United States v. Simon

United States Court of Appeals
For the First Circuit

Nos. 20-1368
20-1412
UNITED STATES OF AMERICA,

Appellee/Cross-Appellant,

v.

RICHARD M. SIMON,

Defendant, Appellant/Cross-Appellee.

Nos. 20-1369
20-1411
UNITED STATES OF AMERICA,

Appellee/Cross-Appellant,

v.

SUNRISE LEE,

Defendant, Appellant/Cross-Appellee.

Nos. 20-1370
20-1413
UNITED STATES OF AMERICA,

Appellee/Cross-Appellant,

v.

JOSEPH A. ROWAN,

Defendant, Appellant/Cross-Appellee.
Nos. 20-1382
20-1409

UNITED STATES OF AMERICA,

Appellee/Cross-Appellant,

v.

JOHN KAPOOR,

Defendant, Appellant/Cross-Appellee.

Nos. 20-1410
20-1457

UNITED STATES OF AMERICA,

Appellee/Cross-Appellant,

v.

MICHAEL J. GURRY,

Defendant, Appellant/Cross-Appellee.

APPEALS FROM THE UNITED STATES DISTRICT COURT
FOR THE DISTRICT OF MASSACHUSETTS

[Hon. Allison D. Burroughs, U.S. District Judge]

Before

Howard, Chief Judge,
Selya, Circuit Judge,
and Gelpí,* District Judge.

William W. Fick, with whom Daniel N. Marx and Fick & Marx LLP
were on brief, for defendant Simon.
Peter Charles Horstmann for defendant Lee.

* Of the District of Puerto Rico, sitting by designation.
 Michael Kendall, with whom Karen Eisenstadt, Alexandra I.
Gliga, and White & Case LLP were on brief, for defendant Rowan.
Martin G. Weinberg and Kosta S. Stojilkovic, with whom Martin
G. Weinberg Law, P.C., Beth A. Wilkinson, Chanakya A. Sethi, and
Wilkinson Walsh LLP were on brief, for defendant Kapoor.
Megan A. Siddall, with whom Tracy A. Miner and Miner Orkand
Siddall LLP were on brief, for defendant Gurry.
David M. Lieberman, Attorney, Appellate Section, United
States Department of Justice, with whom Nicholas L. McQuaid, Acting
Assistant Attorneys General, Criminal Division, Robert A. Zink,
Acting Deputy Assistant Attorney General, Nathaniel R. Mendell,
Acting United States Attorney, Donald C. Lockhart, Appellate
Chief, and Mark T. Quinlivan, Fred Wyshak, K. Nathaniel Yeager,
and David G. Lazarus, Assistant United States Attorneys, were on
brief, for the United States.

August 25, 2021
 SELYA, Circuit Judge. A noted British ethologist once

observed that "[t]he total amount of suffering per year in the

natural world is beyond all decent contemplation." Richard

Dawkins, River Out of Eden 131-32 (Basic Books 1995). Some of

this suffering is unavoidable, but some is caused by those who

callously place profits over principle. The facts of this mammoth

case, as supportably found by the jury, tell a chilling tale of

suffering that did not need to happen. It involves a group of

pharmaceutical executives who chose to shunt medical necessity to

one side and shamelessly proceeded to exploit the sickest and most

vulnerable among us — all in an effort to fatten the bottom line

and pad their own pockets.

The tale told by this case chronicles the pernicious

practices employed by a publicly held pharmaceutical firm, Insys

Therapeutics, Inc. (Insys), with respect to the marketing and sale

of Subsys, a fentanyl-laced medication approved by the United

States Food and Drug Administration (FDA) for use in the treatment

of breakthrough cancer pain. When the government got wind of these

practices, it launched an investigation. That investigation

produced evidence that led a federal grand jury to indict seven of

the company's top executives on charges brought under the Racketeer

Influenced and Corrupt Organizations Act (RICO), 18 U.S.C

§ 1962(d). Two of the executives eventually entered into plea

agreements, but the rest stood their ground. Following a fifty-

- 4 -
one-day trial, the jury convicted the five remaining defendants as

charged (with an exception described below), and the district court

(again with an exception described below) declined to set aside

the jury verdicts. The court then sentenced the defendants to

prison terms of varying lengths, ordered defendant-specific

restitution, and directed the forfeiture of certain assets.

On appeal, the defendants — ably represented — raise a

gallimaufry of claims. The government cross-appeals, assigning

error to the district court's refusal to embrace the whole of the

jury verdicts and to its computation of the forfeiture amounts.

After careful consideration of an amplitudinous record, we uphold

the jury verdicts in full, affirm the defendants' sentences (which

are unchallenged), vacate the restitution and forfeiture orders,

and remand for further proceedings consistent with this opinion.

I

We begin with a snapshot of the relevant facts drawn

from the evidence adduced at trial. We then briefly rehearse the

travel of the case.

A

Insys is a pharmaceutical firm founded by one of the

defendants, Dr. John Kapoor. Under the Insys umbrella, Kapoor

sought to develop sublingual spray drug-delivery formulations.

The firm explored various options, but soon concentrated on

- 5 -
developing a sublingual fentanyl spray. This product came to be

called "Subsys."

In early 2012, the FDA approved Subsys for the treatment

of patients suffering from "breakthrough cancer pain." The term

"breakthrough cancer pain" is a term of art: it refers to brief

spikes in pain (typically lasting less than one hour) in patients

with cancer who are already dealing with constant and relatively

steady pain. All other uses of Subsys were deemed "off-label."

When Subsys went on the market, its FDA-approved label

declared that "[t]he initial dose of Subsys to treat episodes of

breakthrough cancer pain is always 100 micrograms." Moreover, the

label warned that "Subsys contains fentanyl," which is a "Schedule

II controlled substance with an abuse liability similar to other

opioid analgesics." Relatedly, the label carried a limitation on

who could prescribe the drug: due to "the risk for misuse, abuse,

addiction and overdose," Subsys could be prescribed "only through

a restricted program . . . called 'Risk Evaluation and Mitigation

Strategy'" (REMS). This program formed part of the FDA's

Transmucosal Immediate Release Fentanyl REMS Access Program, which

required patients, prescribers, and pharmacists to sign a form

stating that they understood the risks presented by the prescribed

drug.

Subsys made its debut in the marketplace in March of

2012 (shortly after FDA approval was secured). At that point in

- 6 -
time, Kapoor was serving as Insys's executive chairman, Michael

Babich was serving as its chief executive officer, Shawn Simon was

serving as its vice president of sales, and Matthew Napoletano was

serving as its vice president of marketing.

Around the time of the Subsys launch, Insys assembled a

marketing team. It proceeded to provide its sales force with

access to data that ranked physicians "based on their history of

prescribing within the opiate market, in particular, the fentanyl

market." The ranking system assigned a number between 1 and 10 to

each doctor — the higher the number the greater the volume of

prescriptions written. Salespeople were instructed to target

doctors ranked 5 or above and to give their "highest attention" to

those assigned a 10. They were also told to employ a "switch

strategy" aimed at persuading prescribers whose patients already

had been determined to need a similar fentanyl product to jettison

the similar product in favor of Subsys. Although the only approved

use for Subsys was for treatment of breakthrough cancer pain, most

of the prescribers listed in the database were pain-management

specialists, not oncologists.

Notwithstanding Insys's strategic plan, Kapoor was

disappointed with initial sales and revenue figures. He told

colleagues that it was "the worst f*****g launch in pharmaceutical

history he's ever seen." In Kapoor's view, the "main issue" was

that the majority of patients who started on Subsys would stay on

- 7 -
the drug only for the first month and would not refill their

prescriptions. Napoletano hypothesized that patients were

electing not to stick with Subsys because insurance companies were

choosing not to cover it. Patients, he suggested, did not want to

pay out of pocket to refill Subsys prescriptions.

Kapoor, though, had a different take: he attributed the

widespread failure to refill Subsys prescriptions to patients

"starting on too low of a dose." Because the Subsys label

specified the initial dose as 100 micrograms, Kapoor expressed

concern that patients who were used to a higher dose of a competing

product would not be satisfied with the pain management offered by

Subsys at that initial dosage. Consistent with Kapoor's concerns,

sales data (which Insys executives analyzed daily) showed that the

lower a patient's starting dose, the higher the "falloff rate."

By the fall of 2012, Insys had begun to overhaul its

marketing team. Shawn Simon was cashiered, and Alec Burlakoff

(previously a regional manager) replaced him as vice president of

sales. Defendant Joseph A. Rowan was promoted into Burlakoff's

former role. Defendants Sunrise Lee and Richard M. Simon were

installed as regional managers, and defendant Michael J. Gurry

became vice president for managed markets.1

1 To avoid any confusion between Richard Simon and Shawn
Simon, we subsequently refer to Richard Simon — and only Richard
Simon — as "Simon."

- 8 -
 In addition to these executive-suite changes, Insys

revamped its sales and marketing strategy. That fall, it hosted

both a national sales meeting and a national sales call to train

its sales force on a "new plan of attack." This plan had several

components:

• A new "switch program" allowed patients who were

transitioning to Subsys from a competing drug to

receive vouchers to defray the cost of Subsys for

as long as they needed it or until it was covered

by their insurance.

• A new "super voucher" program offered a means of

providing free product to patients.

• A specially crafted "effective dose" message

informed prescribers that, despite the statements

on the FDA-approved labelling, 100- or 200-

microgram doses were not effective. To complement

this "effective dose" messaging, sales

representatives were notified "each and every time"

a prescriber wrote a Subsys prescription for 100-

or 200-micrograms; and they were instructed to

report back within 24 hours both as to the reason

why the doctor had prescribed the low dose and as

to how the doctor planned to titrate the patient to

the "effective dose."

- 9 -
 • A revised compensation structure was put in place.

This structure rewarded sales representatives for

pushing doctors to prescribe higher doses of

Subsys. Under it, larger prescribed doses yielded

salespeople larger bonuses both because bonus

percentages were higher for higher doses and

because higher doses were more costly.

The icing on the cake was Insys's inauguration of a

speaker program in August of 2012. The ostensible "objective of

the program" was to provide "peer-to-peer education." To that

end, Insys would invite physicians whom it envisioned as potential

Subsys prescribers and the speaker (a fellow health-care provider)

would "present the information [about the drug] to them." These

presentations would take place through "online web hosting[s]" or

at "dinner meetings." Each sales region was to host a roughly

equal number of programs.

In its original incarnation, the speaker program never

got off the ground. Instead, Kapoor transmogrified it. About a

month after Napoletano announced the inauguration of the program,

Kapoor "put on hold all speaker programs effective immediately."

This directive emanated from Kapoor's disagreement with Napoletano

about what the objective of the program ought to be: as Kapoor

saw it, the speaker program "was designed for the speakers," not

- 10 -
for the physicians who comprised the audience. Kapoor "wanted

every speaker to write" Subsys prescriptions.

To accomplish this objective, Kapoor asked Napoletano

for a list of the doctors who served as speakers, along with data

as to "how many of them were writing [Subsys]" and data as to "what

percentage of the prescriptions came from them." Napoletano

balked, responding that "it's the attendees that you measure" —

not the speakers. Kapoor "was not in agreement with that" and

continued to insist upon a restructuring of the program.

In September, Kapoor, Burlakoff, Babich, and Napoletano

met to discuss the direction of the speaker program. Consistent

with Kapoor's vision, Burlakoff argued against the original peer-

to-peer education model. When Napoletano pointed out that "in

accordance with pharma code" each event had to have "a minimum of

two to four people" attend, Burlakoff replied that he "d[idn't]

care if there are any attendees" and that "he expect[ed] every

speaker to write" prescriptions. He said that the speaker program

should be "about the speaker and getting return from the speaker."

Although the meeting "was very contentious," Kapoor was satisfied

that his message had been received and proceeded to lift his "hold"

on the speaker program.

Burlakoff then emailed the sales force stating that

speaker programs are "the number one opportunity to grow [their]

business." He predicted that "[t]he hungry, motivated sales

- 11 -
representatives will be facilitating as many speaker programs as

humanly possible." He also suggested that a successful speaker

program would require salespersons to seek out speakers who are

"expert[s] with the utilization of Subsys in [their] clinical

practice" and who "have at least 20 patients on Subsys."

Even with this sharp change in direction, Insys's top

brass disagreed as to how to measure the program's success. In

October, Kapoor, Napoletano, Babich, and Burlakoff met regarding

that issue. Napoletano wanted to "track [the attendees] moving

forward to see if the presentation had any impact and if they

adopted the product in their practice." Burlakoff disagreed and

reiterated that "the metric to track is the speaker." The meeting

concluded with the issue still up in the air.

At a subsequent meeting, Kapoor resolved the issue. He

stated that he "wanted to make sure every speaker wrote" Subsys

prescriptions and "wanted a positive ROI" — a shorthand reference

to return on investment. The ROI, as Kapoor measured it, would be

the ratio between net revenue and the amount paid for speaker

services. After a heated exchange, Napoletano capitulated and

agreed to begin preparing reports tracking speakers and their

corresponding ROIs. These reports allowed Kapoor to "see how

successful [the] speakers were and how much product they were

writing, based on how much money [Insys] had given them so far."

Once this data became available, any speaker who "did not generate

- 12 -
at least two times in revenue what was being paid to them" was

"flagged" for a "temporary hold on programming." Refined to bare

essence, the flagged speakers "wouldn't get programs" and, thus,

would not receive honorariums.

This new protocol transformed the speaker programs from

pedagogical exercises into funding mechanisms for a pay-for-play

fandango. It is, therefore, unsurprising that with the new

protocol in place, Burlakoff sought to identify "whales." He

coined the term "whales" to refer to physicians who "ha[d] agreed

in a very clear and concise manner that they [were] up for the

deal, which [meant that] they [would] be compensated based on the

number of prescriptions of Subsys they wr[ote]." A corollary to

that deal was that "the more they wr[ote] and the more they

increase[d] the dose, the more they'[d] get paid to speak." At

Burlakoff's urging, regional sales managers were to have a "candid

conversation" with each potential whale and make clear that if the

physician was going to receive payments from Insys, he was "going

to write a significant amount of Subsys prescriptions to new

patients as well as increase the doses of current patients."

Burlakoff told sales managers to view speakers as their "business

partner[s]."

Burlakoff's whale hunt was fruitful: he identified many

whales, including Drs. Mahmood Ahmad, Gavin Awerbuch, Steven Chun,

Patrick Couch, Paul Madison, Judson Somerville, and Xiulu Ruan.

- 13 -
These prescribers were frequently mentioned on the daily 8:30 a.m.

management calls, in which Kapoor, Babich, Napoletano, Burlakoff,

and Gurry regularly participated. All of the whales committed to

prescribing large quantities of Subsys. And if a whale failed to

meet prescription expectations, an Insys representative would put

pressure on him to get him back on track.

Without exception, the prescription numbers of these

physicians increased when they joined the speaker program. In an

email, Burlakoff described the doctors as "clueless" because they

"prescribe strictly based on their relationship with the sales

manager." As a result of that relationship and the pressure that

sales representatives exerted, practitioners designated as

"[s]peakers" generated approximately $4,200,000 in net revenue (60

percent of Insys's total net revenue) after receiving more than

$550,000 in speakers' fees. Pleased with the success of the

reconstituted speaker program, Kapoor raised the speaker budget in

subsequent years.

Insys allocated speaker programs primarily to whales and

other prolific Subsys prescribers. These practitioners were paid

between $1,000 and $3,000 per event, depending on the particular

practitioner's "résumé or . . . influence." Speakers' payments

were routinely sent by mail. Multiple speaker events featured the

same practitioner. Insys initially capped annual speaking fees at

$100,000 per practitioner but later raised the ceiling to $125,000.

- 14 -
At a meeting in January of 2014, Babich, Burlakoff, and Richard

Simon compiled a list of "doctors that had the highest potential

to write." Burlakoff then "mobilized the sales force to go out

and make sure that these 19 or 20 doctors reached their [fees]

cap."

Despite the largess shown to speakers, the speaking

events themselves had little to no attendance. Often, only the

speaker, a friend or family member, and the sales representative

were on hand. Even when more people were in attendance, the

speaker programs were mostly "social outings" or "just a reason to

gather people and have dinner and pay [the doctor]." Although

sales representatives were required to submit sign-in forms and

attendee evaluation forms to a third-party compliance firm (Sci

Medica), they frequently submitted inaccurate documentation,

including sign-in sheets with names and signatures of people who

were not present, to give the speaking programs an aura of

legitimacy. And when Kapoor replaced Sci Medica with an in-house

compliance officer, the apocryphal documentation continued to

flow.

While the revamped speakers' program drove up the volume

of Subsys prescriptions, insurance coverage remained a problem.

Medicare, Medicaid, and private insurance companies covered the

cost of Subsys prescriptions only if a practitioner obtained prior

authorization to prescribe the drug. And because of the FDA label,

- 15 -
coverage was limited to patients with a current cancer diagnosis

who both suffered from breakthrough cancer pain and already had

tried other opioid medication.

Nor did the coverage limitations stop there. As a

condition precedent to coverage, insurers required that a patient

had tried a generic fentanyl product that had either failed to

ameliorate the breakthrough cancer pain or proved difficult to

ingest. To seek prior authorization, a practitioner typically

submitted patient and diagnosis information to the insurer, and

the insurer relied upon the accuracy of the submitted information

in its decisionmaking. When Insys launched Subsys, it processed

prior authorization requests through a third party and achieved

only a 30-35 percent success rate for prior authorization

approvals.

To enhance the approval rate, Gurry suggested bringing

the approval process in-house. With Kapoor's blessing, Gurry hired

Elizabeth Gurrieri in October of 2012 to found the Insys

Reimbursement Center (IRC), which operated out of Insys

headquarters. Insys created an opt-in form through which Subsys

prescribers could authorize the IRC to contact insurers and request

prior authorizations. The form listed patient information that

insurers typically would request during the prior authorization

process, such as whether the patient had tried certain medications.

Particular items from the list could be checked off as applying to

- 16 -
a specific case. This streamlined the process: a prescriber would

sign and fax an opt-in form to the IRC; the IRC would call the

insurer; and if the insurer needed additional information, the IRC

would reach out to the sales representative who would then follow

up with the prescriber. Insys encouraged physicians to use the

IRC, knowing that if the prior authorization was approved, "[t]he

sales rep would get paid, Insys would get paid, and the script

would get paid." A pilot program achieved an approval rate of 65-

70 percent. As a result, Insys quickly transitioned the IRC out

of its pilot phase and expanded it. Gurrieri was promoted to

manager of reimbursement services in March of 2013.

The IRC proved to be a rousing success. It owed much of

its success to the sales representatives. They interacted with

the physicians and collected documentation requested by insurers

during the prior-authorization process. A sales representative

would often spend at least one day per week in a physician's

office, reviewing patient files, assisting with authorizations,

and completing the opt-in forms.

Another factor in the IRC's success was the hiring of

"area business liaison[s]." These individuals were assigned to

the physicians who prescribed Subsys in substantial volume. Each

area business liaison worked in a physician's office processing

authorizations, but was paid by Insys, thereby reducing the

physician's overhead.

- 17 -
 The third, and perhaps most impactful, factor in the

IRC's success was Insys's decision to begin collecting data on

each coverage decision. The IRC identified diagnoses and

conditions that historically had prompted particular insurers to

approve Subsys prescriptions. It proceeded to list these diagnoses

and conditions on the opt-in form, and sales representatives

encouraged physicians to employ them when seeking Subsys

authorizations. For example, Gurrieri noted success using "the

terminology 'history of cancer,' which means that they didn't have

cancer at the time but they had a history of cancer." Once

salespeople heard that use of that phrase could help obtain

insurance approval, the IRC, "all of a sudden, saw more opt-ins

having 'history of cancer' on them, which [led] to better approval

ratings."

Management regularly discussed the IRC on the daily 8:30

a.m. calls. All updates about the IRC were communicated by Gurry

during those calls. Although Insys had made great strides in

upping its approval rate, Kapoor put constant pressure on the IRC

to achieve a rate of 90 percent or higher. Striving to attain

this benchmark, the IRC started to offer training sessions to sales

representatives on "how to get the drug approved." Similarly,

Gurry started to advise sales representatives about what diagnoses

and conditions should be checked on the opt-in forms. He famously

directed IRC employees "to ride the gray line," that is, to "work

- 18 -
around the insurance companies" and "find ways around their

questions." Following that direction, the IRC developed

strategies to mislead insurers into granting prior authorizations

for the use of Subsys. Some of these strategies included

misleading the insurer into believing that the caller was calling

from the physician's office rather than from the IRC; representing

that a patient had cancer even if the available information

reflected only a history of cancer; giving the ICD-9 diagnosis

code as "338" to obscure the fact that the diagnosis was chronic

pain (which uses code 338.29 or 338.4) and not cancer pain or

neoplasm-related pain (which uses code 338.3); listing tried-and-

failed medications that the patient had never used; and falsely

stating that patients had dysphagia (difficulty swallowing).

Insys expected insurance companies to ask whether a

physician had prescribed Subsys to treat "breakthrough cancer

pain." Gurrieri instructed IRC staff to respond with "the spiel,"

which was pat phrasing designed to obfuscate the purpose of the

prescription. The essence of the spiel was that "[t]he physician

is aware that the medication is intended for the management of

breakthrough pain in cancer patients, and the physician is treating

the breakthrough pain." Phrased in this way, the expectation was

that "the person on the other end of the phone would be misled to

think the patient had cancer and approve the prior authorization."

- 19 -
 The record makes manifest that the IRC, in practice, was

more interested in transmitting information that would prompt

favorable coverage determinations than it was in transmitting

accurate information. Through the IRC, the insurers were fed a

steady diet of deceptions, evasions, and half-truths.

