Gonzales v. Oregon, 546 U.S. 243 (2006)

The Oregon Death With Dignity Act (ODWDA) exempts from civil or criminal liability state-licensed physicians who, in compliance with ODWDA’s specific safeguards, dispense or prescribe a lethal dose of drugs upon the request of a terminally ill patient. In 2001, the Attor ney General issued an Interpretive Rule to address the implementation and enforcement of the CSA with respect to ODWDA, declaring that using controlled substances to assist suicide is not a legitimate medical practice and that dispensing or prescribing them for this purpose is unlawful under the CSA. The State, a physician, a pharmacist, and [p. 244] some terminally ill state residents challenged the Rule. The District Court permanently enjoined its enforcement. The Ninth Circuit invali dated the Rule, reasoning that, by making a medical procedure author ized under Oregon law a federal offense, it altered the balance between the States and the Federal Government without the requisite clear statement that the CSA authorized the action; and in the alternative, that the Rule could not be squared with the CSA’s plain language, which targets only conventional drug abuse and excludes the Attorney General from medical policy decisions.

(b) The Interpretive Rule at issue is not entitled to Auer deference as an interpretation of 21 CFR § 1306.04. Unlike the underlying regu lations in Auer, which gave specificity to a statutory scheme the Sec retary of Labor was charged with enforcing and reflected the Labor Department’s considerable experience and expertise, the underlying regulation here does little more than restate the terms of the statute itself. The CSA allows prescription of drugs that have a “currently accepted medical use,” 21 U. S. C. § 812(b); requires a “medical purpose” for dispensing the least controlled substances of those on the schedules, § 829(c); and defines a “valid prescription” as one “issued for a legitimate medical purpose,” § 830(b)(3)(A)(ii). Similarly, physicians are consid ered practitioners if they dispense controlled substances “in the course of professional practice.” 21 U. S. C. § 802(21). The regulation just re peats two of these statutory phrases and attempts to summarize the others. An agency does not acquire special authority to interpret its own words when, instead of using its expertise and experience to formu late a regulation, it has elected merely to paraphrase the statutory lan guage. Furthermore, any statutory authority for the Interpretive Rule would have to come from 1984 CSA amendments adding the “public [p. 245] interest” requirement, but 21 CFR § 1306.04 was adopted in 1971. That the current interpretation runs counter to the intent at the time of the regulation’s promulgation is an additional reason why Auer deference is unwarranted. Pp. 256–258.

(c) The Interpretive Rule is also not entitled to Chevron deference. The statutory phrase “legitimate medical purpose” is ambiguous in the relevant sense. However, Chevron deference is not accorded merely because the statute is ambiguous and an administrative official is in volved. A rule must be promulgated pursuant to authority Congress has delegated to the official. The specific respects in which the Attor ney General is authorized to make rules under the CSA show that he is not authorized to make a rule declaring illegitimate a medical standard for patient care and treatment specifically authorized under state law. Congress delegated to the Attorney General only the authority to pro mulgate rules relating to “registration” and “control” of the dispensing of controlled substances, 21 U. S. C. § 821 (2000 ed., Supp. V), and “for the efficient execution of his [statutory] functions,” 21 U. S. C. § 871(b). Control means “to add a . . . substance . . . to a schedule,” § 802(5), following specified procedures. Because the Interpretive Rule does not concern scheduling of substances and was not issued under the required procedures, it cannot fall under the Attorney General’s control authority. Even if “control” were understood to signify something other than its statutory definition, it could not support the Interpretive Rule. Nor can the Interpretive Rule be justified under the CSA’s registration pro visions. It does not undertake the Act’s five-factor analysis for deter mining when registration is “inconsistent with the public interest,” § 823(f), and it deals with much more than registration. It purports to declare that using controlled substances for physician-assisted suicide is a crime, an authority going well beyond the Attorney General’s statu tory power to register or deregister physicians. It would be anomalous for Congress to have painstakingly described the Attorney General’s limited authority to deregister a single physician or schedule a single drug, but to have given him, just by implication, authority to declare an entire class of activity outside the course of professional practice and therefore a criminal violation of the CSA. It is not enough that “public interest,” “public health and safety,” and “Federal law” are used in the part of the Act over which the Attorney General has authority. Cf. Sutton v. United Air Lines, Inc., 527 U. S. 471. The first two terms do not call on the Attorney General, or any executive official, to make an independent assessment of the meaning of federal law. The Attorney General did not base the Interpretive Rule on an application of the five-factor test generally, or the “public health and safety” factor spe cifically. Even if he had, it is doubtful that he could cite those factors [p. 246] to deregister a physician simply because he deemed a controversial practice permitted by state law to have an illegitimate medical purpose. The federal-law factor requires the Attorney General to decide “[c]om pliance” with the law but does not suggest that he may decide what the law is. To say that he can define the substantive standards of medical practice as part of his authority would also put 21 U. S. C. § 871(b) in considerable tension with the narrowly defined control and registration delegation. It would go, moreover, against the plain language of the text to treat a delegation for the “execution” of his functions as a further delegation to define other functions well beyond the Act’s specific grants of authority. The authority desired by the Government is inconsistent with the Act’s design in other fundamental respects, e. g., the Attorney General must share power with, and in some respect defer to, the Secre tary, whose functions are likewise delineated and confined by the Act. Postenactment congressional commentary on the CSA’s regulation of medical practice is also at odds with the Attorney General’s claimed authority. The Government’s claim that the Attorney General’s deci sion is a legal, not medical, one does not suffice, for the Interpretive Rule places extensive reliance on medical judgments and views of the medical community in concluding that assisted suicide is not a legitimate medical purpose. The idea that Congress gave him such broad and un usual authority through an implicit delegation is not sustainable. The importance of the issue of physician-assisted suicide makes the oblique form of the claimed delegation all the more suspect. Pp. 258–269.

(d) The Attorney General’s opinion is unpersuasive under Skidmore. The CSA and this Court’s case law amply support the conclusion that Congress regulates medical practice insofar as it bars doctors from using their prescription-writing powers as a means to engage in illicit drug dealing and trafficking as conventionally understood. Beyond this, the Act manifests no intent to regulate the practice of medicine generally, which is understandable given federalism’s structure and limi tations. The CSA’s structure and operation presume and rely upon a functioning medical profession regulated under the States’ police pow ers. The Federal Government can set uniform standards for regulating health and safety. In connection with the CSA, however, the only pro vision in which Congress set general, uniform medical practice stand ards, 42 U. S. C. § 290bb–2a, strengthens the understanding of the CSA as a statute combating recreational drug abuse, and also indicates that when Congress wants to regulate medical practice in the given scheme, it does so by explicit statutory language. The difficulty in defending the Attorney General’s declaration that the CSA impliedly criminalizes physician-assisted suicide is compounded by the Act’s consistent delega tion of medical judgments to the Secretary and its otherwise careful [p. 247] allocation of powers for enforcing the CSA’s limited objectives. The Government’s contention that the terms “medical” or “medicine” refer to a healing or curative art, and thus cannot embrace the intentional hastening of a patient’s death, rests on a reading of 21 U. S. C. § 829(a)’s prescription requirement without the illumination of the rest of the stat ute. Viewed in context, that requirement is better understood as en suring that patients use controlled substances under a doctor’s super vision so as to prevent addiction and recreational abuse. To read prescriptions for assisted suicide as “drug abuse” under the CSA is dis cordant with the phrase’s consistent use throughout the Act, not to men tion its ordinary meaning. The Government’s interpretation of the pre scription requirement also fails under the objection that the Attorney General is an unlikely recipient of such broad authority, given the Secre tary’s primacy in shaping medical policy under the CSA and the Act’s otherwise careful allocation of decisionmaking powers. Pp. 269–275. 368 F. 3d 1118, affirmed.

