delivered the opinion of the Court.
Section 4-10 of Illinois’s Health Maintenance Organization Act, 215 Ill. Comp. Stat., ch. 125, §4-10 (2000), provides recipients of health coverage by such organizations with a right to independent medical review of certain denials of benefits. The issue in this ease is whether the statute, as applied to health benefits provided by a health maintenance organization under contract with an employee welfare benefit plan, is preempted by the Employee Retirement Income Security Act of 1974 (ERISA), 88 Stat. 832, as amended, 29 U. S. C. § 1001 et seq. We hold it is not.
I
Petitioner, Rush Prudential HMO, Inc., is a health maintenance organization (HMO) that contracts to provide medical services for employee welfare benefit plans covered by ERISA. Respondent Debra Moran is a beneficiary under one such plan, sponsored by her husband’s employer. Rush’s “Certificate of Group Coverage,” issued to employees who participate in employer-sponsored plans, promises that Rush will provide them with “medically necessary” services. The terms of the certificate give Rush the “broadest possible discretion” to determine whether a medical service claimed by a *360beneficiary is covered under the certificate. The certificate specifies that a service is covered as “medically necessary” if Rush finds:
“(a) [The service] is furnished or authorized by a Participating Doctor for the diagnosis or the treatment of a Sickness or Injury or for the maintenance of a person’s good health.
“(b) The prevailing opinion within the appropriate specialty of the United States medical profession is that [the service] is safe and effective for its intended use, and that its omission would adversely affect the person’s medical condition.
“(c) It is furnished by a provider with appropriate training, experience, staff and facilities to furnish that particular service or supply.” Record, PI. Exh. A, p. 21.
As the certificate explains, Rush contracts with physicians “to arrange for or provide services and supplies for medical care and treatment” of covered persons. Each covered person selects a primary care physician from those under contract to Rush, while Rush will pay for medical services by an unaffiliated physician only if the services have been “authorized” both by the primary care physician and Rush’s medical director. See id., at 11,16.
In 1996, when Moran began to have pain and numbness in her right shoulder, Dr. Arthur LaMarre, her primary care physician, unsuccessfully administered “conservative” treatments such as physiotherapy. In October 1997, Dr. LaMarre recommended that Rush approve surgery by an unaffiliated specialist, Dr. Julia Terzis, who had developed an unconventional treatment for Moran’s condition. Although Dr. La-Marre said that Moran would be “best served” by that procedure, Rush denied the request and, after Moran’s internal appeals, affirmed the denial on the ground that the procedure was not “medically necessary.” 230 F. 3d 959, 963 (CA7 *3612000). Rush instead proposed that Moran undergo standard surgery, performed by a physician affiliated with Rush.
In January 1998, Moran made a written demand for an independent medical review of her claim, as guaranteed by §4-10 of Illinois’s HMO Act, 215 Ill. Comp. Stat., ch. 125, § 4-10 et seq. (2000), which provides:
“Each Health Maintenance Organization shall provide a mechanism for the timely review by a physician holding the same class of license as the primary care physician, who is unaffiliated with the Health Maintenance Organization, jointly selected by the patient..., primary care physician and the Health Maintenance Organization in the event of a dispute between the primary care physician and the Health Maintenance Organization regarding the medical necessity of a covered service proposed by a primary care physician. In the event that the reviewing physician determines the covered service to be medically necessary, the Health Maintenance Organization shall provide the covered service.”
The Act defines a “Health Maintenance Organization” as
“any organization formed under the laws of this or another state to provide or arrange for one or more health care plans under a system which causes any part of the risk of health care delivery to be borne by the organization or its providers.” Ch. 125, § 1-2.1
*362When Rush failed to provide the independent review, Moran sued in an Illinois state court to compel compliance with the state Act. Rush removed the suit to Federal District Court, arguing that the cause of action was “completely preempted” under ERISA. 230 F. 3d, at 964.
While the suit was pending, Moran had surgery by Dr. Terzis at her own expense and submitted a $94,841.27 reimbursement claim to Rush. Rush treated the claim as a renewed request for benefits and began a new inquiry to determine coverage. The three doctors consulted by Rush said the surgery had been medically unnecessary.
Meanwhile, the federal court remanded the case back to state court on Moran’s motion, concluding that because Moran’s request for independent review under § 4-10 would not require interpretation of the terms of an ERISA plan, the claim was not “completely preempted” so as to permit removal under 28 U. S. C. § 1441.2 230 F. 3d, at 964. The state court enforced the state statute and ordered Rush to submit to review by an independent physician. The doctor selected was a reconstructive surgeon at Johns Hopkins Medical Center, Dr. A. Lee Dellon. Dr. Dellon decided that Dr. Terzis’s treatment had been medically necessary, based on the definition of medical necessity in Rush’s Certificate of *363Group Coverage, as well as his own medical judgment. Rush’s medical director, however, refused to concede that the surgery had been medically necessary, and denied Moran’s claim in January 1999.
Moran amended her complaint in state court to seek reimbursement for the surgery as “medically necessary” under Illinois’s HMO Act, and Rush again removed to federal court, arguing that Moran’s amended complaint stated a claim for ERISA benefits and was thus completely preempted by ERISA’s civil enforcement provisions, 29 U. S. C. § 1132(a), as construed by this Court in Metropolitan Life Ins. Co. v. Taylor, 481 U. S. 58 (1987). The District Court treated Moran’s claim as a suit under ERISA, and denied the claim on the ground that ERISA preempted Illinois’s independent review statute.3
The Court of Appeals for the Seventh Circuit reversed. 230 F. 3d 959 (2000). Although it found Moran’s state-law reimbursement claim completely preempted by ERISA so as to place the case in federal court, the Seventh Circuit did not agree that the substantive provisions of Illinois’s HMO Act were so preempted. The court noted that although ERISA broadly preempts any state laws that “relate to” employee benefit plans, 29 U. S. C. § 1144(a), state laws that “regulat[e] *364insurance” are saved from preemption, § 1144(b)(2)(A). The court held that the Illinois HMO Act was such a law, the independent review requirement being little different from a state-mandated contractual term of the sort this Court had held to survive ERISA preemption. See 280 F. 3d, at 972 (citing UNUM Life Ins. Co. of America v. Ward, 526 U. S. 358, 375-376 (1999)). The Seventh Circuit rejected the contention that Illinois’s independent review requirement constituted a forbidden “alternative remedy” under this Court’s holding in Pilot Life Ins. Co. v. Dedeaux, 481 U. S. 41 (1987), and emphasized that § 4-10 does not authorize any particular form of relief in state courts; rather, with respect to any ERISA health plan, the judgment of the independent reviewer is only enforceable in an action brought under ERISA’s civil enforcement scheme, 29 U. S. C. § 1132(a). 230 F. 3d, at 971.
