UNITED STATES OF AMERICA v. THOMAS KELLER (2025)

OPINION

Defendant-Appellant Thomas Keller appeals his conviction on four counts of prescribing controlled substances outside the scope of professional practice and his subsequent sentence to a term of 30 months of incarceration and 3 years of supervised release.  1 Keller makes four claims: (1) the district court erred in denying his motion to suppress a journal found at his residence that was seized pursuant to a search warrant justified by neither probable cause nor the “plain view” doctrine; (2) the district court erred in not holding an evidentiary hearing on his suppression motion; (3) the charges against him violated the nondelegation doctrine; and (4) the district court erred in calculating his sentencing range under United States Sentencing Guideline § 2D1.1 by relying on a drug conversion ratio found in the Sentencing Guidelines commentary.

As to Keller's first claim, the district court did not err in denying Keller's motion to suppress, because the seized journal fell within the scope of the search warrant and its seizure was supported by probable cause. 2 As to his second claim, the district court did not abuse its discretion in declining to hold an evidentiary hearing on the seizure of the journal because Keller's conclusory allegations did not establish contested issues of fact. Regarding Keller's third claim, there was no violation of the nondelegation doctrine, as the Attorney General's promulgation of the relevant regulations fell within the scope of the authority intelligibly delegated to the Attorney General by Congress. Finally, Keller's sentencing claim fails because the challenged drug conversion table underwent the same congressional review process as the Guidelines and was expressly incorporated into the relevant Guideline itself.

I. FACTUAL AND PROCEDURAL BACKGROUND

A. Factual Background

Defendant-Appellant Thomas Keller began working as a solo practitioner physician specializing in pain management in Santa Rosa, California, in 2011. Keller possessed a license from the Drug Enforcement Administration (“DEA”) allowing him to prescribe various controlled substances pursuant to 21 U.S.C. § 829(a) and (b). This allowed Keller to prescribe opiates like Percocet and OxyContin.

Between beginning his practice in 2011 and surrendering his license in 2018, Keller was in the 99th percentile of pain specialists “in terms of the amount [of opioids] he [was] prescribing per patient per day.” Keller was known by local pharmacists for prescribing “only narcotics,” as well as for prescribing opiates in exceptionally large quantities. One pharmacy eventually refused to fill prescriptions written by Keller. In one episode that the government alleged was demonstrative, a treating psychiatrist, after speaking with Keller about a patient they were separately treating, filed a complaint with the Medical Board of California:

Dr. Thomas Keller has been prescribing opioids to a pregnant and breastfeeding mother for the last year without any knowledge of her pregnancy or lactation status. I learned this after calling him to ask him about whether he was aware of her breastfeeding status due to my concern for the [Controlled Substance Utilization Review and Evaluation System (CURES)] report that indicate[d] ․ [Keller] refill[ed] a large amount of opioids during her pregnancy and breastfeeding.

In March 2013, Keller began to treat A.M., then 18 years old, for “low back pain.” A.M. presented no physical issues justifying prescribing large amounts of opiates, but Keller nevertheless prescribed many.

A.M. attempted suicide in December 2013. Following A.M.'s suicide attempt, Keller briefly terminated his treatment of A.M., but he ultimately resumed treatment at her mother's request. Contrary to accepted professional practice (given Keller's knowledge of A.M.'s suicide attempt), Keller continued to prescribe A.M. opiates and other controlled substances—including thousands of pills of oxycodone, OxyContin, Carisoprodol, and diazepam—over the next four years. In July 2017, A.M. committed suicide by overdosing on prescription drugs. Some of these drugs, including oxycodone and diazepam, had been prescribed by Keller.

Also in 2017, federal agents began to investigate Keller on suspicion of over-prescribing medications. Besides monitoring Keller's office with a pole camera and utilizing an undercover officer to pose as a potential patient, law enforcement agents executed a search warrant at Keller's personal residence. Among other items, agents seized a journal that contained handwritten notes regarding Keller's “patient information” and “medical information.” Following the raid on his residence, Keller surrendered both his medical license and his DEA license. He also closed his medical practice.

In September 2018, Keller was indicted on various federal charges, including unlawfully dispensing and distributing controlled substances without a legitimate medical purpose under the Controlled Substances Act (“CSA”) (in violation of 21 U.S.C. § 841(a)(l), (b)(l)(C), and (b)(2)) and health care fraud (18 U.S.C. § 1347). In August 2019, Keller was indicted in California state court on charges including second-degree murder (stemming from A.M.'s death) and overprescribing opiates. Keller's state trial—which preceded his federal trial—concluded in March 2020. Keller was acquitted of murder and two drug charges. The jury hung on the other charges.

B. Procedural Background

A superseding federal indictment was filed against Keller in 2021. Keller was charged with ten counts of prescribing various drugs outside the scope of professional practice (in violation of 21 U.S.C. § 841 (a)(1), (b)(1)(C), and (b)(2)) 3 and one count of health care fraud (18 U.S.C. § 1347).

Keller moved to suppress the journal seized during the raid on his residence and requested an evidentiary hearing. The district court denied the motion to suppress the journal 4 and declined to hold an evidentiary hearing. Keller also moved to dismiss the drug charges against him, arguing, among other things, that those counts violated the nondelegation doctrine. The district court also denied that motion.

During the 2022 trial, the government introduced portions of Keller's seized journal. Keller was convicted on four counts of distributing controlled substances outside the scope of professional practice. The jury hung on the remaining six counts of prescribing controlled substances outside the scope of professional practice (and the government opted not to re-try Keller on those counts). The district court granted Keller's motion for judgment of acquittal as to the health care fraud count.

Before sentencing, Keller challenged the Presentence Report (PSR)'s use of a “converted drug weight ratio” of 6,700:1 for oxycodone in calculating his Guidelines range on the ground that the ratio appears not in the Sentencing Guidelines themselves but in commentary appended to the Guidelines. Without this ratio, Keller contends that his Guidelines range would have been 0 to 6 months. Keller also argued that if the ratio were to be applied, Keller should have been found responsible for distributing a “converted” drug weight of 96.51 kilograms (covering only the drugs identified in the counts on which Keller was convicted), rather than the 2,794.5 kilograms that the government argued he should be found responsible for distributing (which covered all the drugs Keller had been charged with distributing, including counts on which Keller was not convicted).

At Keller's sentencing hearing, the district court utilized the challenged ratio in calculating the converted drug weight distributed by Keller but limited its calculations to only those drugs that Keller was convicted of distributing. The district court ultimately adopted a total offense level of 24, which corresponded with a Guidelines range of 51 to 63 months. The court then varied downward and sentenced Keller to 30 months in prison followed by 3 years of supervised release.