Just as sales representatives were incentivized to push

physicians to prescribe higher doses of Subsys, IRC staffers were

incentivized to obtain insurance approvals. Goals known as "gates"

were set weekly. If the gate was opened, the staff member (usually

paid a low hourly wage) would receive a bonus.

The cocktail that Insys had mixed — including its revised

marketing and sales strategies, its use of speaker programs as

vehicles for bribes to physicians, its use of business liaisons,

and its no-holds-barred tactics within the IRC — proved to be

lucrative. Insys was able to go public only a year after

introducing Subsys to the market. Within two years after the

initial public offering, the company reached a market cap of over

$3,000,000,000. And by the end of 2015, Insys's stock price had

nearly quadrupled. Throughout, the defendants received

substantial salaries, bonuses, and stock options.

But Insys's meteoric rise appeared too good to be true,

and the company attracted unwanted attention. When federal

authorities began probing the details of how Insys was marketing

Subsys, the defendants' scheme began to unravel.

- 20 -
 B

In the wake of the federal investigation, a federal grand

jury sitting in the District of Massachusetts charged Kapoor, Lee,

Simon, Gurry, and Rowan with conspiracy to distribute Subsys

through a pattern of racketeering activity.2 See id. The

conspiracy was effected, the indictment said, through acts of mail

fraud, see id. § 1341; honest-services mail fraud, see id. §§ 1341,

1346; wire fraud, see id. § 1343; honest-services wire fraud, see

id. §§ 1343, 1346; and Controlled Substances Act (CSA) violations,

see 21 U.S.C. § 841(a)(1). Following lengthy pretrial

maneuvering, not relevant here, a fifty-one-day trial ensued.

The jury returned guilty verdicts against all of the

defendants. In connection with those verdicts, the jury made a

series of special findings that all the defendants were guilty of

committing predicate acts of mail-fraud and wire-fraud, and that

all the defendants (except Gurry) were guilty of agreeing to

distribute a controlled substance and to commit honest-services

mail fraud and honest-services wire fraud.

The defendants moved for judgments of acquittal and/or

new trials. See Fed. R. Crim. P. 29(a), 33(a). The district court

granted in part the joint motion for judgments of acquittal filed

by Kapoor, Lee, Simon, and Rowan, vacating as to them the adverse

2Babich and Burlakoff were also named as defendants. Both
of them entered guilty pleas before trial.

- 21 -
findings with respect to the CSA and honest-services predicates.

See United States v. Gurry, 427 F. Supp. 3d 166, 222 (D. Mass.

2019). But with respect to all five defendants, the court rejected

their challenges to the mail- and wire-fraud predicates, rejected

their efforts to secure judgments of acquittal, and declined to

order a new trial. See id. The court sentenced the defendants to

terms of immurement of varying lengths and entered a series of

restitution and forfeiture orders.3 See United States v. Babich,

No. 16-CR-10343, 2020 WL 1235536, at *10 (D. Mass. Mar. 13, 2020).

All of the defendants appealed, and the government cross-appealed.

II

In this venue, we are faced with a kaleidoscopic array

of claims. Kapoor, Lee, Simon, and Rowan contend that the evidence

was insufficient to convict on the various mail- and wire-fraud

predicates, assigning error to the district court's denial of their

joint motion for judgment of acquittal. Relatedly, all defendants

claim error in the admission of patient-harm testimony and

3 The court sentenced Kapoor to a sixty-six-month term of
immurement, ordered restitution of $59,755,362.45, and directed
forfeiture of $1,914,771.20. As to Lee, the court imposed a prison
sentence of a year and a day, ordered restitution of $5,000,000,
and directed forfeiture of $1,170,274. As to Simon, the court
imposed a thirty-three-month term of immurement, ordered
restitution of $5,000,000, and directed forfeiture of
$2,338,078.72. Gurry's sentence was identical to Simon's, except
that he was ordered to forfeit $3,390,472.89. Finally, the court
sentenced Rowan to serve a twenty-seven-month prison term, ordered
restitution of $5,000,000, and directed forfeiture of
$2,078,217.66.

- 22 -
prejudicial spillover arising out of the government's efforts to

prove the CSA and honest-services predicates through that

testimony.

Some defendants raise individual claims as well. Lee

challenges the district court's order denying her pretrial motion

for severance, certain of the district court's evidentiary

rulings, and one of the district court's jury instructions. Rowan

claims that the government unlawfully withheld exculpatory

material, and that the district court erred in denying his mid-

trial motion to compel production of that material. The

defendants, jointly and severally, offer a plethora of reasons as

to why they — or some of them — ought to be granted new trials,

including claims relating to allegedly conflicted counsel, weight

of the evidence, and prosecutorial misconduct during closing

arguments.4 And although the defendants do not challenge their

4At various points, some of the defendants purport to
incorporate by reference arguments made by other defendants. See
Fed. R. App. P. 28(i). For example, a footnote in Rowan's brief
purports to "adopt[] and incorporate[] the facts and arguments in
the briefs of co-defendants Dr. John Kapoor, Richard Simon, Michael
Gurry, and Sunrise Lee, whether or not this brief explicitly
mentions them." Lee's and Gurry's briefs each contains similar
statements.
The rule in this circuit is that "[a]doption by
reference . . . cannot occur in a vacuum; to be meaningful, the
arguments adopted must be readily transferrable from the
proponent's case to the adopter's case." United States v. David,
940 F.2d 722, 737 (1st Cir. 1991). Given this rule, the shorthand
adoption by reference attempted by these defendants is partially
an empty gesture. And to the extent that the incorporated
arguments pass through this screen, they fail on the merits (except

- 23 -
prison sentences, they do contest the district court's ancillary

orders awarding restitution and forfeiture. The government cross-

appeals, assigning error to the district court's order vacating

the jury's findings adverse to Kapoor, Lee, Simon, and Rowan on

the CSA and honest-services predicates. It also appeals the

district court's calculation of forfeiture amounts with respect to

Lee, Simon, Gurry, and Rowan.

We start our journey with the parties' competing claims

concerning the sufficiency of the evidence with respect to the CSA

and honest-services predicates. From there, we wend our way

through the remaining sufficiency-of-the-evidence claims, the

admissibility of the patient-harm testimony, questions pertaining

to evidentiary spillover, and a myriad of other claims of trial

error. Our journey ends with an appraisal of the parties' opposing

views regarding issues related to restitution and forfeiture.

III

Under RICO, it is a crime "for any person employed by or

associated with any enterprise engaged in, or the activities of

which affect, interstate or foreign commerce," to conspire "to

conduct or participate, directly or indirectly, in the conduct of

such enterprise's affairs through a pattern of racketeering

activity." See 18 U.S.C. § 1962(c), (d). A pattern of

with respect to certain incorporated arguments, identified in
Parts XIV and XV, infra, regarding restitution and forfeiture).

- 24 -
racketeering activity requires at least two predicate racketeering

acts within ten years of each other. See id. § 1961(5). A

defendant need not have "agree[d] to commit or facilitate each and

every part of the substantive offense" in order to be found guilty.

Salinas v. United States, 522 U.S. 52, 63 (1997). Nor need such

a defendant be capable of committing the substantive offense. See

id. Instead, "[a]ll the government need show is that the defendant

agreed to facilitate a scheme in which a conspirator would commit

at least two predicate acts, if the substantive crime [had]

occurred." United States v. Rodríguez-Torres, 939 F.3d 16, 29

(1st Cir. 2019); see Salinas, 522 U.S. at 65 ("A [RICO] conspirator

must intend to further an endeavor which, if completed, would

satisfy all of the elements of a substantive criminal offense, but

it suffices that he adopt the goal of furthering or facilitating

the criminal endeavor.").

In this case, the critical questions involve whether —

as to each defendant — the record sufficiently supports the jury's

verdict that he or she, directly or through another conspirator,

committed the charged offenses. While the jury answered these

questions in the affirmative (except as to Gurry, who was found

guilty only with respect to the mail- and wire-fraud predicates),

the district court found the government's proof of the CSA and

honest-services predicates wanting. The court ruled that,

although "it would not have been unreasonable for the jury to infer

- 25 -
that the nefarious tacit understanding the Government describes

existed," it "would have been equally reasonable for the jury to

infer from the same evidence that no such tacit understanding

existed." Gurry, 427 F. Supp. 3d at 186. Because the proof "gives

equal or nearly equal circumstantial support to a theory of guilt

and a theory of innocence," the court vacated the jury findings

regarding the CSA and honest-services predicates vis-à-vis Kapoor,

Lee, Simon, and Rowan. Id. (quoting United States v. Burgos, 703

F.3d 1, 10 (1st Cir. 2012)).

The government appeals from this ruling. Our task is

familiar. We afford de novo review to the district court's rulings

on the defendants' joint motion for judgment of acquittal. See

United States v. Kilmartin, 944 F.3d 315, 325 (1st Cir. 2019);

United States v. Olbres, 61 F.3d 967, 970 (1st Cir. 1995). "Where,

as here, the defendant[s] challenge[] the sufficiency of the

evidence, all of the proof 'must be perused from the government's

perspective.'" Kilmartin, 944 F.3d at 325 (quoting United States

v. Gomez, 255 F.3d 31, 35 (1st Cir. 2001)). This lens demands

that "we scrutinize the evidence in the light most compatible with

the verdict, resolve all credibility disputes in the verdict's

favor, and then reach a judgment about whether a rational jury

could find guilt beyond a reasonable doubt." Olbres, 61 F.3d at

970 (quoting United States v. Taylor, 54 F.3d 967, 974 (1st Cir.

1995)).

- 26 -
 In conducting this tamisage, "we must honor the jury's

evaluative choice among plausible, albeit competing, inferences."

United States v. Rodríguez-Vélez, 597 F.3d 32, 40 (1st Cir. 2010).

When all is said and done, "[t]he court need not be convinced that

the verdict is correct; it need only be satisfied that the verdict

is supported by the record." Kilmartin, 944 F.3d at 325.

Consequently, a "verdict must stand unless the evidence is so scant

that a rational factfinder could not conclude that the government

proved all the essential elements of the charged crime beyond a

reasonable doubt." Rodríguez-Vélez, 597 F.3d at 39 (emphasis in

original).

Our next chore is to elaborate the elements of the CSA

predicates. The CSA makes it a crime "for any person knowingly or

intentionally . . . to manufacture, distribute, or dispense . . . a

controlled substance." 21 U.S.C. § 841(a)(1). Even so, licensed

health-care practitioners (typically, physicians) registered under

the CSA are authorized to dispense controlled substances. See id.

§ 822(b). This authorization, though, is not absolute. Such

practitioners face criminal liability "when their activities fall

outside the usual course of professional practice." United States

v. Moore, 423 U.S. 122, 124 (1975); see United States v.

Limberopoulos, 26 F.3d 245, 249 (1st Cir. 1994) ("[T]he statute

applies to a pharmacist's (or physician's) drug-dispensing

activities so long as they fall outside the usual course of

- 27 -
professional practice."). Because a RICO conspiracy conviction

requires proof that defendants "specifically intended that some

conspirator commit each element of" the predicate racketeering

acts, Ocasio v. United States, 136 S. Ct. 1423, 1432 (2016)

(emphasis in original), the government had to prove that the

defendants specifically intended that a licensed practitioner

would prescribe Subsys "with no legitimate medical purpose,"

United States v. Volkman, 797 F.3d 377, 391 (6th Cir. 2015); see

United States v. Feingold, 454 F.3d 1001, 1008 (9th Cir. 2006).

Against this backdrop, we canvass the proof as to Kapoor,

Lee, Simon, and Rowan.

A

Through his motion for judgment of acquittal, Kapoor

challenged, inter alia, the jury's finding that he was guilty of

conspiracy to commit racketeering activities through a pattern of

racketeering acts that included the CSA and honest-services

predicates. As we have said, the district court set aside the

jury's findings with respect to those predicates. The question is

one of evidentiary sufficiency.

The record is replete with support for the proposition

that Kapoor intended physicians to write medically illegitimate

prescriptions. Kapoor sought out pill mill doctors (that is,

doctors who were notorious for their readiness to prescribe drugs

regardless of medical necessity). See, e.g., United States v.

- 28 -
Iriele, 977 F.3d 1155, 1161 (11th Cir. 2020) (describing as "pill

mill" a clinic where people "could get prescriptions for their

controlled substances of choice with few, if any, questions

asked"). For instance, Burlakoff testified that, to increase

sales, Kapoor wanted him to do "[w]hatever it took with whomever

[they] called on, including pill mill doctors."

Perhaps the best illustration of Kapoor's intent is

found in the evidence concerning his attitude toward Dr. Madison.

Kapoor encouraged dealings with Dr. Madison despite having

reviewed an email in which a sales representative wrote that "Dr.

Madison runs a very shady pill mill and only accepts

cash. . . . He basically just shows up to sign his name on the

prescription pad, if he shows up at all." Kapoor also knew that

another sales representative had observed a "shady setup" in Dr.

Madison's office with "many patients . . . going in and out of

there . . . just seeking medication." As one expert witness put

it, this prescribing behavior was inconsistent with a doctor's

duty to carry out "those things that are necessary to reasonably

diagnose the problem," such as "history taking, physical

examination, and the obtaining and evaluation of diagnostic

studies."

Although on unmistakable notice of the kind of operation

that Dr. Madison appeared to be running, Kapoor pursued him.

Babich testified that Kapoor "wanted [Dr. Madison] to write the

- 29 -
drug" and awarded him speaker programs (and, thus, kickbacks) "as

much as once a week." This was consistent both with Babich's

description of Kapoor's avowed "philosophy" and with other

evidence reflecting Kapoor's appetite for whales. The jury

reasonably could have found that Kapoor's decision to continue

courting and compensating Dr. Madison, notwithstanding his

knowledge that the doctor was running a notorious pill mill, was

proof of at least a tacit understanding of Kapoor's culpable role

in the distribution scheme. See United States v. King, 898 F.3d

797, 809 (8th Cir. 2018).

Kapoor complains that this is a bridge too far. He

laments that he received hundreds of emails a day, that he was

busy with other business pursuits and charitable endeavors, and

that Dr. Madison is only one of 13 doctors discussed in the four-

page email. It follows, Kapoor suggests, that a reasonable jury

could not infer that he read the sales representative's description

of Dr. Madison.

This suggestion is little more than whistling past the

graveyard. It conveniently overlooks that the jury heard evidence

that Kapoor demanded information on "every [Subsys] script that

was written" and "every doctor that wrote it." He demanded

spreadsheets to parse doctor-level data and sought to identify

"whales" — doctors who understood that, in exchange for speaker-

- 30 -
program payments, they would prescribe "a significant amount of

Subsys prescriptions."

What is more, Babich testified that Kapoor expressed

great interest in these kinds of sales reports. Kapoor "want[ed]

every single rep every Friday to e-mail [Babich] a list of all

their top physicians and what happened with those top physicians

that week." An assistant "print[ed] these out" and put "them on

[Kapoor's] desk for Monday morning, so he c[ould] review" them.

Given that level of attention to detail vis-à-vis prescribers, the

inference that Kapoor read the email about Dr. Madison seems

compelling.

Last — but surely not least — Babich confirmed that he

gave the four-page email directly to Kapoor. He also testified

that — several months after he had forwarded that email about Dr.

Madison to Kapoor — the same sales representative again reiterated

that Dr. Madison operated a pill mill and added that Dr. Madison

faced possible legal action. Babich described this matter as "a

serious issue" and testified that he personally reviewed this

information with Kapoor. Kapoor responded that Dr. Madison "still

has a medical license. I don't want him taken off the call list"

for speaker programs.

We need not tarry. The evidence, taken in the light

most hospitable to the verdict, plainly supports the inference

that Kapoor knew of Dr. Madison's illegitimate prescribing habits

- 31 -
yet took steps to ensure that he would continue prescribing Subsys.

Indeed, the evidence warrants an inference that Kapoor sought to

recruit Dr. Madison as a speaker (that is, as a kickback recipient)

precisely because of these habits. Such an inference is consistent

with other evidence that pill mill doctors were prized by Kapoor:

he tracked physicians' prescription patterns, gave favorable

treatment to the doctors who prescribed Subsys most profligately,

and — according to Burlakoff — did "whatever it took" to increase

Subsys sales. As Burlakoff put it, Kapoor's message to the sales

force was that "pill mills for [Insys] meant dollar signs."

The evidence also showed that Kapoor led Insys's effort

to influence physicians' prescription decisions through "effective

dose" messaging. The FDA-approved label stated that "[t]he initial

dose of Subsys to treat episodes of breakthrough cancer pain is

always 100 micrograms." Doctors were supposed to "look at one

patient at a time" and "titrate one patient at a time" to the dose

of the drug that achieves "the desired effect." Noting that

patients on higher doses were more likely to refill their Subsys

prescriptions, Kapoor sought to ride roughshod over this regime

and "move patients to higher doses." His mantra was to "push the

dose." To that end, he incorporated into the speaker program

kickbacks for dosage increases — the greater the increase, the

greater the payout. Predictably, Kapoor's campaign to increase

dosages resulted in the sales force negotiating dosage agreements

- 32 -
with doctors. And Insys closely monitored these agreements: for

example, when Dr. Somerville's dosage numbers appeared to be low,

a sales representative was instructed to "[d]rill into [the medical

assistant's] head that every refill has to be 180 to 240

[micrograms]" because "Dr. Somerville agreed to do this."

To sum up, the evidence plainly supports a finding that

Kapoor intended practitioners to prescribe Subsys as much as

possible, even when there was no medical necessity for the drug or

the dosage prescribed. His "effective dose" campaign was designed

to dissuade doctors from prescribing medically appropriate lower

dosages and to accelerate dose titration. A reasonable jury could

infer that, by taking these actions, Kapoor pushed physicians — in

Burlakoff's words — to "initiate a new habit" that would transform

patients into repeat customers, quite apart from medical

necessity. See United States v. Clough, 978 F.3d 810, 820-21 (1st

Cir. 2020) (concluding that giving "opioid-dependent patients high

dosages of this highly-addictive fentanyl drug, even when patients

had no problems with their existing regimen" supported reasonable

inference that defendant's "behavior was not reminiscent of a

physician assistant prescribing based on need, but rather [that]

of a drug pusher"). And having thrown medical necessity to the

wind, Kapoor's "push the dose" message effectively directed Insys

salespersons, who were not health-care professionals, to enforce

mandatory ranges of dosages. Following Kapoor's lead, they shaped

- 33 -
doctors' prescription decisions without regard to any individual

patient information by getting the doctors to commit to meeting

prescription-quantity numbers on a weekly basis. Jurors are

allowed to use common sense and — surveying this unattractive

tableau — a reasonable jury could have inferred that Kapoor, in

"push[ing] the dose," intended doctors to increase doses of Subsys

regardless of who the patient was or what the patient's medical

needs might be. See United States v. Guzman, 571 F. App'x 356,

363 (6th Cir. 2014) ("[T]he jury could reasonably infer the

requisite agreement to distribute controlled substances, as well

as knowledge and participation" from "evidence showing that

[defendants] tried to modify the prescribing practices of another

nurse practitioner," including by directing a "nurse to prescribe

short-acting rather than long-acting medications and to prescribe

prednisone for all customers.").

So, too, a reasonable jury could conclude that when drug

wholesalers reported suspicious Subsys ordering patterns to the

federal Drug Enforcement Administration (DEA), Kapoor sought to

tamp down any suspicions so that Insys could continue its modus

operandi while concealing it from outside scrutiny. Wholesalers

serve as middlemen between pharmaceutical manufacturers and

pharmacies and they impose quantity limits on the amount of

controlled substances that a pharmacy can receive each month. When

wholesalers notice suspicious ordering patterns, they are

- 34 -
obligated to notify the DEA. During the relevant time frame,

several pharmacies that served Insys speakers overshot their

monthly quantity limits. As a result, drug wholesalers froze

Subsys shipments to those pharmacies.

Insys executives knew that the reason for the freeze was

that wholesalers' software algorithms to monitor order patterns

had flagged Subsys orders as "potentially suspicious." In an email

to Kapoor, Christopher Homrich, an Insys executive, told him that

it was "very likely" that the DEA software would flag future orders

from those pharmacies as "suspicious" due to the "material"

increase in projected Subsys sales. Such freezes would be inimical

to Insys's strategic aim of getting doctors to prescribe Subsys in

heavy doses and without regard to medical necessity. Because

"Kapoor wanted to keep doing business" with the physicians

(particularly the whales) associated with the targeted pharmacies,

he demanded that Insys executives "find an alternative to make

sure one of our top customers has the product."

To accomplish this end run around the DEA and to avoid

the imminent freezes, Kapoor decided to explore a direct-ship

option. Such an option would have Insys ship Subsys straight to

the pharmacy associated with the prescribing doctor. Insys's

distribution manager (Dion Reimer) advised against this setup "at

least a dozen times." Given that Subsys was a controlled

substance, "[t]rying to circumvent any of the systems that are out

- 35 -
there could raise red flags." Kapoor disregarded Reimer's protests

and authorized sales representatives to negotiate supply

commitments and direct-ship agreements with individual doctors.

At his direction, Insys proceeded down this crooked path.5

Kapoor insists that the direct-ship agreements were made

only to "maximize sales of Subsys" and that he was "not trying to

evade DEA rules." But something that prowls like a tiger and

growls like a tiger usually is a tiger — and Reimer's assessment

of the direct-ship option as a means of circumventing DEA

guidelines seems spot-on.