Because the decision of the Court of Appeals conflicted with the Fifth Circuit’s treatment of a similar provision of Texas law in Corporate Health Ins., Inc. v. Texas Dept. of Ins., 215 F. 3d 526 (2000), we granted certiorari, 533 U. S. 948 (2001). We now affirm.
II
To “safeguard] ... the establishment, operation, and administration” of employee benefit plans, ERISA sets “minimum standards ... assuring the equitable character of such plans and their financial soundness,” 29 U. S. C. § 1001(a), and contains an express preemption provision that ERISA “shall supersede any and all State laws insofar as they may now or hereafter relate to any employee benefit plan . . . .” § 1144(a). A saving clause then reclaims a substantial amount of ground with its provision that “nothing in this subchapter shall be construed to exempt or relieve any person from any law of any State which regulates insurance, banking, or securities.” § 1144(b)(2)(A). The “unhelpful” drafting of these antiphonal clauses, New York State Confer*365ence of Blue Cross & Blue Shield Plans v. Travelers Ins. Co., 514 U. S. 645, 656 (1995), occupies a substantial share of this Court’s time, see, e. g., Egelhoff v. Egelhoff, 532 U. S. 141 (2001); UNUM Life Ins. Co. of America v. Ward, supra; California Div. of Labor Standards Enforcement v. Dillingham Constr., N. A., Inc., 519 U. S. 316 (1997); Metropolitan Life Ins. Co. v. Massachusetts, 471 U. S. 724 (1985). In trying to extrapolate congressional intent in a case like this, when congressional language seems simultaneously to preempt everything and hardly anything, we “have no choice” but to temper the assumption that “‘the ordinary meaning . . . accurately expresses the legislative purpose,’” id., at 740 (quoting Park ’N Fly v. Dollar Park & Fly, Inc., 469 U. S. 189, 194 (1985)), with the qualification “ ‘that the historic police powers of the States were not [meant] to be superseded by the Federal Act unless that was the clear and manifest purpose of Congress.’” Travelers, supra, at 655 (quoting Rice v. Santa Fe Elevator Corp., 331 U. S. 218, 230 (1947)).
It is beyond serious dispute that under existing precedent § 4-10 of the Illinois HMO Act “relates to” employee benefit plans within the meaning of § 1144(a). The state law bears “indirectly but substantially on all insured benefit plans,” Metropolitan Life, 471 U. S., at 739, by requiring them to submit to an extra layer of review for certain benefit denials if they purchase medical coverage from any of the common types of health care organizations covered by the state law’s definition of HMO. As a law that “relates to” ERISA plans under § 1144(a), §4-10 is saved from preemption only if it also “regulates insurance” under § 1144(b)(2)(A). Rush insists that the Act is not such a law.
A
In Metropolitan Life, we said that in deciding whether a law “regulates insurance” under ERISA’s saving clause, we start with a “common-sense view of the matter,” 471 U. S., *366at 740, under which “a law must not just have an impact on the insurance industry, but must be specifically directed toward that industry.” Pilot Life Ins. Co. v. Dedeaux, supra, at 50. We then test the results of the commonsense enquiry by employing the three factors used to point to insurance laws spared from federal preemption under the McCarran-Ferguson Act, 15 U. S. C. § 1011 et seq,4 Although this is not the place to plot the exact perimeter of the saving clause, it is generally fair to think of the combined “commonsense” and McCarran-Ferguson factors as parsing the “who” and the “what”: when insurers are regulated with respect to their insurance practices, the state law survives ERISA. Cf Group Life & Health Ins. Co. v. Royal Drug Co., 440 U. S. 205, 211 (1979) (explaining that the “business of insurance” is not coextensive with the “business of insurers”).
1
The commonsense enquiry focuses on “primary elements of an insurance contract^ which] are the spreading and underwriting of a policyholder’s risk.” Ibid. The Illinois statute addresses these elements by defining “health maintenance organization” by reference to the risk that it bears. See 215 Ill. Comp. Stat., ch. 125, § 1-2(9) (2000) (an HMO “provide[s] or arrange[s] for . . . health care plans under a system which causes any part of the risk of health care delivery to be borne by the organization or its providers”).
Rush contends that seeing an HMO as an insurer distorts the nature of an HMO, which is, after all, a health care provider, too. This, Rush argues, should determine its characterization, with the consequence that regulation of an HMO is not insurance regulation within the meaning of ERISA.
*367The answer to Rush is, of course, that an HMO is both: it provides health care, and it does so as an insurer. Nothing in the saving clause requires an either-or choice between health care and insurance in deciding a preemption question, and as long as providing insurance fairly accounts for the application of state law, the saving clause may apply. There is no serious question about that here, for it would ignore the whole purpose of the HMO-style of organization to conceive of HMOs (even in the traditional sense, see n. 1, supra) without their insurance element.
“The defining feature of an HMO is receipt of a fixed fee for each patient enrolled under the terms of a contract to provide specified health care if needed.” Pegram v. Herdrich, 530 U. S. 211, 218 (2000). “The HMO thus assumes the financial risk of providing the benefits promised: if a participant never gets sick, the HMO keeps the money regardless, and if a participant becomes expensively ill, the HMO is responsible for the treatment....” Id., at 218-219. The HMO design goes beyond the simple truism that all contracts are, in some sense, insurance against future fluctuations in price, R. Posner, Economic Analysis of Law 104 (4th ed. 1992), because HMOs actually underwrite and spread risk among their participants, see, e. g., R. Shouldice, Introduction to Managed Care 450-462 (1991), a feature distinctive to insurance, see, e. g., SEC v. Variable Annuity Life Ins. Co. of America, 359 U. S. 65, 73 (1959) (underwriting of risk is an “earmark of insurance as it has commonly been conceived of in popular understanding and usage”); Royal Drug, supra, at 214-215, n. 12 (“[U]nless there is some element of spreading risk more widely, there is no underwriting of risk”).
