Opinion for the Court filed by Circuit Judge ROGERS.
Dissenting opinion filed by Circuit Judge GRIFFITH.
ROGERS, Circuit Judge.
The Abigail Alliance for Better Access to Developmental Drugs (“the Alliance”) [472]*472seeks to enjoin the Food and Drug Administration (“FDA”) from continuing to enforce a policy barring the sale of new drugs that the FDA has determined, after Phase I trials on human beings, are sufficiently safe for expanded human testing (hereafter “post-Phase I investigational new drugs”). More specifically, the Alliance seeks access to potentially life-saving post-Phase I investigational new drugs on behalf of mentally competent, terminally ill adult patients who have no alternative government-approved treatment options (hereafter “terminally ill patients”). The Alliance contends that the FDA’s policy violates the substantive due process rights to privacy, liberty, and life of its terminally ill members. The complaint presents the question of whether the Due Process Clause protects the right of terminally ill patients to decide, without FDA interference, whether to assume the risks of using potentially life-saving investigational new drugs that the FDA has yet to approve for commercial marketing but that the FDA has determined, after Phase I clinical human trials, are safe enough for further testing on a substantial number of human beings.
Upon applying the Supreme Court’s test for addressing substantive due process claims set forth in Washington v. Glucksberg, 521 U.S. 702, 710, 117 S.Ct. 2258, 138 L.Ed.2d 772 (1997), we hold that the district court erred in dismissing the Alliance’s complaint pursuant to Federal Rule of Civil Procedure 12(b)(6) for failure to state a claim. First, the right at issue, carefully described, is the right of a mentally competent, terminally ill adult patient to access potentially life-saving post-Phase I investigational new drugs, upon a doctor’s advice, even where that medication carries risks for the patient. Second, we find, upon examining “our Nation’s history, legal traditions, and practices,” Glucksberg, 521 U.S. at 710, 117 S.Ct. 2258, that the government has not blocked access to new drugs throughout the greater part of our Nation’s history. Only in recent years has the government injected itself into consideration of the effectiveness of new drugs. Third, Supreme Court precedent on liberty indicates that the right claimed by the Alliance can be inferred from the Court’s conclusion in Cruzan v. Director, Missouri Department of Health, 497 U.S. 261, 278, 110 S.Ct. 2841, 111 L.Ed.2d 224 (1990), that an individual has a due process right to refuse life-sustaining medical treatment, id. at 279, 110 S.Ct. 2841. Here, the claim implicates a similar right — the right to access potentially life-sustaining medication where there are no alternative government-approved treatment options. In both instances, the key is the patient’s right to make the decision about her life free from government interference.
Because the question remains whether the FDA’s challenged policy has violated that right, we reverse the dismissal of the Alliance’s complaint and remand the case to the district court to determine whether the FDA’s policy “is narrowly tailored to serve a compelling [governmental] interest.” Glucksberg, 521 U.S. at 721, 117 S.Ct. 2258 (quoting Reno v. Flores, 507 U.S. 292, 302, 113 S.Ct. 1439, 123 L.Ed.2d 1 (1993)).
In Part I, we set forth the background to this appeal. In Part II, we examine Supreme Court precedent indicating how substantive due process rights are to be discerned. So guided, we consider, in Part III, whether the Alliance’s claimed right warrants protection under the Due Process Clause.
I.
A.
The Food, Drug, and Cosmetic Act (“FDCA”), Pub.L. No. 75-717, §§ 1-902, [473]*47352 Stat. 1040 (1938) (codified as amended at 21 U.S.C. § 301 et seq. (2000)), prohibits drug manufacturers from introducing any “new drug” into interstate commerce until manufacturers have applied for, and received, FDA approval. 21 U.S.C. § 355(a). A “new drug” is any substance covered by the FDCA not “generally recognized, among experts ... as safe and effective for use under the conditions prescribed in the labeling.” 21 U.S.C. § 321(p)(l); see also United States v. 50 Boxes More or Less, 909 F.2d 24 (1st Cir.1990). Before a new drug is eligible for full approval and marketing, the Secretary of the U.S. Department of Health and Human Services must find “substantial evidence that the drug will have the effect it purports or is represented to have.” 21 U.S.C. § 355(d). Exempted from this general ban are new drugs “intended solely for investigational use by experts . . . .” Id. § 355(i)(l).
