Nos. 24-5384/5385

Appeal from the United States District Court for the Eastern District of Kentucky at Covington. No. 2:21-cr-00002—David L. Bunning, District Judge.

Argued: June 11, 2025

Decided and Filed: November 24, 2025

Before: GILMAN, DAVIS, and MATHIS, Circuit Judges.

ARGUED: Michael Ferrara, DINSMORE & SHOHL LLP, Columbus, Ohio, for Appellant Siefert. Ronald W. Chapman II, CHAPMAN LAW GROUP, Troy, Michigan, for Appellant Ehn. Dermot Lynch, UNITED STATES DEPARTMENT OF JUSTICE, Washington, D.C., for Appellee. ON BRIEF: Michael Ferrara, Lindsay Gerdes, DINSMORE & SHOHL LLP, Columbus, Ohio, for Appellant Siefert. Ronald W. Chapman II, CHAPMAN LAW GROUP, Troy, Michigan, for Appellant Ehn. Dermot Lynch, UNITED STATES DEPARTMENT OF

DAVIS, Circuit Judge. Timothy Ehn and William Siefert executed a health-care-fraud scheme at the heart of the opioid epidemic. Ehn owned and operated a pain clinic in northern Page 2Kentucky. With the help of Siefert, Ehn created a testing scheme that involved using patients’ medical conditions as a front to conduct specialized urine drug tests that were billed at a higher rate than the routine tests most patients needed. This scheme began in 2017 and lasted, for Siefert, until he stopped working at the clinic in 2019 and, for Ehn, until 2021. Both Ehn and Siefert were indicted for conspiracy to distribute a controlled substance, health care fraud, and conspiracy to commit health care fraud. Siefert was also indicted for unlawful distribution of a controlled substance. A jury ultimately convicted Siefert of health care fraud and Ehn of health care fraud and conspiracy to commit the same. Both defendants appeal those convictions on various grounds and challenge the procedural reasonableness of their sentences. For the following reasons, we AFFIRM.

Dr. Timothy Ehn, a chiropractor, owned the Northern Kentucky Center for Pain Relief (the 'Clinic') and operated its chiropractic wing. The Clinic serviced referral-only patients, providing them with multiple modalities of treatment, including chiropractic, pain-management, and injection therapy. And the Clinic offered more medications than just opioids—patients were prescribed “helper medicines” to decrease their pain. (Trial Tr. Vol. 9, R. 247, PageID 6856). Ehn hired Dr. William Siefert, a medical doctor, in August 2014 to run the medical side of the Clinic and to serve as medical director.

As part of its pain-management practice, the Clinic's doctors used urine drug testing ('UDT') to ensure that patients took their medication as prescribed and were not taking medications or substances that could potentially interact with prescribed opioids. The Clinic used two types of testing. The first type—presumptive UDT—provides a positive or negative result to indicate the presence or absence of a particular drug in the urine. The second type— definitive testing—identifies the concentration of the tested-for drugs in a patient’s system.

The Clinic’s physicians enrolled in multiple insurance programs to serve their patients. Some insurers, like Medicare, required the enrolled physicians to abide by Medicare laws, regulations, and program instructions to receive payment for their services. So, to legitimately Page 3receive reimbursement for UDT services, Siefert and Ehn had to abide by the relevant Medicare guidelines for UDT. These guidelines, called local coverage determinations ('LCDs'), are evidence-based policies written and implemented by the local Medicare contractor. Relevant here, LCDs outline when Medicare will reimburse for UDT.

Medicare regularly reimburses for presumptive UDT because those tests may be routinely ordered. But definitive tests must be run on complex equipment, making them more costly to carry out. As a result, they are reimbursed at a much higher rate than presumptive UDT. And because definitive UDT can test for multiple classes of drugs at once, the tests are more expensive the more classes of drugs tested. So, Medicare requires documentation justifying the medical necessity for definitive tests before it will reimburse. This means that physicians must document why the services were provided for each patient on the date of each visit, making an individualized assessment of need. Physicians must determine a patient's likelihood of straying from his or her pain-management protocol; the more likely a patient is to deviate, the more often he will need a definitive UDT. A key limitation is that physicians cannot implement a blanket policy to always order both a presumptive and definitive test—a physician must determine that the definitive test is medically necessary to use in treating the patient before ordering it. This includes ensuring that each class of drugs tested for is medically necessary for the patient’s care.

From 2013 to 2014, the Clinic ran presumptive tests in house but outsourced its definitive testing to Southwest Labs. During this time, Ehn would tell doctors that they needed to order both types of UDT for every patient who came in the door. And Ehn profited from these test referrals, even in the face of concerns from staff that the profits came from illegal testing practices and kickbacks.

Seeing the profits to be earned from definitive testing, Ehn enlisted Siefert to devise a new plan. Ehn and Siefert knew that only the lab that was doing the testing could bill for reimbursement. So, with the help of Karla Cox, the Clinic’s office manager from 2016 to 2019, they planned to finance an in-house definitive testing machine. This would allow the Clinic to collect and process both UDT types on site. Cox created a business plan at Ehn's direction, using a $215 reimbursement rate per definitive test. That number was based on the highest Page 4reimbursement level for definitive UDT offered on Medicare’s 2016 fee schedules. Cox used the highest reimbursement rate because “Ehn said that if we were going to take on bringing in the lab, we needed to make sure that we ensured that it was the highest level of reimbursement.' (Trial Tr. Vol. 5, R. 243, PageID 5649). Both Ehn and Siefert agreed to use this reimbursement rate to calculate the Clinic's projected revenues, which were then provided to the lenders financing the new machine. And yet no doctor documented the medical necessity of presumptively billing every test at the highest reimbursement rate.

As the Clinic was setting up its definitive-testing lab, Ehn’s goal of profiting from lab tests led to two uncharged over-billing schemes. First, Wellcare-a Kentucky Medicaid contractor-overpaid by over $2,000 per definitive UDT from January 2016 to May 2017. Second, insurers overpaid the Clinic for specimen-validity testing from 2016 to 2018. Ehn knew of these windfalls, but he wanted to remain quiet. Ultimately, the insurers caught the Clinic and settled any overpayment disputes.

