United States v. David Ming Pon

Related Cases

    Case: 17-11455 Date Filed: 06/29/2020 Page: 1 of 89 [PUBLISH] IN THE UNITED STATES COURT OF APPEALS FOR THE ELEVENTH CIRCUIT ________________________ No. 17-11455 ________________________ D.C. Docket No. 3:14-cr-00075-BJD-PDB-1 UNITED STATES OF AMERICA, Plaintiff-Appellee, versus DAVID MING PON, Defendant-Appellant. ________________________ Appeal from the United States District Court for the Middle District of Florida ________________________ (June 29, 2020) Case: 17-11455 Date Filed: 06/29/2020 Page: 2 of 89 Before ED CARNES, MARTIN, and ROGERS,* Circuit Judges. CARNES, Circuit Judge: A jury found David Pon guilty of twenty counts of health care fraud, in violation of 18 U.S.C. § 1347, and the district court entered a judgment of conviction on the verdict. After finding that Pon’s fraud scheme resulted in a loss of nearly $7 million, the court sentenced him to 121 months in prison. He appeals his convictions and sentence. I. FACTS Pon was an ophthalmologist.1 As a sole practitioner, he established his practice with a main office in Leesburg, Florida, and a satellite office in Orlando. Many of his patients were elderly. He diagnosed hundreds of them with, and lasered their eyes to purportedly treat, a debilitating and uncurable eye disease known as wet age-related macular degeneration (WMD). Here is how his scheme worked. Pon would run diagnostic tests on a patient. After diagnosing the patient with WMD, he would move on to the “treatment” phase, which involved lasering one or both of the patient’s eyes. Pon would laser his patients’ eyes with the laser set on the lowest power setting and in * Honorable John M. Rogers, United States Circuit Judge for the Sixth Circuit, sitting by designation. 1 Pon was once a licensed doctor but we do not refer to him as one because Florida’s Board of Medicine revoked his license in August 2016 as a result of the jury’s guilty verdict. 2 Case: 17-11455 Date Filed: 06/29/2020 Page: 3 of 89 micropulse mode.2 He would then submit a claim to Medicare for the diagnostic tests and the laser session. As a result, he would receive from Medicare around $1,200 total for each set of diagnostic tests and lasering. Pon would bill his micropulse laser sessions under Medicare code 67220, the code for “laser photocoagulation for [WMD], for a choroidal neovascular membrane,” or in other words, “burning an area of abnormal leaking blood vessels with a laser.” Laser photocoagulation is a treatment for WMD that creates a scar in the eye by “cooking” shut the abnormal blood vessels (feeder vessels) that are characteristic of WMD. But the extremely low power settings that Pon set his laser to before each session were not high enough to achieve coagulation, so his purported treatments left no scars and did not fit under code 67220. One expert testified that Pon’s settings were “way too low” for coagulation purposes, and that his method was tantamount to “jump-start[ing] [a car] off a flashlight. It’s so little energy.” And Pon agreed. He described his purported treatment technique — which he referred to as “the micropulse laser technique for treatment of feeder vessels” — as treating WMD while leaving “no or minimal scarring.” According to Pon, 2 The laser that Pon used has several different user-selectable modes of operation. When micropulse mode is selected, the laser is on for only a predetermined percentage of the exposure time. For example, a 15 percent duty cycle means that the laser is on for only 15 percent of the exposure time and is off for the remaining 85 percent of the exposure time. Pon would set his laser to micropulse mode with a 15 percent duty cycle before lasering his patients’ eyes. 3 Case: 17-11455 Date Filed: 06/29/2020 Page: 4 of 89 “the whole concept” behind his purported treatment was to use the laser to heat up the WMD feeder vessels “without causing a burn.” His intention was “to get the effect from the laser without causing a burn, coagulation.” In fact, according to Pon, he would “virtually never get a scar or a burn” if he did his “technique properly.” But Pon continued to bill Medicare for his laser “treatments” under code 67220 for laser photocoagulation — or laser scarring. And Pon became a top Medicare biller of WMD laser scarring treatment, billing Medicare for his micropulse laser (which is intended not to create a scar), under code 67220 for laser photocoagulation (which is intended to create a scar). The percentage of his patients whom he diagnosed with WMD and billed Medicare under code 67220 for laser photocoagulation treatment substantially increased over the years. Around 2006, drug injections had supplanted laser photocoagulation as the typically favored WMD treatment method, so other ophthalmologists’ laser treatments and billing amounts for laser photocoagulation went down. Pon’s, by contrast, went up dramatically. Pon’s practice produced puzzlement and sowed suspicion. Other doctors who also treated Pon’s patients were puzzled about his WMD diagnoses and laser “treatments.” In the fall of 2008, for example, Virginia-based doctor Robert Vogel was treating his longtime patient, D.M., and noticed that the 83-year-old had several left-eye maladies, but not WMD. Because D.M. would be in Florida for 4 Case: 17-11455 Date Filed: 06/29/2020 Page: 5 of 89 the winter months, Dr. Vogel told him to check in with an eye doctor after he got there. D.M. chose Pon, who diagnosed him with WMD and micropulse lasered his eyes. When D.M. returned to Virginia a few months later, Dr. Vogel was “shocked” when D.M. told him that Pon had lasered both of his eyes. Dr. Vogel examined both eyes, did not see WMD in either of them, and could not understand why either one would have been lasered. Nor did Dr. Vogel see a scar in either eye that would indicate Pon had used a laser at settings that would have treated WMD if D.M. had actually suffered from it. This “unusual” situation prompted Dr. Vogel to tell D.M. to find a doctor other than Pon the next time he went to Florida. Other experts observed similar anomalies involving Pon’s practice and patients. Optometrist Sam Williams referred some of his own patients to Pon, who diagnosed every one of them with WMD. Dr. Williams, who has more than forty- five years of experience as an optometrist, became concerned when some of those patients told him that Pon had lasered their eyes on multiple occasions. As a result, Dr. Williams sent them to other ophthalmologists for second opinions about the medical necessity of the suspicious laser treatments. “[O]n every occasion” the ophthalmologists found that there was no sign Pon had lasered those patients’ eyes in a way that would actually treat WMD or that the patients needed any laser treatment for any eye disease. Dr. Williams stopped referring his patients to Pon. 5 Case: 17-11455 Date Filed: 06/29/2020 Page: 6 of 89 Ophthalmologist and retinal specialist Elias Mavrofrides discovered much the same thing. He examined at least thirty of Pon’s patients and determined that, although many reported having undergone repeated laser treatment by Pon for WMD, their eyes showed no signs of the disease. Many of Pon’s patients told Dr. Mavrofrides that they were not sure why Pon was lasering their eyes, but “were told that they would lose vision without treatment.” Dr. Mavrofrides thought that Pon’s reported use of lasers on his patients “over and over and over [was] extremely atypical or unusual.” In 2008 an optometrist referred to Dr. Mavrofrides a patient Pon had diagnosed with WMD and lasered eight months in a row. After examining the patient, Dr. Mavrofrides wrote a letter to the optometrist stating that he “honestly d[id] not see any necessity for the [laser] treatments [the patient] has had.” Sometime before the fall of 2011 the government discovered Pon’s scheme. It happened when Special Agent Christian Jurs conducted a data analysis to determine whether any doctors were billing Medicare under codes associated with what he was told were outdated WMD treatment methods, including laser photocoagulation.3 That analysis revealed that Pon was a “significant outlier” with 3 Special Agent Jurs is a Medicare fraud investigator who became an agent with the United States Department of Health and Human Services after having worked as a naval intelligence officer and as an agent for the Naval Criminal Investigative Service. He has worked with Medicare data on numerous occasions since he became a HHS special agent nearly two decades ago. 6 Case: 17-11455 Date Filed: 06/29/2020 Page: 7 of 89 respect to the Medicare claims he submitted under 67220, the billing code for laser photocoagulation. In 2010, for example, Pon had submitted claims under that code for approximately 93 percent of his Medicare patients, while his ophthalmologist peers had submitted claims under that code for an average of only seven- hundredths of one percent of their patients. That is a disparity of about 132-to-1. The disparity prompted Agent Jurs to run Pon’s name through a complaint database, which showed that an unidentified person had lodged a complaint about Pon’s WMD treatment. And that, in turn, prompted him to interview approximately thirty doctors who had seen patients whom Pon had diagnosed with WMD and micropulse lasered. After Agent Jurs conducted some of those interviews, the government obtained a warrant to search Pon’s offices. In September 2011 federal law enforcement officers executed the search warrant and seized Pon’s patient files along with thousands of photographs and videos of his patients’ eyes. The next month the Centers for Medicare and Medicaid Services sent Pon a letter notifying him that it had suspended his Medicare payments based on what it identified as “credible allegations of [health care] fraud.” The suspension letter stated that between 2004 and 2011 Pon had submitted Medicare claims under codes associated with laser photocoagulation treatment and a type of WMD diagnostic test “that were disproportionate to claims submitted by other ophthalmologists for these codes,” and “that many of [Pon’s] 7 Case: 17-11455 Date Filed: 06/29/2020 Page: 8 of 89 patients did not have the underlying medical conditions that would support the procedures represented by these codes.” Sometime after Pon’s Medicare payments were suspended, Agent Jurs retained an expert, Dr. Thomas Friberg, to review the photos and videos of the eyes of about 500 patients whom Pon had diagnosed with WMD. Dr. Friberg was asked to review whether Pon’s patients did, in fact, have WMD. Dr. Friberg had obtained his medical degree from the University of Minnesota, completed an ophthalmology residency at Stanford University Medical Center, and held fellowships at the Harvard Medical School and the Duke University Eye Center. He is a professor of ophthalmology and a professor of bioengineering at the University of Pittsburgh, has authored or co-authored more than 175 articles in peer-reviewed publications, and has, over the course of his four-decade career, received more than $7 million in grants to study age-related eye diseases, including WMD. Dr. Friberg’s review of the more than 10,000 images of the eyes of patients whom Pon had diagnosed with WMD and “treated” took him about a year to complete. He found what he saw “shocking.” He realized that he “was looking at hundreds of images of patients that had nothing wrong with their eye[s].” According to Dr. Friberg, maybe “five to ten” –– only one or two percent –– of the 500 patients whom Pon had diagnosed with WMD actually had any form of 8 Case: 17-11455 Date Filed: 06/29/2020 Page: 9 of 89 macular degeneration. Dr. Friberg concluded that Pon had shown a “reckless disregard for his patients.” II. PROCEDURAL HISTORY In April 2014 a federal grand jury returned a twenty-count indictment against Pon. Each count charged him with health care fraud, in violation of 18 U.S.C. § 1347.4 The indictment alleged that he committed health care fraud by falsely diagnosing eleven of his patients with WMD and using those false diagnoses as a basis for submitting a total of twenty Medicare reimbursement claims (a different one described in each count) for performing both additional WMD testing that he knew was medically unnecessary and laser sessions that could not actually treat the disease even if the patients had been suffering from it, which they had not.5 Pon faced a statutory maximum of ten years per count. 18 U.S.C. § 1347(a). He went to trial. 4 Section 1347 makes it a crime to “‘knowingly and willfully’ engage in a scheme (1) ‘to defraud any health care benefit program’ or (2) to use false pretenses to obtain money from ‘any health care benefit program,’ ‘in connection with the delivery of or payment for health care benefits, items, or services.’” United States v. Bergman, 852 F.3d 1046, 1065 (11th Cir. 2017) (quoting 18 U.S.C. § 1347). 5 To protect their privacy, the indictment does not refer to the patients by name, and we will follow that lead by referring to them as the eleven patients, or individually as Patient One, Patient Two, and so on. 9 Case: 17-11455 Date Filed: 06/29/2020 Page: 10 of 89 A. The Daubert Hearing Before trial Pon notified the government that he intended to offer the expert testimony of Giorgio Dorin, a former director of development at the company that manufactured the laser Pon used. Dorin’s proposed testimony boiled down to two main points. First, he would testify about the general concepts of lasers and their application to eye diseases. Second, he would testify about a “newer method” of treating WMD that Pon said he had used on his patients: “subthreshold micropulse laser photostimulation.” Dorin would opine that the newer method could be used to close feeder vessels in a way that, unlike laser photocoagulation, would not leave a scar. Dorin chose to use the term “photostimulation” because the term “photocoagulation” suggests that the treatment necessarily produces a visible scar. The object of Dorin’s testimony would be to show that the low-power laser sessions Pon had subjected his patients to could actually treat WMD and could do so without leaving a telltale scar. The government moved to exclude Dorin’s proposed testimony under Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993), after its medical and laser experts advised it that “Dorin’s assertion regarding subthreshold micropulsed laser ‘photostimulation’ is not a medically accepted standard of care for the treatment of [WMD].” The government stated in its motion “that the term now used by the defense, laser ‘photostimulation[,]’ is apparently an attempt to 10 Case: 17-11455 Date Filed: 06/29/2020 Page: 11 of 89 somehow legitimatize the fact that there is no scarification in the retinas of [Pon’s] victims.” The district court held a three-day Daubert hearing. Dorin was the only witness for the defense. The government put on two witnesses — Dr. Friberg and David Buzawa, the co-founder of the company that manufactured Pon’s laser. All three of the witnesses testified about whether subthreshold micropulse laser photostimulation (which left no scar) could treat WMD. At the hearing, Dorin testified that the use of subthreshold micropulse laser photostimulation to treat WMD was “[t]heoretical[].” He conceded that he was unaware of any journal article stating that it is possible to use subthreshold micropulse photostimulation with Pon’s laser to treat WMD. But he talked about previous research that he said at least supported the idea. He described a study showing that subthreshold micropulse laser treatment was effective for diabetic macular edema, but that is a different medical condition than WMD. Defense counsel then asked Dorin whether he could, as a result of that one study involving diabetic macular edema, infer that subthreshold micropulse laser treatment was effective for WMD. Dorin responded: “I don’t see why . . . not, but we don’t have so many paper[s] published yet. I have not seen the clinical data.” Dorin also discussed a paper he authored that a peer-reviewed ophthalmology journal published in 2004, when Dr. Friberg was the journal’s 11 Case: 17-11455 Date Filed: 06/29/2020 Page: 12 of 89 editor-in-chief. In the paper Dorin stated that “laser pulses” could treat WMD by closing feeder vessels and could do so without causing “retinal burns.” Dorin’s paper concluded, however, that “[i]t does take a ‘leap of faith’ to accept that a retina with serious disease can be treated with . . . sub-visible-threshold protocols and without the use of additional pharmacological agents.” At the Daubert hearing, the government called Dr. Friberg. Dr. Friberg testified that Dorin’s proposed theory about subthreshold micropulse laser photostimulation had not been tested and had “absolutely not” gained general acceptance in the scientific community. He also testified that Dorin’s statement that subthreshold micropulse laser photostimulation can