The Minnesota Association of Nurse Anesthetists brought this qui tam suit as relator for the United States, alleging that the defendant hospitals2 and anesthesiologists3 had knowingly made false claims on the United States government by mischaracterizing services they had provided to Medicare patients. The Association pleaded that the defendants violated the False Claims Act, 31 U.S.C. § 3729 (1994), by overcharging the government for their services, and that they had conspired among each other to do so. The district court entered summary judgment for the defendants, holding that the Association's own earlier, public disclosure of the information on which this suit is based precluded subject-matter jurisdiction of this suit. The court also held that the Association lacked standing to bring this suit as relator because it had not shown that the mischaracterizations of the services resulted in pecuniary injury to the government. In addition to the jurisdictional and standing rulings, the court also made three holdings on the merits of the Association's case: that there was no showing of intent to defraud the government because the defendants billed in accordance with the advice given them by the Medicare carriers; that the "overwhelming majority of the evidence on the record" established that the defendants did not mischaracterize the services they provided; and that the Association adduced no evidence of conspiracy. We reverse except as to the judgment on the conspiracy allegations, which we affirm.

This case alleges false claims for services rendered under Part B of the Medicare program. The Medicare program is administered by the Department of Health and Human Services through the Health Care Financing Administration, or HCFA. Medicare Part B is a federally subsidized medical insurance program that pays a portion of the insured's medical expenses. The United States reimburses the medical expenses through the HCFA, which, in turn, contracts with private insurance companies to administer and pay claims from the Medicare Trust Fund. United States v. Mackby, 261 F.3d 821, 824 (9th Cir. 2001).

The Association represents the certified nurse anesthetists of Minnesota. Nurse anesthetists are registered nurses who administer anesthesia, either alone or under the supervision of an anesthesiologist. The Association claims that the defendant anesthesiologists and hospitals presented false claims for payment by mischaracterizing anesthesia services rendered to Medicare patients from about 1989 to 1997. Four kinds of mischaracterizations are alleged: billing on a reasonable charge basis when the services the anesthesiologists provided did not meet the criteria for reasonable charge reimbursement; billing services as personally performed by the anesthesiologist when the services did not meet the criteria for personal performance; billing as if the anesthesiologist involved were directing fewer concurrent cases than he or she actually did direct; and certifying that it was medically necessary for both an anesthesiologist and anesthetist to personally perform cases that in fact an anesthetist alone personally performed. Understanding the significance of these alleged mischaracterizations requires some understanding of the Medicare regulations as they existed at the various times in question. We will therefore briefly explain the nature of each allegation before considering the questions of jurisdiction and standing and the merits of the case.

The first type of mischaracterization alleged is that anesthesiologists billed services for reasonable charge reimbursement when they did not render services eligible for such reimbursement. In the early 1980s Congress became concerned that hospital-based physicians were charging Medicare for work performed by hospital employees. S. Rep. No. 97-494, at 22 (1982), reprinted in 1982 U.S.C.C.A.N. 781, 797-98. To stop this, Congress directed the Department of Health and Human Services to adopt regulations governing Medicare payments to physicians working in hospitals. Tax Equity and Fiscal Responsibility Act of 1982, Pub.L. No. 97-248, Title I, § 108, 96 Stat. 324, 337 (codified as amended at 42 U.S.C. § 1395xx(a)(1) (1994)). The regulations were to establish criteria for distinguishing between services rendered by a physician to an individual patient, which could be reimbursed on a reasonable charge basis, and services rendered to the provider or to the provider's patient population as a whole, which would be reimbursed on a reasonable cost basis. Id. Accordingly, the Department adopted regulations in 1983 outlining when physicians providing anesthesia services would be reimbursed on a reasonable charge basis. Conditions for payment of charges: Anesthesiology services, 48 Fed. Reg. 8902, 8926-28 (March 2, 1983). A physician could be reimbursed for anesthesiology services in a hospital on a reasonable charge basis if:

(1) For each patient, the physician

(i) Perform[ed] a pre-anesthetic examination and evaluation;

(ii) Prescrib[ed] the anesthesia plan;

(iii) Personally participat[ed] in the most demanding procedures in the anesthesia plan, including induction and emergence;

(iv) Ensure[d] that any procedures in the anesthesia plan that he or she d[id] not perform [were performed] by a qualified individual;

(v) Monitor[ed] the course of anesthesia administration at frequent intervals;

(vi) Remain[ed] physically present and available for immediate diagnosis and treatment of emergencies; and

(vii) Provide[d] indicated postanesthesia care.

(2) The physician either perform[ed] the procedure directly, without the assistance of an anesthetist, or direct[ed] no more than four anesthesia procedures concurrently and [did] not perform any other services while he or she [was] directing those concurrent procedures. 42 C. F. R. § 405.552(a) (1983).

If the physician's services did not meet the criteria outlined above, then they were reimbursable only on a reasonable cost basis, as physician services to the provider. 42 C.F.R. § 405.552(b) (1983); 48 Fed. Reg. at 8927.

In its complaint in this case, the Association claimed that the anesthesiologists billed on a reasonable charge basis when they had not met the criteria for reasonable charge reimbursement and that the hospitals actively aided the anesthesiologists in the false billing. Specifically, the Association alleged that anesthesiologists at Unity, Mercy, and North Memorial hospitals commonly billed for medical direction of cases in which they never entered the operating room and of cases in which they were not present at the patient's emergence from anesthesia. At Abbott Northwestern and St. Cloud hospitals, the Association alleged that, while the anesthesiologists were usually present at emergence, they sometimes billed on a reasonable charge basis even though they were unavailable for emergencies (as shown by their failure to answer pages) and were absent at emergence.

Second, the Association alleged that the anesthesiologists billed for personally performing cases when they did not meet the criteria for personal performance of the case.

Effective in 1992, HCFA adopted a three-tier system of payment for anesthesiologists providing anesthesia. 56 Fed. Reg. 59502, 59628 (Nov. 25, 1991) (codified at 42 C.F.R. § 414.46(c) (1992)). Under that system, Medicare would pay the highest rate when the anesthesiologist either personally performed the entire anesthesia case or was "continuously involved" in only one case in which an anesthetist was also involved. When an anesthesiologist billed a case as personally performed, the anesthetist involved would not be entitled to any Medicare reimbursement, unless there were special conditions requiring attendance of both an anesthesiologist and an anesthetist at once. Sec. 414.46(c)(2). On the second tier, for medically directing two to four concurrent cases, the anesthesiologist would receive a lower rate, which in turn was diminished (during part of the relevant time) by a set percentage for each additional concurrent case. 42 C.F.R. § 414.46 (d). At the third tier, the lowest rate was paid for supervision of more than four concurrent cases. 42 C.F.R. § 414.46(e).

The Association alleged that the anesthesiologists billed with the highest-rate "AA" modifier, designating that they had performed cases personally, when they were not continuously present during the case and in fact were simultaneously engaged in other activities, including medical direction of concurrent cases.