So Congress has understood from the start, when the phrase “Health Maintenance Organization” was established and defined in the HMO Act of 1973. The Act was intended to encourage the development of HMOs as a new form of health care delivery system, see S. Rep. No. 93-129, pp. 7-9 *368(1973), and when Congress set the standards that the new health delivery organizations would have to meet to get certain federal benefits, the terms included requirements that the organizations bear and manage risk. See, e. g., Health Maintenance Organization Act of 1973, § 1301(c), 87 Stat. 916, as amended, 42 U. S. C. § 300e(c); S. Rep. No. 93-129, at 14 (explaining that HMOs necessarily bear some of the risk of providing service, and requiring that a qualifying HMO “as-sum[e] direct financial responsibility, without benefit of reinsurance, for care ... in excess of the first five thousand dollars per enrollee per year”). The Senate Committee Report explained that federally qualified HMOs would be required to provide “a basic package of benefits, consistent with existing health insurance patterns,” id., at 10, and the very text of the Act assumed that state insurance laws would apply to HMOs; it provided that to the extent state insurance capitalization and reserve requirements were too stringent to permit the formation of HMOs, “qualified” HMOs would be exempt from such limiting regulation. See § 1311, 42 U. S. C. §300e-10. This congressional understanding that it was promoting a novel form of insurance was made explicit in the Senate Report’s reference to the practices of “health insurers to charge premium rates based upon the actual claims experience of a particular group of subscribers,” thus “raising costs and diminishing the availability of health insurance for those suffering from costly illnesses,” S. Rep. No. 93-129, at 29-30. The federal Act responded to this insurance practice by requiring qualifying HMOs to adopt uniform capitation rates, see § 1301(b), 42 U. S. C. § 300e(b), and it was because of that mandate “pos[ing] substantial competitive problems to newly emerging HMOs,” S. Rep. No. 93-129, at 30, that Congress authorized funding subsidies, see §1304, 42 U. S. C. § 300e-4. The Senate explanation left no doubt that it viewed an HMO as an insurer; the subsidy was justified because “the same stringent requirements do not apply to other indemnity or service benefits insurance plans.” *3695. Rep. No. 93-129, at 30. In other words, one year before it passed ERISA, Congress itself defined HMOs in part by reference to risk, set minimum standards for managing the risk, showed awareness that States regulated HMOs as insurers, and compared HMOs to “indemnity or service benefits insurance plans.”
This conception has not changed in the intervening years. Since passage of the federal Act, States have been adopting their own HMO enabling Acts, and today, at least 40 of them, including Illinois, regulate HMOs primarily through the States’ insurance departments, see Aspen Health Law and Compliance Center, Managed Care Law Manual 31-32 (Supp. 6, Nov. 1997), although they may be treated differently from traditional insurers, owing to their additional role as health care providers,5 see, e. g., Alaska Ins. Code § 21.86.010 (2000) (health department reviews HMO before insurance commissioner grants a certificate of authority); Ohio Rev. Code Ann. § 1742.21 (West 1994) (health department may inspect HMO). Finally, this view shared by Congress and the States has passed into common understanding. HMOs (broadly defined) have “grown explosively in the past decade and [are] now the dominant form of health plan coverage for privately insured individuals.” Gold & Hurley, The Role of Managed Care “Products” in Managed Care “Plans,” in Contemporary Managed Care 47 (M. Gold ed. 1998). While the original form of the HMO was a single corporation employing its own physicians, the 1980’s saw a variety of other types of structures develop even as traditional insurers altered their own *370plans by adopting HMO-like cost-control measures. See Weiner & de Lissovoy, Razing a Tower of Babel: A Taxonomy for Managed Care and Health Insurance Plans, 18 J. of Health Politics, Policy and Law 75, 83 (Spring 1993). The dominant feature is the combination of insurer and provider, see Gold & Hurley, supra, at 47, and “an observer may be hard pressed to uncover the differences among products that bill themselves as HMOs, [preferred provider organizations], or managed care overlays to health insurance,” Managed Care Law Manual, supra, at 1. Thus, virtually all commentators on the American health care system describe HMOs as a combination of insurer and provider, and observe that in recent years, traditional “indemnity” insurance has fallen out of favor. See, e. g., Weiner & de Lissovoy, supra, at 77 (“A common characteristic of the new managed care plans was the degree to which the roles of insurer and provider became integrated”); Gold, Understanding the Roots: Health Maintenance Organizations in Historical Context, in Contemporary Managed Care, supra, at 7,8,13; Managed Care Law Manual, supra, at 1; R. Rosenblatt, S. Law, & S. Rosenbaum, Law and the American Health Care System 552 (1997); Shouldice, Introduction to Managed Care, at 13, 20. Rush cannot checkmate common sense by trying to submerge HMOs’ insurance features beneath an exclusive characterization of HMOs as providers of health care.
2
On a second tack, Rush and its amici dispute that § 4-10 is aimed specifically at the insurance industry. They say the law sweeps too broadly with definitions capturing organizations that provide no insurance, and by regulating noninsurance activities of HMOs that do. Rush points out that Illinois law defines HMOs to include organizations that cause the risk of health care delivery to be borne by the organization itself, or by “its providers.” 215 Ill. Comp. Stat., ch. 125, § 1-2(9) (2000). In Rush’s view, the reference to “its *371providers” suggests that an organization may be an HMO under state law (and subject to §4-10) even if it does not bear risk itself, either because it has “devolve[d]” the risk of health care delivery onto others, or because it has contracted only to provide “administrative” or other services for self-funded plans. Brief for Petitioner 38.
These arguments, however, are built on unsound assumptions. Rush’s first contention assumes that an HMO is no longer an insurer when it arranges to limit its exposure, as when an HMO arranges for capitated contracts to compensate its affiliated physicians with a set fee for each HMO patient regardless of the treatment provided. Under such an arrangement, Rush claims, the risk is not borne by the HMO at all. In a similar vein, Rush points out that HMOs may contract with third-party insurers to protect themselves against large claims.
The problem with Rush’s argument is simply that a reinsurance contract does not take the primary insurer out of the insurance business, cf. Hartford Fire Ins. Co. v. California, 509 U. S. 764 (1993) (applying McCarran-Ferguson to a dispute involving primary insurers and reinsurers); id., at 772-773 (“[Pjrimary insurers ... usually purchase insurance to cover a portion of the risk they assume from the consumer”), and capitation contracts do not relieve the HMO of its obligations to the beneficiary. The HMO is still bound to provide medical care to its members, and this is so regardless of the ability of physicians or third-party insurers to honor their contracts with the HMO.
