlemen’s Ass’n v. United States E.P.A., 773 F.2d 268, 271 (10th Cir.1985), granting deference to EPA decision regarding environmental effects of Coyote predacide; Sierra Club v. Costle, 657 F.2d 298, 333-334 (D.C.Cir.1981), deferring to EPA’s use of a computer model to forecast results of emission limitations.

Judged against this minimum standard of rationality, we cannot say that the Advisory Committee’s selection of studies to review was arbitrary and capricious. At the June 24-25, 1985 meeting, the Committee reviewed seven studies. These seven studies included the “Ranch Hand” study, which directly analyzed the health effects of Agent Orange on approximately 1000 veterans who flew cargo in and out of Vietnam and were heavily exposed to the defoliant. 8 The committee also reviewed several studies specifically addressing the incidence of soft tissue sarcoma among veterans, as well as studies of veteran populations located in Massachusetts, New York, and New Zealand. Minutes of June 24-25, 1985 transcripts at 1-2. These studies were selected because 1) they were recently conducted, 2) included studies on veteran populations, and 3) included studies supporting the connection between Agent Orange and various diseases. Declaration of Theodore Colton, Chairman of the Dioxin Panel of the Advisory Committee at 1114.

Moreover, the Committee has reviewed numerous other studies following enactment of the regulation. During the March 3-4,1986 meeting, the Committee reviewed seven studies on Vietnam veterans; on the November 17-18,1986 meeting, it reviewed eight studies, at least two of which concerned Vietnam veterans. On April 27-28, 1987, it reviewed twelve more studies, including ones conducted by EPA and the Center for Disease Control. Plaintiffs do not argue that any of these studies are biased against veterans, or do not represent a fair sample of studies on human populations.

Instead, plaintiffs claim that the members of the Advisory Committee were required to personally read all studies on dioxin either prior to or after adoption of the regulation. According to plaintiffs, that task would require reading the 735 documents listed in a study that the EPA performed on dioxin, as well as numerous studies identified by plaintiffs’ experts.

We cannot agree that Congress required the Committee to perform this Herculean task. As Senator Simpson states, “[t]he conferees recognize that it would be impossible for the Advisory Committee ... to review or consider all of the thousands of studies which have been conducted on the subject of dioxin.” 130 Cong.Rec. S 13592 (Oct. 4, 1984). Instead, the Act commands the Committee to select the studies it deems most pertinent, and transmit its findings to the Administrator. Section 6(d)(3). Moreover, such a comprehensive review would have largely duplicated prior efforts of the Administrator. Under P.L. 96-151, § 307(a)(1) as amended, 38 U.S.C. § 219 note, Congress instructed the VA to “conduct a comprehensive review and scientific analysis” of the relevant literature. The VA has performed that analysis, and has updated the review on an approximately yearly basis. See Statement of Senator Cranston, 130 Cong.Rec. S él46 (May 22, 1984). It is difficult to believe that Congress intended that this formidable task be performed twice.

We also agree with defendants’ common sense observation that these committee experts did bring a wealth of background knowledge to the task. For example, Committee member Dr. Colton states in a declaration that he has reviewed numerous studies on dioxin as an editor of a scientific journal, as an expert witness in Agent Orange product liability litigation, and as a member of government-sponsored committees on Vietnam Veterans’ health. Colton Declaration at ¶¶ 5-10. Similarly, committee member Dr. Melvin studied the effects of dioxin for many years at the invitation of numerous businesses, foreign governments, and the United States government. He has presented several papers on dioxin; his work has required him to periodically review the pertinent literature on dioxin. Melvin Declaration at ¶¶ 1, 5-11.

While we agree with plaintiffs that Congress intended a “collegial” review of the pertinent studies, and the committee could not therefore rely exclusively on individual members’ expertise as an alternative to conducting a group review of the scientific literature, we believe Congress did not intend to preclude the committee from relying on individual members’ expertise to narrow and streamline the committee’s work to some extent. We would be troubled if the committee relied exclusively on individual members’ knowledge and failed to review a reasonable number of studies collectively. But that is not the case here; the committee has reviewed numerous relevant studies, and individual members’ expertise has supplemented, rather than replaced, that collegial review. 9

Applying due deference to the Committee’s determination of which studies to review, the court finds that the Committee’s selection of studies did not violate the Dioxin Act.

B. Animal Studies.

Plaintiffs also argue that the Committee erred by failing to review studies that traced the effects of Dioxin exposure on animals. Plaintiffs have submitted several declarations of expert witnesses who state that the exclusion of animal studies resulted in an incomplete review of the scientific literature. Plaintiffs’ experts agree that the human population exposed to dioxin is too small to provide a basis for an informed conclusion. See declarations of Thomas A. Gasiewicz at ¶¶ 12,15; declaration of Peter C. Kahn at ¶¶ 8, 10; declaration of James R. Olson at 1112 (noting that the Environmental Protection Agency reviewed numerous animal studies in its assessment of the effects of Dioxin).