In arguing to the contrary, Kapoor points to Babich's

testimony recounting how Insys retained "outside attorneys who

have expertise with the DEA rules" to ensure the direct-ship

arrangements were done "the right way." Babich also testified,

though, that Insys "did not tell the lawyers who drafted the

agreement[s] that [Insys was] providing kickbacks to the

physicians" associated with these pharmacies. According to

Babich, that omission was deliberate: the company feared that the

bribes contravened federal anti-kickback law. Under certain

circumstances, a party's retention of counsel may (as Kapoor

Unscrupulous practitioners apparently welcomed the direct-
5

ship option. In one of the weekly sales representative reports,
Kapoor was informed that Dr. Ahmad had committed to write "more
scripts than [Insys] can handle . . . once the pharmacy issue is
resolved." Other practitioners, including Drs. Couch, Ruan, and
Awerbuch, also benefited from direct-ship agreements.

- 36 -
suggests) ground an inference of benevolent motive. See United

States v. Powers, 702 F.3d 1, 9 (1st Cir. 2012). But viewing what

happened here in context, a jury instead could reasonably infer

that direct-ship agreements were evidence of Kapoor's efforts to

have doctors continue to prescribe Subsys illegitimately. See id.

(explaining that advice-of-counsel defense "is not available to

one who omits to disclose material information to advisors"

(quoting Janeiro v. Urological Surgery Prof. Ass'n, 457 F.3d 130,

140 (1st Cir. 2006))).

Taking the evidence in the light most favorable to the

verdict, a jury reasonably could conclude — as this jury did —

that Kapoor relentlessly pursued pill mill doctors, pressured

health-care practitioners to increase dosages regardless of

medical need (through financial incentives and upfront

prescription commitments), knew of and encouraged certain

physicians' illegitimate prescribing habits, and — facing

regulatory scrutiny for the burgeoning sales generated through

these tactics — tried to hide the true state of affairs by cutting

out the middleman. Cf. Volkman, 797 F.3d at 391 (holding that

evidence that defendants "were aware of the reality that the

prescriptions from their clinic had no legitimate medical purpose"

and that "[i]nstead of rectifying the . . . issues with [the]

prescriptions, [defendants] exacerbated the problem

by . . . cutting out the middleman" sufficed "for a jury to find

- 37 -
that [defendants] executed a plan to unlawfully distribute

controlled substances with no legitimate medical purpose").

Consequently, we are satisfied that the adverse finding against

Kapoor as to the CSA predicates was supported by the record and,

therefore, should have been allowed to stand.

This holding also dictates that we reinstate the jury's

verdict against Kapoor as to the honest-services predicates.

Federal law prohibits a "scheme or artifice to defraud," 18 U.S.C.

§§ 1341, 1343, including "a scheme or artifice to deprive another

of the intangible right of honest services," id. § 1346. A person

contravenes this provision if, "in violation of a fiduciary duty,

[he] participate[s] in bribery or kickback schemes." Skilling v.

United States, 561 U.S. 358, 407 (2010). Kapoor disputes the

"fiduciary duty" element and contends that the government failed

to prove that he specifically intended health-care professionals

to write medically illegitimate Subsys prescriptions.

As the district court noted, the "overlap" between the

CSA and honest-services predicates is striking. Gurry, 427 F.

Supp. 3d at 187-88. Just as the jury instructions for the CSA

predicates required proof beyond a reasonable doubt that a

particular defendant agreed that a health-care practitioner would

prescribe Subsys outside the usual course of medical practice, the

honest-services predicates required evidence that "the Defendant

agreed and specifically intended that health-care practitioners

- 38 -
would breach their fiduciary duty to their patients by prescribing

Subsys or a particular dose of Subsys outside the usual course of

professional practice and not for a legitimate purpose."

Accordingly, the evidence supporting the intent element of the CSA

predicates was "coextensive" with the evidence supporting the

fiduciary duty element of the honest-services predicates. Id. at

188.

That is game, set, and match. Because we already have

concluded that the evidence supports the jury's finding with

respect to Kapoor's guilt regarding the CSA predicates, we must

perforce conclude that the evidence supports the findings with

respect to Kapoor's guilt regarding the honest-services

predicates. It follows that we must reverse the district court's

partial grant of the Rule 29(c) motion in favor of Kapoor and

reinstate the jury's findings as to him insofar as they pertain to

both the CSA and honest-services predicates.

B

The district court set aside the jury's finding that Lee

was guilty of conspiracy to commit racketeering activities through

a pattern of racketeering acts that included the CSA and honest-

services predicates. The jury heard evidence, though, that Lee

supervised the sales representative who reported that Dr. Madison

had a "shady setup" and that patients at Dr. Madison's office "were

just seeking medication." When the sales representative spoke to

- 39 -
Lee about her concerns with Dr. Madison's potential law-

enforcement issues, Lee replied that "[i]t was okay." Like Kapoor,

she appeared unfazed by Dr. Madison's potential criminal liability

and "ensured that Dr. Madison understood that he would

speak . . . as much as [Insys] can utilize him" — which meant, of

course, that Dr. Madison would continue to receive kickbacks. The

only condition was that "he would prescribe a significant amount

of Subsys, more and more as time went on, and increas[e] the dose."

This condition had nothing to do with medical necessity.

Lee's hot pursuit of Dr. Madison supports the conclusion

that getting doctors to write illegitimate prescriptions was not

merely an unforeseeable risk of her work for Insys but, rather, an

integral part of the business model that she assiduously followed

while doing that work. As Babich explained, Dr. Madison was made

a speaker notwithstanding that his clinic was a pill mill

"[b]ecause he was the biggest writer of the type of product in the

Chicago land area, and getting that revenue was very important to

[Insys] as a company." A jury could reasonably infer that Insys

knowingly counted on revenue from illegitimate prescriptions and

that Lee (as a regional sales manager who benefitted handsomely

from greater sales) intended to keep that revenue stream flowing

even if it meant prescribing Subsys to patients who did not

legitimately need it.

- 40 -
 Other evidence corroborated the conclusion that Lee

intended prescribing doctors to expand the company's customer base

to people who did not qualify medically to use Subsys. The

regional managers were instructed by Simon "to get a specific

number of scripts per week that is mutually agreed to and an

outline of how [the representatives who reported to them] will

hold [them]selves and [their] customers to this plan." In other

words, salespeople were to negotiate prescription quotas with the

doctors in their territories. These quotas had no apparent

relationship to either medical necessity or patient needs, and the

jury had an ample basis for inferring that Lee followed Simon's

instructions.

Here, too, Lee's experience with Dr. Madison exemplifies

the point. As a speaker, Dr. Madison was expected to maintain or

exceed previous prescription-writing numbers. When Dr. Madison

fell short, Lee would order a sales representative "to continue to

put pressure on [Dr. Madison]" and tell the doctor "that if he's

going to keep doing these programs, he needs to keep his writing

up." There was no medically informed rationale for Dr. Madison's

quota, and his agreement to abide by such a quota is a surefire

sign that Lee knew that, under that agreement, Dr. Madison would

be prescribing Subsys illegitimately. Cf. United States v. Hughes,

895 F.2d 1135, 1142 (6th Cir. 1990) (reaching similar conclusion

regarding quota for blood tests). Her incessant enforcement of

- 41 -
the quota therefore is evidence that she intended for Dr. Madison

to write those illegitimate prescriptions. The way that Lee did

business with Dr. Madison is emblematic of her intent to have

health-care practitioners forsake medical necessity for financial

gain.

We conclude that the adverse finding against Lee as to

the CSA predicates was supported by the record and, therefore,

should have been allowed to stand. And as with Kapoor, see supra

Part III(A), this holding dictates that we reinstate the jury's

findings as to the honest-services predicates. It follows that we

must reverse the district court's partial grant of the Rule 29(c)

motion in favor of Lee as it pertains to both the CSA and honest-

services predicates.

C

The district court set aside the jury's finding that

Simon was guilty of conspiracy to commit racketeering activities

through a pattern of racketeering acts that included the CSA and

honest-services predicates. Once again, we disagree. On this

record, a reasonable jury could conclude that Simon (also a

regional manager) intended health-care providers to prescribe

Subsys outside the usual course of professional practice.

Simon encouraged the sales force to agree with each

practitioner on a "specific number of scripts per week" — a quota

— and to "push the dose." Relatedly, sales representatives under

- 42 -
his supervision pressured health-care practitioners to write

medically illegitimate prescriptions. For example, nurse Heather

Alfonso agreed with her sales representative "to do one to two

scripts per week." She later admitted that she "had come to rely

on th[e] extra money" and had "broke[n her] duty to patients."

So, too, Dr. Awerbuch was informed, at Simon's behest, that "the

average of his prescriptions was very low, within the one to 200

microgram range." As he recalled it, he then "started prescribing

[Subsys] to patients who didn't really even need to be on it just

to increase [his] numbers."

A reasonable jury could infer from the evidence that

illegitimate prescriptions were not an unintended consequence of

Simon's sales techniques but, rather, were a goal. With an eye on

revenue, Simon specifically sought to have practitioners prescribe

Subsys to patients who did not need it. It was Simon, for instance,

who endeavored to enforce a minimum-dosage agreement with Dr.

Somerville. As a result, Dr. Somerville entered into a Faustian

bargain with Insys that required, in Simon's own words, "every

refill" to be for at least 180 micrograms. In facilitating this

arrangement, Simon not only knew that prescriptions would

thereafter be untethered from patients' medical histories but also

solicited precisely that outcome. As one defense expert explained,

a doctor "decide[s]" whether the medication is warranted "at the

moment while [she's] seeing the patient," not "a week in advance."

- 43 -
Accordingly, a reasonable jury could find that Simon intended

doctors to prescribe Subsys outside the course of professional

practice because his quota arrangements required them to commit

both to specific numbers of Subsys prescriptions and to specific

dosages before they had a chance either to examine their patients

or to assess patients' medical needs. See Hughes, 895 F.2d at

1142; United States v. Mahar, 801 F.2d 1477, 1487 (6th Cir. 1986)

("[T]hat patients were regularly sold controlled substances

. . . selected by non-physician[s] . . . would further support a

finding that controlled substances were issued outside the usual

course of medical practice and for no legitimate medical

purpose.").

We conclude that the adverse finding against Simon as to

the CSA predicates was supported by the record and, therefore,

should have been allowed to stand. And as with Kapoor, see supra

Part III(A), this holding dictates that we reinstate the jury's

findings as to the honest-services predicates. It follows that we

must reverse the district court's partial grant of the Rule 29(c)

motion in favor of Simon as it pertains to both the CSA and honest-

services predicates.

D

As with Kapoor, Lee, and Simon, the district court set

aside the jury's finding that Rowan was guilty of conspiracy to

commit racketeering activities through a pattern of racketeering

- 44 -
acts that included the CSA and honest-services predicates. Here,

too, the evidence supports a finding that Rowan intended doctors

to write medically illegitimate prescriptions. Rowan worked hard

to develop quota agreements. For example, he was not shy about

communicating his prescription expectations to Dr. Couch.

According to one witness, Rowan gave Dr. Couch "a hard time about

the fact he hadn't been prescribing enough" and threatened to

"take[] away" the speaking programs (and, thus, the kickback

payments) if Dr. Couch "wasn't prescribing enough." Rowan's

aggressive enforcement of prescription quotas is evidence that he

knew that he was soliciting prescriptions that were not medically

necessary. See Hughes, 895 F.2d at 1135. Given that knowledge,

the kickbacks that Rowan was arranging constituted incentives for

prescribers to prescribe Subsys illicitly. See id.

The jury also heard evidence that Rowan had reason to

believe that successful performance of his job depended on

promoting illicit prescription-writing. His dealings with Dr.

Ruan illustrate this point. Rowan spoke directly with Dr. Ruan to

make clear that Insys would "pay [Dr. Ruan] as much as we possibly

and humanly can in exchange to write as much Subsys as [Dr. Ruan]

humanly can." In the same vein, the government introduced evidence

that Rowan understood that Dr. Ruan would find a way to prescribe

more as long as the dollars kept flowing. The facts on the ground

confirmed Rowan's understanding: Dr. Ruan ultimately wrote enough

- 45 -
Subsys prescriptions to boost Rowan into a position as Insys's top

sales representative anywhere in the country. Moreover, Rowan's

soaring sales figures exemplified the success of the kickback

scheme, and he was repeatedly mentioned in the 8:30 a.m. management

calls as a poster child for the proposition that "if you give these

[doctors] programs, they're going to write the drug for you."

There was more. After the DEA froze opioid shipments to

the pharmacy that principally filled Dr. Ruan's prescriptions,

Rowan learned that the pharmacy had access to an "unlimited supply"

of a competing opioid. He learned as well that the pharmacy wanted

not only a similar arrangement for Subsys in order to circumvent

"limits on [Schedule II drugs] in place by [the] current

wholesaler" but also "to ensure uninterrupted delivery to patients

of Dr. [Ruan]." Although by then Rowan either knew or was

willfully ignorant of Dr. Ruan's pill mill tendencies, he

nonetheless became involved in negotiating a direct-ship agreement

for Dr. Ruan. He (along with Kapoor and Babich) attended the

dinner meeting with Dr. Ruan at which the direct-ship agreement

was finalized.

Taking this proof in the light most favorable to the

verdict, a reasonable jury could conclude that Rowan intended Drs.

Couch and Ruan to prescribe Subsys outside the accepted course of

medical practice. See Volkman, 797 F.3d at 391. Because the

record supports a determination that Rowan agreed to commit CSA

- 46 -
violations, the jury's finding to that effect should not have been

vacated. See Kilmartin, 944 F.3d at 325. And as with Kapoor, see

supra Part III(A), this holding dictates that we reinstate the

jury's findings as to the honest-services predicates. It follows

that we must reverse the district court's partial grant of the

Rule 29(c) motion in favor of Rowan as it pertains to both the CSA

and honest-services predicates.

E

We add a coda. As said, the district court rested its

vacation of certain predicates on the so-called equipoise

principle, holding that because the proof "gives equal or nearly

equal circumstantial support to a theory of guilt and a theory of

innocence," Gurry, 427 F. Supp. 3d at 186 (quoting Burgos, 703

F.3d at 10), those predicate-act findings should be set aside. We

conclude that the equipoise principle was inapposite: the

evidence, viewed in the requisite light, was not so evenly

balanced. We summarize our reasoning.

We start with common ground: we agree with the district

court that the equipoise principle is entrenched in this circuit's

jurisprudence. When "the 'evidence viewed in the light most

favorable to the verdict gives equal or nearly equal circumstantial

support to a theory of guilt and a theory of innocence of the crime

charged,' this court must reverse the conviction." United States

v. Flores-Rivera, 56 F.3d 319, 323 (1st Cir. 1995) (quoting United

- 47 -
States v. Sanchez, 961 F.2d 1169, 1173 (5th Cir. 1992)). But "this

equal-evidence rule takes hold only after [the inquiring court]

ha[s] drawn all reasonable inferences in favor of the verdict."

Magraw v. Roden, 743 F.3d 1, 5 (1st Cir. 2014) (emphasis in

original). Here — as we already have explained — the evidence,

viewed in the light most favorable to the jury verdicts, clearly

favors a finding that the defendants conspired to distribute Subsys

even when the drug served no legitimate medical purpose. See supra

Parts III(A)-(D). "When the[se] pieces of evidence are layered,

with inferences taken in the government's favor, this is not a

case in equipoise." United States v. Ortiz, 447 F.3d 28, 34 (1st

Cir. 2006). Rather, it is a case in which a jury could find,

beyond a reasonable doubt, that the four affected defendants

(Kapoor, Lee, Simon, and Rowan) conspired with health-care

practitioners to write Subsys prescriptions outside the course of

accepted medical practice and without any medical justification.

We conclude, therefore, that the equipoise principle simply did

not apply. To the contrary, this is a case in which the jury

supportably found that the government had proved the CSA and

honest-services predicates beyond a reasonable doubt. It follows,

then, that the district court erred in vacating those findings.

IV

The jury found all five defendants guilty with respect

to the mail- and wire-fraud predicates. In their joint Rule 29(c)

- 48 -
motion, four of the defendants (Kapoor, Lee, Simon, and Rowan)

challenge the sufficiency of the evidence underlying these

findings.6 As the district court saw it, however, the evidence

supported those portions of the jury's findings and, thus, left

them intact. The four named defendants appeal that ruling.

The mail-fraud statute prohibits use of the mails in

connection with a "scheme or artifice to defraud." 18 U.S.C.

§ 1341. To establish the commission of this offense, the

government must show a scheme to defraud using false pretenses,

the defendant's knowing and willing participation in the scheme

with the intent to defraud, and the use of the mails in furtherance

of that scheme. See United States v. Soto, 799 F.3d 68, 92 (1st

Cir. 2015).

The mail- and wire-fraud offenses share common elements.

They differ only in that, to prove a violation of the wire-fraud

statute, the government must establish the use of wires (rather

than the use of the mails) in furtherance of the alleged scheme.

See United States v. Arif, 897 F.3d 1, 9 (1st Cir. 2018).

A

We start with Kapoor's claims of evidentiary

insufficiency with respect to the mail-fraud predicate. We can

make short shrift of them.

6 Gurry separately challenges these findings in his motion
for a new trial. See infra Part XII.

- 49 -
 1

Kapoor disputes the sufficiency of the evidence in

connection with the mail-fraud predicate only with respect to the

third element: whether the use of the mails furthered the alleged

scheme. To this end, he argues that the mailed bribes to

practitioners did not further the misrepresentations to insurers

regarding patients' conditions. We do not agree.

The "in furtherance" requirement is to be read broadly.

United States v. Hebshie, 549 F.3d 30, 36 (1st Cir. 2008). "[T]he

use of the mails need not be an essential element of the scheme."

Schmuck v. United States, 489 U.S. 705, 710 (1989). To prove this

element, the government need only show that the mailing was

"incident to an essential part of the scheme," id. at 711 (quoting

Pereira v. United States, 347 U.S. 1, 8 (1954)), or "a step in

[the] plot," id. (quoting Badders v. United States, 240 U.S. 391,

394 (1916)). We therefore parse the record to determine whether

the evidence shows some "connection or relationship" between the

mailing and the fraudulent scheme. Hebshie, 549 F.3d at 36

(quoting United States v. Pimental, 380 F.3d 575, 587 n.5 (1st

Cir. 2004)).

We find that the record shows a sufficient connection.

The mailed bribes generated prescriptions, which were fraudulently

processed through the IRC's authorization scheme. And in order to

facilitate the fraudulent processing of prior-authorization

- 50 -
requests, Insys offered business liaisons to whales (prolific

prescribers) who received bribes through the mail. Such bribe

recipients included Drs. Awerbuch, Chun, and Ahmad.

"The relevant question at all times is whether the

mailing is part of the execution of the scheme." Schmuck, 489

U.S. at 715. Because the scheme alleged here involved mailing

bribes for writing medically illegitimate Subsys prescriptions and

then obtaining insurance payments for those prescriptions, we

conclude that a jury reasonably could answer this question in the

affirmative.

Kapoor resists this conclusion. He argues that because

the IRC processed all the prescriptions that it received

(regardless of whether the prescribing doctor was bribed), the

scheme to defraud insurers did not "depend" on the bribes mailed

to doctors. In support, Kapoor notes that the bribed doctors were

only a "small fraction" of the doctors whose prescriptions were

fraudulently processed through the IRC.

This argument misses the mark: the government need not

show that the fraudulent scheme would have petered out without the

bribes. The mail-fraud statute does "not require[] a 'but-for'

link between a mailing and the fraudulent scheme." Hebshie, 549

F.3d at 36 (quoting Pimental, 380 F.3d at 587). It requires only

a connection between the two, and the record, read in the light

most favorable to the government, supports an inference that the

- 51 -
bribes, in increasing the volume of prescriptions, facilitated the

scheme. No more is exigible to uphold the jury's mail-fraud

finding.

2

As a fallback, Kapoor essays a constructive amendment

claim. In mounting this claim, he says that the mail-fraud scheme

described by the district court differed from that charged in the

indictment. The government, he adds, failed to adduce sufficient

proof of the latter scheme.

Specifically, he calls our attention to paragraph 31 of

the indictment, in which the grand jury alleged that "[h]ad the

insurers known that the defendants gave bribes and kickbacks to

the targeted practitioners, the insurers would not have authorized

payment for Subsys." Because the government did not show that

Kapoor "intended for the IRC to affirmatively misrepresent"

Insys's bribes to practitioners, his thesis runs, it failed to

substantiate the scheme it alleged. Although Kapoor gets high

marks for ingenuity, his claim fails the constructive-amendment

test.

"A constructive amendment 'occurs when the charging

terms of the indictment are altered, either literally or in effect,

by prosecution or court after the grand jury has last passed upon

them.'" United States v. Dunn, 758 F.2d 30, 35 (1st Cir. 1985)

(quoting Gaither v. United States, 413 F.2d 1061, 1071-72 (D.C.

- 52 -
Cir. 1969)). "[O]ur practice has been to look to statutory

elements in response to claims by defendants that 'the crime

charged' has been changed." United States v. Mubayyid, 658 F.3d

35, 51 (1st Cir. 2011). Therefore, "[s]o long as the statutory

violation remains the same, the jury can convict even if the facts

found are somewhat different than those charged." Id. (quoting

United States v. Twitty, 72 F.3d 228, 231 (1st Cir. 1995)); see

United States v. Dowdell, 595 F.3d 50, 68 (1st Cir. 2010).

In this case, the putative amendment occurred after

trial (when the court denied the defendants' joint Rule 29(c)

motion). Thus, Kapoor had no realistic opportunity to assert his

constructive amendment claim below. Consequently, this claim of

error engenders de novo review. See United States v. Rodriguez,

919 F.3d 629, 635 (1st Cir. 2019); United States v. Hernández, 490

F.3d 81, 83 (1st Cir. 2007).