The FDCA directs the Secretary to promulgate regulations for testing new drugs. Id. Pursuant to this authority, the FDA has promulgated regulations that require three phases of government testing on humans before investigational new drugs can receive FDA approval and enter the commercial marketplace. In Phase I, new drugs are tested on 20 to 80 human subjects to determine “the side effects associated with increasing doses, and, if possible, to gain early evidence on effectiveness.” 21 C.F.R. § 312.21(a). It takes approximately one year to conduct Phase I testing.1 FDA counsel acknowledged at oral argument that drugs that survive this phase have been deemed “sufficiently safe for substantial human testing, but [are] not yet proven to be safe and effective to the satisfaction of the FDA [to be commercially marketed].” Oral Argument Tape of Oct. 21, 2005 at 15:57-15:59. Phase II involves targeted, controlled clinical studies of up to several hundred human subjects “to evaluate the effectiveness of the [Phase I investigational new] drug ... and to determine the common short-term side effects and risks associated with the drug.” 21 C.F.R. § 312.21(b). Phase III expanded trials, which can include several thousand human subjects, are “performed after preliminary evidence suggesting effectiveness of the drug has been obtained, and are intended to gather the additional information about effectiveness and safety that is needed to evaluate the overall benefit-risk relationship of the drug .... ” Id. § 312.21(c). With narrow exceptions, FDA regulations require informed consent to be obtained from clinical trial participants. Id. §§ 50.1-50.27.
B.
On January 16, 2003, the Alliance submitted a proposal to the FDA for new regulations to render post-Phase I investigational new drugs available to terminally ill patients who were not admitted to the FDA’s clinical trials. The FDA rejected the proposal by letter dated April 25, 2003, outlining the FDA’s policy. On June 11, 2003, Alliance filed a Citizen Petition, pursuant to 21 C.F.R. § 10.30, challenging the FDA’s policy barring the sale of investigational new drugs that have successfully completed Phase I trials to terminally ill patients not selected for clinical trials. The FDA acknowledged receipt of the Citizen Petition but otherwise did not respond within 180 days, thereby entitling the Alliance to seek judicial review of the challenged policy. See id. § 10.30(e)(2).
The Alliance filed suit against the FDA Commissioner and the Secretary of the [474]*474Department of Health and Human Services, seeking to enjoin the FDA from enforcing the policy barring the sale of post-Phase I investigational new drugs to terminally ill patients not in Phase II clinical trials. Noting that the FDA has administrative discretion to define several stages for human testing of new drugs after animal testing has been conducted, the complaint alleges that it takes, on average, just under seven years for investigational new drugs to complete the three phases of clinical human trials and receive FDA approval for commercial marketing and thus become eligible for purchase by persons not in FDA clinical trials. Compl. ¶ 12.2 The complaint also alleges that non-commercial options provide relief only to a very small number of terminally ill patients as spaces in clinical trials are “very limited ... in relation to the need.” Compl. ¶ 15. The Alliance asserts that clinical human trials are limited in number and by type of patient who qualifies. Further, the FDA’s “compassionate use” programs, which permit drug companies voluntarily to provide new drugs at cost during the pre-approval period, are available only to “a fraction of those in desperate need.” Id. Although the FDA may permit “treatment use” of unapproved new drugs, see 21 C.F.R. § 312.34 (1999), and has allowed access for limited groups of persons with AIDS,3 the FDA has refused as a general matter to allow terminally ill patients to have access to investigational new drugs that have successfully completed Phase I trials. Consequently, the complaint alleges, the effect of the FDA policy, as illustrated by the examples of the deaths of four terminally ill patients, has been to deny terminally ill patients the choice to use post-Phase I investigational new drugs despite the patients’ willingness “to assume risks if their physicians advise them that a treatment may save or prolong their lives and if they have no other viable options.” Compl. ¶¶ 16, 18. Prior to discovery, the FDA moved to dismiss the complaint, and, alternatively, for summary judgment. The Alliance responded by filing an opposition and its own motion for summary judgment.
The district court dismissed the complaint pursuant to Rule 12(b)(6) for failure to state a claim. The court rejected the Alliance’s argument that it sought no “new” right but only recognition and enforcement of the right to life that is explicitly guaranteed by the Due Process Clause, observing that no court decision has “extended the Due Process Clause to cover a terminally ill patient’s right to receive medical treatment.” Mem. Op. of Aug. 30, 2004, at 18, 2004 WL 3777340 (emphasis deleted). Although acknowledging “the Nation’s longstanding legal tradition ... to attempt to preserve life,” id., the district court stated that in Glucksberg, the Supreme Court had distinguished some “personal” decisions from others, 521 U.S. at 727, 117 S.Ct. 2258, and that the Alliance could not “possibl[y] claim that the specific right claimed has a long-standing tradition.” Mem. Op. at 18. The district court also rejected the Alliance’s argument that the Supreme Court’s recognition in Cruzan of the right to choose death by refusing medical treatment implied a complementary right to choose life by obtaining potentially life[475]*475saving medication. In the district court’s view, the Alliance sought recognition of “an entirely different sort of right [from that recognized in Cruzan ] — not freedom from government imposition, but an affirmative right of access to medical treatment.” Id. at 19. In the absence of due process protection for terminally ill patients seeking access to potentially life saving post-Phase I drugs, the district court concluded that the challenged FDA policy is rationally related to a legitimate governmental interest.