Leaving aside the question of whether these expert declarations are admissible evidence, 10 the court must again defer to the scientific judgment of the Committee. Defendants argue that the Committee made a sound judgment that animal studies “have little relevance to human reaction.” Deposition of Committee Member Leonard T. Kurland at 140. Members of the Committee agreed that animals “don’t provide a good model” in light of the “500-fold variation” in animal responses to Dioxin. Kur-land deposition at 101-102. Moreover, the Committee members believed that they had “a considerable number of human experience type studies that are far more relevant and far more pertinent to the question of Vietnam veteran exposure.” Deposition of Frederic R. Conway III, Executive Secretary to the Committee, at 191-192.

Defendants have submitted declarations of several Committee members providing additional reasons for the exclusion of animal studies. Dr. Yanders, who has reviewed “hundreds” of animal studies, states that while animal studies are “helpful in showing the range of biological responses possible from exposure to dioxin,” they are “not very helpful in showing what to expect in terms of human responses at the very low levels of exposure which may have occurred in Vietnam.” Yanders Dee-laration at II19. Committee member Dr. Lathrop states that “we now know that the human system reacts quite differently to dioxin exposure than do various species of animals.” Declaration at II29. Similarly, Committee member Walter N. Melvin states that “the problems of extrapolation [from animals to humans] make [animal] studies only marginally useful” in determining dioxin’s health effect on humans. Declaration at II15. See also Judge Wein-stein’s decision in In Re Agent Orange Product Liability Litigation, 611 F.Supp. 1223, 1241 (E.D.N.Y.1985), aff'd 818 F.2d 145 (2d Cir.1987), stating that animal studies “are of so little probative force and are so potentially misleading as to be inadmissible” to support claims of Vietnam veterans against Agent Orange manufacturers.

Plaintiffs deride the committee members’ declarations as “post hoc ” rationalizations that cannot be used to support the agency’s action. Burlington Truck Lines v. U.S., 371 U.S. 156, 168-169, 83 S.Ct. 239, 246, 9 L.Ed.2d 207 (1962) (holding that post hoc rationalizations offered by counsel may not be used to justify agency decisions). We do not construe these declarations as post hoc justifications developed for the purposes of this litigation. Instead, the declarations elucidate the basis for the Committee’s determinations. Asarco, supra, 616 F.2d at 1160. In addition, contrary to plaintiffs’ contention, the declarations are unanimous on the question of whether consideration of animal studies would be useful to their task.

These declarations from Committee members provide a rational basis for the Committee’s decision. Plaintiffs’ experts disagree with these scientists, but it is not this court’s role to weigh the conflicting expert testimony and come to an independent conclusion. Asarco, supra, 616 F.2d at 1160-1161, holding that district court erred by weighing the conflicting evidence and substituting its judgment for that of the agency. Therefore, applying the mandated deference to the committee, we uphold the Committee’s exclusion of animal studies.

Plaintiffs have a fallback position: they contend that the committee was required to make a formal finding that animal studies were not relevant to their task. Their failure to make a decision to exclude these studies on the record cannot be cured, they contend, with post hoc declarations and deposition testimony explaining the basis of the Committee’s action.

Plaintiffs are correct that the Committee never formally “voted” to exclude animal studies or issued findings to explain the exclusion of animal studies, though the Committee did receive presentations which discussed animal studies, and reviewed human studies which discussed the findings of animal studies. See, e.g., Transcript of April 22, 1985 meeting at 140-170 (statement by Dr. Barclay Shephard), declaration of Dr. Melvin at ¶ 21, declaration of Dr. Yanders at ¶ 19. However, there is nothing in the Dioxin Act that requires the Advisory Committee to make findings about its selection of pertinent studies.

Moreover, the Administrative Procedure Act does not require the Committee to make formal findings on the record — for the simple reason that the APA does not apply to the Committee. The test for whether an entity is considered an “agency” under the APA is whether the entity has “substantial independent authority in the exercise of specific functions.” Soucie v. David, 448 F.2d 1067, 1073 (D.C.Cir.1971). The Advisory Committee has no substantial independent authority; its sole role is to review scientific literature and provide recommendations to the Administrator, who alone has the authority to draft and adopt the Dioxin regulation. See Washington Research Proj. Inc. v. Department of H., E. & W., 504 F.2d 238, 246-248 (D.C.Cir.1974) (holding that panel of scientists who evaluate research grant applications is not an agency under the APA); Wolfe v. Weinberger, 403 F.Supp. 238, 241 (D.D.C.1975) (scientific panel appointed to provide recommendations to the Food and Drug Administration regarding over-the-counter drugs is not an agency under the APA).

Finally, plaintiffs argue that even if the Act did not require formal findings, this court should require such findings as necessary for the task of reviewing the Committee’s action. Again, we disagree; the declarations and deposition testimony of Committee members provide an adequate explanation of why animal studies were excluded. Therefore, the court will not invalidate the regulation due to the Committee’s failure to review studies on animals.

V. Improper Guidelines re Evaluation of Studies.

Plaintiffs also argue that the Administrator erred by promulgating a guideline that precluded the Administrator from reviewing animal studies.