Nor do we see anything standing in the way of applying the saving clause if we assume that the general state definition of HMO would include a contractor that provides only administrative services for a self-funded plan.6 Rush points *372out that the general definition of HMO under Illinois law includes not only organizations that “provide” health care plans, but those that “arrange for” them to be provided, so long as “any part of the risk of health care delivery” rests upon “the organization or its providers.” 215 Ill. Comp. Stat., ch. 125, § 1-2(9) (2000). See Brief for Petitioner 38. Rush hypothesizes a sort of medical matchmaker, bringing together ERISA plans and medical care providers; even if the latter bear all the risks, the matchmaker would be an HMO under the Illinois definition. Rush would conclude from this that §4-10 covers noninsurers, and so is not directed specifically to the insurance industry. Ergo, ERISA’s saving clause would not apply.
It is far from clear, though, that the terms of § 4-10 would even theoretically apply to the matchmaker, for the requirement that the HMO “provide” the covered service if the independent reviewer finds it medically necessary seems to assume that the HMO in question is a provider, not the mere arranger mentioned in the general definition of an HMO. Even on the most generous reading of Rush’s argument, however, it boils down to the bare possibility (not the likelihood) of some overbreadth in the application of § 4-10 beyond orthodox HMOs, and there is no reason to think Congress would have meant such minimal application to noninsurers to remove a state law entirely from the category of insurance regulation saved from preemption.
In sum, prior to ERISA’s passage, Congress demonstrated an awareness of HMOs as risk-bearing organizations subject to state insurance regulation, the state Act defines HMOs by reference to risk bearing, HMOs have taken over much business formerly performed by traditional indemnity insurers, and they are almost universally regulated as insurers under state law. That HMOs are not traditional “indem*373nity” insurers is no matter; “we would not undertake to freeze the concepts of ‘insurance’... into the mold they fitted when these Federal Acts were passed.” SEC v. Variable Annuity Life Ins. Co. of America, 359 U. S., at 71. Thus, the Illinois HMO Act is a law “directed toward” the insurance industry, and an “insurance regulation” under a “commonsense” view.
B
The McCarran-Ferguson factors confirm our conclusion. A law regulating insurance for McCarran-Ferguson purposes targets practices or provisions that “ha[ve] the effect of transferring or spreading a policyholder’s risk;... [that are] an integral part of the policy relationship between the insurer and the insured; and [are] limited to entities within the insurance industry.” Union Labor Life Ins. Co. v. Pireno, 458 U. S. 119, 129 (1982). Because the factors are guideposts, a state law is not required to satisfy all three McCarran-Ferguson criteria to survive preemption, see UNUM Life Ins. Co. v. Ward, 526 U. S., at 373, and so we follow our precedent and leave open whether the review mandated here may be described as going to a practice that “spread[s] a policyholder’s risk.” For in any event, the second and third factors are clearly satisfied by § 4-10.
It is obvious enough that the independent review requirement regulates “an integral part of the policy relationship between the insurer and the insured.” Illinois adds an extra layer of review when there is internal disagreement about an HMO’s denial of coverage. The reviewer applies both a standard of medical care (medical necessity) and characteristically, as in this case, construes policy terms. Cf Pegram v. Herdrich, 530 U. S., at 228-229. The review affects the “policy relationship” between HMO and covered persons by translating the relationship under the HMO agreement into concrete terms of specific obligation or freedom from duty. Hence our repeated statements that the interpretation of insurance contracts is at the “core” of the business of *374insurance. E. g., SEC v. National Securities, Inc., 393 U. S. 453, 460 (1969).
Rush says otherwise, citing Union Labor Life Ins. Co. v. Pireno, supra, and insisting that that case holds external review of coverage decisions to be outside' the “policy relationship.” But Rush misreads Pireno. We held there that an insurer’s use of a “peer review” committee to gauge the necessity of particular treatments was not a practice integral to the policy relationship for the purposes of McCarran-Ferguson. 458 U. S., at 131-132. We emphasized, however, that the insurer’s resort to peer review was simply the insurer’s unilateral choice to seek advice if and when it cared to do so. The policy said nothing on the matter. The insurer’s contract for advice from a third party was no concern of the insured, who was not bound by the peer review committee’s recommendation any more, for that matter, than the insurer was. Thus it was not too much of an exaggeration to conclude that the practice was “a matter of indifference to the policyholder,” id., at 132. Section 4-10, by contrast, is different on all counts, providing as it does a legal right to the insured, enforceable against the HMO, to obtain an authoritative determination of the HMO’s medical obligations.
The final factor, that the law be aimed at a “practice . . . limited to entities within the insurance industry,” id., at 129, is satisfied for many of the same reasons that the law passes the commonsense test. The law regulates application of HMO contracts and provides for review of claim denials; once it is established that HMO contracts are, in fact, contracts for insurance (and not merely contracts for medical care), it is clear that §4-10 does not apply to entities outside the insurance industry (although it does not, of course, apply to all entities within it).
Even if we accepted Rush’s contention, rejected already, that the law regulates HMOs even when they act as pure administrators, we would still find the third factor satisfied. *375That factor requires the targets of the law to be limited to entities within the insurance industry, and even a matchmaking HMO would fall within the insurance industry. But the implausibility of Rush’s hypothesis that the pure administrator would be bound by § 4-10 obviates any need to say more under this third factor. Cf. Barnett Bank of Marion Cty., N. A. v. Nelson, 517 U. S. 25, 39 (1996) (holding that a federal statute permitting banks to act as agents of insurance companies, although not insurers themselves, was a statute regulating the “business of insurance” for McCarran-Ferguson purposes).
Ill
Given that §4-10 regulates insurance, ERISA’s mandate that “nothing in this subchapter shall be construed to exempt or relieve any person from any law of any State which regulates insurance,” 29 U. S. C. § 1144(b)(2)(A), ostensibly forecloses preemption. See Metropolitan Life, 471 U. S., at 746 (“If a state law ‘regulates insurance,’ ... it is not preempted”). Rush, however, does not give up. It argues for preemption anyway, emphasizing that the question is ultimately one of congressional intent, which sometimes is so clear that it overrides a statutory provision designed to save state law from being preempted. See American Telephone & Telegraph Co. v. Central Office Telephone, Inc., 524 U. S. 214, 227 (1998) (AT&T) (clause in Communications Act of 1934 purporting to save “the remedies now existing at common law or by statute,” 47 U. S. C. § 414 (1994 ed.), defeated by overriding policy of the filed-rate doctrine); Adams Express Co. v. Croninger, 226 U. S. 491, 507 (1913) (saving clause will not sanction state laws that would nullify policy expressed in federal statute; “the act cannot be said to destroy itself” (internal quotation marks omitted)).