The Alliance appeals, and our review is de novo4 See Cicippio-Puleo v. Islamic Republic of Iran, 353 F.3d 1024, 1031-32 (D.C.Cir.2004). We treat the dismissal of the complaint as occurring pursuant to Rule 12(b)(6), notwithstanding the district court’s consideration of the FDA’s April 23, 2003 letter because the letter’s conclusion was alleged in the complaint and the FDA does not dispute its contents. See Gryl ex rel. Shire Pharms. Group PLC v. Shire Pharms. Group PLC, 298 F.3d 136, 140 (2d Cir.2002); Pryor v. Nat’l Collegiate Athletic Ass’n, 288 F.3d 548, 560 (3d Cir.2002) (citing 62 Fed. Proc. L.Ed. § 62:508). Cf. Settles v. United States Parole Comm’n, 429 F.3d 1098, 1107 (D.C.Cir .2005).
A court should not dismiss a complaint pursuant to Rule 12(b)(6) for failure to state a claim “unless it appears beyond doubt that the plaintiff can prove no set of facts in support of his claim which would entitle him to relief.” Conley v. Gibson, 355 U.S. 41, 45-46, 78 S.Ct. 99, 2 L.Ed.2d 80 (1957); Warren v. District of Columbia, 353 F.3d 36, 37 (D.C.Cir.2004). In determining the sufficiency of the complaint, this court reviews questions of law de novo while treating the complaint’s factual allegations as true and granting the plaintiff the benefit of all reasonable inferences from the facts alleged. See Conley, 355 U.S. at 45-46, 78 S.Ct. 99; Sparrow v. United Air Lines, Inc., 216 F.3d 1111, 1114 (D.C.Cir.2000).
II.
The Due Process Clause of the Fifth Amendment to the United States Constitution provides that “[n]o person shall be ... deprived of life, liberty, or property, without due process of law.” U.S. Const. Amend. V. The Supreme Court has held that the Clause “guarantees more than fair process” and accords substantive protection to the rights it guarantees. See Troxel v. Granville, 530 U.S. 57, 65, 120 S.Ct. 2054, 147 L.Ed.2d 49 (2000) (plurality opinion); Glucksberg, 521 U.S. at 719, 117 S.Ct. 2258; Flores, 507 U.S. at 301-02, 113 S.Ct. 1439. Substantive due process claims can present difficulties for courts. See Michael H. v. Gerald D., 491 U.S. 110, 121, 109 S.Ct. 2333, 105 L.Ed.2d 91 (1989) (plurality opinion); Moore v. City of East Cleveland, 431 U.S. 494, 502, 97 S.Ct. 1932, 52 L.Ed.2d 531 (1977). In a case of first impression where fundamental rights may be at stake, determining the limits of the government’s authority over an individual’s freedom to make certain personal decisions unavoidably entails a careful and possibly arduous assessment of that personal decision’s objective characteristics in order to determine whether it warrants protection under the Due Process Clause. Cf. Roberts v. U.S. Jaycees, 468 U.S. 609, 620, 104 S.Ct. 3244, 82 L.Ed.2d 462 (1984). Nonetheless, the district court appears to have viewed its role as unduly constrained. Pointing to an advisory cautioning in Dro[476]*476nenburg v. Zech, 741 F.2d 1388, 1396 (D.C.Cir.1984), that lower courts “should [not] freely create new constitutional rights” without “guidance from the Constitution or ... from articulated Supreme Court principle,” the district court focused on the absence of binding precedent recognizing the substantive due process right claimed by the Alliance. Since Dronenburg, the Supreme Court has provided guideposts to enable a court to assess the merits of the Alliance’s claim.5
Although the Supreme Court has never explicitly said so, and we need not decide the matter here, it appears the Supreme Court has employed two distinct approaches when faced with a claim to a fundamental right. In some cases, the Court has discerned the existence of fundamental rights by probing what “personal dignity and autonomy” demand. See Planned Parenthood of Southeastern Pa. v. Casey, 505 U:S. 833, 851, 112 S.Ct. 2791, 120 L.Ed.2d 674 (1992) (citations omitted). In other cases, the Court has derived fundamental rights by reference to the Nation’s history and legal tradition, see, e.g., Glucksberg, 521 U.S. 702, 117 S.Ct. 2258.6 The line of cases beginning with Griswold v. Connecticut, 381 U.S. 479, 85 S.Ct. 1678, 14 L.Ed.2d 510 (1965), and continuing through Eisenstadt v. Baird, 405 U.S. 438, 92 S.Ct. 1029, 31 L.Ed.2d 349 (1972), Roe v. Wade, 410 U.S. 113, 93 S.Ct. 705, 35 L.Ed.2d 147 (1973), and Casey, 505 U.S. 833, 112 S.Ct. 2791, follow the first approach with their heavy reliance on the concepts of individual rights to autonomy and self-determination, and in their unwillingness to countenance state intrusion into certain protected domains such as the bedroom, the clinic, and the womb. This approach is succinctly captured by Casey’s characterization of substantive due process rights as those that involve “the most intimate and personal choices a person may make in a lifetime, choices central to personal dignity and autonomy.” Casey, 505 U.S. at 851, 112 S.Ct. 2791.