Ehn learned no lessons from these settlements. To maximize profits from the newly installed definitive-testing machine, Ehn continued to encourage providers to regularly order both types of UDT for patients. Because Ehn was a chiropractor, he could not prescribe opioids or order UDT, so he relied on Siefert, who was authorized by the Drug Enforcement Administration ('DEA') to prescribe controlled substances, to do so. According to various Clinic staff members, it was Clinic practice to order a presumptive and a definitive test for each patient every time they came to the Clinic—once a month per patient. Patient medical records also reflected this practice. At trial, the government used five patients' treatment records to establish Ehn and Siefert’s fraudulent billing scheme. The billing data for these patients showed that the Clinic billed their insurers monthly for both types of UDT. And the definitive test was nearly always billed for the most expensive test.

The Clinic's added profits from in-house definitive testing came with operational troubles. Because of the definitive UDT machine’s complexity, the Clinic had to hire a qualified technician to run and maintain it. The Clinic’s first lab technician failed to perform the routine maintenance necessary to ensure that the machine properly tested the urine samples. That failure affected the results the machine produced. Indeed, the results of tests run on the malfunctioning Page 5machine could not be used in patient care because the machine returned false positives, showing a fatal variety of drugs present in living patients’ systems.

Clinic staff raised concerns to both Siefert and Ehn about the faulty results and the malfunctioning machine. But even after Ehn observed the machine malfunctioning and agreed that it was malfunctioning, the Clinic continued to use it for testing without changing testing practices. The Clinic even billed insurers for tests run on the malfunctioning machine and kept the profits.

On August 23, 2018-two years after the Clinic began processing its own definitive tests-Siefert wrote a memo to Ehn reflecting his realization that the Clinic was likely overtesting patients without documenting the medical necessity for the tests. Siefert stated that the Clinic “need[ed] to be doing a risk stratification” before deciding to order any definitive UDT.¹ (Trial Tr. Vol. 5, R. 243, PageID 5707). He also acknowledged that the Clinic might need to reimburse insurers for potential over-billing. Cox weighed in with a memo of her own, reporting Medicare’s requirement that providers document the rationale for each definitive test administered. That rationale must focus on the patient’s needs and substances relevant to his or her treatment. She cited Medicare guidelines explaining risk stratification and cautioned that definitive tests are rarely needed absent a limited set of circumstances, like an unexpected presumptive result.

Ehn responded to these memos by having Cox perform an internal audit of the Clinic’s patients to conduct after-the-fact risk stratification. To control the amount of repayment necessary, Ehn specifically told Cox that the audit should reveal a higher percentage of high-risk patients and a lower percentage of low-risk patients. But Ehn was a chiropractor, not a medical doctor, and lacked any authority or training to order UDT, so his directions were not grounded in professional medical judgment. And a patient’s risk level must be established on a patient-bypatient basis, not based on predetermined risk categorization. Yet nearly all patients received a high-risk classification, limiting the Clinic’s cost of repayment for over-testing and allowing it to continue testing the vast majority of patients every month.

Page 6Siefert left the Clinic in early 2019. By that time, the Clinic had lost its latest lab technician. Finding no one to run the definitive UDT machine, Ehn ran it himself with assistance from an outside contractor. But Ehn did not have time to work as a full-time chiropractor and run the machine, so he fell behind on preparing urine samples and running tests. This resulted in samples being left untested beyond their shelf lives, so Clinic staff had to discard some. The Clinic fell so behind on maintenance that it had to outsource testing once again because definitive UDT results from the Clinic's machine could not be trusted. The government executed a search warrant on the Clinic in the summer of 2020. After this, Ehn continued to process definitive tests. But he could not overcome the accumulated backlog, and many urine samples were not tested in time or at all, meaning providers could not use the results to treat patients. These late-processed tests were not medically necessary because the results were unreliable and could not be used in treating patients. Yet the Clinic still billed for them.

A grand jury charged both defendants with conspiracy to commit health care fraud, in violation of 18 U.S.C. § 1349 (Count 13) and health care fraud, in violation of 18 U.S.C. § 1347 (Counts 14–15). These counts alleged that the defendants engaged in a fraudulent UDT scheme to over-bill insurers. The indictment also charged both defendants with conspiracy to distribute controlled substances, in violation of 21 U.S.C. § 846 (Count 1), and charged Siefert with eleven specific instances of unlawful distribution of a controlled substance, in violation of 21 U.S.C. § 841(a)(1) (Counts 2–12). Both defendants elected to exercise their jury-trial right. Each orally moved for a directed verdict of acquittal at the close of the government’s case. The district court denied both motions.

The jury found Siefert guilty of health care fraud and Ehn guilty of both conspiracy to commit health care fraud and health care fraud. Both defendants renewed their motions for acquittal and moved for a new trial under Federal Rules of Criminal Procedure 29 and 33 after the jury returned its verdict. The district court denied both motions, finding there was “ample, competent and admissible evidence presented for a reasonable jury to” convict. (Mem. Op. and Order, R. 280, PageID 8218). Both defendants’ pre-sentencing reports (“PSRs”) calculated their total offense levels as 37 with a criminal history category of I. After adjudicating the parties’Page 7 objections to the PSRs, the district court calculated a total offense level of 24 for Siefert and 26 for Ehn. This gave Siefert a Guidelines range of 51 to 63 months' imprisonment and Ehn a range of 63 to 78 months. The court sentenced both defendants well below their respective Guidelines ranges: 18 months for Siefert and 30 months for Ehn.

Both defendants appeal various aspects of their convictions and sentences. Ehn argues that there was insufficient evidence to support either of his convictions. He further claims that the district court improperly admitted propensity evidence. He also challenges the district court's denial of his requested jury instruction on LCDs. He claims that the government materially varied from the fraud charged in the superseding indictment. He next contends his conflict-of-interest waiver was constitutionally invalid. Lastly, he claims that the district court improperly calculated the loss and restitution amounts, rendering his sentence procedurally unreasonable.