We discern nothing resembling a constructive amendment

here. The crime charged was not altered because the language in

paragraph 31 did not implicate the statutory elements of the RICO

conspiracy. To prove a RICO conspiracy, "it is enough to prove

that a defendant agreed with one or more others that two predicate

offenses be committed." Aetna Cas. Sur. Co. v. P & B Autobody, 43

F.3d 1546, 1562 (1st Cir. 1994) (emphasis in original). The

predicate offenses themselves, however, are not elements required

to be proved. See id. Since Kapoor was not "convicted of a crime

- 53 -
other than that charged in the indictment," no constructive

amendment occurred.7 United States v. Day, 700 F.3d 713, 720 (4th

Cir. 2012) (quoting United States v. Randall, 171 F.3d 195, 203

(4th Cir. 1999)).

"[T]he rule against constructive amendments . . . is

focused not on particular theories of liability but on the offenses

charged in an indictment." Id. at 720 (emphasis in original).

Although the district court's order "eliminated a theory of

liability" alleged in paragraph 31, United States v. Celestin, 612

F.3d 14, 25 (1st Cir. 2010), "the statutory violation remain[ed]

the same," Mubayyid, 658 F.3d at 51 (quoting Twitty, 72 F.3d at

231). For that reason, the district court's order did not

constructively amend the indictment in any forbidden way. See

Celestin, 612 F.3d at 25; United States v. Miller, 471 U.S. 130,

145 (1985); cf. United States v. Mueffelman, 470 F.3d 33, 38 (1st

Cir. 2006) (distinguishing forbidden constructive amendment from

7 For the sake of completeness, we note that the challenged
language tracked one of the government's earlier theories of fraud
liability. The government originally alleged that each time Insys
submitted a Subsys authorization request on behalf of a bribed
doctor, the defendants committed fraud just by omitting
information about the bribe. The district court rejected this
theory, ruling from the bench that not "every prescription is bogus
just because there was a kickback behind it." Hence, the court
said, there was "[no] obligation to disclose [the kickback]." From
that point forward, the government elected to pursue only the
remaining mail-fraud allegations in the indictment.

- 54 -
one alleging "a scheme similar to but somewhat narrower in breadth

and malignity than that charged in the indictment").

B

Lee's challenge to the sufficiency of the evidence

underlying the mail- and wire-fraud predicates need not detain us.

We previously have limned the elements of those offenses. Lee

disputes only the second element of each offense — her knowing and

willing participation in the scheme with the intent to defraud —

and claims that the government failed to show that she had the

requisite knowledge of, or involvement in, the scheme. The record

belies her protestations.

Wire transmission of authorization requests and

approvals was essential to the operation of the IRC, and the

government's proof showed that Lee had both extensive interactions

with the IRC and a working knowledge of the approval process. A

few examples will suffice to hammer home the point:

• During the IRC's pilot phase, Gurrieri communicated

directly with Lee about Dr. Awerbuch's

prescriptions. Lee received a list of over one

hundred prescriptions that the IRC was attempting

to process on Dr. Awerbuch's behalf.

• Lee supervised the representative assigned to Dr.

Awerbuch, who helped process the authorization

requests; she also arranged the hiring of Dr.

- 55 -
 Awerbuch's niece as a business liaison to "fill out

the forms," "get the prescriptions pushed through,"

and "work[] with th[e] IRC."

• Lee was "very close" to an IRC authorization

specialist and lobbied for her promotion.

• Lee tried very hard to maximize the authorization

rate because she understood that Insys got paid

(and her own compensation increased) only if

insurers approved the drug.

We add, moreover, that the record supports an inference

that Lee pushed for prior authorizations with knowledge that the

information that the IRC relayed to insurers was inaccurate. She

worked closely with Drs. Madison and Awerbuch, who were two of the

most prolific prescribers of Subsys in the country. The record

likewise supports an inference that Lee knew that these prescribers

were writing medically illegitimate prescriptions. Because these

prescriptions would not get insurance approval organically, sales

representatives had to be "coach[ed]" on the misleading diagnosis

codes to be provided to insurers, and Lee was aware of this

coaching because she was copied in emails that discussed it.

On this record, a jury unquestionably could conclude

that Lee knew that the IRC was processing medically illegitimate

prescriptions by deliberately providing insurers misleading

information. The jury also could conclude that Lee agreed to

- 56 -
facilitate the fraudulent scheme by generating more prescriptions

for the IRC to process through mailed bribes and by streamlining

Dr. Madison's and Dr. Awerbuch's insurance-authorization

processes. The district court, therefore, did not err in denying

Lee's motion for judgment of acquittal on the mail- and wire-fraud

predicates.

C

Simon's evidentiary insufficiency claim with respect to

the mail- and wire-fraud predicates is easily dispatched. Like

Lee, Simon disclaims knowledge of or involvement in the insurance-

fraud scheme. See Soto, 799 F.3d at 92; Arif, 897 F.3d at 9. The

record, however, tells a different tale: it supports the jury's

findings as to both predicates.

The evidence shows that Simon understood, assisted, and

furthered the IRC's fraudulent activities. The sales

representatives who reported to him informed him whenever a doctor

granted the IRC permission to contact an insurer for an

authorization, and he was copied on emails reporting denials by

insurers. To convey this information to senior management, Simon

created "charts in progress" reports which documented the IRC's

efforts to obtain authorization for each Subsys prescription. In

addition, it was Simon who created the business liaison program.

A jury reasonably could envision these efforts as knowing

facilitation of the IRC's corrupt authorization processes.

- 57 -
 The record also supports an inference that Simon sought

to maximize the IRC's success despite knowing that the information

the authorization specialists supplied to insurers was misleading

and/or false. He was an occasional participant in the 8:30 a.m.

daily management calls, during which Kapoor and other senior

executives regularly discussed the IRC and strategies for

obtaining insurer authorizations. Such strategies included the

use of misleading words, phrases, and diagnosis codes, as well as

the "spiel." What is more, the government introduced evidence

that Simon saw these strategies in action when he visited the IRC

and listened to calls during which employees contacted insurers

and requested Subsys authorizations. From this evidence, a jury

reasonably could find that Simon had knowledge of the IRC's

fraudulent activities, yet chose to feed the IRC more prescriptions

by bribing doctors through the mail. It follows that the district

court did not err in denying Simon's motion for judgment of

acquittal on the mail- and wire-fraud predicates.

D

Rowan's claim of error with respect to the sufficiency

of the evidence underlying the mail- and wire-fraud predicates is

bootless. He, too, challenges only the intent element of the

jury's adverse findings on the mail- and wire-fraud predicates.8

8 The government asserts that Rowan failed to preserve this
claim of error and that, therefore, review is only "for clear and

- 58 -
See Soto, 799 F.3d at 92; Arif, 897 F.3d at 9. Our review

discloses, however, that the record is shot full of evidence that

Rowan monitored, facilitated, and participated in the IRC

authorization process. For instance, he personally arranged a

liaison for Dr. Chun and he instructed his subordinate (a sales

representative) to have the "[prior authorization] form filled out

every day with update to [Gurry]" and "to do whatever we could to

help and assist in getting that insurance pull-through." Various

of his subordinates reported to Rowan to confirm that doctors had

completed IRC opt-in forms and to alert him when doctors

encountered difficulty obtaining insurance approvals. A jury

reasonably could conclude that these were deliberate efforts to

support the corrupt IRC authorization procedure.

To cinch the matter, a jury reasonably could conclude

that Rowan undertook these efforts notwithstanding his knowledge

that the IRC was deliberately misleading insurers. Rowan had every

reason to believe that Dr. Ruan was prescribing Subsys

illegitimately, and a reasonable jury could infer that Rowan knew

that the IRC's efforts to get prior authorization for many of Dr.

Ruan's prescriptions were likewise illegitimate. His attendance

at an IRC training session corroborates such an inference. At

gross injustice." United States v. Foley, 783 F.3d 7, 12 (1st
Cir. 2015). We assume, without deciding, that the claim was
preserved and, therefore, engenders de novo review. See Kilmartin,
944 F.3d at 325.

- 59 -
that session, Rowan heard about the IRC's "history of cancer"

practice, including an explicit instruction to the authorization

specialists to include a reference to cancer even if "that's not

what we're seeing them for" because such a reference meant a "sure

approval" from insurers. So, too, Rowan learned that the IRC

maintained a list of drugs for authorization specialists to include

as tried-and-failed medications — a list that was to be used

liberally even if particular patients had not furnished any

information about prior medications. It thus seems nose-on-the-

face plain that, after this session, Rowan knew that the IRC was

defrauding insurers both because it cited bogus medical rationales

in support of prescriptions and because it provided apocryphal

lists of tried-and-failed medications. Yet, he continued to work

hand-in-glove with the IRC.

We do not gainsay that the jury was free to conclude, as

Rowan argues, that the IRC training session was innocuous. But

there were two sides to this particular story, and "it [was] within

the jury's purview to evaluate [these] competing factual

inferences." United States v. Ridolfi, 768 F.3d 57, 61 (1st Cir.

2014). Rowan's efforts to bribe doctors through the mail and to

push through Dr. Ruan's prescriptions despite Rowan's knowledge of

what was going on supports the conclusion that he knowingly and

willingly participated in the scheme with the intent to defraud

insurers. It follows that the district court did not err in

- 60 -
denying Rowan's motion for judgment of acquittal on the mail- and

wire-fraud predicates.

V

The district court admitted at trial testimony of nine

patients who had received Subsys prescriptions from doctors who

participated in the kickback scheme. All of the defendants

challenge the admission of their testimony as irrelevant and unduly

prejudicial. Some stage-setting is needed.

The defendants had anticipated the government's

presentation of evidence that patients had suffered harm from

taking Subsys. Prior to the trial, they moved to exclude such

evidence in its totality. The district court granted their motion

in part, leaving the government free to present testimony about

"the medical care that patients received from co-conspirator

physicians" and their "medical status." This evidence was allowed

for the purpose of showing "that prescribing was not medically

necessary or was in excess of what was medically necessary, or

that a patient's medical status was different from what was

represented to insurers in furtherance of claims for

reimbursement." The court also allowed the introduction of

evidence showing "that a patient became addicted to Subsys, the

medical consequences of that addiction, and whether and how

prescribing practices changed thereafter." Striving to strike a

balance, though, the court prohibited "evidence concerning the

- 61 -
social consequences to the patient of wrongful prescribing or

addiction, such as loss of employment, erosion of familial

relationships, and the like."

At the final pretrial conference, the defendants renewed

their objections to patient-harm evidence. The government argued

that it should be allowed to elicit testimony as to patients'

medical histories (e.g., whether a patient had cancer) "because

the IRC, which is run by Insys, is telling the pharmacy benefit

managers and other insurers that patients have cancer when the

patient doesn't have cancer." This testimony, the government said,

was intended "primarily to prove the fraud." So, too, the

government wanted to adduce testimony about the effects that Subsys

had on patients — that they "couldn't function[,] [t]hey slept all

day[,] [t]hey became addicted."

The district court essentially reaffirmed its earlier

ruling. The court noted that the charged conspiracy involved "not

just defrauding the insurance company," but also

"overprescri[ption] and increase [in] prescriptions." It

therefore concluded that the government should be "allowed to put

that evidence on to show that [the defendants] succeeded in their

objective, which is evidence of the fact that it was their

objective."

During trial, the defendants objected for a third time

to evidence of patient harm. In response, the district court

- 62 -
reiterated that it would not broadly preclude such testimony. When

the defendants renewed their objections yet again, the court

reiterated that testimony regarding addiction was "fair game."

All in all, nine patients testified about the

debilitating effects of addiction that they experienced while

ingesting Subsys. We offer a representative sampling of this

testimony:

• Cathy Avers testified that, as a result of taking

Subsys, she "bec[a]me an addict" such that "[n]o

matter how much [she] took, eventually it just

wasn't enough." She testified to side effects such

as "having a hard time functioning, standing up,

going to sleep. It was such an impact on [her]

being able to get up, out of bed, get dressed, and

do anything." She confirmed that the information

Insys had provided to her insurer — that she had a

current cancer diagnosis, was taking morphine and

hydromorphone, and was using a fentanyl patch — was

apocryphal.

• Paul Lara testified that, while taking Subsys, he

wound up "not finding [his] way home in a town

[he'd] lived in all [his] life" and having "to call

[his] wife to get directions home." He repeatedly

hallucinated and "thought [he] was going crazy."

- 63 -
 He could not follow what customers were saying to

him at work and once "literally three or four" of

his teeth "[fell] out right there [while] talking

to a customer." He also confirmed that Insys's

representations to his insurer that he had a

current diagnosis of cancer were spurious.

• Sara Dawes testified that, while taking Subsys, she

was "unable to function" and spent "most of [her]

time in bed." When she stopped taking Subsys, she

"was very, very, very sick and mentally couldn't

hold it together" to the point that she had "a

breakdown" and "drove off and left [her] kids on

Christmas." She also testified that, contrary to

what Insys had told her insurer, she never had

cancer, never had taken methadone, and did not have

difficulty ingesting generic fentanyl products.

• Betty Carrera testified that, while taking Subsys,

she began having such phantasmagoric hallucinations

that the police had to be called several times.

She could not function and spent her days sleeping.

She said that, when withdrawing from Subsys, she

had nightmares and hallucinations, and she would

"[wake] up at night screaming." She also

- 64 -
 contradicted Insys's representations to her insurer

and testified that she never had issues swallowing.

• Woodrow Chestang described "slobber . . . just

run[ning] down [his] mouth," watching the clock,

and craving more Subsys between doses. When he was

unable to get Subsys, he experienced delirium

tremens, nausea, and inability to eat or drink. He

sometimes curled "into a fetal position" and

realized that he was "burn[ing] up with fever." He

added that, contrary to the information that Insys

had given to his insurer, he neither had a history

of cancer nor had previously been prescribed

generic opioid containing fentanyl.

• Scott Byrd testified that Subsys was "life-

changing" because "[i]t put [him] into an addiction

state that [he] almost couldn't come out of."

Because he used more than the quantity that his

doctor had prescribed, he ran out early and

experienced major withdrawal. He also swore that

the signature on the opt-in form purportedly

authorizing Insys to contact his insurer on his

behalf was not his and, in fact, misspelled his

name.

- 65 -
• Kendra Skalnican testified that she developed an

addiction after starting on Subsys, and, as a

result, began to take more of the medication than

had been prescribed. When she ran out, she

experienced severe withdrawal, sweating, vomiting,

diarrhea, and pain all over her body. She told the

jury that Subsys "made [her] addicted" and "[she]

slept a lot of [her] life away." She also testified

— contrary to information provided by Insys to her

insurer — that she never had issue swallowing pills

and never had tried other fentanyl products.

• Michelle DiLisio (previously Kamzyuk) testified

that, while taking Subsys, she was lethargic,

fatigued, dizzy, and felt "out of it." She reported

that she suffered from severe withdrawal symptoms

after she stopped taking the medication. And she

made clear that the information that Insys had

furnished to her insurer was false: she never had

"any cancer ever" and, specifically, she never had

ovarian cancer (indeed, she had undergone ovary-

removal surgery years before Subsys had been

prescribed for her).

• Alicia Hinesley testified that Subsys made her

"extremely sleepy" and led to difficulty in

- 66 -
 thinking. Sometimes she would sit or sleep all

day. Belying Insys's statements that she was

experiencing breakthrough cancer pain, she flatly

denied that she ever had cancer.

After the jury verdicts had been returned, the

defendants moved for a new trial. They argued that the admission

of the patient-harm testimony constituted reversible error. The

district court thought not: it concluded that "[t]he patient

testimony at trial conformed to the Court's motion in limine ruling

in which it allowed only limited use of patient testimony and

carved out most inflammatory aspects, such as the social

consequences of addiction." Gurry, 427 F. Supp. 3d at 203. The

testimony, the court said, "was relevant to show the medical care

that patients received from co-conspirator prescribers, to

demonstrate that certain prescriptions were not medically

necessary or were excessive, and to support claims that a patient's

medical status was different from what was represented to

insurers." Id.

It hardly bears repeating that a trial court enjoys

considerable discretion with respect to its evidentiary rulings.

See United States v. Zaccaria, 240 F.3d 75, 78 (1st Cir. 2001).

We review the rulings that the defendants challenge here only for

abuse of that discretion. See Iacobucci v. Boulter, 193 F.3d 14,

20 (1st Cir. 1999).

- 67 -
 A

We start with the defendants' claim that the challenged

testimony was irrelevant. The standard for relevancy is not

exacting. See United States v. Rivera Calderón, 578 F.3d 78, 97

(1st Cir. 2009). The patient-harm testimony is relevant if it has

the "tendency to make the existence of any fact that is of

consequence to the determination of the action more probable or

less probable than it would be without the evidence." Id. (quoting

Fed. R. Evid. 401).

The district court appropriately found that the patient-

harm testimony was relevant. To prove the CSA predicates, the

government had to show that the defendants agreed that a health-

care practitioner would prescribe Subsys outside the usual course

of medical practice and without any legitimate medical purpose.

See Volkman, 797 F.3d at 391. The evidence of the patients'

altered behavior, addiction, and withdrawal symptoms was plainly

relevant to show that the doctors' treatment was outside the course

of professional practice. This is particularly true where, as

here, each doctor continued to prescribe Subsys to his or her

patient despite knowing of the patient's addiction. Taking a

practical view of what had transpired, the jury reasonably could

have regarded the patient-harm testimony as powerful proof both

that the coconspirator doctors prescribed Subsys in the absence of

any medical necessity and that they failed to minimize the risk of

- 68 -
adverse effects when setting dosages. In fine, the patient-harm

testimony was relevant to show that the doctors contravened their

professional obligations. See United States v. Singh, 54 F.3d

1182, 1187 (4th Cir. 1995). And we think it obvious that evidence

that the doctors prescribed Subsys outside the usual course of

professional practice while receiving kickbacks constitutes

evidence relevant to show that the defendants had entered into an

agreement to bring about exactly that result. See United States

v. Rivera-Santiago, 872 F.2d 1073, 1079 (1st Cir. 1989) (explaining

that "[t]he actions, as well as the words of the [coconspirators],

are evidence of the existence and scope of a conspiracy"). On

this record, evidence about the exploitation of addiction was

relevant to show that all of the coconspirators, including the

defendants, viewed addiction less as a societal problem and more

as a pathway to predatory profits.

On the same basis, we dismiss Gurry's contention that he

"had no connection to prescribers' medical decision-making." The

patient-harm testimony is relevant as to Gurry because it helped

establish the scope of the conspiracy. See id. "The fact that

[Gurry] participated in one retail link of the distribution chain,

knowing that it extended beyond his individual role, was

sufficient" to establish relevancy as to him. Id.

The defendants erect a straw man. They submit that the

patient-harm testimony "said nothing about what [they], who had no

- 69 -
contact with any of these patients and no knowledge of [what] they

were affected by Subsys, specifically intended." But as we already

have discussed, a core component of the conspiracy to distribute

Subsys was influencing doctors to "push the dose." The most

logical reason for the defendants' unremitting efforts to increase

dosages was their knowledge that patients on higher doses would

refill their Subsys prescriptions while patients on lower doses

would not. The patient-harm testimony showed vividly just how the

"effective dose" messaging furthered the scheme.

At the risk of carting coal to Newcastle, we add that

the patient-harm testimony also helped to explain how the

defendants could expect doctors to fulfill their commitments to

Insys representatives, that is, to meet quotas obligating them to

prescribe inordinately high amounts of Subsys. The patients

trusted the doctors; the doctors provided a limited explanation of

the drug to the patients; and by the time the patients realized

they were addicted, they were powerless to refrain from seeking

more and more Subsys.

To say more about relevancy would be to paint the lily.

Because the patient-harm testimony tended to show the ins and outs

of the defendants' scheme, it was within the district court's

discretion to deem this evidence relevant. See United States v.

Hale, 857 F.3d 158, 171 (4th Cir. 2017).

- 70 -
 B

The defendants next argue that the patient-harm

testimony, even if relevant, was unfairly prejudicial. In

examining this claim, we begin with evidentiary bedrock. A

district court may exclude relevant evidence if an objecting party

can show that "its probative value is substantially outweighed by

a danger of . . . unfair prejudice." Fed. R. Evid. 403. But under

Rule 403, one size does not fit all. Thus, we afford the district

court "considerable latitude in steadying the balance which Rule

403 demands." United States v. Cadden, 965 F.3d 1, 22 (1st Cir.

2020) (quoting United States v. Rodriguez-Estrada, 877 F.2d 153,

156 (1st Cir. 1989).

Through serial rulings, the court below exercised care

in weighing the considerations affecting the Rule 403 balance.

From the beginning, the court precluded evidence concerning the

social consequences of addiction and — in its own words — took

pains to "carve[] out most inflammatory aspects" of the testimony.

Gurry, 427 F. Supp. 3d at 203. Even so, the defendants complain

that the court did not carve out a sufficiently wide exclusionary

swath.

The defendants' argument relies primarily on our

decision in Kilmartin, 944 F.3d at 315. Kilmartin, though, is a

horse of a different hue. There, the government prosecuted — for

fraud-related crimes — a defendant who advertised cyanide to

- 71 -
suicidal individuals, collected their money, and sent them Epsom

salts instead. See id. at 323-24. At trial, the government

offered as "anecdotal background evidence" testimony from victims

(other than those named in the charged counts) who had tried to

purchase cyanide from the defendant. Id. at 333. This testimony

"went into excruciating detail about the . . . victims' personal

lives, medical issues, histories of depression, earlier suicide

attempts, suicidal motivations, and the like." Id. at 335. We

later described the testimony as "copious," "emotionally charged,"

and as having "virtually no probative value." Id. at 337. Because

the inordinate potential for prejudice "substantially outweighed"

the dubious probative value of the anecdotal evidence, we held

that the district court abused its discretion in admitting that

evidence. Id. at 338.