The other approach for determining whether a claimed right warrants substantive due process protection, which appears to be more restrictive,7 has two “features.” See Glucksberg, 521 U.S. at 720, 117 S.Ct. 2258. Under Glucksberg, courts must inquire whether the fundamental right asserted is “objectively, ‘deeply rooted in this Nation’s history and tradition,’ ” id. at 721, 117 S.Ct. 2258 (quoting Moore, 431 U.S. at 503, 97 S.Ct. 1932; Snyder v. Massachusetts, 291 U.S. 97, 54 S.Ct. 330, 78 L.Ed. 674 (1934)),8 and “implicit in the concept of [477]*477ordered liberty, such that neither liberty nor justice would exist if [it] were sacrificed,” Glucksberg, 521 U.S. at 721, 117 S.Ct. 2302 (quoting Palko v. Connecticut, 302 U.S. 319, 325-26, 58 S.Ct. 149, 82 L.Ed. 288 (1937)) (internal quotation marks omitted). Additionally, in order to ensure that courts do not multiply rights without principled boundaries, courts must provide a “careful description of the fundamental liberty interest.” Id. at 721-23, 117 S.Ct. 2302. If a court concludes that the claimed right is a fundamental right entitled to protection under the Due Process Clause, then the burden shifts to the government to show that its encroachment upon the right “is narrowly tailored to serve a compelling [governmental] interest.” Id. at 721, 117 S.Ct. 2302 (quoting Flores, 507 U.S. at 302, 113 S.Ct. 1439).
Because we conclude, upon applying the seemingly more restrictive analysis of Glucksberg, that the claimed right warrants protection under the Due Process Clause, we need not decide whether the line of cases construing the concept of “personal dignity and autonomy” would also lend protection to the claimed right.
III.
The question presented by the Alliance’s complaint is whether the Due Process Clause protects the right of terminally ill patients to make an informed decision that may prolong life, specifically by use of potentially life-saving new drugs that the FDA has yet to approve for commercial marketing but that the FDA has determined, after Phase I clinical human trials, are safe enough for further testing on a substantial number of human beings. The Due Process Clause, as Glucksberg makes clear, protects those liberties “deeply rooted in this Nation’s history and tradition.” 521 U.S. at 721, 117 S.Ct. 2258 (citation omitted). The Supreme Court has variously referred to these rights as principles “so rooted in the traditions and conscience of our people as to be ranked as fundamental,” Snyder, 291 U.S. at 105, 54 S.Ct. 330, and as immunities “implicit in the concept of ordered liberty,” Palko, 302 U.S. at 325, 58 S.Ct. 149. Thus, a court’s examination of our Nation’s history and tradition cannot be based on so specific a description of the claimed right as would undercut the interests protected by the Due Process Clause.
One feature of the Glucksberg analysis requires courts to compose a “careful description” of the asserted fundamental liberty interest before extending due process protection to it. 521 U.S. at 721, 117 S.Ct. 2258. The Supreme Court has not settled on how precisely formulated the right must be. Two Justices have interpreted the “careful description” requirement as indicating that courts should identify fundamental rights at the “most specific level at which a relevant tradition protecting, or denying protection to, the asserted right can be identified.” Michael H., 491 U.S. [478]*478at 127 n. 6, 109 S.Ct. 2333 (1989) (Scalia, J., with Rehnquist, C.J., concurring). Two other Justices have indicated that asserted rights not expressed at “ ‘the most specific level’ [of generality] available” can nonetheless be recognized. Id. at 132, 109 S.Ct. 2333 (O’Connor and Kennedy, JJ., concurring). The “careful description” requirement was first invoked by the Court in Flores, 507 U.S. at 302, 113 S.Ct. 1439 (1993), which relied on Collins v. City of Harker Heights, 503 U.S. 115, 125, 112 S.Ct. 1061, 117 L.Ed.2d 261 (1992), where the notion of careful description was expressed as a pleading requirement. Since Glucksberg, the Court has applied this requirement once without elaboration. See Chavez v. Martinez, 538 U.S. 760, 775-76, 123 S.Ct. 1994, 155 L.Ed.2d 984 (2003).