The Dioxin Act requires the Administrator to establish guidelines for the evaluation of the findings of scientific studies on the health effects of dioxin. 38 U.S.C. § 354 note section 5(b)(1)(A). In compliance with this section, the Administrator adopted a five factor test for evaluating scientific studies. These factors are whether the results of the particular study are 1) statistically significant and replicable, 2) able to withstand peer review, 3) supported by well-described methodology, 4) applicable to the veteran population of interest, and 5) accord with the view of the Advisory Committee. 38 C.F.R. § 1.17(b)(lM5).

The first three factors are drawn from section 5(b)(1)(A) of the Dioxin Act itself, and plaintiffs do not criticize these factors. Nor do plaintiffs criticize the fifth factor— whether the Advisory Committee viewed the particular study as valid. However, plaintiffs do challenge the fourth factor— whether the study is “applicable to the veteran population of interest.” Plaintiffs contend that this factor operated to exclude VA review of animal studies, and that the exclusion of animal studies was arbitrary and capricious.

This argument is identical to plaintiffs’ claim against the Advisory Committee— that exclusion of animal studies was arbitrary and capricious. The Committee experts’ conclusions that animal studies are not probative due to wide interspecies variation provide a rational basis for the Administrator’s decision. We do not hold that the decision was arbitrary and capricious. 11

VI. Improper Use of Cause and Effect Standard.

Plaintiffs next argue that the VA erred by requiring proof of a “cause and effect” relationship between dioxin exposure and various diseases. Plaintiffs contend that Congress intended only that there be a significant “statistical association” between dioxin and a particular disease in order to grant service connection for that disease.

The difference between these two standards is elucidated by both plaintiffs’ and defendants’ experts. A statistical association “means that the observed coincidence in variations between exposure to the toxic substance and the adverse health effects is unlikely to be a chance occurrence or happenstance.” Declaration of David Ozanoff at ¶ 5. The cause and effect relationship “describes a much stronger relationship between exposure to a particular toxic substance and the development of a particular disease than ‘statistically significant association’ does”. Declaration of Christopher Cox at 117. Cox adds that “it is often not possible, based on the sound scientific and medical evidence that currently exists, to demonstrate that there is a cause and ef-feet relationship” between human exposure and disease, because of “society’s aversions to experimentation on humans.” Id.

For a causal relationship to be found, a scientist would require a strong and consistent level of association and a plausible explanation of the biological mechanism at work. Ozanoff Declaration at 116. Or, as committee member Dr. Lathrop states, “finding a higher than expected incidence of a disease among a particular population alone does not demonstrate a causal relationship ...” Lathrop Declaration at 1125. To find a causal relationship, “[i]t is necessary to look for many [other] factors,” such as “a linear progression between the greater appearance of symptoms and the increase in dose levels of exposure.” Id.

Defendants do not dispute that the VA required proof of a causal relationship to grant service connection. 12 The regulation explicitly states that “[s]ound scientific and medical evidence does not establish a cause and effect relationship between dioxin exposure” and any diseases except some cases of chloracne. 38 C.F.R. § 3.311(d), emphasis added. Based on that finding, the VA denied service connection status for all other diseases.

Defendants argue instead that the causal relationship test conformed to congressional intent. Thus, this dispute hinges on statutory interpretation: did Congress intend for the Administrator to use a causal relationship test, or a statistical association test?

Before turning to this issue, an initial observation is appropriate: the resolution of this matter is primarily a function of statutory interpretation. 13 Plaintiffs are not alleging that the use of a cause and effect test is scientifically invalid, and therefore arbitrary and capricious; they contend instead that this test violated express congressional command. Thus, while some deference is appropriate to an agency’s interpretation of a statute that it is authorized to administer, Chevron U.S.A. v. Natural Res. Def. Council, 467 U.S. 837, 844, 104 S.Ct. 2778, 2782, 81 L.Ed.2d 694 (1984), this deference is “constrained by [the court’s] obligation to honor the clear meaning of a statute as revealed by its language, purpose and history,” Teamsters v. Daniel, 439 U.S. 551, 566 n. 20, 99 S.Ct. 790, 800 n. 20, 58 L.Ed.2d 808 (1979). See also Abramowitz v. U.S.E.P.A., 832 F.2d 1071, 1077 (9th Cir.1987): while the “reviewing court cannot substitute its judgment for that of the administrative agency, it is equally well-established that a court cannot defer to agency discretion when the intent of the Act is clear.”

A. Statutory Language.

We turn first to the language of the Act. Because there is no single section of the Act which directly addresses this disputed issue, we cite various sections in an effort to piece together Congressional intent.

In the findings section of the Act, Congress notes that “[t]here is scientific and medical uncertainty regarding” the long-term health effects of Dioxin exposure. 38 U.S.C. § 354 note section 2(2). Congress states that “[t]here is some evidence that chloracne, porphyria cutanea tarda, and soft tissue sarcoma are associated with exposure to certain levels of dioxin as found in some herbicides.” Section 2(5), emphasis added. Congress then notes that the Ranch Hand study, conducted on veterans directly exposed to dioxin, “contained the conclusion ‘that there is insufficient evidence to support a cause and effect relationship between herbicide exposure and adverse health in the Ranch Hand group at this time.’ ” Section 2(7), emphasis added.

