United States v. Universal Rehab

Opinions of the United
1999 Decisions States Court of Appeals
for the Third Circuit

2-11-1999

USA v. Universal Rehab
Precedential or Non-Precedential:

Docket 97-1412

Follow this and additional works at: http://digitalcommons.law.villanova.edu/thirdcircuit_1999

Recommended Citation
"USA v. Universal Rehab" (1999). 1999 Decisions. Paper 38.
http://digitalcommons.law.villanova.edu/thirdcircuit_1999/38

This decision is brought to you for free and open access by the Opinions of the United States Court of Appeals for the Third Circuit at Villanova
University School of Law Digital Repository. It has been accepted for inclusion in 1999 Decisions by an authorized administrator of Villanova
University School of Law Digital Repository. For more information, please contact Benjamin.Carlson@law.villanova.edu.
Filed February 11, 1999

UNITED STATES COURT OF APPEALS
FOR THE THIRD CIRCUIT

No. 97-1412

UNITED STATES OF AMERICA,
Appellant

v.

UNIVERSAL REHABILITATION SERVICES
(PA), INC.,

On Appeal from the United States District Court
for the Eastern District of Pennsylvania
D.C. Criminal Action No. 94-cr-00147-1
District Judge: Hon. Robert F. Kelly

Nos. 97-1413 & 97-1467

UNITED STATES OF AMERICA

v.

ATTILA HORVATH,
Appellant/Cross-Appellee

On Appeal from the United States District Court
for the Eastern District of Pennsylvania
D.C. Criminal Action No. 94-cr-00147-2
District Judge: Hon. Robert F. Kelly
Nos. 97-1414 & 97-1468

UNITED STATES OF AMERICA

v.

RICHARD J. LUKESH,
Appellant/Cross-Appellee

On Appeal from the United States District Court
for the Eastern District of Pennsylvania
D.C. Criminal Action No. 94-cr-00147-3
District Judge: Hon. Robert F. Kelly

Argued: May 22, 1998

Before: ROTH, MCKEE, and GARTH, Circuit Judges

(Opinion filed February 11, 1999)

Michael R. Stiles
United States Attorney
Walter S. Batty, Jr.
Assistant United States Attorney
Chief of Appeals

Valli F. Baldassano (Argued)
Suzanne B. Ercole (Argued)
Assistant United States Attorneys
Suite 1250
615 Chestnut Street
Philadelphia, PA 19106

Attorneys for Appellee/
and Cross-Appellant

2
 Thomas Colas Carroll, Esquire
(Argued)
Carroll & Cedrone
The Curtis Center, Suite 750
Independence Square West
Philadelphia, PA 19106

Attorney for Appellants/
and Cross-Appellees

OPINION OF THE COURT

ROTH, Circuit Judge:

Defendants, Universal Rehabilitation Services, Inc.
(Universal), Attila Horvath, and Richard Lukesh, were
charged with mail fraud and false claims in a 39 count
indictment. The government alleged that they conducted a
Medicare fraud scheme over a three-year period. The jury
convicted the defendants only on a mail fraud count, Count
One, the earliest count of the indictment. The government
now argues that evidence of illegal activity, which occurred
after the date of the offense in Count One but still within
the three-year period of the originally charged scheme, can
be used to sustain the verdict of guilty on Count One.

In this appeal, we must determine to what extent post-
offense evidence can be considered to support the
conviction on Count One. We must also decide whether the
District Court properly admitted into evidence the guilty
pleas of two testifying mid-level managers after defendants
had represented to the court that they would not, on cross
examination, challenge the credibility of these witnesses or
otherwise make the pleas admissible.

I. Facts

The corporate defendant, Universal, provided
rehabilitative services, including speech therapy, to
Medicare patients in nursing homes. The patients served
were primarily the elderly. Universal submitted claims for
the patients it treated to Independence Blue Cross (IBC),
which administered Medicare coverage. IBC would then
submit claims to Medicare.

3
Medicare is authorized to reimburse speech therapy
treatments that are both medically reasonable and
necessary. Pursuant to 42 U.S.C. S 1395 et seq. and its
implementing regulations, manuals are published that
provide Medicare coverage guidelines. These include four
criteria to determine if the treatment is medically
reasonable and necessary, and therefore reimbursable: 1)
the therapy must be safe and effective for treatment of the
patient's condition; 2) the services provided to the patient
must be of a level of complexity that the services can only
be provided by a certified speech pathologist; 3) where
restorative treatment is ordered, there must be an
expectation that the services being provided to the patient
will bring about significant improvement in a reasonable
period of time; and 4) the frequency and duration of
services must be reasonable and necessary for the patient's
condition.

Universal therapists would evaluate patients in accord
with the physician's orders and prepare a plan of
treatment. The evaluation and plan of treatment must state
whether the treatment is in fact necessary. The evaluation
and plan is then approved (signed) by the doctor. As
treatment progresses, speech therapists create progress
notes. If treatment is to continue, Medicare requires that
the need for it be certified every 30 days by a medical
doctor. Medicare will continue to pay for speech therapy as
long as patients are making progress toward the goals
established in the plan of treatment. The doctor executes
the 30 day certification by signing a Medical Information
Form (MIF), which contains a summary, primarily from the
therapists' progress notes, of the previous 30 days of
treatment, as well as a recommended course of treatment
for next 30 days.

Universal submitted electronic Medicare bills to IBC.
Essentially, IBC would review Universal's reimbursement
claims. Universal and IBC often disagreed about the
interpretation of Medicare guidelines and the propriety of
the bills submitted. At some point, IBC began requesting
that Universal supply the following documentation to
support the electronic bills: a) the evaluation and plan of
treatment, b) the progress notes, and c) the MIFs.

4
Claims that were rejected would be "rewritten" by
Universal staff. Depending on the reason for rejection, the
"rewriting" to obtain reimbursement might be fraudulent.
For example, if information was missing from a form,
Universal personnel would "rewrite" the form to include the
missing information. Other "rewriting," however, was not as
innocuous. For example, a mid-level corporate employee
pled guilty to mail fraud for falsifying the performance
record of a patient on certain speech therapy tests to create
the "appearance" that the patient had made progress,
thereby rendering the services reimbursable. In fact, the
patient was not making progress and the services should
not have been reimbursable. Another type of "rewriting"
consisted of changing evaluations and MIFs after the doctor
had signed them. The doctor's signature would then be
taped and xeroxed onto the "rewritten" forms.

The government alleged that a mail fraud scheme
occurred from the summer of 1988 through September 21,
1991, during which 1) evaluations and plans of treatment
were rewritten to create the "appearance" of patients who
could properly receive reimbursable therapy; 2) MIFs
containing physician's certifications were falsified, that is,
altered after the doctor originally signed them and the
signature photocopied back on; and 3) progress notes were
altered to conform with the falsified evaluation and MIFs. A
grand jury indicted Attila Horvath, Vice President and
Director of Finance at Universal; Richard Lukesh, Director
of Operations at Universal; Universal, the corporation itself;
and three other Universal employees for mail fraud, in
violation of 18 U.S.C. S 1341 (Counts One through
Seventeen, charging 17 mailings in furtherance of the
fraud), and false claims, in violation of 18 U.S.C.S 287
(Counts Eighteen through Thirty-nine).

After a jury trial, Horvath, Lukesh and Universal were
convicted on Count One and acquitted on each of the other
38 counts, representing 16 other mailings and false claims
that arose from the practice of falsifying speech therapy
documents. Count One charged that, in furtherance of the
scheme, Blue Cross mailed a check, dated May 10, 1989, to
Universal. The check represented a claim for treatment

5
rendered to a patient, Mildred Hynds, from February 15 to
28, 1989.1

Before trial, defendants had moved in limine to prevent
the government from presenting evidence that government
witnesses Judy Blum Bonjo and Penny Martin, employees
of Universal, had pled guilty to mail fraud.2 Defendants
promised that they would not challenge these witnesses'
credibility on cross-examination. Defendants argued that
the guilty pleas were not relevant under Rule 401 of the
Federal Rules of Evidence and, if relevant, were more
prejudicial than probative under Rule 403. The District
Court denied the motion in limine. Defendants appeal this
evidentiary ruling.

After the verdict, defendants moved for a Fed. R. Crim. P.
Rule 29 judgment of acquittal, arguing that there was
insufficient evidence to support the jury's verdict. The court
denied the motion. Defendants appeal this ruling as well.3

The District Court had subject matter jurisdiction over
federal indictments, charging violations of the federal
criminal code. We have appellate jurisdiction over
defendants' direct appeal from a judgment of conviction. 28
U.S.C. S 1291.

II. Sufficiency of the Evidence

Immediately after the jury returned its verdict of
conviction on Count One, defendants moved for acquittal
under Rule 29, arguing that there was insufficient evidence
_________________________________________________________________

1. Counts Two, Three, and Four also involved the mailing of checks that
pertained to Mildred Hynds. Count Twenty-seven charged a false claim
arising out of a bill that was submitted for the treatment of Hynds.

2. Blum Bonjo testified that she had pled guilty to Count One of the
indictment before the jury. App. 2190. Martin was charged in a separate
indictment and testified that she had pled guilty to one count of mail
fraud.

3. The government also filed a cross-appeal, claiming that the District
Court erred at sentencing in refusing to consider the fraud loss from the
"virtually identical" conduct charged in the other 38 acquitted counts as
relevant conduct under Section 1B1.3 of the Sentencing Guidelines.
Because of our holding on defendants' appeal, we do not reach the
government's cross-appeal.

6
to sustain the verdict. The District Court denied the motion
as to all three defendants. United States v. Universal
Rehabilitation Services, Inc., Attila Horvath, and Richard
Lukesh, No. 94-147, E.D.Pa., slip op. 3-12, May 31, 1996.
We review de novo the district court's denial of a post-
verdict judgment of acquittal. United States v. Iafelice, 978
F.2d 92 (3d. Cir. 1992).

We apply the same legal standard as the District Court.
We must view the evidence in the light most favorable to
the jury verdict and presume that the jury properly
evaluated the credibility of the witnesses, found the facts,
and drew rational inferences. Id. at 94. In that light, we
"must affirm the convictions if a rational trier of fact could
have found defendant[s] guilty beyond a reasonable doubt,
and the verdict is supported by substantial evidence." U.S.
v. Coyle, 63 F.3d 1239, 1243 (3d Cir. 1995).

"The mail fraud statute, 18 U.S.C. S 1341, proscribes any
scheme or artifice to defraud in which the defendant
participated with the specific intent to defraud and in
which the mails were used `in furtherance of the fraudulent
scheme.' " Coyle, 63 F.3d at 1243."Proof of specific intent
. . . may be found from a material misstatement of fact
made with reckless disregard for the truth." United States
v. Hannigan, 27 F.3d 890, 892 n.1 (3d Cir. 1994). As such,
the government must prove that a scheme to defraud
existed, that each defendant participated in that scheme,
and that each participated with specific intent to defraud.
United States v. Pflaumer, 774 F.2d 1224, 1233 (3d Cir.
1985).

It follows from the above that an individual must join a
scheme while the scheme exists; that is, that the scheme
and a defendant's knowing participation in it must occur
over the same period of time. In this case, the government
alleged in the thirty-nine count indictment that a scheme
took place from the summer of 1988 through September
21, 1991. However, the jury convicted the defendants only
on Count One, the May 10, 1989, mailing. The jury
acquitted the defendants on the remaining 38 counts,
relating to mailings and false claims after May 10, 1989.
Thus, we must define the period of time which constitutes
the scheme for which the jury convicted the defendants.

7
Because of the acquittals on Counts Two through Thirty-
nine, the scheme could not have run through the entire
period which the government charged in the indictment.
Nevertheless, May 10, 1989, the date of the offense charged
in Count One, is not necessarily the evidentiary cut-off. We
must determine whether events that took place after May
10, 1989, up until September 1991, can be used to support
the jury verdict on the May 10 mailing.

The government argues that the scheme is not bound in
time by the May 10, 1989, mailing -- that the Hynds check,
the subject of Count One, was a mailing in furtherance of
a broad three year scheme. However, as we made clear in
United States v. Pflaumer, 774 F.2d 1224 (3d Cir. 1985), in
the context of a conspiracy to commit mail fraud in
violation of 18 U.S.C. S 371, the conspiracy at issue "was a
conspiracy to violate the mail fraud statute, which end[s]
with the last mailing." Id. at 1232. As such, willful
membership in the scheme must have existed on or before
May 10, 1989, the date of the mailing for the count of
conviction. There is no concept of retroactive joinder in mail
fraud schemes after the last culpable mailing. Moreover,
because the defendants were acquitted of the later-
occurring mail fraud counts, those mailings cannot be
considered to be "culpable mailings."

Furthermore, the fact that Pflaumer concerned conspiracy
to commit mail fraud, not just mail fraud, is a distinction
without a difference under the circumstances of this case.
A conspiracy to violate a substantive prohibition in a
federal statute ends when the unlawful object of the
conspiracy has been accomplished. Grunewald v. United
States, 353 U.S. 391, 406-15 (1957). In Pflaumer, the object
of the conspiracy was the use of the mails to effectuate a
scheme to defraud states of tax revenues. Although the
scheme to defraud the states may have continued past the
last mailing, the use of the mails to effectuate the scheme
ceased upon the last mailing. 774 F.2d at 1232.

As was the situation in Pflaumer, the object of the
scheme here was mail fraud, the mailing was in
furtherance of the scheme, and the mailing marks the end
of the scheme. To be sure, subsequent words and deeds
may bear on whether a person was participating and

8
participating willfully in a scheme. For example, a
statement by a defendant after the end of the scheme may
make clear that he was in fact willfully participating in, or
had knowledge of, the scheme before it ended. There must,
however, be substantial evidence of participation before the
mailing. United States v. Pearlstein, 576 F.2d 531, 537 (3d
Cir. 1978).

A. Evidence That a Crime Was Committed

The question then is whether there is sufficient evidence
of events prior to the May 10, 1989, mailing to sustain the
jury verdict? To sustain the jury verdict, in addition to
demonstrating that each defendant willfully joined the mail
fraud scheme, the government must produce sufficient
evidence that a mail fraud scheme existed. This means that
there must be sufficient evidence that the Hynds mailing
was in furtherance of a mail fraud scheme, i.e. , that the
services to Mildred Hynds for which defendants billed and
were paid $1,411 by IBC's check # 0249905, dated May 10,
1989, were unauthorized under Medicare coverage
guidelines and that defendants' misrepresentation of these
services was part of the scheme to defraud.

Defendants argue that there was no evidence that a
crime was in fact committed. They argue that the evidence
indicates that the treatment rendered to Hynds was
authorized by her physician as medically necessary and
appropriate. The record reveals a different story. Carol
Pomilio, a speech therapist at Universal, testified that
Medicare initially rejected the Hynds billing. In order to
obtain reimbursement, the February 15, 1989, evaluation
and plan of treatment, which the doctor had signed, was
destroyed and replaced by one with an altered evaluation
that did not accurately reflect Hynds' treatment plan. The
doctor's signature was then xeroxed onto the altered
document. As documentary support for this treatment plan,
Pomilio lowered the scoring percentages on the progress
notes. The fact that the treatment was originally ordered by
a doctor does not neutralize the fraud; i.e., that the billing
for the continued treatment, based on falsified progress
notes and a xeroxed doctor's signature, was fraudulent.

There is no doubt that, as defendants argue, the
rewriting of or resubmission of rewritten documents is not

9
per se fraudulent. In fact, Julia Blum Bonjo, another
Universal speech therapist, testified that a large part of her
job consisted of legitimate rewriting or redrafting of
documents. However, when the rewriting consisted of
misrepresentation of a patient's test results with the intent
of obtaining otherwise unauthorized reimbursement, there
is sufficient evidence to find an underlying scheme to
defraud, which is required as an element of the offense of
mail fraud.

The next question we must consider is whether there is
sufficient evidence that each of the defendants willfully
joined the scheme to commit mail fraud. United States v.
Pearlstein, 576 F.2d at 537. In this connection, the
government must prove that each defendant willfully joined
the scheme before May 10, 1989, the date of the last
mailing in furtherance of the scheme. It must show that
each defendant possessed the requisite intent to defraud.
Proof is required that "defendants must either have devised
the fraudulent scheme themselves, or have wilfully
participated in it with knowledge of its fraudulent intent."
Id. A defendant need not personally be involved with the
actual mailing to be liable so long as there was knowledge
that the use of mails would follow in the ordinary course of
business or that such use can reasonably be foreseen.
United States v. Sturm, 671 F.2d 749, 751 (3d Cir. 1982);
United States v. Funt, 896 F.2d 1288, 1294 (11th Cir. 1990).4

B. Evidence Linking Defendant Horvath to the Crime

We will first examine the evidence against defendant
Attila Horvath up to and through the May 10, 1989, mailing
of the count of conviction. The government must prove that
Horvath willfully joined the scheme before that date. The
District Court enumerated specific record references to
Horvath's knowledge and willful participation in the scheme
upon which the jury could have relied. However, once we
_________________________________________________________________

4. Defendants also contend in their appeals that there was insufficient
evidence of the use of the mails. We have reviewed the evidence of the
business practices and of specific reference to the correspondence in
question, and we will affirm the District Court on this point. See United
States v. Burks, 867 F.2d 795 (3d Cir. 1989); United States v. Hannigan,
27 F.3d 890 (3d Cir. 1994).

10
eliminate the post-May 1989 evidence, there is little of any
significance.

(1) There is the testimony of Wendy Gold, who was
Director of Speech Therapy before Penny Martin. Gold
testified that on April 28, 1989, she discussed with Horvath
and Lukesh the problem of over-utilization, i.e., treating
patients unnecessarily to increase production. Moreover,
Gold discussed with Horvath the fact that many of the
patients in the nursing homes with which Universal
contracted would not qualify for Medicare therapy. At this
meeting and in subsequent memos, dated May 8 and May
15, 1989, there was extensive discussion about ways to
increase productivity. There was, however, no evidence that
anyone suggested falsification in billing. Gold took the
position that asking therapists to do what Universal wanted
done in order to increase production would not be ethical.
Gold subsequently quit due to pressure to increase
production.

(2) Judy Blum Bonjo, Director of Utilization Review,
testified that her department was directed by Horvath in an
April 15, 1991, memo to bill the "risk patients" in order to
reduce documentation backlogs. This evidence is not
relevant, however, because it refers to an event more than
2 years after the Count One mailing. Additionally, Horvath
argues that the memo refers to billing prior to receipt of all
the paperwork, not to falsifying information.

Blum Bonjo also testified that Horvath directed her to bill
without physician signatures and without some requisite
documentation in place. This evidence is problematic
because there is no reference to its date. Horvath asserts,
and the Government does not controvert, that this
reference is to the same April 15, 1991, memo and for that
reason occurred after the Count One mailing.

(3) Penny Martin, Director of Speech Therapy after Gold,
testified that Horvath wanted patient documentation
billable no matter what had to be done. Martin explained
that Horvath ran weekly meetings and closely monitored
the results of speech therapy practices at Universal.
Furthermore, Martin explained that between 1989 and
1991 Blum Bonjo met with Lukesh on a daily basis and

11
with Horvath and Lukesh on a weekly basis. Once again,
Horvath asserts, and the Government does not controvert,
that Martin's testimony refers to her observations of what
Horvath told her in June 1990, one year after the Count
One mailing. In fact, Martin did not join Universal until
August 1989, several months after the Count One mailing.

Examined as a whole, the record is extremely sparse as
to Horvath's willful participation in the scheme before May
10, 1989. The only evidence that clearly concerns pre-
mailing events is the Gold testimony. This testimony,
however, only establishes pressure to bill, not pressure to
bill falsely. Although we hesitate to overturn a jury verdict,
we conclude in this instance that, after we exclude the
post-mailing evidence, upon which the jury likely relied,
there is little to support the jury's verdict. As such, we hold
that the District Court erred in denying Horvath's motion
for judgment of acquittal. We will remand the case to the
District Court to enter a judgment of acquittal for Horvath.

C. Evidence Linking Defendant Lukesh to Crime

Contrary to Horvath, the relevant evidence against
Lukesh is stronger.

1) Lukesh admitted he was aware that documents were
being rewritten in order to "augment" the billing record. A
memo, dated Oct. 28, 1988, directed therapists "to be
creative" in getting Medicare reimbursements. On cross-
examination, Lukesh testified that one such "creative way"
to get reimbursed was to rewrite the document to change
the patient profile. Defendants point out that no witness
ever testified to interpreting the term "creative" to imply
falsification. Nevertheless, the jury could infer Lukesh's
state of mind from his choice of words.

2) Lukesh authored a document in the spring of 1988
known as the "billing hold" which became an administrative
measure at Universal. This directive was designed to gather
information on what billing documents were being held up
and for what reason. It was provided to the speech
therapists, often with specific directions on how to alter
speech therapy documents. While it is possible that Lukesh
did not intend that the rewriting of documents, pursuant to
this directive, would be fraudulent (Lukesh testified that he

12
believed documents were merely being corrected or
augmented), the jury was certainly entitled to infer that
Lukesh had knowledge of the fraudulent billing practices.

3) Lukesh was present at an April, 1989, meeting where
Blue Cross alerted Universal that documents "were being
improperly altered and inappropriate patients were being
treated." Lukesh's contention as to the billing hold directive
described above, that he believed documents were merely
being corrected or augmented, is belied by this meeting
with Blue Cross. The information gathered by Lukesh at the
meeting, together with his authoring of the directive, lead to
the permissible inference that Lukesh knew, well before the
Count One mailing, that his employees were falsifying
documents.

4) Karen Lightman Pallies, a speech therapist from
January 1989 through September 1989, testified that she
was required to falsify her therapy documentation and that
Lukesh told the therapists at a pre-May 1989 meeting that
reports were being rejected by Blue Cross and the
therapists would have to keep rewriting these reports until
they were accepted. Pallies did admit that re-writing could
involve simply "documenting accurate historical facts" as
distinguished from "making something up."

5) Blum Bonjo and Martin testified that Lukesh was
aware that documents were being altered before being
submitted to Blue Cross for payment. The testimony by
Martin is problematic because she started working after the
Count One mailing. Furthermore, Lukesh asserts, and the
government does not controvert, that Blum Bonjo's
testimony refers to events around April 1990, after the
mailing.

6) Therapist Audrey Isaak told the jury that she attended
regular staff meetings between 1988 and 1991 and at one,
Lukesh told the staff that "our documentation need[s] to be
written in such a way that payment would be granted."

Clearly, the evidence against Lukesh is more substantial
than that against Horvath. Certainly Lukesh's authoring of
documents designed to effectuate a policy of resubmitting
bills and rewriting the documents underlying them, a policy
that led therapists to submit fraudulent documents,

13
provides evidence of intent on Lukesh's part. When this
evidence is coupled with the fact that Lukesh was informed
by Blue Cross of the fraudulent practices, there was
sufficient evidence that Lukesh willfully joined the scheme
to commit mail fraud.

D. Evidence Linking the Corporation to the Crime

Universal, the corporate defendant, argues as the basis
for its post-verdict judgment of acquittal motion that there
was not substantial evidence that it directed its employees
to falsify claims. "A corporation is criminally responsible for
the unlawful acts of its employees or other agents, provided
such unlawful acts are done on behalf of the corporation
and within the scope of the agent's employment or apparent
authority." United States v. American Radiator & Standard
Sanitary Corp., 433 F.2d 174, 205 (3d Cir. 1970).

Universal argues that there was never a corporate policy
of falsifying claims and that no individual was directed to
falsify claims on behalf of the company. Universal contends
that the individual defendants were acting on their own.
This argument has no merit. As the District Court
explained, without even considering the guilt of the high-
level officials Horvath and Lukesh, the evidence was
sufficient for the jury to have concluded that Carol Pomilio
was acting within the scope of her authority and in the
course of her employment with the intent to benefit
Universal when she knowingly misled Medicare by altering
Hynds' patient documentation. Additionally, numerous
witnesses testified that they were pressured to bill for
treatment to inappropriate patients and to rewrite patient
documentation in order to make claims reimbursable, i.e.,
to benefit the corporation. For these reasons, wefind
sufficient evidence to support the conviction of defendant
Universal.

III. Admissibility of the Guilty Plea of a Testifying Co-
Defendant

Julia Blum Bonjo, a speech therapist at Universal, pled
guilty to Count One, the mail fraud count on which the
other defendants were later convicted. Penny Martin also
pled guilty in a separate information to one count of mail
fraud. In exchange for the pleas, Blum Bonjo and Martin

14
agreed to testify as government witnesses. Before the trial,
defendants moved to bar testimony of the pleas. In their
motion, defendants argued that the plea agreements were
irrelevant, and, even if they were relevant, they were more
prejudicial than probative. Defendants represented that
they would not examine Blum Bonjo or Martin in any way
that would make this evidence admissible, i.e., they would
not challenge their credibility. Moreover, defense counsel
asserted that, even if the government was skeptical about
his representation that he would not challenge the
credibility of Blum Bonjo or Martin on cross examination,
the proper procedure was for the court to then permit the
government to prove the pleas on redirect.

The District Court denied defendants' motion. Relying on
United States v. Gaev, 24 F.3d 473 (3d Cir. 1994), the court
explained that, although a guilty plea cannot be used to
establish a co-conspirator's guilt, it can be introduced for
some valid purpose. After enumerating the proper purposes
for which the pleas could be admitted, the trial judge found
that the probative value of the evidence outweighed its
prejudicial effect and admitted it, denying defendants'
motion. The defendants assert that this ruling was
erroneous and that the error was not harmless, arguing
that "there is a high probability that the jury's conviction
on Count One -- in light of its acquittal verdicts on close to
40 other counts -- was influenced by the Bonjo plea to
Count One."5

The decision to admit or exclude the evidence is
committed to the sound discretion of the district court. In
re Merritt Logan Inc., 901 F.2d 349, 359 (3d Cir. 1990). We
use an abuse of discretion standard to review the District
Court's decision not to exclude evidence under Rule 403.
United States v. Gaev, 24 F.3d 473, 476 (3d Cir. 1994).

We note at the outset the well-settled proposition that
plea agreements of co-conspirators cannot be used as
evidence of a defendant's guilt. Gaev, 24 F.3d at 476. We
stated in United States v. Gambino, 926 F.2d 1355 (3d Cir.
1991) that:
_________________________________________________________________

5. In her testimony, Blum Bonjo stated that she had pled guilty to
"Count One."

15
 [t]here are strong considerations against using a co-
conspirator's guilt as substantive evidence of another
defendant's guilt. "The foundation of [this] policy is the
right of every defendant to stand or fall with the proof
of the charge made against him, not against somebody
else . . . . The defendant ha[s] a right to have his guilt
or innocence determined by the evidence presented
against him, not by what has happened with regard to
a criminal prosecution against someone else.

926 F.2d at 1363 (quoting in part Bisaccia v. Attorney
General of New Jersey, 623 F.2d 307, 312 (3d Cir. 1980)).

However, we have recognized certain valid purposes for
which co-conspirators' guilty pleas can be admitted. See,
e.g., Gaev, 24 F.3d 476-77 (enumerating proper purposes).
The District Court here explained that a plea was
admissible for a number of valid purposes. First, the court
explained that a plea was admissible to assist the jury in
assessing the credibility of the testifying witness. As we
explained in Gambino, "by eliciting the witness' guilty plea
on direct examination, the government dampens attacks on
credibility, and forecloses any suggestion that it was
concealing evidence." 926 F.2d at 1363. But, defendants in
their pre-trial motion pledged not to challenge Blum Bonjo's
and Martin's credibility based on their plea agreements.
Thus, there was no reason to allow the government this
pre-emptive strike on direct examination. See United States
v. Thomas, 998 F.2d 1202, 1205 (3d Cir. 1993).

The District Court also justified its ruling on the motion
in limine after the trial by referring to defendants' cross-
examination of Blum Bonjo. The court explained that
defendants challenged critical aspects of Blum Bonjo's
participation in the activities that formed the basis for the
mail fraud charge. One cannot, however, use defense
counsel's behavior after denial of the motion as any
indication of what he would have done had his motion been
granted by the court.

The District Court also explained that the pleas were
admissible to make clear that defendants were not being
singled out for selective prosecution. Indeed in United
States v. Inadi, 790 F.2d 383 (3d Cir.), rev'd on other

16
grounds, 106 S.Ct. 1121 (1986), we allowed the admission
of a co-conspirator's guilty plea "in order to rebut defense
counsel's persistent attempts on cross-examination to raise
an inference that the co-conspirators had not been
prosecuted, and that [the defendant] had been singled out
for prosecution." Id. at 384, n.2. But that is not the case
here. Defense counsel in his motion in limine pledged not
to engage in questioning on cross examination that would
render the pleas admissible. Certainly, if the defense had
raised an inference of selective prosecution, this tactic
would have been covered by counsel's pledge. However,
selective prosecution never became an issue in this case. In
Thomas, where selective prosecution was also not an issue,
we explained that "there was no need to mention the guilty
pleas to deter any concern that Thomas was being singled
out for prosecution." 998 F.2d at 1205. Moreover, even if
the implication of selective prosection had become an issue,
it could have been dealt with by instructing the jury that
the issue of selective prosecution was not their concern. Id.

The trial court suggested a further reason to admit the
plea -- to explain the witnesses' firsthand knowledge of the
defendant's misdeeds. See United States v. Halbert, 640
F.2d 1000, 1005 (9th Cir. 1981). But a witness's guilty plea
doesn't establish her basis of knowledge. The witness's
testimony itself will establish that basis.

After the valid purposes for which the plea may be
admissible are proffered, the trial court must balance the
probative value of the testimony with its prejudicial effect.
The standard for the balancing is that of Federal Rule of
Evidence 403.6 That is, the prejudice that inures to the
defendant by the admission of a co-conspirator's guilty plea
may be overcome by its probative value. Furthermore, in
some circumstances, a limiting instruction against the use
of the plea as substantive evidence of defendant's guilt, as
_________________________________________________________________

6. Fed. R. Evid. 403 provides:

Although relevant, evidence may be excluded if its probative value
is
substantially outweighed by the danger of unfair prejudice,
confusion of the issues, or misleading the jury, or by
considerations
of undue delay, waste of time, or needless presentation of
cumulative evidence.

17
was given in this case, reduces that prejudice. United States
v. Werme, 939 F.2d 108, 114 (3d Cir. 1991). However, in
Gaev, we recognized that "there may be cases where the
inference of guilt from the co-conspirator's plea agreement
is sufficiently strong that even limiting instructions will not
effectively contain it." 24 F.3d at 478. This is one of those
cases.

Notwithstanding the limiting instruction here, the jury,
having heard evidence accusing defendants of 39 counts of
mail fraud and false claims, acquitted all 3 defendants on
38 of the counts. The only count of conviction was the one
to which the government's star witness, Blum Bonjo, a mid-
level employee, had pled guilty. In addition, the patient
involved in Count One, Mildred Hynds, was also involved in
four other counts on which defendants were acquitted. It
would appear then that whether a particular patient needed
treatment, at least in the Hynds case, was not a
determining factor in the jury verdict. Moreover, the
government has conceded in its cross-appeal that the
conduct charged in the other 38 counts was virtually
identical to the conduct charged in Count One. The
evidence cited by the government, in arguing that there is
sufficient evidence to support the convictions on Count
One, is for a large part the same evidence that the jury
heard and then found not to support conviction on 38 of
the counts. There is, therefore, the strong possibility that
Blum Bonjo's guilty plea to that count was considered by
the jury as direct evidence against the defendants.

Concerning the probative value of the pleas, as we have
described above, the proffered reasons for admitting the
evidence were weak. Therefore, notwithstanding the
discretion due a district court judge making an evidentiary
determination, we find that the probative value of the pleas
was substantially outweighed by the danger of unfair
prejudice. Despite the fact that we have found sufficient
evidence to support Lukesh and Universal's convictions, the
evidence on the count of conviction was very similar to the
evidence supporting the thirty-eight acquitted counts. In
view of the fact that the only distinguishing evidence on the
count of conviction is the consideration of the co-
conspirators' guilty pleas, we must conclude that the denial

18
of defendants' motion in limine and the subsequent
introduction at trial of evidence of Blum Bonjo's and
Martin's guilty pleas was reversible error. We will reverse
the judgment of conviction as regards Lukesh and Universal.7

CONCLUSION

For the reasons stated above, we will reverse the
conviction of Attila Horvath for the insufficiency of the
evidence against him and remand his case to the District
Court for entry of a judgment of acquittal. We will reverse
the convictions of defendants Lukesh and Universal and
remand their cases to the District Court for a new trial.
_________________________________________________________________

7. As regards Horvath, our holding that the post-verdict judgment of
acquittal should have been granted renders moot his appeal of the denial
of the motion in limine.

19
GARTH, Circuit Judge, dissenting:

The majority of the panel (my colleagues Judges Roth
and McKee) have reviewed the record and reversed the
conviction of the defendant Attila Horvath ("Horvath"), who
had been charged with mail fraud in conducting a Medicare
fraud scheme over a three-year period.1 In doing so, they
have directed Horvath's acquittal on Count One2 despite the
credible evidence found against him by the jury and despite
the District Court's application of the undeniably correct
standard, under United States v. Glasser, 315 U.S. 60
(1942), which requires affirmance of a conviction so long as
the verdict is supported by substantial evidence, with all
reasonable inferences viewed in a light most favorable to
the Government. That standard, to which this Court
adheres, requires an affirmance of the defendants'
convictions on this appeal.

A.

In light of the majority's rejection of the jury's verdict,
which is tested by the Glasser standard just adverted to,
and in light of an even more erroneous holding that
completely eviscerates the deference which we are required
to give to a District Court's discretionary ruling to admit
pleas of codefendants,3 I must respectfully dissent.
_________________________________________________________________

1. The damage from Medicare fraud has been extensive. See, e.g.,
Medicare Contractors Aren't Pursuing Fraud, Audit Shows, USA TODAY,
Dec. 2, 1998, at A1; Probers Allege Medicare Fraud by Columbia/HCA,
CHICAGO TRIBUNE, Feb. 11, 1998, at N3 (disclosing investigation of
nationwide Medicare fraud by large health care organization); Fraud and
Waste in Medicare, N.Y. TIMES, Aug. 1, 1997, at A30 (stating
government estimates approximately $23 billion per year lost to Medicare
fraud).

2. Count One charged that Horvath, Lukesh and Universal violated 18
U.S.C. section 1341, which proscribes use of the mails in furtherance of
a fraudulent scheme.

3. See General Elec. Co. v. Joiner, 522 U.S. 136 (1997) (recognizing, in
context of expert testimony, that trial courts have discretion to admit
such testimony, and rejecting standard that "fail[s] to give the trial
court
the deference that is the hallmark of abuse of discretion review"). A long
line of cases from this Court has respected the District Court's
discretion

20
B.

Indeed, I dissent not only from the majority's insistence
on refusing the admission of the guilty pleas of
codefendants Penny Martin and Judy Blum Bonjo on the
record presented here, but I strongly urge the full court to
grant en banc consideration to what can only be
characterized as an aberration in our jurisprudence-- an
aberration dictated in the first instance by the suspect
holding in United States v. Thomas, 998 F.2d 1202 (3d Cir.
1993). In Thomas, in the face of a compelling and reasoned
dissent by Judge Rosenn of this Court, our own precedents,
decisions from our sister Circuits,4 and fully documented
and detailed discretionary rulings by the District Court
Judge, the same distorted result as to the exclusion of a
plea agreement gave rise to the same miscarriage of justice
that we now see here.

Our court -- the full court -- should undertake to clarify
this significant aspect of our criminal jurisprudence by
vacating the instant majority opinion, by rehearing the
issue en banc, and by reaffirming the standard that Judge
Scirica of this Court found appropriate in Gaev . It should
re-align the Third Circuit on the correct and proper course,
which provides for the acceptance of guilty pleas at trial by
codefendants when the correct standard of a District
Court's discretion is applied to the facts and when the
proper purposes of credibility, selective prosecution and
establishing firsthand knowledge are found, as they were
unquestionably applied by the District Court and found
here. Such a course would comport with the jurisprudence
_________________________________________________________________

in admitting pleas of codefendants. See United States v. Gaev, 24 F.3d
473 (3d Cir.), cert. denied, 513 U.S. 1015 (1994); United States v.
Gambino, 926 F.2d 1355 (3d Cir.), cert. denied, 502 U.S. 956 (1991);
United States v. Newman, 490 F.2d 139 (3d Cir. 1974); United States v.
Gullo, 502 F.2d 759 (3d Cir. 1974). Indeed, this Court has held
repeatedly that admission of a plea, without more, is not ordinarily
reversible. See, e.g., United States v. Restaino, 369 F.2d 544 (3d Cir.
1966).

4. See, e.g., United States v. Tse, 135 F.3d 200 (1st Cir. 1998); United
States v. Casto, 889 F.2d 562 (5th Cir. 1989), cert. denied, 493 U.S.
1092 (1990); United States v. Louis, 814 F.2d 852 (2d Cir. 1987).

21
of our sister Circuits that have wisely, and in accord with
prevailing legal standards, accepted such evidence.

Based on the record before us, I believe (1) that the
evidence was sufficient to support Horvath's conviction, and
(2) that the District Court did not abuse its discretion in
admitting the codefendants' guilty pleas, and that the guilty
pleas were admitted for a proper purpose and did not
invalidate the convictions of all the defendants. Therefore,
I would hold the majority's disregard of the District Court's
rulings and its reversal of those convictions is unwarranted
and violates both the District Court's and this Court's
proper standards of review.

C.

Although not addressed in the majority opinion, see
Majority Op. at 6 n.3, I further conclude that pursuant to
the teaching of United States v. Watts, 519 U.S. 148 (1997),
the District Court grievously erred in not considering
acquitted and uncharged conduct pursuant to Section
2F1.1 of the Guidelines in calculating the Sentencing
Guidelines range of each of the defendants.

Accordingly, I dissent from the majority's decision to
remand Horvath's case to the District Court to enter a
judgment of acquittal, and to reverse the convictions of
Lukesh and Universal and remand for a new trial on Count
One.

I

As the majority opinion notes, see Majority Op. at 6,
Horvath, Lukesh and Universal appealed the denial by the
District Court of their post-conviction motion for a
judgment of acquittal, or in the alternative, a new trial.
These defendants had moved for relief on the basis of
insufficiency of the evidence and claimed that the
admission of the codefendants' guilty pleas were unduly
prejudicial.

Prior to that motion, however, these defendants had
made an in limine motion to exclude testimony concerning
the plea agreements entered into by the codefendants. The

22
District Court took the matter under advisement, accepted
submissions by the parties and heard argument as to the
applicable legal principles. The District Court, in its
discretion, denied the motion, concluding that

if [Bonjo and Martin] testify the jury is going to
certainly wonder whether or not they have been
charged. It's going to wonder perhaps what they have
been promised by the prosecutor if anything and what
they may be getting in return for their testimony.

I think in weighing all of those factors with the possible
prejudice that I am going to allow the Government to
bring out the fact of the guilty plea and the fact of the
guilty plea agreement.

The District Court considered the issue again after the
conclusion of the trial. In denying these defendants' post-
conviction motion, the District Court first noted that, in a
motion challenging the sufficiency of the evidence, the court

must view the evidence in the light most favorable to
the government. A claim of insufficiency places a very
heavy burden on the [defendants]. [The court] must
affirm the convictions if a rational trier of fact could
have found defendant[s] guilty beyond a reasonable
doubt, and the verdict is supported by substantial
evidence.

The mail fraud statute, 18 U.S.C. S 1341, proscribes
any "scheme or artifice to defraud" in which the
defendant participated with the specific intent to
defraud and in which the mails were used "in
furtherance of the fraudulent scheme." The scheme
"need not be fraudulent on its face but must involve
some sort of fraudulent misrepresentations or
omissions reasonably calculated to deceive persons of
ordinary prudence and comprehension." Proof of
specific intent is required, which "may be found from a
material misstatement of fact made with reckless
disregard for the truth." United States v. Coyle, 63 F.3d
1239, 1243 (3d Cir. 1995) (citations omitted). Thus, a
mail fraud conviction stands where the evidence
demonstrates a defendant's willful participation in a
scheme to defraud with knowledge of its fraudulent

23
 nature. See United States v. Pearlstein, 576 F.2d 531,
537 (3d Cir. 1978).

In denying the motion, the District Court conducted a
careful review of the evidence indicating a crime had been
committed, and the evidence connecting each defendant to
the crime. After outlining the testimony and documentation
connecting Horvath and the other defendants to the crime
charged in Count One, the District Court summarized the
evidence as follows:

In the instant case, the individual defendants' guilty
knowledge could rationally be inferred beyond a
reasonable doubt from (1) the active parts Horvath and
Lukesh played in the daily operation of the speech
therapy practice, (2) [d]efendants' knowledge and
monitoring of Universal's billing problems -- the
number of patient cases being billed and met with
record requests, denials and increased scrutiny by
Blue Cross . . . (3) the testimony of key administrators,
Julia Blum Bonjo, Penny Martin and Wendy Gold,
recalling how Horvath and Lukesh wanted patient
documentation billable no matter what had to be done
to it, and (4) the directives given to speech therapists
by their supervisors, under tremendous pressure from
Lukesh and Horvath, to pick up patients, even though
inappropriate, to warrant skilled services, and to
rewrite patient documentation to make patient services
"billable" and reimbursable by Medicare.

The District Court found this evidence sufficient to support
the jury's finding that a crime had been committed and that
Horvath and Lukesh (and, through their actions, Universal)
had the intent required under the mail fraud statute.
Neither the appellants nor the majority have given any
reason to question that conclusion by the District Court or
the jury's verdict.

The District Court next addressed the issue of admission
of the plea agreements. It reiterated its concern over
selective prosecution, and noted as well that the testimony
from Bonjo and Martin would assist the jury in assessing
credibility. I believe that the majority has artfully finessed
the issue of admitting the codefendants' guilty pleas,

24
without ever providing reasons or presenting evidence that
the District Court had abused its discretion - indeed,
without ever even holding that the District Court Judge did
so. In so doing, the majority has not dealt fairly with the
jurisprudence dealing with this subject, and has ignored a
completely developed record. I thus turn first to the District
Court's rulings admitting the guilty pleas. I do so observing
that the admission of the guilty pleas at trial does not
impact upon the other, independently sufficient evidence
against the defendants.

II

The panel majority has reversed the Count One
convictions of Lukesh and Universal because of the putative
prejudicial effect that the introduction of Bonjo's and
Martin's guilty pleas had at trial. Although the majority
does not address this issue with respect to Horvath, it is
obvious that the admission of the guilty pleas did not affect
the other evidence presented to the jury to support
Horvath's conviction. As the defendants were acquitted of
thirty-eight (38) counts of the indictment but were
convicted of the very count to which Bonjo pleaded guilty,
defendants contend -- and the majority concurs-- that the
guilty pleas were not admitted for a proper purpose and
were so prejudicial that their admission constitutes
reversible error. Majority Op. at 17-19.

I strongly disagree. In my view, the pleas of both Bonjo
and Martin,5 after careful consideration and discussion in
accordance with our requirements set forth in Gaev, were
properly admitted. Even if not properly admitted, the
evidence adduced at trial, combined with the curative
instructions given by the District Court, rendered any error
harmless. Indeed, the majority admits as much when it
declares that the evidence against Lukesh and Universal is
sufficient to support the jury's verdict. Majority Op. at 13-
14.
_________________________________________________________________

5. Bonjo was charged in the same indictment as Horvath, Lukesh and
Universal. Martin, it appears from the record, was charged separately by
information and pled guilty to one count of mail fraud.

25
The District Court, at the time of both Bonjo's and
Martin's testimony, gave a curative instruction, advising the
jury that the purpose for admitting testimony concerning
their respective plea agreements was so that they could
"adequately assess the credibility" of the testimony. The
jury was also instructed that the jury was not to consider
the plea agreements as evidence of the guilt of Horvath,
Lukesh or Universal. The District Court, while charging the
jury at the end of the trial, instructed the jury again, as
follows:

Julia Blum Bonjo and Penny Martin entered into plea
agreements with the Government. Such plea
agreements are expressly approved as lawful and
proper by the United States Supreme Court and are
appropriate, are proper. Each witness' decision to plead
guilty is a personal decision about her own guilt. You
may not consider this evidence against the defendants
on trial nor may you draw any conclusions or inferences
of any kind about the guilt of the defendants on trial
from the fact that a prosecution witness pled guilty to
similar charges.

The testimony of such witnesses, as I indicated, should
be scrutinized with caution and give it the weight that
you think should be given under all of the
circumstances.

And I indicated to you during the trial that the fact
that [Bonjo and Martin] entered pleas of guilty could
not be considered by you in determining the guilt or
innocence of any of the people on trial here. The only
reason the plea and the plea agreement were brought
out was so that you would know all of the
circumstances surrounding the entry of the plea, you'd
know the terms under which the plea was entered and
you could judge for yourselves whether the witness in
the trial is testifying truthfully or whether the witness
has a motive to embellish testimony or vary from the
truth.

That is the only basis or the only reason why the plea
and the plea agreement were admitted.

26
(Emphasis added.) This Court reviews a District Court's
decision to admit evidence of plea agreements for abuse of
discretion, and we have been directed in no uncertain
terms to defer to that discretion. See Joiner, 522 U.S. at
136; Gaev, 24 F.3d at 476.

Although a co-conspirator's guilty plea cannot be used as
substantive evidence of a defendant's guilt, evidence of a
guilty plea or plea agreement may be introduced for other
permissible purposes. Judge Rosenn, in a cogent analysis
of this issue in United States v. Thomas, included as proper
purposes "(1) to bolster the credibility of the co-conspirators
as prosecution witnesses; (2) to quell the inference that the
co conspirators were not punished and that [the
defendants] w[ere] thus `singled-out' for punishment; and
(3) to establish the basis for the co-conspirators'firsthand
knowledge of the crime about which they testified. Each of
these is a proper purpose for admitting a guilty plea." 998
F.2d at 1208 (Rosenn, J., dissenting). See also Gaev, 24
F.3d at 476; Gambino, 926 F.2d at 1363; United States v.
Werme, 939 F.2d 109, 113 (3d Cir. 1991), cert. denied, 502
U.S. 1092 (1992); United States v. Inadi, 790 F.2d 383, 384
n.2 (3d Cir. 1986). To this list should be added the
overarching principle that trial judges have broad discretion
to admit testimony under Rule 403 that "discloses the
purpose, knowledge or design of a particular person."
Glasser, 315 U.S. at 80; cf. Joiner, 522 U.S. 136
(recognizing, in context of expert testimony, that trial courts
have discretion to admit such testimony, to which this
Court must defer). The mere admission of a plea is not
ordinarily reversible. See, e.g., Restaino, 369 F.2d 544.
Other Circuits have also held that admission of a plea by a
co-conspirator is appropriate for purposes other than to
persuade a jury of a defendant's guilt. See, e.g., Tse, 135
F.3d 200; Casto, 889 F.2d 562; Louis, 814 F.2d 852.

In Gaev, this Court summarized that the underlying
principle concerning the admissibility of a plea agreement
as follows: "If a co-conspirator who appears as a witness
has pleaded guilty, the trier of fact should know about the
plea agreement in order properly to evaluate the witness's
testimony, unless that would unduly prejudice the
defendant." 24 F.3d at 476 (emphasis added). Moreover, as

27
the credibility of the Government's witnesses normally is
critical to the successful prosecution of a case, this Court
has acknowledged that the strategic admission of a guilty
plea "dampens attacks on credibility, and forecloses any
suggestion that [the Government] was concealing evidence.
Such disclosure is appropriate." Gambino, 926 F.2d at
1363.

Here, the District Court permitted the introduction of the
guilty pleas because of credibility concerns that might arise
in the minds of the jurors as to Bonjo's and Martin's
testimony. In addition, the District Court determined that
the admission of the guilty pleas would allay the jurors'
concerns -- that might arise in the absence of such
admission -- that Horvath and Lukesh were subjected to
selective prosecution. In making these rulings, the District
Court, following this Court's direction in Gaev, balanced the
probative and prejudicial impact of the pleas. The District
Court found that, as in Gaev, credibility was an issue
because the witnesses' testimony was challenged, and that
the limiting instructions cured any prejudice. Even more so
than in Thomas, here the District Court found "specific
issues of credibility" that warranted admission of the pleas.

These determinations were well within the District
Court's discretion, and this Court must (and I emphasize
the word must) accord the proper deference such rulings on
admissibility by the District Court. Cf. Joiner, 522 U.S. 136
(finding discretionary decisions of trial court must be
upheld in absence of manifest error).

A fair reading of the record reveals that the District Court
admitted the pleas for a proper purpose, and the
Government's limited use of the guilty pleas at trial caused
no undue prejudice that would not have been corrected by
the District Court's curative instructions. This is not a case
in which the prosecution placed "undue emphasis" upon
the pleas, Restaino, 369 F.2d at 545, to convince the jury
of the defendants' guilt. Nor is this a case in which the
curative instruction failed to advise the jury not to consider
the admission of guilt by the witness against the
defendants. See Newman, 490 F.2d 139; Gullo, 502 F.2d
759. Here, another co-defendant (Vicki Meitus) was
acquitted of the charge alleged in Count One. This fact,

28
when combined with the curative instruction, more than
amply supports the conclusion that admission of the pleas
was not an abuse of discretion. See Restaino, 369 F.2d at
546.

Indeed, the majority at no point claims that the District
Court Judge abused his discretion -- nor could it in light
of the careful and detailed consideration that Judge Kelly
gave to the evidence in accordance with this Court's
direction in Gaev. See Appendix 1768-72; Memorandum of
May 31, 1996 at 16-20. Nor does the majority point to any
evidence in the record that could support a determination
that the District Court abused its discretion.

The only reason I can discern that the majority gives in
holding that the guilty pleas were not admitted for a proper
purpose was because Horvath, Lukesh and Universal
pledged not to raise the issue of the accomplices' guilty
pleas. Majority Op. at 16-17. Thus, the majority concludes
that there was no need preemptively to bolster the
credibility of the witnesses with the admission of their
guilty pleas because the defendants would not have
attacked Blum Bonjo and Martin on credibility grounds.

In so holding, however, the majority totally ignores prior
directions of this Court that guilty pleas can be admissible
even in the absence of an attack on a witness' credibility.
See Gambino, 926 F.2d at 1363 ("[E]ven in the absence of
this attack [on the Government witness' credibility], the
elicited testimony [i.e., the guilty plea] was proper here.")
See also Thomas, 998 F.2d at 1208 (Rosenn, J., dissenting).
As Gambino has not been overruled, it remains the rule in
this Circuit. The absence of an attack on a witness's
credibility is simply insufficient for a court to find that a
judge abused his discretion in admitting the plea; here, the
District Court cited Gambino in its post-trial ruling for
precisely this proposition. To hold otherwise, as the
majority does, not only presents a conflict with Gambino,6
but it would for all time foreclose the government's
admission of evidence of a guilty plea on direct examination
of its witnesses by the preemptive promise by defense
_________________________________________________________________

6. See Internal Operating Procedure 9.1 ("[N]o subsequent panel
overrules the holding in a published opinion of a previous panel").

29
counsel not to question the witness concerning the plea.
That rule is not the rule of this Circuit. See Gaev, 24 F.3d
at 477-78 ("While plea agreements have often been
admitted in response to actual or anticipated attacks on a
witness's credibility, an attack is not always necessary to
justify their introduction") (emphasis added).

I urge that the Court consider this issue en banc because
the majority opinion here -- as the majority opinion did in
Thomas, which concerned an almost identical set of facts --
in effect rules out the use of any guilty plea, without regard
to the discretion of the District Court, so long as defense
counsel promises not to question the witness concerning
the plea. Thomas and the analysis of the majority in the
present case, when seen through the lens of Gaev,
Gambino, Newman, Gullo and Restaino, makes a mockery of
the holdings in those cases. To resolve the conflict between
these precedents and Thomas and this case, the Court as
a whole should confront this issue.

III

The majority has concluded that the evidence presented
at trial was insufficient to support the conviction of
Horvath, Majority Op. at 12, although there was sufficient
evidence to support the convictions of Lukesh and
Universal. Majority Op. at 13-14. My reading of the record
leads me to a contrary conclusion with regard to Horvath.
Because I conclude that the jury's verdict can be easily
sustained by the evidence produced at trial, and that the
admission of Bonjo's and Martin's guilty pleas does not
affect this conclusion, I would affirm the conviction of
Horvath.

The District Court enunciated and applied the correct
standard in reviewing a challenge to a jury verdict based on
insufficiency of evidence. In reviewing a jury verdict, the
court "must view the evidence in the light most favorable to
the jury verdict and presume that the jury verdict properly
evaluated credibility of the witnesses, found the facts, and
draw rational inferences." United States v. Iafelice, 978 F.2d
92, 94 (3d Cir. 1992). A conviction must be sustained if the
verdict is supported by substantial evidence, and all

30
reasonable inferences must be viewed in a light most
favorable to the Government. Id. This Court must simply
determine whether "the conclusion chosen by the fact
finders was permissible." United States v. McGill, 964 F.2d
222, 230 (3d Cir. 1992), cert. denied, 506 U.S. 1023 (1992).
If there is substantial evidence to support the verdict, we
will not reverse even though this Court may have decided
the case differently. United States v. Sain, 141 F.3d 463,
470 (3d Cir. 1998).

Despite the majority's emphasis upon the fact that much
of the evidence that the District Court relied upon to link
Horvath to the fraudulent scheme did not directly relate to
the Hynds incident charged in Count One, the elements of
mail fraud do not necessitate that a defendant participate
in every act executed in furtherance of that scheme. As the
District Court noted, relying on Pearlstein, 576 F.2d at 541,
the jury can infer the requisite intent from circumstantial
evidence.

To obtain a conviction under section 1341, the
Government must establish (1) the existence of a scheme to
defraud; (2) the participation by the defendant in the
particular scheme charged with the specific intent to
defraud; and (3) the use of the United States mails in
furtherance of the fraudulent scheme. See United States v.
Hannigan, 27 F.3d 890, 892 (3d Cir. 1994).

Here, the evidence showed that Horvath, the Director of
Finance, and Lukesh, the Director of Operations, actively
participated in the daily operations of Universal and worked
closely together. Both knew that Universal was having
difficulty with Independence Blue Cross in getting its
speech therapy services reimbursed. In addition, the
testimony from the Universal administrators (i.e., Martin,
Blum Bonjo, and Wendy Gold) indicated that Horvath
and Lukesh created a "coercion" culture in which pressure
was placed upon Universal's employees to obtain
reimbursement in any way possible.

Specifically, Gold testified that Horvath and Lukesh had
placed her under enormous pressure to increase
production, which included pressure to rewrite the patient
documentation. Blum Bonjo testified that she met with

31
Horvath and Lukesh together weekly for discussions about
cases in which Medicare denied reimbursement for the
speech services rendered and also testified that Universal
handled non-billable cases by rewriting patient
documentation to make such treatment billable.

Similarly, Martin testified that Horvath said that "he
wanted the documentation billable no matter what had to
be done to it." Some of the cases were put on hold (meaning
that they would not be submitted for billing) because they
did not seem appropriate for billing to Medicare. Martin
testified that when the number of claims on hold were high,
Horvath instructed that the claims be billed without
doctors' signatures. Further, Martin stated that the
documentation had to be rewritten and altered and
originals had to be destroyed in order to avoid
discrepancies in the documentation. Finally, the District
Court reviewed the evidence presented at trial and found
that Horvath was very active in the daily operations of
Universal from the outset of the scheme, and closely
monitored the results of the speech therapy practice.
Horvath, through his approval of Universal's "rewriting"
policy, sanctioned the fraudulent scheme alleged in Count
One, for which he was found guilty.

In my view, there is no question that the jury's verdict
linking Horvath to the fraudulent scheme is supported by
the record. There was sufficient evidence to convince the
jury, and this Court should not "weigh evidence or
determine the credibility of witnesses." United States v.
Casper, 956 F.2d 416, 421 (3d Cir. 1992) (quoting Glasser,
415 U.S. at 80). Reversal for insufficiency of evidence
should not be granted except where the failure of the
prosecution is "clear." Id. That situation is not present here.
See also United States v. Anderson, 108 F.3d 478, 480-81
(3d Cir. 1997) ( "Only when the record contains no
evidence, regardless of how it is weighed, from which the
jury could find guilt beyond a reasonable doubt, may an
appellate court overturn the verdict"); United States v.
McGlory, 968 F.2d 309, 321 (3d Cir.), cert. denied, 506 U.S.
956 (1992).

The question, then, for the jury to answer was whether
the evidence sufficiently established that Horvath was

32
connected to the conduct charged in Count One. The jury
determined that there was, and the record more than fairly
supports that determination.

IV

As the majority has reversed the convictions on all three
defendants, its opinion does not discuss the sentence
imposed by the District Court, which was the subject of the
government's cross-appeal. Although I would affirm the
respective convictions of Horvath, Lukesh and Universal, I
would reverse the District Court's sentence calculation
under the Sentencing Guidelines ("the Guidelines"), as
presumably the issue of uncharged and acquitted conduct
can arise again in this case.

The majority opinion, oddly enough, does not reflect the
sentences imposed upon the defendants. Let me do so. The
District Court sentenced Universal to two years' probation
and a $25,000 fine. Horvath and Lukesh were sentenced to
three years of probation each. Horvath was fined $10,000,
plus $705.20 in restitution. Lukesh was fined $15,000,
plus restitution of $705.20.

Under Section 1B1.3 of the Guidelines, a District Court
can consider relevant conduct when calculating a
defendant's guideline range whether or not that conduct
was formally charged. See United States v. Watts, 519 U.S.
148, 152-53 (1997); see also U.S.S.G. S 1B1.3 cmt. 1 ("The
principles and limits of sentencing accountability under
this guideline are not always the same as the principles
and limits of criminal liability. [T]he focus is on the specific
acts and omissions for which the defendant is to be held
accountable . . . rather than on whether the defendant is
criminally liable.") In addition, under Watts, a defendant's
guideline range is affected even by acquitted conduct, as
long as that conduct has been proved by a preponderance
of the evidence. See Watts, 519 U.S. at 156. See also United
States v. Cianci, 154 F.3d 106 (3d Cir. 1998); United States
v. Baird, 109 F.3d 856 (3d Cir. 1997) (finding conduct
uncharged pursuant to plea agreement may be considered
at sentencing).

33
Notwithstanding the Guidelines and Watts, at sentencing,
the District Court declined to consider uncharged and
acquitted conduct when calculating defendants' guidelines
ranges. Instead, the District Court only considered the loss
associated with the conviction obtained in Count One--
totaling $1,411.20 -- without explaining why it was not
considering the uncharged and acquitted conduct that the
Government sought to be included. I believe that this
omission was an abuse of discretion.

While I conclude that the evidence proved that Horvath,
Lukesh, and Universal could be held accountable for the
relevant (i.e., uncharged or acquitted) conduct, at the very
least, the District Court should have made findings of fact
as to why it declined to consider that conduct in the
calculation of the sentences of Horvath, Lukesh and
Universal. See, e.g., E.C. Ernst, Inc. v. Koppers Co. Inc., 626
F.2d 324 (3d Cir. 1980) (remanding because District Court
failed to make factual findings and thus this Court could
not determine how District Court assessed evidence). The
District Court merely identified the standard under Watts
and its conclusion not to consider the relevant conduct
without stating its reasons, stating "I find under the Watts
case that the burden is preponderance of the evidence and
decline to include that conduct in the specific offense
characteristics."

Given that the District Court ruled in denying the post-
conviction motion that there was substantial evidence to
support the underlying conviction in Count One, I fail to
understand how the District Court could conclude that the
Government had not met its burden by a preponderance of
the evidence of Horvath, Lukesh, and Universal's
involvements in the other related counts of the very same
scheme to defraud.

The inclusion of uncharged and acquitted conduct, which
may have amounted to a total loss of $343,500, would have
resulted in an increase of eight levels under the Guidelines.
See U.S.S.G. S 2F1.1(I). Horvath had an offense level of 8,
providing a guidelines range of 0-6 months, and Lukesh
had an offense level of 10, providing a guideline range of 6-
12 months. If uncharged and acquitted conduct were
considered at sentencing, however, Horvath's sentencing

34
range could have increased to 21-27 months (i.e., to offense
level 16) and Lukesh's range could have increased to 27-33
months (i.e., offense level 18).

Accordingly, I would hold the defendants accountable at
sentencing for the uncharged and acquitted conduct with
which they were involved in executing their fraudulent
scheme.

V

In its opinion, the majority has reversed a jury verdict
founded on sufficient evidence and has reversed detailed
evidentiary rulings by the District Court. In doing so, the
majority has violated four appellate strictures: 1) it has
ignored our established standard of review under Glasser,
which requires that all reasonable inferences be resolved in
favor of the government in an appeal challenging sufficiency
of the evidence; 2) it has substituted its own "jury verdict"
for that of the enpanelled jury; 3) it has refused to give the
required deference to a District Court Judge's discretionary
rulings; and 4) it has perpetuated a jurisprudential conflict
over the admission of codefendant guilty pleas at trial.

The trial transcript and post-conviction order reveal that
the District Court carefully considered the arguments
raised by the defendants, and rejected them in accordance
with principles long established by prior panels of this
Court. Under these circumstances, when the issues raised
on appeal concern evidentiary issues addressed by the
District Court in a careful analysis that considered the
arguments for both sides, I am unable to agree that the
District Court abused its discretion -- a claim that not even
the majority justifies.

In sum, therefore, I cannot subscribe to the majority
opinion because I believe that the guilty pleas were
admitted for a proper purpose and because there was
sufficient evidence to support Horvath's conviction. The
evidence revealed an elaborate scheme of fraud on the
medical insurance system of this country, which, although

35
not acknowledged by the majority, is a continuing problem
that has cost our country dearly.7

I also conclude that the District Court erred in failing to
state its reasons for not considering the uncharged and
acquitted conduct of the defendants at sentencing.

Accordingly, I respectfully dissent.

A True Copy:
Teste:

Clerk of the United States Court of Appeals
for the Third Circuit
_________________________________________________________________

7. See note 1, supra.

36