produce the same beneficial effect as laser photocoagulation was “crazy.” When asked whether there was even one study showing that the type of laser Pon used has closed a feeder vessel without leaving a scar while in micropulse mode, Dr. Friberg responded: “No.” He believed that there was “absolutely no acceptance in the ophthalmological community for Dorin’s conjecture.” The government’s other witness at the Daubert hearing, David Buzawa, held nine patents related to laser technology and had nearly forty years of experience with lasers. He was also the co-founder of the company that manufactured Pon’s laser. Buzawa testified that the power settings that Pon said he set his laser to before each laser session were “unusually low for micropulse.” Buzawa also stated 12 Case: 17-11455 Date Filed: 06/29/2020 Page: 13 of 89 that he had “never heard of or read a publication of successful feeder vessel closure using micropulse [mode] at [Pon’s] settings or any settings.” He concluded that feeder vessel treatment always creates a visible scar, according to “every paper and presentation that [he had] heard or read.” At the end of the Daubert hearing, defense counsel admitted that his proffered witness “Dorin [was] drawing conclusions that have not yet been scientifically tested.” The district court ruled that Dorin could testify about the general concepts of lasers and their applications. But he could not offer his opinion that subthreshold micropulse laser photostimulation could treat WMD by closing feeder vessels. The court found that Dorin’s proposed opinion was “conjecture” and ruled that it did not satisfy either Federal Rule of Evidence 702 or Daubert. Then came the jury trial. B. The Trial The jury trial began in September of 2015. The government presented nineteen witnesses and introduced more than 700 exhibits during its seven-day case-in-chief. Dr. Friberg was one of the government’s first witnesses. He began with an overview of the science behind WMD and how it is diagnosed and treated. He explained, for example, that laser photocoagulation, which is billed under Medicare code 67220, is a WMD treatment that can be used to “cook” feeder 13 Case: 17-11455 Date Filed: 06/29/2020 Page: 14 of 89 vessels. He testified that laser photocoagulation did not “have much uptake” in the medical community because the laser leaves behind a scar that can itself impair vision. Dr. Friberg also testified that he was “very certain” that if a doctor uses “enough [laser] energy . . . to cook the feeder vessel, you’re going to leave a [scar].” And he stated that he was “[a] hundred percent” certain that Pon’s micropulse laser technique could not close a feeder vessel. Dr. Friberg also explained the science behind drug injections, another WMD treatment that was introduced after laser photocoagulation. He testified that “there is no [WMD] treatment that we know of that is better than these [drug injections].” Dr. Friberg then described how Agent Jurs had retained him to review the photos and videos of the eyes of the approximately 500 patients whom Pon had diagnosed with WMD and “treated” with his micropulse laser technique. Dr. Friberg described how his review of those photos and videos revealed that the vast majority of those patients did not actually have WMD. For that reason, what he saw in his review “got [him] kind of mad.” According to Dr. Friberg, Pon was unnecessarily lasering the eyes of patients for a disease they did not have. Dr. Friberg discussed in front of the jury hundreds of images of the eyes of the eleven patients identified in the indictment — patients Pon had diagnosed with WMD and for whom Pon had billed Medicare under code 67220 for laser photocoagulation treatment for that disease. Dr. Friberg testified that based on his 14 Case: 17-11455 Date Filed: 06/29/2020 Page: 15 of 89 review of the images he did not see any indication that any of those eleven patients had WMD or any evidence of the scarring that necessarily results from laser photocoagulation treatment. See infra pp. 41–56. The government also presented the testimony of other doctors who had personally examined the eleven patients. With each of those doctors the government went patient by patient, eliciting testimony about each patient; the doctors explained to the jury the medical records they had created to document their examinations of the patients. See id. None of those doctors concluded that any of the eleven patients had WMD when Pon diagnosed them with it. Some of those patients did not even have dry age-related macular degeneration, which typically precedes WMD. And several doctors echoed Dr. Friberg’s conclusion; they testified that treating WMD by using laser photocoagulation always leaves a scar indicating the use of that treatment method — a scar that the doctors testified they did not see in the eyes of the eleven patients listed in the indictment, even though Pon had billed Medicare under code 67220 for the laser photocoagulation treatment of each patient. One doctor read from a medical record of his own examination of one of the eleven patients. Pon had billed Medicare more than ten times under code 67220 for the laser photocoagulation of that patient. In his own medical record, the doctor stated that both of the patient’s retinas were “completely normal.” He 15 Case: 17-11455 Date Filed: 06/29/2020 Page: 16 of 89 testified that he had seen many of Pon’s patients for a second opinion, and all of them “had a history of multiple [billed laser photocoagulation] procedures without any visible indication for such procedures or any clinical evidence that such procedures had been performed.” The jury also heard from some of the eleven patients themselves. One of them testified that only after a second doctor confirmed that she did not actually have WMD was she relieved of the shock and fear she felt as a result of her diagnosis. Another testified that Pon never gave her the option of drug injections instead of laser treatments, and she “felt terrible” after another doctor told her that she did not have WMD. Despite the fact that Pon had diagnosed both of those patients with WMD years before the trial, and they had never received any medically recognized treatment for it, each of them testified that they still had the requisite vision to (and did) drive a car. After the government rested, Pon himself took the stand. He testified that he ran so many diagnostic tests (and billed Medicare for them) because he wanted to be “as comfortable as possible not to miss a diagnosis.” He claimed that the reason the patients did not have a scar in their eyes showing that they had undergone the laser photocoagulation treatment he had billed Medicare for was his use of what he termed his “miraculous treatment.” As had Dorin at the Daubert hearing, Pon testified that his micropulse laser technique could not only treat WMD but could 16 Case: 17-11455 Date Filed: 06/29/2020 Page: 17 of 89 do so without leaving a telltale scar. His technique, Pon explained, was to set his laser to the lowest power settings before each session and then adjust those settings during the session “until [he] s[aw] the reaction that [he] want[ed] to get.” He said that he had attended a presentation “in the early 2000s sometime” that discussed this purported treatment, which he described as “the most fantastic news [he had] ever heard,” and “a major breakthrough.” He did not, however, have any memory of when or where the presentation took place. None. After hearing the presentation, Pon was “real enthusiastic” that he could use the technique described “to help a lot more patients.” Pon said that he believed the technology used to perform the technique described in the presentation was “the greatest thing since sliced bread,” so he purchased the necessary equipment. But he later testified that “it all happened about the same time,” so he couldn’t “recall exactly” whether he bought the new equipment before or after the presentation he claimed had inspired him. Pon said that once he had the equipment, he began to practice the technique he had heard about in the presentation. He first attempted the technique “on the patients that [we]re very advanced, so there’s really not very much downside risk there.” Eventually he “got the procedure to evolve” so that the risk was “almost zero.” He did not say whether he had billed Medicare under code 67220 for laser 17 Case: 17-11455 Date Filed: 06/29/2020 Page: 18 of 89 photocoagulation on these early, experimental patients as he had billed Medicare for laser photocoagulation on the patients listed in the indictment. Pon testified that an important part of his purported treatment was his ability to identify and diagnose WMD in its very early stages. He claimed that by using the new technology he could “visualize, directly visualize, these blood vessels” that caused WMD. But, even with the new technology, “finding the feeder vessel . . . is extremely difficult, even for someone” with a lot of experience diagnosing WMD. Pon testified that learning how to accurately interpret the diagnostic images and identify the feeder vessel locations “takes a lot of learning,” and “takes years, literally years, to learn how to do.” Pon admitted he did not “know of anybody specifically” who used the same micropulse laser technique to treat WMD. He said that one reason he might be the only one doing so is that the technique has “a steep learning curve.” But he was “too busy” — despite taking at least two months of vacation each year — and didn’t “feel obligated” to publish anything describing his technique or the treatment results he was achieving. After Pon’s testimony, the defense called thirteen of his patients and the spouse of one deceased patient. Only one of those patients was among the eleven listed in the twenty counts of the indictment. They generally testified that they thought Pon was generous and trustworthy and that their vision had improved after 18 Case: 17-11455 Date Filed: 06/29/2020 Page: 19 of 89 seeing him. They were not, of course, qualified to testify whether they had ever actually had WMD. After the government presented some rebuttal testimony, Pon presented some surrebuttal testimony, and the case eventually went to the jury, which found him guilty on all twenty counts. C. The Sentencing At Pon’s sentence hearing, the district court rejected the probation office’s finding and the government’s argument that the amount of loss attributable to Pon’s fraud scheme was more than $11 million. Instead, the court found that the loss amount was approximately $7 million, which resulted in an 18-level enhancement to the base offense level. The court’s application of that and other enhancements yielded an advisory guidelines range of 121 months to 151 months in prison. The court sentenced Pon to 121 months in prison on each count, to run concurrently. III. THE CONVICTION ISSUES Pon raises two contentions about his convictions, neither of which questions the sufficiency of the evidence to convict him. First, he contends that under Daubert and Federal Rule of Evidence 702, the district court should have allowed his expert to testify about the use of subthreshold micropulse photostimulation as a treatment for WMD. Second, Pon contends that the district court should not have 19 Case: 17-11455 Date Filed: 06/29/2020 Page: 20 of 89 allowed the government to present rebuttal evidence showing that he billed Medicare for performing services on patient J.L.’s blind left eye. In the alternative, he argues that even if that rebuttal evidence was properly admitted, the court should have allowed him to present all of his surrebuttal evidence. A. The Daubert Issue It is not easy to persuade a court of appeals to reverse a district court’s judgment on Daubert grounds. United States v. Brown, 415 F.3d 1257, 1264 (11th Cir. 2005). Doing so is tough toil because the “theme that shapes appellate review in this area is the limited nature” of that review. Id. We review evidentiary decisions under the abuse of discretion standard. Id. at 1264–65 (citing Gen. Elec. Co. v. Joiner, 522 U.S. 136, 141 (1997)). And under that standard district courts have a “significant” range of choice, which is to say that we defer to their evidentiary “decisions to a considerable extent.” Id. at 1265; accord McCorvey v. Baxter Healthcare Corp., 298 F.3d 1253, 1257 (11th Cir. 2002) (“[O]ur review of evidentiary rulings by trial courts . . . is very limited.”) (quotation marks omitted). The deference we show trial courts on evidentiary rulings is especially pronounced in the Daubert context, where the abuse of discretion standard places a “heavy thumb” — “really a thumb and a finger or two” — “on the district court’s side of the scale.” Brown, 415 F.3d at 1268. That’s done for a number of reasons. The district court occupies the best position to rule on Daubert issues given its 20 Case: 17-11455 Date Filed: 06/29/2020 Page: 21 of 89 familiarity “with the procedural and factual details” of the trial, which it presides over and is immersed in. Id. at 1266. The rules that control the admission of expert testimony “must be applied in case-specific evidentiary circumstances that often defy generalization.” Id. And deference maintains the importance of the trial and discourages appeals of rulings about expert witness testimony. See id. As a result, “the task of evaluating the reliability of expert testimony is uniquely entrusted to the district court,” and we must grant “the district court considerable leeway in the execution of its duty.” Rink v. Cheminova, Inc., 400 F.3d 1286, 1291 (11th Cir. 2005) (citations and quotation marks omitted). We do so mindful, of course, that granting that leeway “is not the same thing as abdicating appellate responsibility.” Brown, 415 F.3d at 1266. After holding a three-day Daubert hearing, the district court found that Dorin’s theory that subthreshold micropulse photostimulation could treat WMD by closing feeder vessels was unreliable. As a result, it allowed Dorin to testify about only the general concepts of lasers and their application to eye diseases. That ruling was not an abuse of discretion. Federal Rule of Evidence 702 “controls the admission of expert testimony.” United States v. Frazier, 387 F.3d 1244, 1259 (11th Cir. 2004) (en banc). Under that rule expert witnesses may testify if, among other things, their “testimony is the product of reliable principles and methods.” Fed. R. Evid. 702(c). Determining 21 Case: 17-11455 Date Filed: 06/29/2020 Page: 22 of 89 whether expert testimony is the product of “reliable principles and methods” is the province of the Daubert test. See City of Tuscaloosa v. Harcros Chems., Inc., 158 F.3d 548, 562 (11th Cir. 1998) (stating that courts must determine whether “the methodology by which the expert reaches his conclusions is sufficiently reliable as determined by the sort of inquiry mandated in Daubert”). Daubert instructs that a reliability determination involves four main inquiries about the expert’s theory or technique: “(1) whether it can be (and has been) tested; (2) whether it has been subjected to peer review and publication; (3) what its known or potential rate of error is, and whether standards controlling its operation exist; and (4) whether it is generally accepted in the field.” Brown, 415 F.3d at 1267 (citing Daubert, 509 U.S. at 593–94). In Joiner the Supreme Court added another inquiry to gauge reliability: whether there is “an analytical gap between the data and the opinion proffered.” 522 U.S. at 146; accord McDowell v. Brown, 392 F.3d 1283, 1300 (11th Cir. 2004) (“[A]n expert opinion is inadmissible when the only connection between the conclusion and the existing data is the expert’s own assertions . . . .”) (citing Joiner, 522 U.S. at 146). If the analytical distance between the data and the opinion proffered “is simply too great,” a court may conclude that the opinion is unreliable. Joiner, 522 U.S. at 146. 22 Case: 17-11455 Date Filed: 06/29/2020 Page: 23 of 89 Here, three of the four Daubert factors weigh against the reliability of Dorin’s theory. First, Dorin testified that although his theory could be tested, it has not been. Dr. Friberg agreed. So did defense counsel, who conceded that “Dorin [was] drawing conclusions that have not yet been scientifically tested.” And like Pon, Dorin acknowledged that he did not “know of anybody doing it.” Second, Pon failed to provide evidence about the theory’s known or potential rate of error and whether any standards exist to control for error. Dorin himself acknowledged that he “ha[d] not seen . . . clinical data” about it. And third, the record shows that Dorin’s theory is not generally accepted in the ophthalmology field. Dr. Friberg testified that the theory has “[a]bsolutely not” gained that acceptance. Buzawa’s testimony that “[f]eeder vessel treatment has always been superthreshold,” instead of subthreshold, according to “every paper and presentation that [he had] heard or read” confirmed the point. That leaves only one factor that weighs — ever so slightly — in favor of reliability: Dorin’s peer-reviewed paper mentioning the theory. But just as “[p]ublication . . . is not a sine qua non of admissibility,” Daubert, 509 U.S. at 593, publication alone is not enough to conclude that a district court abused its discretion in not admitting expert testimony, see Allison v. McGhan Med. Corp., 184 F.3d 1300, 1313 (11th Cir. 1999). 