Congress next states that the purpose of the Act is to ensure compensation for all “disabilities ... that are connected, based on sound scientific and medical evidence, to military service in Vietnam.” Section 3. To carry out this purpose, Congress requires the VA to establish guidelines for the review of the scientific evidence. Section 5(b)(1)(A). The guidelines require the Administrator to take into account “whether the results [of the scientific studies] are statistically significant, are capable of replication, and withstand peer review.” Id.

These sections of the Act do not clearly reveal congressional intent. The findings advert to scientific evidence on the subject, noting that there is some evidence of an “association”, but a lack of evidence for a causal relationship. Thus, Congress uses both standards, without explicitly choosing between them. Similarly, the scientific guidelines language in section 5(b)(1)(A) does not demonstrate a choice among the competing standards. The requirements that a study be statistically significant, capable of replication, and able to withstand peer review are “general grounds upon which a scientist would evaluate any study.” Declaration of Dr. Yanders at II22. They instruct the VA and the Committee on the methodology to be used in evaluating the probative value of a study; they are silent on the standard to be used in determining the requisite quantum of proof necessary to grant service connection.

B. Legislative History.

Because the language of the Act is at best ambiguous, at worst silent on this issue, we turn to applicable legislative history. 14 Blum v. Stenson, 465 U.S. 886, 896, 104 S.Ct. 1541, 1547, 79 L.Ed.2d 891 (1984); State of Alaska v. Lyng, 797 F.2d 1479, 1483 (9th Cir.1986). That history supports plaintiffs’ position. For example, Senator Simpson, states that “[i]t is impossible ... to make any absolute determination as to cause and effect.” 130 Cong. Rec. S 6156-6157 (May 22, 1984). Senator Byrd noted that prior to the Act, “overly onerous burdens of proof ... have been required of veterans regarding their exposure to dioxin.” 130 Cong.Rec. S 6166. Senator Spector added that “[t]he Federal Government has imposed an impossible burden of proof” on veterans attempting to demonstrate service connection. 130 Cong. Rec. S 6152.

In response to these and other comments, Senator Simpson proclaimed that the Act “would provide very real and substantive change in consideration and adjudication of claims based on exposure to agent orange ...” Id. at 6155. That change would come about through rulemaking on dioxin, which in turn would rely upon a scientific determination of “the possibility of an increased risk of certain diseases based on the many ongoing studies.” Id. at 6157, emphasis added.

These comments were echoed in the House. Representative Applegate, chairman of a House subcommittee with jurisdiction over the Act, stated that the Advisory Committee’s task was to “identify medical conditions that could be related to exposure.” 130 Cong.Rec. H 11164 (Oct. 3, 1984). Similarly, House Veterans’ Affairs Committee Chairman Montgomery stated that the Act requires the Administrator to determine whether the evidence “indicates a correlation” between exposure and diseases. Id. at H 11161.

Chairman Montgomery amplified this remark in subsequent comments to the VA. He informed the VA that “service connection has always been the product of temporal rather than causal relationship” and that the VA should “not introduce a causal relationship requirement into the law.” AR Vol. 1. tab G.

These statements, and others that the Court has discovered in the legislative history, amply support the claim that Congress did not intend the VA to use a causal relationship. Instead, the comments overwhelmingly suggest that Congress intended service connection to be granted on the basis of “an increased risk of incidence”, or a “significant correlation” between dioxin and various diseases. 15

C. Prior VA and Congressional Practice.

The cause and effect test is also inconsistent with prior VA and congressional practice: both the VA and Congress have used a “statistical association” standard to grant service connection status for other types of diseases. 16 For example, the VA granted service connection for cardiovascular diseases incurred by veterans who suffered amputations of legs or feet. 38 C.F. R. § 3.310(b). The VA based its service connection determination on a report pre pared by the National Academy of Sciences.

The VA describes the report as showing that these amputations “result in a significant increase in the incidence of cardiovascular disease.” 49 Fed.Reg. at 50340. The report did not demonstrate a causal relationship between amputation and heart disease; as the author states, it “was not designed to test specific hypotheses about the causes of the increased risk. Causation could well be multifactorial.” Richard A. Ryder, Amputations of Extremities and Cardiovascular Diseases at 17. Indeed, Ryder states that “[t]he reasons for the statistically significant relationship demonstrated between limb amputation and cardiovascular disorder are not obvious.” Id. at 18. He adds that cardiovascular diseases could well be caused by factors other than amputations, such as “sedentary lifestyle” and “psychosocial tension.” Id. Thus, the VA based service connection status on a study that demonstrated a statistically significant relationship, not a cause and effect relationship.

Similarly, when Congress itself granted service-connection status for diseases suffered by former prisoners of war, 38 U.S.C. § 312(b), Congress based its findings in part on a VA study of prisoners of war. That study merely showed “a higher incidence” of physical and psychological disorders that are “related” to the conditions of their imprisonment. H.Rep. No. 28, 97th Cong., 1st Sess., reprinted in 1981 U.S. Code Cong. & Ad.News at 1410, 1413-1414. The study did not try to establish a causal relationship.