In Hutchins v. District of Columbia, 188 F.3d 531 (D.C.Cir.1999), the en banc court applied the careful description requirement in its substantive due process analysis. The court viewed the careful description requirement as a means of constraining the inadvertent creation of rights that could fall within the scope of loosely worded descriptions and thus threaten the separation of powers. See id. at 542-45. Despite reaching different conclusions about the appropriate level of generality in describing the claimed right, compare id. at 538 (citing Michael H., 491 U.S. at 127 n. 6, 109 S.Ct. 2333) (Scalia, J., with Rehnquist, C.J., concurring), with id. at 555-57 (Rogers, J., dissenting) (citing Moore, 431 U.S. at 502-03, 97 S.Ct. 1932), the court concluded that the animating principle underlying the careful description requirement is that courts should proceed with care in examining substantive due process claims. See id. at 538.
The Alliance’s complaint contains the careful description we seek, allowing this court to consider whether the challenged FDA policy impinges upon one or more of the interests protected by the Due Process Clause. The FDA characterizes the Alliance’s claimed right as a broadly stated prerogative to access post-Phase I investigational new drugs and to receive treatment, but the Alliance has defined the right more narrowly. The Alliance claims neither an unfettered right of access to all new or investigational new drugs nor a right to receive treatment from the government or at government expense. The Alliance’s claim also does not challenge the Controlled Substances Act, 21 U.S.C. §§ 801 et seq., or the government’s authority to regulate substances deemed harmful to public health, safety, and welfare. Rather, the Alliance contends that the fundamental due process rights to privacy, liberty, and life include the right of terminally ill patients, acting on a doctor’s advice, to obtain potentially life-saving medication when no alternative treatment approved by the government is available. Recognizing that the effectiveness and side effects of the investigational new drugs may still be in question after the Phase I trials have been completed, the Alliance asks only that the decision to assume these known or unknown risks be left to the terminally ill patient and not to the FDA. This description of the claimed right conforms to the demands of even the narrowest interpretation of the Glucksberg “careful description” requirement.9
[479]*479B.
The other feature of the Glucksberg inquiry requires courts to determine whether there exists a long-standing tradition in our Nation that would protect individual access to potentially life-saving medication. Courts must focus on discerning those constitutionally protected interests whose existence can be inferred from the Due Process Clause and Supreme Court precedent construing the Clause. See Cruzan, 497 U.S. at 278, 110 S.Ct. 2841. Although it is relevant to the substantive due process analysis that the government has never proscribed the desired conduct, this is not dispositive. The absence of regulation could be attributable to a liberty interest that is deeply rooted in this Nation’s history and tradition, and there[480]*480fore characterized by a history of liberty from governmental interference, but there may be another explanation. For example, a lack of regulation might indicate only that the technology of yesteryear did not warrant it.
The FDA’s discussion of the merits of this question consists of a single sentence: “[The] FDA has had statutory authority to regulate drugs for almost a century, and that authority is now firmly ingrained in our understanding of the appropriate role of government.” Appellee’s Br. at 19.10 We offer the following observations, mindful of the fact that the Alliance is complaining only of obstacles to post-Phase I investigational new drugs erected by the FDA and not obstacles that might be erected by state consumer protection or other laws.11
A right of control over one’s body has deep roots in the common law. The venerable commentator on the common law William Blackstone wrote that the right to “personal security” includes “a person’s legal and uninterrupted enjoyment of his life, his limbs, his body, [and] his health,” as well as “the preservation of a man’s health from such practices as may prejudice or annoy it.” William Blackstone, 1 Commentaries *125, *130. This right included the .right to self-defense and the right to self-preservation. “For whatever is done by a man, to save either life or member, is looked upon as done upon the highest necessity and compulsion.” Id. at *127. As recognized throughout Anglo-American history and law, when a person is faced with death, necessity often warrants extraordinary measures not otherwise justified. Indeed the principle holds even when that action impinges upon the rights of others. See, e.g., Ploof v. Putnam, 81 Vt. 471, 475, 71 A. 188 (1908) (“This doctrine of necessity applies with special force to the preservation of human life____ One may sacrifice the personal property of another to save his life or the lives of his fellows.”) (internal citation omitted); Mouse’s Case, 77 Eng. Rep. 1341, 1342 (K.B.1609) (deciding that it is lawful to throw overboard property of another for safety of lives of passengers); Restatement (First) of Torts § 197 (1934); see generally George C. Christie, The Defense of Necessity Considered from the Legal and Moral Points of View, 48 Duke L.J. 975 (1996). But see The Queen v. Dudley and Stephens, 14 Q.B.D. 273 (1884) (holding that the defense of necessity did not justify taking of innocent life). Barring a terminally ill patient from the use of a potentially life-saving treatment impinges on this right of self-preservation.