Keller now appeals the district court's denial of his motion to suppress the journal seized from his residence; the district court's failure to hold an evidentiary hearing; the district court's denial of his nondelegation doctrine motion to dismiss; and his sentence.

II. JURISDICTION

We have jurisdiction under 28 U.S.C. § 1291 and 18 U.S.C. § 3742(a).

III. STANDARD OF REVIEW

We review de novo the trial court's denial of a motion to suppress. See United States v. Ramirez, 976 F.3d 946, 951 (9th Cir. 2020). “We review for an abuse of discretion a court's decision whether to conduct an evidentiary hearing on a motion to suppress.” United States v. Howell, 231 F.3d 615, 620 (9th Cir. 2000). We review de novo the trial court's denial of a motion to dismiss an indictment on constitutional grounds. See United States v. McCalla, 545 F.3d 750, 753 (9th Cir. 2008). “We review the district court's ․ construction of the United States Sentencing Guidelines de novo, and its application of the Guidelines to the facts for abuse of discretion.” United States v. Harris, 999 F.3d 1233, 1235 (9th Cir. 2021).

IV. DISCUSSION

A. The search warrant covered the seizure of Keller's journal and the supporting affidavit established probable cause sufficient to justify the seizure of the journal.

“Probable cause exists where the totality of the circumstances indicates a ‘fair probability that ․ evidence of a crime will be found in a particular place.’ ” United States v. Elmore, 917 F.3d 1068, 1074 (9th Cir. 2019) (alteration in original) (quoting Illinois v. Gates, 462 U.S. 213, 238 (1983)); see also United States v. Wong, 334 F.3d 831, 836 (9th Cir. 2003) (“Probable cause exists if it would be reasonable to seek the evidence in the place indicated in the affidavit.” (cleaned up) (quoting United States v. Peacock, 761 F.2d 1313, 1315 (9th Cir. 1985), abrogated on other grounds by Gomez v. United States, 490 U.S. 858 (1989))). An “issuing judge's finding of probable cause is entitled to ‘great deference.’ ” Elmore, 917 F.3d at 1074 (quoting United States v. Krupa, 658 F.3d 1174, 1177 (9th Cir. 2011)).

Affidavits filed in support of search warrants must provide “a substantial basis for determining the existence of probable cause.” Gates, 462 U.S. at 239. The DEA agent's affidavit in support of the search warrant on Keller's residence highlighted that Keller prescribed a “surprisingly” high number of controlled substances;  5 received various professional documents at his home address, including both financial and medical licensure documents; and was seen traveling between his home and office carrying a briefcase. The DEA agent also stated in the affidavit that based on her extensive experience with similar investigations, practitioners “often retain personal and business notes, letters, and correspondence relating to their narcotics/prescription orders at their residences.” Thus, both direct surveillance of Keller and the agent's expertise in comparable investigations provided probable cause that documents relevant to the crimes being investigated would be located at Keller's residence. And as the district court found, it is also a “commonsensical” fact “that doctors, and perhaps especially doctors who commit crimes and wish to shield their activities from their colleagues or potential inspectors, bring medical records home.” Accordingly, there was probable cause to believe that the incriminatory documentary evidence described in the warrant might be found at Keller's residence. 6

Keller's journal also fell squarely within the scope of materials subject to seizure. The search warrant expressly authorized the seizure of “[d]ocuments, including but not limited to ․ journals, books, [and] records ․ that refer or relate to ․ the ordering, prescribing, or dispensing of any controlled drug” in the locations to be searched, including Keller's residence (emphasis added). The relevant seized document was immediately identifiable as a journal, which Keller's counsel conceded at oral argument. And while Keller attempts to distinguish between “professional ‘journals and ledgers’ ” and what he describes as his “personal, handwritten diary,” the warrant itself makes no such distinction. The search warrant thus permitted the seizure of journals with information relevant to the crimes being investigated—such as the journal at issue—from Keller's home, regardless of whether Keller also used those journals for recording personal, non-professional information. 7

B. The district court did not abuse its discretion in declining to hold an evidentiary hearing on the admissibility of the journal.

The district court did not conduct an evidentiary hearing regarding the legibility of Keller's journal.  8 Keller now argues that we should, at minimum, “conditionally vacate the judgment and remand for an evidentiary hearing” because of an alleged “material factual dispute” regarding the legibility of Keller's journal.

Keller argues that the district court must hold such a hearing “[w]henever there is a disputed material fact concerning a suppression motion.” For this, Keller invokes our holding in United States v. Mejia, 69 F.3d 309 (9th Cir. 1995), that a district court is “require[d] ․ to conduct an evidentiary hearing when the moving papers filed in connection with a pre-trial suppression motion show that there are contested issues of fact relating to the lawfulness of a search.” Id. at 318. But for moving papers to show that there are contested issues of fact warranting an evidentiary hearing, they must “allege facts with sufficient definiteness, clarity, and specificity to enable the trial court to conclude that contested issues of fact exist.” Howell, 231 F.3d at 620. A simple desire to cross-examine agents that a movant has accused of being untruthful does not itself create grounds for an evidentiary hearing. See Franks v. Delaware, 438 U.S. 154, 171 (1978) (“[T]he challenger's attack must be more than conclusory and must be supported by more than a mere desire to cross-examine.”).

“We review for an abuse of discretion a court's decision whether to conduct an evidentiary hearing on a motion to suppress.” Howell, 231 F.3d at 620. Abuse of discretion is a “highly deferential standard, under which the appellate court cannot substitute its view of what constitutes substantial justification for that of the district court” but must rather merely “assur[e] that the district court's determination has a basis in reason.” Gonzales v. Free Speech Coal., 408 F.3d 613, 618 (9th Cir. 2005) (cleaned up) (quoting Bay Area Peace Navy v. United States, 914 F.2d 1224, 1230 (9th Cir. 1990)). Accordingly, we look to whether the district court had a basis in reason for declining to find that Keller “allege[d] facts with sufficient definiteness, clarity, and specificity to enable the trial court to conclude that contested issues of fact exist” to warrant an evidentiary hearing. Howell, 231 F.3d at 620.

Keller emphasizes that it took law enforcement agents “many months of study to be able to decipher” the journal. But the fact that it took such study to completely decipher the journal does not negate the fact that—as the district court found—relevant target words in the journal were immediately discernible to agents perusing it. The issue of total legibility advanced by Keller is distinct from whether specific words or phrases would have demonstrated to an agent that at least some of the journal's contents related to the prescription of relevant drugs and thus made the journal seizable pursuant to the warrant.