Siefert raises two challenges on appeal. First, he claims that the government's introduction of uncharged patient death evidence amounted to prosecutorial misconduct, rendering his trial fundamentally unfair. Second, like Ehn, Siefert claims the district court procedurally erred in calculating the amount of loss attributable to him.

We first address Ehn’s sufficiency-of-the-evidence challenges. Ehn claims that the jury lacked sufficient evidence to support either of his convictions. We review Ehn’s challenges de novo. United States v. Ray , 803 F.3d 244, 262 (6th Cir. 2015). We ask “whether any rational trier of fact could have found the elements of the offense beyond a reasonable doubt.” United States v. Sumlin , 956 F.3d 879, 891 (6th Cir. 2020) (quoting United States v. Maliszewski , 161 F.3d 992, 1005 (6th Cir. 1998)). To prevail, Ehn must establish that no rational juror could have found the government met its burden on either charge. Ehn’s burden is a heavy one because we examine “the evidence in the light most favorable to the government and draw all inferences in the government's favor.' Id. (quoting Maliszewski , 161 F.3d at 1005). In so doing, we also draw all inferences and resolve any credibility determinations in favor of preserving the jury’s verdict. Id.

We begin with Ehn's health-care-fraud conviction. To prove Ehn committed that offense, the government had to show that Ehn “knowingly (1) devised and (2) executed a scheme to defraud a healthcare benefit program (3) with intent to defraud.” United States v. Betro , 115 F.4th 429, 443–44 (6th Cir. 2024) (citation omitted); 18 U.S.C. § 1347. Evidence presented at trial amply satisfies the government’s burden.

Ehn knew that Medicaid and Medicare would reimburse only for medically necessary definitive testing. After all, he agreed to abide by Medicaid and Medicare regulations, which explained that each would reimburse only for medically necessary care. Yet his instructions to Clinic staff about testing practices disregarded those regulations. The regulations clearly state that, to receive reimbursement for testing, doctors must document the medical necessity of definitive testing for each patient and the classes of drugs to be tested. But the Clinic often billed and received payment for definitive testing with no documented need for a definitive test, let alone one that tests for the maximum number of drug classes. And the Clinic billed tests run on malfunctioning equipment even though those tests were unreliable and thus not medically necessary. Indeed, Ehn knew that the definitive machine was malfunctioning, but he did nothing to stop the Clinic from billing for the unreliable, medically unnecessary tests the machine produced. Instead, Ehn micromanaged billing practices and encouraged providers to over-test their patients in an effort to use definitive UDT to increase the Clinic’s profits. And Ehn had no training or expertise to discern whether these tests were medically necessary; as a chiropractor, he was not authorized to prescribe opioids or order UDT.

Despite his lack of expertise, Ehn ignored concerns that the Clinic was testing too frequently. Even when Siefert told Ehn that the Clinic had been over-testing and needed to engage in risk stratification, Ehn asked Cox to run an internal audit with results showing primarily high-risk patients to limit repayment. But Ehn could not have known whether patients were high or low risk, or whether those tests were medically necessary.

Ehn's consistent choice to put UDT profits above ensuring that insurers paid only for Clinic patients' medically necessary care readily supports his health-care-fraud conviction.

Page 9After all, the government need not present '[d]irect evidence of fraudulent intent' to defeat Ehn's sufficiency challenge. Betro , 115 F.4th at 444 (alteration in original) (quoting United States v. Anderson , 67 F.4th 755, 770 (6th Cir. 2023) (per curiam)). The jury could have inferred Ehn’s guilt from his instructions to bill for the highest level of care even if that care was not medically necessary. See id. at 444–45. Or from the fact that he allowed the Clinic to bill for tests run so long after collecting the sample that it could no longer rely on the provider's certification that the tests were medically necessary when ordered. See United States v. Bertram , 900 F.3d 743, 749 (6th Cir. 2018). So, too, would Ehn's desire to maximize profits (shown through the 2016 business plan) or to skew the 2018 audit results to high-risk patients have been sufficient for the jury to infer fraud.

Ample testimony from Clinic employees and expert witnesses explained how UDT should have been used at the Clinic and when insurers would reimburse for definitive UDT. Testimony also established that Ehn strayed from those parameters at nearly every opportunity. Drawing all inferences in the government’s favor, as we must, see Sumlin , 956 F.3d at 891, the record shows that the Clinic ordered and billed for UDT greatly exceeding what Ehn should have known Medicare and Medicaid would reimburse for. From this, the jury could have inferred that Ehn knowingly and intentionally devised and executed a health-care-fraud scheme to over-bill for medically unnecessary UDT, in violation of 18 U.S.C. § 1347.

Ehn’s challenge to his conspiracy conviction fares no better. To prove Ehn engaged in a conspiracy to commit health care fraud, the government had to show that “two or more persons conspired, or agreed, to commit the crime,” and “that the defendant knowingly and voluntarily joined the conspiracy.' United States v. Rogers , 769 F.3d 372, 377 (6th Cir. 2014) (citation omitted); 18 U.S.C. § 1349. The government readily established that an agreement existed through circumstantial evidence tending to show both Siefert and Ehn agreed to participate in the common scheme. See United States v. Hughes , 505 F.3d 578, 593 (6th Cir. 2007).

The same evidence supporting Ehn's health-care-fraud conviction also shows that the jury reasonably convicted him of conspiracy to commit health care fraud. The 2016 business Page 10plan shows that Ehn and Siefert intended to bring definitive testing in house and to bill for those tests at the highest reimbursement level to maximize the Clinic’s profits, notwithstanding the fact that no medical need could justify that plan. And Ehn and Siefert agreed to share the profits such testing brought in, again with no regard to whether they lawfully obtained those proceeds from insurers. Drawing all inferences in the government’s favor, a jury could have found that Ehn and Siefert agreed to engage in a conspiracy to commit health care fraud.