The anesthesiologists contend that when an anesthesiologist and anesthetist were both involved in one case, with no concurrent procedures, an anesthesiologist was entitled to designate a case as personally performed so long as he or she met the criteria for medical direction of the anesthetist. Specifically, the anesthesiologists contend that personal performance of a single case involving an anesthetist did not require the anesthesiologist's continuous presence in the operating room, as long as the anesthesiologist was present in the operating suite. Thus, according to the anesthesiologists, when they had only one case at a time, they were entitled to be paid at the higher, "personal performance" rate even though they were performing the same duties otherwise compensated at the lower, "medical direction" rate. The corollary of this theory is that anesthetists would be paid for their work on a case if the anesthesiologist happened to be performing a concurrent case, but would not be entitled to Medicare reimbursement for performing the very same duties in a case in which the anesthesiologist had no concurrent case.

The Association, on the other hand, contends that in order to be paid the higher rate for personally performing a case, an anesthesiologist had to be more closely involved than was necessary to satisfy the medical direction criteria of section 405.552, and in fact, had to be continuously present with the patient.

The Association also alleged that the anesthesiologists misrepresented the number of cases they were performing concurrently. As we said in section B, the 1992 regulations reduced an anesthesiologist's medical direction fee per case as the number of cases he or she directed increased (up to the maximum of four). 42 C.F.R. § 414.46(d) (1992). This was apparently a short-lived policy, which was changed as of January 1, 1994 to a flat rate per case, no matter whether the anesthesiologist was directing two, three or four concurrent procedures. Omnibus Reconciliation Act of 1993, Pub. L. No. 103-66, § 13516, 107 Stat. 312, 583-84 (1993). The Association alleged that while the 1992 regulations were in force, the anesthesiologists regularly understated the number of concurrent cases they were directing simultaneously in order to get a bigger payment from Medicare than they were entitled to.

Finally, the Association alleges that the hospitals represented in many cases that it was necessary for both an anesthesiologist and anesthetist to personally perform the case, whereas only the anesthetist actually satisfied the criteria for personal performance.

Congress decided in 1986 that anesthetists' services should be reimbursable under Part B of Medicare on a reasonable charge basis, starting on January 1, 1989. Omnibus Reconciliation Act of 1986, Pub. L. No. 99-509, § 9320, 100 Stat. 1874, 2013-16 (1986). Before this, Medicare did not pay for anesthetists' anesthesia services on a reasonable charge basis. See 48 Fed. Reg. at 8927. Even after the law was changed to allow direct reimbursement for anesthetists' services, when an anesthesiologist was involved with only one case at a time and an anesthetist worked on that case as well, Medicare would not ordinarily pay for the anesthetist's services because it was considered inefficient to have both an anesthesiologist and an anesthetist wholly engaged in one case. 57 Fed. Reg. 33878, 33887 (July 31, 1992) ("[W]e are concerned that recognizing medical direction in a single anesthesia procedure would encourage inefficiencies in anesthesia practice arrangements. Our policies should not encourage the involvement of both practitioners in a single anesthesia procedure if either practitioner could appropriately furnish the service alone."). However, if there was some unusual medical necessity requiring the attendance of both anesthesiologist and anesthetist on the same case, Medicare would then pay for both at 100% of their usual personal performance rate. 42 C.F.R. § 414.46(c)(3) (1992); see HCFA Transmittal No. B-98-2 (1998). Later, the regulation was amended to allow both anesthesiologist and anesthetist to bill for these "one-on-one" cases even without special medical necessity, but the total payout was to be limited to the amount that would have been paid to the anesthesiologist for doing the case alone. 60 Fed. Reg. 38400, 38416 (July 26, 1995) (proposing payment scheme to begin in 1998).

The Association alleges that in order to collect payment for both an anesthesiologist and an anesthetist performing a single case, the hospitals sometimes certified that there was a medical necessity for both to attend, whereas in fact the anesthesiologist did not assist throughout the case, demonstrating that no necessity existed.

On December 28, 1994, the Association brought this suit on behalf of the United States alleging violation of the False Claims Act, 31 U.S.C. §§ 3729(a) (1) and (2) and (7), conspiracy to violate the Act, and violation of the hospitals' Medicare provider agreements in connection with their anesthesia billing practices. The United States declined to intervene.

At the threshold, we must decide whether we have subject-matter jurisdiction over this case. On November 8, 1994, some seven weeks before filing this case, the Association and several individual anesthetists sued many of the same defendants alleging various federal antitrust and state law violations, again in connection with their anesthesia billing practices. The antitrust complaint alleged:

[T]he defendant anesthesiology groups and their co-conspirators have engaged in a wide-spread practice of fraudulent billing of anesthesia services in violation of . . . Federal statutes, including § 1128(a)(1)(A). Such violations include, but are not limited to, billing for services that they did not render, billing for operations at which they were not present and inaccurately designating operations as one-on-one for Medicare purposes.

The allegations in the Association's antitrust case were immediately reported in the local newspapers in St. Paul and St. Cloud on November 10 and 11. The Association also provided a copy of the antitrust case to the United States government. Only after this publicity did the Association file this False Claims Act case, under seal, as provided by statute.

The district court held that the disclosure of allegations of fraud in the Association's antitrust suit and in newspaper articles about the antitrust suit precluded subject-matter jurisdiction over this qui tam case because of a special jurisdictional limitation in the False Claims Act disallowing suits based on publicly disclosed information, 31 U.S.C. § 3730(e)(4)(1994). In addition to the disclosures the district court cited, the defendants contend that instances of one type of claim alleged by the Association, improper billing of one-on-one cases, were disclosed in the course of an administrative audit before the Association had assembled its allegations, likewise depriving the district court of subject-matter jurisdiction over that claim.

The level of our review of a district court's ruling on subject-matter jurisdiction depends on whether the district court based its determination on the complaint, on undisputed facts outside the complaint, or on findings of fact. See Osborn v. United States, 918 F.2d 724, 729-30 (8th Cir. 1990). In this case, the district court limited its jurisdictional inquiry to the complaint and undisputed facts, styling its order as a summary judgment. See generally id. at 729-30 (discussing distinction between subject-matter jurisdiction determination under Fed.R.Civ.P. 12(b)(1) and summary judgment). Therefore, we exercise de novo review, limited to "determining whether the district court's application of the law is correct and, if the decision is based on undisputed facts, whether those facts are indeed undisputed." Id. at 730 (quotations omitted).

The limitations on subject-matter jurisdiction over False Claims qui tam cases were enacted as part of the False Claims Amendments Act of 1986, Pub. L. No. 99-562, 100 Stat. 3153, 3157 (1986). The 1986 amendments were an avowed attempt to reinvigorate the False Claims Act after a 1943 amendment and judicial decisions interpreting the 1943 amendment had emasculated the 1863 law. Understanding the Congressional intent expressed in the 1986 amendments therefore requires a review of False Claims Act history.