Defendants’ efforts to rebut this evidence are unavailing. They first argue that the presumption for cardiovascular diseases was based on demonstration of a causal relationship, even though the causal mechanism was unknown. We must reject that conclusory claim, given the explicit findings of the report and the VA’s assessment of the report stating that only a significant increase in incidence was found.

Second, defendants argue that the POW presumption is distinguishable because it too rested upon a finding of a causal connection. We reject this assertion for the same reason; the legislative history does not reveal demonstration of a causal connection.

Defendants next argue that even if the service connection designations in these examples were based on proof less stringent than causality, Congress intended that the VA determine service connection status for Agent Orange related diseases under a different standard. The reason for this different treatment, the argument goes, is that the Dioxin Act is the first time Congress has instructed the VA to decide whether to grant service connection to a specific disease. Congress has previously given the VA the authority to add to the list of statutory presumptions for any disease, as long as the Administrator’s decision is based on “sound judgment.” 38 U.S.C. § 313(b). In contrast, under the Act, the VA has been explicitly charged with determining service connection for Agent Orange related diseases, and the standard to be used is not “sound judgment”, but “sound scientific or medical evidence” indicating a “connection” to Agent Orange exposure. 38 U.S.C. § 354 note section 5(b)(2)(A).

There are two problems with this argument. First, it is inconsistent with defendants’ prior admission that Congress intended the VA to act in accordance with prior principles and procedure. Second, defendants do not explain why these different formulations imply that Congress wanted dioxin-related diseases to be judged under a more stringent burden of proof. This Court discerns no meaningful difference between the standard of “sound judgment” and “ ‘sound scientific medical evidence’ demonstrating a ‘connection’ between exposure and disease”. We therefore reject this argument as well.

Finally, at oral argument defendants asserted that the use of the statistical association test would engender unscientific results by mandating a finding of service connection if even one study showed an association between Agent Orange and various diseases. We disagree: even under the statistical association test, the Advisory Committee and Administrator would be required to carefully examine the methodology of each study and determine whether its findings are “statistically significant, are capable of replication, and withstand peer review.” Section 5(b)(1)(A). Moreover, the Committee and the Administrator would still be required to weigh the scientific evidence cumulatively to avoid giving undue weight to a particular study. Thus, use of the less demanding standard would not contravene proper scientific method.

For all these reasons, we find that defendants misinterpreted the degree and type of proof that Congress required for service connection status. The legislative history, and prior VA and congressional practice, support our finding that Congress intended that the Administrator predicate service connection upon a finding of a significant statistical association between dioxin exposure and various diseases. We hold that the VA erred by requiring proof of a causal relationship. 17

VII. Benefit of the Doubt Policy.

Plaintiffs also argue that the Administrator violated the Dioxin Act by failing to apply a “benefit of the doubt” policy in the determination of which diseases are caused by Agent Orange. Plaintiffs contend if the Administrator faced an approximate balance of positive and negative evidence regarding whether a particular disease was caused by dioxin, the Act requires the VA to give the benefit of the doubt to veterans and grant service connection status to the particular disease. 18 Defendants concede that the Administrator did not apply a benefit of the doubt policy to their determinations, but argue that the Dioxin Act did not require him to do so.

When the VA adjudicates individual benefits claims and “there is an approximate balance of positive and negative evidence regarding the merits of an issue material to the determination of a claim, the benefit of the doubt” is given to the claimant. 38 U.S.C. § 354 note section 2(13); 38 C.F.R. § 3.102. The Dioxin Act states that when the Administrator determines through the rulemaking process which diseases are presumptively caused by dioxin, the Administrator “shall give due regard to the need to maintain the [benefit of the doubt policy] with respect to [individual claims adjudication].” 38 U.S.C. § 354 note section 5(b)(2)(A)(i).

Plaintiffs contend that this “due regard” provision means that the Administrator must apply the benefit of the doubt policy not only to individual claims adjudication, but to the aggregate rulemaking determination of which diseases are caused by Agent Orange.

Defendants argue that the “due regard” provision does not mandate that the VA apply the benefit of the doubt in the aggregate rulemaking proceeding. Instead, the due regard provision merely evinces congressional intent to preserve the benefit of the doubt policy for individual claims adjudication following enactment of the Act.