Justice Kennedy delivered the opinion of the Court. The question before us is whether the Controlled Sub stances Act allows the United States Attorney General to [p. 249] prohibit doctors from prescribing regulated drugs for use in physician-assisted suicide, notwithstanding a state law per mitting the procedure. As the Court has observed, “Ameri cans are engaged in an earnest and profound debate about the morality, legality, and practicality of physician-assisted suicide.” Washington v. Glucksberg, 521 U. S. 702, 735 (1997). The dispute before us is in part a product of this political and moral debate, but its resolution requires an in quiry familiar to the courts: interpreting a federal statute to determine whether executive action is authorized by, or otherwise consistent with, the enactment.

To prevent diversion of controlled substances with medical uses, the CSA regulates the activity of physicians. To issue [p. 251] lawful prescriptions of Schedule II drugs, physicians must “obtain from the Attorney General a registration issued in accordance with the rules and regulations promulgated by him.” 21 U. S. C. § 822(a)(2). The Attorney General may deny, suspend, or revoke this registration if, as relevant here, the physician’s registration would be “inconsistent with the public interest.” § 824(a)(4); § 822(a)(2). When deciding whether a practitioner’s registration is in the public interest, the Attorney General “shall” consider:

Oregon voters enacted ODWDA in 1994. For Oregon residents to be eligible to request a prescription under [p. 252] ODWDA, they must receive a diagnosis from their attending physician that they have an incurable and irreversible dis ease that, within reasonable medical judgment, will cause death within six months. Ore. Rev. Stat. §§ 127.815, 127.800(12) (2003). Attending physicians must also deter mine whether a patient has made a voluntary request, ensure a patient’s choice is informed, and refer patients to counsel ing if they might be suffering from a psychological disorder or depression causing impaired judgment. §§ 127.815, 127.825. A second “consulting” physician must examine the patient and the medical record and confirm the attending physician’s conclusions. § 127.800(8). Oregon physicians may dispense or issue a prescription for the requested drug, but may not administer it. §§ 127.815(1)(L), 127.880.

In 1997, Members of Congress concerned about ODWDA invited the DEA to prosecute or revoke the CSA registra tion of Oregon physicians who assist suicide. They con tended that hastening a patient’s death is not legitimate med ical practice, so prescribing controlled substances for that purpose violates the CSA. Letter from Sen. Orrin Hatch and Rep. Henry Hyde to Thomas A. Constantine (July 25, 1997), reprinted in Hearing on S. 2151 before the Senate Committee on the Judiciary, 105th Cong., 2d Sess., 2–3 (1999) [p. 253] (hereinafter Hearing). The letter received an initial, favor able response from the director of the DEA, see Letter from Thomas A. Constantine to Sen. Orrin Hatch (Nov. 5, 1997), Hearing 4–5, but Attorney General Reno considered the matter and concluded that the DEA could not take the pro posed action because the CSA did not authorize it to “dis place the states as the primary regulators of the medical profession, or to override a state’s determination as to what constitutes legitimate medical practice,” Letter from Attor ney General Janet Reno to Sen. Orrin Hatch, on Oregon’s Death with Dignity Act (June 5, 1998), Hearing 5–6. Legis lation was then introduced to grant the explicit authority Attorney General Reno found lacking; but it failed to pass. See H. R. 4006, 105th Cong., 2d Sess. (1998); H. R. 2260, 106th Cong., 1st Sess. (1999).

On November 9, 2001, without consulting Oregon or appar ently anyone outside his Department, the Attorney General [p. 254] issued an Interpretive Rule announcing his intent to restrict the use of controlled substances for physician-assisted sui cide. Incorporating the legal analysis of a memorandum he had solicited from his Office of Legal Counsel, the Attorney General ruled:

Executive actors often must interpret the enactments Congress has charged them with enforcing and implement ing. The parties before us are in sharp disagreement both as to the degree of deference we must accord the Interpre tive Rule’s substantive conclusions and whether the Rule is authorized by the statutory text at all. Although balancing the necessary respect for an agency’s knowledge, expertise, and constitutional office with the courts’ role as interpreter of laws can be a delicate matter, familiar principles guide us. An administrative rule may receive substantial deference if it interprets the issuing agency’s own ambiguous regulation. Auer v. Robbins, 519 U. S. 452, 461–463 (1997). An interpre tation of an ambiguous statute may also receive substantial deference. Chevron U. S. A. Inc. v. Natural Resources De fense Council, Inc., 467 U. S. 837, 842–845 (1984). Defer ence in accordance with Chevron, however, is warranted only “when it appears that Congress delegated authority to the agency generally to make rules carrying the force of law, [p. 256] and that the agency interpretation claiming deference was promulgated in the exercise of that authority.” United States v. Mead Corp., 533 U. S. 218, 226–227 (2001). Other wise, the interpretation is “entitled to respect” only to the extent it has the “power to persuade.” Skidmore v. Swift & Co., 323 U. S. 134, 140 (1944).

In Auer, the underlying regulations gave specificity to a statutory scheme the Secretary of Labor was charged with enforcing and reflected the considerable experience and ex pertise the Department of Labor had acquired over time with respect to the complexities of the Fair Labor Standards [p. 257] Act. Here, on the other hand, the underlying regulation does little more than restate the terms of the statute itself. The language the Interpretive Rule addresses comes from Congress, not the Attorney General, and the near equiva lence of the statute and regulation belies the Government’s argument for Auer deference.

Furthermore, as explained below, if there is statutory au thority to issue the Interpretive Rule it comes from the 1984 amendments to the CSA that gave the Attorney General au [p. 258] thority to register and deregister physicians based on the public interest. The regulation was enacted before those amendments, so the Interpretive Rule cannot be justified as indicative of some intent the Attorney General had in 1971. That the current interpretation runs counter to the “intent at the time of the regulation’s promulgation” is an additional reason why Auer deference is unwarranted. Thomas Jef ferson Univ. v. Shalala, 512 U. S. 504, 512 (1994) (internal quotation marks omitted). Deference under Auer being in appropriate, we turn to the question whether the Interpre tive Rule, on its own terms, is a permissible interpretation of the CSA.

The starting point for this inquiry is, of course, the lan guage of the delegation provision itself. In many cases au thority is clear because the statute gives an agency broad power to enforce all provisions of the statute. See, e. g., Na [p. 259] tional Cable & Telecommunications Assn. v. Brand X In ternet Services, 545 U. S. 967, 980 (2005) (explaining that a Federal Communications Commission regulation received Chevron deference because “Congress has delegated to the Commission the authority to . . . ‘prescribe such rules and regulations as may be necessary in the public interest to carry out the provisions’ of the Act” (quoting 47 U. S. C. § 201(b))); Household Credit Services, Inc. v. Pfennig, 541 U. S. 232, 238 (2004) (giving Chevron deference to a Federal Reserve Board regulation where “Congress has expressly delegated to the Board the authority to prescribe reg ulations . . . as, in the judgment of the Board, ‘are neces sary or proper to effectuate the purposes of ’ ” the statute (quoting 15 U. S. C. § 1604(a))). The CSA does not grant the Attorney General this broad authority to promulgate rules.