The original False Claims Act was enacted in 1863 in order to strike back against the fraud of unscrupulous Civil War defense contractors. S. Rep. No. 99-345, at 8 (1986), reprinted in 1986 U.S.C.C.A.N. 5266, 5273. The Act contained a qui tam provision allowing private persons to sue as relators representing the government's interests, and it rewarded relators who prevailed in their suits with a bounty of half the damages and forfeitures they recovered for the government.4 Id. at 10. The large size of the relator's share, which came out of the government's ultimate recovery, had an obvious potential to put the relator and the government at odds with each other.

During World War II there were many qui tam cases against defense contractors, and in one notorious case, United States ex rel. Marcus v. Hess, 317 U.S. 537 (1943), the government contended that the relator had simply copied allegations from a criminal indictment already on file. The Supreme Court held that, even if the relator, Marcus, had simply taken allegations from a criminal indictment, the False Claims Act would nevertheless permit him to proceed as relator. Id. at 545. In reaction to Hess, Attorney General Francis Biddle asked Congress to repeal the qui tam provisions of the False Claims Act. S. Rep. No. 99-345, at 11. Congress refused to go so far, but it did amend the Act to provide that there would be no jurisdiction over qui tam suits "whenever it shall be made to appear that such suit was based upon evidence or information in the possession of the United States, or any agency, officer or employee thereof, at the time such suit was brought."5 31 U.S.C. § 232(C) (1946); S. Rep. No. 99-345, at 12. The provision was explained as an attempt to curtail parasitical suits in which the informer "rendered no service" to the government. 89 Cong. Rec. 10846 (1943). In United States ex rel. Wisconsin v. Dean, 729 F.2d 1100 (7th Cir. 1984), the State of Wisconsin brought a qui tam suit based on Medicaid fraud that it had already disclosed to the federal government. The Seventh Circuit, interpreting the 1943 amendment, held that even though the discovery of the fraud was entirely due to the State's investigation, qui tam jurisdiction was barred because the federal government knew of the fraud before Wisconsin filed suit. Id. at 1104-07.

Within months of the Dean decision, the National Association of Attorneys General adopted a resolution urging Congress "to rectify the unfortunate result of the Wisconsin v. Dean decision." S. Rep. No. 99-345, at 13. Congress responded. Senate Bill 1562, which became the 1986 False Claims Amendments Act, was introduced shortly after and was "aimed at correcting restrictive [court] interpretations" of the False Claims Act which "tend to thwart the effectiveness of the statute."6 Id. at 4, 13. The goals of the 1986 Amendments Act were (1) to encourage those with information about fraud against the government to bring it into the public domain; (2) to discourage parasitic qui tam actions by persons simply taking advantage of information already in the public domain; and (3) to assist and prod the government into taking action on information that it was being defrauded. United States ex rel. Mistick PBT v. Hous. Auth., 186 F.3d 376, 401 (3d Cir. 1999) (Becker, C.J., dissenting) (citing S. Rep. No. 99-345, at 1-8, 23-24 ), cert. denied, 529 U.S. 1018 (2000).

The 1943 amendments could bar qui tam suits on the ground of information technically in the government's possession, even if no one in the government knew about the information. 132 Cong. Rec. 22340 (1986) (remarks of Rep. Bedell). The supporters of the 1986 Amendments Act believed that this prevented relators from bringing suits in situations in which their participation was needed and in which the fraud would not be prosecuted without their intervention. In an apparent attempt to correct this shortcoming of the 1943 version of the Act, Congress switched from barring suit on the ground of government possession of information before the relator filed suit to barring suit on the ground of public disclosure of such information. See id. (referring to House bill). Congress evidently assumed that if information was publicly disclosed, the government was likely to discover it on its own, without the need for a qui tam relator. See United States ex rel. Stinson, Lyons, Gerlin & Bustamonte, P.A. v. Prudential Ins. Co., 944 F.2d 1149, 1169 (3d Cir. 1991) (Scirica, J., dissenting).

The 1986 Act also added an important exception to the jurisdictional bar for relators who are "an original source of such information." The relevant section is 31 U.S.C. § 3730(e)(4):

(A) No court shall have jurisdiction over an action under this section based upon the public disclosure of allegations or transactions in a criminal, civil, or administrative hearing, in a congressional, administrative, or Government Accounting Office report, hearing, audit, or investigation, or from the news media, unless the action is brought by the Attorney General or the person bringing the action is an original source of the information.

(B) For purposes of this paragraph, "original source" means an individual who has direct and independent knowledge of the information on which the allegations are based and has voluntarily provided the information to the Government before filing an action under this section which is based on the information.

Section 3730(e)(4) is crucial to resolving the jurisdictional issue in the present case. Applying the section requires us to answer three questions: (1) Have allegations made by the relator been "publicly disclosed" before the qui tam suit was brought? (2) If so, is the qui tam suit "based upon" the public disclosure? and (3) If so, was the relator an "original source" of the information on which the allegations were based? See United States v. Bank of Farmington, 166 F.3d 853, 859 (7th Cir. 1999)(applying tripartite test). Jurisdiction exists only if the answer to one of the first two questions is "no" or the answer to the third question is "yes." The original source inquiry, in turn, has three parts; the relator's knowledge of the information must be (1) direct and (2) independent, and (3) the relator must have voluntarily provided the information to the Government before filing suit.

The Association contends that the goals of the 1986 Amendments, including the intent to revise the law after the unsatisfactory result in Dean, would be thwarted by interpreting the 1986 Amendments Act to bar suits by relators who caused the public disclosure of the fraud. However, the various components of the statute have been interpreted by the courts in such a way that, when the pieces are put together, the result is sometimes to bar such actions. See, e.g., United States ex rel. Dhawan v. New York Med. Coll., 252 F.3d 118, 121-22 (2d Cir. 2001) (relator who disclosed information in state court suit was thereby barred); United States ex rel. Hafter v. Spectrum Emergency Care, Inc., 190 F.3d 1156, 1163 (10th Cir. 1999) (relator whose information was "impetus" for investigation was barred by disclosure of results of investigation); United States ex rel. Mistick PBT, 186 F.3d at 389 (relator whose FOIA request brought fraud to light was barred because the government's response to the request was public disclosure); United States ex rel. Jones v. Horizon Healthcare Corp., 160 F.3d 326, 335 (6th Cir. 1998) (relator's state lawsuit disclosed claims and therefore barred federal qui tam suit); United States ex rel. Devlin v. California, 84 F.3d 358, 360, 363 (9th Cir. 1996) (relator who fed allegations to newspaper barred because resulting article disclosed fraud); United States ex rel. Kreindler & Kreindler v. United Tech. Corp., 985 F.2d 1148, 1159 (2d Cir. 1993) (relator who caused information to be revealed in the course of discovery in a lawsuit barred thereby). The Association cites a letter by the sponsors of the 1986 Amendments Act reviewing the career of the public disclosure bar in the courts: "Certain courts have exploded this limited bar in ways that mock the very purpose and intent of the 1986 Amendments." Letter from Rep. Howard L. Berman and Sen. Charles E. Grassley to Janet Reno, 145 Cong. Rec. E1540, 1546 (July 14, 1999).7

Our task, of course, is to effectuate Congress's intent, and we must interpret the three subsidiary components of section 3730(e)(4) with a view to how they contribute to the effect of the statute as a whole.