The ambiguity of this provision is puzzling. If Congress wished to mandate the benefit of the doubt policy for rulemaking, it could have simply inserted language such as “the VA shall apply the benefit of the doubt policy of 38 C.F.R. § 3.102 to the rulemaking required by this Act.” Similarly, if Congress wanted to ensure that the Act did not alter the benefit of the doubt policy for individual claims adjudication, it could have stated “nothing in this Act shall be construed to alter the benefit of the doubt policy of 38 C.F.R. § 3.102.” Given how easy it would have been for Congress to clearly pronounce its intent, we are puzzled that the language is so ambiguous. Unable to resolve the dispute by reading the statute, we turn to the legislative history-

B. Floor Statements.

Plaintiffs rely heavily on Senator Cran-ston’s discussion of the provision; Senator Cranston offered the amendment which inserted the due regard provision. Senator Cranston explained that the purpose of the provision was to ensure that no veteran would be made procedurally “worse off” as a result of the substitution of aggregate rulemaking for individual claims adjudication. 130 Cong.Rec. S 6158 (May 22, 1984). Senator Cranston explained that unless the Administrator is “guided” by the benefit of the doubt policy in the rulemaking process, the “individual veterans who might prevail” on particular claims “where the evidence ... is relatively evenly balanced and the reasonable doubt standard tips the balance in the claimant’s favor, would be worse off after the rulemaking.” Id. This is so because the Administration could determine in the rulemaking process that whole categories of diseases are not service related without applying the benefit of the doubt policy to any of those diseases. As Senator Cranston explains, “if the Administrator does not take that reasonable doubt standard into account, such a veteran claimant will no longer receive the benefit of the doubt on a critical issue in his or her case.” Id.

However, defendants also submit floor statements that favor their position. For example, Senator Simpson notes that the individual adjudicatory proceeding is “the only level at which we presently consider the approximate balance of positive and negative evidence regarding the merits of [an] issue material to the determination of the veteran’s claim.” Id. The due regard provision is not “intended to alter this present policy — either expand it or restrict it.” 130 Cong.Rec. S 6155.

Similarly, Representative Montgomery states “[w]hat this legislation says is that the Administrator must give due regard to the need to maintain, unchanged, the doctrine of reasonable doubt as it applies to individual cases.” 130 Cong.Rec. H 11161 (October 3,1984). However, the Act “does not say that the doctrine controls determinations made by the Administrator ... which are the subject of rulemaking procedures undertaken by the Administrator pursuant to this legislation.” Id.

These contradictory floor statements suggest that one legislative faction wanted to mandate benefit of the doubt, another faction did not. Congress may have resolved this split in an increasingly familiar way — by “fudging” the issue with ambiguous language.

Given the ambiguous language and contradictory legislative history, the court turns to a third tool of statutory construction: which interpretation accords with other language and the underlying purpose of the Act? 19

It is here that plaintiffs have a decided advantage. Defendants’ interpretation of the language (that due regard merely meant that Congress did not intend to change prior VA policy regarding individual claims adjudication) would only make sense if there were provisions in the Act which would otherwise cause the VA to change that policy. However, there are no such provisions in the Act. Thus, defendants’ interpretation would make the provision ineffectual surplusage. In contrast, plaintiffs’ interpretation gives the language vitality and meaning; the language adds a new instruction to the VA which is not provided in any other sections of the Act. When construing a statute, the court is bound to give effect “to all provisions, so that no part will be inoperative or superfluous, void or insignificant.” Sutherland, Statutory Construction, Vol. 2A, § 46.06 at 104 (4th Ed.); U.S. Army Eng. Center v. Fed. Lab. Rel. Auth., 762 F.2d 409, 416-417 (4th Cir.1985). This consideration counsels us to accept plaintiffs’ interpretation.

Second, plaintiffs’ interpretation best accords with the remedial purpose of the legislation. Congress replaced the prior system of individual claims adjudication with an aggregate rulemaking process because it did not believe that veterans were well served by this process. See section 2(12), noting that dioxin claims “present adjudicatory issues which are significantly different from issues generally presented in claims based upon the usual types of injuries incurred in medical service.” As Senator Simpson states, the Act was intended to ensure that veterans “have their exposure claims adjudicated under uniform and consistent regulations that incorporate rational scientific judgments”, as opposed to the prior system, in which the claims are “committed to the sound judgment of the VA’s adjudication officers” who decide them on “a case-by-case basis.” 130 Cong. Rec. S 13591 (October 4, 1984).

Given Congress’ view that an aggregate rulemaking process would better serve the interest of the veterans than would individual claims adjudication, it is very unlikely that Congress would legislate a system in which veterans would actually be made worse through the rulemaking process. Yet as Senator Cranston notes, the failure to apply benefit of the doubt at the rule-making stage would make veterans worse off; whole categories of diseases could be deemed not connected to military service even if there were an approximate balance of positive and negative evidence. In contrast, if the Act had never been passed, veterans could at least present their individual claims and receive the benefit of the doubt on whether Agent Orange exposure caused their particular affliction. Thus, under defendants’ interpretation, Congress would take away with one hand (benefit of the doubt) what it tried to give with the other (a more rational and uniform process governing dioxin exposure claims).

We cannot ascribe this result to the Act. Thus, even granting deference to the VA’s construction of this provision, Alaska v. Lyng, 797 F.2d 1479, 1481 (9th Cir.1986), cert denied 480 U.S. 945, 107 S.Ct. 1603, 94 L.Ed.2d 789 (1987), we find that defendants’ interpretation of this provision is not plausible, and therefore violates the Act. Accordingly, we find that defendants violated the Act by failing to apply the benefit of the doubt policy to the aggregate rule-making determination. 20

VIII. Inadequate Minutes.

Plaintiffs' final claim is that the Advisory Committee failed to send complete and accurate minutes of their meetings to the Administrator. As a result, the Administrator allegedly did not receive an accurate report of the Committee’s views.