Turning first to the Attorney General’s authority to make regulations for the “control” of drugs, this delegation cannot sustain the Interpretive Rule’s attempt to define standards of medical practice. Control is a term of art in the CSA. [p. 260] “As used in this subchapter,” § 802––the subchapter that in cludes § 821––

Even if “control” in § 821 were understood to signify some thing other than its statutory definition, it would not support the Interpretive Rule. The statutory references to “con trol” outside the scheduling context make clear that the At torney General can establish controls “against diversion,” e. g., § 823(a)(1), but do not give him authority to define diver sion based on his view of legitimate medical practice. As explained below, the CSA’s express limitations on the Attor ney General’s authority, and other indications from the statu tory scheme, belie any notion that the Attorney General has been granted this implicit authority. Indeed, if “control” were given the expansive meaning required to sustain the Interpretive Rule, it would transform the carefully described [p. 261] limits on the Attorney General’s authority over registration and scheduling into mere suggestions.

By this logic, however, the Attorney General claims ex traordinary authority. If the Attorney General’s argument were correct, his power to deregister necessarily would in clude the greater power to criminalize even the actions of registered physicians, whenever they engage in conduct he deems illegitimate. This power to criminalize—unlike his power over registration, which must be exercised only after considering five express statutory factors—would be unre strained. It would be anomalous for Congress to have so painstakingly described the Attorney General’s limited au thority to deregister a single physician or schedule a single drug, but to have given him, just by implication, authority to declare an entire class of activity outside “the course of professional practice,” and therefore a criminal violation of the CSA. See Federal Maritime Comm’n v. Seatrain Lines, Inc., 411 U. S. 726, 744 (1973) (“In light of these specific [p. 263] grants of . . . authority, we are unwilling to construe the ambiguous provisions . . . to serve this purpose [of creating further authority]—a purpose for which it obviously was not intended”).

The same principle controls here. It is not enough that the terms “public interest,” “public health and safety,” and “Federal law” are used in the part of the statute over which the Attorney General has authority. The statutory terms “public interest” and “public health” do not call on the Attor ney General, or any other executive official, to make an in dependent assessment of the meaning of federal law. The Attorney General did not base the Interpretive Rule on an [p. 264] application of the five-factor test generally, or the “public health and safety” factor specifically. Even if he had, it is doubtful the Attorney General could cite the “public inter est” or “public health” to deregister a physician simply be cause he deemed a controversial practice permitted by state law to have an illegitimate medical purpose.

The limits on the Attorney General’s authority to define medical standards for the care and treatment of patients bear also on the proper interpretation of § 871(b). This sec tion allows the Attorney General to best determine how to execute “his functions.” It is quite a different matter, how ever, to say that the Attorney General can define the sub stantive standards of medical practice as part of his author ity. To find a delegation of this extent in § 871 would put that part of the statute in considerable tension with the nar rowly defined delegation concerning control and registration. It would go, moreover, against the plain language of the text to treat a delegation for the “execution” of his functions as a further delegation to define other functions well beyond [p. 265] the statute’s specific grants of authority. When Congress chooses to delegate a power of this extent, it does so not by referring back to the administrator’s functions but by giving authority over the provisions of the statute he is to interpret. See, e. g., National Cable & Telecommunications Assn., 545 U. S. 967; Household Credit Services, 541 U. S. 232.

In a similar vein the 1970 Act’s regulation of medical prac tice with respect to drug rehabilitation gives the Attorney General a limited role; for it is the Secretary who, after con sultation with the Attorney General and national medical groups, “determine[s] the appropriate methods of profes sional practice in the medical treatment of . . . narcotic addic tion.” 42 U. S. C. § 290bb–2a; see 21 U. S. C. § 823(g) (2000 ed. and Supp. II) (stating that the Attorney General shall register practitioners who dispense drugs for narcotics treat ment when the Secretary has determined the applicant is qualified to treat addicts and the Attorney General has con [p. 266] cluded the applicant will comply with recordkeeping and se curity regulations); Moore, 423 U. S., at 144 (noting that in enacting the addiction-treatment provisions, Congress sought to change the fact “that ‘criminal prosecutions’ in the past had turned on the opinions of federal prosecutors”); H. R. Rep. No. 93–884, p. 6 (1974) (“This section preserves the distinctions found in the [CSA] between the functions of the Attorney General and the Secretary . . . . All decisions of a medical nature are to be made by the Secretary . . . . Law enforcement decisions respecting the security of stocks of narcotic drugs and the maintenance of records on such drugs are to be made by the Attorney General”).

The structure of the CSA, then, conveys unwillingness to cede medical judgments to an executive official who lacks medical expertise. In interpreting statutes that divide authority, the Court has recognized: “Because historical familiarity and policymaking expertise account in the first instance for the presumption that Congress delegates in terpretive lawmaking power to the agency rather than to the reviewing court, we presume here that Congress in tended to invest interpretive power in the administrative actor in the best position to develop these attributes.” Mar [p. 267] tin v. Occupational Safety and Health Review Comm’n, 499 U. S. 144, 153 (1991) (citations omitted). This presumption works against a conclusion that the Attorney General has authority to make quintessentially medical judgments.

The importance of the issue of physician-assisted suicide, which has been the subject of an “earnest and profound de bate” across the country, Glucksberg, 521 U. S., at 735, makes the oblique form of the claimed delegation all the more sus [p. 268] pect. Under the Government’s theory, moreover, the med ical judgments the Attorney General could make are not limited to physician-assisted suicide. Were this argument accepted, he could decide whether any particular drug may be used for any particular purpose, or indeed whether a phy sician who administers any controversial treatment could be deregistered. This would occur, under the Government’s view, despite the statute’s express limitation of the Attorney General’s authority to registration and control, with attend ant restrictions on each of those functions, and despite the statutory purposes to combat drug abuse and prevent illicit drug trafficking.

If, in the course of exercising his authority, the Attorney General uses his analysis in the Interpretive Rule only for guidance in deciding when to prosecute or deregister, then the question remains whether his substantive interpretation is correct. Since the Interpretive Rule was not promul gated pursuant to the Attorney General’s authority, its inter pretation of “legitimate medical purpose” does not receive Chevron deference. Instead, it receives deference only in accordance with Skidmore. “The weight of such a judgment in a particular case will depend upon the thoroughness evi dent in its consideration, the validity of its reasoning, its con sistency with earlier and later pronouncements, and all those factors which give it power to persuade, if lacking power to control.” 323 U. S., at 140; see also Mead, 533 U. S., at 235 (noting that an opinion receiving Skidmore deference may “claim the merit of its writer’s thoroughness, logic, and ex pertness, its fit with prior interpretations, and any other [p. 269] sources of weight”). The deference here is tempered by the Attorney General’s lack of expertise in this area and the ap parent absence of any consultation with anyone outside the Department of Justice who might aid in a reasoned judg ment. In any event, under Skidmore, we follow an agency’s rule only to the extent it is persuasive, see Christensen v. Harris County, 529 U. S. 576, 587 (2000); and for the reasons given and for further reasons set out below, we do not find the Attorney General’s opinion persuasive.