[A] section of a statute should not be read in isolation from the context of the whole Act, and . . . in fulfilling our responsibility in interpreting legislation, we must not be guided by a single sentence or member of a sentence, but [should] look to the provisions of the whole law, and to its object and policy.

State Highway Comm'n v. Volpe, 479 F.2d 1099, 1111-12 (8th Cir. 1973) (citations omitted).

The first question, whether the allegations have been publicly disclosed, must be answered "yes." The Association essentially concedes that the antitrust suit and the resulting newspaper articles amounted to public disclosure, as the district court held.

However, on appeal the defendants further contend that the Association's allegations were also publicly disclosed through a 1991 audit performed for Medicare by The Travelers, which, as an administrative audit, would qualify as one of the types of public disclosure that could trigger the jurisdictional bar. In the audit, Travelers notified Mercy Medical Center that it had been billing for anesthetists' services in cases in which the anesthetist and anesthesiologist were both working on the same case and no other concurrent cases, known as one-on-one cases.

Review of the Association's claims indicates that this is not one of the practices the Association alleges is fraudulent. The Association contends that some of the alleged fraudulent practices arose as a response to the 1991 audit, when the hospitals realized they could no longer bill for both the anesthesiologist and an anesthetist's service in routine one-on-one situations, and so began falsely certifying that such cases required the involvement of both an anesthesiologist and an anesthetist. Therefore, the 1991 audit was not a public disclosure of this allegation.

The Association also contends that the anesthesiologists billed for personally performing cases in which they were not "continuously involved" with the patient or case. In contrast, the audit letter does not say that the anesthesiologists were not performing the cases, but instead concludes that they were performing them, which meant that the hospital could not bill for the anesthetist's involvement in the same case. Therefore, the audit did not publicly disclose this allegation of fraud, either. The defendants contend that the transactions pointed out in the audit are some of the same transactions on which the Association now bases its claims. The audit may have revealed the fact that the defendants represented that the anesthesiologists performed these cases, but the audit did not state that the anesthesiologists had not in fact done so. In order to bar jurisdiction, a public disclosure must reveal both the true state of facts and that the defendant represented the facts to be something other than what they were. United States ex rel. Rabushka v. Crane Co., 40 F.3d 1509, 1514 (8th Cir. 1994). The audit did not reveal what the Association now contends was the true state of the facts, i.e., that the anesthesiologists were not performing the cases they billed for. Therefore, the audit was not a public disclosure that could bar jurisdiction over the Association's claims.

The second question under section 3730(e)(4) is whether the allegations in the qui tam case were "based upon" the public disclosure. This requires us to address the meaning of those words, a question which has split the federal circuits, but which our court has not yet explicitly addressed.

The minority view, shared only by the Fourth Circuit and one panel of the Seventh Circuit (in schism with another panel),8 is that "based upon" should be given its ordinary meaning of "derived from," so that the qui tam allegation must have resulted from the disclosure in order to bar jurisdiction. United States ex rel. Siller v. Becton Dickinson & Co., 21 F.3d 1339, 1348 (4th Cir. 1994); United States v. Bank of Farmington, 166 F.3d 853, 863 (7th Cir. 1999). The minority bases its reading on a "straightforward textual exegesis" of the phrase "based upon," Siller, 21 F.3d at 1348, as well as the policy consideration that a suit not derived from the public disclosure is not "parasitic," and so is not the kind of suit the 1986 Amendments Act was meant to prevent, Bank of Farmington, 166 F.3d at 863.

The majority view is that a qui tam suit is "based upon" a public disclosure whenever the allegations in the suit and in the disclosure are the same, "regardless of where the relator obtained his information." United States ex rel. Doe v. John Doe Corp., 960 F.2d 318, 324 (2d Cir. 1992). Accord United States ex rel. Findley v. FPC-Boron Employees' Club, 105 F.3d 675, 682-85 (D.C. Cir. 1997); United States ex rel. Mistick PBT v. Housing Auth., 186 F.3d 376, 385-88 (3d Cir. 1999) (qui tam suit based upon disclosure if the disclosure "sets out" allegations or all essential elements of qui tam claim), cert. denied, 529 U.S. 1018 (2000); United States ex rel. McKenzie v. BellSouth Telecom., Inc., 123 F.3d 935, 940 (6th Cir. 1997) ("based upon" public disclosure means "supported by" disclosure); United States ex rel. Lamers v. City of Green Bay, 168 F.3d 1013, 1017 (7th Cir. 1999) (relevant facts disclosed in media after relator filed administrative complaint and before relator filed qui tam suit; therefore qui tam jurisdiction barred unless relator an original source); United States ex rel. Biddle v. Board of Trustees of the Leland Stanford, Jr., Univ., 161 F.3d 533, 536-40 (9th Cir. 1998); United States ex rel. Precision Co. v. Koch Indus., Inc., 971 F.2d 548, 552-53 (10th Cir. 1992) ("As a matter of common usage, the phrase 'based upon' is properly understood to mean 'supported by.'"); Cooper v. Blue Cross and Blue Shield, 19 F.3d 562, 567 (11th Cir. 1994) (per curiam) ("based upon" means "supported by").

The majority view has a powerful argument to commend it: if a suit is only based upon a public disclosure if it results from the disclosure, as the minority interpretation would have it, then the statute's additional provision allowing suit if the relator is "an original source" of the underlying information is of no effect, because no one could be an original source if his knowledge was derived from public disclosure.9 More specifically, a relator's knowledge could not be "independent" of the public disclosure, sec. 3730(e)(4)(B), if it was derived from the public disclosure. The majority of courts have considered it inconceivable that Congress would have drafted the statute so poorly as to have included a provision that could never have any effect.

There are two considerable objections to this majority rule. First, its reading distorts the plain meaning of the words "based upon the public disclosure," since if the qui tam allegations are not derived from the public disclosure itself, they are not based upon the public disclosure, but rather on the facts which have been publicly disclosed. Elsewhere in the statute, Congress used the phrase "based on the facts underlying the pending action," 31 U.S.C. § 3730(b)(5), which suggests the drafters distinguished between basing a suit on facts and basing it on a disclosure of such facts. The majority's interpretation also distorts the words "based upon" by taking away the causal relation inherent in the phrase. As the Fourth Circuit remarked, "We are unfamiliar with any usage, let alone a common one or a dictionary definition, that suggests that 'based upon' can mean 'supported by.'" Siller, 21 F.3d at 1349. This objection may well be unanswerable. The Third Circuit, in adopting the majority rule, candidly admitted that the "in ordinary usage the phrase 'based upon' is not generally used to mean 'supported by,'" Mistick PBT, 186 F.3d at 386, but concluded that there was elsewhere evidence that the statute was not carefully drafted. Id. at 387. Evidently, the words "based upon" were simply not well chosen to express Congress's meaning.