What little Ehn argues on the merits—that he merely trusted Siefert’s medical judgment in ordering tests-ignores the substantial evidence showing his involvement in ordering and billing UDT to maximize the Clinic’s (and, in turn, his) profits. Ehn’s primary argument is a legal one: He argues that the rule of consistency bars his conviction because the jury acquitted Siefert of the same charge. This common law rule “at one time required ‘that, where all possible co-conspirators are tried together, and all but one are acquitted, the remaining conspirator's conviction must be reversed for lack of sufficient evidence.’” United States v. Crayton , 357 F.3d 560, 564 (6th Cir. 2004) (quoting United States v. Walker , 871 F.2d 1298, 1304 n.5 (6th Cir. 1989)). However, this rule is no longer good law after United States v. Powell , 469 U.S. 57 (1984). See Getsy v. Mitchell , 495 F.3d 295, 306–07 (6th Cir. 2007) (en banc) (acknowledging abrogation of the rule). So “the acquittal of all but one co-conspirator during the same trial does not necessarily indicate that the jury found no agreement to act.' Crayton , 357 F.3d at 565. Because the rule of consistency is no longer good law, Ehn is left with no argument to challenge his conspiracy conviction.

Next we consider Ehn's argument that the district court erroneously allowed the government to present '[s]ignificant and unfairly prejudicial evidence' of his prior conduct. (ECF 19, Ehn’s Br. at 33). Ehn challenges the government’s introduction of evidence that Ehn (1) accepted kickbacks from Southwest Labs, (2) settled an overpayments dispute with Wellcare, and (3) resolved a double-billing dispute concerning specimen-validity testing. The district court found the proposed evidence was 'admissible to establish the defendants' knowledge, intent, [and] lack of mistake pursuant to 404(b).” (Pre-trial Hr’g Tr., R. 157, PageID 3107). The court provided a limiting instruction to the jury, allowing it to “consider the evidence only as it relates Page 11to . . . Ehn's plan, motive, intent, or knowledge.' (Jury Instrs., R. 218, PageID 3712). We review the district court's decision to admit Rule 404(b) evidence under a modified abuse-ofdiscretion standard. United States v. Fairley , 137 F.4th 503, 517 (6th Cir. 2025). 'First, we review for clear error the factual determination that other acts occurred,” and then “we review de novo the legal determination that the acts were admissible for a permissible 404(b) purpose.” Id. (quoting United States v. Adams , 722 F.3d 788, 810 (6th Cir. 2013)). Finally, 'we review for abuse of discretion the determination that the probative value of the evidence is not substantially outweighed by unfair prejudicial impact.” Id. (quoting Adams , 722 F.3d at 810–11).

Federal Rule of Evidence 404 prohibits the use of prior-acts evidence 'to prove a person’s character in order to show that on a particular occasion the person acted in accordance with the character.' Fed. R. Evid. 404(b)(1). However, it allows such evidence to show “motive, opportunity, intent, preparation, plan, knowledge, identity, absence of mistake, or lack of accident.” Fed. R. Evid. 404(b)(2).

The district court acted within its discretion in allowing the government to present this evidence. Ehn does not dispute that these prior acts occurred. And the evidence of these prior acts is probative of a material issue other than character because it 'relates to conduct that is substantially similar and reasonably near in time to the specific intent offense at issue.” United States v. Barnes , 822 F.3d 914, 922 (6th Cir. 2016) (citation omitted). It tends to show that, immediately before or concurrently to the fraud scheme charged, Ehn knew how to use UDT to fraudulently increase profits and had the motive and intent to keep ill-gotten gains from overbilling insurers. And the district court allowed the government to present the challenged evidence to combat Ehn’s good-faith defense of inadvertent over-billing—a material issue other than Ehn’s character. Lastly, Ehn does not explain how this evidence was unduly prejudicial. If anything, Ehn claims in his reply brief that the instruction was confusing. But the instruction specifically identified the other-acts evidence introduced, explained what the jury had to find to consider those acts, and limited the purpose for which the jury could consider the acts. Thus, the instructions faithfully followed the rule explaining when other-acts evidence is admissible, mitigating any potential prejudice to Ehn. See United States v. Bartholomew , 310 F.3d 912, 922 (6th Cir. 2002).

Page 12We move next to Ehn’s contention that the district court erroneously denied his requested LCD jury instruction. As part of its argument that Siefert and Ehn billed for UDT while intentionally disregarding the relevant Medicaid and Medicare standards for reimbursement, the government introduced expert testimony regarding the relevant LCDs that determined when the Clinic could seek Medicare reimbursement for UDT. Ehn contends that, through this testimony, the government tried to prove that he committed fraud by violating LCDs because the government’s case rested on the definition of “medically necessary,” which LCDs define. So, according to Ehn, a limiting instruction was necessary to ensure that the jury did not convict Ehn based solely on his violation of any LCDs. The proposed instruction would explain the limited role of LCDs in the Medicare-reimbursement process and inform the jury about how a provider could appeal the accuracy of an LCD.

The district court rejected Ehn's proposed instruction, reasoning that it 'leads to confusion” and does not involve definitions that the jury will be asked to decide. (Charge Conf. Tr., R. 326, PageID 8906). We review that decision for abuse of discretion. Anderson , 67 F.4th at 764. We may reverse “only if the proposed instruction is (1) a correct statement of the law, (2) not substantially covered by the charge actually delivered to the jury, and (3) concerns a point so important in the trial that the failure to give it substantially impairs the defendant’s defense.” United States v. Volkman , 797 F.3d 377, 385 (6th Cir. 2015) (internal quotation marks and citation omitted).

We need go no further than the first prong of this test to determine that no abuse of discretion occurred here. Ehn’s proposed instruction incorrectly states the law. The instruction states that “decision-makers at every level of the appeal process can decline to apply the local coverage determination. Thus the only person or entity bound by a local coverage determination is the contractor that issued it.' (Proposed Instrs., R. 214, PageID 3667). This is incorrect. When Ehn agreed to serve Medicare patients, he entered a contract binding himself to abide by the relevant Medicare rules and LCDs issued by Medicare contractors. So regardless of the LCD-appeal structure, Clinic staff, including Ehn, were required to abide by LCDs when billing. And LCDs are binding unless superseded by an administrative law judge on appeal of the denial Page 13of a claim or by rulemaking from the Department of Health and Human Services. See Agendia, Inc. v. Becerra , 4 F.4th 896, 902 (9th Cir. 2021). So the instruction incorrectly states that only the contractor who issued the LCD is bound by it.