This case is a world apart from Kilmartin. The patient-

harm testimony here was relatively brief and squarely probative,

established that the patients became addicted to Subsys and

suffered withdrawal symptoms, shed a bright light on the

prescribing habits of the coconspirator physicians, tied the

"effective dose" messaging into the scheme, and catalogued (in

checklist fashion) many of the ways in which the IRC misrepresented

patient information. Perhaps most importantly, the patient-harm

testimony explained how the charged conspiracy was able to function

and how it generated product demand. And, finally, the testimony

- 72 -
was concise: no testifying patient was permitted to dwell unduly

on the harm that he or she suffered. Viewed in this perspective,

the patient-harm testimony was less like the challenged testimony

in Kilmartin and more like the victim testimony in Cadden, 965

F.3d at 22 — the admission of which we approved because it was

relatively brief and the trial court precluded more graphic

details.

To be sure, the patient-harm testimony packed a punch.

Nevertheless, the issue is not prejudice simpliciter but, rather,

whether particular evidence crosses the line into the forbidden

realm of unfair prejudice.9 See United States v. Pitrone, 115 F.3d

1, 8 (1st Cir. 1997) ("[I]t is only unfair prejudice against which

the law protects." (emphasis in original)). The fact that

addiction is ugly does not bar the government from offering

evidence about it when — as in this case — the defendants' scheme

has made addiction relevant and probative. See, e.g., United

States v. Morales-Aldahondo, 524 F.3d 115, 120 (1st Cir. 2008)

(holding that, although admitted images of child pornography

"undoubtedly had an emotional impact on jurors," district court

"properly balanced the competing concerns of Rule 403" when

evidence was probative and court "limit[ed] the number of images

9We have observed before that "all evidence is meant to be
prejudicial." Rodriguez-Estrada, 877 F.2d at 156. If it was not
intended to influence the jury in one way or another, it is
unlikely that any party would seek to introduce it.

- 73 -
presented"). In the last analysis, a "court is not required to

scrub the trial clean of all evidence that may have an emotional

impact, where the evidence is part of the Government's narrative."

Id. at 120 (internal quotation omitted).

We are aware that the defendants offered to stipulate

that none of the testifying patients had cancer. But such a

stipulation was not an acceptable proxy for the patients'

testimony. The scope of the proffered stipulation was much

narrower than the scope of the testimony, and the government was

entitled to show (for example) other misrepresentations made by

the IRC. We consistently have rejected parties' attempts to insist

that district courts accept stipulations that are not commensurate

substitutes for live proof, see, e.g., Cadden, 965 F.3d at 22, and

we do so here.

To sum up, we discern no abuse of discretion in the

court's construction of the Rule 403 balance. The patient-harm

testimony bore on the government's theory of the case in salient

ways, and the court took prudent steps to soften the emotional

impact of the testimony. We have stated before that "[o]nly rarely

— and in extraordinarily compelling circumstances — will we, from

the vista of a cold appellate record, reverse a district court's

on-the-spot judgment concerning the relative weighing of probative

value and unfair effect." United States v. Mehanna, 735 F.3d 32,

59 (1st Cir. 2013) (quoting Freeman v. Package Mach. Co., 865 F.2d

- 74 -
1331, 1340 (1st Cir. 1988)). This is not so rare an instance, and

the district court acted within the encincture of its discretion

under Rule 403 in allowing the challenged testimony.

C

The defendants' challenge to the admissibility of the

patient-harm testimony incorporates one last point. They contend

that the patient-harm testimony was cumulative of other proof.

They note, for example, that Dr. Awerbuch and Nurse Alfonso

testified that their Subsys prescriptions were not medically

necessary and that Gurrieri and other IRC staffers testified that

they lied to insurers about patients' conditions. Since the

defendants did not raise this objection below, plain error review

obtains. See Taylor, 54 F.3d at 972-73; United States v. Nivica,

887 F.2d 1110, 1116 (1st Cir. 1989).

"The plain error hurdle is high," United States v.

Hunnewell, 891 F.2d 955, 956 (1st Cir. 1989), and a purported error

must (among other things) be "clear or obvious" in order to be

"plain." United States v. Duarte, 246 F.3d 56, 60 (1st Cir. 2001).

Cumulativeness is almost always a matter of degree, and the

defendants' claim of cumulativeness — if it suggests an error at

all — at most suggests an error that is neither clear nor obvious.

See United States v. Sepulveda, 15 F.3d 1161, 1185 (1st Cir. 1993)

("We have routinely found cumulative evidence impotent when

- 75 -
accidentally uncorked."). Plain error is, therefore, plainly

absent.

VI

Gurry — whom the jury acquitted with respect to the CSA

and honest-services predicates — contends that the evidence

admitted with respect to those predicates unfairly influenced the

jury's findings against him on the mail- and wire-fraud

predicates.10

This is, for all intents and purposes, a claim of

prejudicial spillover. As relevant here, prejudicial spillover

occurs when the evidence admitted to prove a charge as to which

the defendant was acquitted "was so extensive, inflammatory, and

prejudicial that it necessarily spilled over into the jury's

consideration of [his] guilt on other charges." Mubayyid, 658

F.3d at 72.

To determine whether an unacceptable threat of

prejudicial spillover materialized, we must evaluate whether the

record evinces "a 'serious risk' that the joinder of offenses

compromised a specific trial right or 'prevent[ed] the jury from

10Although the other four defendants advanced similar
contentions in their briefs, those contentions have been rendered
moot by our vacatur of the district court's partial grant of their
Rule 29(c) motions. See supra Part III (A)-(D); see also Mubayyid,
658 F.3d at 73 (holding claim of prejudicial spillover without
merit after appellate court reinstated the previously vacated
conviction).

- 76 -
making a reliable judgment about guilt or innocence.'" Id.

(quoting United States v. Houle, 237 F.3d 71, 75-76 (1st Cir.

2001)). The devoir of persuasion rests with the defendant to show

"prejudice so pervasive that a miscarriage of justice looms."

United States v. Trainor, 477 F.3d 24, 36 (1st Cir. 2007) (quoting

United States v. Levy-Cordero, 67 F.3d 1002, 1008 (1st Cir. 1995)).

In the court below, Gurry argued that the government's

"accusation" that he and the other defendants intended to coax

doctors to prescribe Subsys illegitimately, coupled with the

patient-harm testimony, tainted the jury's findings against him on

other matters. The district court rejected this argument and

refused to order a new trial on this ground. See Gurry, 427

F. Supp. 3d at 196-97. It found that the patient-harm testimony

was properly admitted as to all the defendants and all the charged

predicates and observed that its jury instructions had been custom

tailored to guard against prejudicial spillover. See id.

We review the district court's denial of a new trial

based on allegations of prejudicial spillover for abuse of

discretion. See United States v. Neal, 36 F.3d 1190, 1205 (1st

Cir. 1994). We discern none.

At the outset, it bears mentioning that Gurry's argument

repastinates much of the same ground covered in our discussion of

the admissibility of the patient-harm testimony. See supra Part

V. He was charged as a coconspirator and, thus, almost all of the

- 77 -
evidence properly admitted against other coconspirators was

relevant to and independently admissible against him. See United

States v. O'Bryant, 998 F.2d 21, 26 (1st Cir. 1993). And because

the patient-harm testimony was independently admissible against

Gurry, he hardly can be heard to complain about an untoward

spillover effect. See id. Simply put, the government's case

against Gurry would have comprised essentially the same evidence

even if the government had not seen fit to charge him with the

acquitted predicates.

We add that Gurry's argument that patient-harm testimony

likely "incited [the jury's] ire" is severely wounded by his

acquittal with respect to the CSA and honest-services predicates.

That the jury's findings distinguished among defendants and

differentiated among proposed predicates is strong evidence that

no spillover prejudice occurred. See United States v. Williams,

809 F.2d 75, 88 (1st Cir. 1986); cf. United States v. Natanel, 938

F.2d 302, 308 (1st Cir. 1991) ("The introduction of evidence

against other defendants cannot realistically be viewed as having

jeopardized [the defendant's] chances on [one count] when the jury

proved willing to treat the case against [him] on its own merits

by acquitting him on the other counts."). Here, moreover, the

jury differentiated not only between counts but among defendants

— and that selectivity is "strong evidence" that the jury was not

blinded by raw emotion but, rather, properly compartmentalized and

- 78 -
applied the law to the facts. United States v. Bailey, 405 F.3d

102, 112 (1st Cir. 2005); see United States v. Dworken, 855 F.2d

12, 29 (1st Cir. 1988) (giving credence to "jury's ability to

segregate the evidence and carefully weigh against which defendant

it was applicable" (quoting United States v. Richman, 600 F.2d

286, 299-300 (1st Cir. 1979)).

Much of the credit for the jury's discernment must go to

the district court. The court excluded the most inflammatory

evidence about the effects of Subsys and prudently instructed the

jury both to treat each defendant individually and to weigh

separately the evidence as to each defendant. As a general rule,

"instructing the jury to consider each charged offense, and any

evidence relating to it, separately as to each defendant"

constitutes an "adequate measure[] to guard against spillover

prejudice." United States v. Casas, 425 F.3d 23, 50 (1st Cir.

2005); see, e.g., United States v. Figueroa, 976 F.2d 1446, 1454

(1st Cir. 1992) (holding that "district court minimized any danger

from prejudicial spillover through its repeated instructions that

the jury was to give separate consideration to each charge against

each defendant"). Gurry has not pointed to anything that would

take this case out of the general rule.

Little more need be said. The jury acquitted Gurry with

respect to the CSA and honest-services predicates while at the

same time finding the four other defendants guilty of those

- 79 -
charges. This result constituted "an uncommonly convincing 'ex

post validation' of the jury instructions." Figueroa, 976 F.2d at

1454. In the circumstances of this case, Gurry's claim of

prejudicial spillover lacks force, and the district court acted

well within the ambit of its discretion in refusing to grant him

a new trial on that ground.

VII

During pretrial proceedings, Lee moved for a severance

of the charges against her. See Fed. R. Crim. P. 14(a). The

district court denied her motion. Lee assigns error.

A

In support of severance, Lee argued below that the

government charged two distinct conspiracies: one to bribe doctors

who would prescribe Subsys indiscriminately and another to defraud

insurers to pay for those prescriptions. From this starting point,

she asserted that a joint trial would prejudice her because she

was not personally involved in the second of these conspiracies.

The district court denied her motion, concluding that Lee had

failed to make a sufficient showing of potential prejudice.

Specifically, the court found that Lee had "fail[ed] to identify

any evidence or argument that would not be admissible against her

in a separate trial" and that her allegations of prejudice were

wholly conclusory.

- 80 -
 On appeal, Lee traverses the same terrain. Her case

should have been severed, she submits, because she was a stranger

to the IRC portion of the wrongdoing. We review the district

court's denial of her motion for abuse of discretion. See United

States v. Azor, 881 F.3d 1, 12 (1st Cir. 2017).

When — as in this case — an indictment charges a criminal

conspiracy among multiple defendants, the government enjoys the

benefit of a rebuttable presumption that a joint trial is

appropriate. See United States v. Soto-Beníquez, 356 F.3d 1, 29

(1st Cir. 2003) (explaining that "the general rule is that those

indicted together are tried together"); see also Zafiro v. United

States, 506 U.S. 534, 537 (1993) (noting Supreme Court's

"repeated[] . . . approv[al] of joint trials" for coconspirators).

And in cases where joinder is proper, "[w]e must affirm the

district court's denial of a motion to sever unless the defendant

makes a strong and convincing showing of prejudice." United States

v. Richardson, 515 F.3d 74, 81 (1st Cir. 2008) (internal citations

omitted); see Azor, 881 F.3d at 12.

Here, we uphold the district court's refusal to sever

for two reasons. First, the record contains substantial evidence

showing Lee's involvement with the IRC (for instance, evidence

showing that Lee sought to maximize the number of opt-in forms to

be transmitted to the IRC and evidence showing that she supervised

some of Insys's IRC authorization specialists). Second, because

- 81 -
the government charged and proved a single conspiracy and because

Lee was charged and convicted as a coconspirator, virtually all of

the evidence properly admitted against the other defendants

(including evidence showing that the IRC was an integral part of

the single conspiracy) was also admissible against Lee. See

O'Bryant, 998 F.2d at 26; see also Richardson, 515 F.3d at 82

("[T]his Court has repeatedly refused to overrule a denial of

severance if substantially the same evidence would have been

admitted in separate trials.").

Straining to show that she did not belong in the case,

Lee identifies 34 witnesses who — she speculates — would not have

been called to testify had she been tried alone. But the unadorned

fact that additional witnesses will be called in a joint trial is

not a cognizable basis for severance. The right to a severance

necessarily entails a showing of prejudice, and Lee offers no

explanation as to why the testimony of these witnesses (who, in

her brief's words, "had nothing relevant or incriminating to say

about Lee") prejudiced her in any way.

B

Lee plucks out of thin air a new assault on the denial

of her motion for a severance. She contends, for the first time

on appeal, that joinder was improper under Federal Rule of Criminal

Procedure 8(b). This misjoinder, she says, independently demanded

severance. Although we normally review the propriety of joinder

- 82 -
de novo, see Azor, 881 F.3d at 12, Lee's unpreserved contention

engenders — at most — plain error review,11 see United States v.

Greenleaf, 692 F.2d 182, 187 n.4 (1st Cir. 1982); see also United

States v. Ackerly, 981 F.3d 70, 74 (1st Cir. 2020).

Whatever the standard of review, a claim of misjoinder

"requires reversal only if the misjoinder results in actual

prejudice." United States v. Lane, 474 U.S. 438, 449 (1986); see

United States v. Bruck, 152 F.3d 40, 44 (1st Cir. 1998). The

movant must show that her joinder had a "substantial and injurious

effect or influence in determining the jury's verdict." Bruck,

152 F.3d at 44 (quoting Kotteakos v. United States, 328 U.S. 750,

776 (1946)). Lee's feeble effort to show that her joinder was

prejudicial falls far short.

Lee starts with the uncontroversial proposition that

prejudicial joinder may entitle a defendant to a severance. See

Natanel, 938 F.2d at 306. Lee has very little to say, though,

about why her joinder was prejudicial. Her only argument seems to

rest on her self-serving conclusion that "the Government misled

the Court into believing that Lee 'dealt extensively' with the

11 In all likelihood, the claim of misjoinder — which was
available before trial but not raised by any pretrial motion — was
waived. See Fed. R. Crim. P. 12(b)(3)(B)(iv). Here, however, the
government has not suggested waiver, and we assume for argument's
sake that the misjoinder claim is subject to appellate review
(albeit only for plain error).

- 83 -
IRC." This conclusion, in turn, circles back to her protest that

she had "no criminal association" with the IRC side of the venture.

As we already have explained, this protest is at odds

with the record. In certain cases, evidence at trial may "serv[e]

as an ex post assurance that joinder was a step founded on a

reasonable, good faith basis in fact." Id. at 307. So it is here,

and we hold unhesitatingly that Lee's joinder was appropriate.

VIII

Lee's employment history was unusual for a

pharmaceutical executive: her most relevant prior work experience

seems to have been as an exotic dancer at a Chicago-area strip

club. Before trial, the government sought leave to introduce

evidence about Lee's past work and her unorthodox professional

behavior with Dr. Madison (the notorious pill mill operator). The

district court ruled, over Lee's objection, that the proffered

evidence was "not admissible to prove the Defendants' character,

but such evidence may be admissible for other purposes, including

to establish the nature of the relationships between the co-

conspirators, duress, or relevant corporate culture." Along the

same lines, the court ruled that Lee's employment history "[was]

not admissible to prove [her] character," but "may be intrinsic to

aspects of the charged offense" and, to that extent, might be

admissible. In the end, the court temporized, stating that the

- 84 -
proffered evidence "will be admitted if it is otherwise admissible"

under the federal rules of evidence.

At trial, the government offered evidence of Lee's

employment history. Burlakoff testified that he first met Lee

while she was working at a strip club, that he invited her to apply

for a sales manager position at Insys, and that he sent her résumé

to Babich. In order to bolster its theory that Kapoor knew of

Lee's lack of credentials in either management or pharmaceutical

sales, the government discussed at sidebar its intention to

introduce an email that suggested that Lee had run an escort

service. The court refused to admit the email but allowed

testimony about whether portions of Lee's résumé were incomplete.

In front of the jury, the prosecutor asked Burlakoff

whether someone had provided Insys with information that Lee was

"running an escort service." The court sustained an objection and

struck the question. But that was not the end of the matter. When

the parties returned to sidebar, the district court ruled that the

email contained "relevant information," but directed the

prosecutor "to keep the salacious aspect to an absolute minimum."

Acceding to a defense request for an instruction that the contents

of the email were not being admitted for the truth of the matter

asserted, the court told the jurors that they would "hear testimony

. . . that the company got some information about Ms. Lee that

suggested that she might not be qualified for the job." Because

- 85 -
"the letter that [Insys] received is anonymized," the court

cautioned:

The letter does not — and I cannot emphasize
this strongly enough — does not come in for
the truth of the matter asserted. . . . [T]he
person that wrote . . . the letter . . . is
certainly not here. They're not testifying.
There may be issues of bias. We don't have
any way to know if what they're saying is true
or not. You're to consider this information
only to the effect that [it] had on the company
and what they did in response to receiving
this information.

Burlakoff then testified that Babich had received an

email about Lee from an "ex-fiancé . . . who had a bone to pick

with her." According to Burlakoff, the email questioned why Insys

would hire someone with Lee's background and listed several

websites. He checked the websites and found topless photos of

Lee. After he informed Babich, Babich consulted with Kapoor, who

"s[aw] no issue with it" but asked that "those pictures come down

immediately." Burlakoff relayed Kapoor's wishes to Lee, who took

the topless photos down.

Separately, two sales representatives testified that

they went to a Chicago nightclub with Lee and Dr. Madison after a

speaker event. One testified that Lee "was sitting on [Dr.

Madison's] lap, kind of bouncing around, and he had his hand sort

of inappropriately all over her on her chest." The other sales

representative testified that he observed "[v]ery inappropriate

contact" between Lee and Dr. Madison, such as Dr. Madison placing

- 86 -
"[h]is hands . . . all over her, her front and her pants, in her

shirt" and "heavily kissing" Lee.

Lee objected to all of this testimony and moved for a

mistrial, which the district court denied. She argues that the

court erred in admitting this evidence because it constituted

"salacious propensity evidence" that should have been excluded

under Federal Rule of Evidence 404(b). She suggests that because

"the jury heard questions that gave an inference that if Lee worked

as an escort or operated an escort service for financial gain in

the past and had topless photos on the internet, it is more likely

that she committed the charged offense for financial gain." In

the alternative, she suggests that the evidence should have been

excluded under Rule 403.

We review the district court's admission of the

challenged evidence for abuse of discretion. See Iacobucci, 193

F.3d at 20. We start with Lee's Rule 404(b) challenge. Rule

404(b)'s propensity bar "excludes only extrinsic evidence —

'evidence of other crimes, wrongs, or acts' — whose probative value

exclusively depends upon a forbidden inference of criminal

propensity." United States v. Manning, 79 F.3d 212, 218 (1st Cir.

1996) (quoting United States v. Hadfield, 918 F.2d 987, 994 n.5

(1st Cir. 1990)). Evidence intrinsic to the crime charged is not

precluded under Rule 404(b). See id.

- 87 -
 Following these guideposts, we conclude that Rule

404(b)'s proscription of propensity evidence is inapposite here.

The probative value of the challenged evidence does not depend

exclusively on a forbidden inference of propensity but, rather, is

intrinsic to the crime charged. Burlakoff's testimony about Lee's

qualifications (or lack of them) tends to show that neither Kapoor

nor Lee could reasonably think that Lee was hired as a sales

manager due to either her executive excellence or her marketing

skill set. Instead, the evidence suggests that the defendants'

scheme to bribe doctors into prescribing Subsys indiscriminately

offered doctors both money (through the speaker programs) and

sexual favors.

So, too, the sales representatives' testimony about

Lee's physical interactions with Dr. Madison has independent

probative value: that testimony confirms Lee's willingness to

influence doctors' prescription habits through sexual

interactions. As Burlakoff made clear, the doctors "prescribe[d]

strictly based on their relationship with the sales manager."

Here, the challenged evidence was relevant because it explained

the background and development of the relationship between two of

the coconspirators (Lee and Dr. Madison) inasmuch as it showed

Lee's tactics for getting Dr. Madison "to keep his writing up" and

because it revealed some of the unprofessional motivations

underlying Dr. Madison's prescription habits. See United States

- 88 -
v. Escobar-de-Jesus, 187 F.3d 148, 169 (1st Cir. 1999). As the

district court noted, the evidence is "illustrative of [Lee's]

relationship with [Dr. Madison] and how she's interacting with

him" to motivate the doctor to prescribe more and more Subsys.

We also reject Lee's contention that the jury

necessarily inferred that she was likely to have committed a crime

from evidence that she ran an escort service and that topless

photos of her floated on the internet. The record contains no

indication of the evidence being offered or used for that purpose.

Perhaps more importantly, the district court carefully limited the

ways in which the jury could put that information to use. The

email came in only to show "the effect that [it] had on the company

and what [the company] did in response to receiving this

information." We long have held that courts may presume that

jurors will follow the judge's instructions, United States v.