The second objection to the majority view is that the policy justification sometimes given by courts in the majority, if taken to its logical conclusion, would return us to the rule of the Dean case, which Congress was specifically attempting to overrule by means of the 1986 Amendments Act. For instance, the District of Columbia Circuit justified its adoption of the majority rule as follows:

[T]he blocking of freeloading relators who copy their complaints directly from public disclosures is not the FCA's only concern. From its inception, the qui tam provisions of the FCA were designed to inspire whistleblowers to come forward promptly with information concerning fraud so that the government can stop it and recover ill-gotten gains. Once the information is in the public domain, there is less need for a financial incentive to spur individuals into exposing frauds.

FPC-Boron Employees' Club, 105 F.3d at 685. By this reasoning, once the information is available to the government, the government has no need to pay a relator for disclosing it, even if the relator discovered the fraud in the first instance. Against this, the minority view reasons that by attempting to overrule the Dean case legislatively and especially by enacting the "original source" exception to the jurisdictional bar, Congress obviously rejected this one-shot view of its financial interests in favor of a fairness policy. Rather than biting the hand that fed it, Congress apparently chose to take a longer view, reasoning that its interests over time would be served by rewarding informants rather than confiscating their claims whenever it could do so.10

The minority objects that the majority's reading of "based upon" throws up a jurisdictional bar in some suits that are not parasitical, in the sense of being cribbed from the public disclosure, whereas the avowed goal underlying the 1986 Amendments Act was to eliminate parasitical suits. See FPC-Boron Employees' Club, 105 F.3d at 685 (acknowledging that "our interpretation of the jurisdictional bar may on occasion prevent qui tam lawsuits that may not be truly 'parasitic'"). The minority view contends that by trying to rectify Dean, Congress showed a desire to treat relators fairly, which would be frustrated by kicking relators out of court when their claim was not parasitical, but was merely disclosed before the relator had filed suit.

In our view, however, these policy objections disappear if one considers the overall design of the public disclosure provision. Congress's fairness concern is not effectuated by each part of the statute read in isolation, but rather by the statute as a whole. The "based upon" clause serves the concern of utility, that is of paying only for useful information, and the "original source" exception serves the concern of fairness, that is of not biting the hand that fed the government the information. If the "based upon" clause threatens to kick relators out of court because the government does not need them, the "original source" exception reopens the courthouse door for certain deserving relators. Therefore, the majority view reaches the correct result, not because Congress cared nothing for fairness and everything for utility, but because it used two different provisions to strike a balance between these concerns.

Thus, the majority reading of section 3730(e)(4) is consistent with Congress's apparent policy. We also conclude that the majority view, though not free of strain, gives a more coherent meaning to the confusing language of the section than the minority view does. A final factor supporting the majority's reading is that it fits with the drift of our circuit precedent.

Our court has not expressly considered the meaning of the "based upon" clause, although Judge Magill has announced his support of the majority rule in a dissent. United States ex rel. Rabushka v. Crane Co., 40 F.3d 1509, 1527-28 (8th Cir. 1994). However, in United States ex rel. Barth v. Ridgedale Electric, Inc., 44 F.3d 699, 702 (8th Cir. 1995), we held that a suit was barred because of a public disclosure in newspaper articles that were not published until after one would-be relator had completed its investigation and reported the allegations to the County Attorney. From the chronology of these events, it is obvious that the allegations of the qui tam complaint in Barth were not derived from the newspaper articles. Nevertheless, Barth parsed the original source provisions, which would have been utterly unnecessary if the suit had not been "based upon" public disclosures, and concluded that the relators could not prosecute the suits because they were not original sources of the information. Id. at 704. Because the result in Barth would have been different if "based upon" meant "derived from," that case suggests that the issue has been resolved in this circuit, albeit implicitly, consistently with the majority rule.

Having concluded that the majority rule makes better sense of the 1986 Amendments Act and better effectuates the policy goals of that Act, we now explicitly endorse the majority view and hold that the allegations in this case were "based upon" the antitrust case and accompanying newspaper accounts.

Finally, we come to the question of whether the Association was an "original source" of the information disclosed. If not, then there is no jurisdiction over this case. Since we know from the history of the False Claims Act that the original source provision was added in 1986 to permit claims like the one in Dean, in which a claimant investigated the fraud and then revealed it to the government before filing suit, we would expect that the effect of the original source provision is to protect from the public disclosure bar those who first bring a claim to light. However, "original source" is defined in section 3730(e)(4)(B) in a way that does not distinguish between those who first bring a claim to light and others who later make the same discovery independently,11 and it does not always protect those responsible for the initial disclosure of a fraud claim, e.g., Barth, 44 F.3d at 702-04.

Under section 3730(e)(4)(B), a claimant is deemed an original source if he or she (1) has "direct and independent knowledge of the information on which the allegations are based" and (2) has voluntarily provided the information to the "Government" before filing the qui tam suit.12

We have determined that the words "direct" and "independent" were intended to express two ideas, rather than one. Barth, 44 F.3d at 703. We have interpreted "independent knowledge" to mean knowledge not derived from the public disclosure. Id. But see United States ex rel. Fine v. Advanced Sciences, Inc., 99 F.3d 1000, 1006-07 (10th Cir. 1996)(independent means independent of anyone else-i.e., the same thing as direct). The independent knowledge requirement clearly serves the congressional goal of barring parasitic actions, but it is worth noting that it does not bar actions based on old news, in which the relator independently discovers information already known to the public. See Fed. Recovery Servs., Inc. v. United States, 72 F.3d 447, 452 (5th Cir. 1995) (this situation addressed in 31 U.S.C. § 3730(d)(1), which reduces size of relator's award). There is no doubt that the Association's knowledge was independent of the Association's antitrust case and the newspaper articles based on that case.

But did the Association also have direct knowledge? This term is more problematic. Courts have used various formulations, sometimes looking at the words without any reference to what Congress hoped to accomplish by using the term, and sometimes focusing on the policy of avoiding parasitism without paying much attention to the actual words Congress employed. The Third Circuit cited a dictionary definition of "direct" as "marked by absence of an intervening agency, instrumentality or influence: immediate." Stinson, Lyons, 944 F.2d at 1160 (quoting Webster's Third International Dictionary 640 (1976)). We reiterated this definition in Barth, 44 F.3d at 703 (quoting United States ex rel. Springfield Term. Ry. Co. v. Quinn, 14 F.3d 645, 656 (D.C. Cir. 1994)). Also in Barth we quoted the Ninth Circuit's definition of direct knowledge as "unmediated by anything but [the plaintiff's] own labor," id. (quoting Wang , 975 F.2d at 1417), which reflects the congressional intent to avoid parasitical suits in which the plaintiff contributed nothing.