And much of the journal is immediately legible. As the district court found, “patient names and medical terms are discernible in a quick flip through the journal's pages.”  9 Considering the obvious and immediate relevance of such references to the crimes for which Keller was under investigation, Keller's broad allegations of general illegibility did not suffice to create “contested issues of fact” that would have warranted an evidentiary hearing. Id. There were many relevant entries in the journal that were legible to a reader, justifying its seizure pursuant to the warrant. No hearing was necessary to make those determinations. And beyond his unavailing legibility challenge, Keller does not otherwise “allege facts with sufficient definiteness, clarity, and specificity” to demonstrate a dispute over any material fact. Id. The district court thus did not abuse its discretion in not conducting an evidentiary hearing on Keller's journal.

C. Keller's charges did not violate the nondelegation doctrine.

Keller's charges included ten counts of violating the CSA—specifically, 21 U.S.C. § 841(a)(1), (b)(1)(C), and (b)(2). Keller moved to dismiss the indictment, arguing that “the drug counts violate the Constitution's nondelegation doctrine because criminal liability only attaches to prescribing physicians by virtue of a federal regulation promulgated by the Executive Branch, an improper delegation of Congress's power to define crimes with insufficient guidance to the Executive Branch.”  10 The district court rejected Keller's motion.

Keller contends that “the scope of the criminal liability against Keller was defined solely by an administrative regulation, and the statute authorizing its promulgation lacked any intelligible principle.” Keller also argues that it is an “open question” whether something more than an intelligible principle is required when the government promulgates regulations leading to criminal sanctions, and “maintains that more than a showing of a mere intelligible principle is required.”

We hold that the relevant sections of the CSA satisfy the nondelegation doctrine's intelligible principle test. The Attorney General's promulgation of the challenged regulation thus reflects a constitutionally permissible exercise of authority. And under precedent, even when promulgated regulations implicate criminal punishments, the correct standard is the intelligible principle test.

1. The nondelegation doctrine and the intelligible principle test.

The nondelegation doctrine reflects the separation-of-powers principles inherent in the Constitution. The doctrine arises from Article I, which establishes that “[a]ll legislative Powers ․ shall be vested in a Congress of the United States.” U.S. Const. art. I, § 1. Congress “may not transfer to another branch ‘powers which are strictly and exclusively legislative.’ ” Gundy v. United States, 588 U.S. 128, 135 (2019) (plurality opinion) (quoting Wayman v. Southard, 23 U.S. (10 Wheat.) 1, 42–43 (1825)).

But “the nondelegation doctrine does not prevent Congress from seeking assistance, within proper limits, from its coordinate Branches.” Touby v. United States, 500 U.S. 160, 165 (1991). Rather, delegation is permissible so long as Congress articulates an “intelligible principle” by which the actor empowered to exercise delegated authority can conform. Mistretta v. United States, 488 U.S. 361, 372 (1989). The intelligible principle test reflects “a practical understanding that in our increasingly complex society, replete with ever changing and more technical problems, Congress simply cannot do its job absent an ability to delegate power under broad general directives.” Id.

The standards for satisfying this test are “not demanding.” Gundy, 588 U.S. at 146 (plurality opinion). “Only twice in this country's history,” both times in 1935, has the Supreme Court “found a delegation excessive—in each case because Congress had failed to articulate any policy or standard to confine discretion.” Id. (cleaned up) (quoting Mistretta, 488 U.S. at 373 n.7) (citing A.L.A Schechter Poultry Corp. v. United States, 295 U.S. 495 (1935); Panama Refining Co. v. Ryan, 293 U.S. 388 (1935)). A delegation of authority will be upheld so long as Congress delineates “the general policy, the public agency which is to apply it, and the boundaries of th[e] delegated authority.” Am. Power & Light Co. v. SEC, 329 U.S. 90, 105 (1946). “[A] nondelegation inquiry always begins (and often almost ends) with statutory interpretation,” and “ ‘[i]t is a fundamental canon of statutory construction that the words of a statute must be read in their context and with a view to their place in the overall statutory scheme.’ ” Gundy, 588 U.S. at 135, 141 (plurality opinion) (quoting Nat'l Ass'n. of Home Builders v. Defs. of Wildlife, 551 U.S. 644, 666 (2007)).

2. 21 U.S.C. §§ 821 and 871 and 21 C.F.R. § 1306.04.

Under 21 U.S.C. § 821, the Attorney General has the authority to “promulgate rules and regulations ․ relating to the registration and control of the manufacture, distribution, and dispensing of controlled substances.” 21 U.S.C. § 821. Section 871(b) authorizes the Attorney General to “promulgate and enforce any rules, regulations, and procedures which he may deem necessary and appropriate for the efficient execution of his functions under this subchapter.” Id. § 871(b).

Accordingly, the Attorney General has promulgated 21 C.F.R. § 1306.04. That regulation establishes, in part, that for “[a] prescription for a controlled substance to be effective [it] must be issued for a legitimate medical purpose by an individual practitioner acting in the usual course of his professional practice,” and that violators “shall be subject to the penalties provided for violations of the provisions of law relating to controlled substances.” 21 C.F.R. § 1306.04(a).

3. The Attorney General's promulgation of 21 C.F.R. § 1306.04 satisfies the intelligible principle test.

Keller was charged with violations of 21 U.S.C. § 841(a)(1), (b)(1)(C), and (b)(2)—that is, the “provisions of law relating to controlled substances” referenced in 21 C.F.R. § 1306.04(a)—based on his distribution of controlled substances “outside the scope of professional practice.” Keller was subject to the penalties of the CSA (a criminal statute) because of his failure to comply with § 1306.04(a) (a regulation promulgated by the executive branch pursuant to 21 U.S.C. §§ 821 and 871(b)).

In determining the constitutionality of delegated authority, we look to whether Congress has clearly articulated a general policy by which implementing agencies or officials may abide. See Am. Power, 329 U.S. at 105. The general policy outlined within the CSA is “the efficient execution” of the Attorney General's “functions under” that statute. 21 U.S.C. § 871(b). Those functions are defined as “control[ling] ․ the manufacture, distribution, and dispensing of controlled substances.” Id. § 821 (emphasis added); see also Oregon v. Ashcroft, 368 F.3d 1118, 1129 (9th Cir. 2004) (identifying the CSA's “mandate” as “combat[ting] prescription drug abuse and addiction”), aff'd sub nom. Gonzales v. Oregon, 546 U.S. 243 (2006). This kind of mandate—limited in terms of the subject area in which delegated authority may be exercised, yet broad as to how the delegated authority may be exercised—is the type of delegation of authority long upheld as permissible. See, e.g., Whitman v. Am. Trucking Ass'n, 531 U.S. 457, 473 (2001) (allowing the Environmental Protection Agency to “establish uniform national [air pollution] standards at a level that is requisite to protect public health”); Nat'l Broad. Co. v. United States, 319 U.S. 190, 215, 225–26 (1943) (allowing the Federal Communications Commission to regulate broadcast licensing as the “public interest, convenience, or necessity” require); see also Loving v. United States, 517 U.S. 748, 771 (1996) (“[Since 1935,] we have ․ upheld, without exception, delegations under standards phrased in sweeping terms.”).