Such a bar also puts the FDA in the position of interfering with efforts that could save a terminally ill patient’s life. Although the common law imposes no general duty to rescue or to preserve a life, it does create liability for interfering with such efforts. Section 326 of the Restatement (First) of Torts, first published in 1934, explained that
[481]*481[o]ne who, without a privilege to do so, intentionally prevents a third person from giving to another aid necessary to his bodily security, is liable for bodily harm caused to the other by the absence of aid which he has prevented the third person from giving.
While infrequently invoked, this common law rule is of venerable vintage. See id.; see also Soldano v. O’Daniels, 141 Cal. App.3d 443, 190 Cal.Rptr. 310, 313, 316-18 (1983); Miller v. Arnal Corp., 129 Ariz. 484, 632 P.2d 987, 993 (App.1981).12
In contrast to these ancient principles, regulation of access to new drugs has a history in this country that is of recent origin. Prior to 1906, there was essentially no drug regulation in the United States.13 In that year Congress enacted the Pure Food and Drug Act (“1906 Act”), Pub.L. No. 59-384, 34 Stat. 768 (repealed 1938), which prohibited misbranded and adulterated foods or drugs from entering interstate commerce, 34 Stat. at 768, and prohibited false and misleading labeling, id. at 770. For a small number of particularly dangerous drugs, the 1906 Act required the labels to identify the drug’s ingredients and quantities. Id. The statute also authorized the Bureau of Chemistry, a predecessor of the FDA, to seize nonconforming goods and to recommend federal prosecution of those who violated the 1906 Act. Id. at 769 § 4. The 1906 Act did not, however, limit individual access to new drugs or regulate therapeutic claims by drug manufacturers. Cf. United States v. Johnson, 221 U.S. 488, 31 S.Ct. 627, 55 [482]*482L.Ed. 823 (1911). It thus appears that a patient still could obtain access to any new drug for medicinal use, even if the drug had no therapeutic benefit, albeit subject to the controls placed on narcotics in 1914 by the Harrison Narcotic Act. Act of Dec. 17,1914, 38 Stat. 785.14
In 1938, Congress enacted the FDCA in response to the deaths of more than one hundred people, many of them children, from ingesting Elixir Sulfanilamide, which had been marketed as an antibiotic. See Report of the Secretary of Agriculture on Deaths Due to Elixir Sulfanilamide, S. Doc. No. 124, 75th Cong., 2d Sess. 1, 1-3 (1937) (“1937 Report”).15 For the first time, Congress required that drug manufacturers test, and the FDA review, all new drugs for safety prior to their commercial distribution. Pub.L. No. 75-717, 52 Stat. 1040 (1938) (codified as amended at 21 U.S.C. §§ 301 et seq.); 1937 Report at 1-3. Under the 1938 Act, a new drug could be commercially marketed only after the manufacturer filed a New Drug Application (“NDA”) with the FDA that set forth medical and scientific information attesting to the drug’s safety. The 1938 Act did not, however, require drug manufacturers to receive affirmative FDA approval before marketing the drug.16 Rather, an NDA became automatically effective within a time frame set by the FDA unless the FDA determined that the drug was unsafe and barred its commercial distribution.17 It was not until 1951, in the Durham-Humphrey Amendment, that Congress created the category of prescription drugs, i.e., drugs that are unsafe for self-medication but which can be used while under a doctor’s supervision. See Act of Oct. 25, 1951, 65 Stat. 648 (1951) (codified at 21 U.S.C. § 353(b)).
Only in 1962 did Congress require drug manufacturers to provide empirical evidence of the effectiveness of a drug as opposed to merely the drug’s safety.18 The Kefauver-Harris Amendments, Pub.L. No. 87-781, 76 Stat. 780 (1960) (codified in scattered sections of 21 U.S.C. §§ 301-81 (1982 & Supp. IV 1986)), were enacted in response to the rash of birth defects discovered in babies whose mothers had taken Thalidomide to ease morning sickness caused by pregnancy.19 The Kefauver-Harris Amendments transformed drug regulation and the approval process in several respects. First, the Amendments required the FDA to review a new drug for both safety and effectiveness and specified that to demonstrate effectiveness manufacturers were required to submit data from “adequate and well-controlled investigations.” 21 U.S.C. § 355(d). Second, the Amendments authorized the FDA to approve human clinical trials, regulate drug advertising, inspect drug-manufacturing facilities, and promulgate good manufacturing practices. The Amendments also required drug man[483]*483ufacturers to disclose to the FDA any information they received regarding the adverse consequences of approved drugs.20 This legislation set the framework for the system of drug regulation currently in place.
Despite the increased federal scrutiny of new drugs, important aspects of patient access to drugs are unregulated by the government and appear always to have been unregulated. “The FDA’s regulatory authority extends to manufacturers of drugs but not to the physicians who dispense them.”21 Thus, a doctor may prescribe a drug to a patient for a purpose other than that for which the FDA has approved the use of the drug. Such “off-label” use may occur even if the drug is not deemed safe or effective for that use. Further, it appears that the FDA has never prohibited either off-label prescription or off-label use of drugs.22 In recent years, the FDA has been moving to permit drug manufacturers to promote the use of their drugs for off-label purposes in limited circumstances.23 See Food and Drug Administration Modernization Act of 1997, Pub.L. No. 105-115, 111 Stat. 2296 (codified in scattered sections of 21 U.S.C. §§ 301-81).
For over half of our Nation’s history, then, until the enactment of the 1906 Act, a person could obtain access to any new drug without any government interference whatsoever. Even after enactment of the FDCA in 1938, Congress imposed no limitation on the commercial marketing of new drugs based upon the drugs’ effectiveness. Rather, at that time, the FDA could only interrupt the sale of new drugs based on its determination that a new drug was unsafe. Government regulation of drugs premised on concern over a new drug’s efficacy, as opposed to its safety, is of recent origin. And even today, a patient may use a drug for unapproved purposes even where the drug may be unsafe or ineffective for the off-label purpose. Despite the FDA’s claims to the contrary, therefore, it cannot be said that government control of access to potentially lifesaving medication “is now firmly ingrained in our understanding of the appropriate role of government,” Appellee’s Br. at 19, so as to overturn the long-standing tradition of the right of self-preservation.24
C.
The Alliance’s claim also falls squarely within the realm of rights the Supreme Court has held are “implicit in the concept [484]*484of ordered liberty.” Palko, 302 U.S. at 325, 58 S.Ct. 149. Specifically, the claimed right is implied by the Court’s conclusion in Cruzan that due process protects a person’s right to refuse life-sustaining treatment. See Cruzan, 497. U.S. at 279, 110 S.Ct. 2841. Writing for the Court, Chief Justice Rehnquist noted in examining the origins of the doctrine of informed consent that the Court had observed early on that “[n]o right is held more sacred, or is more carefully guarded, by the common law, than the right of every individual to the possession and control of his own person, free from all restraint or interference of others, unless by clear and unquestionable authority of law.” Id. at 269, 110 S.Ct. 2841 (quoting Union Pacific R. Co. v. Botsford, 141 U.S. 250, 251, 11 S.Ct. 1000, 35 L.Ed. 734 (1891)). The Court reasoned that “[t]he logical corollary of the doctrine of informed consent is that the patient generally possesses the right not to consent, that is, to refuse treatment.” Id. at 270, 110 S.Ct. 2841. Confronting for the first time what it described as a “perplexing question with unusually strong moral and ethical overtones,” id. at 277, 110 S.Ct. 2841, the Court turned to the language of the Fourteenth Amendment and its precedent to determine whether “the United States Constitution grants what is in common parlance referred to as a ‘right to die,’ ” id. The Court reasoned that “[t]he principle that a competent person has a constitutionally protected liberty interest in refusing unwanted medical treatment may be inferred from our prior decisions.” Id. Without qualification, the Court stated: “It cannot be disputed that the Due Process Clause protects an interest in life as well as an interest in refusing life-sustaining medical treatment.” Id. at 281, 110 S.Ct. 2841.
A similar analysis leads to the conclusion that the Due Process Clause protects the liberty interest claimed by the Alliance for its terminally ill members. See supra Part III.A. The text of the Due Process Clause refers to protecting “liberty” and “life.” Although there is no similarly clear textual basis for a “right to die” or refusing life-sustaining medical treatment, the Supreme Court in Cruzan recognized, in light of the common law and constitutionally protected liberty interests based on the inviolability of one’s body, that an individual has a due process right to make an informed decision to engage in conduct, by withdrawing treatment, that will cause one’s death.25 The logical corollary is that an individual must also be free to decide for herself whether to assume any known or unknown risks of taking a medication that might prolong her life.
Like the right claimed in Cruzan, the right claimed by the Alliance to be free of FDA imposition does not involve treatment by the government or a government subsidy. Rather, much as the guardians of the comatose patient in Cruzan did, the Alliance seeks to have the government step aside by changing its policy so the individual right of self-determination is not violated. The Alliance claims that there is a protected right of terminally ill patients to choose to use potentially life-saving investigational new drugs that have successfully cleared Phase I. If there is a protected liberty interest in self-determination that includes a right to refuse life-sustaining [485]*485treatment, even though this will hasten death, then the same liberty interest must include the complementary right of access to potentially life-sustaining medication, in light of the explicit protection accorded “life.”26 Our reasoning is not unlike that of the Supreme Court in Eisenstadt, 405 U.S. 438, 92 S.Ct. 1029, where the Court held that the right to be free from unwanted government intrusion into the fundamental decision whether to have children establishes a right of access to contraception.