In deciding whether the CSA can be read as prohibiting physician-assisted suicide, we look to the statute’s text and design. The statute and our case law amply support the [p. 270] conclusion that Congress regulates medical practice insofar as it bars doctors from using their prescription-writing pow ers as a means to engage in illicit drug dealing and traffick ing as conventionally understood. Beyond this, however, the statute manifests no intent to regulate the practice of medicine generally. The silence is understandable given the structure and limitations of federalism, which allow the States “ ‘great latitude under their police powers to legislate as to the protection of the lives, limbs, health, comfort, and quiet of all persons.’ ” Medtronic, Inc. v. Lohr, 518 U. S. 470, 475 (1996) (quoting Metropolitan Life Ins. Co. v. Massa chusetts, 471 U. S. 724, 756 (1985)).

The structure and operation of the CSA presume and rely upon a functioning medical profession regulated under the States’ police powers. The Attorney General can register a physician to dispense controlled substances “if the applicant is authorized to dispense . . . controlled substances under the laws of the State in which he practices.” 21 U. S. C. § 823(f). When considering whether to revoke a physician’s regis tration, the Attorney General looks not just to violations of federal drug laws; but he “shall” also consider “[t]he recom mendation of the appropriate State licensing board or pro fessional disciplinary authority” and the registrant’s com pliance with state and local drug laws. Ibid. The very definition of a “practitioner” eligible to prescribe includes physicians “licensed, registered, or otherwise permitted, by the United States or the jurisdiction in which he practices” to dispense controlled substances. § 802(21). Further cau tioning against the conclusion that the CSA effectively dis places the States’ general regulation of medical practice is the Act’s pre-emption provision, which indicates that, absent a positive conflict, none of the Act’s provisions should be “construed as indicating an intent on the part of the Con gress to occupy the field in which that provision operates . . . to the exclusion of any State law on the same subject mat [p. 271] ter which would otherwise be within the authority of the State.” § 903.

On its own, this understanding of medicine’s boundaries is at least reasonable. The primary problem with the Govern ment’s argument, however, is its assumption that the CSA [p. 273] impliedly authorizes an executive officer to bar a use simply because it may be inconsistent with one reasonable under standing of medical practice. Viewed alone, the prescription requirement may support such an understanding, but stat utes “should not be read as a series of unrelated and isolated provisions.” Gustafson v. Alloyd Co., 513 U. S. 561, 570 (1995). The CSA’s substantive provisions and their arrange ment undermine this assertion of an expansive federal au thority to regulate medicine.

The statutory scheme with which the CSA is intertwined further confirms a more limited understanding of the pre scription requirement. When the Secretary considers Food and Drug Administration approval of a substance with “stim ulant, depressant, or hallucinogenic effect,” he must forward the information to the Attorney General for possible schedul ing. Shedding light on Congress’ understanding of drug abuse, this requirement appears under the heading “Abuse [p. 274] potential.” 21 U. S. C. § 811(f). Similarly, when Congress prepared to implement the Convention on Psychotropic Sub stances, it did so through the CSA. § 801a.

The Government’s interpretation of the prescription re quirement also fails under the objection that the Attorney General is an unlikely recipient of such broad authority, given the Secretary’s primacy in shaping medical policy under the CSA, and the statute’s otherwise careful allocation of decisionmaking powers. Just as the conventions of ex pression indicate that Congress is unlikely to alter a statute’s obvious scope and division of authority through muffled hints, the background principles of our federal system also belie the notion that Congress would use such an obscure grant of authority to regulate areas traditionally supervised by the States’ police power. It is unnecessary even to con sider the application of clear statement requirements, see, e. g., United States v. Bass, 404 U. S. 336, 349 (1971); cf. BFP v. Resolution Trust Corporation, 511 U. S. 531, 544–546 (1994), or presumptions against pre-emption, see, e. g., Rush Prudential HMO, Inc. v. Moran, 536 U. S. 355, 387 (2002), to reach this commonsense conclusion. For all these reasons, we conclude the CSA’s prescription requirement does not au [p. 275] thorize the Attorney General to bar dispensing controlled substances for assisted suicide in the face of a state medical regime permitting such conduct.

Contrary to the Court’s analysis, this case involves not one but three independently sufficient grounds for reversing the Ninth Circuit’s judgment. First, the Attorney General’s in terpretation of “legitimate medical purpose” in 21 CFR § 1306.04 (2005) (hereinafter Regulation) is clearly valid, given the substantial deference we must accord it under Auer v. Robbins, 519 U. S. 452, 461 (1997), and his two re maining conclusions follow naturally from this interpreta tion. See Part I, infra. Second, even if this interpretation of the Regulation is entitled to lesser deference or no defer [p. 276] ence at all, it is by far the most natural interpretation of the Regulation—whose validity is not challenged here. This in terpretation is thus correct even upon de novo review. See Part II, infra. Third, even if that interpretation of the Reg ulation were incorrect, the Attorney General’s independent interpretation of the statutory phrase “public interest” in 21 U. S. C. §§ 824(a) and 823(f), and his implicit interpretation of the statutory phrase “public health and safety” in § 823(f)(5), are entitled to deference under Chevron U. S. A. Inc. v. Nat ural Resources Defense Council, Inc., 467 U. S. 837 (1984), and they are valid under Chevron. See Part III, infra. For these reasons, I respectfully dissent.

The Directive thus purports to do three distinct things: (1) to interpret the phrase “legitimate medical purpose” in the Regulation to exclude physician-assisted suicide; (2) to deter mine that prescribing, dispensing, and administering feder ally controlled substances to assist suicide violates the CSA; and (3) to determine that participating in physician-assisted suicide may render a practitioner’s registration “inconsistent with the public interest” within the meaning of 21 U. S. C. §§ 823(f) and 824(a)(4) (which incorporates § 823(f) by refer [p. 277] ence). The Court’s analysis suffers from an unremitting fail ure to distinguish among these distinct propositions in the Directive.

To begin with, it is doubtful that any such exception to the Auer rule exists. The Court cites no authority for it, be cause there is none. To the contrary, our unanimous deci sion in Auer makes clear that broadly drawn regulations are entitled to no less respect than narrow ones. “A rule requir [p. 278] ing the Secretary to construe his own regulations narrowly would make little sense, since he is free to write the regula tions as broadly as he wishes, subject only to the limits im posed by the statute.” 519 U. S., at 463 (emphasis added).

As used in this section, “prescription” is susceptible of at least three reasonable interpretations. First, it might mean any oral or written direction of a practitioner for the dispen sation of drugs. See United States v. Moore, 423 U. S. 122, 137, n. 13 (1975) (“On its face § 829 addresses only the form that a prescription must take. . . . [Section] 829 by its terms does not limit the authority of a practitioner”). Second, in light of the requirement of a “medical purpose” for the dis pensation of Schedule V substances, see § 829(c), it might mean a practitioner’s oral or written direction for the dispen sation of drugs that the practitioner believes to be for a legit imate medical purpose. See Webster’s New International Dictionary 1954 (2d ed. 1950) (hereinafter Webster’s Second) (defining “prescription” as “[a] written direction for the prep aration and use of a medicine”); id., at 1527 (defining “med icine” as “[a]ny substance or preparation used in treating disease”) (emphasis added). Finally, “prescription” might [p. 279] refer to a practitioner’s direction for the dispensation of drugs that serves an objectively legitimate medical purpose, regardless of the practitioner’s subjective judgment about the legitimacy of the anticipated use. See ibid.