In any event, the proposed instruction does not reflect the point of law that Ehn contends that the district court should have instructed the jury on. Nothing in the instruction clearly conveys to the jury that criminal liability may not be imposed simply because Ehn violated an LCD. The district court was correct in concluding that this instruction would 'lead[] to confusion” because it does not inform the jury on things it would be called upon to consider. (Charge Conf. Tr., R. 326, PageID 8906). Ehn provides no argument to persuade us otherwise. Because he cannot meet even the first prong of this test, the district court did not abuse its discretion in denying the proposed instruction.

Ehn next argues that the government materially varied from the indictment to convict him of both health care fraud and conspiracy to commit health care fraud. He claims that the superseding indictment alleged only that he orchestrated a scheme of fraud premised on billing insurers for medically unnecessary tests. But, according to Ehn, the government’s theory at trial shifted to prove that Ehn committed a different fraud scheme premised on billing for tests run on a broken machine and billing for useless delayed tests. This different theory, says Ehn, prejudicially varied from the conduct charged in the indictment.

We review claims of “variances from an indictment de novo .” United States v. Mize , 814 F.3d 401, 408 (6th Cir. 2016) (citation omitted). “[A] variance is ‘not per se prejudicial.’” Id. at 409 (quoting United States v. Budd , 496 F.3d 517, 521 (6th Cir. 2007)). It occurs 'if the evidence ‘proves facts materially different from those alleged in the indictment.’” United States v. Davis , 970 F.3d 650, 659 (6th Cir. 2020) (quoting Budd , 496 F.3d at 521). We will reverse only if Ehn demonstrates that '(1) a variance occurred and (2) that the variance affected' his substantial rights. Mize , 814 F.3d at 409 (quoting United States v. Kuehne , 547 F.3d 667, 683 (6th Cir. 2008)). A defendant’s substantial rights “are affected only when the defendant shows prejudice to his ability to defend himself at trial, to the general fairness of the trial, or to the Page 14indictment's sufficiency to bar subsequent prosecutions.' Kuehne , 547 F.3d at 683 (quoting United States v. Hynes , 467 F.3d 951, 962 (6th Cir. 2006)).

Ehn has shown no such variance. The indictment clearly charged Ehn with a single health-care-fraud scheme committed through various means from 2017 through 2021. And it charged a single conspiracy to carry out that scheme from 2017 through 2019. The indictment alleges that Ehn executed the scheme to defraud by knowingly submitting claims for payment for 'medical services that were not performed or were not medically necessary[.]' (Superseding Indictment, R. 58, PageID 346). It identified specific definitive UDT billed (most at the highest reimbursement level) in violation of 18 U.S.C. § 1347. The “methods, manner, and means” used to carry out the conspiracy and health-care-fraud scheme included: (1) submitting claims to health-benefit programs for UDT insurance reimbursement; (2) directing Clinic employees to obtain urine samples during office visits; (3) performing both presumptive and definitive UDT on the specimens regardless of patients' individualized needs pursuant to a secret 'blanket order'; (4) submitting 'false and fraudulent claims' for UDT reimbursement and falsely representing that the tests were medically necessary; (5) submitting bills for definitive UDT at a higher level of complexity than medically necessary; and (6) causing the submission of bills for definitive UDT not used to treat patients. ( Id. at 340–41). These allegations adequately notified Ehn of the charges brought against him and the facts through which the government intended to prove its case.

Start with the substantive health-care-fraud charge. The stated method, manner, and means of fraud put Ehn on notice that the scheme could involve any number of medically unnecessary tests, including those conducted on broken machines. After all, neither Medicare nor Medicaid would reimburse for those useless tests. The indictment even identified specific tests billed in violation of federal law. Several of those tests were conducted either while the lab technician failed to properly maintain the machine from 2017 to 2018 or when Ehn fell behind in running the definitive tests in 2020. Ehn was on notice that he would need to defend against the government's contention that these tests either were not performed or were medically unnecessary. That the government’s evidence at trial went into further detail about how and why those tests were not performed or were medically unnecessary (i.e., because they were performed Page 15on broken machinery or too long after the sample had been collected) does not mean the government varied from the indictment. After all, “[t]he presentation of additional evidence to substantiate charged offenses . . . does not constitute facts materially different from those charged in the indictment.” Kuehne , 547 F.3d at 686. The fraud scheme charged was the same regardless of which specific evidence the government used to prove it.

Next consider Ehn’s conspiracy conviction. At trial, the government sought to prove that Siefert and Ehn agreed to execute a UDT scheme to maximize the Clinic's profits and their personal profits. The scheme involved over-testing and over-billing by requiring both presumptive and definitive UDT for every patient at every visit, billing at the highest reimbursement level even when it was not medically necessary, and billing for testing on malfunctioning equipment. This mirrors the conspiracy alleged in the indictment. We will find a variance in conspiracy cases only where 'the indictment alleged one conspiracy, but the evidence can reasonably be construed only as supporting a finding of multiple conspiracies.' Adams , 722 F.3d at 805-06 (alteration in original) (quoting United States v. Caver , 470 F.3d 220, 236 (6th Cir. 2006)). Ehn does not identify any other conspiracy alleged at trial; his qualms focus only on the different means that the government raised for how he and Siefert executed their scheme. But proving the conspiracy using different means is not the same as varying from the indictment. Because the indictment encompassed all relevant evidence that the government presented to prove the crimes of conviction, Ehn’s variance challenge fails.

The last of Ehn's pre-sentencing challenges concerns the conflict-of-interest waiver he signed. Prior to trial, the government raised a potential conflict of interest regarding Ehn’s trial counsel. Ehn’s chosen trial counsel, Squire Patton Boggs (US) LLP, had represented Wellcare of Kentucky (now owned by Centene Corp.) in its overpayment dispute and settlement with the Clinic and represented Centene-which helps run Kentucky Medicaid-throughout the trial. Ehn's attorney responded to the government's notice, dispelling any concerns that the firm's conflict would impact his performance in this case. The district court held a conflict-of-interest hearing, where Ehn stated his desire to keep his retained trial counsel. Ehn signed a written waiver acknowledging the same.