Spencer, 873 F.3d 1, 16 (1st Cir. 2017), and Lee has provided no

reason for us to deviate from that norm.

Nor did the district court abuse its discretion in

concluding that the probative value of the challenged evidence was

not substantially outweighed by its unfairly prejudicial effects.

We afford district courts appreciable discretion in striking the

balance that Rule 403 demands. See Mehanna, 735 F.3d at 59;

Freeman, 865 F.2d at 1340. The evidence challenged here was

probative of one of the ways in which Lee and her superiors

- 89 -
attempted to influence prescribers, and it was also probative of

the defendants' intent to downplay traditional sales strategies

that focus on patients' needs. Here, moreover, the district court

was sensitive to the potential for prejudice, cautioning the

government to "tone it down" and to avoid the specific details of

Lee's encounter with Dr. Madison. In the same spirit, the court

made certain that the information derived from the email was

presented to the jury as suspect: it told the jurors that there

was no way to find out if the information in the email was true

and instructed them not to take it for the truth of the matter.12

We conclude, therefore, that the district court held the Rule 403

balance steady and true, and that Lee's claim of error is

impuissant.

Lee's appeal from the denial of her motion for a mistrial

is equally unavailing. "Declaring a mistrial is a last resort,

only to be implemented if the taint is ineradicable, that is, only

if the trial judge believes that the jury's exposure to the

evidence is likely to prove beyond realistic hope of repair."

Sepulveda, 15 F.3d at 1184. We review the district court's denial

12Lee argues in passing that the "[a]dmission" of the email
"would offend" the Confrontation Clause. See U.S. Const. amend.
VI. This argument collapses of its own weight: the email was
never admitted into evidence and, in any event, the court told the
jury that it could not consider the contents of the email for the
truth of the matter asserted. Consequently, the right to
confrontation was not implicated. See United States v. Cabrera-
Rivera, 583 F.3d 26, 33 (1st Cir. 2009).

- 90 -
of a mistrial for abuse of discretion. See United States v.

Chisholm, 940 F.3d 119, 126 (1st Cir. 2019).

In the case at hand, the district court supportably found

that a mistrial was not required. Its clear limiting instructions

and prompt striking of extraneous matter, combined with the

presumption that juries follow the trial court's instructions,

leads inexorably to a conclusion that the district court did not

abuse its discretion.

IX

Lee requested a jury instruction on supervisory

condonation. She asked that the jury be instructed that while

"Burlikoff [sic] and Babich's knowledge or condoning of activities

does not by itself constitute a defense or an excuse," evidence of

their "actions or omissions, or evidence of deficiencies in the

manner in which they implemented or enforced [Insys's] policies

and procedures, may be considered . . . to the extent that such

evidence bears on the issue of whether or not defendant Lee formed

the required intent to commit the crimes with which [s]he is

charged." The district court did not give the requested

instruction. Lee preserved her objection and now assigns error.

Our review of the district court's eschewal of this

proposed instruction is for abuse of discretion. See United States

v. De La Cruz, 514 F.3d 121, 139 (1st Cir. 2008). A district court

is, of course, under no obligation to honor a party's word choices

- 91 -
or to parrot proposed language when delivering jury instructions.

See United States v. DeStefano, 59 F.3d 1, 2-3 (1st Cir. 1995).

As a result, we will not second-guess the trial court's rejection

of a proposed instruction unless the proposed instruction is itself

substantively correct, was not covered (at least in substance) in

the charge as given, and touched upon a salient point (such that

the refusal so to instruct seriously undercut the proponent's

ability to mount a particular claim or defense and caused

substantial prejudice). See id.

Lee's proposed instruction fails under the second and

third prongs of this formulation. The district court's charge, as

rendered, contained a good-faith instruction, which informed the

jury that "[t]he 'good faith' of a Defendant is a complete defense

to the charge in the indictment because good faith on the part of

the Defendant is, simply, inconsistent with both knowingly and

willfully agreeing to become a member of the alleged conspiracy

and specifically intending that a member of the alleged conspiracy

would commit criminal conduct." The court added that "[a]n honest

mistake in judgment or an honest error in management does not rise

to the level of criminal conduct." So, the court said, "[i]f the

evidence in the case leaves . . . a reasonable doubt as to whether

a Defendant acted with criminal intent or in good faith," the jury

should "find the Defendant not guilty."

- 92 -
 This instruction fully permitted Lee to present her

supervisory condonation defense and, thus, forestalled any

cognizable claim of prejudice. Lee demurs, maintaining that the

court's good-faith instruction did not accommodate her two-pronged

argument that she "was lawfully following the instructions of her

employer" and that "Insys condoned her conduct."

Lee's claim of error depends on an unrealistically

cramped reading of the court's good-faith instruction. Under this

instruction, Lee was free to argue that she acted in good faith

because she subjectively believed that her conduct was lawful and

that she based that belief on her employer's orders, its

condonation of her conduct, or both. Because of her employer's

guidance and approval, she might say, her mistake was an honest

one. The court's good-faith instruction focused the jury on Lee's

"actual, subjective beliefs," so the "charge basically did what

[Lee] wanted it to do." United States v. Denson, 689 F.3d 21, 26

(1st Cir. 2012). Because the instruction actually given

accommodated both prongs of Lee's argument, the district court's

refusal to use Lee's proposed language was well within its

discretion.

X

Rowan assigns error to the district court's denial of

his mid-trial motion to compel the disclosure of allegedly

exculpatory information. See Brady v. Maryland, 373 U.S. 83

- 93 -
(1963). This claim of error harks back to a prosecutor's comment

to Rowan's counsel, allegedly made during a break in Gurrieri's

testimony, supposedly mentioning that the government had discussed

a recording used as an IRC training tool with Gurrieri. Asserting

that this recording was a critical piece of evidence in the

government's case against him, Rowan moved to compel the government

to produce all communications between Gurrieri and the government

concerning the recording.

In response, the government vouchsafed that it "has

consistently met and exceeded its ethical and legal discovery

obligations in this case." There were no further communications

that were subject to production, the government said, because it

had "fully complied with all of its obligations," including

disclosure of all of its interview reports and rough notes. The

government added that "[i]f [it] was aware of any exculpatory or

Brady information in any form, it would have disclosed that

information in a report, in agent notes, verbally, via email, or

in some other form."

The district court denied Rowan's motion "[b]ased on the

government[']s representations" and its own "understanding of the

issues in the case as a result of a lengthy trial." The court

took the opportunity, though, to remind the government "that its

[Brady] obligations continue through sentencing." Rowan moved for

reconsideration, but to no avail.

- 94 -
 We review the district court's denial of a motion to

compel discovery for abuse of discretion. See United States v.

Flete-Garcia, 925 F.3d 17, 33 (1st Cir.), cert. denied, 140 S. Ct.

388 (2019). This standard of review is not one-dimensional. See

Akebia Therapeutics, Inc. v. Azar, 976 F.3d 86, 92 (1st Cir. 2020);

United States v. Lewis, 517 F.3d 20, 24 (1st Cir. 2008). Within

it, we review for clear error the district court's factual finding

that no further document subject to production existed. See United

States v. Padilla-Galarza, 990 F.3d 60, 79-80 (1st Cir. 2021).

Under Brady, the government is obligated "to disclose

evidence in its possession that is favorable to the accused and

material to guilt or punishment." United States v. Prochilo, 629

F.3d 264, 268 (1st Cir. 2011) (citing, inter alia, Brady, 373 U.S.

at 87). Where, as here, a claim of Brady error is advanced, the

defendant bears the burden of showing "a likelihood of prejudice

stemming from the government's nondisclosure." Flete-Garcia, 925

F.3d at 33. To make such a showing, he must "articulate with some

specificity what evidence he hopes to find in the requested

materials, why he thinks the materials contain this evidence, and,

finally, why this evidence would be both favorable to him and

material." Id. (quoting Prochilo, 629 F.3d at 269). And in

determining whether the evidence sought is material, "[t]he

question is not whether the defendant would more likely than not

have received a different verdict with the evidence, but whether

- 95 -
in its absence he received a fair trial, understood as a trial

resulting in a verdict worthy of confidence." United States v.

Josleyn, 206 F.3d 144, 152 (1st Cir. 2000) (quoting Strickler v.

Greene, 527 U.S. 263 (1999)).

Rowan has utterly failed to make the requisite showing.

The most prominent fly in the ointment is that he has failed to

establish that the evidence he seeks actually exists. Although

Rowan conclusorily asserts that "such communications must have

occurred," all three prosecutors (including the prosecutor whom

the defense identified as having mentioned the government's

purported discussion with Gurrieri) signed the pleading in which

the government insisted that it had "withheld nothing." Given the

unequivocal nature of the government's representations and the

experience gleaned by the court in presiding over this case

(including protracted pretrial proceedings, discovery disputes,

and a lengthy trial), we decry no clear error in the court's

determination that the claimed evidence did not exist. A

defendant's naked assertion that a particular communication "must

have occurred," no matter how vociferously expressed, is

insufficient to undermine a reasoned judicial determination that

no such communication actually exists. See United States v. Duval,

496 F.3d 64, 75 (1st Cir. 2007); cf. Padilla-Galarza, 990 F.3d at

80 (holding, in Jencks Act context, that "the government cannot be

- 96 -
expected to produce that which has never existed"). We therefore

reject Rowan's claim of Brady error.13

XI

Following the adverse jury verdicts, Simon — represented

by successor counsel — moved for a new trial. See Fed. R. Crim.

P. 33. Among the grounds asserted in support of this motion, he

averred that his trial counsel had been handicapped by a conflict

of interest. Specifically, he averred that his trial counsel,

Steven Tyrrell, was conflicted because the law firm in which

Tyrrell was a principal — Weil Gotshal & Manges LLP (Weil) — was

representing Insys in a bankruptcy restructuring at the same time

that Tyrrell was representing Simon in this case. The district

court disagreed and refused to order a new trial. Simon appeals

that ruling.

Simon's conflict-of-interest claim has its roots in an

internal investigation that Insys conducted some three years prior

13 We add that Rowan's explanations for why the alleged
evidence would be exculpatory and material are unconvincing: they
are woven with nothing more than wispy threads of speculation and
surmise. Mere conjecture that certain communications "might
contain exculpatory evidence" without "any supporting evidence or
arguments to indicate this was, in fact, the case," is inadequate
to ground a claimed Brady violation. United States v. Brandon, 17
F.3d 409, 456 (1st Cir. 1994); see Flete-Garcia, 925 F.3d at 34
(concluding that "district court's refusal to compel production of
requested information is not an abuse of discretion" when "theory
of materiality is based entirely on conjecture"); Prochilo, 629
F.3d at 269 (explaining that defendant's Brady showing "cannot
consist of mere speculation").

- 97 -
to the start of Tyrrell's representation of Simon. In December of

2013, Insys received a subpoena from the Department of Justice.

Insys immediately retained Skadden, Arps, Slate, Meagher & Flom

LLP (Skadden) to serve as its outside investigations and compliance

counsel. Skadden conducted a thorough investigation, interviewed

numerous Insys employees, reviewed a wide range of company

practices, and offered advice to Insys's board of directors.

Years passed and — in 2017 — Simon retained Tyrrell to

represent him in the case at hand. The following year, Insys

turned to Weil in connection with anticipated chapter 11

proceedings. When Tyrrell became aware of his firm's potential

representation of Insys, he discussed the matter with Simon.

Tyrrell informed Simon that — should his representation of Simon

continue — he would be "walled off" from the Weil team handling

Insys's bankruptcy reorganization. Simon assented to this

arrangement.

In due course, Weil signed an engagement letter with

Insys, which explicitly permitted Tyrrell to act adversely to Insys

in connection with his representation of Simon. Weil quickly

instituted screens to prohibit the two teams from reviewing,

discussing, or sharing information.

We fast-forward to June of 2019. After the jury returned

its verdicts, Simon queried Tyrrell about Weil "representing Insys

in its bankruptcy case." Tyrrell reminded Simon of their earlier

- 98 -
conversation, described the "wall" that was in place, and assured

Simon that "there is no sharing of information or interaction."

Simon renewed his queries the following month, calling Tyrrell's

attention specifically to the internal investigation that Skadden

had overseen. Tyrrell responded that the internal investigation

had ended before the criminal case began and reiterated that Weil's

representation of Insys in the bankruptcy proceedings was

unrelated to the criminal case.

Unassuaged, Simon retained fresh counsel and moved for

a new trial on the ground that Tyrrell had been laboring under a

conflict of interest. He alleged that Weil's representation of

Insys had inhibited Tyrrell and prevented him from seeking to

obtain the findings of Insys's internal investigation into the

marketing and sale of Subsys. Although Insys had consistently

asserted that those materials were shielded by the attorney-client

privilege, Simon argued that a different (conflict-free) attorney

could have pierced the privilege. The government opposed the

motion. The district court denied relief, concluding that Simon's

proffered alternative strategy was not plausible. See Gurry, 427

F. Supp. 3d at 217.

We review the district court's factual findings in

connection with the conflict-of-interest claim for clear error but

afford de novo review to the court's ultimate conclusion. See

Reyes-Vejerano v. United States, 276 F.3d 94, 97 (1st Cir. 2002).

- 99 -
Under the Sixth Amendment, a defendant has a right to conflict-

free counsel. See United States v. Ponzo, 853 F.3d 558, 574 (1st

Cir. 2017); U.S. Const. amend. VI. That right, though, does not

protect a defendant from an attorney's "mere theoretical division

of loyalties." Id. at 575 (quoting Mickens v. Taylor, 535 U.S.

162, 171 (2002)). To prevail on a conflict-of-interest claim, a

defendant must show that "'a conflict of interest actually

affected' the lawyer's 'performance.'" Id. (quoting Mickens, 535

U.S. at 171). Such a showing requires a demonstration "that (1)

the lawyer could have pursued a plausible alternative defense

strategy or tactic and (2) the alternative strategy or tactic was

inherently in conflict with or not undertaken due to the attorney's

other interests or loyalties." Id. (quoting United States v.

Colón-Torres, 382 F.3d 76, 88 (1st Cir. 2004)).

We conclude — as did the court below, see Gurry, 427 F.

Supp. 3d at 217-19 — that no actual conflict of interest existed

because piercing attorney-client privilege to lay bare the

findings of Skadden's internal investigation was not a plausible

defense strategy. According to Simon, this proposed strategy would

have offered "material from Skadden's internal investigation to

substantiate a good-faith defense." This must be so, he muses,

because Skadden "apparently . . . did not advise Insys to shut

down the ISP, to close the IRC, or to fire or discipline Mr.

Simon." Building on this rickety foundation, Simon argues that

- 100 -
the seeming absence of such advice must mean that Skadden concluded

that Insys's operations were beyond reproach. So, Simon's thesis

runs, Skadden's internal investigatory materials "would have

revealed the evidentiary basis — facts, documents, witness

testimony — underlying Skadden's findings and advice, which [his]

defense counsel could have marshaled to use at trial."

This proposed strategy is both substantively and

strategically bankrupt. First and foremost, Simon's allegation

that the findings reached during the investigation must be

favorable to him is anchored on abject speculation. Simon

consistently has acknowledged that he has never been "privy to the

details of [outside counsel's] findings and advice." Skadden's

findings are, he confesses, "unknown to [him]." Knowledge is

essential to the making of value judgments, and saying that

something is "unknown" is tantamount to an admission that its

favorability cannot be ascertained.

Despite this void, Simon self-servingly surmises that

the materials generated during the investigation must bolster his

defense because Skadden interviewed him and — subsequent to that

interview and the completion of Skadden's investigation — "nobody

ever counseled [him] to modify his own practices or imposed any

discipline or punishment on him for wrongdoing." Simon also

suggests that since the IRC did not shut down, an inference is

warranted that Skadden did not advise Insys to cease operations.

- 101 -
Piling inference upon inference, he then suggests that Skadden

must have refrained from giving such advice because it found

Insys's business practices aboveboard. In other words, Simon asks

us to assume that the materials would be exculpatory simply because

the internal investigation neither "resulted in [any] adverse

employment action against [him]" nor brought about any changes in

day-to-day IRC operations. Arriving at that assumption, though,

elevates hope over reason. Given the complicity of so many company

hierarchs in the scheme, the unknown time span covered by the

internal investigation, and the lack of congruity between that

time span and the life of the conspiracy, Insys's failure to either

take adverse action against Simon or to modify the IRC's modus

operandi may well have other more compelling explanations.

In all events, the district court had ample reason to

infer that the findings of the internal investigation were likely

detrimental to Simon's defense. The government and the defendants

engaged in considerable pretrial skirmishing as to whether the

government could elicit testimony from an Insys compliance officer

who coordinated the investigation. See id. at 218. Her testimony

would have focused on her conclusion that the IRC was engaging in

insurance fraud, id. — a conclusion that Simon would just as soon

have the jury not hear. So, too, other evidence in the record

makes it likely that the evidence Simon seeks would not have been

exculpatory. As we already have pointed out, see supra Part

- 102 -
III(C), the record includes substantial evidence of Simon's

knowledge of illegitimate Subsys prescriptions and his attempts to

increase their volume, his knowledge of the IRC's fraudulent

representations to insurers, and the like. Viewing the record in

its entirety, Simon's notion that Insys permitted him to continue

working because his work was legitimate seems far less plausible

than the notion that he was kept in place because his work

furthered the ongoing criminal scheme. Cf. Gurry, 427 F. Supp.

3d. at 220 ("The evidence at trial indicated that although Insys

hired compliance personnel and a general counsel after receiving

the subpoena in December 2013, these individuals were largely

viewed as obstacles to the success of the sales force and the

company."). Considering the improbability of Simon's assumption,

his afterthought defense strategy cannot be said to possess even

a patina of plausibility and, thus, cannot be considered a viable

strategy. See United States v. Cardona-Vicenty, 842 F.3d 766, 773

(1st Cir. 2016); see also Ponzo, 853 F.3d at 577 ("[M]ere

speculation does not suffice to show a Sixth Amendment

infraction.")

To complete the picture, we note that the proffered

strategy was not only implausible but also entailed significant

strategic risks. It is hornbook law that forgoing "a strategy

that could inculpate the defendant does not constitute an actual

- 103 -
conflict." Ponzo, 853 F.3d at 576. That is precisely the sort of

strategy that Simon now embraces. We explain briefly.

It is luminously clear that piercing the attorney-client

privilege would have been fraught with peril. Success in that

endeavor would have opened the floodgates for damaging testimony

from Insys's compliance officer, in-house counsel, and others

involved in the internal investigation. The potentially dire

consequences of such a strategy explain why the other defendants

— even though most of them would have had at least as good a chance

as Simon to benefit from the allegedly exculpatory evidence —

chose, through independent and highly skilled counsel, not to buck

Insys's attorney-client privilege. Instead, they banded together

and asked the district court, in their own words, to "preclude the

government from eliciting at trial any testimony regarding

privileged communications between Insys or its Board of Directors

. . . and the company's in-house or outside counsel." To put it

bluntly, they all went to the mat to block the government from

introducing the findings of the internal investigation. The fact

that no other defendant sought to pierce Insys's attorney-client

privilege is a telling indication that this strategy was neither

likely to be helpful to the defendants nor free from significant

risks of further inculpating them. Cf. Brien v. United States,

695 F.2d 10, 16 (1st Cir. 1982) (giving weight to "the fact that

none of [defendant's] other co-defendants, even though they had

- 104 -
independent counsel," sought the particular evidence). This is

far removed from the kind of alternative defense strategy that can

undergird a Sixth Amendment claim. See Ponzo, 853 F.3d at 576.

If more were needed — and we do not think that it is —

Simon also has failed to establish a meaningful relationship

between the findings of the internal investigation and his

proffered good-faith defense. Such a defense asks the jury to

determine what the defendant's "actual, subjective beliefs" may

have been. Denson, 689 F.3d at 26. Because Simon has never been

privy to the findings of the investigation, those findings could

not have informed his subjective beliefs.14 See United States v.

Zayyad, 741 F.3d 452, 461 (4th Cir. 2014); United States v.

Dynalectric Co., 859 F.2d 1559, 1574 n.19 (11th Cir. 1988).

14At trial, attorney Tyrrell did press a condonation defense
on Simon's behalf: he argued that "when [Simon] started, the
actions that he took were in line with the strategies that were
mapped out by the company's leaders and communicated to the entire
sales force, and there's no evidence that [he] knew or understood
that any aspect of those strategies was illegal." Because this
defense substantially covered the defense that Simon now says was
impaired and because the findings of the internal investigation
remain largely shrouded in mystery, it is apparent to us that Simon
has failed to articulate any benefit that his proposed strategy
plausibly might have achieved. Thus, there is no basis to conclude
that Tyrrell's choice to refrain from trying to pierce the
attorney-client privilege "actually affected the adequacy of
[Simon's] representation." Familia-Consoro v. United States, 160
F.3d 761, 764 (1st Cir. 1998) (quoting Cuyler v. Sullivan, 446
U.S. 335, 349 (1980)); cf. Brien, 695 F.2d at 15 (finding no actual
conflict of interest when "the tactics [defendant] suggests that
his attorney could have pursued appear to be merely hypothetical
choices that in reality could not have benefited [him]").

- 105 -
 That ends this aspect of the matter. To prevail on a

Sixth Amendment conflict-of-interest claim, "the conflict must be

real." Brien, 695 F.2d at 15. The conflict of interest that Simon

ascribes to his trial counsel is purely theoretical and, thus,

does not come close to supporting a claim of constitutional

dimension. See id. We are not in the business of granting

"undeserved windfall[s]" to defendants who merely point to any

course of action not taken by their attorney and cry foul.

Cardona-Vicenty, 842 F.3d at 774 (internal quotation omitted). It

is exactly that kind of windfall that Simon is seeking. His quest

goes begging because the district court was on solid ground in

denying his conflict-of-interest claim.

XII

Gurry contends that the district court blundered in

denying his motion for a new trial. He argues that the evidence

against him was "remarkably thin" and that the government's case

turned on the "uncorroborated" word of one cooperating witness —

Gurrieri.

Where, as here, a new trial motion is based upon the

weight of the evidence, a district court should not grant a new

trial "unless it is quite clear that the jury has reached a

seriously erroneous result." United States v. Rothrock, 806 F.2d

318, 322 (1st Cir. 1986) (quoting Borras v. Sea-Land Serv., Inc.,

586 F.2d 881, 887 (1st Cir. 1978)). In a nutshell, such a remedy

- 106 -
should be granted sparingly and only when the evidence

preponderates heavily against the jury's verdict or a miscarriage

of justice otherwise looms. See United States v. Merlino, 592

F.3d 22, 32 (1st Cir. 2010). We review a district court's denial

of such a motion solely for abuse of discretion. See United States

v. Ruiz, 105 F.3d 1492, 1501 (1st Cir. 1997).

The record comfortably supports Gurry's convictions on

the mail- and wire-fraud predicates. He advised employees to "ride

the gray line" with insurers and use the "spiel" to obscure the

patients' lack of a cancer diagnosis. In addition, he led

strategic planning for the IRC, attended the daily 8:30 a.m.

management calls as the IRC's "mouthpiece," listened to accounts

of the IRC's deceptive practices during those daily calls, directly

supervised Gurrieri (who instructed employees to report false

medical rationales for prescriptions and bogus lists of tried-and-

failed medications), approved spurious patient-specific reports of

difficulty swallowing, and enforced IRC authorization quotas.

This evidence supports the jury's conclusion that Gurry

deliberately participated in Insys's defrauding of insurers — a

scheme that involved bribing doctors (through the mails) to

generate prescriptions and misrepresenting (through the wires)

patients' medical histories and needs.

In resisting this conclusion, Gurry focuses single-

mindedly on Gurrieri's credibility. Without that testimony, he

- 107 -
suggests, the evidence against him would be weakened to a point

where the adverse jury verdict would topple.

Gurry's single-minded focus means that he has left

himself with a steep uphill climb: "the district court must

generally defer to a jury's credibility assessments" when

evaluating a motion for a new trial. Merlino, 592 F.3d at 32. On

appeal, we may not "second guess the [district] judge's refusal of

a new trial and the jury's willingness to accept the essentials of

[a government witness's] account of the events." United States v.

Pitocchelli, 830 F.2d 401, 403 (1st Cir. 1987) (affirming denial

of new trial when trial court elected to leave "to the jury the

ultimate decision as to whether it believed" disputed testimony).

Even if we set to one side the steepness of this slope,

Gurry has not shown that the jury's verdict was seriously flawed.

He offers nothing that is sufficient to discredit the inference

that he purposefully bought into the IRC's tactics. We briefly

inspect his main contention — that Gurrieri was not to be believed

— and explain why we find that contention wanting.

First, Gurry emphasizes Gurrieri's decision to testify

as a cooperating witness. He rates this as a reason to disbelieve

her testimony. But Gurry's rating system is out of kilter: he

fails to take into account the jury's prerogatives. The district

court appropriately instructed the jury that Gurrieri was

cooperating with the government and that her testimony therefore

- 108 -
ought to be considered "with particular care and caution." Given

this cautionary instruction, it was within the jury's province to

choose whether to believe or disbelieve Gurrieri's testimony. See

United States v. Appolon, 695 F.3d 44, 55 (1st Cir. 2012).

Next, Gurry declares that Gurrieri's testimony was

"uncorroborated." This declaration is specious.15 Other witnesses

and documents substantiated the inference that Gurry both knew of

and supported the IRC's corrupt tactics. For instance, a sales

manager testified that she toured the IRC with Gurry, and that

they listened as an employee used deceptive tactics to obtain

Subsys authorization from an insurer over the telephone. Then,

too, Babich testified that Gurry was part of the "primary group"

of senior executives who participated in the daily 8:30 a.m.

management calls, that Gurry was in charge of communicating to

that group "any highlights both positive and negative that they're

seeing in the IRC," that those highlights were informed by Gurry's

communications with Gurrieri, and that those daily calls discussed

the IRC's deceptive tactics (including the promiscuous use of

15We do not mean to imply that corroboration was a sine que
non to a conviction. It was not. See United States v. Martínez-
Medina, 279 F.3d 105, 115 (1st Cir. 2002) (holding that the
"uncorroborated testimony of a government informant is . . . enough
to convict" because "the law of this circuit . . . leaves in the
hands of the jury decisions about credibility of witnesses 'so
long as the testimony is not incredible or insubstantial on its
face'" (quoting United States v. Andujar, 49 F.3d 16, 21 (1st Cir.
1995))).

- 109 -
"dysphagia" references and the "spiel"). Babich also testified

that the dysphagia gambit was discussed by Gurrieri in front of a

group that included Gurry.

There was also documentary corroboration. More than one

piece of this corroboration originated with Gurry, who (for

example) sent himself an email reminder about employee training on

the difference between breakthrough cancer pain and breakthrough

pain. Similarly, he sent a detailed email to sales managers

enumerating strategies that were crafted to prompt unwarranted

insurer approvals. Additionally, he was copied on several

inculpatory emails, including emails about "the issue that arose

with Dr. Chun's pharmacy" and the direct shipments of Subsys to

Dr. Ruan's pharmacy for the purpose of ensuring "uninterrupted

delivery to patients." Corroboration may come in various forms

and shapes, and we find significant corroboration for Gurrieri's

testimony in this record.

Gurry presses his attack on Gurrieri's credibility in

other ways as well. For instance, he makes a frontal assault on

Gurrieri's testimony that he maintained an office near hers at the

IRC. In this regard, he notes that two witnesses testified

otherwise. That may be so, but it is up to the jury to decide who

to believe — and that is especially true when witnesses offer

inconsistent versions of the facts. See United States v. Patel,

370 F.3d 108, 112 (1st Cir. 2004). And to tie a bow on it, even

- 110 -
if we assume, for argument's sake, that Gurrieri's recollection

was inaccurate on this one point, the jury was still entitled to

credit other aspects of her testimony that were unfavorable to

Gurry. Because a witness's testimony is not a monolith, it was

within the jury's purview to "credit some parts of [Gurrieri's]

testimony and disregard other potentially contradictory portions."

United States v. Sabean, 885 F.3d 27, 37 (1st Cir. 2018) (quoting

United States v. Alicea, 205 F.3d 480, 483 (1st Cir. 2000)).

Gurry also posits that Gurrieri's testimony that she was

following Gurry's directions is contradicted by her "eagerness to

take credit for the IRC's success" in the moment. Gurry's argument

rests on a kernel of truth: Gurrieri did claim credit for the

"creat[ion] [of] the IRC." But nothing about that claim undercuts

her testimony that she consulted with Gurry on key decisions, that

he sanctioned the IRC's deceptive tactics, and that he directed

her to undertake specific acts of fraud (including the submission

of authorization requests containing fictious medication lists).

Gurry has one last shot in his sling. He complains that

only Gurrieri characterized him as dishonest. Other witnesses, he

says, described him as disciplined, quiet, polite, respectful,

supportive, and stiff. These traits, he tells us, are inconsistent

with the government's attempted depiction of him as a racketeer.

- 111 -
 We are not so sanguine. A quiet, polite, and respectful

demeanor is simply not a warranty of good behavior.16 Choir boys

and curmudgeons alike can commit conspicuously corrupt crimes. It

was the jury's task to weigh the salience, if any, of Gurry's

positive traits against the specific evidence of his less-than-

savory actions. Given the deference that we afford juries in

regard to credibility calls, we cannot say that the jury in this

case either misweighed the evidence or reached a seriously

erroneous result.

This door is shut. The jury was entitled to credit

Gurrieri's testimony, and the district court did not err in denying

Gurry's motion for a new trial.

XIII

The defendants sought a new trial on the ground that

prosecutorial misconduct infected the government's closing

argument. The district court denied their motion, and all of them

— Gurry directly, and the rest by adoption — now appeal.

We set the stage. During the rebuttal portion of his

closing argument, the prosecutor sought to establish that the

defendants specifically intended physicians to prescribe Subsys

This verity has been part and parcel of the human experience
16

from time immemorial. Over four centuries ago, the Bard of Avon
famously wrote "To beguile the time, look like the time — bear
welcome in your eye, your hand, your tongue. Look like the
innocent flower, but be the serpent under't." William Shakespeare,
Macbeth, act 1, sc. 5 (circa 1606).

- 112 -
outside the usual course of professional practice. He told the

jury:

People intend [the] reasonably foreseeable
consequences of their actions. It is as
though, if I took a gun and fired it into the
audience, which I'm not going to do, I don't
intend to shoot any particular individual, but
I know somebody's going to get hit. And when
the defendants arm these doctors with all
these bribes and all these incentives, they
were creating a loaded gun.

None of the defendants interposed a contemporaneous objection.

In the same phase of his closing argument, the prosecutor

referred to evidence that defendants had hired a compliance

officer. He noted that the defendants "had no interest in

compliance prior to that" and that the compliance officer "told

you, when she was hired in April of 2014, that she was being

frustrated in her efforts." The prosecutor then stated, "regarding

Mr. Gurry, who was running the IRC, who is responsible for the

IRC, that's his job. As a corporate officer, he bears the

responsibility." This time, the defendants objected.

The prosecutor also stated:

After nine weeks of trial, there should be no
doubt, in anybody's mind here, that there was
a massive insurance fraud here, happened
every day, day in and day out. And there was
a massive bribery scheme involved. I think
the defendants concede as much, but what they
want to sit here and say to you is that these
men and women who ran this company, who were
the managers, had no idea what was going on.
Sort of like that scene from Casablanca, I'm

- 113 -
 shocked to find out there's illegal gambling
in this place.

Along this same line, the prosecutor argued that the defendants

"incentivized these doctors" to prescribe Subsys frequently and at

high doses, "and they can't sit here and tell you, now, that they

didn't intend for that to happen." The defendants did not

contemporaneously object to either of these comments.

At the conclusion of the government's rebuttal, the

district court gave a curative instruction in response to the

objection relating to Gurry's corporate-officer status. It told

the jury that "the corporation, Insys, is not on trial here. The

individuals are on trial and your verdict must turn on your

assessment of the culpability of them as individuals and not as

corporate officers." Neither side objected to this instruction.

Several days later — but before jury deliberations began

— the defendants sought additional curative instructions or in the

alternative, a mistrial. In support, they identified several

instances of alleged prosecutorial misconduct:

• They alluded to the comment about Gurry's

corporate-officer status and argued that they could

not be held criminally liable merely for the

wrongdoing of subordinates.

• They calumnized the prosecutor's "loaded gun"

analogy and asserted that the statement that

- 114 -
 "[p]eople intend [the] reasonably foreseeable

consequences of their actions" deviated materially

from the specific-intent element of a RICO

conspiracy charge.

• Observing that none of them had elected to testify,

the defendants raised the specter that the

prosecutor's rebuttal argument had "made veiled

reference to the fact that Defendants had pressed

various factual arguments at trial without taking

the witness stand."

The district court responded with an offer to give additional

curative instructions. The court then circulated draft

instructions; defense counsel proposed revisions; and the court

accepted all but one of the proposed revisions.17 The court read

its prepared charge to the jury and followed up by reading the

supplemental instructions. In pertinent part, the supplemental

instructions admonished:

At least some of the defendants were at
relevant times corporate officers or managers
with responsibility for their departments
and/or subordinates. The fact that a
defendant had an executive or managerial
position at Insys is not alone enough to
convict the defendant of the RICO conspiracy
charge in the indictment.

17 Rowan requested that the court tell the jury that the
challenged comment "was not a correct statement of the law." The
court declined that request.

- 115 -
A healthcare company executive's or manager's
failure to correct or prevent misconduct at
the company does not alone constitute a
violation of the RICO statute. In other
words, even if you think that a defendant
should have known about certain conduct,
should have done more to correct or prevent
such conduct or should be responsible for the
conduct of company employees, you cannot
convict the defendant on this basis.

As I already told you bribes and kickbacks
alone are insufficient to convict in this
case. For you to find an agreement regarding
the racketeering act of illegal distribution
of a controlled substance, honest services
mail fraud or honest services wire fraud, you
must find that defendants agreed to and
specifically intended for healthcare
practitioners to write Subsys prescriptions
outside of the usual course of professional
practice and without legitimate medical
purpose. Under the law, knowledge of
foreseeable consequences without more is not
enough to establish that someone specifically
intended certain conduct. Rather, the
government must prove that the defendant acted
with a bad purpose or with the object of
committing a prohibited act, here, for the
controlled substance and honest services
predicates, having healthcare practitioners
prescribe Subsys outside of the usual course
of professional practice and without
legitimate medical purpose.

. . .

Finally, you should not interpret anything
that was said in this case as a comment on the
fact that defendants chose not to testify. As
I've already instructed you, defendants have
an absolute constitutional right not to
testify. And you cannot draw any inference
from the fact that they exercised their
rights. You cannot consider or discuss
defendants' choices not to testify during your
deliberations.

- 116 -
After giving these supplemental instructions, the district court

asked if any party wanted to be heard at sidebar. Receiving no

affirmative response, the court instructed the jury to start its

deliberations.

Following the adverse jury verdicts, the defendants

renewed their prosecutorial misconduct claims in their new-trial

motions. Those motions were uniformly denied. See Gurry, 427 F.

Supp. 3d at 201.

Although we review the district court's order denying a

new trial for abuse of discretion, see Merlino, 592 F.3d at 32

n.5, we evaluate de novo their claims of error involving the

propriety of the government's closing argument, see United States

v. Kuljko, 1 F.4th 87, 94 (1st Cir. 2021); United States v.

Carpenter, 736 F.3d 619, 626 (1st Cir. 2013). We start with the

claims of error arising out of the government's comments about

Gurry's corporate-officer status and the alleged allusions to the

defendants' failure to testify. Those claims of error share a

common characteristic: the defendants do not assert that the

challenged comments were so toxic that no cautionary instructions

could have saved the day but, rather, assert only that the

cautionary instructions given by the district court were

insufficient.

- 117 -
 The architecture of the defendants' assertions shapes

the contours of our inquiry. This architecture places waiver

principles front and center. We have explained that "when the

'subject matter [is] unmistakably on the table, and the defense's

silence is reasonably understood only as signifying agreement that

there was nothing objectionable,' the issue is waived on appeal."

Soto, 799 F.3d at 96 (quoting United States v. Christi, 682 F.3d

138, 142 (1st Cir. 2012)). One application of this rule occurs

when "the district court informed the [parties] exactly how it was

planning to instruct the jury" and "sought their feedback," with

the result that a party's counsel "affirmatively stated there was

no objection" or "remained silent." Soto, 799 F.3d at 96. In

that circumstance, an appellate court is free to consider the

instructions approved by that party. See id. Any claim that the

instructions are inadequate is deemed waived. See id.

With respect to the corporate-officer comment and the

alleged references to the defendants' failure to testify, this is

such a case. The defendants sought curative instructions

addressing specific components of the government's rebuttal

argument and the district court obliged by circulating proposed

instructions. The court invited edits and — in so far as the

proposed instructions pertained to the corporate-officer comment

and the comments allegedly touching upon the defendants' failure

to testify — accepted all the proposed edits. The court then read

- 118 -
the edited instructions to the jury. After doing so, the court

invited counsel to approach sidebar, yet counsel declined the

invitation. That declination unambiguously signified approval of

the supplemental instructions as given and constituted a waiver of

the defendants' arguments on those points. See id.

To be sure, the defendants now argue that waiver

principles apply only to "the court's instruction-in-chief, [but]

not to curative instructions." This is so, they say, because only

the former "result[s] from an iterative process of give and take

between the parties and the court." Here, however, the transcript

shows beyond hope of contradiction that such an iterative process

took place with respect to the curative instructions. In addition,

we previously have found that waiver principles apply with

undiminished force to claims of error targeting curative

instructions. See, e.g., United States v. Charriez-Rolón, 923

F.3d 45, 53 (1st Cir. 2019). We hold, therefore, that the

defendants' claims of error regarding the corporate-officer

comment and the alleged comments on the defendants' failure to

testify are unavailing.

This leaves the claim of error relating to the

prosecutor's use of the "loaded gun" metaphor. The government

concedes that this metaphor was inconsistent with the specific-

intent element of a RICO conspiracy offense and, thus, improper.

Given this concession, we are left to determine whether the

- 119 -
impropriety was harmless. For that purpose, "[t]he bottom-line

question is whether the impropriety 'so poisoned the well that the

trial's outcome was likely affected.'" Kuljko, 1 F.4th at 94

(quoting United States v. Mejia-Lozano, 829 F.2d 268, 274 (1st

Cir. 1987)).

"In this context, harmless error review takes into

account a multiplicity of factors." Id. Those factors include

"the severity of the impropriety, the nature of the impropriety

(that is, whether or not it was deliberate, whether or not it was

isolated, and the like), the strength of the government's case

against the defendant, and how the district court responded to the

impropriety (especially the timing, nature, and force of any

curative instructions)." Id. The district court, looking at the

"loaded gun" metaphor through this prism, concluded that each of

the pertinent factors "counsel[ed] against a finding that the

Government's misstatement 'so poisoned the well' as to warrant a

new trial." Gurry, 427 F. Supp. 3d at 201. We agree.

This inquiry is, of course, case-specific. As we already

have explained, see supra Parts III(A)-(D), the evidence of the

defendants' guilt was copious. The unseemly metaphor itself played

only a bit part in the case: the prosecutor used it only once in

a rebuttal that lasted around thirty minutes and in a trial that

lasted for over seven weeks. Importantly, the prosecutor made no

attempt to weave the metaphor into other portions of either his

- 120 -
closing argument or the trial as a whole. Considering that the

"loaded gun" imagery occupies only a few lines in a compendious

transcript, the infelicitous comment can fairly be described as

"isolated." United States v. Alcantara, 837 F.3d 102, 110 (1st

Cir. 2016).

The defendants disagree. They argue that the

prosecutor's improper metaphor was a deliberate effort to portray

them "as indiscriminate drug dealers." In support, they rely on

United States v. Carpenter, 494 F.3d 13 (1st Cir. 2007) — a case

in which the defendant was convicted of defrauding investors by

misrepresenting his investment strategy, id. at 16. The prosecutor

used "some permutation of the word 'gamble'" in "eighteen

instances" during closing argument, as well as "numerous

references to other gambling terms" like "cashing in chips,"

"doubling down," and "river boat gambler." Id. at 23. The

district court granted the defendant a new trial, concluding that

these persistent references reflected a deliberate (and ultimately

successful) attempt to inflame the jury, and we affirmed. See id.

at 22.

Except, perhaps, to the extent that it illustrates the

wide margins of the district court's discretion with respect to

the granting of a new trial based on an out-of-bounds closing

argument, Carpenter is not a fair congener. That case involved a

series of improper references and a pattern of abuse. In contrast,

- 121 -
the prosecutor in this case used the "loaded gun" metaphor once,

and the district court supportably found that it was an isolated

instance and not a continuing theme. Moreover, the district court

in Carpenter found that the prosecutor's misconduct was

prejudicial, whereas in this case the district court found that

the misconduct, in light of the curative instructions, was

harmless. Given these significant discrepancies, comparing this

case to Carpenter is like comparing cabbages to cantaloupes.

Here, moreover, the district court's curative

instructions were carefully crafted and went to the heart of the

matter. The content and timing of those instructions argue

persuasively against a finding that the government's misstatement

irretrievably poisoned the well. Importantly, the instructions

unambiguously debunked the prosecutor's mistaken view of the

specific-intent element of the charged offense. The prosecutor

had told the jury that people intend the reasonably foreseeable

consequences of their actions. To ensure that the jurors did not

get the wrong impression, the court told them that this proposition

had nothing to do with the case at hand. Furthermore, the court

told them in no uncertain terms that "knowledge of foreseeable

consequences without more is not enough to establish that someone

specifically intended certain conduct." These pointed

instructions cleared the air and kept the jurors focused on the

real issues in the case.

- 122 -
 Grasping at straws, the defendants say that the curative

instructions were insufficient because they failed to tell the

jury that the prosecutor's argument was improper. But a trial

court is not required to use magic words in framing curative

instructions: it is only required to convey, in clear language,

a message adequate to redress the perceived harm. See United

States v. Riccio, 529 F.3d 40, 45 (1st Cir. 2008) ("This court has

repeatedly held that a strong, explicit and thorough curative

instruction to disregard improper comments by the prosecutor is

sufficient to cure any prejudice from prosecutorial misconduct.").

The curative instructions given by the court below satisfied this

standard, and the court — exercising its discretion — determined

that adding a specific indictment of the prosecutor's misstatement

was unnecessary. The substantial deference that we afford trial

courts in matters of this sort reflects an awareness that the

"trial judge . . . listened to the tone of the argument as it was

delivered," had an opportunity to "observe[] the apparent reaction

of the jurors," and was "more conversant with the factors relevant

to the determination." Carpenter, 494 F.3d at 24. We think that

the district court's determination that its curative instructions

would set the jury straight, without any need to place a scarlet

letter on the prosecutor, was within the broad compass of its

discretion.

- 123 -
 One further observation should be made. Although the

district court's curative instructions are adequate on their face,

the record also offers an external validation of their efficacy.

As the district court noted, the "loaded gun" metaphor "related

primarily" to the intent element of the CSA and honest-services

predicates. Gurry, 427 F. Supp. 3d at 199 n.94. Thus, Gurry's

acquittal on these two predicates lends considerable credence to

the conclusion that the district court's curative instructions

ensured that any damage done by the prosecutor's improper metaphor

did not affect the outcome of the trial. See Kuljko, 1 F.4th at

95.

We summarize succinctly. In view of the isolated nature

of the gun metaphor, the timely and effective curative instructions

given by the district court, the government's independently strong

case against the defendants, and the jury's acquittal of Gurry on

the CSA and honest-services predicates, we hold that the

prosecutor's comment, though unacceptable, was harmless. See

Kuljko, 1 F.4th at 95.

XIV

The penultimate leg of our odyssey brings us to the

defendants' challenges to the district court's restitution orders.

They argue that the district court's calculation of the restitution

amounts reflected only "a kind of rough justice," unsupported by

- 124 -
the record. The government defends the district court's

calculations.

We paint the backdrop. In the wake of the jury verdicts,

the government sought $306,000,000 in restitution. This figure

reflected the value of all Subsys prescriptions written during the

racketeering period (2012-2015). The defendants objected,

challenging the government's method of computation and asserting

that the government's suggested price tag was exorbitant. The

district court found a middle ground, ordering restitution in

lesser amounts. See United States v. Babich, No. 16-CR-10343,

2020 WL 759380, at *6 (D. Mass. Feb. 14, 2020); see also supra

note 3 (listing inter alia per-defendant restitution amounts).

En route the court made five specific rulings. First,

the court awarded restitution to six patient victims. See Babich,

2020 WL 759380, at *3-4. Second, the court declined the

government's invitation to base restitution on the totality of

Subsys prescriptions written during the life of the conspiracy.

See id. at *6. Even so, the court acknowledged that sifting the

legitimate prescriptions from the fraudulent ones would "be too

complicated and unduly prolong and burden the sentencing process."

Id. With that in mind, the court made its third ruling, limiting

restitution to losses traceable to prescriptions written solely by

thirteen bribed coconspirator doctors identified by the

government. See id.

- 125 -
 Fourth, the court awarded as restitution 100 percent of

the insurers' paid claims for Subsys prescriptions written by those

thirteen coconspirator-prescribers. See id. In making these

awards, the court refused to apply two reductions urged by the

defendants. See id. One requested reduction was "to account for

only those claims that passed through the IRC." Id. The other

was "to account for only those prescriptions made for non-cancer

patients." Id. Figures reported by the government for these two

categories, the defendants argued, should be deemed a cap for

permissible restitution.18 The district court rejected this two-

pronged argument, stating that "[a]lthough the Court finds the

amount of restitution owed beyond the thirteen co-conspirator

doctors to be too complicated to calculate, it is clear that the

amount that would be owed is at least equal to the total value of

prescriptions written by the bribed doctors." Id.

Fifth, the court apportioned restitution. It held

Kapoor fully responsible for the total amount of restitution owed

— $59,755,362.45 — and capped the restitution obligations of the

18 According to a government expert, "approximately 80.9% of
all Subsys prescriptions" were processed by the IRC. And according
to a second government expert, prescriptions written for non-
cancer patients accounted for approximately 73 percent of Subsys
prescriptions written by the thirteen coconspirator-prescribers.

- 126 -
other defendants at lesser levels.19 See Babich, 2020 WL 1235536,

at *10.

The central restitution-related issue on appeal revolves

around the district court's decision to award insurers 100 percent

of paid claims for Subsys prescriptions written by the thirteen

coconspirator-prescribers. "We review restitution orders for

abuse of discretion, examining the court's subsidiary factual

findings for clear error and its answers to abstract legal

questions de novo." United States v. Chiaradio, 684 F.3d 265, 283

(1st Cir. 2012); see Padilla-Galarza, 990 F.3d at 92.

A defendant convicted of certain federal crimes

(including, as relevant here, crimes "committed by fraud or

deceit," 18 U.S.C. § 3663A(c)(1)(A)(ii), "must make restitution to

victims commensurate with the victims' actual losses," United

States v. Naphaeng, 906 F.3d 173, 179 (1st Cir. 2018).

"[R]estitution is designed to compensate the victim, not to punish

the offender." Id. In awarding restitution, the court's goal is

"to make the victim whole again." United States v. Innarelli, 524

F.3d 286, 293 (1st Cir. 2008). Thus, a restitution order should

19Of course, liability for restitution under federal law may
be joint and several and may be apportioned by the court among the
responsible parties. See 18 U.S.C. § 3664(h). In this instance,
the court apportioned that liability among the defendants who went
to trial and those that pleaded guilty before trial (Burlakoff and
Babich).

- 127 -
"not confer a windfall upon [the] victim." Naphaeng, 906 F.3d at

179.

For the purpose of calculating restitution, actual loss

is the beacon by which federal courts must steer. See id. In

this context, actual loss is "limited to [the] pecuniary harm that

would not have occurred but for the defendant's criminal activity."

Id. (quoting United States v. Alphas, 785 F.3d 775, 786 (1st Cir.

2015)). This standard obligates the government to show both that

the particular loss would not have occurred but for the conduct

undergirding the offense of conviction and that a causal nexus

exists between the loss and the conduct — a nexus that is neither

too remote factually nor too remote temporally. See United States

v. Cutter, 313 F.3d 1, 7 (1st Cir. 2002).

Restitution is serious business, but hearings to

quantify restitution amounts should not be allowed to spawn mini-

trials. As we previously have explained, we do not expect a

sentencing court to "undertake a full-blown trial" in order to

arrive at an appropriate restitution amount. Naphaeng, 906 F.3d

at 179. Nor do we hold a sentencing court to a standard of

"absolute precision" when fashioning restitution orders. Id.

(quoting United States v. Mahone, 453 F.3d 68, 74 (1st Cir. 2006));

see United States v. Sánchez-Maldonado, 737 F.3d 826, 828 (1st

Cir. 2013). In the end, we will uphold a sentencing court's

restitution award "[a]s long as the court's order reasonably

- 128 -
responds to some reliable evidence." Sánchez-Maldonado, 737 F.3d

at 828; see Naphaeng, 906 F.3d at 179 ("[A] restitution award

requires only 'a modicum of reliable evidence.'" (quoting United

States v. Vaknin, 112 F.3d 579, 587 (1st Cir. 1997))).

Although this standard is "relatively modest in

application," Padilla-Galarza, 990 F.3d 60 at 92, it has some

teeth. A sentencing court's "[m]ere guesswork will not suffice."

Naphaeng, 906 F.3d at 179; see Vaknin, 112 F.3d at 587. Similarly,

"rough approximation[s]" that do not "sufficiently

reflect[] . . . the losses" of the victims are not appropriate

grist for the restitution mill. Innarelli, 524 F.3d at 294. The

court must resolve any genuine and material disputes about "the

fact, cause, or amount of the loss" by a preponderance of the

evidence. Vaknin, 112 F.3d at 582-83; see 18 U.S.C. § 3664(d).

Given this framework, we conclude that the district

court's determination to award as restitution 100 percent of Subsys

claims linked to the thirteen coconspirator-prescribers is

insupportable. To be specific, the court's determination that all

of the claims traceable to the thirteen coconspirator-prescribers

constituted actual losses caused by the defendants' fraudulent

conduct was not borne out by the preponderance of the evidence.

For one thing, no party offered evidence that supported the 100-

percent figure. In fact, a government expert opined, without

contradiction, that "approximately 80.9% percent of all Subsys

- 129 -
prescriptions passed through the IRC." 80.9 percent is not 100

percent, and the government represented to the court that the

expert's figure was "a fair and consistent, reasonable approach

for the court to use." According this figure due weight, it is

evident that the government did not establish but-for causation

for all of the claims traceable to the thirteen coconspirator-

prescribers. Indeed, the government's steadfast reliance on the

expert's calculations is functionally equivalent to an admission

that not every Subsys prescription written by these doctors

received prior authorization as a result of IRC fraud.

For another thing, the district court appears to have

taken a shortcut to compensate for the difficulty of calculating

restitution with respect to Subsys prescriptions written by

unbribed physicians. See Babich, 2020 WL 759380, at *6. In

justifying its finding of actual loss generated through

coconspirator-prescribers, the district court pointedly referred

to the incalculable losses caused by non-bribed doctors. See id.

This reference, though, was out of step with the court's earlier

determination that restitution would take account only of the

losses caused by the coconspirator-prescribers. See id. To this

extent, then, the court's award was internally inconsistent: on

the one hand, the court appears to have found that the losses

generated by non-bribed doctors were incalculable but, on the other

- 130 -
hand, to have found that those losses nonetheless justified more

munificent restitution awards.

These infirmities doom the restitution orders. Every

loss that factors into the restitutionary amount must "have an

adequate causal link to the defendant[s'] criminal conduct."

Alphas, 785 F.3d at 786. The blending of two distinct sets of

losses, one of which was incalculable, fails to satisfy the

causality requirement. Consequently, the challenged restitution

orders must be vacated. On remand, the district court should

recalculate the amounts of restitution consistent with its earlier

determination that restitution should be limited to prescriptions

written by the coconspirator-prescribers. What remains is for the

court to "tak[e] into account the extent (if at all) to which the

[coconspirator-prescribers'] claims encompassed legitimate

losses" not processed through the IRC, id., and to refashion the

restitution orders accordingly. Although the court's "reasoning

and the calculations leading to the amounts ordered" must be clear,

Innarelli, 524 F.3d at 295, its bottom-line determination need

only amount to a reasonable response to reliable evidence in the

record, see Sánchez-Maldonado, 737 F.3d at 828.

XV

The finish line is in sight. The district court ordered

monetary forfeitures in varying amounts, see supra note 3, and the

affected parties (including the government) ask us to resolve

- 131 -
dueling claims of error pertaining to these forfeiture orders. In

evaluating forfeiture orders, we assay the court's legal

conclusions de novo and examine its factual findings for clear

error. See United States v. George, 886 F.3d 31, 39 (1st Cir.

2018).

The baseline rule is uncontroversial. A defendant who

has been convicted of RICO conspiracy is liable to forfeit "any

property constituting, or derived from, any proceeds which the

person obtained, directly or indirectly, from racketeering

activity." 18 U.S.C. § 1963(a)(3). Following the defendants'

convictions, the government sought forfeitures equaling the gross

proceeds obtained by Lee, Simon, Gurry, and Rowan, respectively,

during the racketeering period. Ruling that "any proceeds obtained

from Insys during the time of the conspiracy are forfeitable," the

district court obliged. Babich, 2020 WL 1235536, at *5. The court

went on to hold that "the Defendants' salaries and exercised stock

options constitute 'proceeds' that were obtained 'directly or

indirectly' from the RICO conspiracy."20 Id. As an offset, though,

the court held that the income taxes that each defendant had paid

were not "proceeds" under section 1963(a)(3) because those amounts

never "ended up in the Defendants' pockets for them to spend in

20 Insofar as the forfeiture orders are based upon the
monetization of exercised stock options, neither side has
challenged the district court's calculations.

- 132 -
the way in which they wanted." Id. at *7 (alterations omitted).

Accordingly, the court — in shaping its forfeiture orders as to

Lee, Simon, Gurry, and Rowan — deducted from their respective gross

incomes "the amount of the tax withheld" during the racketeering

period. Id.

Gurry lands the first blow. He contends that the

district court erred as a matter of law because "it declined to

determine what portion of [his] income was tainted by racketeering

activity." The government counterpunches. In a cross-appeal, it

contends that the tax offsets were erroneous as a matter of law.21

We deal with each contention in turn.

A

A defendant's proceeds from racketeering activity are

"subject to a rule of proportionality." Cadden, 965 F.3d at 37

(quoting United States v. Angiulo, 897 F.2d 1169, 1211 (1st Cir.

1990)). This guardrail ensures that proceeds are subject to

forfeiture only to "the extent they are tainted by the racketeering

activity." Id. (quoting Angiulo, 897 F.2d at 1212). It follows

that a district court's forfeiture order must determine "the

portion of [the defendant's] earnings . . . over the relevant time

21 Due to his unique compensation package, Kapoor neither
sought nor received a tax offset. See Babich, 2020 WL 1235536, at
*6 n.6. As a result, the government's cross-appeal does not
implicate his forfeiture order.

- 133 -
period that were tainted by the racketeering activity and therefore

subject to forfeiture." Id. at 38.

Gurry advances three arguments as to why certain

portions of his work at Insys cannot be linked to the racketeering

activity and as to why, as a result, the forfeiture of his entire

salary was in error. Lee, Simon, and Rowan adopt these arguments.

Gurry first notes that although he was an Insys employee

until 2016, his work at the IRC ended in May of 2014. Because

"[t]here is no evidence that his job responsibilities after May

2014 included any racketeering activity," he posits, any

subsequent proceeds are not subject to forfeiture. This is too

crabbed a view of the facts: Gurry's relinquishment of the

responsibility for supervising the IRC did not end his furtherance

of, participation in, and profiting from the racketeering scheme.

By 2014, Gurry had negotiated with insurance companies to add

Subsys to their compendia of approved drugs. Those efforts helped

the IRC to continue its fraudulent scheme and garner additional

revenue for Insys even after Gurry's responsibilities changed. To

the extent that Gurry's racketeering activities on behalf of the

IRC generated profits for him after his departure from the IRC,

that revenue constitutes proceeds "obtained from the racketeering

activity . . . that formed the basis of [his] convictions." Id.

at 37. Those proceeds were, therefore, forfeitable. See id. And

in any event, "[m]ere cessation of activity in furtherance of the

- 134 -
conspiracy does not constitute withdrawal" from the conspiracy.

United States v. Leoner-Aguirre, 939 F.3d 310, 319 (1st Cir. 2019)

(quoting United States v. Ciresi, 697 F.3d 19, 27 (1st Cir. 2012)).

Next, Gurry maintains that his work for the IRC comprised

only 20 percent of his job responsibilities. But he cites no

authority to support a reduction in his forfeiture amount based on

the percentage of his time devoted to the scheme. It would be

perverse to provide an incentive for racketeering efficiency, and

we do not think that a racketeer can limit his forfeiture liability

by the simple expedient of devoting some of his time to legitimate

work. Forfeiture calculations depend on the proceeds gained

directly or indirectly from racketeering activity, see 18 U.S.C.

§ 1963(a)(3), not on the percentage of a defendant's time devoted

to the conspiracy.22

Gurry also contends that his forfeiture order should

reflect only the percentage of fraudulent Subsys sales during the

racketeering period, not all Subsys sales during that period. The

government confesses error and agrees that a remand on this ground

is appropriate. That confession is premised upon our opinion in

22 At any rate, Gurry has not established whether the 80
percent of his work allegedly unrelated to the racketeering
activity generated earnings for him that were independent of
fraudulent Subsys sales. What counts is that the record supports
the conclusion that Gurry knowingly joined and furthered the
insurance-fraud scheme and that his earnings during that time for
the "non-IRC work" flowed at least indirectly from his IRC efforts.

- 135 -
Cadden, 965 F.3d at 37-38, which was decided while these appeals

were pending. There, we vacated a forfeiture order because "the

government failed to prove that all [drug] sales over the period

in question were generated by fraud." Id. Profits from non-

fraudulent sales, we said, are not proceeds obtained (directly or

indirectly) from the racketeering activity. See id. at 37-38.

We ordered the district court, on remand, "to assess . . . the

portion of [the defendant's] earnings . . . that were tainted by

racketeering activity." Id. at 38.

The same instruction is warranted here. As a matter of

law, any Subsys prescription processed independently of the IRC

falls outside the scope of the fraudulent scheme. And since the

IRC did not seek prior authorization for every Subsys prescription,

the district court must determine the percentage of Subsys prior

authorizations that were successful through the IRC's efforts.

Forfeiture of the whole of Gurry's earnings was, therefore, in

error. The forfeiture orders pertaining to Lee, Simon, and Rowan

suffer from the same defect, and those orders also must be

revisited.

Gurry is barking up the wrong tree, however, when he

tries to convince us that "the IRC did not lie about every

prescription it processed." The defendants agreed below that 73

percent of the IRC's authorizations involved prescriptions for

non-cancer patients and the district court found that the IRC

- 136 -
"misled insurers in a number of ways," even when the patients had

cancer. Babich, 2020 WL 1235536, at *6. The IRC's deceptions

included dissembling about patients experiencing breakthrough

cancer pain, having a history of cancer, having tried-and-failed

other medications, and having difficulty swallowing. See id.

These tactics were systematically employed by the IRC and did not

become honest or accurate by virtue of a patient having cancer.

See id. Mendacity was a hallmark of the IRC's operations — a

hallmark that permeated its prior authorization efforts.

We agree with the district court that "the fact that a

prescription was requested for a cancer patient is insufficient to

establish that it was not fraudulent." Id. Based on the

overwhelming evidence that these sleazy tactics were business as

usual at the IRC, we find that the district court's determination

that each prescription processed by the IRC during the racketeering

period was tainted by fraud is grounded upon reasonable inferences

drawn from adequately established facts. The district court's

determination was not clearly erroneous.

B

We turn next to the government's cross-appeal. We

conclude that the district court's decision to offset the

defendants' forfeiture obligations based on the income taxes they

paid on those earnings constituted error. Two recent cases inform

this conclusion.

- 137 -
 In Cadden, the defendant argued that "the District Court

erred in calculating the forfeiture amount without deducting the

amount in taxes that he paid on those proceeds." 965 F.3d at 38.

We disagreed, holding that "the word 'proceeds' in the forfeiture

statute refers to gross proceeds, not net profits." Id. (quoting

United States v. Hurley, 63 F.3d 1, 21 (1st Cir. 1995)). Because

the defendant "clearly 'obtained' the amount of funds subject to

forfeiture before they were subject to taxation," that amount was

"subject to forfeiture, even though the amount he obtained was

itself taxable." Id.

Our decision in United States v. Chin, 965 F.3d 41 (1st

Cir. 2020), is to like effect. There, we concluded that "the fact

that the offender is required to pay a certain portion of his

salary to the federal government as taxes does not affect the fact

that he 'obtained' that portion," id. at 57. Taken together,

Cadden and Chin resolve the issue. The defendants in this case

were taxed on the proceeds subject to forfeiture precisely because

they had "obtained" those proceeds.

C

Consistent with these rulings, we vacate the district

court's forfeiture orders as to Lee, Simon, Gurry, and Rowan. The

district court must assess what percentage of Subsys prior

authorizations were successful independently of the IRC, and

reduce the forfeiture amounts of each defendant by that percentage.

- 138 -
See Cadden, 965 F.3d at 38. It should not, however, apply any tax

offset. We remand for the purpose of recalculating these

forfeiture amounts.

XVI

We need go no further.23 Insys and Kapoor deserve great

credit for developing Subsys — a medication which, appropriately

dispensed, would have been an important weapon in society's

continuing battle to alleviate breakthrough cancer pain. But

Subsys was not appropriately dispensed. Instead, the defendants

— driven by unalloyed greed — marketed the medication through a

pattern of racketeering activity and conspired to ensure that it

would be dispensed outside the usual course of medical practice

and without a legitimate medical purpose. "Pill mills for us meant

dollar signs" and — from the defendants' coign of vantage — Subsys

prescriptions, like snake oil on the frontier, became above all

else a means of generating revenue. In taking this cynical

approach, the defendants turned what should have been a blessing

into a curse.

The jury, after a protracted trial presided over with

great care and circumspection by a no-nonsense judge, heard

23To the extent, if at all, that particular defendants have
alluded to other potential claims of error in their extensive
briefing, those claims are either insufficiently developed or
patently meritless. Thus, we reject them without further
elaboration.

- 139 -
detailed evidence with respect to the defendants' pernicious

practices regarding the marketing of Subsys. The jury found the

evidence sufficient to hold the defendants guilty beyond a

reasonable doubt on virtually all of the charges lodged in the

indictment. The jury's findings and verdicts are, we think, fully

supportable, and the defendants' multifaceted challenges to them,

though skillfully presented, are without force. We conclude,

therefore, that the findings and verdicts must stand.

We reach a different result with respect to certain

monetary awards made by the district court ancillary to sentencing.

Although the defendants do not challenge their sentences as such

(and those sentences must remain intact), the restitution and

forfeiture orders are attacked (some by the defendants, some by

the government, and some by both). We find that the challenged

amounts were not properly calculated in certain respects. Thus,

certain restitution and forfeiture orders, identified above, must

be vacated, and the case must be remanded for further proceedings

consistent with this opinion.

To summarize, we set aside the district court's vacation

of certain of the jury's special findings regarding the guilt of

Kapoor, Lee, Simon, and Rowan vis-à-vis the CSA and honest-services

predicates and order reinstatement of those findings. We affirm

the jury's special findings and verdicts as to all defendants. We

also affirm the district court's denial of the defendants' sundry

- 140 -
motions for judgments of acquittal and/or new trials. So, too, we

affirm the district court's orders with respect to challenged

pretrial and mid-trial rulings. Finally, we affirm the defendants'

sentences,24 but vacate the district court's restitution and

forfeiture orders (except for the forfeiture order regarding

Kapoor) and remand for recalculation of the appropriate

restitution and forfeiture amounts.

Affirmed in part, reversed in part, vacated in part, and remanded.

24The government has not requested that, upon reinstatement
of the special findings concerning the CSA and honest-services
predicate, see supra Part III, we remand for resentencing of the
four affected defendants (Kapoor, Lee, Simon, and Rowan). In the
absence of such a request, we see no need to do so.

- 141 -