We also look to whether Congress has designated an agency to exercise the relevant authority. See Am. Power, 329 U.S. at 105. The CSA is unambiguous in granting such authority to the Attorney General. 21 U.S.C. §§ 821, 871.

The test also requires that we analyze whether Congress has “clearly delineate[d] ․ the boundaries of [the] delegated authority.” See Am. Power, 329 U.S. at 105. The scope of the authority delegated by 21 U.S.C. §§ 821 and 871(b) is both clear and bounded: the Attorney General is authorized to promulgate rules and regulations “relating to the registration and control of the manufacture, distribution, and dispensing of controlled substances and to listed chemicals,” 21 U.S.C. § 821, and to promulgate rules in support of the execution of his functions under the CSA, see id. § 871(b).

Furthermore, Section 841 makes it “unlawful for any person knowingly or intentionally” to “distribute ․ a controlled substance,” “[e]xcept as authorized by [the CSA].” 21 U.S.C. § 841(a)(1). The CSA permits prescriptions “by a practitioner” for controlled substances, id. § 829(a)–(b), and the statute defines a “practitioner” as a “physician” authorized “to distribute ․ a controlled substance in the course of professional practice,” id. § 802(21) (emphasis added). The CSA also defines a “valid prescription” as “a prescription that is issued for a legitimate medical purpose in the usual course of professional practice.” 21 U.S.C. § 829(e)(2)(A) (emphasis added); see also id. § 830(b)(3)(A)(ii) (defining a “valid prescription” as “a prescription which is issued for a legitimate medical purpose by an individual practitioner ․ acting in the usual course of the practitioner's professional practice”).

The text of 21 C.F.R. § 1306.04 borrows directly from the CSA in declaring that for “[a] prescription for a controlled substance to be effective” it “must be issued for a legitimate medical purpose by an individual practitioner acting in the usual course of his professional practice.” 21 C.F.R. § 1306.04(a); see 21 U.S.C. §§ 802(21), 802(56)(C), 829(e)(2)(A), 830(b)(3)(A)(ii). This reinforces our conclusion that section 1306.04 operates within the limits of delegated authority. It delineates the standards by which a “practitioner” must abide when prescribing controlled substances and allows for criminal prosecution (pursuant to the CSA) of a physician's failure to abide by “professional practice[s].” 21 C.F.R. § 1306.04(a). It neither extends beyond, nor conflicts with, the authority to (1) regulate controlled substances and (2) enforce the provisions of the CSA delegated to the Attorney General by 21 U.S.C. §§ 821 and 871(b).

Thus, the requirements for the delegation of authority to be constitutional are met. See Am. Power, 329 U.S. at 105. Those sections of the CSA that authorize the promulgation and enforcement of § 1306.04 establish a clear general policy; the implementing agency (or here, official—the Attorney General) is specifically identified; and the boundaries of the authority delegated to the Attorney General under the CSA are clearly outlined. Section 1306.04 operates within the boundaries of the authority delegated to the Attorney General under the CSA.

4. No heightened standard beyond the intelligible principle test is required for delegations of authority implicating criminal sanctions.

Keller claims that it “is an open question” whether Congress must provide guidance beyond an intelligible principle when dealing with the criminal law and criminal penalties. Even assuming that was true at the time of briefing and argument, in United States v. Pheasant, 129 F.4th 576 (9th Cir. 2025), we held that “[e]ven in the criminal context, the ‘intelligible principle’ test provides the controlling legal standard for evaluating non-delegation challenges.” Id. at 583. Keller's argument that regulations relating to criminal penalties require something greater than an intelligible principle is thereby foreclosed by our precedent. And because § 1306.04 satisfies the intelligible principle test, Keller's challenge fails.

D. The district court properly calculated Keller's Sentencing Guidelines range.

1. Keller's sentence under the drug conversion ratio.

At sentencing, the district court agreed with Keller that the conduct forming the basis of his conviction limited the relevant drug quantities for sentencing purposes. This consisted of 14.4 grams of oxycodone, 84 grams of carisoprodol, and 2.4 grams of diazepam, before applying any drug weight conversion factors. As relevant here, § 2D1.1(a)(5) of the Guidelines directs that the offense level for illegally distributed controlled substances is “specified in the Drug Quantity Table set forth in subsection (c).” In turn, § 2D1.1(c) lists various controlled substances (e.g., heroin, cocaine, methamphetamine) and then provides a base offense level for different quantities of each substance. For example, 90 kilograms or more of heroin leads to a base offense level of 38.

Some controlled substances, including oxycodone, are not specifically mentioned in § 2D1.1(c). To determine the base offense level for these substances, § 2D1.1(c) provides a residual category based on “Converted Drug Weight.” Thus, for example, for a base offense level of 26, the defendant must be responsible for “[a]t least 400 KG but less than 700 KG of Converted Drug Weight.” U.S.S.G. § 2D1.1(c)(7). Determining the proper converted weight, and thus base offense level, for controlled substances that are not specifically listed in § 2D1.1(c) requires further consideration of both § 2D1.1(c) and Guidelines commentary that it specifically incorporates.

At the top of § 2D1.1(c)'s Drug Quantity Table, an asterisk directs the reader to a lower portion of text entitled “Notes to Drug Quantity Table,” which sets forth various definitions. Importantly, these Notes are part of the text of § 2D1.1(c) itself. Subsection (B) of these Notes confirms that oxycodone is among the controlled substances covered by § 2D1.1(c), because it provides that “[t]he terms ‘Hydrocodone (actual)’ and ‘Oxycodone (actual)’ refer to the weight of the controlled substance, itself, contained in the pill, capsule, or mixture.” Subsection (K) of the Notes then explains how to calculate the “Converted Drug Weight” for controlled substances like oxycodone that are not specifically listed in § 2D1.1(c). Note (K) states:

The term “Converted Drug Weight,” for purposes of this guideline, refers to a nominal reference designation that is used as a conversion factor in the Drug Conversion Tables set forth in the Commentary below, to determine the offense level for controlled substances that are not specifically referenced in the Drug Quantity Table or when combining differing controlled substances.