The district court held that the Association had no direct knowledge of the information because its knowledge came from its members. The court cited two cases in which corporations were formed after the information had been discovered and disclosed by people who became shareholders of the corporations; in these cases, the corporations were not original sources of the information. Federal Recovery Servs., 72 F.3d at 451-52; Precision Co., 971 F.2d at 554. These cases are easily distinguishable, because they involved a corporate plaintiff that did not exist at the time the information was discovered. No courts have held that corporations responsible for the discovery of information cannot have "direct knowledge" because they have to act through agents. In fact, corporate plaintiffs have been held to have direct knowledge making them an original source. In Springfield Terminal Railway Co., 14 F.3d at 657, the District of Columbia Circuit held that a corporate relator had sufficiently direct knowledge of information to be an original source. Accord United States ex rel Durcholz v. FKW Inc., 997 F. Supp. 1159, 1166 (S.D. Ind. 1998). Moreover, in Barth one relator was a labor union; although we held that the union had no direct knowledge of the information because its representative did not have such knowledge, we did not suggest organizations can never be original sources. 44 F.3d at 703-04. There is no hint in the history of the 1986 Amendments Act that Congress intended to disqualify organizational relators.13 To the contrary, any such rule would have disqualified the State of Wisconsin from proceeding as relator in Dean and so would defeat one of the announced motivations behind the 1986 Amendments Act. Though organizations must, of course, act through agents, this does not render their knowledge parasitical or their agency "intervening" in the sense of interrupting the causal connection between the corporation's efforts and the knowledge. See Black's Law Dictionary 212 (7th ed. 1999) ("intervening cause" or "intervening agency" is "An event that comes between the initial event in a sequence and the end result, thereby altering the natural course of events that might have connected a wrongful act to an injury").

In further contrast to the corporate relators in Precision Co. and Federal Recovery Services, the Association is an unincorporated association. Unlike a corporation, a voluntary unincorporated association has no legal status separate from its members. See St. Paul Typothetae v. St. Paul Bookbinders' Union, 102 N.W. 725 (Minn.1905) ("Such [unincorporated] societies, in the absence of statutes recognizing them, have no legal entity distinct from that of their members."). By statute Minnesota altered the common law to permit persons associated under a common name to sue under that name, Minn. Stat. Ann. § 540.151 (2000), but this statute is only procedural. Unincorporated associations derive their rights from the rights of their members. See Federal Election Comm'n v. Colo. Republican Fed. Campaign Comm., 121 S. Ct. 2351, 2362 n.10 (2001) (First Amendment rights). Thus, associations can have standing to assert their members' rights in court, see United Food & Commercial Workers Union Local 751 v. Brown Group, Inc., 517 U.S. 544, 551-53 (1996), whereas a corporation has no standing to assert rights belonging to its shareholders, Waseca Co. Bank v. McKenna, 21 N.W. 566 (Minn. 1884). An association's knowledge is in no way parasitic of its members and is "direct" within the meaning of the original source clause.

In this case, the Association pleaded that its members have "personal knowledge that defendant anesthesiologists have routinely billed Medicare for personal performance of anesthesia procedures in which they were not continuously involved or present." It further pleaded that its members "have personal knowledge of defendants' false claims by virtue of communications with defendants themselves, participation in the anesthesia procedures which were later fraudulently billed by the defendant anesthesiologist, and familiarity with hospital records disclosing defendants' fraud." The defendants' response is that the anesthetists did not have direct knowledge of the anesthesiologists' billing practices, which came to light in an audit.

There are two problems with the defendants' argument. First, the record shows that the anesthetists often did see the anesthesiologist filling out forms used for billing with misleading information. These observations would support an inference that the anesthesiologist submitted false bills.

Second, to qualify as an original source, a relator does not have to have personal knowledge of all elements of a cause of action. Springfield Term. Ry., 14 F.3d at 656-67. Direct knowledge of the anesthesiologists' operating room practices would be enough. A false claim consists of a representation contrary to fact, made knowingly or recklessly. If the relator has direct knowledge of the true state of the facts, it can be an original source even though its knowledge of the misrepresentation is not first-hand. Id. We therefore conclude that the Association has not only independent, but also direct knowledge of the information in question within the meaning of section 3730(e)(4)(B).

The last statutory condition for qualifying as an original source is that the relator must have voluntarily provided the information to the government before filing suit. The defendants concede that the Association sent a copy of its antitrust complaint to the local Medicare Part B office several days after it filed the antitrust suit. The Association's attorney filed an affidavit saying he received a call from a Medicare representative within a few days after mailing the complaint. The representative said he was referring the complaint to the Justice Department. We conclude that the Association fulfilled the requirement that it provide the information to the government before filing suit.

The defendants urge us to adopt an additional requirement that the relator must have revealed the allegations to the government before the public disclosure in order to be an original source. This rule has been adopted by the District of Columbia and Sixth Circuits. FPC-Boron Employees' Club, 105 F.3d at 690-91; McKenzie, 123 F.3d at 943. This additional requirement has no textual basis in the statute. Moreover, the courts adopting this requirement have justified it by arguing that after public disclosure, the relator has no utility to the government. FPC-Boron Employees' Club, 105 F.3d at 691 ("Once the information has been publicly disclosed, however, there is little need for the incentive provided by a qui tam action."). However, as we have seen, through the original source provisions Congress chose to reward persons who discovered and revealed fraud, rather than confiscating their claims. At the same time, Congress limited that beneficence by denying the bounty even to those who uncovered the fraud unless they had revealed it to the government before filing suit. Sec. 3730(e)(4)(B). We would change the balance Congress struck if we were to further restrict the class of those whose discoveries had been made public but who were nevertheless permitted to proceed as relators. We decline to adopt the proposed additional requirement.

We hold that the Association qualifies as an original source of the information on which its allegations are based. We have subject-matter jurisdiction over this case.

The district court held that the Association lacked standing to pursue the qui tam claim because there was no pecuniary injury to the United States when the anesthesiologists allegedly billed for medically directing or personally performing cases without fulfilling the requirements for medical direction or personal performance. The district court went beyond the pleadings to examine the evidence on the injury issue. We therefore review this issue under the summary judgment standard, rather than limiting our inquiry to the pleadings. See Lujan v. Defenders of Wildlife, 504 U.S. 555, 561 (1992).

We review the district court's entry of summary judgment de novo, applying the same standard appropriate in the district court. Breeding v. Arthur J. Gallagher & Co., 164 F.3d 1151, 1156 (8th Cir. 1999). Summary judgment is proper only if, taking the evidence in the light most favorable to the non-moving party, there is no genuine issue of material fact and the moving party is entitled to judgment as a matter of law. Id.

After the date of the district court's decision, the Supreme Court decided Vermont Agency of Natural Resources v. United States ex rel. Stevens, 529 U.S. 765, 778 (2000), in which it held that "a qui tam relator under the [False Claims Act] has Article III standing." The Court identified two discrete injuries to the United States that are redressed through False Claims cases: "both the injury to its sovereignty arising from violation of its laws (which suffices to support a criminal lawsuit by the Government) and the proprietary injury resulting from the alleged fraud." Id. at 771. The Court held that a qui tam relator gained standing to assert the government's rights through a "partial assignment of the Government's damages claim." Id. at 773. The defendants contend that if there are no damages, a qui tam relator does not have standing to assert the government's claim for penalties. The United States, appearing as amicus, contends that the relator can pursue such a claim. We have no occasion to address this argument, since the district court erred in holding that the United States would have suffered no pecuniary injury even if the Association proved false claims. That holding was based on a misunderstanding of the payment rules.

The Association alleges that the anesthesiologists billed for medical direction in cases in which they fell short of the requirements for that designation. The district court concluded that if the anesthesiologists had not billed these cases as medical direction, "[p]resumably" they could only have billed them as personally performed, which would have been even more expensive. This does not follow. The services of an anesthesiologist who was involved in some way with a case but failed to meet the requirements of personal performance or medical supervision were not reimbursable on a reasonable charge basis under Medicare as services to the patient. Instead, the anesthesiologist's services would be considered supervisory services furnished to the hospital, reimbursable to the hospital on a reasonable cost basis only. 42 C.F.R. § 405.552(b) (1983); 48 Fed. Reg. 8902, 8927 (March 2, 1983). The defendants contend that if the anesthesiologists were not billed as performing or directing the cases, the government would have had to pay the anesthetists for the same service and therefore there would have been no net loss to the government. Before 1989, anesthetists were not eligible for reasonable charge reimbursements for anesthesia services. 48 Fed. Reg. at 8927. After January 1, 1989, Medicare reimbursed for anesthetists' services, but at least for some of that time it paid a lower rate than it paid for anesthesiologists' anesthesia services. United States General Accounting Office Report to Congressional Committees, Medicare Payments for Medically Directed Anesthesia Services Should Be Reduced 24-27 (March 1992). The Association filed an affidavit asserting that anesthetists were paid less than anesthesiologists for performing cases personally during much of the relevant time period. Therefore, it is incorrect to conclude that the government would have paid the same for the services no matter whether they were billed as personally performed by an anesthesiologist or not. Additionally, the Association claims that in some cases the hospitals certified that it was medically necessary for both an anesthesiologist and anesthetist to perform anesthesia on a single patient, with no concurrent cases, when in fact the anesthesiologist did not personally perform the case. In such cases, if the government paid the personal performance rate to both the anesthetist and the anesthesiologist, the government's cost would be doubled. These sorts of pecuniary injury plainly confer standing on the relator who alleges them. The district court's legal conclusion that the Association lacked standing was premised on a faulty understanding of the applicable regulations.

The Association also appeals the district court's entry of summary judgment against it on the merits of its suit.

The district court entered summary judgment against the Association on the merits of its claims to the extent the claims were based on anesthesiologists billing cases as "personally performed." The district court held that Medicare regulations in effect at the time in question were "susceptible" to the interpretation that an anesthesiologist need not have been continuously physically present in the operating room to bill a case as "personally performed." The court concluded that this ambiguity ruled out the possibility that the defendants knew they were presenting false claims when they billed for personal performance of the cases in which they were not continuously present.

The United States has filed an amicus brief taking strong exception to the proposition that one cannot make a false statement by verifying compliance with an ambiguous regulation so long as one's actions satisfied any possible interpretation of the regulation. The government's argument finds support in the Ninth Circuit's recent case of United States ex rel. Oliver v. Parsons Co., 195 F.3d 457, 460, 463 (9th Cir. 1999) (court's interpretation of ambiguous regulation determines whether claim of compliance with regulation was false), cert. denied, 530 U.S. 1228 (2000). The defendants fall back from this position, arguing that the district court "did not . . . hold that the ambiguity of the anesthesia regulations negated a finding of falsity." Instead, the defendants contend the district court held that "many considerations precluded a finding of intent."

The False Claims Act prohibits the knowing presentation of false claims for government payment or approval. 31 U.S.C. § 3729(a). The Act defines "knowing" and "knowingly" to mean that the actor had actual knowledge of the pertinent information or acted in deliberate ignorance or in reckless disregard of the truth or falsity of that information. Sec. 3729(b). The question on intent here is whether the defendants knew (or would have known absent deliberate blindness or reckless disregard) that their bills would lead the government to believe that they had provided services that they actually did not provide. If a statement alleged to be false is ambiguous, the government (or here, the relator) must establish the defendant's knowledge of the falsity of the statement, which it can do by introducing evidence of how the statement would have been understood in context. See United States v. Garfinkel, 29 F.3d 1253, 1256 (8th Cir. 1994) ("evidence offered at trial could potentially resolve any ambiguity on the face of the document"); United States v. Anderson, 579 F.2d 455, 460 (8th Cir. 1978) ("In light of these ambiguities . . . the government must negative any reasonable interpretation that would make the defendant's statement factually correct."); United States v. Mackby, 261 F.3d 821, 827 (9th Cir. 2001) (False Claims Act violation consisted of filling in Medicare claim form contrary to instructions received in Medicare bulletins). If the Association shows the defendants certified compliance with the regulation knowing that the HCFA interpreted the regulations in a certain way and that their actions did not satisfy the requirements of the regulation as the HCFA interpreted it, any possible ambiguity of the regulations is water under the bridge. However, it is important to remember that the standard for liability is knowing, not negligent, presentation of a false claim. Oliver, 195 F.3d at 464-65.

The alleged ambiguity is limited to the meaning of the requirement in the 1992 regulation 42 C.F.R. § 414.46(c)(2)(ii), that an anesthesiologist must be "continuously involved" in a case in order to have personally performed an anesthesia case in which an anesthetist was also "involved." The defendants contend that there was confusion about what was required of an anesthesiologist in order to bill a case as personally performed or "AA." The record shows that up until September 1993, while there may have been some uncertainty about the interpretation of "continuously involved," the defendants were on notice of the possibility that they were expected to be present with the anesthetist in order to represent that they had personally performed a case. There is at least a question of fact as to their state of mind during this early period. During this time frame, defendant Allina's in-house lawyer advised that it was his understanding that an anesthesiologist had to be "continuously present" with the anesthetist to bill for personal performance. The same lawyer inquired of Travelers, the Medicare carrier for the Twin Cities area, whether an anesthesiologist had to be in the operating room the whole time to bill a single case as personally performed. Travelers agreed to get an answer from HCFA to this question. The defendants contend that they considered this question settled by an HCFA memo, which Travelers relayed to its provider community in September 1993, and which Blue Cross, carrier for the St. Cloud area, relayed to its providers in April 1994. The memo stated:

It has been reported that anesthesiologists will bill using the AA modifier even though they are outside the operating room performing other activities, such as pain blocks, doing pre or post operative evaluations, or administering and /or monitoring a labor epidural. For the anesthesiologist to bill using the AA modifier [for personally performed case] under these circumstances, he must be physically present in the operating suite while the [anesthetist] is attending to the case. If the anesthesiologist is not continuously involved with the case, then it is considered neither personally performed nor medically directed.

(emphasis added). The defendants introduced evidence that at least some of them relied on this memo in forming the belief that they could bill for personally performing cases despite leaving the operating room, so long as they were present in the operating suite, which they define as the area in the hospital where surgery takes place.

In response, the Association contends that it was not reasonable to read this memo as authorizing anesthesiologists to bill for personal performance when they were not in the room with the patient. The Association argues that the 1993 HCFA memorandum was not intended to authorize anesthesiologists to leave while an anesthetist performed their one-on-one cases, so long as they stayed in the operating suite. Instead, as the HCFA pointed out in April 1996, the intended point of the 1993 memorandum was to emphasize that, whoever physically performed the work in a one-on-one case, the anesthesiologist had to be solely devoted to that case in order to bill it as personally performed. The 1996 HCFA memo reasoned: "It should be assumed that, if the physician leaves the operating room, he/she is performing other duties. If the physician leaves the operating room to perform any other duties, the anesthesia procedure may not be billed as personally performed."

A few months after Travelers' dissemination of the 1993 HCFA memorandum, a further memorandum from HCFA on the subject of personally performed procedures was published in the American Society of Anesthesiologists newsletter of April 1994. This memorandum made it clear that anesthesiologists were not to leave a patient during a personally performed procedure. The memo stated that an anesthesiologist performing medical direction of concurrent procedures could "momentarily leave that procedure and perform another physician service" so long as this did not occur during a demanding part of the procedure. The memo contrasted the requirements for medically directed procedures with those for personally performed procedures: "Of course, we have not extended this policy to the case in which the anesthesiologist is personally performing the case. The reason for this is rather obvious. The anesthesiologist who is billing for personal performance of the case must personally perform the case. In theory, there is no one else to hand the case to." The memo concluded by saying that if the anesthetist, rather than the anesthesiologist was actually performing the case, then the anesthetist, rather than the anesthesiologist, should be paid for it. Thus, the time frame within which the 1993 HCFA memorandum could have been thought to have given the anesthesiologists permission to bill cases as personally performed when they were not immediately involved in the procedure was quite brief.

Even assuming that, for six months or so, the 1993 HCFA memorandum misled some defendants into believing that anesthesiologists could leave the operating room and still represent that they had personally performed the case, this would only rule out claims in which leaving the operating room was the only respect in which the anesthesiologist fell short of fulfilling the personal performance standard. The Association amassed a record that would support the conclusion that the anesthesiologists regularly fell short of the standard in other respects.

First, whether or not the anesthesiologists disqualified themselves per se from billing at the personal performance rate by absenting themselves from the operating room, various anesthetists testified by deposition and affidavit that the anesthesiologists did not merely step out of the room, but in fact were often gone for large periods of time and at crucial times.14 For instance, anesthetist Drew Mathews testified that in heart operations billed as personally performed, for over ninety percent of the cases he worked on, the anesthesiologist was present less than fifty percent of the time. Mathews kept extensive records of anesthesiologists' presence during the cases he worked on. Genevieve Crofoot testified that in one-on-one cases, the anesthesiologists would perform the induction and never come back. Kathleen Antoline also said that the anesthesiologists "circled 1 [denoting personally performed] and that was the end of their participation directly with that patient with me."

Second, the Association's witnesses stated that the anesthesiologists billed as personally performing cases when they were unavailable for emergencies on the case. Sometimes they were in a completely different part of the hospital during the case or even left the hospital. Mary Buchanan said anesthesiologists would circle one but would be unavailable for emergencies or would even leave the building. Kathleen Antoline said that a quarter of the time when she paged for emergencies, no one would come. Bart Barry testified that he had a case in which the anesthesiologist billed one-to-one despite being gone for three hours; when he returned, he had a blanket and appeared to have just woken up. Barry said that when someone is sleeping, the person may not hear a page. One anesthesiologist, John Magdsick, testified that he would consider himself in personal attendance of a patient as long as he was anywhere in the hospital. Nothing in the HCFA memorandum could have led the defendants to think that they could bill a case as personally performed when they were not present in the operating suite or when they were not available for emergencies in the case.

Third, many of the cases cited in the complaint involved anesthesiologists billing for personal performance while doing other duties inconsistent with personal performance, such as billing a concurrent case. Gayle McKay tesified that anesthesiologists at Abbott-Northwestern would routinely bill cases as personally performed when they had left the room to do other billable procedures with other patients. Other times, they would leave to do post-operative rounds. The Association also filed copies of records it contends were altered to conceal the fact that a case done concurrently with another was actually billed as personally performed. Additionally, the evidence of the anesthesiologists' protracted absences from the operating rooms may give rise to the reasonable inference that the anesthesiologists were actually engaged in other duties while they were gone from the room. The 1996 HCFA memo on this subject assumes that significant absence from the room during the surgery would indicate performance of other duties that would render personal performance billing inappropriate.

Defendants have certainly made no showing that they were led to believe that the kind of conduct outlined above qualified as the personal performance of an anesthesia case. They were not entitled to summary judgment on the allegations that they knowingly billed cases as personally performed when their services did not satisfy HCFA criteria for that designation.

The district court also entered summary judgment on the merits of the Association's claim that the anesthesiologists failed to participate in patients' "emergence" from anesthesia. As outlined earlier, the Medicare regulations required that anesthesiologists billing for "medical direction" of anesthetists must satisfy seven requirements in each case. One of those requirements was personal participation in "the most demanding procedures in the anesthesia plan, including induction and emergence." 42 C. F. R. § 405.552(a)(1)(iii) (1983).

The Association produced witness after witness who said that the anesthesiologists at the hospitals where they worked routinely left after induction of anesthesia and did not return for emergence. The district court summarized this evidence: "[T]he record shows that anesthesiologists routinely left the operating room before the end of the procedure, often speaking with patients in the recovery room or by telephone hours or even days after their procedures."

The Association's witnesses said that emergence occurs at the end of the surgery, in the operating room, and involves removing the breathing tube, allowing the patient to wake up, determining that the patient is stable, and finally taking the patient to the recovery room and relinquishing him to the care of a non-anesthesia caregiver. The Association produced letters from the American Society of Anesthesiologists and an excerpt from an anesthesia textbook, which were all consistent with this definition of emergence. The Association also presented the expert report of William Birnie, who worked for the HCFA when it was drafting the regulations in question. Birnie testified that the regulations were based on advice from the American Society of Anesthesiologists that emergence was a particularly demanding part of the anesthesia process and it occurred at "the end of the case when the surgical procedure has been completed and the patient is being prepared by the anesthesiologist to be turned over to a non-anesthesia provider."

Despite this extensive record, the district court held that "the overwhelming majority of the evidence on the record relating to the medical definition of [emergence]" supports the defendants' contention that emergence goes on for days. Therefore, according to the district court, the anesthesiologists did not need to participate in the extubation, stabilization, and transfer to the recovery room in order to fulfill the requirements of medical direction. Apparently the district court chose to disregard a record full of evidence contrary to its factual conclusion. This is impermissible on summary judgment, and we must therefore reverse.

The district court also entered summary judgment against the Association on its conspiracy claim, holding that there was no evidence that the anesthesiologists and hospitals had conspired to present false claims. We agree. The Association limits its attack on this holding to a footnote, and it presents no significant evidence for reversal. Accordingly, the judgment of the district court is reversed, except insofar as it enters summary judgment for the defendants on Count III of the Third Amended Complaint, alleging conspiracy, and on that part of Count V alleging conspiracy. In those two respects, the summary judgment is affirmed.