Contrary to the FDA’s position, nothing in this court’s precedent or that of the other circuit courts of appeal conflicts with our analysis. Although the district court concluded, in reliance upon our decision in Dronenbivrg, 741 F.2d at 1396, that lower courts may not consider claims to new substantive due process rights and principles not previously identified by the Supreme Court, see supra page 475, this court has addressed substantive due process claims on a number of occasions. See, e.g., N.Y. State Ophthalmological Soc’y v. Bowen, 854 F.2d 1379 (D.C.Cir.1988). Most pertinently, in Butera v. District of Columbia, 235 F.3d 637 (D.C.Cir.2001), the court confronted, in the context of a qualified immunity defense, the claim of a substantive due process right to life, personal security, and bodily integrity. Butera involved a suit under 42 U.S.C. § 1983 brought by the mother of a man who was shot while working undercover for the police department. The court in Butera did not suggest that the advisory admonition in Dronenburg, 741 F.2d at 1396, precluded either the substantive due process inquiry or the conclusion that a fundamental right was implicated.
The decisions in the other circuits on which the FDA relies likewise fail to support its position that there is no substantive due process right of access to potentially life-saving treatment. United States v. Burzynski Cancer Research Institute, 819 F.2d 1301 (5th Cir.1987), which held that the doctor and patient had not stated a constitutional tort based on the allegedly improper seizure of the doctor’s patient records and thus that they did not overcome the defendant’s claim of qualified immunity, id. at 1310-11, bears no legal or factual relevance to the question before this court. The statement in Camohan v. United States, 616 F.2d 1120, 1122 (9th Cir.1980), that “[c]onstitutional rights of privacy and personal liberty do not give individuals the right to obtain [the cancer drug] laetrile free of the lawful exercise of government police power,” was dictum; the Ninth Circuit never reached the merits of the claimed fundamental right of access [486]*486as the complaint was dismissed for failure to exhaust administrative remedies.
Further, as the Alliance pointed out in its brief, the terminally ill patients in Rutherford v. United States, 616 F.2d 455 (10th Cir.1980), like those in Camohan, sought access to laetrile, a new cancer drug that had not cleared FDA’s Phase I safety hurdle and thus had not been approved for expanded testing on humans in ongoing clinical trials, see id. at 456-57. The Tenth Circuit rejected a right to laetrile, reasoning that the choice of a particular treatment or medication is “within the area of governmental interest in protecting public health.” Id. at 457. Of course, the government’s interest in regulating has no bearing upon the identification of a fundamental right. Rather, its interest is to be considered only if, and after, a court recognizes a fundamental right; at that point, the burden shifts to the government to demonstrate a narrowly tailored “compelling interest” in burdening that right. Because the FDA had neither eliminated the possibility that laetrile was a poison nor approved the drug for basic human testing in Phase I trials, the government’s interest in Rutherford might well have been sufficiently compelling to warrant restricting access to the drug. In this case, the government’s interest may prove to be weaker because the Alliance seeks only access to investigational new drugs that the FDA, after Phase I human trials, has deemed sufficiently safe for human testing on a substantial number of human beings. In other words, the Alliance seeks for its members the same right of access enjoyed by those terminally ill patients lucky enough to secure a spot in Phase II trials.
Accordingly, we hold that the district court erred in dismissing the Alliance’s complaint pursuant to Rule 12(b)(6) for failure to state a claim. We conclude, upon applying the Glucksberg analysis and heeding the protected liberty interests articulated by the Supreme Court, that where there are no alternative government-approved treatment options, a terminally ill, mentally competent adult patient’s informed access to potentially life-saving investigational new drugs determined by the FDA after Phase I trials to be sufficiently safe for expanded human trials warrants protection under the Due Process Clause. The prerogative asserted by the FDA — to prevent a terminally ill patient from using potentially life-saving medication to which those in Phase II clinical trials have access — thus impinges upon an individual liberty deeply rooted in our Nation’s history and tradition of self-preservation. See Glucksberg, 521 U.S. at 721, 117 S.Ct. 2258; Flores, 507 U.S. at 302, 113 S.Ct. 1439. The district court never reached the question of whether the challenged FDA policy violates this protected liberty interest, and we therefore remand the case to the district court to determine whether the FDA’s policy barring access to post-Phase I investigational new drugs by terminally ill patients is narrowly tailored to serve a' compelling governmental interest.