The Court points out that the Regulation adopts some of the phrasing employed in unrelated sections of the statute. See ante, at 257. This is irrelevant. A regulation that sig nificantly clarifies the meaning of an otherwise ambiguous statutory provision is not a “parroting” regulation, regard less of the sources that the agency draws upon for the clari fication. Moreover, most of the statutory phrases that the Court cites as appearing in the Regulation, see ibid. (citing 21 U. S. C. §§ 812(b) (“ ‘currently accepted medical use’ ”), 829(c) (“ ‘medical purpose’ ”), 802(21) (“ ‘in the course of pro fessional practice’ ”)), are inapposite because they do not “parrot” the only phrase in the Regulation that the Directive purported to construe. See 66 Fed. Reg. 56608 (“I hereby [p. 280] determine that assisting suicide is not a ‘legitimate medical purpose’ within the meaning of 21 CFR § 1306.04 . . . ”). None of them includes the key word “legitimate,” which gives the most direct support to the Directive’s theory that § 829(c) presupposes a uniform federal standard of medical practice.2

Since the Regulation does not run afowl (so to speak) of the Court’s newly invented prohibition of “parroting”; and since the Directive represents the agency’s own interpreta tion of that concededly valid regulation; the only question remaining is whether that interpretation is “plainly errone ous or inconsistent with the regulation”; otherwise, it is “con trolling.” Auer, 519 U. S., at 461 (internal quotation marks omitted). This is not a difficult question. The Directive is assuredly valid insofar as it interprets “prescription” to re quire a medical purpose that is “legitimate” as a matter of federal law—since that is an interpretation of “prescription” that we ourselves have adopted. Webb v. United States, 249 U. S. 96 (1919), was a prosecution under the Harrison Act of a doctor who wrote prescriptions of morphine “for the pur pose of providing the user with morphine sufficient to keep him comfortable by maintaining his customary use,” id., [p. 281] at 99 (internal quotation marks omitted). The dispositive issue in the case was whether such authorizations were “pre scriptions” within the meaning of § 2(b) of the Harrison Act, predecessor to the CSA. Ibid. We held that “to call such an order for the use of morphine a physician’s prescription would be so plain a perversion of meaning that no discussion of the subject is required.” Id., at 99–100. Like the Direc tive, this interprets “prescription” to require medical pur pose that is legitimate as a matter of federal law. And the Directive is also assuredly valid insofar as it interprets “le gitimate medical purpose” as a matter of federal law to ex clude physician-assisted suicide, because that is not only a permissible but indeed the most natural interpretation of that phrase. See Part II, infra.

Setting aside the implicit delegation inherent in Congress’s use of the undefined term “prescription” in § 829, the Court’s [p. 282] reading of “control” in § 821 is manifestly erroneous. The Court urges, ante, at 260, that “control” is a term defined in part A of the subchapter (entitled “Introductory Provi sions”) to mean “to add a drug or other substance . . . to a schedule under part B of this subchapter,” 21 U. S. C. § 802(5) (emphasis added). But § 821 is not included in “part B of this subchapter,” which is entitled “Authority to Con trol; Standards and Schedules,” and consists of the sections related to scheduling, 21 U. S. C. §§ 811–814 (2000 ed. and Supp. V), where the statutory definition is uniquely appro priate. Rather, § 821 is found in part C of the subchapter, §§ 821–830, entitled “Registration of Manufacturers, Distrib utors, and Dispensers of Controlled Substances,” which in cludes all and only the provisions relating to the “manufac ture, distribution, and dispensing of controlled substances,” § 821. The artificial definition of “control” in § 802(5) has no conceivable application to the use of that word in § 821. Under that definition, “control” must take a substance as its direct object, see 21 U. S. C. § 802(5) (“to add a drug or other substance . . . to a schedule”)—and that is how “control” is consistently used throughout part B. See, e. g., §§ 811(b) (“proceedings . . . to control a drug or other substance”), 811(c) (“each drug or other substance proposed to be con trolled or removed from the schedules”), 811(d)(1) (“If con trol is required . . . the Attorney General shall issue an order controlling such drug . . . ”), 812(b) (“Except where control is required . . . a drug or other substance may not be placed in any schedule . . . ”). In § 821, by contrast, the term “con trol” has as its object, not “a drug or other substance,” but rather the processes of “manufacture, distribution, and dis pensing of controlled substances.” It could not be clearer that the artificial definition of “control” in § 802(5) is inappli cable. It makes no sense to speak of “adding the manufac turing, distribution, and dispensing of substances to a sched ule.” We do not force term-of-art definitions into contexts where they plainly do not fit and produce nonsense. What [p. 283] is obviously intended in § 821 is the ordinary meaning of “control”—namely, “[t]o exercise restraining or directing in fluence over; to dominate; regulate; hence, to hold from ac tion; to curb,” Webster’s Second 580. “Control” is regularly used in this ordinary sense elsewhere in part C of the sub chapter. See, e. g., 21 U. S. C. §§ 823(a)(1), (b)(1), (d)(1), (e)(1), (h)(1) (“maintenance of effective controls against di version”); §§ 823(a)(5), (d)(5) (“establishment of effective control against diversion”); § 823(g)(2)(H)(i) (“to exercise supervision or control over the practice of medicine”); § 830(b)(1)(C) (“a listed chemical under the control of the reg ulated person”); § 830(c)(2)(D) (“chemical control laws”) (em phasis added).

When the word is given its ordinary meaning, the Attor ney General’s interpretation of the prescription requirement of § 829 plainly “relat[es] to the . . . control of the . . . dispens ing of controlled substances,” 21 U. S. C. § 821 (2000 ed., Supp. V) (emphasis added), since a prescription is the chief requirement for “dispensing” such drugs, see § 829. The same meaning is compelled by the fact that § 821 is the first section not of part B of the subchapter, which deals entirely with “control” in the artificial sense, but of part C, every section of which relates to the “registration and control of the manufacture, distribution, and dispensing of controlled substances,” § 821. See §§ 822 (persons required to regis ter), 823 (registration requirements), 824 (denial, revocation, or suspension of registration), 825 (labeling and packaging), 826 (production quotas for controlled substances), 827 (rec ordkeeping and reporting requirements of registrants), 828 (order forms), 829 (prescription requirements), 830 (regula tion of listed chemicals and certain machines). It would be peculiar for the first section of this part to authorize rule making for matters covered by the previous part. The only sensible interpretation of § 821 is that it gives the Attorney General interpretive authority over the provisions of part C, all of which “relat[e] to the registration and control of the [p. 284] manufacture, distribution, and dispensing of controlled sub stances.” These provisions include both the prescription re quirement of § 829, and the criteria for registration and deregistration of §§ 823 and 824 (as relevant below, see Part III, infra).3