Page 16Ehn now claims that the representation of Wellcare by his trial-counsel’s firm “prevented counsel from fully exploring the Wellcare issue at trial, significantly hindering Dr. Ehn’s trial defense” and deprived him of his Sixth Amendment rights. (ECF 19, Ehn’s Br. at 57). Whether Ehn's conflict waiver was constitutionally valid and whether the district court properly concluded that no conflict existed are both legal questions that we review de novo. United States v. Osborne , 402 F.3d 626, 630 (6th Cir. 2005). But we review the factual findings underlying the district court’s legal conclusions for clear error. Id.

To remedy any conflict-of-interest concerns, a court 'must be convinced that the defendant[] understand[s] the rights being waived and the consequences of the waiver of those rights.' Id. at 631 (citing Fed. R. Crim. P. 44(c)). In other words, a binding conflict waiver must be “knowing, intelligent, and voluntary.” Id. at 630 (citation omitted). In determining whether the district court adequately informed the defendant of his right to conflict-free counsel, we consider whether the court explained the hazards of representation by the conflicted attorney and informed him of his right to unconflicted counsel. Id. at 631. This includes whether the court ensured that the defendant 'has discussed the matter with his attorney or if he wishes with outside counsel.” Fed. R. Crim. P. 44(c) advisory committee’s note to 1979 amendment (citation omitted).

Ehn knowingly, intelligently, and voluntarily waived any conflict. At the conflict-ofinterest hearing, Ehn affirmed that he discussed the conflict issue with his counsel, understood that counsel could not use any information that his firm learned while representing Wellcare against him, declined independent representation on the conflict issue three separate times, and chose to retain his counsel. Further, he signed a conflict waiver recognizing his right to conflictfree counsel in which he acknowledged the potential conflict of interest and expressed his “wish to continue with [his] current counsel.” (Conflict Waiver, R. 163, PageID 3149). The waiver concludes with Ehn's acknowledgement that he is 'making this decision knowingly and voluntarily.' ( Id. ). Considering all this evidence, Ehn waived his right to challenge the constitutionality of his counsel’s representation. See Osborne , 402 F.3d at 630.

We now turn to Siefert's prosecutorial misconduct claim. Before trial, Siefert filed a motion in limine to exclude evidence of seven uncharged patient deaths. The government intended to use the evidence to show that Siefert and Ehn knew that their prescribing methods harmed patients and led to overdose deaths, yet they continued to prescribe in that manner anyway. It would also show, according to the government, that Siefert lied to the medical board when he said that he “never had a patient overdose and die during [his] many years of practice.” (Trial Tr. Vol. 3, R. 241, PageID 4991). The district court denied the motion, determining that 'targeted' evidence could be relevant to show Siefert's knowledge that the patients were possibly misusing their prescriptions. (Pre-trial Hr'g Tr., R. 157, PageID 3040). The court reasoned that such evidence could support the government's theory that continuing to issue prescriptions was outside the scope of usual professional practice or for no legitimate medical purpose—i.e., that he was prescribing in an unauthorized manner. See Ruan v. United States , 597 U.S. 450, 454 (2022) (holding that the government must prove that an authorized medical prescriber, in issuing controlled substances, knowingly or intentionally acted in an unauthorized manner).

The district court acknowledged that ' Ruan didn't change the rules of evidence,' but reasoned that circumstantial evidence about the patients’ deaths was 'critical' for the government to meet its burden under Ruan. ( Id. ). Therefore, said the district court, the probative value of the evidence was “not substantially outweighed by the danger of unfair prejudice.” ( Id. at 3042). Nonetheless, to minimize the risk of any potential unfair prejudice, it adopted a limiting instruction providing that the jury could consider evidence of death only 'for the purpose of [determining] the defendant’s knowledge, [or] intent, and that the distribution of the prescriptions were or were not outside the scope of professional practice or not for legitimate medical purpose[s].” ( Id. ).

At trial, the government presented evidence about all seven patients' deaths. For instance, one expert opined that Siefert’s prescribing methods contributed to the patients’ deaths.

Page 18And family members of some the deceased patients testified that their loved ones showed signs of addiction while in Siefert’s care. But it offered specific evidence of Siefert’s knowledge of the deaths of only three patients. Therefore, as a curative measure, the court instructed the jury to disregard death-related evidence “[i]f the defendant did not know about certain evidence of deaths,” because that evidence would “not [be] relevant to that defendant’s knowledge” for the distribution counts. (Trial Tr. Vol. 2, R. 240, PageID 4697).

In denying Siefert’s post-trial motion for acquittal or new trial, the district court clarified that the prosecution “was not required to prove that Defendants were aware of each death for the evidence of the patient deaths to be relevant.' (Mem. Op. and Order, R. 280, PageID 8212). Instead, it emphasized that the evidence of patient deaths was relevant if it had “some tendency to suggest that a provider defendant knew or was deliberately ignorant of the fact that patients might be misusing prescriptions and overdosing.” ( Id. (quoting United States v. Hofstetter , No. 3:15-cr-27, 2019 WL 6718489 (E.D. Tenn. Dec. 9, 2019)). Considering the limiting instruction that it provided to the jury, the district court found that the prosecution did not violate the court’s pretrial order, and even if it had, vacatur was not warranted. After all, the jury acquitted Siefert of the drug distribution counts for which the challenged evidence was admitted. Accordingly, it denied Siefert’s motion for acquittal or new trial.

Siefert challenges the district court’s decision on appeal, leaning heavily on his assertion that the evidence of patient deaths was “irrelevant and prejudicial.” (ECF 15, Siefert’s Br. at 37). Specifically, he argues that the government violated the district court’s pretrial order by offering evidence of the other four patients’ deaths without having a good-faith belief that it could show Siefert’s knowledge of the deaths. In proceeding in this fashion, says Siefert, the government committed misconduct entitling him to a new trial.