U.S.S.G. § 2D1.1(c), Note (K). Note (K) thus directs the reader to Drug Conversion Tables in the Guidelines commentary, which are found in the commentary's Application Note 8(D). And Application Note 8(D) provides that 1 gram of oxycodone has a converted drug weight of 6,700 grams. U.S.S.G. § 2D1.1 cmt. n.8(D).

In Keller's case, his 14.4 grams of oxycodone, multiplied by a 6,700 conversion factor, led to a converted drug weight of approximately 96 kilograms (the other substances did not materially increase the converted drug weight based on the applicable formulae). This placed Keller at a base offense level of 22 under U.S.S.G. § 2D1.1(c)(9). The addition of a 2-level increase under U.S.S.G. § 3B1.1 for abusing a position of trust increased the base offense level to 24, which together with Keller's criminal history score led to a Guidelines range of 51–63 months. As discussed above, the district court sentenced Keller to 30 months in prison.

Keller argues that the district court erred in calculating his sentencing range by relying on the 6,700:1 conversion ratio for oxycodone because the ratio appears in the Sentencing Guidelines commentary and not in the text of the Guidelines itself. See U.S.S.G. § 2D1.1(c) & cmt. n.8(D). Keller argues that this reliance on the commentary is no longer proper after the Supreme Court's decision in Kisor v. Wilkie, 588 U.S. 558 (2019), and this court's recent decision in United States v. Castillo, 69 F.4th 648 (9th Cir. 2023). In Keller's view, the commentary impermissibly expands upon the Guidelines. And if the commentary's conversion ratio is invalid, Keller says (and the government does not dispute) that the applicable Guidelines range would instead be 0–6 months imprisonment.

2. Any error in calculating Keller's sentence was not harmless.

As an initial matter, we agree with Keller that if the district court erred in relying on the oxycodone conversion ratio, that error would not be harmless. “To establish harmlessness, the Government must show that ‘it is more probable than not’ that the error did not affect the sentence.” United States v. Dominguez-Caicedo, 40 F.4th 938, 963 (9th Cir. 2022) (quoting United States v. Morales, 108 F.3d 1031, 1040 (9th Cir. 1997) (en banc)). And “[i]n most cases a defendant who has shown that the district court mistakenly deemed applicable an incorrect, higher Guidelines range has demonstrated a reasonable probability of a different outcome.” Molina-Martinez v. United States, 578 U.S. 189, 200 (2016).

In this case, the district court relied on the converted drug weight ratio to calculate a sentencing range of 51–63 months before departing downward from that starting point to impose a sentence of 30 months' imprisonment. At sentencing the district court, referencing a Guideline range of 51 to 63 months, stated: “[t]hat is the Guideline calculation adopted in this case.” Later, at the end of the hearing, the court reiterated that “the Guideline range that I've adopted is 51 to 63 months,” but that it was “appropriate to vary downward from the Guideline range, and[,] applying all of the 3553(a) factors[,] ․ to impose a sentence of 30 months in prison.” The government does not challenge Keller's contention that without the converted drug weight ratio, his Guidelines range would have been 0–6 months. Under these circumstances, we do not find it “more probable than not” that the district court's use of the conversion ratio had no bearing on Keller's sentence. See Dominguez-Caicedo, 40 F.4th at 963 (quoting Morales, 108 F.3d at 1040). We do not think it is probable that if the applicable range were 0–6 months, the district court would have still sentenced Keller to 30 months in prison—which would be five times the upper end of the range. In fact, this is highly improbable, given that the district court found it “appropriate to vary downward from the Guideline range” and imposed a sentence well below its calculated range.

The concurrence argues that the district court considered the alternative 0–6 month range when it stated that “it is very difficult to imagine a non-custodial sentence in this case, applying all of the 3553(a) factors.” But this statement does not shed light on what Guidelines range the court had in mind before applying the § 3553(a) factors, and we know the Guidelines range it had calculated was 51–63 months. There is likewise no basis to conclude that the district court's skepticism towards “a non-custodial sentence” suggests that the court had in mind a Guidelines range of 0-6 months as a starting point, from which it deviated upwards to impose a 30-month sentence. The district court plainly believed that a custodial sentence was warranted, but we cannot conclude that the court would have selected a 30-month sentence if the Guidelines range were 0–6 months.

The government briefly argues that the district court would have imposed the same sentence even if the Guidelines range had been different, pointing to the court's statement that “even if I did disagree with the drug amount, I would still not impose a sentence of less than 30 months in this case considering all of the 3553(a) factors.” The concurrence also argues that this language indicates that the court implicitly considered the 0–6 month sentencing range. But in context, the district court was not referring to the possibility that the commentary reflected an impermissible interpretation of the Guideline under Kisor. Instead, based on the colloquy during the hearing, the court was referring to the possibility that it could deviate from the advisory Guidelines based on a policy disagreement with the ratio. See Kimbrough v. United States, 552 U.S. 85, 101 (2007) (“The Government acknowledges that the Guidelines ‘are now advisory’ and that, as a general matter, ‘courts may vary [from Guidelines ranges] based solely on policy considerations, including disagreements with the Guidelines.’ ” (alteration in original) (quoting Rita v. United States, 551 U.S. 338, 351 (2007))). Thus, the district court did not address the Kisor issue or state that it would have imposed the same sentence even if the Guidelines range had been 0–6 months rather than 51–63 months.

In any case, even if we agreed with the government's reading of the record, the statement in question still would not establish harmlessness. “A ‘district court's mere statement that it would impose the same ․ sentence no matter what the correct calculation cannot, without more, insulate the sentence from remand.’ ” Dominguez-Caicedo, 40 F.4th at 963 (ellipsis in original) (quoting United States v. Munoz-Camarena, 631 F.3d 1028, 1031 (9th Cir. 2011) (per curiam)). “This is because a district court's analysis must ‘flow from an initial determination of the correct Guidelines range,’ and the district court must keep that range ‘in mind throughout the process.’ ” Id. (citation omitted) (quoting Munoz-Camarena, 631 F.3d at 1030–31). For this reason as well, any error in calculating the Guidelines range cannot be regarded as harmless.