Even if the Directive were entitled to no deference what ever, the most reasonable interpretation of the Regulation and of the statute would produce the same result. Virtually every relevant source of authoritative meaning confirms that the phrase “legitimate medical purpose”4 does not include intentionally assisting suicide. “Medicine” refers to “[t]he science and art dealing with the prevention, cure, or allevia tion of disease.” Webster’s Second 1527. The use of the word “legitimate” connotes an objective standard of “medi cine,” and our presumption that the CSA creates a uniform federal law regulating the dispensation of controlled sub stances, see Mississippi Band of Choctaw Indians v. Holy [p. 286] field, 490 U. S. 30, 43 (1989), means that this objective stand ard must be a federal one. As recounted in detail in the memorandum for the Attorney General that is attached as an appendix to the Directive (OLC Memo), virtually every medical authority from Hippocrates to the current American Medical Association (AMA) confirms that assisting suicide has seldom or never been viewed as a form of “prevention, cure, or alleviation of disease,” and (even more so) that as sisting suicide is not a “legitimate” branch of that “science and art.” See OLC Memo, App. to Pet. for Cert. 113a–130a. Indeed, the AMA has determined that “ ‘[p]hysician-assisted suicide is fundamentally incompatible with the physician’s role as healer.’ ” Washington v. Glucksberg, 521 U. S. 702, 731 (1997). “[T]he overwhelming weight of authority in ju dicial decisions, the past and present policies of nearly all of the States and of the Federal Government, and the clear, firm and unequivocal views of the leading associations within the American medical and nursing professions, establish that assisting in suicide . . . is not a legitimate medical purpose.” OLC Memo, supra, at 129a. See also Glucksberg, supra, at 710, n. 8 (prohibitions or condemnations of assisted suicide in 50 jurisdictions, including 47 States, the District of Colum bia, and 2 Territories).

In the face of this “overwhelming weight of authority,” the Court’s admission that “[o]n its own, this understanding of medicine’s boundaries is at least reasonable,” ante, at 272 (emphasis added), tests the limits of understatement. The only explanation for such a distortion is that the Court con fuses the normative inquiry of what the boundaries of medi cine should be—which it is laudably hesitant to undertake—with the objective inquiry of what the accepted definition of “medicine” is. The same confusion is reflected in the Court’s remarkable statement that “[t]he primary problem with the Government’s argument . . . is its assumption that the CSA impliedly authorizes an executive officer to bar a use simply [p. 287] because it may be inconsistent with one reasonable under standing of medical practice.” Ante, at 272–273 (emphasis added). The fact that many in Oregon believe that the boundaries of “legitimate medicine” should be extended to include assisted suicide does not change the fact that the overwhelming weight of authority (including the 47 States that condemn physician-assisted suicide) confirms that they have not yet been so extended. Not even those of our Eighth Amendment cases most generous in discerning an “evolution” of national standards would have found, on this record, that the concept of “legitimate medicine” has evolved so far. See Roper v. Simmons, 543 U. S. 551, 564–567 (2005).

The Court contends that the phrase “legitimate medical purpose” cannot be read to establish a broad, uniform fed eral standard for the medically proper use of controlled sub stances. Ante, at 268. But it also rejects the most plausi ble alternative proposition, urged by the State, that any use authorized under state law constitutes a “legitimate medical purpose.” (The Court is perhaps leery of embracing this po sition because the State candidly admitted at oral argument that, on its view, a State could exempt from the CSA’s cover age the use of morphine to achieve euphoria.) Instead, the Court reverse-engineers an approach somewhere between a uniform national standard and a state-by-state approach, holding (with no basis in the CSA’s text) that “legitimate medical purpose” refers to all uses of drugs unrelated to “addiction and recreational abuse.” Ante, at 274. Thus, though the Court pays lipservice to state autonomy, see ante, at 269–271, its standard for “legitimate medical purpose” is in fact a hazily defined federal standard based on its pur posive reading of the CSA, and extracted from obliquely rel evant sections of the Act. In particular, relying on its observation that the criteria for scheduling controlled sub stances are primarily concerned with “addiction or abnormal effects on the nervous system,” ante, at 273 (citing 21 [p. 288] U. S. C. §§ 811(c)(7), 812(b), 811(f), 801a), the Court concludes that the CSA’s prescription requirement must be interpreted in light of this narrow view of the statute’s purpose.

The scheduling provisions of the CSA on which the Court relies confirm that the CSA’s “design,” ante, at 269, is not as narrow as the Court asserts. In making scheduling deter minations, the Attorney General must not only consider a drug’s “psychic or physiological dependence liability” as the Court points out, ante, at 273 (citing 21 U. S. C. § 811(c)(7)), but must also consider such broad factors as “[t]he state of current scientific knowledge regarding the drug or other sub stance,” § 811(c)(3), and (most notably) “[w]hat, if any, risk there is to the public health,” § 811(c)(6). If the latter factor were limited to addiction-related health risks, as the Court supposes, it would be redundant of § 811(c)(7). Moreover, in making registration determinations regarding manufactur ers and distributors, the Attorney General “shall” consider “such other factors as may be relevant to and consistent with the public health and safety,” §§ 823(a)(6), (b)(5), (d)(6), (e)(5) (emphasis added)—over and above the risk of “diversion” of controlled substances, §§ 823(a)(1), (a)(5), (b)(1), (d)(1), (d)(5), (e)(1). And, most relevant of all, in registering and deregis tering physicians, the Attorney General “may deny an appli [p. 289] cation for such registration if he determines that the issu ance of such registration would be inconsistent with the pub lic interest,” § 823(f); see also § 824(a)(4), and in making that determination “shall” consider “[s]uch other conduct which may threaten the public health and safety,” § 823(f)(5). All of these provisions, not just those selectively cited by the Court, shed light upon the CSA’s repeated references to the undefined term “abuse.” See §§ 811(a)(1)(A), (c)(1), (c)(4), (c)(5); §§ 812(b)(1)(A), (b)(2)(A), (b)(3)(A), (b)(4)(A), (b)(5)(A).

Although, as I have described, the Court’s opinion no more defers to state law than does the Directive, the Court relies on two provisions for the conclusion that “[t]he structure and operation of the CSA presume and rely upon a functioning medical profession regulated under the States’ police pow ers,” ante, at 270—namely, the registration provisions of § 823(f) and the nonpre-emption provision of § 903. Reli ance on the former is particularly unfortunate, because the Court’s own analysis recounts how Congress amended § 823(f) in 1984 in order to liberate the Attorney General’s power over registration from the control of state regulators. See ante, at 261; 21 U. S. C. § 823(f); see also Brief for Peti tioners 34–35. And the nonpre-emption clause is embar rassingly inapplicable, since it merely disclaims field pre emption, and affirmatively prescribes federal pre-emption [p. 290] whenever state law creates a conflict.5 In any event, the Directive does not purport to pre-empt state law in any way, not even by conflict pre-emption—unless the Court is under the misimpression that some States require assisted suicide. The Directive merely interprets the CSA to prohibit, like countless other federal criminal provisions, conduct that hap pens not to be forbidden under state law (or at least the law of the State of Oregon).