23 Case: 17-11455 Date Filed: 06/29/2020 Page: 24 of 89 A further indication that Dorin’s theory was unreliable is the analytical gap between it and the research that Dorin said supports it. See Joiner, 522 U.S. at 146. At the hearing, Dorin suggested that because a study showed subthreshold micropulse laser treatment can treat diabetic macular edema, his theory that it could treat WMD is sound. That is a “leap[] from an accepted scientific premise to an unsupported one.” Allison, 184 F.3d at 1314. Diabetic macular edema is a different condition than WMD. Between the premise that subthreshold micropulse laser treatment can treat the first condition and Dorin’s theory that it can treat the second one, there is an analytical gap. Instead of properly bridging that gap, Dorin tried to ipse dixit over it; but a bald assertion cannot carry the Daubert burden. See Williams v. Mosaic Fertilizer, LLC, 889 F.3d 1239, 1249 (11th Cir. 2018) (“[N]othing in either Daubert or the Federal Rules of Evidence requires a district court to admit opinion evidence that is connected to existing data only by the ipse dixit of the expert.”) (quoting Joiner, 522 U.S. at 146). The district court concluded that “there [was] simply too great an analytical gap between the data and the opinion proffered.” Joiner, 522 U.S. at 146. That was not an abuse of discretion but a proper exercise of the “considerable leeway” the court had. Kumho Tire Co. v. Carmichael, 526 U.S. 137, 152 (1999).6 6 Pon also argues that the district court violated his Sixth Amendment right to put on a meaningful defense because it “mechanistically” applied the Daubert factors. We disagree. The 24 Case: 17-11455 Date Filed: 06/29/2020 Page: 25 of 89 B. The Rebuttal Issue Pon also contends that his convictions must be reversed because the district court erred in two of the evidentiary rulings it made during his fourteen-day trial: (1) allowing the government to present rebuttal evidence showing that he billed Medicare for performing certain medical services on J.L., who was one of his patients, and (2) partially limiting the scope of his surrebuttal evidence about J.L. We review both of those evidentiary rulings only for an abuse of discretion. See Frazier, 387 F.3d at 1270; United States v. Haimowitz, 706 F.2d 1549, 1560 (11th Cir. 1983). And an abuse of discretion does “not warrant reversal where the resulting error was harmless.” United States v. Barton, 909 F.3d 1323, 1330 (11th Cir. 2018). One of the fourteen defense witnesses (not counting Pon) was J.L., who took the stand on the eleventh day of the trial. Like some of Pon’s other patients who appeared as defense witnesses, J.L. testified about his history as Pon’s patient and described him as generous and trustworthy. He also told the jury that in 1994 both of his retinas had detached because of a complication from diabetes and Pon had performed surgery that year on each eye — the left eye in July and the right eye in August. Pon knew J.L. was unemployed and uninsured at the time and performed district court gave ample reasons why three of those factors weighed against the reliability of Dorin’s theory and outweighed the fourth factor. 25 Case: 17-11455 Date Filed: 06/29/2020 Page: 26 of 89 the surgeries without any guarantee he would be paid. The 1994 surgery on J.L.’s right eye was successful, but the one on his left eye was not; he lost all of his sight in it soon after that surgery. Defense counsel also put into evidence excerpts of Pon’s treatment records for J.L. Those records included logs in which Pon documented procedures he had performed on J.L.’s right eye between 2004 and 2015. They also indicated that Pon had diagnosed J.L. with WMD in 2009 and had micropulse lasered J.L’s right eye several times. The excerpts from Pon’s records of treating J.L. did not list any tests or procedures that Pon had performed on J.L.’s blind left eye. During cross-examination, the prosecutor asked J.L. when Pon had last performed any procedure on his left eye, the one in which he had been completely blind for more than twenty years. J.L. said it had “been a couple years”; he said Pon did a “regular eye check on it,” which involved an exam with “the eye charts,” an “ultrasound,” and “pictures,” but that Pon had done no “major procedures” on his left eye. The prosecutor then asked J.L. whether Pon had ever done “any kind of an injection” or any “dye tests” in J.L.’s left eye. J.L. responded: “No. I’ve never had anything done in my left eye.” On redirect, J.L. reiterated that Pon had not done “any tests on [his] left eye,” including fluorescein angiograms, but Pon had examined that eye “a couple times” by looking at it “through a lens” to see if he could improve the vision in it. J.L.’s firm testimony that Pon had never done 26 Case: 17-11455 Date Filed: 06/29/2020 Page: 27 of 89 any fluorescein angiograms on his left eye was significant because it contradicted the bills Pon submitted to Medicare on at least a half dozen occasions — bills for performing fluorescein angiogram tests on J.L.’s totally blind left eye. See infra pp. 30–32. In light of J.L.’s testimony and the medical records Pon’s attorney had put into evidence, the government sought to introduce through one of its agents rebuttal evidence in the form of a spreadsheet and related testimony. Agent Jurs had created the spreadsheet after J.L.’s testimony by looking through Pon’s Medicare claims history for billings related to J.L.’s left eye after 1994, when he had lost all of his sight in it. The billing records for the decade between 1994 and 2004 were not available,7 but the records from 2004 until the trial in 2015 were. The spreadsheet showing the billings and Agent Jurs’ testimony established that Pon had billed Medicare for performing services on J.L.’s left eye –– in which he had been blind since 1994 –– 52 times between 2004 and 2015. Those 52 billings totaled approximately $19,500 and, aside from three billings for a January 2009 surgery on the blind eye (which Pon would later testify were merely the 7 Agent Jurs testified that he wasn’t certain why he had not been able to access Pon’s Medicare billing records before 2004. He thought the most likely explanation was that the Medicare system did not give records that “go back, you know, to when the earth cooled. They usually go [back] a five- or ten-year time frame.” 27 Case: 17-11455 Date Filed: 06/29/2020 Page: 28 of 89 result of clerical errors), all of the billings were for ophthalmic ultrasounds, fluorescein angiograms, and fundus photography, which are diagnostic tests. Pon billed Medicare all of those times and for all of that money claiming that he had done procedures on J.L.’s blind left eye, including fluorescein angiograms, even though his own witness, J.L., testified that Pon never performed a fluorescein angiogram or any test involving the injection of dye on his left eye. All Pon had ever done on that eye, J.L. insisted, was perform a “regular eye check on it,” and examine it by looking through a lens “a couple of times.” His testimony evidenced, at the very least, that Pon’s six billings for fluorescein angiograms, which involve the injection of dye, on J.L.’s left eye were fraudulent. Defense counsel moved to exclude the spreadsheet and Agent Jurs’ testimony about it, arguing that the evidence was not proper rebuttal. The district court allowed the government to present the rebuttal evidence because the court recalled (incorrectly, it turned out) that the defense, during J.L.’s testimony, had been the first to bring up the fact that Pon had treated J.L.’s blind left eye. Pon asserts that ruling was reversible error. It wasn’t. Rule 611(a) vests district courts with authority to “exercise reasonable control over the mode and order of examining witnesses and presenting evidence.” Fed. R. Evid. 611(a). District courts have “broad discretion” in exercising that authority and will not be reversed except for abuse of that discretion. United States 28 Case: 17-11455 Date Filed: 06/29/2020 Page: 29 of 89 v. Hill, 643 F.3d 807, 845 (11th Cir. 2011); accord Haimowitz, 706 F.2d at 1560. Pon presented J.L.’s testimony as an example of how he had treated a patient out of the goodness of his heart and not for a profit motive. In light of that, the district court did not abuse its discretion in admitting billings Pon had generated for services he claimed to have rendered on that patient’s blind eye. C. The Surrebuttal Issue 1. The Procedural Facts Pon alternatively asked the district court, if it was not going to keep out the government’s rebuttal evidence, to let Pon retake the stand and present as surrebuttal evidence his explanation for billing Medicare for services on J.L.’s blind left eye. The court reserved a ruling on that alternative request until after the government presented its rebuttal evidence. After the government did so, Pon renewed his request to present surrebuttal testimony, arguing that he had not had an “opportunity to respond” to the “impression” the government created that Pon improperly conducted and billed for treatments on a blind eye. The district court characterized the government’s rebuttal evidence as “very damning” and stated that it had “this idea of fairness tug[ging]” at it, but its “inclination” was to deny the defense’s request that Pon be allowed to testify again. The court also asked “the further question of whether, in deference to . . . Dr. Pon’s Sixth Amendment right . . . he should be given an 29 Case: 17-11455 Date Filed: 06/29/2020 Page: 30 of 89 opportunity to offer an explanation” about the billed procedures for J.L.’s blind left eye. The court gave the attorneys the weekend to research the surrebuttal issue and indicated it would do the same. On Monday, the court heard argument from both sides about whether it had erred by allowing the government’s rebuttal testimony and, if so, how any error should be remedied. Neither side mentioned the Sixth Amendment or any other constitutional right that Pon might have to present surrebuttal testimony. Before ruling, the court allowed Pon to proffer the testimony he wished to present in response to the government’s rebuttal evidence. He took the stand in a testimonial proffer, stating that he had performed services on J.L.’s blind left eye to determine whether it had a problem that could lead to complications that might result in a complete loss of vision in the right eye. Pon testified that ophthalmic ultrasounds, which made up 41 of the 52 billings in the spreadsheet, are used to look for abnormalities in the eye. And in J.L.’s case, “it’s very important to examine his left eye and continue to examine his left eye” to make sure he was not developing sympathetic ophthalmia — a condition that can lead to blindness in both eyes. For that reason, Pon said that it was necessary to examine J.L.’s left eye “periodically.” But Pon didn’t explain what “periodically” meant, and he didn’t explain why he needed to examine J.L.’s left eye as frequently as he billed Medicare for doing. The records proved that Pon 30 Case: 17-11455 Date Filed: 06/29/2020 Page: 31 of 89 sometimes billed Medicare for performing ophthalmic ultrasounds on J.L.’s left eye multiple times in one month and, on at least four occasions, twice on the same day. When asked about the six fluorescein angiograms he had billed Medicare for conducting on J.L.’s blind left eye, Pon said they are “a very useful test to help diagnose sympathetic ophthalmia.” Of course, J.L. himself had emphatically testified that Pon never performed a fluorescein angiogram on his left eye. Pon didn’t offer any explanation for that contradiction. Finally, Pon also proffered that the January 2009 surgery was done on J.L.’s right eye, not his left as Pon had billed Medicare. He said that billing for surgery on the left eye resulted from a “clerical error.” After the proffer, defense counsel argued that it would be “incorrect” and “misleading” to prevent Pon from testifying about the clerical error in the billing of the January 2009 surgery. The district court agreed and decided to let Pon testify that he billed for surgery on the wrong eye as a result of a clerical error. The court would not allow Pon to testify about any of the numerous non-surgical services on the blind left eye and why he claimed they were necessary. After the jury returned, Pon took the stand and testified that the billing for surgery on J.L’s left eye in 2009 was a clerical error, and that surgery actually had been performed on J.L.’s right eye. On cross-examination, Pon admitted that those 31 Case: 17-11455 Date Filed: 06/29/2020 Page: 32 of 89 three 2009 clerical errors about the surgical services were “just three entries out of two pages of entries,” and that the total billing for those three entries was “less than $3,000.” The two pages of entries showed that Pon had billed $16,441 for the 49 non-surgical services on J.L.’s blind left eye — billings that Pon never claimed were the result of clerical error. During closing arguments to the jury, the government spent its time and aimed its arguments at the evidence involving the fraudulent billings for treatments Pon claimed to have rendered to the eleven patients listed in the indictment — evidence it had presented through nineteen witnesses and nearly 760 exhibits over seven days of the fourteen-day trial. It mentioned J.L. (who had testified for less than one hour of the trial), but just barely. It described him only as someone Pon had incorrectly diagnosed with WMD. It said nothing at all about any services that Pon had billed for J.L.’s blind left eye. Not one word. Defense counsel, by contrast, discussed J.L. at length in his closing argument. Referring to the surgery Pon had performed on J.L. in 1994, for which Pon did not know if he would get paid, counsel asked the jury: “Is that, ladies and gentlemen, what a fraudster would do?” And addressing the rebuttal evidence that Pon billed for services on J.L.’s blind left eye, counsel argued: I submit to you that the diagnostic tests and examinations that Dr. Pon did, whether it’s on the right eye, which still was viable, or the left eye, which he was legally blind in, are still tests that are appropriate for a 32 Case: 17-11455 Date Filed: 06/29/2020 Page: 33 of 89 doctor to do, and there’s absolutely nothing wrong with doing those tests and billing Medicare for it. However, he did not offer the jury any explanation for why J.L., the witness he called who would have been the subject of those tests — including tests involving dye injections — would testify under oath that at least some of them had never been performed. After closing arguments, the court instructed the jury, among other things, that Pon was “on trial only for the specific crimes charged in the indictment,” and the jury’s task was to determine whether Pon was “guilty or not guilty of those specific crimes.” The jury found him guilty of all twenty counts of health care fraud, none of which mentioned J.L. We review rulings about whether to allow surrebuttal evidence only for an abuse of discretion. See Frazier, 387 F.3d at 1270; Haimowitz, 706 F.2d at 1560. The district court’s ruling partially granted and partially denied Pon’s request to present surrebuttal evidence. The court granted his request to re-take the stand and testify that the three billings for surgery on J.L.’s left eye were clerical errors, and he had actually performed the surgery on J.L’s functioning right eye. But the court denied Pon’s request to testify that the reason he billed for so many diagnostic tests on J.L.’s blind left eye was to determine if there was a problem with it that could lead to complications that might result in a loss of vision in the right eye. 33 Case: 17-11455 Date Filed: 06/29/2020 Page: 34 of 89 2. The Preservation Issue Pon contends the court’s refusal to let him testify to that was both trial error and constitutional error. The constitutional error he claims is a violation of his Sixth Amendment right “to defend against the government’s evidence” and denial of a fair opportunity to respond to the government’s rebuttal testimony. Pon clearly preserved the trial error issue in the district court, but it is far from clear that he preserved the constitutional issue. “No procedural principle is more familiar . . . than that a constitutional right may be forfeited in criminal as well as civil cases by the failure to make timely assertion