We find little merit in this claim. We have reviewed all of the alleged omissions detailed in paragraphs 96-111 of Plaintiff’s’ Statement of Undisputed Facts, and find that the vast majority of the committee discussions omitted in the minutes deal with issues such as committee procedure and background briefing on the Dioxin Act. The failure of the minutes to mention these discussions is not significant; the Administrator did not need to read the minutes of these matters to be appraised of the Committee’s scientific conclusions about dioxin’s health effects.

Moreover, a large portion of the more substantive discussions that were omitted were unfavorable to plaintiffs. For example, the minutes omitted a discussion by Barclay Shephard, Director of the VA Agent Orange Projects Office, who discounted the probative value of animal studies and noted a “general consensus” that Vietnam Veterans exposed to Agent Orange do not face a higher risk of fathering children with birth defects. Plaintiffs’ Statement of Undisputed Facts at II 97(t). Finally, while the minutes do fail to mention Committee review of studies that are favorable to plaintiffs’ position, such as two Swedish studies that show a link between dioxin and soft tissue sarcoma, these two studies did not alter the Committee’s overall conclusion that the evidence did not support a cause and effect relationship between dioxin and soft tissue sarcoma. The Administrator was informed of the Committee’s overall conclusion; the failure of the minutes to specifically mention the Committee’s evaluation of the two Swedish studies is harmless error.

IX. Conclusion and Remedy.

We hold that the Administrator misinterpreted two important provisions of the Act. The Administrator both imposed an imper-missibly demanding test for granting service connection for various diseases and refused to give veterans the benefit of the doubt in meeting that demanding standard. These errors compounded one another, as they increased both the type and the level of proof needed for veterans to prevail during the rulemaking proceedings. We find that these errors, especially compounded with one another, sharply tipped the scales against veteran claimants. As the Act was passed amidst “substantial uncertainty” over the health effects of Agent Orange, we do not find that these errors were harmless; there is a substantial possibility that the errors shaped the conclusions reached by the Advisory Committee and the Administrator. Accordingly, we hereby invalidate the portion of the Dioxin regulation which denies service connection for all other diseases but chloracne. 38 C.F.R. section 311a(d). We also void all benefit denials made under section 311(d), and remand this matter to the Advisory Committee and the VA for further proceedings not inconsistent with this opinion.

IT IS SO ORDERED.

1 . There is no dispute that dioxin is "one of the most highly toxic substances known to the scientific community." House Report at 4451. What is disputed is "how much exposure .to dioxin was experienced by Vietnam veterans, how much exposure can be expected to produce long-term health effects, and at what rate, or frequency, if any, are these effects being experienced” by Vietnam veterans. Id.

2 . Under the VA regulations, “service-connected” means that a disability “was incurred or aggravated .. .in the line of duty in the active military, naval, or air service.” 38 U.S.C. § 101(16). When a disease is designated as presumptively service connected, the individual veteran does not need to prove that the disease was incurred during service. Instead, the burden shifts to the VA to prove lack of service connection, which the VA can demonstrate, inter alia, by showing that some other intervening event caused the disability. 38 C.F.R. § 3.03.

3 . In addition to prescribing rulemaking for dioxin-related claims, the Act also mandates rule-making to govern claims brought by survivors of atmospheric nuclear testing. Section 5(a)(1)(B). Because the nuclear testing rule is not challenged in this proceeding, the Court will make no further reference to this important aspect of the Act.

4 . The Advisory Committee, in turn, was divided into two eight member panels; one panel focused on Agent Orange, the other panel focused on nuclear testing. To avoid confusion, the court will refer only to the Advisory Committee, and not to the subpanel, even though some members of the Advisory Committee did not serve on the subpanel which worked on the Agent Orange issue.

5 . Defendants' efforts to distinguish Evergreen are unavailing. Defendants contend that Evergreen did not involve "challenges to VA regulations brought by individual Veterans attempting to collaterally attack the denial of benefit claims.” Defendant's Memorandum of Points and Authorities at 28-29. This argument is wrong in its premise and conclusion. Some of the plaintiffs in Evergreen were individual veterans challenging benefit denials. 621 F.2d at 1007. Even if no individual veterans had brought claims in Evergreen, section 211(a) does not hinge its preclusive effects on what types of plaintiffs challenge VA action.

6 . We also note as a practical matter that little would be accomplished by dismissing this suit on subject matter jurisdiction grounds. Under the newly enacted statute, plaintiffs could merely wait until September 1989 and file this identical case in the Federal Circuit. The policies of section 211(a) are not served by delay and venue transfer.

7 . Indeed, defendants claim that we are powerless to review the choice at all because the Committee’s selection is "committed to agency discretion by law” under APA section 701(a)(2). We need not decide whether this narrow exception to judicial review applies here, because we agree with defendants that the selection is entitled to great deference and accords with a minimum standard of rationality. Ethly Corp. v. Environmental Protection Agcy., 541 F.2d 1, 36 (D.C.Cir.1976) (en banc).