With regard to the CSA’s registration provisions, 21 U. S. C. §§ 823(f), 824(a), the Court argues that the statute cannot fairly be read to “ ‘hide elephants in mouseholes’ ” by delegating to the Attorney General the power to determine the legitimacy of medical practices in “ ‘vague terms or ancil lary provisions.’ ” Ante, at 267 (quoting Whitman v. Ameri can Trucking Assns., Inc., 531 U. S. 457, 468 (2001)). This case bears not the remotest resemblance to Whitman, which held that “Congress . . . does not alter the fundamental de tails of a regulatory scheme in vague terms or ancillary pro visions.” Ibid. (emphasis added). The Attorney General’s power to issue regulations against questionable uses of con trolled substances in no way alters “the fundamental details” of the CSA. I am aware of only four areas in which the Department of Justice has exercised that power to regulate uses of controlled substances unrelated to “addiction and recreational abuse” as the Court apparently understands that phrase: assisted suicide, aggressive pain management therapy, anabolic-steroid use, and cosmetic weight-loss ther apy. See, e. g., In re Harline, 65 Fed. Reg. 5665, 5667 (2000) (weight loss); In re Tecca, 62 Fed. Reg. 12842, 12846 (1997) (anabolic steroids); In re Roth, 60 Fed. Reg. 62262, 62263, 62267 (1995) (pain management). There is no indication that [p. 291] enforcement in these areas interferes with the prosecution of “drug abuse” as the Court understands it. Unlike in Whitman, the Attorney General’s additional power to ad dress other forms of drug “abuse” does absolutely nothing to undermine the central features of this regulatory scheme. Of course it was critical to our analysis in Whitman that the language of the provision did not bear the meaning that respondents sought to give it. See 531 U. S., at 465. Here, for the reasons stated above, the provision is most naturally interpreted to incorporate a uniform federal standard for le gitimacy of medical practice.6

Finally, respondents argue that the Attorney General must defer to state-law judgments about what constitutes legitimate medicine, on the ground that Congress must speak clearly to impose such a uniform federal standard upon the States. But no line of our clear-statement cases is applica ble here. The canon of avoidance does not apply, since the Directive does not push the outer limits of Congress’s com merce power, compare Solid Waste Agency of Northern Cook Cty. v. Army Corps of Engineers, 531 U. S. 159, 172 (2001) (regulation of isolated ponds), with United States v. Sullivan, 332 U. S. 689, 698 (1948) (regulation of labeling of drugs shipped in interstate commerce), or impinge on a core aspect of state sovereignty, cf. Atascadero State Hospital v. Scanlon, 473 U. S. 234, 242 (1985) (sovereign immunity); Gregory v. Ashcroft, 501 U. S. 452, 460 (1991) (qualifications of state government officials). The clear-statement rule based on the presumption against pre-emption does not [p. 292] apply because the Directive does not pre-empt any state law, cf. id., at 456–457; Rush Prudential HMO, Inc. v. Moran, 536 U. S. 355, 359 (2002). And finally, no clear statement is required on the ground that the Directive intrudes upon an area traditionally reserved exclusively to the States, cf. BFP v. Resolution Trust Corporation, 511 U. S. 531, 544 (1994) (state regulation of titles to real property), because the Fed eral Government has pervasively regulated the dispensation of drugs for over 100 years. See generally Brief for Pro-Life Legal Defense Fund et al. as Amici Curiae 3–15. It would be a novel and massive expansion of the clear statement rule to apply it in a commerce case not involving pre-emption or constitutional avoidance, merely because Congress has chosen to prohibit conduct that a State has made a contrary policy judgment to permit. See Sullivan, supra, at 693.

Sections 823(f) and 824(a) explicitly grant the Attorney General the authority to register and deregister physicians, and his discretion in exercising that authority is spelled out in very broad terms. He may refuse to register or deregis ter if he determines that registration is “inconsistent with the public interest,” 21 U. S. C. § 823(f), after considering five factors, the fifth of which is “[s]uch other conduct which may threaten the public health and safety,” § 823(f)(5). See also In re Arora, 60 Fed. Reg. 4447, 4448 (1995) (“It is well estab lished that these factors are to be considered in the disjunc [p. 293] tive, i. e., the Deputy Administrator may properly rely on any one or a combination of factors, and give each factor the weight he deems appropriate”). As the Court points out, these broad standards were enacted in the 1984 amendments for the specific purpose of freeing the Attorney General’s dis cretion over registration from the decisions of state authori ties. See ante, at 261.

Three considerations make it perfectly clear that the stat ute confers authority to interpret these phrases upon the Attorney General. First, the Attorney General is solely and explicitly charged with administering the registration and deregistration provisions. See §§ 823(f), 824(a). By mak ing the criteria for such registration and deregistration such obviously ambiguous factors as “public interest” and “public health and safety,” Congress implicitly (but clearly) gave the Attorney General authority to interpret those criteria—whether or not there is any explicit delegation provision in the statute. “Sometimes the legislative delegation to an agency on a particular question is implicit rather than ex plicit. In such a case, a court may not substitute its own [p. 294] construction of a statutory provision for a reasonable inter pretation made by the administrator of an agency.” Chev ron, 467 U. S., at 844. The Court’s exclusive focus on the explicit delegation provisions is, at best, a fossil of our preChevron era; at least since Chevron, we have not conditioned our deferral to agency interpretations upon the existence of explicit delegation provisions. United States v. Mead Corp., 533 U. S. 218, 229 (2001), left this principle of implicit delega tion intact.

The Attorney General is thus authorized to promulgate regulations interpreting §§ 823(f) and 824(a), both by implicit delegation in § 823(f) and by two grounds of explicit delega tion in § 821. The Court nevertheless holds that this triply [p. 295] unambiguous delegation cannot be given full effect because “the design of the statute,” ante, at 265, evinces the intent to grant the Secretary of Health and Human Services exclusive authority over scientific and medical determinations. This proposition is not remotely plausible. The Court cites as au thority for the Secretary’s exclusive authority two specific areas in which his medical determinations are said to be binding on the Attorney General—with regard to the “scien tific and medical evaluation” of a drug’s effects that precedes its scheduling, § 811(b), and with regard to “the appropriate methods of professional practice in the medical treatment of the narcotic addiction of various classes of narcotic addicts,” 42 U. S. C. § 290bb–2a; see also 21 U. S. C. § 823(g) (2000 ed. and Supp. II). See ante, at 265–266. Far from establishing a general principle of Secretary supremacy with regard to all scientific and medical determinations, the fact that Congress granted the Secretary specifically defined authority in the areas of scheduling and addiction treatment, without other wise mentioning him in the registration provisions, sug gests, to the contrary, that Congress envisioned no role for the Secretary in that area—where, as we have said, interpre tive authority was both implicitly and explicitly conferred upon the Attorney General.

Even if we could rewrite statutes to accord with sensible “design,” it is far from a certainty that the Secretary, rather than the Attorney General, ought to control the registration of physicians. Though registration decisions sometimes re quire judgments about the legitimacy of medical practices, the Department of Justice has seemingly had no difficulty making them. See In re Harline, 65 Fed. Reg. 5665; In re Tecca, 62 Fed. Reg. 12842; In re Roth, 60 Fed. Reg. 62262. But unlike decisions about whether a substance should be scheduled or whether a narcotics addiction treatment is le gitimate, registration decisions are not exclusively, or even primarily, concerned with “medical [and] scientific” factors. See 21 U. S. C. § 823(f). Rather, the decision to register, or [p. 296] to bring an action to deregister, an individual physician im plicates all the policy goals and competing enforcement pri orities that attend any exercise of prosecutorial discretion. It is entirely reasonable to think (as Congress evidently did) that it would be easier for the Attorney General occasionally to make judgments about the legitimacy of medical practices than it would be for the Secretary to get into the business of law enforcement. It is, in other words, perfectly consist ent with an intelligent “design of the statute” to give the Nation’s chief law enforcement official, not its chief health official, broad discretion over the substantive standards that govern registration and deregistration. That is especially true where the contested “scientific and medical” judgment at issue has to do with the legitimacy of physician-assisted suicide, which ultimately rests, not on “science” or “medi cine,” but on a naked value judgment. It no more depends upon a “quintessentially medical judgmen[t],” ante, at 267, than does the legitimacy of polygamy or eugenic infanticide. And it requires no particular medical training to undertake the objective inquiry into how the continuing traditions of Western medicine have consistently treated this subject. See OLC Memo, App. to Pet. for Cert. 113a–130a. The Sec retary’s supposedly superior “medical expertise” to make “medical judgments,” ante, at 266, is strikingly irrelevant to the case at hand.