In ERISA law, we have recognized one example of this sort of overpowering federal policy in the civil enforcement provisions, 29 U. S. C. § 1132(a), authorizing civil actions for *376six specific types of relief.7 In Massachusetts Mut. Life Ins. Co. v. Russell, 473 U. S. 134 (1985), we said those provisions amounted to an “interlocking, interrelated, and interdependent remedial scheme," id., at 146, which Pilot Life described as “representing] a careful balancing of the need for prompt and fair claims settlement procedures against the public interest in encouraging the formation of employee benefit plans,” 481 U. S., at 54. So, we have held, the civil enforcement provisions are of such extraordinarily preemptive power that they override even the “well-pleaded complaint” rule for establishing the conditions under which a cause of action may be removed to a federal forum. Metropolitan Life Ins. Co. v. Taylor, 481 U. S., at 63-64.
*377A
Although we have yet to encounter a forced choice between the congressional policies of exclusively federal remedies and the “reservation of the business of insurance to the States,” Metropolitan Life, 471 U. S., at 744, n. 21, we have anticipated such a conflict, with the state insurance regulation losing out if it allows plan participants “to obtain remedies ... that Congress rejected in ERISA,” Pilot Life, supra, at 54.
In Pilot Life, an ERISA plan participant who had been denied benefits sued in a state court on state tort and contract claims. He sought not merely damages for breach of contract, but also damages for emotional distress and punitive damages, both of which we had held unavailable under relevant ERISA provisions. Russell, supra, at 148. We not only rejected the notion that these common law contract claims “regulat[ed] insurance,” Pilot Life, 481 U. S., at 50-51, but went on to say that, regardless, Congress intended a “federal common law of rights and obligations” to develop under ERISA, id., at 56, without embellishment by independent state remedies. As in AT&T, we said the saving clause had to stop short of subverting congressional intent, clearly expressed “through the structure and legislative history!,] that the federal remedy . . . displace state causes of action.” 481 U. S., at 57.8
Rush, says that the day has come to turn dictum into holding by declaring that the state insurance regulation, § 4-10, is preempted for creating just the kind of “alternative remedy” we disparaged in Pilot Life. As Rush sees it, the inde*378pendent review procedure is a form of binding arbitration that allows an ERISA beneficiary to submit claims to a new decisionmaker to examine Rush’s determination de novo, supplanting judicial review under the “arbitrary and capricious” standard ordinarily applied when discretionary plan interpretations are challenged. Firestone Tire & Rubber Co. v. Bruch, 489 U. S. 101, 110-112 (1989). Rush says that the beneficiary’s option falls within Pilot Life’s notion of a remedy that “supplement^] or supplant[s]” the remedies available under ERISA. 481 U. S., at 56.
We think, however, that Rush overstates the rule expressed in Pilot Life. The enquiry into state processes alleged to “supplemen[t] or supplan[t]” the federal scheme by allowing beneficiaries “to obtain remedies under state law that Congress rejected in ERISA,” id., at 54, has, up to now, been far more straightforward than it is here. The first case touching on the point did not involve preemption at all; it arose from an ERISA beneficiary’s reliance on ERISA’s own enforcement scheme to claim a private right of action for types of damages beyond those expressly provided. Russell, 473 U. S., at 145. We concluded that Congress had not intended causes of action under ERISA itself beyond those specified in § 1132(a). Id., at 148. Two years later we determined in Metropolitan Life Ins. Co. v. Taylor, supra, that Congress had so completely preempted the field of benefits law that an ostensibly state cause of action for benefits was necessarily a “creature of federal law” removable to federal court. Id., at 64 (internal quotation marks omitted). Russell and Taylor naturally led to the holding in Pilot Life that ERISA would not tolerate a diversity action seeking monetary damages for breach generally and for consequential emotional distress, neither of which Congress had authorized in § 1132(a). These monetary awards were claimed as remedies to be provided at the ultimate step of plan enforcement, and even if they could have been characterized as products of “insurance regulation,” they would have signifi*379cantly expanded the potential scope of ultimate liability imposed upon employers by the ERISA scheme.
Since Pilot Life, we have found only one other state law to “conflict” with § 1132(a) in providing a prohibited alternative remedy. In Ingersoll-Rand Co. v. McClendon, 498 U. S. 133 (1990), we had no trouble finding that Texas’s tort of wrongful discharge, turning on an employer’s motivation to avoid paying pension benefits, conflicted with ERISA enforcement; while state law duplicated the elements of a claim available under ERISA, it converted the remedy from an equitable one under § 1132(a)(3) (available exclusively in federal district courts) into a legal one for money damages (available in a state tribunal). Thus, Ingersoll-Rand fit within the category of state laws Pilot Life had held to be incompatible with ERISA’s enforcement scheme; the law provided a form of ultimate relief in a judicial forum that added to the judicial remedies provided by ERISA. Any such provision patently violates ERISA’s policy of inducing employers to offer benefits by assuring a predictable set of liabilities, under uniform standards of primary conduct and a uniform regime of ultimate remedial orders and awards when a violation has occurred. See Pilot Life, supra, at 56 (“‘The uniformity of decision ... will help administrators ... predict the legality of proposed actions without the necessity of reference to varying state laws’ ” (quoting H. R. Rep. No. 93-533, p. 12 (1973))); 481 U. S., at 56 (“The expectations that a federal common law of rights and obligations under ERISA-regulated plans would develop ... would make little, sense if the remedies available to ERISA participants and beneficiaries under [§ 1132(a)] could be supplemented or supplanted by varying state laws”).
But this case addresses a state regulatory scheme that provides no new cause of action under state law and authorizes no new form of ultimate relief. While independent review under §4-10 may well settle the fate of a benefit claim under a particular contract, the state statute does not en*380large the claim beyond the benefits available in any action brought under § 1132(a). And although the reviewer’s determination would presumably replace that of the HMO as to what is “medically necessary” under this contract,9 the relief ultimately available would still be what ERISA authorizes in a suit for benefits under § 1132(a).10 This case therefore does not involve the sort of additional claim or remedy exemplified in Pilot Life, Russell, and Ingersoll-Rand, but instead bears a resemblance to the claims-procedure rule that we sustained in UNUM Life Ins. Co. of America v. Ward, 526 U. S. 358 (1999), holding that a state law barring enforcement of a policy’s time limitation on submitting claims did not conflict with § 1132(a), even though the state “rule of decision,” id., at 377, could mean the difference between success and failure for a beneficiary. The procedure provided by § 4-10 does not fall within Pilot Life’s categorical preemption.
*381B
Rush still argues for going beyond Pilot Life, making the preemption issue here one of degree, whether the state procedural imposition interferes unreasonably with Congress’s intention to provide a uniform federal regime of “rights and obligations” under ERISA. However, “[s]uch disuniformities ... are the inevitable result of the congressional decision to ‘save’ local insurance regulation.” Metropolitan Life, 471 U. S., at 747.11 Although we have recognized a limited exception from the saving clause for alternative causes of action and alternative remedies in the sense described above, we have never indicated that there might be additional justifications for qualifying the clause’s application. Rush’s arguments today convince us that further limits on insurance regulation preserved by ERISA are unlikely to deserve recognition.
To be sure, a State might provide for a type of “review” that would so resemble an adjudication as to fall within Pilot Life’s categorical bar. Rush, and the dissent, post, at 394 (opinion of Thomas, J.), contend that § 4-10 fills that bill by imposing an alternative scheme of arbitral adjudication at *382odds with the manifest congressional purpose to confine adjudication of disputes to the courts. It does not turn out to be this simple, however, and a closer look at the state law reveals a scheme significantly different from common arbitration as a way of construing and applying contract terms.
In the classic sense, arbitration occurs when “parties in dispute choose a judge to render a final and binding decision on the merits of the controversy and on the basis of proofs presented by the parties.” 1 I. MacNeil, R. Speidel, & T. Stipanowich, Federal Arbitration Law §2.1.1 (1995) (internal quotation marks omitted); see also Uniform Arbitration Act § 5, 7 U. L. A. 173 (1997) (discussing submission evidence and empowering arbitrator to “hear and determine the controversy upon the evidence produced”); Commercial Dispute Resolution Procedures of the American Arbitration Association ¶ ¶ R33-R35 (Sept. 2000) (discussing the taking of evidence). Arbitrators typically hold hearings at which parties may submit evidence and conduct cross-examinations, e. g., Uniform Arbitration Act §5, and are often invested with many powers over the dispute and the parties, including the power to subpoena witnesses and administer oaths, e. g., Federal Arbitration Act, 9 U. S. C. § 7; 28 U. S. C. § 653; Uniform Arbitration Act § 7, 7 U. L. A., at 199; Cal. Civ. Proc. Code Ann. §§ 1282.6, 1282.8 (West 1982).
Section 4-10 does resemble an arbitration provision, then, to the extent that the independent reviewer considers disputes about the meaning of the HMO contract12 and receives “evidence” in the form of medical records, statements from *383physicians, and the like. But this is as far as the resemblance to arbitration goes, fór the other features of review under §4-10 give the proceeding a different character, one not at all at odds with the policy behind § 1132(a). The Act does not give the independent reviewer a free-ranging power to construe contract terms, but instead, confines review to a single term: the phrase “medical necessity,” used to define the services covered under the contract. This limitation, in turn, implicates a feature of HMO benefit determinations that we described in Pegram v. Herdrich, 530 U. S. 211 (2000). We explained that when an HMO guarantees medically necessary care, determinations of coverage “cannot be untangled from physicians’ judgments about reasonable medical treatment.” Id., at 229. This is just how the Illinois Act operates; the independent examiner must be a physician with credentials similar to those of the primary care physician, 215 Ill. Comp. Stat., ch. 125, §4-10 (2000), and is expected to exercise independent medical judgment in deciding what medical necessity requires. Accordingly, the reviewer in this case did not hold the kind of conventional evidentiary hearing common in arbitration, but simply received medical records submitted by the parties, and ultimately came to a professional judgment of his own. Tr. of Oral Arg. 30-32.
Once this process is set in motion, it does not resemble either contract interpretation or evidentiary litigation before a neutral arbiter, as much as it looks like a practice (having nothing to do with arbitration) of obtaining another medical opinion. The reference to an independent reviewer is similar to the submission to a second physician, which many health insurers are required by law to provide before denying coverage.13
The practice of obtaining a second opinion, however, is far removed from any notion of an enforcement scheme, and *384once §4-10 is seen as something akin to a mandate for second-opinion practice in order to ensure sound medical judgments, the preemption argument that arbitration under § 4-10 supplants judicial enforcement runs out of steam.
Next, Rush argues that § 4-10 clashes with a substantive rule intended to be preserved by the system of uniform enforcement, stressing a feature of judicial review highly prized by benefit plans: a deferential standard for reviewing benefit denials. Whereas Firestone Tire & Rubber Co. v. Bruch, 489 U. S., at 115, recognized that an ERISA plan could be designed to grant “discretion” to a plan fiduciary, deserving deference from a court reviewing a discretionary judgment, § 4-10 provides that when a plan purchases medical services and insurance from an HMO, benefit denials are subject to apparently de novo review. If a plan should continue to balk at providing a service the reviewer has found medically necessary, the reviewer’s determination could carry great weight in a subsequent suit for benefits under § 1132(a),14 depriving the plan of the judicial deference a fiduciary’s medical judgment might have obtained if judicial review of the plan’s decision had been immediate.15
Again, however, the significance of §4-10 is not wholly captured by Rush’s argument, which requires some perspec*385tive for evaluation. First, in determining whether state procedural requirements deprive plan administrators of any right to a uniform standard of review, it is worth recalling that ERISA itself provides nothing about the standard. It simply requires plans to afford a beneficiary some mechanism for internal review of a benefit denial, 29 U. S. C. § 1133(2), and provides a right to a subsequent judicial forum for a claim to recover benefits, § 1132(a)(1)(B). Whatever the standards for reviewing benefit denials may be, they cannot conflict with anything in the text of the statute, which we have read to require a uniform judicial regime of categories of relief and standards of primary conduct, not a uniformly lenient regime of reviewing benefit determinations. See Pilot Life, 481 U. S., at 56.16
Not only is there no ERISA provision directly providing a lenient standard for judicial review of benefit denials, but there is no requirement necessarily entailing such an effect even indirectly. When this Court dealt with the review standards on which the statute was silent, we held that a general or default rule of de novo review could be replaced *386by deferential review if the ERISA plan itself provided that the plan’s benefit determinations were matters of high or unfettered discretion, see Firestone Tire, supra, at 115. Nothing in ERISA, however, requires that these kinds of decisions be so “discretionary” in the first place; whether they are is simply a matter of plan design or the drafting of an HMO contract. In this respect, then, §4-10 prohibits designing an insurance contract so as to accord unfettered discretion to the insurer to interpret the contract’s terms. As such, it does not implicate ERISA’s enforcement scheme at all, and is no different from the types of substantive state regulation of insurance contracts we have in the past permitted to survive preemption, such as mandated-benefit statutes and statutes prohibiting the denial of claims solely on the ground of untimeliness.17 See Metropolitan Life Ins. Co. v. Massachusetts, 471 U. S. 724 (1985); UNUM Life Ins. Co. of America v. Ward, 526 U. S. 358 (1999).