Generally, we review questions of prosecutorial misconduct de novo. United States v. Gardiner , 463 F.3d 445, 459 (6th Cir. 2006). The government argues that plain-error review applies here because Siefert failed to raise his misconduct objection during trial. True, Siefert objected to the government’s use of patient death evidence solely on evidentiary grounds during Page 19trial. But Siefert contends that he still preserved his misconduct objection because he raised the issue in post-trial proceedings in the district court, and the court addressed it in a post-trial order. Although the government points to caselaw stating that failure to lodge a misconduct objection during trial means that the issue receives plain-error review on appeal, see, e.g., Betro , 115 F.4th at 447, the cases it relies on for this proposition involve instances where the objecting party failed to raise the issue with the district court at any stage. Less clear is the standard to be applied when, as here, the specific objection was not raised during trial but was later ruled upon by the district court. Regardless, we need not resolve this disagreement because we conclude that Siefert’s objection fails under any applicable standard. Cf. United States v. Pacheco , Nos. 23-5762, 5819, 2025 WL 2060779, at *4 (6th Cir. July 23, 2025).

“We employ a two-step test in evaluating a claim of prosecutorial misconduct.” United States v. Carson , 560 F.3d 566, 574 (6th Cir. 2009). First, we ask whether the challenged statement or action was improper. Id. If it was improper, we next consider whether the conduct was flagrant, warranting reversal. Id. Because Siefert has not shown that the government's actions were improper, he fails at step one.

Siefert makes four interrelated arguments. According to him, the government acted improperly by: (1) defying the court's pre-trial order concerning the use of uncharged death evidence; (2) introducing evidence relating to patient overdose deaths despite lacking a goodfaith basis for doing so; and (3) introducing irrelevant and prejudicial death-related evidence. He also claims that (4) he is entitled to a new trial because the 'cumulative effect' of the prosecution's misconduct deprived him of due process. ( Id. at 23). The district court rejected these arguments, as do we.

As an initial matter, the district court did not forbid the introduction of death-related evidence absent a showing of Siefert's direct knowledge. Rather, the court allowed for the possibility that the government could show constructive knowledge or deliberate ignorance. True, there was seeming incongruity between the district court's pre- and post-trial orders. Pretrial, the district court said that the government was 'going to have to show thatPage 20 [the defendants] knew about the deaths,” while post-trial it stated that the government “was not required to prove that Defendants were aware of each death for the evidence of the patient deaths to be relevant.” (Pre-trial Hr’g Tr., R. 157, PageID 3035; Mem. Op. and Order, R. 280, PageID 8212). Still, we reject Siefert’s invitation to read the pretrial order as a hard-and-fast prohibition against the introduction of death-related evidence absent a direct knowledge link. Had the district court meant to bar the introduction of such evidence, it would have granted Siefert's motion to exclude it. Indeed, all parties knew during the pretrial conference that the government did not have direct evidence establishing Siefert’s knowledge of four of the seven patient deaths. But that fact alone did not render all evidence of “red flags” irrelevant for the drug distribution counts. So Siefert's contention that '[b]ecause [direct] evidence simply did not exist, the government’s conduct must be viewed as deliberate” is off base. (Siefert’s Br. at 34–35). At no point did the district court find that the government violated its pretrial order. Nor do we.

With this clarification in mind, Siefert's good-faith argument buckles. He asserts the government lacked 'good faith' because it had no reason to 'believe that it could meet the knowledge burden[.]” ( Id. at 30). Yet, Siefert points us to no specific actions to show that the government lacked a reason to believe that it could show that Siefert knew of or deliberately ignored evidence of patient deaths. Rather, he seems to argue that the government’s failure to definitively show that Siefert knew about four of the deaths is evidence of bad faith in and of itself. But we do not automatically equate this lack of direct proof with a lack of good faith. As noted, the government had previewed evidence of patient deaths at the pretrial hearing, and the district court indicated that '[t]he knowledge . . . described here by the United States is sufficient” for introduction of the evidence. (Pre-trial Hr’g Tr., R. 157, PageID 3042). Under these circumstances, the government’s inability to ultimately show Siefert’s knowledge does not constitute lack of good faith.

Next, we agree with the district court that the evidence relating to patient deaths was relevant and not unduly prejudicial. Siefert resists this conclusion by tethering relevance to Siefert's knowledge of the deaths. But relevance under Ruan involves a different kind of “knowledge”: the mens rea with which the government must prove Siefert acted when issuing prescriptions. Ruan , 597 U.S. at 457–58. To satisfy Ruan , evidence of patient overdose deaths Page 21can be relevant to questions of whether the doctor knew or had reason to know that his prescribing practices were unauthorized and whether he intended to prescribe in an unauthorized manner. Introduction of evidence for this end was not improper.

Siefert's prejudice argument also fails. As discussed, the district court expressly instructed the jury to consider death-related evidence only on the issue of Siefert’s knowledge or intent—and even then, only to the extent that Siefert knew or should have known of the deaths. Subject to narrow exceptions not applicable here, 'our legal system presumes . . . that jurors follow limiting instructions[.]' Samia v. United States , 599 U.S. 635, 646 (2023). And here there is no reason to conclude that they did not. While such instructions are not 'a sure-fire panacea” to mitigate prejudice, we agree with the district court that Siefert “c[ould not] show that he was prejudiced by the admittance of this evidence.” United States v. Haywood , 280 F.3d 715, 724 (6th Cir. 2002); (Mem. Op. and Order, R. 280, PageID 8213). After all, the jury acquitted Siefert of every count to which the death-related evidence was relevant. This makes Siefert’s suggestion that the evidence was so prejudicial that it could have infected the jury's decision making as to the health-care-fraud count speculative at best.