8 . Dr. Lathrop, a member of the Committee who was involved in the Ranch Hand study, estimated that these veterans received 1000 times the exposure of the typical veteran. Transcript of June 24, 1985 meeting at 8.

9 . Plaintiffs also suggest that because the Committee relied to some extent on prior individual expertise, the record is not sufficiently complete for the court to review the Committee’s decisions. Once again, plaintiffs have overstated their case: the court has before it the transcripts and minutes of the Committee’s meetings, and deposition testimony and declarations that explain the committee’s selection process. We find these materials are adequate to review the Committee’s work.

10 . Plaintiffs contend that these declarations were submitted merely to inform the court about the relevant matters that the Committee failed to consider, and are therefore admissible under Asarco Inc. v. U.S.E.P.A., 616 F.2d 1153, 1160 (9th Cir.1980). Defendants, on the other hand, argue that these declarations are "expert testimony not in the record which seeks to challenge the wisdom of agency action.” Nehmer v. United States Veterans Administration (January 4 Order at 12-13); Asarco, 616 F.2d at 1160-1161. Actually, the declarations serve both purposes; they alert the court to the fact that some experts consider review of animal studies to be essential to performing a thorough examination of the literature. They also challenge the wisdom of the Committee's decision to exclude the studies. Under Asarco, we review these declarations for the former, but not the latter, purpose.

11 . Plaintiffs argue that the VA is selective and unprincipled in its use of animal studies, alleging that the VA relies on animal studies when they refute the claims of veterans, and ignores animal studies when they support veterans' claims. The Court agrees that the VA seems to stand on both sides of the fence on this issue, but nevertheless does not believe that plaintiffs have proven that the VA’s exclusion of animal studies in this context was irrational.

12 . However, defendants take two somewhat contradictory positions in their briefs. They first seem to argue that causal relationship and statistical association tire synonymous, citing VA testimony that “statistically significant correlations” between Agent Orange exposure and a disease would serve to prove a “causal relationship." Defendant’s memorandum at 45, citing Senate Hearing 98-511 at 500-501. However, defendants do not stick with this position, and their experts state in declarations that there are important differences between the two tests. The court finds that the two terms are not synonymous, and therefore rejects defendants’ first argument.

13 . Indeed, committee members appeared to recognize that the extent and type of proof required for service connection status was a legal, not a scientific issue. See Statement of John Murphy, June 25, 1985 meeting at 7, statement of Dr. Colton, June 25, 1985 meeting at 4-5.

14 . Defendants argue that when statutory language does not clearly resolve the question, the court must automatically defer to the agency’s interpretation under Chevron, supra. We do not agree that a court’s inquiry ends upon review of statutory language. We must also examine legislative history and underlying legisla tive purpose to ascertain congressional intent. Portland Cement Association v. Ruckelshaus, 486 F.2d 375, 379-380 (D.C.Cir.1973), cert. denied 417 U.S. 921, 94 S.Ct. 2628, 41 L.Ed.2d 226 (1974): "[i]n ascertaining legislative intent we begin with the language of a statute, but this is subject to an overriding requirement of looking to all sources including purpose and legislative history to ascertain discernible legislative purpose." It would be improper for us to defer to the agency’s interpretation without first determining whether these other materials of statutory construction clear up the ambiguity of statutory language.

15 . Defendants cite one contrary statement. Senator Grassley stated that it would be "imprudent” to grant service connection without a showing of a causal link. 130 Cong.Rec. S 6170 (May 22, 1984). However, Senator Grassley was discussing S 786, a different bill than the one actually passed by the Senate. Moreover, even the language of S.786 does not directly require a causal connection; the bill would grant service connection for diseases that "medical research has shown may be due” to dioxin exposure.

16 . The burdens imposed for other service-related diseases is relevant to the question presented here, because Congress intended the VA to promulgate a dioxin regulation “consistent with present Veterans’ Administration standards and practices.” Statement of Senator Mitchell, 130 Cong.Rec. S 6167. Both sides agree that prior VA practices elucidate the practices required for dioxin adjudication both sides submit evidence of those prior practices.

17 . Because we have decided in favor of plaintiffs on their statutory challenge to use of the cause and effect test, we need not resolve the claim that use of that test also violated plaintiffs’ equal protection rights.

18 . Representative Montgomery likens this policy to a well-known baseball rule: "this [policy] has sometimes been described in baseball terms as the tie goes to the runner." 130 Cong.Rec. H 11161 (October 3, 1984).

19 . We once again reject defendant’s contention that the court must uphold an agency’s reasonable interpretation of a statute whenever the statutory language is ambiguous and leaves room for policy-making by an agency. Deference to an agency is appropriate, but the court must still make its own judgment about congressional intent by using other tools of statutory interpretation. Portland Cement, supra, 486 F.2d at 379-380.

20 . We have considered defendants’ argument that this violation of the Act was harmless error; we address the argument infra at 1422-1423.