* * *
In sum, §4-10 imposes no new obligation or remedy like the causes of action considered in Russell, Pilot Life, and Ingersoll-Rand. Even in its formal guise, the State Act bears a closer resemblance to second-opinion requirements than to arbitration schemes. Deferential review in the HMO context is not a settled given; §4-10 operates before the stage of judicial review; the independent reviewer’s de novo examination of the benefit claim mirrors the general or *387default rule we have ourselves recognized; and its effect is no greater than that of mandated-benefit regulation.
In deciding what to make of these facts and conclusions, it helps to go back to where we started and recall the ways States regulate insurance in looking out for the welfare of their citizens. Illinois has chosen to regulate insurance as one way to regulate the practice of medicine, which we have previously held to be permissible under ERISA, see Metropolitan Life, 471 U. S., at 741. While the statute designed to do this undeniably eliminates whatever may have remained of a plan sponsor’s option to minimize scrutiny of benefit denials, this effect of eliminating an insurer’s autonomy to guarantee terms congenial to its own interests is the stuff of garden variety insurance regulation through the imposition of standard policy terms. See id., at 742 (“[S]tate laws regulating the substantive terms of insurance contracts were commonplace well before the mid-70’s”). It is therefore hard to imagine a reservation of state power to regulate insurance that would not be meant to cover restrictions of the insurer’s advantage in this kind of way. And any lingering doubt about the reasonableness of § 4-10 in affecting the application of § 1132(a) may be put to rest by recalling that regulating insurance tied to what is medically necessary is probably inseparable from enforcing the quintessentially state-law standards of reasonable medical care. See Pegram v. Herdrich, 530 U. S., at 236. “[I]n the field of health care, a subject of traditional state regulation, there is no ERISA preemption without clear manifestation of congressional purpose.” Id., at 237. To the extent that benefit litigation in some federal courts may have to account for the effects of § 4-10, it would be an exaggeration to hold that the objectives of § 1132(a) are undermined. The saving clause is entitled to prevail here, and we affirm the judgment.
It is so ordered.
In the health care industry, the term “Health Maintenance Organization” has been defined as “[a] prepaid organized delivery system where the organization and the primary care physicians assume some financial risk for the care provided to its enrolled members. ... In a pure HMO, members must obtain care from within the system if it is to be reimbursed.” Weiner & de Lissovoy, Razing a Tower of Babel: A Taxonomy for Managed Care and Health Insurance Plans, 18 J. of Health Politics, Policy and Law 75, 96 (Spring 1993) (emphasis in original). The term “Managed Care Organization” is used more broadly to refer to any number of systems combining health care delivery with financing. Id., at 97. The Illinois definition of HMO does not appear to be limited to the tradi*362tional usage of that term, but instead is likely to encompass a variety of different structures (although Illinois does distinguish HMOs from pure insurers by regulating “traditional” health insurance in a different portion of its insurance laws, 215 Ill. Comp. Stat., ch. 5 (2000)). Except where otherwise indicated, we use the term “HMO” because that is the term used by the State and the parties; what we intend is simply to describe the structures covered by the Illinois Act.
In'light of our holding today that §4-10 is not preempted by ERISA, the propriety of this ruling is questionable; a suit to compel compliance with §4-10 in the context of an ERISA plan would seem to be akin to a suit to compel compliance with the terms of a plan under 29 U. S. C. § 1132(a)(3). Alternatively, the proper course may have been to bring a suit to recover benefits due, alleging that the denial was improper in the absence of compliance with § 4-10. We need not resolve today which of these options is more consonant with ERISA.
No party has challenged Rush’s status as defendant in this case, despite the fact that many lower courts have interpreted ERISA to permit suits under § 1132(a) only against ERISA plans, administrators, or fiduciaries. See, e. g., Everhart v. Allmerica Financial Life Ins. Co., 275 F. 3d 751, 754-756 (CA9 2001); Garren v. John Hancock Mut. Life Ins. Co., 114 F. 3d 186, 187 (CA11 1997); Jass v. Prudential Health Care Plan, Inc., 88 F. 3d 1482, 1490 (CA7 1996). Without commenting on the correctness of such holdings, we assume (although the information does not appear in the record) that Rush has failed to challenge its status as defendant because it is, in fact, the plan administrator. This conclusion is buttressed by the fact that the plan’s sponsor has granted Rush discretion to interpret the terms of its coverage, and by the fact that one of Rush’s challenges to the Illinois statute is based on what Rush perceives as the limits that statute places on fiduciary discretion. Whatever Rush’s true status may be, however, it is immaterial to our holding.
The McCarran-Ferguson Act requires that the business of insurance be subject to state regulation, and, subject to certain exceptions, mandates that “[n]o Act of Congress shall be construed to invalidate . . . any law enacted by any State for the purpose of regulating the business of insurance ... .” 15 U. S. C. § 1012(b).
We have, in a limited .number of cases, found certain contracts not to be part of the “business of insurance” under McGarran-Ferguson, notwithstanding their classification as such for the purpose of state regulation. See, e. g., SEC v. Variable Annuity Life Ins. Co. of America, 359 U. S. 65 (1959). Even then, however, we recognized that such classifications are relevant to the enquiry, because Congress, in leaving the “business of insurance” to the States, “was legislating concerning a concept which had taken on its coloration and meaning largely from state law, from state practice, from state usage.” Id., at 69.
ERISA’s “deemer” clause provides an exception to its saving clause that prohibits States from regulating self-funded plans as insurers. See 29 U. S. C. § 1144(b)(2)(B); FMC Corp. v. Holliday, 498 U. S. 52, 61 (1990). Therefore, Illinois’s Act would not be “saved” as an insurance law to the *372extent it applied to self-funded plans. This fact, however, does not bear on Rush’s challenge to the law as one that is targeted toward non-risk-bearing organizations.
Title 29 U. S. C. § 1132(a) provides in relevant part:
“A civil action may be brought—
“(1) by a participant or beneficiary—
“(A) for the relief provided for in subsection (c) of this section [concerning requests to the administrator for information], or
“(B) to recover benefits due to him under the terms of his plan, to enforce his rights under the terms of the plan, or to clarify his rights to future benefits under the terms of the plan;
“(2) by the Secretary, or by a participant, beneficiary or fiduciary for appropriate relief under section 1109 of this title [breach of fiduciary duty];
“(3) by a participant, beneficiary, or fiduciary (A) to enjoin any act or practice which violates any provision of this subchapter or the terms of the plan, or (B) to obtain other appropriate equitable relief (i) to redress such violations or (ii) to enforce any provisions of this subchapter or the terms of the plan;
“(4) by the Secretary, or by a participant, or beneficiary for appropriate relief in the case of a violation of 1025(c) of this title [information to be furnished to participants];
“(5) except as otherwise provided in subsection (b) of this section, by the Secretary (A) to enjoin any act or practice which violates any provision of this subchapter, or (B) to obtain other appropriate equitable relief (i) to redress such violation or (ii) to enforce any provision of this subchapter;
“(6) by the Secretary to collect any civil penalty under paragraph (2), (4), (5), or (6) of subsection (c) of this section or under subsection (i) or (i) of this section.”
Rush and its amici interpret Pilot Life to have gone a step further to hold that any law that presents such a conflict with federal goals is simply not a law that “regulates insurance,” however else the “insurance” test comes out. We believe the point is largely academic. As will be discussed further, even under Rush’s approach, a court must still determine whether the state law at issue does, in fact, create such a conflict. Thus, we believe that it is more logical to proceed as we have done here.
The parties do not dispute that § 4-10, as a matter of state law, purports to make the independent reviewer’s judgment dispositive as to what is “medically necessary.” We accept this interpretation of the meaning of the statute for the purposes of our opinion.
This is not to say that the court would have no role beyond ordering compliance with the reviewer’s determination. The court would have the responsibility, for example, to fashion appropriate relief, or to determine whether other aspects of the plan (beyond the “medical necessity” of a particular treatment) affect the relative rights of the parties. Rush, for example, has chosen to guarantee medically necessary services to plan participants. For that reason, to the extent § 4-10 may render the independent reviewer the final word on what is necessary, see n. 9, supra, Rush is obligated to provide the service. But insurance contracts do not have to contain such guarantees, and not all do. Some, for instance, guarantee medically necessary care, but then modify that obligation by excluding experimental procedures from coverage. See, e. g., Tillery v. Hoffman Enclosures, Inc., 280 F. 3d 1192 (CA8 2002). Obviously, §4-10 does not have anything to say about whether a proposed procedure is experimental. There is also the possibility, though we do not decide the issue today, that a reviewer’s judgment could be challenged as inaccurate or biased, just as the decision of a plan fiduciary might be so challenged.
Thus, we do not believe that the mere fact that state independent review laws are likely to entail different procedures will impose burdens on plan administration that would threaten the object of 29 U. S. C. § 1132(a); it is the HMO contracting with a plan, and not the plan itself, that will be subject to these regulations, and every HMO will have to establish procedures for conforming with the local laws, regardless of what this Court may think ERISA forbids. This means that there will be no special burden of compliance upon an ERISA plan beyond what the HMO has already provided for. And although the added compliance cost to the HMO may ultimately be passed on to the ERISA plan, we have said that such “indirect economic effectfs],” New York State Conference of Blue Cross & Blue Shield Plans v. Travelers Ins. Co., 514 U. S. 645, 659 (1995), are not enough to preempt state regulation even outside of the insurance context. We recognize, of course, that a State might enact an independent review requirement with procedures so elaborate, and burdens so onerous, that they might undermine § 1132(a). No such system is before us.
Nothing in the Act states that the reviewer should refer to the definitions of medical necessity contained in the contract, but the reviewer did, in this case, refer to that definition. Thus, we will assume that some degree of contract interpretation is required under the Act. Were, no interpretation required, there would be a real question as to whether § 4-10 is properly characterized as a species of mandated-benefit law of the type we approved in Metropolitan Life Ins. Co. v. Massachusetts, 471 U. S. 724 (1985).
See, e. g., Cal. Ins. Code Ann. § 10123.68 (West Supp. 2002); Ind. Code §27-13-37-5 (1999); N. J. Stat. Ann. § 17B:26-2.3 (1996); Ókla. Admin. Code §365:10-5-4 (1996); R. I. Gen. Laws §27-39-2 (1998).
See n. 10, supra.
An issue implicated by this ease but requiring no resolution is the degree to which a plan provision for unfettered discretion in benefit determinations guarantees truly deferential review. In Firestone Tire itself, we noted that review for abuse of discretion would home in on any conflict of interest on the plan fiduciary’s part, if a conflict was plausibly raised.
. That last observation was underscored only two Terms ago in Pegram v. Herdrich, 530 U. S. 211 (2000), when we again noted the potential for conflict when an HMO makes decisions about appropriate treatment, see id., at 219-220. It is a fair question just how deferential the review can be when the judicial eye is peeled for conflict of interest. Moreover, as we explained in Pegram, “it is at least questionable whether Congress would have had mixed eligibility decisions in mind when it provided that decisions administering a plan were fiduciary in nature.” Id., at 232. Our decision today does not require us to resolve these questions.
Rush presents the alternative argument that § 4-10 is preempted as conflicting with ERISA’s requirement that a benefit denial be reviewed by a named fiduciary, 29 U. S. C. §1133(2). Rush contends that §4-10 interferes with fiduciary discretion by forcing the provision of benefits over a fiduciary’s objection. Happily, we need not decide today whether § 1133(2) carries the same preemptive force of § 1132(a) such that it overrides even the express saving clause for insurance regulation, because we see no conflict. Section 1133 merely requires that plans provide internal appeals of benefit denials; §4-10 plays no role in this process, instead providing for extra review once the internal process is complete. Nor is there any conflict in the removal of fiduciary “discretion”; as described below, ERISA does not require that such decisions be discretionary, and insurance regulation is not preempted merely because it conflicts with substantive plan terms. See UNUM Life Ins. Co. of America v. Ward, 526 U. S. 358, 376 (1999) (“Under [Petitioner’s] interpretation ... insurers could displace any state regulation simply by inserting a contrary term in plan documents. This interpretation would virtually rea[d] the saving clause out of ERISA” (internal quotation marks and citations omitted)).
We do not mean to imply that States are free to create other forms of binding arbitration to provide de novo review of any terms of insurance contract^; as discussed above, our decision rests in part on our recognition that the disuniformity Congress hoped to avoid is not implicated by decisions that are so heavily imbued with expert medical judgments. Rather, we hold that the feature of §4-10 that provides a different standard of review with respect to mixed eligibility decisions from what would be available in court is not enough to create a conflict that undermines congressional policy in favor of uniformity of remedies.