of the right before a tribunal having jurisdiction to determine it.” Yakus v. United States, 321 U.S. 414, 444 (1944). To preserve an error in a criminal trial, a party must “inform[] the court — when the court ruling or order is made or sought — of the action the party wishes the court to take, or the party’s objection to the court’s action and the grounds for that objection.” Fed. R. Crim. P. 51(b). Failing to contemporaneously object “ordinarily precludes the raising on appeal of the unpreserved claim of trial error.” Puckett v. United States, 556 U.S. 129, 135 (2009). Although a contemporaneous objection preserves an issue for appellate review, “not every objection is a constitutional objection.” United States v. Candelario, 240 F.3d 1300, 1304 (11th Cir. 2001). We have held over and over 34 Case: 17-11455 Date Filed: 06/29/2020 Page: 35 of 89 again that to preserve an issue, a litigant must “first clearly present it to the district court, that is, in such a way as to afford the district court an opportunity to recognize and rule on it.” See, e.g., Juris v. Inamed Corp., 685 F.3d 1294, 1325 (11th Cir. 2012) (quotation marks omitted). While Pon did preserve through objection and argument the issue of whether the district court’s partial limitation on his surrebuttal evidence violated the rules governing the presentation of rebuttal and surrebuttal evidence, he never once mentioned the Sixth Amendment or argued to the district court that the limitation violated that or any other constitutional provision. Our precedent indicates that an objection on nonconstitutional grounds is not enough to preserve a constitutional issue. For example, in United States v. Chau, 426 F.3d 1318 (11th Cir. 2005), we held that the defendant’s hearsay objection in the district court did not preserve the Confrontation Clause issue he pressed on appeal. Id. at 1321–22 (“[A] hearsay objection does not preserve the [Confrontation Clause] issue . . . .”). In United States v. Hawkins, 934 F.3d 1251, 1264 (11th Cir. 2019), we concluded that the “tepid objections made by defense counsel” and the “rumblings of concern about the phrasing of questions” did not preserve the argument the defendants made on appeal that the trial court had admitted improper opinion testimony. And in United States v. Elbeblawy, 899 F.3d 925, 938 (11th Cir. 2018), we held that a defendant had not preserved an argument when he only “mentioned it, in passing, in a post- 35 Case: 17-11455 Date Filed: 06/29/2020 Page: 36 of 89 trial reply motion” because the “post-trial remark was neither timely nor sufficiently developed” to preserve the issue. Pon did not even mention in passing to the district court the constitutional issue he wants to pursue before us. It is true that at one point the district court — not Pon — stated: “[T]here’s the further question of whether, in deference to . . . Dr. Pon’s Sixth Amendment right . . . he should be given an opportunity to offer an explanation at that point.” That probably is not enough to preserve the Sixth Amendment issue for appeal, even under the view of the treatise that the dissent relies on to reach the contrary conclusion. See Dissent at 8 n.4. The relevant part of that treatise states: [I]f the record reveals that the parties and the court were aware of the claim or issue and litigated it, then whether or not it served as the basis for determination the claim or issue was raised and is reviewable on appeal. Moreover, if the district court sua sponte raised an issue of law and explicitly resolved the issue on the merits, that ruling is fully reviewable on appeal even though no party raised it below. 19 James Wm. Moore et al., Moore’s Federal Practice § 205.05(1) (3d ed. 2019) (emphases added) (footnotes omitted). Like our precedent, the treatise states that the mere mention of an issue does not preserve it. Instead, the issue must have been “decide[d],” “litigated,” and “explicitly resolved . . . on the merits” to be preserved. Id. As our discussion of the surrebuttal argument in the district court has shown, the Sixth Amendment issue was not decided, litigated, or explicitly resolved on the 36 Case: 17-11455 Date Filed: 06/29/2020 Page: 37 of 89 merits there. Defense counsel did not even mention it; the government did not mention it; and the district court mentioned it only in passing in a single sentence. Even after the district court referred to it, defense counsel did not argue that Pon had a Sixth Amendment right to present surrebuttal evidence in these circumstances. The failure to do so is all the more significant because counsel objected on Sixth Amendment grounds to five other rulings against Pon on evidentiary issues during the trial.8 But he did not assert or even mention the Sixth Amendment in connection with the partial limitation on surrebuttal evidence. On this record, the author of the treatise might well say that the Sixth Amendment issue does not “fairly appear[] in the record as having been raised or decided.” Moore’s § 205.05(1). But we don’t have to decide if that Sixth Amendment issue was presented to the district court. We can assume that it was. We can make that assumption because even if partially limiting Pon’s surrebuttal did violate his Sixth 8 For example, in objecting to the district court’s ruling that he could not admit exhibits during cross-examination, counsel argued that “the restriction impinge[d] Dr. Pon’s Sixth Amendment constitutional right to cross-examination”; he moved for a mistrial after information came out about what a nontestifying doctor said, invoking the Sixth Amendment and arguing that Pon was “deprived of [his] right to confrontation”; in offering an exhibit that the government sought to have excluded, he argued that Pon’s “Sixth Amendment constitutional right to present evidence” entitled him to have the exhibit accepted; he sought the court’s permission for Pon to confer with counsel during breaks in his testimony, arguing that Pon was entitled to do so because of “the right to counsel under the Sixth Amendment”; and he again argued that one of Pon’s exhibits should have been admitted because the district court’s “exclusion of [the proffered exhibit was] a denial of Dr. Pon’s Sixth Amendment right to present evidence.” 37 Case: 17-11455 Date Filed: 06/29/2020 Page: 38 of 89 Amendment rights, that error was harmless beyond a reasonable doubt, which necessarily means that any nonconstitutional error from that limitation was harmless as well. 3. The Harmless Error Standard We review preserved assertions of error — both constitutional and nonconstitutional error — for harmlessness. See Fed. R. Crim. P. 52(a) (“Any error, defect, irregularity, or variance that does not affect substantial rights must be disregarded.”); United States v. Olano, 507 U.S. 725, 734 (1993) (“When the defendant has made a timely objection to an error and Rule 52(a) applies, a court of appeals normally engages in a specific analysis of the district court record — a so-called ‘harmless error’ inquiry — to determine whether the error was prejudicial.”). As this Court sitting en banc has recognized, the Supreme Court has repeatedly held that “the vast majority of constitutional errors that occur at a criminal trial, including Sixth Amendment violations, should be examined for prejudicial effect and those errors do not require reversal if they are harmless.” United States v. Roy, 855 F.3d 1133, 1167 (11th Cir. 2017) (en banc). The harmless error doctrine is important because it “promotes public respect for the criminal process by focusing on the underlying fairness of the trial.” Neder v. United States, 527 U.S. 1, 18 (1999) (quotation marks omitted). Review for 38 Case: 17-11455 Date Filed: 06/29/2020 Page: 39 of 89 harmlessness “is also essential to avoid a ‘sporting theory of justice’ and a regime of gotcha review.” Roy, 855 F.3d at 1142 (quoting United States v. Agurs, 427 U.S. 97, 108 (1976)). A constitutional error is harmless if the government proves “beyond a reasonable doubt that the error complained of did not contribute to the verdict obtained.” Chapman v. California, 386 U.S. 18, 22 (1967). And “[t]o say that an error did not contribute to the verdict is . . . to find that error unimportant in relation to everything else the jury considered on the issue in question, as revealed in the record.” Yates v. Evatt, 500 U.S. 391, 403 (1991); accord Cape v. Francis, 741 F.2d 1287, 1294–95 (11th Cir. 1984) (“If, upon its reading of the trial record, the appellate court is firmly convinced that the evidence of guilt was so overwhelming that the trier of fact would have reached the same result without the tainted evidence, then there is insufficient prejudice to mandate the invalidation of the conviction.”). A nonconstitutional error, on the other hand, is harmless unless it “resulted ‘in actual prejudice because it had substantial and injurious effect or influence in determining the jury’s verdict.’” United States v. Guzman, 167 F.3d 1350, 1353 (11th Cir. 1999) (quoting United States v. Lane, 474 U.S. 438, 449 (1986)). If a reviewing court “can say with fair assurance . . . that the judgment was not substantially swayed by the [nonconstitutional] error,” the court must affirm even 39 Case: 17-11455 Date Filed: 06/29/2020 Page: 40 of 89 if the district court erred. United States v. Hornaday, 392 F.3d 1306, 1315–16 (11th Cir. 2004). Under both harmless error standards, the reviewing court examines the trial record in its entirety to make its prejudice determination. See Neder, 527 U.S. at 19 (constitutional standard); United States v. Sweat, 555 F.3d 1364, 1367 (11th Cir. 2009) (nonconstitutional standard). The government bears the burden of showing harmlessness in both situations when the issue was properly preserved by timely objection. See, e.g., United States v. Vonn, 535 U.S. 55, 62 (2002) (noting that it is the government’s “burden of showing that any error was harmless” under harmless-error review). As may be apparent, the constitutional harmless error hurdle is the higher of the two for the government to clear. See generally United States v. Mathenia, 409 F.3d 1289, 1291 (11th Cir. 2005) (noting “the less demanding [harmless error] test that is applicable to non-constitutional errors”); United States v. Robles, 408 F.3d 1324, 1327 (11th Cir. 2005) (noting that “[w]hen the error is of the constitutional variety, a higher standard is applied” than when the error is nonconstitutional). For that reason, a holding that a constitutional error is harmless necessarily means that it is also harmless if it happens to be nonconstitutional error. 40 Case: 17-11455 Date Filed: 06/29/2020 Page: 41 of 89 4. The Harmlessness of the Assumed Error We are persuaded that, even if the district court erred in partially limiting Pon’s surrebuttal evidence, and that error violated the Sixth Amendment, it was harmless beyond a reasonable doubt. See United States v. Willner, 795 F.3d 1297, 1322 (11th Cir. 2015) (“One circumstance in which courts find constitutional errors harmless beyond a reasonable doubt is when the evidence of the defendant’s guilt is ‘so overwhelming.’”) (quoting Harrington v. California, 395 U.S. 250, 254 (1969)). Here’s why. a. The Eleven Patients Listed in the Indictment First, the government presented a slew of compelling evidence that not a single patient out of the eleven identified in the indictment had WMD, yet Pon had diagnosed every one of them with that degenerative eye disease anyway. And those patients’ eyes showed no signs at all of having undergone laser photocoagulation treatment for WMD, though Pon had billed Medicare under code 67220 for laser photocoagulation treatment for each patient. The government presented the testimony of not one, but a dozen doctors about the patients listed in the indictment. Dr. Friberg and eleven other doctors who had examined the eleven patients identified in the indictment testified against Pon. All told, the twelve of 41 Case: 17-11455 Date Filed: 06/29/2020 Page: 42 of 89 them collectively had more than 330 years of experience.9 None of those doctors could find any evidence that any of the eleven patients identified in the indictment had WMD when Pon diagnosed them with it, and they all concluded that the patients did not have the telltale scars associated with the laser photocoagulation treatment that Pon had billed Medicare for performing on each patient. Because the overwhelming amount of the evidence is important, we recount in detail the doctors’ testimony about each of the eleven patients. (1) Patient One About Patient One, the jury heard testimony from three ophthalmologists: Dr. Friberg and two others. Dr. Friberg testified that before trial he reviewed images of both of Patient One’s eyes, each of which Pon had diagnosed with WMD and for each of which he had billed Medicare under code 67220 as though he had performed laser photocoagulation treatment. The images of the patient’s left eye (the basis for Count One) were taken both on and after the “treatment” date, and the images of the patient’s right eye (the basis for Count Two) were taken before, on, and after the “treatment” date. Dr. Friberg told the jury about his review of those images and explained that he did not see any indication at all that 9 Dr. Friberg: 32 years; Dr. Williams: 45 years; Dr. Berger: 31 years; Dr. Magruder: 26 years; Dr. Gills: 47 years; Dr. Pennachio: 30 years; Dr. Wehrly: 21 years; Dr. Mavrofrides: 11 years; Dr. Schwenk: 31 years; Dr. Beneke: 24 years; Dr. Vogel: 21 years; Dr. Kraut: 19 years. 42 Case: 17-11455 Date Filed: 06/29/2020 Page: 43 of 89 the patient had WMD in either of her eyes or a scar showing she had actually received laser photocoagulation treatment in either eye. The second ophthalmologist testified about five examinations he had conducted on Patient One after Pon’s diagnosis and purported treatment. He was “confident that [the patient] was not” suffering from WMD when he examined her, and he said she did not have the disease when Pon billed Medicare for laser photocoagulation treatment on her eyes. Patient One went to the third testifying ophthalmologist to get a second opinion about whether she had WMD. That ophthalmologist had examined Patient One on three separate occasions after Pon’s diagnosis and purported laser treatment of her eyes. He discussed with the jury the medical records of his examinations of Patient One. Based on those three examinations, he testified that she did not have WMD when Pon billed Medicare for laser photocoagulation treatment of her eyes nor did she have a laser-related scar in her eyes afterwards. None of Pon’s excluded surrebuttal evidence, which is the assumed error we are talking about, had anything to do with Patient One. (2) Patient Two About Patient Two, the jury heard from Dr. Friberg and two other ophthalmologists. Before trial, Dr. Friberg reviewed images of this patient’s left eye, which Pon had diagnosed with WMD and for which he had billed Medicare 43 Case: 17-11455 Date Filed: 06/29/2020 Page: 44 of 89 under code 67220 as though he had performed laser photocoagulation treatment on two separate occasions. The images Dr. Friberg reviewed were taken before, on, and after the first “treatment” date (the basis for Count Three), and before and on the second “treatment” date (the basis for Count Four). Dr. Friberg testified that based on his review there was “a high degree of medical certainty” that the patient had not had WMD. He also testified that he did not see any indication that the patient had a scar in her eye showing she had actually undergone laser photocoagulation treatment. The second ophthalmologist testified about six examinations he had conducted on Patient Two’s eye after Pon’s diagnosis and supposed treatment. After reviewing for the jury the medical records of his examinations, he testified that this patient did not have WMD when Pon billed Medicare for laser photocoagulation treatment of her eye. The third ophthalmologist testified that he had examined Patient Two on two separate occasions just months before Pon diagnosed her with WMD. He had not seen any evidence that the patient had WMD, and he testified that the chance of her developing the disease in the brief period between the time he examined her and the time Pon billed Medicare for laser photocoagulation of her eye was “very unlikely.” None of Pon’s excluded surrebuttal evidence, which is the assumed error we are talking about, had anything to do with Patient Two. 44 Case: 17-11455 Date Filed: 06/29/2020 Page: 45 of 89 (3) Patient Three Patient Three was also the subject of testimony from Dr. Friberg and two other ophthalmologists. Before trial, Dr. Friberg reviewed images of both of this patient’s eyes, each of which Pon had diagnosed with WMD and for each of which he had billed Medicare under code 67220 as though he had performed laser photocoagulation treatment. The images of the patient’s left eye (the basis for Count Five) were taken on and after the “treatment” date, and the images of the patient’s right eye (the basis for Count Six) were taken before, on, and after that date. Dr. Friberg testified that based on his review he did not see any indication that on any of those dates the patient had WMD or a scar indicating he had actually undergone laser photocoagulation treatment. The second ophthalmologist explained to the jury the medical records documenting some of Patient Three’s visits with him. He testified to his “shock[]” of learning that the patient had purportedly undergone laser photocoagulation treatment at Pon’s office just months after he himself had concluded that the patient did not have WMD. He also stated that he had examined the patient after Pon did and had concluded that the patient did not have WMD or any laser photocoagulation scars in his eyes. The third ophthalmologist testified that he examined this patient on three separate occasions more than four years after Pon’s diagnosis and purported 45 Case: 17-11455 Date Filed: 06/29/2020 Page: 46 of 89 treatment and, on each occasion, he saw no indication that the patient had WMD. He also testified that he did not see any laser photocoagulation scars in the patient’s eyes. None of Pon’s excluded surrebuttal evidence, which is the assumed error we are talking about, had anything to do with Patient Three. (4) Patient Four The jury heard from Dr. Friberg and another ophthalmologist about Patient Four. Before trial, Dr. Friberg reviewed images of this patient’s right eye, which Pon had diagnosed with WMD and for which he had billed Medicare under code 67220 as though he had performed laser photocoagulation treatment on two separate occasions. The images Dr. Friberg reviewed were taken before, on, and after the two “treatment” dates (the basis for Counts Seven and Eight). Dr. Friberg testified that based on his review he did not see any indication that the patient had WMD or a scar in her eyes showing that she had actually received laser photocoagulation treatment for that disease. Instead, the patient’s eye that Pon said suffered from WMD “look[ed] pristine” and there was “[n]o medical reason” to laser it. The other ophthalmologist testified about his examination of Patient Four several years after Pon had diagnosed her with WMD and supposedly treated her. He stated that she did not have WMD when Pon diagnosed her with it and billed Medicare for laser photocoagulation of it, that he did not see any scarring in her 46 Case: 17-11455 Date Filed: 06/29/2020 Page: 47 of 89 right eye that would indicate she had undergone that treatment, and that she still did not have WMD at the time he had examined her eyes. He added that the patient’s eyesight was “[v]ery good” for her age. None of Pon’s excluded surrebuttal evidence, which is the assumed error we are talking about, had anything to do with Patient Four. (5) Patient Five The jury heard from Dr. Friberg and another ophthalmologist about Patient Five. Before trial, Dr. Friberg reviewed images of the patient’s right eye, which Pon had diagnosed with WMD and for which he had billed Medicare under code 67220 as though he had performed laser photocoagulation treatment. The images Dr. Friberg reviewed were taken on and after the purported treatment date (the basis for Count Nine). Dr. Friberg testified that he was “[v]ery certain” the patient did not have WMD. He added that her blood vessels looked “[e]xcellent.” And although Pon had billed Medicare for laser photocoagulation of this patient’s right eye six times, Dr. Friberg saw no scars indicating she had ever received any laser photocoagulation in that eye. The second ophthalmologist testified about seven examinations he conducted on Patient Five after her optometrist referred her to him for a macular degeneration evaluation. All seven of the examinations were after Pon had diagnosed Patient Five with WMD and purportedly treated her. Based on his 47 Case: 17-11455 Date Filed: 06/29/2020 Page: 48 of 89 examinations, this ophthalmologist testified that there was no way that Patient Five had WMD when Pon diagnosed her with the disease. And he testified that he saw no scarring in her right eye that would indicate she had received laser photocoagulation treatment for WMD. None of Pon’s excluded surrebuttal evidence, which is the assumed error we are talking about, had anything to do with Patient Five. (6) Patient Six The jury heard about Patient Six from Dr. Friberg and two other ophthalmologists. Before trial, Dr. Friberg reviewed images of both of the patient’s eyes, each of which Pon had diagnosed with WMD and for which he had billed Medicare under code 67220 as though he had performed laser photocoagulation treatment. The images Dr. Friberg reviewed were taken before and after the “treatment” date for this patient’s right eye (the basis for Count Ten) and before, on, and after the “treatment” date for this patient’s left eye (the basis for Count Eleven). Dr. Friberg testified that based on his review the patient had not had WMD in either eye. And he testified that the patient’s eyes had no scars indicating that she had ever received any laser photocoagulation treatment. The second ophthalmologist testified about his treatment of Patient Six over a nine-year period that overlapped with the time Pon had treated her. This 48 Case: 17-11455 Date Filed: 06/29/2020 Page: 49 of 89 ophthalmologist reviewed for the jury his medical records from twelve examinations of the patient and testified that she had never showed signs of WMD. The third ophthalmologist testified about examinations he had conducted on Patient Six after Pon’s diagnosis and purported treatment. He stated that she did not have WMD and did not have any scarring from laser photocoagulation treatments. None of Pon’s excluded surrebuttal evidence, which is the assumed error we are talking about, had anything to do with Patient Six. (7) Patient Seven The jury heard from Dr. Friberg and two other ophthalmologists about Patient Seven. Before trial, Dr. Friberg reviewed images of both of the patient’s eyes, each of which Pon had diagnosed with WMD and for each of which he had billed Medicare under code 67220 as though he had performed laser photocoagulation treatment. The images Dr. Friberg reviewed were taken before, on, and after the “treatment” date for this patient’s left eye (the basis for Count Twelve) and on the “treatment” date for this patient’s right eye (the basis for Count Thirteen). Dr. Friberg testified that based on his review the patient never had WMD in either eye and did not have a laser scar indicating that he ever received laser photocoagulation treatment in either eye. The second ophthalmologist examined Patient Seven after Pon had billed Medicare numerous times for laser photocoagulation treatment of his right eye. 49 Case: 17-11455 Date Filed: 06/29/2020 Page: 50 of 89 This ophthalmologist testified that he found “no evidence whatsoever of any previous laser treatment.” He was “[a] hundred percent” certain that the patient did not have WMD or any scars indicating laser treatment. The third ophthalmologist examined Patient Seven after the patient was referred to him for a cataract evaluation. He went over for the jury the five examinations he had conducted on Patient Seven after Pon’s diagnosis and purported treatment. None of the evaluations showed any sign of WMD. None of Pon’s excluded surrebuttal evidence, which is the assumed error we are talking about, had anything to do with Patient Seven. (8) Patient Eight The jury heard about Patient Eight from Dr. Friberg, two other ophthalmologists, and an optometrist with 45 years of experience. Before trial, Dr. Friberg reviewed images of both of the patient’s eyes, each of which Pon had diagnosed with WMD and for each of which he had billed Medicare under code 67220 as though he had performed laser photocoagulation treatment. Those images were taken before, on, and after the “treatment” date for this patient’s left eye (the basis for Count Fourteen) and before and on the “treatment” date for this patient’s right eye (the basis for Count Fifteen). Dr. Friberg testified that based on his review the patient did not have WMD in his left eye and did not have a scar indicating he had ever received laser photocoagulation treatment in either eye. Dr. 50 Case: 17-11455 Date Filed: 06/29/2020 Page: 51 of 89 Friberg also testified that there was no WMD in the area of the patient’s right eye that Pon had marked for treatment. That area, Dr. Friberg added, was “quite pristine.” The second ophthalmologist testified about five examinations conducted on Patient Eight — one before and four after Pon’s diagnosis and purported laser photocoagulation treatment of the patient’s eyes. He testified that none of the examinations showed any indication of WMD or scarring from laser photocoagulation treatment. The third ophthalmologist performed cataract surgery on Patient Eight after Pon had diagnosed the patient with WMD and billed Medicare for laser photocoagulation treatment of both eyes. This ophthalmologist discussed his pre- and post-operation examinations of the patient’s eyes and testified that he never saw any evidence of WMD. The optometrist was another witness who testified about his examinations of Patient Eight’s eyes after Pon had diagnosed them with WMD and purportedly treated them. He testified that he saw no evidence of the scarring that would accompany laser photocoagulation treatment, and no evidence of WMD in the patient’s eyes. And he explained that because the patient did not have WMD when he examined him after Pon had diagnosed and supposedly treated his eyes, there was no way the patient had WMD when Pon diagnosed him with it. None of Pon’s 51 Case: 17-11455 Date Filed: 06/29/2020 Page: 52 of 89 excluded surrebuttal evidence, which is the assumed error we are talking about, had anything to do with Patient Eight. (9) Patient Nine The jury heard from Dr. Friberg and another ophthalmologist about Patient Nine. Before trial, Dr. Friberg reviewed images of both of this patient’s eyes, each of which Pon had diagnosed with WMD and for each of which he had billed Medicare under code 67220 as though he had performed laser photocoagulation treatment. The images Dr. Friberg reviewed were taken on and after the “treatment” date for this patient’s left eye (the basis for Count Sixteen) and before, on, and after the “treatment” date for this patient’s right eye (the basis for Count Seventeen). Dr. Friberg testified that based on his review the patient did not have WMD in either eye at the time Pon diagnosed her with that disease, and she did not have scarring from laser photocoagulation treatment in either eye. The other ophthalmologist examined Patient Nine after Pon had diagnosed her with WMD and purportedly treated her eyes. This ophthalmologist testified that he was “[a] hundred percent” certain that she had not had WMD or any scarring from laser photocoagulation treatment for that disease. None of Pon’s excluded surrebuttal evidence, which is the assumed error we are talking about, had anything to do with Patient Nine. 52 Case: 17-11455 Date Filed: 06/29/2020 Page: 53 of 89 (10) Patient Ten The jury heard about Patient Ten from Dr. Friberg, from another ophthalmologist, and also from an optometrist with 45 years of experience. Before trial, Dr. Friberg reviewed images of the patient’s left eye, which Pon had diagnosed with WMD and for which he had billed Medicare under code 67220 as though he had performed laser photocoagulation treatment. The images Dr. Friberg reviewed were taken on the “treatment” date for this patient’s left eye (the basis for Count Eighteen). He testified that based on his review the patient did not have WMD at the time Pon diagnosed her with that disease, nor did she have a scar indicating that she had ever received laser photocoagulation treatment for the disease. The other ophthalmologist testified about his treatment of Patient Ten after Pon had diagnosed her with WMD and purportedly treated her. Patient Ten had come to him for a second opinion about whether she had macular degeneration. He testified that he was “[a] hundred percent certain” that she did not have WMD or any scars from laser treatment for that disease. The optometrist told the jury about four examinations he had conducted on Patient Ten after Pon’s diagnosis and purported laser treatment. He testified that he never saw any indication that she had WMD and he had never seen any scarring from laser photocoagulation treatment for that disease. None of Pon’s excluded 53 Case: 17-11455 Date Filed: 06/29/2020 Page: 54 of 89 surrebuttal evidence, which is the assumed error we are talking about, had anything to do with Patient Ten. (11) Patient Eleven The jury heard about Patient Eleven from Dr. Friberg, from another ophthalmologist, and from the optometrist with 45 years of experience. Before trial, Dr. Friberg reviewed images of this patient’s right eye that Pon had diagnosed with WMD and for which he had billed Medicare under code 67220 as though he had performed laser photocoagulation treatment on two separate occasions. The images he reviewed were taken before, on, and after the first “treatment” date (the basis for Count Nineteen) and on and after the second “treatment” date (the basis for Count Twenty). Dr. Friberg testified that based on his review, he was “[c]ertain” that the patient did not have WMD when Pon diagnosed her with that disease, and that she did not have a scar indicating she had received laser photocoagulation treatment for the disease. The other ophthalmologist told the jury that Patient Eleven had been referred to him for a second opinion about whether Pon’s frequent lasering of her eye was necessary. He testified that he examined the patient twice and both examinations showed no evidence of WMD and no scar indicating she had received laser photocoagulation treatment for that disease. 54 Case: 17-11455 Date Filed: 06/29/2020 Page: 55 of 89 The optometrist testified about his treatment of Patient Eleven over a period of six years, some of which overlapped with Pon’s treatment of her. He testified that none of his examinations revealed any sign of WMD. None of Pon’s excluded surrebuttal evidence, which is the assumed error we are talking about, had anything to do with Patient Eleven. (12) Summary In regard to the eleven patients whose cases were the basis for all of the counts in the indictment, the jury heard from twelve different doctors, each of whom testified that Pon had diagnosed patients with WMD when those patients clearly had never had it. Those doctors also testified that none of those patients had the telltale scarring that necessarily results from the laser photocoagulation treatment that Pon billed Medicare for performing on them. Other than himself, Pon presented no ophthalmologist or optometrist to testify that any one of the eleven patients actually had WMD at the time he purportedly treated them for it or had the scarring that would necessarily have resulted from the laser photocoagulation treatment he billed Medicare for performing. Not only that, but he billed Medicare for laser photocoagulation treatment of the eleven patients, which necessarily causes a scar, even though Pon himself conceded that his treatment did not cause any scarring. And nobody (other than Pon) testified to support his claim that WMD could be laser-treated without scarring. None of 55 Case: 17-11455 Date Filed: 06/29/2020 Page: 56 of 89 Pon’s excluded surrebuttal evidence, which is the assumed error we are talking about, had anything to do with any of the eleven patients listed in the indictment or the bills he submitted to Medicare for the laser photocoagulation treatment that he never administered to those patients. b. The Hundreds of Other Patients The overwhelming proof of Pon’s guilt did not stop there. There was also strong evidence — uninfluenced and unaffected by the partial limitation on Pon’s surrebuttal evidence (the assumed error we are talking about) — that Pon incorrectly diagnosed and improperly “treated” not just the eleven patients listed in the indictment but also hundreds of other patients. As discussed, Dr. Friberg testified that almost none of the 500 patients whom Pon had diagnosed with WMD actually had the disease. He recounted how, during his close review of the files of approximately 500 of Pon’s patients, it was “rare” for him to see any indication that any patient had any form of macular degeneration — either dry or wet. Only a “very minimal minority” — one or two percent — of the 500 patients whose charts he reviewed had any sort of macular degeneration. Other eye doctors corroborated Dr. Friberg’s diagnoses for dozens of Pon’s patients not named in the indictment. Only Pon testified that his WMD diagnoses were correct, and that he believed the treatments he administered were helpful