In sum, we agree with the district court that the prosecution did not act improperly. As discussed above, the government planned to offer the patient-death evidence to prove knowledge for purposes of the drug counts. At trial, all parties agreed that the government had offered circumstantial evidence that Siefert knew that three of the seven patients had died. And because the government was unsuccessful in its effort to link Siefert by knowledge or deliberate ignorance to the other four patients’ deaths, the district court curtailed the government’s ability to reference them during closing. Siefert points to no examples of the government disregarding this admonition. Accordingly, we see no impropriety in the government’s actions. And because the government did not act improperly by offering evidence relating to uncharged patient deaths, we end our prosecutorial-misconduct analysis at step one.

Finally, Siefert argues that he deserves a new trial because the “cumulative effect” of the prosecution's misconduct deprived him of due process. (ECF 15, Siefert's Br. at 23). We review the district court’s denial of his motion for acquittal or new trial on this basis under the abuse-of-discretion standard. See United States v. Pierce , 62 F.3d 818, 823 (6th Cir. 1995). On Page 22this front, we ask whether the alleged improper conduct “so infected the trial with unfairness as to make the resulting conviction a denial of due process.” Slagle v. Bagley , 457 F.3d 501, 515 (6th Cir. 2006) (quoting Darden v. Wainwright, 477 U.S. 168, 181 (1986)). Isolated errors that “might not be so prejudicial as to amount to a deprivation of due process . . . may cumulatively produce a trial setting that is fundamentally unfair.” United States v. Hernandez , 227 F.3d 686, 697 (6th Cir. 2000) (citation omitted).

Because we find that the government did not engage in even isolated instances of misconduct, we agree with the district court that Siefert is not entitled to a new trial. See United States v. Trujillo , 376 F.3d 593, 614 (6th Cir. 2004). Like the district court, we note that the death-related evidence implicated only charges for which Siefert was acquitted.² Siefert insists that the introduction of the death-related evidence affected the jurors' decision-making on the health care fraud scheme. But he points us to no facts tending to demonstrate such a correlation. And any overlap between the offenses is not obvious. For these reasons, we find that the district court did not abuse its discretion by denying Siefert’s due process claim.

Lastly, we address Siefert and Ehn’s challenges to the reasonableness of their sentences. Both defendants challenge the district court's calculation of the respective loss amounts underlying their Guidelines ranges.³ These amount to challenges to the procedural reasonableness of the defendants’ sentences. See United States v. Rayyan , 885 F.3d 436, 440 (6th Cir. 2018). We review procedural reasonableness under the abuse-of-discretion standard.

Page 23Betro , 115 F.4th at 454. But we review the district court’s factual determination of the amount of loss under the clear-error standard and its methodology for arriving at that amount de novo. Id.

The Federal Sentencing Guidelines enhance “a defendant’s sentence to correlate to the amount of loss caused [by] his fraud.' United States v. Triana , 468 F.3d 308, 319 (6th Cir. 2006). We calculate the amount of loss as “the greater of actual or intended loss.” Betro , 115 F.4th at 454 (quoting United States v. Wendlandt , 714 F.3d 388, 393 (6th Cir. 2013)). “‘[A]ctual loss’ is ‘the reasonably foreseeable pecuniary harm that resulted from the offense[.]’” Triana , 468 F.3d at 320 (quoting U.S.S.G. § 2B1.1 cmt. (n.2(A)(I) and (ii)) (2002)). When, like here, the losses stemming from financial frauds are difficult to quantify, “the district court need only make a reasonable estimate of the loss, given the available information.” Id. (citation omitted). “Such estimates need not be determined with precision.' Id. (internal quotation marks and citation omitted). But the district court must “actually find facts, and it must do so by a preponderance of the evidence.” United States v. White , 492 F.3d 380, 416 (6th Cir. 2007) (emphasis omitted).

Ehn and Siefert argue that the burden to prove the loss amount remained on the government because the district court did not make a finding that the fraud was pervasive. Because of that, say the defendants, the government had to prove and the district court had to find the reasonable estimated loss by a preponderance of the evidence. They also argue that the district court's calculation and methodology was flawed, and that it failed to articulate a reasonable method for calculating the loss amount as required by U.S.S.G. § 2B1.1.

The district court found the loss amount by a preponderance of the evidence and explained its loss calculation at sentencing. The defendants' presentence reports from the probation department followed U.S.S.G. § 2B1.1 to calculate the loss amount: The sum was reached by first determining the total amount billed for the definitive tests and then subtracting twenty percent to account for any legitimate tests. So, the total loss amount was around $47 million for Siefert and around $53 million for Ehn.

Page 24Both the defendants and the government objected to this loss calculation, agreeing that the amount paid by insurance, not the amount billed, should serve as the basis for the calculation. The district court agreed that the loss calculation should use the paid amount. It adopted the remainder of the probation department’s methodology, though, subtracting twenty percent from the paid amount to account for legitimate definitive tests. The court noted that “virtually every patient . . . received both a presumptive and a definitive test” irrespective of whether the patient had an unexpected result. (Sent. Hr'g Tr., R. 323, PageID 8711). The court reasoned that calculating loss based only on definitive UDT—which the evidence showed were nearly always ordered in a fraudulent manner-avoided any objections regarding presumptive tests because presumptive UDT is often medically necessary. And the court reduced the loss and restitution amounts further to account for the settlement repayment to Wellcare in 2019, making the loss amount attributable to Siefert $1,968,765.69 and to Ehn $4,079,182.93. This was a “reasonable estimate of the loss[.]” Triana , 468 F.3d at 320.

Ehn and Siefert claim that this determination lacks any basis because the district court did not explain or justify why it found the twenty percent reduction appropriate to reasonably estimate the loss amount. We disagree. That number is supported by defendants’ own expert. The expert testified that definitive testing may sometimes be medically necessary, but only at a rate of one to two tests per patient per year. That rate is approximately eight or sixteen percent of the Clinic’s testing rate. Extrapolating from that, at most sixteen percent of the Clinic’s tests were medically necessary. So, discounting twenty percent of the amount paid for definitive tests more generously credits Siefert and Ehn for potentially necessary testing. Based on the information available, this was a reasonable estimate of the financial loss accumulated from the defendants’ fraud. See id. at 319-20. So, the district court did not clearly err in calculating each defendant’s loss amounts, and the sentences are procedurally reasonable.

We AFFIRM .