The concurrence argues that the district court's statement was sufficient to establish harmlessness because it identified the Guidelines calculation as only “part of the mix of deciding what the appropriate sentence should be,” and repeatedly justified a custodial sentence under the § 3553(a) factors. The district court's consideration of the § 3553(a) factors is insufficient under our precedent. We have found a Guidelines calculation error not harmless even where a court expressly states that “based on [the] 3553(a) factors” it would impose the same sentence regardless of the Guidelines range, because such conclusory statements “do[ ] not demonstrate that the district court conducted the sentencing a second time starting with the correct range and keeping it in mind throughout the process.” Dominguez-Caicedo, 40 F.4th at 964 (first alteration in original).

3. The distinction between the Sentencing Guidelines and the commentary.

We thus must resolve Keller's sentencing challenge on the merits. This challenge rests on the distinction between the Sentencing Guidelines and their accompanying commentary, and the reach of relatively recent case law addressing when courts may rely on Guidelines commentary.

The Sentencing Reform Act of 1984, 18 U.S.C. § 3551 et seq., created the Sentencing Commission, which “promulgate[d] the guidelines by virtue of an express congressional delegation of authority for rulemaking.” Castillo, 69 F.4th at 655 (alteration in original) (quoting Stinson v. United States, 508 U.S. 36, 44–45 (1993)). The Guidelines must be reviewed by Congress, and “cannot become effective until after [a] 6-month review period for congressional modification or disapproval.” Stinson, 508 U.S. at 44; see also 28 U.S.C. § 994(p).

The Guidelines commentary is also prepared by the Sentencing Commission, see Stinson, 508 U.S. at 44–45, and may help “interpret the guideline[s] or explain how [they are] to be applied,” U.S.S.G. § 1B1.7. The “commentary is not subject to mandatory congressional review,” and has been analogized to “an agency's interpretation of its own legislative rule.” Castillo, 69 F.4th at 655 (quoting Stinson, 508 U.S. at 44).

In Stinson v. United States, the Supreme Court held that “commentary in the Guidelines Manual that interprets or explains a guideline is authoritative unless it violates the Constitution or a federal statute, or is inconsistent with, or a plainly erroneous reading of, that guideline.” 508 U.S. at 38. Relying on Bowles v. Seminole Rock & Sand Co., 325 U.S. 410 (1945), Stinson described the guidelines as “the equivalent of legislative rules adopted by federal agencies,” explaining that “this type of commentary is akin to an agency's interpretation of its own legislative rules” and should generally be given controlling weight. 508 U.S. at 45 (citing Seminole Rock, 325 U.S. at 414). This level of deference was referred to as “Seminole Rock deference,” or later, as “Auer deference.” See Auer v. Robbins, 519 U.S. 452, 461 (1997); see also Kisor, 588 U.S. at 563. Beyond Seminole Rock deference, Stinson further advised that the commentary in the Sentencing Guidelines “provides concrete guidance as to how even unambiguous guidelines are to be applied in practice.” 508 U.S. at 44.

In Kisor, however, the Supreme Court narrowed the deference afforded to an agency's interpretation of its own regulations, holding that such deference is proper only if the regulation is “genuinely ambiguous,” which is determined by “exhaust[ing] all the ‘traditional tools’ of construction” and examining the regulation's “text, structure, history, and purpose.” Kisor, 588 U.S at 574–75 (quoting Chevron U.S.A. Inc. v. Nat. Res. Def. Council, Inc., 467 U.S. 837, 843 n.9 (1984)). If a regulation is shown to be genuinely ambiguous, a court may not defer to the agency's interpretation unless the court independently determines that the interpretation is “reasonable,” i.e., that it “come[s] within the zone of ambiguity the court has identified after employing all its interpretive tools.” Id. at 575–76 (quoting Thomas Jefferson Univ. v. Shalala, 512 U.S. 504, 515 (1994)).

Following Kisor, the question arose as to whether the Kisor standard should replace Stinson's more deferential approach to the Guidelines commentary. We recently answered that question in Castillo, holding that “[t]he more demanding deference standard articulated in Kisor applies to the Guidelines' commentary.” 69 F.4th at 655. We explained that because “Stinson deference is directly grounded in Seminole Rock and Auer deference[,] ․ ‘the only way to harmonize Kisor and Stinson is to conclude that Kisor's gloss on Auer and Seminole Rock applies to Stinson.’ ” Id. at 656 (quoting United States v. Dupree, 57 F.4th 1269, 1275 (11th Cir. 2023) (en banc) (brackets omitted); see also United States v. Scheu, 83 F.4th 1124, 1127–28 (9th Cir. 2023) (tracking the development in the law from Stinson to Castillo).

4. The district court did not err in relying on the 6,700:1 converted drug weight ratio for oxycodone.

Keller argues that the district court could not defer to the 6,700:1 ratio in Application Note 8 because the ratio is not a reasonable interpretation of an ambiguous Guideline under Kisor and Castillo. But we conclude that we need not and should not address whether the ratio is a permissible interpretation of the Guideline under Kisor because, in assessing the weight to be given to the commentary, the converted drug weight ratio for oxycodone is more properly regarded as part of the Guideline itself.

As we explained above, the Guideline covers oxycodone, U.S.S.G. § 2D1.1(c), Note B, it provides base offense levels for “Converted Drug Weight”, id. § 2D1.1(c), and it expressly directs that “ ‘Converted Drug Weight,’ for purposes of this guideline, refers to a nominal reference designation that is used as a conversion factor in the Drug Conversion Tables set forth in the Commentary below,” id. § 2D1.1(c), Note K. As drafted, the ratios in the drug conversion tables are effectively part of § 2D1.1 itself because this section of the Guidelines expressly incorporates them.

Furthermore, and critically, although the “commentary is not subject to mandatory congressional review,” Castillo, 69 F.4th at 655, the 6,700:1 oxycodone ratio underwent the same notice-and-comment and congressional-review process as the Guideline itself. The Sentencing Commission submitted this ratio to Congress for review on May 1, 2003, as part of Amendment 657. Addendum-19–22 (2003 Amendments to the Sentencing Guidelines, at 57–58, available at https://www.ussc.gov/sites/default/files/pdf/amendment-process/reader-friendly-amendments/20030512_RF_Amendments_0.pdf). Congress did not act, and the ratio went into effect on November 1, 2003. Id.; see also United States v. Landron-Class, 696 F.3d 62, 75–76 (1st Cir. 2012). Because the Commission expressly incorporated the Drug Conversion Tables into the text of § 2D1.1(c) and adopted them using the procedures required for enacting the Guidelines themselves, we believe the better view is that the Drug Conversion Tables should be regarded as part of the Guidelines, and no further Kisor inquiry is required before a district court may apply the drug ratios.