and medically necessary. But 56 Case: 17-11455 Date Filed: 06/29/2020 Page: 57 of 89 Pon had strong motivation to say that –– a substantial interest in the outcome of the trial. He acknowledged that most of his patients were Medicare beneficiaries and that the vast majority of the money he made was from Medicare. If convicted, he would lose his medical license and livelihood. And, of course, he faced a prison sentence –– ten years as it turned out. The jury was entitled to take Pon’s interest into account in evaluating his testimony, as it undoubtedly did. We have even held that because a jury is free to infer from a testifying defendant’s demeanor that he is not telling the truth, “a statement by a defendant, if disbelieved by the jury, may be considered as substantive evidence of the defendant’s guilt” when combined with other evidence. United States v. Brown, 53 F.3d 312, 314 (11th Cir. 1995); accord United States v. McCarrick, 294 F.3d 1286, 1293 (11th Cir. 2002) (“In Brown, we held that, in combination with other evidence, the jury’s disbelief of a defendant’s testimony may be used to help establish his guilt.”); see also United States v. Hough, 803 F.3d 1181, 1188 (11th Cir. 2015) (“Having seen and heard [the defendant’s] testimony, the jury was free to discredit her explanation, to infer that the opposite of what she said was true, and to consider that inference as substantive evidence of her guilt.”). This is especially true in regard to “highly subjective elements” such as “the defendant’s intent or knowledge.” Brown, 53 F.3d at 315. The jury observed Pon for more than three days on the witness stand and had ample time to 57 Case: 17-11455 Date Filed: 06/29/2020 Page: 58 of 89 evaluate his demeanor and credibility. See United States v. Devorso, 518 F.3d 1250, 1258 (11th Cir. 2008) (“Given the opportunity to evaluate [the defendant’s] demeanor and credibility, the jury was entitled not only to disbelieve his testimony but, in fact, to find that the opposite of his testimony was true.”). The government also offered overwhelming evidence — again, completely separate from the assumed error about limiting Pon’s surrebuttal evidence — that Pon knew the patients he diagnosed with WMD did not have that disease and knew that his micropulse laser technique did not treat that disease. Almost every doctor who testified — even Pon himself — said that a person with untreated WMD would suffer a substantial decline in the quality of his vision. And yet, the evidence at trial showed that Pon abruptly stopped all WMD treatments on at least four patients he had diagnosed with WMD and on whom he had been using his micropulse laser technique. But Pon continued to treat these patients for other eye conditions, so he would have known that the WMD he had diagnosed in them did not progress even though they were not receiving any treatment for it, which had to mean there was no WMD to begin with and Pon knew it. c. Different Treatment for Patients Who Actually Had WMD The evidence also showed that Pon used very different WMD treatment methods for different patients. Recall the testimony that at the time of Pon’s trial, injecting drugs directly into the eye had become the most widely used, accepted 58 Case: 17-11455 Date Filed: 06/29/2020 Page: 59 of 89 treatment for WMD. See supra pp. 4, 14. But evidence at trial showed that Pon used those drug injections as a WMD treatment for only some of the patients he diagnosed with WMD, at least of those he called as witnesses at trial. Significantly, the evidence showed that he used drug injections to treat WMD largely, if not only, for those patients whose WMD diagnoses were corroborated by another doctor. In other words, he used drug treatments for the rare patients of his who actually had WMD. For other patients, ones for whom there was no evidence of WMD but he billed as if there were, Pon didn’t use drug injections as a WMD treatment or he used it only rarely. For example, four of the thirteen patients who testified on Pon’s behalf had a WMD diagnosis that had been confirmed by another doctor. Pon gave all four of these patients drug injections. For two of these patients, Pon used the accepted drug injection treatment as well as his micropulse laser “treatment,” all in the same eye. For the other two of these four, he administered injections often and consistently, and he never discontinued the injections for either of them. The sum of it is that for the few patients whose WMD was confirmed by other doctors, Pon treated with the accepted drug injections, as well as his micropulse laser technique, and billed for both. At least seven of the patients who testified on Pon’s behalf had WMD diagnoses from Pon that Dr. Friberg rejected. For these patients, Pon either did not 59 Case: 17-11455 Date Filed: 06/29/2020 Page: 60 of 89 use injections to treat what only he had diagnosed as WMD, or he used injections only occasionally. And, as we discussed, for four of these patients, Pon stopped administering any WMD treatment — laser or injections — but continued to treat them for other eye diseases. That, of course, is evidence that Pon knew those patients did not actually have WMD, but he nonetheless diagnosed them with it and administered his micropulse laser “treatment,” while billing Medicare under code 67220 for laser photocoagulation treatment. d. The Ineffectiveness of Pon’s Micropulse Laser “Treatment” The government also offered testimony from other doctors — uninfluenced by Pon’s excluded surrebuttal evidence — that Pon’s micropulse laser technique could not treat WMD and could not coagulate a feeder vessel. One ophthalmologist testified that he was “certain far beyond a reasonable medical . . . certainty” that Pon’s micropulse laser technique would not close a feeder vessel. Another testified that he was not aware of any way that Pon’s micropulse laser technique could coagulate a feeder vessel, which was necessary to treat WMD, and was what Pon billed Medicare for doing. And yet another ophthalmologist testified that there was no way to achieve “subthreshold coagulation” of a feeder vessel, as Pon claimed, because those two “terms are mutually exclusive.” And so it went. 60 Case: 17-11455 Date Filed: 06/29/2020 Page: 61 of 89 Numerous doctors testified that WMD could not be treated with a laser without leaving a scar. Dr. Friberg was “very certain” that using a laser at a power high enough to “cook the feeder vessel,” which is how WMD is treated with a laser, would leave a mark in the eye. He testified that even if the scars did not show up in the eye immediately, “you could see them down the road.” His certainty about this came from his experience directing a clinical trial in which the doctors “used very minimal laser” and thought they “didn’t leave any marks.” But in “[a] month or two,” the patients’ eyes showed signs of scarring. Another ophthalmologist had participated in a clinical trial in which the doctors attempted to treat dry macular degeneration — which usually precedes WMD — using “very, very light laser treatment.” But the trial was stopped early because the patients being treated were developing WMD more quickly than their counterparts who were receiving no treatment at all. This ophthalmologist testified that laser photocoagulating leaking blood vessels, the treatment Pon was billing Medicare for, would “by definition, pretty much create a full-thickness burn” and leave a scar in the eye. He testified that laser treatment for WMD leaves scars that don’t ever completely heal, “and that’s the point actually,” because that is how WMD is treated with a laser. Other doctors agreed. Another ophthalmologist testified that if a feeder vessel had been closed with a laser, “you would see a scar.” And another testified 61 Case: 17-11455 Date Filed: 06/29/2020 Page: 62 of 89 that laser treatment for WMD that has “any kind of effect on the underlying blood vessels” would cause “some kind of scar.” And another testified that “the purpose of” laser treatment for WMD is to “cook[]” the tissue in the eye, which stops the blood vessels from leaking and necessarily results in a visible scar in the eye. And another testified that laser treatment for WMD results in “tiny focal laser scars” that are generally visible with a standard eye exam and that would always show up on a fluorescein angiogram. And yet another testified that “there’s no description of [its] being possible” to close a feeder vessel without scarring, and “there’s a lot of theoretical and scientific reasons why that would seem extremely unlikely to be possible.” Nonetheless, as we discussed, every single doctor who examined the eyes of a patient listed in the indictment testified that the patient’s eyes showed no indication of the scarring that necessarily accompanies laser photocoagulation — the WMD treatment that Pon billed Medicare for administering. See supra pp. 41– 56. e. Filling Out Charts in Advance The record also shows that Pon filled out portions of some patients’ charts with WMD diagnoses and planned diagnostic tests before he had even seen the patients. The government admitted patient notes for three of Pon’s patients who had not shown up for a scheduled appointment. Even though Pon didn’t examine the patients on the date listed in their charts, parts of their charts were filled out as 62 Case: 17-11455 Date Filed: 06/29/2020 Page: 63 of 89 though he had seen them, including diagnoses and treatment plans. For each of the three patients, the prefilled patient notes diagnosed the patient with WMD and indicated that fluorescein angiogram and ICG tests (the tests Pon used in making his WMD diagnoses) would be performed in both of the patient’s eyes. Pon testified that the patient notes were filled out ahead of time likely because his technicians were “trying to save time.” He stated that “if anything needed to be corrected” after he actually examined the patient, he “would have crossed it off” to “make sure everything was consistent with [his] examination.” And he identified seven patients whose prefilled notes he had modified to replace the diagnosis or treatment plan with one he thought was more appropriate. But the government had a response to that. Agent Jurs testified about three patients whose prefilled charts Pon had changed to indicate that fluorescein angiogram and ICG tests would be done only on one eye (instead of on both eyes, as the prefilled charts had originally indicated). For each of those three patients, the government introduced an exhibit listing the procedures that Pon had billed Medicare for. Those exhibits showed that Pon had billed Medicare for fluorescein angiograms and ICGs on both eyes for each of the patients. In other words, although Pon had modified the three patients’ prefilled charts to show that those tests were done on only one eye, he still billed Medicare as though he had done the tests on both eyes. 63 Case: 17-11455 Date Filed: 06/29/2020 Page: 64 of 89 f. The Sound of Silence Finally, the record shows that Pon was professionally silent about his purported treatment. He acknowledged that he didn’t know of any other doctor anywhere who used subthreshold micropulse laser to treat WMD. Yet even though he claimed to have discovered a “miraculous treatment” for WMD, he did absolutely nothing to present, publish, or even talk with other doctors about what he thought of as a cure for the leading cause of irreversible blindness in older people. His silence spoke volumes. g. Summary All of this great volume of evidence we have just recounted was presented before and was completely unrelated to and uninfluenced by the exclusion of any of Pon’s proposed surrebuttal evidence. And it was that great volume of evidence that the government discussed in its closing, not anything about Pon billing Medicare for testing he did on J.L.’s blind left eye. In view of the totality of the evidence presented, what the jury heard about billing for testing on J.L.’s left eye was miniscule. J.L. was one of 34 witnesses who testified at trial, one of fifteen who testified for Pon. J.L.’s testimony about the procedures done on his blind left eye took up only fifteen pages of transcript. And Agent Jurs’ rebuttal testimony concerning that subject, which is what Pon wanted to present surrebuttal testimony 64 Case: 17-11455 Date Filed: 06/29/2020 Page: 65 of 89 about, took up only eleven pages. That’s eleven out of more than 2,000 pages, or about one half of one percent, of the testimony that was presented during the trial. And it bears repeating that J.L. was not one of the patients listed in the crimes charged in the indictment. As we noted, the district court instructed the jury that Pon was “on trial only for the specific crimes charged in the indictment,” and that it had to determine whether Pon was “guilty or not guilty of those specific crimes.” “[T]he Supreme Court has repeatedly held that we must presume that juries follow their instructions,” and this Court has “obediently followed” that direction. Roy, 855 F.3d at 1186–87 (collecting cases). Following that direction, we presume that the jury’s guilty verdict was based on its determination that Pon was guilty as charged of fraudulently billing Medicare for laser photocoagulation for the eleven patients listed in the indictment, not for fraudulently billing Medicare for procedures on J.L.’s blind left eye. For all these reasons, we have no doubt, much less a reasonable doubt, that if the district court had not partially limited Pon’s surrebuttal evidence about J.L., the jury would still have found Pon guilty as charged. The ruling we are assuming was an error did not contribute to the jury’s guilty verdict. See Chapman, 386 U.S. at 22. Any error was harmless beyond a reasonable doubt. 65 Case: 17-11455 Date Filed: 06/29/2020 Page: 66 of 89 5. Our Application of the Harmless Error Rule is Faithful to Precedent Our conclusion that the partial limitation on Pon’s surrebuttal testimony was harmless beyond a reasonable doubt is in keeping with the Supreme Court’s understanding, expressed in Neder, that “[a] reviewing court making th[e] harmless-error inquiry does not, as Justice Traynor put it, become in effect a second jury to determine whether the defendant is guilty.” 527 U.S. at 19 (quotation marks omitted). And it is consistent with the Court’s description of the harmless-error determination as a task done in “typical appellate-court fashion.” Id. Of course harmless error inquiries are typical: “We are, after all, talking about ‘the harmless error rule,’ not ‘the harmless error exception.’ Because errorless trials are not expected, much less required, harmless error analysis is the rule, not the exception.” Roy, 855 F.3d at 1143. And for good reason. “The harmless error rule serves vital interests, chief of which is conserving scarce judicial resources by avoiding pointless retrials.” Id. at 1142. And, as we have mentioned, the rule “is also essential to avoid a ‘sporting theory of justice’ and a regime of gotcha review.” Id. (quoting Agurs, 427 U.S. at 108). The Supreme Court has recognized and relied on these important considerations in holding — over and over again — that the harmless error rule applies in a wide variety of circumstances. The Court has explained that the harmless error rule “promotes public respect for the criminal process by focusing 66 Case: 17-11455 Date Filed: 06/29/2020 Page: 67 of 89 on the underlying fairness of the trial.” Neder, 527 U.S. at 18 (quotation marks omitted). It has directed that when “the record developed at trial establishes guilt beyond a reasonable doubt, the interest in fairness has been satisfied and the judgment should be affirmed.” Rose v. Clark, 478 U.S. 570, 579 (1986). And it has noted that “[r]eversal for error, regardless of its effect on the judgment, encourages litigants to abuse the judicial process and bestirs the public to ridicule it.” Delaware v. Van Arsdall, 475 U.S. 673, 681 (1986) (quoting R. Traynor, The Riddle of Harmless Error 50 (1970)). The Supreme Court has instructed us that “if the defendant had counsel and was tried by an impartial adjudicator, there is a strong presumption that any other errors that may have occurred are subject to harmless-error analysis.” Rose, 478 U.S. at 579 (emphasis added). The Court has “consistently made clear that it is the duty of a reviewing court to consider the trial record as a whole and to ignore errors that are harmless, including most constitutional violations.” 10 United States v. Hasting, 461 U.S. 499, 509 (1983) (emphasis added); accord Neder, 527 U.S. at 10 As the Court’s use of the word “duty” indicates, when the Supreme Court applies harmless error analysis to a certain kind of error, it does not merely suggest that we do the same — it dictates that we are bound to do so. See, e.g., Mathis v. United States, 136 S. Ct. 2243, 2254 (2016) (“[A] good rule of thumb for reading [Supreme Court] decisions is that what they say and what they mean are one and the same . . . .”); United States v. Johnson, 921 F.3d 991, 1001 (11th Cir. 2019) (“[W]e must apply Supreme Court precedent neither narrowly nor liberally — only faithfully.”); Prison Legal News v. Sec’y, Fla. Dep’t of Corr., 890 F.3d 954, 966 (11th Cir. 2018) (“The only Court that can properly cut back on Supreme Court decisions is the Supreme Court itself.”). 