The Court also reasons that, even if the CSA grants the Attorney General authority to interpret § 823(f), the Direc tive does not purport to exercise that authority, because it “does not undertake the five-factor analysis” of § 823(f) and does not “on its face purport to be an application of the registration provision in § 823(f).” Ante, at 261 (empha sis added). This reasoning is sophistic. It would be im proper—indeed, impossible—for the Attorney General to “undertake the five-factor analysis” of § 823(f) and to “appl[y] the registration provision” outside the context of an actual enforcement proceeding. But of course the Attorney Gen [p. 297] eral may issue regulations to clarify his interpretation of the five factors, and to signal how he will apply them in fu ture enforcement proceedings. That is what the Directive plainly purports to do by citing § 824(a)(4), and that is why the Directive’s conclusion on deregistration is couched in conditional terms: “Such conduct by a physician . . . may ‘render his registration . . . inconsistent with the public in terest’ and therefore subject to possible suspension or rev ocation under 21 U. S. C. [§] 824(a)(4).” 66 Fed. Reg. 56608 (emphasis added).

In concluding to the contrary, the Court merely presents the conclusory assertion that “it is doubtful the Attorney General could cite the ‘public interest’ or ‘public health’ to deregister a physician simply because he deemed a contro versial practice permitted by state law to have an illegiti mate medical purpose.” Ante, at 264. But why on earth not?—especially when he has interpreted the relevant statu tory factors in advance to give fair warning that such a prac tice is “inconsistent with the public interest.” The Attorney General’s discretion to determine the public interest in this [p. 298] area is admittedly broad—but certainly no broader than other congressionally conferred executive powers that we have upheld in the past. See, e. g., National Broadcasting Co. v. United States, 319 U. S. 190, 216–217 (1943) (“public interest”); New York Central Securities Corp. v. United States, 287 U. S. 12, 24–25 (1932) (same); see also Mistretta v. United States, 488 U. S. 361, 415–416 (1989) (Scalia, J., dissenting).

The Court’s decision today is perhaps driven by a feeling that the subject of assisted suicide is none of the Federal Government’s business. It is easy to sympathize with that position. The prohibition or deterrence of assisted suicide is certainly not among the enumerated powers conferred on the United States by the Constitution, and it is within the realm of public morality (bonos mores) traditionally ad dressed by the so-called police power of the States. But then, neither is prohibiting the recreational use of drugs or discouraging drug addiction among the enumerated powers. From an early time in our national history, the Federal Gov ernment has used its enumerated powers, such as its power to regulate interstate commerce, for the purpose of protect ing public morality—for example, by banning the interstate shipment of lottery tickets, or the interstate transport of women for immoral purposes. See Hoke v. United States, [p. 299] 227 U. S. 308, 321–323 (1913); Lottery Case, 188 U. S. 321, 356 (1903). Unless we are to repudiate a long and well established principle of our jurisprudence, using the federal commerce power to prevent assisted suicide is unquestion ably permissible. The question before us is not whether Congress can do this, or even whether Congress should do this; but simply whether Congress has done this in the CSA. I think there is no doubt that it has. If the term “legitimate medical purpose” has any meaning, it surely excludes the prescription of drugs to produce death.

Today the majority beats a hasty retreat from these con clusions. Confronted with a regulation that broadly re [p. 300] quires all prescriptions to be issued for a “legitimate medical purpose,” 21 CFR § 1306.04(a) (2005), a regulation recognized in Raich as part of the Federal Government’s “closed . . . system” for regulating the “manner” in “which controlled substances can be utilized for medicinal purposes,” 545 U. S., at 13, 27, the majority rejects the Attorney General’s admit tedly “at least reasonable,” ante, at 272, determination that administering controlled substances to facilitate a patient’s death is not a “ ‘legitimate medical purpose.’ ” The majority does so based on its conclusion that the CSA is only con cerned with the regulation of “medical practice insofar as it bars doctors from using their prescription-writing powers as a means to engage in illicit drug dealing and trafficking as conventionally understood.” Ante, at 270. In other words, in stark contrast to Raich’s broad conclusions about the scope of the CSA as it pertains to the medicinal use of con trolled substances, today this Court concludes that the CSA is merely concerned with fighting “ ‘drug abuse’ ” and only insofar as that abuse leads to “addiction or abnormal effects on the nervous system.”1 Ante, at 273.

The majority’s newfound understanding of the CSA as a statute of limited reach is all the more puzzling because it rests upon constitutional principles that the majority of the Court rejected in Raich. Notwithstanding the States’ “ ‘traditional police powers to define the criminal law and to protect the health, safety, and welfare of their citizens,’ ” 545 U. S., at 30, n. 38, the Raich majority concluded that the CSA applied to the intrastate possession of marijuana for medici nal purposes authorized by California law because “Congress could have rationally” concluded that such an application was necessary to the regulation of the “larger interstate mari juana market.” Id., at 30, 32. Here, by contrast, the major [p. 301] ity’s restrictive interpretation of the CSA is based in no small part on “the structure and limitations of federalism, which allow the States ‘ “great latitude under their police powers to legislate as to the protection of the lives, limbs, health, comfort, and quiet of all persons.” ’ ” Ante, at 270 (quoting Medtronic, Inc. v. Lohr, 518 U. S. 470, 475 (1996), in turn quoting Metropolitan Life Ins. Co. v. Massachusetts, 471 U. S. 724, 756 (1985)). According to the majority, these “background principles of our federal system . . . belie the notion that Congress would use . . . an obscure grant of au thority to regulate areas traditionally supervised by the States’ police power.” Ante, at 274.

I agree with limiting the applications of the CSA in a man ner consistent with the principles of federalism and our constitutional structure. Raich, supra, at 74 (Thomas, J., dissenting); cf. Whitman, supra, at 486–487 (Thomas, J., concurring) (noting constitutional concerns with broad dele gations of authority to administrative agencies). But that is now water over the dam. The relevance of such considera tions was at its zenith in Raich, when we considered whether the CSA could be applied to the intrastate possession of a controlled substance consistent with the limited federal pow [p. 302] ers enumerated by the Constitution. Such considerations have little, if any, relevance where, as here, we are merely presented with a question of statutory interpretation, and not the extent of constitutionally permissible federal power. This is particularly true where, as here, we are interpreting broad, straightforward language within a statutory frame work that a majority of this Court has concluded is so comprehensive that it necessarily nullifies the States’ “ ‘traditional . . . powers . . . to protect the health, safety, and welfare of their citizens.’ ”2 Raich, supra, at 30, n. 38. The Court’s reliance upon the constitutional principles that it re jected in Raich—albeit under the guise of statutory in terpretation—is perplexing to say the least. Accordingly, I respectfully dissent.