This case thus presents a different situation from Castillo. There, we considered whether a defendant's conviction for conspiracy to distribute methamphetamine was a “controlled substance offense” under U.S.S.G. § 4B1.2(b), which in turn qualified the defendant as a career offender under U.S.S.G. § 4B1.1. 69 F.4th at 653. While the Sentencing Guidelines at the time did not define “controlled substance offense” to “include inchoate crimes like conspiracies,” the “commentary extend[ed] the definition to such crimes.” Id. In Castillo's case, the career offender enhancement substantially elevated his advisory Guidelines range. Id. at 651.

Applying Kisor, Castillo held that the commentary in question should not be followed. We explained that “[a]pplying the traditional tools of statutory construction to the text of the guideline, as Kisor instructs, ․ § 4B1.2(b) unambiguously identifies a list of crimes that does not include inchoate offenses.” Id. at 657–58. Because the Guideline itself was “unambiguous, the Supreme Court's decision in Kisor now ma[de] it impermissible to defer” to the application note.  11 Id. at 663. This case is different because here, unlike in Castillo, the Guideline itself expressly incorporates the commentary, a relatively rare occurrence in the Guidelines.

Keller protests that no prior case has treated commentary as part of a Guideline and thus exempt from a further Kisor inquiry. But when Stinson governed, the distinction between the Guidelines and commentary was not especially significant. To the extent today's decision is novel, that is only because the question we confront is itself novel, now that Kisor applies. See Castillo, 69 F.4th at 655.

Keller also suggests that our approach has no “limiting principle” because the Sentencing Commission could promulgate a guideline that defines the offense level for a particular offense by reference to the commentary, and then rely entirely on ever-changing commentary to determine the offense level. Keller is incorrect. The limiting principle is that the commentary incorporated by reference in the Guidelines must itself have undergone notice and comment and congressional review, as the commentary in Amendment 657 did. The commentary, in other words, must be enacted as if it were an express part of a Guideline. If the Sentencing Commission had changed the oxycodone ratio to 10,000:1 without subjecting it to notice and comment and congressional review, we do not think it could be properly regarded as part of the Guideline itself, even if it purported to incorporate it. At that point, a Kisor inquiry would be required.

It is true that in Castillo we found that Kisor deference applied even though the commentary at issue there had also undergone notice and comment. Castillo, 69 F.4th at 652; see also United States v. Dupree, 57 F.4th 1269, 1281 (11th Cir. 2023) (en banc) (Pryor, C.J., concurring). But this was so because the commentary at issue in that case “expand[ed] unambiguous Guidelines.” Castillo, 69 F.4th at 664. By contrast, Application Note 8 does not expand unambiguous Guidelines. Rather, it was expressly incorporated into the Guidelines.

The Commission could have moved the Drug Conversion Tables into the Guideline by pasting them into that part of the text. The Commission has undertaken revisions along these lines in recent years, as courts have grappled with whether the commentary is an appropriate interpretation of the Guidelines under Kisor. See, e.g., Amendment 822 to U.S.S.G. § 4B1.2, effective Nov. 1, 2023. Although the Commission could do that here for the Drug Conversion Tables, requiring it to do so would be unduly formalistic when the Commission has already expressly incorporated the commentary into the Guidelines and put it through the same notice and comment and congressional review process as the Guidelines themselves. Therefore, Kisor is not implicated, and we find no error by the district court.

* * *

For the foregoing reasons, Keller's conviction and sentence are

AFFIRMED.

I believe that any alleged error in sentencing was harmless. I would accordingly affirm Thomas Keller's sentence on the basis of harmless error. I would not reach whether the relevant Sentencing Guidelines commentary was incorporated into the text of the Guidelines itself. I join the rest of the per curiam opinion.

If a district court utilizes an inappropriate Sentencing Guidelines range, that error is subject to harmless-error review. See United States v. Prigan, 8 F.4th 1115, 1122 (9th Cir. 2021). “Errors that impact Guideline calculations typically require remand unless the Government establishes the error was harmless.” United States v. Klensch, 87 F.4th 1159, 1166 (9th Cir. 2023).

For harmlessness, “[a] district court's analysis must ‘flow from an initial determination of the correct Guidelines range,’ and the district court must keep that range ‘in mind throughout the process.’ ” United States v. Dominguez-Caicedo, 40 F.4th 938, 963 (9th Cir. 2022) (citation omitted) (quoting United States v. Munoz-Camarena, 631 F.3d 1028, 1030, 1031 (9th Cir. 2011) (per curiam)). In the majority's understanding, the district court failed to make that determination because it “did not ․ state that it would have imposed the same sentence even if the Guidelines range had been 0–6 months rather than 51–63 months.” Op. at 27–28. I disagree that the district court did not consider the allegedly correct range.

The district court did not explicitly reference the potential 0–6 month Guidelines range. But I believe the district court did consider that range. As the district court stated, “it is very difficult to imagine a non-custodial sentence in this case, applying all of the [§] 3553(a) factors.” This demonstrates consideration of the 0–6 month range: that range was the only one that would have entailed a noncustodial sentence, and it was expressly rejected by the district court. I do not believe the district court must engage in a formulaic recitation of the allegedly correct Guidelines range to satisfy the requirement that it keep such range “in mind” during sentencing proceedings. Dominguez-Caicedo, 40 F.4th at 963 (quoting Munoz-Camarena, 631 F.3d at 1030). And the district court also stated: “But even if I did disagree with the drug amount, I would still not impose a sentence of less than 30 months in this case considering all of the [§] 3553(a) factors.”

Of course, even when the district court considers but rejects the allegedly correct Guidelines range, “[a] ‘district court's mere statement that it would impose the same ․ sentence no matter what the correct calculation cannot, without more, insulate the sentence from remand.’ ” Id. (ellipsis in original) (emphasis added) (quoting Munoz-Camarena, 631 F.3d at 1031). Rather, “[f]or the district court's calculation error to be harmless, the district court ‘must explain, among other things, the reason for the extent of a variance’ from the correct Guidelines range.” Prigan, 8 F.4th at 1122 (quoting Munoz-Camarena, 631 F.3d at 1031). Here, there was “more,” and in my view significantly more.

First, the court identified the Guidelines calculation as “just ․ part of the mix of deciding what the appropriate sentence should be.” Second, the district court repeatedly noted that its justification for a custodial sentence relied on its consideration of various alternative factors under 18 U.S.C. § 3553(a): “[A]pplying the [§] 3553(a) factors, the idea of a non-custodial sentence in this case, it seems virtually impossible”; “it is very difficult to imagine a non-custodial sentence in this case, applying all of the [§] 3553(a) factors”; “you do need to think about general deterrence in this case”; “you need to think about ․ what would be a just punishment for Mr. Keller”; “I think that, you know, applying all of the [§] 3553(a) factors, this is absolutely a case where the Defendant must go to prison.” See 18 U.S.C. § 3553(a)(1), (a)(2)(A), (a)(2)(B).