67 Case: 17-11455 Date Filed: 06/29/2020 Page: 68 of 89 8 (noting that “most constitutional errors can be harmless” and only “a very limited class of cases” evade harmless error review) (quotation marks omitted); Arizona v. Fulminante, 499 U.S. 279, 306 (1991) (noting that “the Court has applied harmless-error analysis to a wide range of errors and has recognized that most constitutional errors can be harmless”); Van Arsdall, 475 U.S. at 681 (“[W]e have repeatedly reaffirmed the principle that an otherwise valid conviction should not be set aside if the reviewing court may confidently say, on the whole record, that the [error was harmless].”). The Fulminante decision is the paramount example of the breadth of the harmless error rule because it held that even where the error was the admission of an unconstitutionally coerced confession, courts must still assess the totality of the evidence and determine if it was harmless beyond a reasonable doubt. See 499 U.S. at 310 (“When reviewing the erroneous admission of an involuntary confession, the appellate court, as it does with the admission of other forms of improperly admitted evidence, simply reviews the remainder of the evidence against the defendant to determine whether the admission of the confession was harmless beyond a reasonable doubt.”). The Supreme Court has also strictly limited the power of federal courts to circumvent a harmless-error inquiry and has rebuked courts who shirk their harmless-error review duties. See, e.g., Bank of Nova Scotia v. United States, 487 U.S. 250, 254 (1988) (“[A] federal court may not invoke supervisory power to 68 Case: 17-11455 Date Filed: 06/29/2020 Page: 69 of 89 circumvent the harmless-error inquiry prescribed by Federal Rule of Criminal Procedure 52(a).”); Hasting, 461 U.S. at 507 (stating that the harmless-error doctrine “cannot be so lightly and casually ignored in order to chastise what the court view[s] as prosecutorial overreaching”). We are not saying, of course, that courts shouldn’t be careful with the harmless error rule. Courts should be careful in the application of all rules. Carelessness is not desirable in any field. But it is not careless to rely on overwhelming evidence of guilt to find an error harmless. The Supreme Court itself has done it. See, e.g., Neder, 527 U.S. at 17 (“In this situation, where a reviewing court concludes beyond a reasonable doubt that the omitted element was uncontested and supported by overwhelming evidence, such that the jury verdict would have been the same absent the error, the erroneous instruction is properly found to be harmless.”) (emphasis added); cf. Yates, 500 U.S. at 405 (describing the Chapman harmless-error inquiry as requiring a court to “ask[] whether the force of the evidence presumably considered by the jury in accordance with the instructions is so overwhelming as to leave it beyond a reasonable doubt that the verdict resting on that evidence would have been the same”) (emphasis added). And in doing so, the Supreme Court has found errors harmless based on overwhelming evidence without displaying special wariness. 69 Case: 17-11455 Date Filed: 06/29/2020 Page: 70 of 89 Our own precedent is thick with decisions finding errors, even serious errors, harmless, oftentimes because of overwhelming evidence of guilt. See United States v. Esquenazi, 752 F.3d 912, 931 (11th Cir. 2014), superseded by statute in non-relevant part, 18 U.S.C. §§ 1956, 1957 (finding an error in the trial court’s jury instructions harmless “[b]ecause overwhelming evidence support[ed] the jury’s finding” of guilt); United States v. Jones, 601 F.3d 1247, 1264 (11th Cir. 2010) (finding the constitutional error the district court allegedly committed harmless because even without the error “the government’s case against [the defendant] was strong”); United States v. Phaknikone, 605 F.3d 1099, 1109 (11th Cir. 2010) (affirming, in spite of an error, the conviction “in the light of the overwhelming evidence of [the defendant’s] guilt”). We do that here. IV. THE SENTENCE ISSUES Pon challenges the procedural reasonableness of his sentence, arguing that the district court improperly calculated his guidelines range by erroneously applying an 18-level enhancement to his base offense level. The primary issue at sentencing was the amount of loss that resulted from Pon’s health care fraud scheme. The government said the loss amount was $11.5 million, which would result in a 20-level enhancement. See U.S.S.G. § 2B1.1(b)(1)(K) (Nov. 2010) (providing for a 20-level enhancement when the total loss amount is more than $7 70 Case: 17-11455 Date Filed: 06/29/2020 Page: 71 of 89 million but not more than $20 million). The district court, however, rejected the government’s number and instead estimated the loss amount at $6.97 million, which resulted in an 18-level enhancement. See id. § 2B1.1(b)(1)(J) (providing for an 18-level enhancement when the total loss amount is more than $2.5 million but not more than $7 million). The guidelines range those calculations produced was 121 to 151 months. Pon contends that the district court did not base the loss amount on reliable and specific evidence. We review the district court’s loss determination only for clear error. United States v. Cobb, 842 F.3d 1213, 1218 (11th Cir. 2016). For that determination to be clearly erroneous, “we must have a definite and firm conviction that a mistake has been made.” United States v. Ford, 784 F.3d 1386, 1396 (11th Cir. 2015). Under the guidelines, “loss is the greater of actual loss or intended loss.” U.S.S.G. § 2B1.1, cmt. n.3(A). Although the district court “may not speculate about the existence of facts and must base its estimate on reliable and specific evidence, [it] is required only to make a reasonable estimate of the loss.” Ford, 784 F.3d at 1396. That is “because often the amount of loss caused by fraud is difficult to determine accurately.” United States v. Medina, 485 F.3d 1291, 1304 (11th Cir. 2007) (quotation marks omitted). And the district court “is in a unique position to assess the evidence and estimate the loss based upon that evidence,” so 71 Case: 17-11455 Date Filed: 06/29/2020 Page: 72 of 89 its “loss determination is entitled to appropriate deference.” United States v. Campbell, 765 F.3d 1291, 1301 (11th Cir. 2014) (quoting U.S.S.G. § 2B1.1 cmt. n.3(C)). Contrary to what Pon says, the district court did base its loss estimate on reliable and specific evidence. It considered a spreadsheet that Agent Jurs prepared for the sentence hearing, which shows the total amount of money that Medicare paid Pon for the thousands of claims he billed on behalf of the patients Dr. Friberg determined did not have WMD. The spreadsheet lists each patient’s name, the notes that Dr. Friberg took when he reviewed the photos and videos of the patient’s eyes, and the number of Medicare claims that Pon had submitted for that patient. The district court also heard extensive testimony from Agent Jurs, who at the time had nearly seventeen years of experience working with Medicare data. He testified that the spreadsheet’s total loss amount was accurate because it was calculated on a patient-by-patient basis. He also explained that for each of the patients Dr. Friberg determined did not have WMD, he “went into the Medicare records, pulled the claims that were specific to that [patient], went into the diagnosis code section of the claims history, pulled the [WMD] diagnosis code that [HHS] knew to be fraudulent and incorrect, and then totaled up only those claims, the numbers amount billed and the amount paid for only those claims.” 72 Case: 17-11455 Date Filed: 06/29/2020 Page: 73 of 89 Agent Jurs also testified that the amount billed, the amount paid, and the number of actual claims came directly from Medicare’s database. That evidence is reliable and it is specific. And based on it, the district court made a reasonable estimate of the loss. See Cobb, 842 F.3d at 1218–19. There was no error in calculating the loss amount. 11 There is one other sentence issue. Pon has not raised it, but the government has. The district court imposed concurrent 121-month terms of imprisonment on each of Pon’s twenty counts of conviction. The government concedes that this was error because the statutory maximum penalty for each count is only 120 months. See 18 U.S.C. § 1347(a). We agree, and we commend the government for bringing the error to our attention. Section 5G1.2(d) of the guidelines provides that where, as here, there are multiple counts of conviction, [i]f the sentence imposed on the count carrying the highest statutory maximum is less than the total punishment, then the sentence imposed on one or more of the other counts shall run consecutively, but only to the extent necessary to produce a combined sentence equal to the total punishment. In all other respects, sentences on all counts shall run concurrently, except to the extent otherwise required by law. 11 Relying on out-of-circuit decisions, Pon also argues that due process required the district court to use the beyond a reasonable doubt standard instead of the preponderance of the evidence standard in determining the loss amount. But “it is the settled law of this circuit that at sentencing, a federal defendant’s due process rights are . . . satisfied by the preponderance of the evidence standard.” United States v. Patti, 337 F.3d 1317, 1323 n.9 (11th Cir. 2003) (quoting United States v. Jackson, 57 F.3d 1012, 1019 (11th Cir. 1995)). 73 Case: 17-11455 Date Filed: 06/29/2020 Page: 74 of 89 U.S.S.G. § 5G1.2(d). We have interpreted § 5G1.2(d) to require “multiple sentences to be served consecutively if the sentence specified by the guidelines is longer than the sentence authorized for any individual count of conviction.” United States v. Pressley, 345 F.3d 1205, 1213 (11th Cir. 2003); accord United States v. Sarras, 575 F.3d 1191, 1209 n.22 (11th Cir. 2009) (concluding that § 5G1.2(d) remains binding even “[]though the guidelines are now advisory”). Those multiple sentences must, of course, run consecutively “to the extent necessary to reach the defendant’s guidelines range.” United States v. Johnson, 451 F.3d 1239, 1243 (11th Cir. 2006). The district court erred in imposing concurrent 121-month terms on each of Pon’s twenty counts. While the guidelines specify a sentence of 121 to 151 months in prison, the statutory maximum for the count with the highest maximum (of any of the twenty counts) is 120 months in prison. Because the statutory maximum is one month less than the 121-month bottom of the guidelines range, § 5G1.2(d) called for at least some of the sentences to run consecutively “to the extent necessary to reach [Pon’s] guidelines range” of 121 to 151 months in prison. Id. The sentence structure must be modified to reach that range. We vacate Pon’s sentences on each count and remand the case to the district court for the limited purpose of letting it modify Pon’s sentence structure to bring 74 Case: 17-11455 Date Filed: 06/29/2020 Page: 75 of 89 it in line with § 5G1.2(d). We leave to the court’s discretion how it will do so.12 Because we do not set aside Pon’s “entire sentencing package” or the time he will remain in prison, the modification does not require a resentencing hearing at which Pon must be present. See United States v. Tamayo, 80 F.3d 1514, 1518, 1519 & n.7 (11th Cir. 1996). V. CONCLUSION We AFFIRM Pon’s convictions but VACATE his sentences on Counts One through Twenty and REMAND the case for resentencing. 12 The government has argued on appeal that one of Pon’s twenty 120-month sentences should be made to run consecutive to the other nineteen by one month. Pon has not addressed that issue. But in any event it’s an issue for the district court to decide. 75 Case: 17-11455 Date Filed: 06/29/2020 Page: 76 of 89 MARTIN, Circuit Judge, concurring in part and dissenting in part: Like the majority, I see no abuse of discretion in the District Court’s exclusion of David Pon’s expert witness at trial. Nor do I see clear error in the District Court’s loss amount calculation. I agree, too, that this Court’s precedent forecloses Mr. Pon’s due process challenge to the loss amount calculation. But I part ways with the conclusion reached in the majority opinion that any error in denying Mr. Pon a surrebuttal was harmless. I believe the denial of a surrebuttal violated Mr. Pon’s constitutional right to present a complete defense. And I do not believe this error was harmless. Rather than affirm his conviction, I would give Mr. Pon a new jury trial. I therefore dissent from the majority opinion’s ruling on the harmlessness of any surrebuttal error. I. “[T]he Constitution guarantees criminal defendants a meaningful opportunity to present a complete defense.” Holmes v. South Carolina, 547 U.S. 319, 324, 126 S. Ct. 1727, 1731 (2006) (quotation marks omitted). Whether this right springs from the fundamentals of due process or from the Sixth Amendment, see Crane v. Kentucky, 476 U.S. 683, 690, 106 S. Ct. 2142, 2146 (1986), it encompasses “the right to present the defendant’s version of the facts as well as the prosecution’s to the jury so it may decide where the truth lies.” Washington v. 76 Case: 17-11455 Date Filed: 06/29/2020 Page: 77 of 89 Texas, 388 U.S. 14, 19, 87 S. Ct. 1920, 1923 (1967). This right is not boundless. If it were, trials might be never-ending. District courts wield broad discretion to control the mode and order of presenting evidence, both to preserve the trial’s truth-seeking function and to avoid wasting time. See Fed. R. Evid. 611; United States v. Jeri, 869 F.3d 1247, 1262 (11th Cir. 2017). But this “discretion does not . . . extend to the exclusion of crucial relevant evidence necessary to establish a valid defense.” United States v. Kelly, 888 F.2d 732, 743 (11th Cir. 1989) (quotation marks omitted). When the government presents rebuttal evidence in a criminal case, the right to present a defense sometimes requires district courts to allow criminal defendants a surrebuttal. See, e.g., United States v. Murray, 736 F.3d 652, 656–59 (2d Cir. 2013); United States v. Barnette, 211 F.3d 803, 821–24 (4th Cir. 2000); United States v. Moody, 903 F.2d 321, 330–31 (5th Cir. 1990). “The purpose of rebuttal evidence is to explain, repel, counteract, or disprove the evidence of the adverse party.” United States v. Frazier, 387 F.3d 1244, 1269 (11th Cir. 2004) (en banc) (quotation marks and alteration omitted). The same is true of a surrebuttal, except that its function is to explain, repel, counteract, or disprove evidence presented in a rebuttal. Though narrower in scope than a defense to the government’s case-in- chief, surrebuttal nonetheless serves an important purpose. It allows criminal defendants to present their version of any new issues that arise in the course of the 77 Case: 17-11455 Date Filed: 06/29/2020 Page: 78 of 89 government’s rebuttal. Juries sometimes cannot “decide where the truth lies” if they hear only the government’s side of a rebuttal issue. Washington, 388 U.S. at 19, 87 S. Ct. at 1923. I say Mr. Pon was entitled to a surrebuttal. The rule in this Circuit, as in others, is that a “surrebuttal is merited where (1) the government’s rebuttal testimony raises a new issue, which broadens the scope of the government’s case, and (2) the defense’s proffered surrebuttal testimony is not tangential, but capable of discrediting the essence of the government’s rebuttal testimony.” Moody, 903 F.2d at 331; see also United States v. Durnin, 632 F.2d 1297, 1301 n.8 (5th Cir. Unit A 1980). 1 Under this standard, this Court’s sister circuits have ruled that the Constitution requires a surrebuttal where the government’s rebuttal raised new, uncharged allegations of fraud in a fraud case, Moody, 903 F.2d at 330–31, advanced a new mental health diagnosis as evidence of future dangerousness to support a death sentence, Barnette, 211 F.3d at 821–24, and introduced new evidence placing a defendant more frequently in the vicinity of the crime, Murray, 736 F.3d at 658–59. Rightly so. Without a surrebuttal, the government could use rebuttal in those cases both to meet the defense case and to make a new, unchallenged case of, respectively, fraud, future dangerousness, and frequent 1 In Bonner v. City of Prichard, 661 F.2d 1206 (11th Cir. 1981) (en banc), this Court adopted as binding precedent all decisions of the former Fifth Circuit handed down before October 1, 1981. Id. at 1209. 78 Case: 17-11455 Date Filed: 06/29/2020 Page: 79 of 89 presence at the scene of the crime. That result would eviscerate the right to present a defense. Both conditions for a surrebuttal were plainly met here. The government’s rebuttal certainly introduced a new issue about the treatment of J.L.’s left eye for conditions wholly unrelated to wet macular degeneration.2 Mr. Pon’s proffered surrebuttal testimony about that treatment would have discredited the essence of the government’s rebuttal case. Mr. Pon’s right to present his defense required the District Court to allow him to testify to his reasons for treating J.L.’s left eye. I discern no real dispute about whether the government’s rebuttal evidence introduced a new issue. The rebuttal evidence concerned treatment of J.L., a patient who testified during the defense case. The substance of J.L.’s testimony was that Mr. Pon treated J.L. for detached retinas even knowing J.L. did not have insurance, that Pon restored vision in J.L.’s right eye but not his left, and that Pon later diagnosed J.L. with wet macular degeneration and treated him with the laser method. On cross examination, the government asked J.L. whether Mr. Pon had performed any tests on his left eye. J.L. said he had not. J.L. reiterated on redirect examination that Mr. Pon hadn’t treated his left eye. After the defense rested, the 2 Mr. Pon contends the District Court abused its discretion in allowing a rebuttal. Not so. The District Court had discretion to allow the government to rebut the impression J.L.’s testimony created of Mr. Pon’s altruism and to clarify J.L.’s treatment history. 79 Case: 17-11455 Date Filed: 06/29/2020 Page: 80 of 89 District Court allowed the government to put on rebuttal, 3 including a spreadsheet made by the government’s investigator showing Mr. Pon billed Medicare for treating J.L.’s left eye on more than fifty occasions between 2004 and 2015. All told, the services added up to $19,350 worth of Medicare charges over eleven years. Up to the point the government put the spreadsheet into evidence, nothing would have indicated to Mr. Pon that he had to defend his reasons for treating J.L.’s left eye. For that matter, he would have had no reason to prove that those treatments even happened. The government’s case-in-chief revolved around Mr. Pon’s false diagnosis and treatment of wet macular degeneration, see Maj. Op. at 13–16, while the defense case depended on showing Mr. Pon lacked the necessary intent to defraud Medicare. With the spreadsheet, the government injected an entirely new, unrelated, and uncharged fraud. The clear implication of the rebuttal evidence was that Mr. Pon fraudulently billed Medicare not only for diagnosing and treating nonexistent wet macular degeneration, but for other procedures as well. This broadened the scope of the government’s case. See Moody, 903 F.2d at 3 The District Court’s decision to allow rebuttal seemed to rest in part on the government telling the District Court that it was Mr. Pon who first put the matter in issue by asking about J.L.’s left eye on redirect. The District Court sided with the government, believing the defense first asked about treatments on J.L.’s left eye on redirect. The government’s representation and the District Court’s belief in this regard is contradicted by the record. It was the government who first introduced the issue of J.L.’s left eye by inquiring about it on cross-examination. 80 Case: 17-11455 Date Filed: 06/29/2020 Page: 81 of 89 331 (holding the government broadened its fraud case when it brought up new, uncharged frauds in rebuttal). Nor is there any real disputing that Mr. Pon’s proffered surrebuttal testimony would have gone right to the heart of the government’s rebuttal. In his proffer, he explained that he did in fact treat J.L.’s left eye and that he had a medical reason for doing so—namely, preventing blindness in J.L.’s right eye. This evidence, if believed, would squarely rebut the government’s contentions that Mr. Pon never treated J.L.’s left eye but billed for it anyway. The limitation of surrebuttal on this record violated Mr. Pon’s right to present a defense. The government stresses that the District Court limited rather than outright denied Mr. Pon’s surrebuttal. And the District Court did allow Mr. Pon to testify in surrebuttal that three of the more than fifty entries were the result of a clerical error. However, limiting Mr. Pon to this explanation may have exacerbated the problem. By allowing Mr. Pon to explain only three of the over fifty treatments, the jury may have been left with the impression that Pon had no explanations for the remaining treatments. Indeed, the government’s cross-examination of Mr. Pon played up the fact that Mr. Pon only had an explanation for “just three entries out of two pages of entries.” I question the propriety of this argument by the government when it knew Mr. Pon had an explanation for the other entries the court forbade him from testifying about. Mr. Pon’s full explanation (treating the 81 Case: 17-11455 Date Filed: 06/29/2020 Page: 82 of 89 left eye helped prevent issues in the right eye) would have rebutted a key contention of the government’s rebuttal case. The exclusion of this testimony violated the right to present a defense just as surely as a complete denial of surrebuttal would have. The limited surrebuttal allowed the government to exploit the District Court’s ruling, creating the impression that Mr. Pon had no explanation at all for more than forty treatments. In fact he did. The short of it is that the District Court allowed the government to inject a whole new allegation of fraud into the trial, then hobbled Mr. Pon’s ability to respond to it. Mr. Pon had no chance to defend himself against allegations that he billed Medicare for treatment he never provided on an eye that couldn’t see. Rather than decline to decide the issue, see Maj. Op. at 37–38, I would hold the District Court violated Mr. Pon’s constitutional right to present a defense. II. Not all errors require reversal. This one does. Under the harmless error doctrine, this Court will not reverse a district court’s constitutional trial error if it is “clear beyond a reasonable doubt that a rational jury would have found the defendant guilty” regardless of the error. Neder v. United States, 527 U.S. 1, 18, 119 S. Ct. 1827, 1838 (1999); see Chapman v. 82 Case: 17-11455 Date Filed: 06/29/2020 Page: 83 of 89 California, 386 U.S. 18, 24, 87 S. Ct. 824, 828 (1967). 4 Following the Supreme Court’s lead, this Court has identified cases in which “the evidence of the defendant’s guilt is ‘so overwhelming’” as “[o]ne circumstance in which courts find constitutional errors harmless beyond a reasonable doubt.” United States v. Willner, 795 F.3d 1297, 1322 (11th Cir. 2015) (quoting Harrington v. California, 395 U.S. 250, 254, 89 S. Ct. 1726, 1728 (1969)). But the Supreme Court has equally cautioned us to be careful with the harmless error doctrine. Take, for example, Neder. It raised the question of whether the harmless error doctrine applies in cases where a district court omits an element of the offense from the jury instructions. Neder, 527 U.S. at 8, 119 S. Ct. at 1833. The Court held the error “is subject to harmless-error analysis.” Id. at 15, 119 S. Ct. at 1837. But it warned that courts “should not find [an] error harmless” where the defendant contested the omitted element and “raised evidence sufficient to support a contrary finding.” Id. at 19, 119 S. Ct. at 1838. Those circumstances would, the Court said, preclude the reviewing court from “conclud[ing] beyond a 4 As the majority recognizes, the parties dispute which harmless error standard applies. See Maj. Op. at 37–38. The government says Mr. Pon did not preserve his constitutional error because he did not object on constitutional grounds below. I think the issue was sufficiently preserved. The District Court specifically raised Mr. Pon’s Sixth Amendment right to offer an explanation when considering whether to grant a surrebuttal. The issue seems to me to “fairly appear[] in the record as having been raised or decided” and thus is before the Court. 19 James Wm. Moore et al., Moore’s Federal Practice § 205.05(1) (3d ed. 2019). 83 Case: 17-11455 Date Filed: 06/29/2020 Page: 84 of 89 reasonable doubt that the jury verdict would have been the same absent the error.” Id. It is important to remember that harmless error review is no substitute for a jury trial. The Sixth Amendment demands no less. U.S. Const. amend. VI (“In all criminal prosecutions, the accused shall enjoy the right to a speedy and public trial, by an impartial jury.”). A reviewing court treads on the right to a jury trial when it uses harmless error review to “‘become in effect a second jury to determine whether the defendant is guilty.’” Neder, 527 U.S. at 19, 119 S. Ct. at 1839 (quoting Roger J. Traynor, The Riddle of Harmless Error 21 (Ohio State Univ. Press 1970)). And there are practical consequences, too. Whenever we invoke harmless error, “the deterrent force of a reversal remains unfelt by those who caused the error.” Harry T. Edwards, To Err Is Human, But Not Always Harmless: When Should Legal Error Be Tolerated?, 70 N.Y.U. L. Rev. 1167, 1170 (1995). We should be particularly wary of invoking “overwhelming evidence” to hold an error harmless. “[T]he Constitution does not trust judges to make determinations of criminal guilt.” Neder, 527 U.S. at 32, 119 S. Ct. at 1844 (Scalia, J., concurring in part and dissenting in part) (emphasis omitted). The right to a jury trial forbids us from doing so. This is why a federal judge may direct a judgment of acquittal but never a judgment of guilt. See Rose v. Clark, 478 U.S. 570, 578, 106 S. Ct. 3101, 3106 (1986); Fed. R. Crim. P. 29. We owe it to 84 Case: 17-11455 Date Filed: 06/29/2020 Page: 85 of 89 defendants who come before us to ask ourselves always whether a rational jury could acquit, and never whether we ourselves think the defendant guilty. Taking up this task, I believe a rational jury might have voted to acquit Dr. Pon. The majority opinion concludes the government’s evidence against Mr. Pon proved beyond a reasonable doubt that “Pon had diagnosed patients with [wet macular degeneration] when those patients clearly had never had it” and that Pon “billed Medicare for laser photocoagulation treatment of the[se] . . . patients.” Maj. Op. at 55. But the government had to prove more than misdiagnosis and unnecessary treatment to win a conviction. To win, it had to show Mr. Pon intended to defraud Medicare by submitting claims he knew “were, in fact, false.” United States v. Crabtree, 878 F.3d 1274, 1285 (11th Cir. 2018) (quotation marks omitted); see also Pattern Crim. Jury Instr. 11th Cir. O53 (2019). The focus on whether there was any doubt that the patients named in the indictment had wet macular degeneration or needed laser treatment elides this point. Of course, I do not contend the evidence the government presented about Mr. Pon’s diagnoses and treatments lacked probative value. The evidence of misdiagnosis and unnecessary treatment supplied circumstantial evidence of Mr. Pon’s intent, and enough of it to support a conviction. See United States v. Clay, 832 F.3d 1259, 1309 (11th Cir. 2016); United States v. Bradley, 644 F.3d 1213, 85 Case: 17-11455 Date Filed: 06/29/2020 Page: 86 of 89 1239 (11th Cir. 2011). But Mr. Pon also met the government’s case with enough evidence to support an acquittal, had the jury so decided. Mr. Pon spent more than three days on the stand. At the outset, he testified he “absolutely [did] not” intend to defraud Medicare. He spent the next several days explaining in granular detail how he diagnosed and treated patients. He said he got the idea for his diagnostic method from a talk given by Dr. Robert Murphy, a leading ophthalmologist. Dr. Murphy proposed using an emerging technology to visualize “occult neovascularization”—essentially, new, difficult-to-see blood vessels forming in the eye. Mr. Pon said he thought the technique was “the greatest thing since sliced bread, because this technology was going to allow direct visualization of those tiny, tiny little blood vessels.” Mr. Pon said he believed based on Dr. Murphy’s presentation that he could detect wet macular degeneration in its earliest stages using a cutting-edge technique. As for treatment, Mr. Pon said he got some ideas about that from Dr. Murphy’s presentation, too. Dr. Murphy reported using a laser to create a “thermal gradient”—a heat differential—in the eye to close the feeder vessels that caused wet macular degeneration. This technique did not cause scarring, as do traditional laser techniques for treating wet macular degeneration. Dr. Pon also thought this was “the greatest thing since sliced bread.” 86 Case: 17-11455 Date Filed: 06/29/2020 Page: 87 of 89 Mr. Pon bought the equipment he believed he needed to put Dr. Murphy’s diagnostic and treatment techniques into practice. He later upgraded the equipment with a purchase he explained he would not have made if his intent had been to defraud. Mr. Pon began treating patients with it and said he found lower power settings provided better results than Dr. Murphy’s thermal-gradient method. And the jury didn’t have to take Mr. Pon’s word for it. Mr. Pon called twelve of his patients and the spouse of another to testify that the lasers improved their (or her spouse’s) vision. One witness, who has advanced public health degrees and background as a nurse, testified she learned from Mr. Pon to identify feeder vessels on a diagnostic monitor during Pon’s treatment of her husband, who Pon diagnosed with wet macular degeneration. Another, a laser nurse, likewise testified she could see the problem areas Mr. Pon identified and that her vision improved with laser treatments. The government agreed some of these thirteen patients actually had wet macular degeneration. It disagreed as to others. But for all the patients who testified for Mr. Pon, their testimony was uniform that Mr. Pon’s laser treatment helped. With this evidence before it, a rational jury could acquit Mr. Pon. This hypothetical acquitting jury could do so without any belief that: (1) Mr. Pon indeed diagnosed his patients with wet macular degeneration no one else could see or (2) treated it using miraculous new technologies in ways no other doctor could. Yet 87 Case: 17-11455 Date Filed: 06/29/2020 Page: 88 of 89 these are the two issues the majority opinion focuses on. To acquit, a jury would have to conclude only that Mr. Pon believed in good faith he could do these things. His own testimony and that of his patients, some of whom even the government admits had wet macular degeneration, would suffice for a rational jury to believe Mr. Pon acted in good faith. If a patient who couldn’t read a menu before says he can after a doctor gave him a laser treatment, reasonable people could agree the doctor believed the treatment worked. I do not dispute that reasonable jurors could also convict Mr. Pon. But the government’s evidence of intent, as distinct from the evidence that Mr. Pon misdiagnosed and unnecessarily treated his patients, is not so overwhelming that this Court should affirm the judgment of guilt. The case for acquittal would have been even stronger if Mr. Pon had the chance to give his full surrebuttal. The government’s rebuttal evidence was highly prejudicial, or “very damning,” as the District Court saw it. In a case all about Mr. Pon’s intent, the government’s rebuttal certainly created the impression that Mr. Pon either 1) billed Medicare for treatments he never provided, or 2) billed Medicare for dozens of useless treatments on a blind eye. And then the government nailed Mr. Pon during his surrebuttal, emphasizing how limited an explanation he gave, just three entries, in the face of a spreadsheet showing years of treatments. The government implied Mr. Pon had nothing to say about the other entries, when of course it knew he had an explanation. The jury may well have 88 Case: 17-11455 Date Filed: 06/29/2020 Page: 89 of 89 taken this as essentially a confession to an unrelated fraud. A “defendant’s own confession is probably the most probative and damaging evidence that can be admitted against him.” Arizona v. Fulminante, 499 U.S. 279, 296, 111 S. Ct. 1246, 1257 (1991) (quotation marks omitted). Between the spreadsheet and the limited surrebuttal, so far as the jury knew Mr. Pon had no explanation at all for treating a blind eye. And so far as we know, the jury went back to their deliberations thinking Mr. Pon was a crook because of it. A rational jury allowed to hear the case with Mr. Pon’s explanation could render a different verdict than this one did. That being the case, I must say this error was harmful. I believe Mr. Pon should get a new trial. III. I believe a second jury should have decided Mr. Pon’s guilt or innocence. For this reason, I dissent from the majority opinion’s holding that any error in Mr. Pon’s case was harmless. 89