To establish harmlessness, the government ultimately must show that “ ‘it is more probable than not’ that [any] error did not affect the sentence.” Dominguez-Caicedo, 40 F.4th at 963 (quoting United States v. Morales, 108 F.3d 1031, 1040 (9th Cir. 1997) (en banc)). The majority finds it “highly improbable” “that if the applicable range were 0–6 months, the district court would have still sentenced Keller to 30 months in prison.” Op. at 26. But I disagree. As the record shows, the district court acknowledged the possibility of a noncustodial, 0–6 month sentence under the Guidelines; rejected that possibility;  1 and identified many factors justifying a custodial sentence. These included deterrence, 18 U.S.C. § 3553(a)(2)(B); “just punishment,” id. § 3553(a)(2)(A); and “the nature and circumstances of the offense and the history and characteristics of the defendant,” id. § 3553(a)(1). For these reasons, I believe it is “more probable than not” that even if the district court employed an incorrect Guidelines range, any such error was harmless, and accordingly would hold as much. Dominguez-Caicedo, 40 F.4th at 963 (quoting Morales, 108 F.3d at 1040).

FOOTNOTES

1.   Keller is no longer incarcerated. Keller's release from custody does not, however, moot his sentencing claim, given that he is still subject to the terms of supervised release. See United States v. Verdin, 243 F.3d 1174, 1178 (9th Cir. 2001).

2.   Because we find that the journal was properly seized pursuant to the search warrant, we do not reach whether the seizure was alternatively justified under the plain view doctrine.

3.   21 U.S.C. § 841(a)(1) criminalizes the act of “knowingly or intentionally ․ manufactur[ing], distribut[ing], or dispens[ing], or possess[ing] with intent to manufacture, distribute, or dispense, a controlled substance.” Subsections 841(b)(1)(C) and 841(b)(2) delineate penalties specific to the scheduling level of the drugs at issue.

4.   The district court granted the motion to suppress as to other seized items.

5.   The search warrant affidavit contained detailed descriptions of Keller's extraordinarily high numbers of controlled substance prescriptions. For example:CURES also revealed the top five products prescribed by KELLER were: oxycodone, hydrocodone, methadone, morphine sulfate and hydromorphone. All of these top five produced are Schedule II controlled substances. Of the 5,230 prescriptions that KELLER wrote for controlled substances, 3,588 were for Schedule II controlled substances. In my training and experience, that is an extraordinarily high percentage of Schedule II controlled substances․The CURES data also shows that KELLER prescribes a large amount of controlled substances in a combination known as “The Holy Trinity”—a drug regimen that includes at least one opioid, a benzodiazepine, and Carisoprodol for one patient. In my training and experience, the Holy Trinity is frequently sought by opiate addicts because that combination enhances the high that they experience. However, based on my training and experience, and my consultation with medical experts, there is no legitimate medical reason to prescribe this combination of drugs, which carries enhanced risk of drowsiness, respiratory depression, confusion, tremor, and seizure.

6.   The warrant provided:The items to be seized are evidence, contraband, fruits, or instrumentalities of violations of 21 U.S.C. §§ 841(a)(l) and 846 (conspiracy to distribute and distribution of controlled substances); 21 U.S.C. § 843(a)(3) (acquiring a controlled substance by misrepresentation, fraud, forgery or deception[)]; 18 U.S.C. §§ 1347, 1349 (health care fraud)[;] 42 U.S.C. § l320a-7b(b) (payment, solicitation, and offering of remuneration for the receipt of a benefit from a federal health care program) ․

7.   Keller made a similar argument regarding other documents that were seized, which the district court accepted in partially granting the motion to suppress:Keller argues that ․ letters to his sister were outside the scope of the warrant and therefore not properly seized. The letters were outside the scope of the warrant. Even a quick skim shows they do not relate to “(1) the prescribing, dispensing, or other distribution of any controlled drug or to any person to whom a controlled substance was prescribed or dispensed; (2) the submission of any billing to Medicare; [ ] (3) the receipt of payment of any compensation in exchange for the act of writing or filling a prescription to a Medicare beneficiary” or any other category of documents listed in Attachment B of the warrant application.

8.   Keller claims “legibility” was important because if the journal were illegible, it would have fallen outside the scope of the search warrant.

9.   These include references to “Suboxone,” “Suboxone + Norco,” “Narcotics,” “A**** M****! [A.M.'s full name, capitalized],” “FU**ING PSYCHO DRUGGIE,” and the phrase “see me as a legal drug dealer.”

10.   Keller also argued below, and reiterates here, that the superseding indictment improperly charged him with “distributing” controlled substances, rather than “dispensing” controlled substances as prohibited by the CSA. The implication of this argument is that Keller's indictment was facially defective for not charging a proper offense. As Keller concedes, however, this court has held that “[i]f the prescription is not lawful, the ‘practitioner’ does not dispense; rather ․ he ‘distributes.’ ” United States v. Black, 512 F.2d 864, 866 (9th Cir. 1975). Certain other circuits have found that it is possible to unlawfully “dispense” controlled substances. See, e.g., United States v. Azmat, 805 F.3d 1018, 1034 (11th Cir. 2015); id. at 1032–34 (collecting cases). Keller “preserves for possible en banc or certiorari review his contention that a licensed physician alleged to have unlawfully prescribed controlled substances has not, as a matter of law, ‘distributed’ controlled substances, as charged in the superseding indictment.” Given that this panel is bound by circuit precedent, we do not address the propriety of the superseding indictment's use of “distributed” instead of “dispensed.”

11.   The Sentencing Commission has since moved the inchoate offense commentary into the Guideline itself. See Amendment 822 to U.S.S.G. § 4B1.2, effective Nov. 1, 2023; see also U.S.S.G. § 4B1.2(d).

1.   And again, the district court made multiple statements about exactly why a noncustodial sentence was essentially “impossible” and why a sentence of 30 months was the absolute minimum. Among the factors the court focused on were A.M.'s tragic death and the fact that “[i]f Mr. Keller had exercised any restraint in how he prescribed medicine, how he prescribed drugs to [A.M.], we might not be here.” The district court also emphasized that “[Keller's] journal clearly reflect[ed] that he did not care what was going to happen to [A.M.] or ․ any number of other patients.”

PER CURIAM: