Description: Diana Gumila ran a home-healthcare company that defrauded the federal government of several million dollars. She was convicted of multiple counts of healthcare fraud and making false statements in connection
∗ Circuit Judge Posner retired on September 2, 2017, and did not partici-pate in the decision of this case, which is being resolved by a quorum of the panel under 28 U.S.C. § 46(d).
2 No. 16-3111
with a healthcare matter. The district judge imposed a
below-guidelines prison sentence of 72 months followed by
24 months of supervised release.
Gumila appeals, raising several challenges to her sentence.
She first argues that the judge miscalculated the
financial loss attributable to her offenses. She also contends
that the 72-month prison term is substantively unreasonable.
Finally, she claims that the judge did not adequately explain
the term and conditions of supervised release. The first two
arguments are meritless. The third is waived. We affirm.
Diana Gumila was head of clinical operations for
Suburban Home Physicians, LLC, which did business under
the name “Doctor at Home.” The company employed doctors
and other medical personnel to provide home medical
care to the elderly in and around Chicago. Gumila was
indicted on 21 counts of healthcare fraud in violation of
18 U.S.C. § 1347 and three counts of making a false statement
in a healthcare matter in violation of 18 U.S.C. § 1035. The
indictment alleged that Doctor at Home (1) overbilled
Medicare for medical home visits; (2) billed Medicare for
unwarranted skilled-nursing services; and (3) billed
Medicare for care-plan oversight services that were never
At trial the government introduced testimony from more
than 20 witnesses and a trove of documentary evidence
establishing that Gumila played a central role in Doctor at
Home’s scheme to defraud the government. The evidence
showed that she regularly overruled physicians who wanted
to discharge patients from their care. She instructed nonphyNo.
sician employees to bill medical services at unjustifiably high
rates (a practice known as “upcoding”). She instructed
employees to claim that patients were homebound even
when they weren’t. And she instructed employees to process
orders authorizing skilled-nursing services even if the
attending doctor did not believe the patient qualified for that
service and even when no doctor had ever examined the
patient. A jury found her guilty on all counts.
Before the sentencing hearing, the government proposed
figures for three categories of financial loss suffered by
Medicare: (1) approximately $2.375 million for unnecessary
and upcoded home visits; (2) at least $9.45 million for
skilled-nursing services that did not meet Medicare’s requirements
and were unnecessary; and (3) $3.779 million in
claims for care-plan oversight services that did not qualify
for payment or were never performed.
In the presentence report (“PSR”), the probation officer
substantiated those figures for the three categories of loss
and estimated the total financial loss stemming from
Gumila’s unlawful conduct to be $15.6 million. The corresponding
guidelines range was 151 to 188 months in prison.
The probation officer recommended a below-guidelines
sentence of 84 months in prison and a 24-month term of
supervised release. The PSR also recommended 18 specific
conditions of supervision.
Gumila filed written objections to the PSR, challenging
the loss calculation and arguing that the loss should be
limited to Medicare payments for the eight patients specifically
mentioned in the indictment—for a total loss of only
$14,449. She argued for a prison sentence of 12 to 18 months.
She did not object to the recommended term or conditions of
4 No. 16-3111
supervised release. The government recommended a belowguidelines
sentence of 120 months in prison, a 24-month
term of supervised release, and $15.6 million in restitution.
At sentencing the judge concluded that the evidence established
an “overwhelming and massive scheme” to defraud
the Medicare program. He rejected Gumila’s argument
that the government was required to present specific evidence
to prove the fraudulent nature of each individual
transaction contributing to the total financial loss. He also
determined that the PSR’s loss estimate of $15.6 million was
reasonable. The judge imposed a sentence of 72 months in
prison (less than half the low end of the guidelines range)
and 24 months of supervised release. He also imposed the
18 conditions of supervision recommended by the PSR and
ordered Gumila to pay $15.6 million in restitution.
On appeal Gumila raises three arguments: (1) the district
judge erred in calculating the financial loss attributable to
her; (2) the 72-month prison term is substantively unreasonable;
and (3) the judge committed procedural error by failing
to explain the term and conditions of supervised release by
reference to the relevant factors listed in 18 U.S.C. § 3553(a).
A. Loss Calculation
We review the judge’s loss calculation deferentially and
will reverse only if we find clear error. United States v.
Littrice, 666 F.3d 1053, 1060 (7th Cir. 2012). Gumila must
show that the judge’s calculation “was not only inaccurate
but outside the realm of permissible computations.” Id.
(quoting United States v. Al-Shahin, 474 F.3d 941, 950 (7th Cir.
2007)). At sentencing the government bears the burden of
No. 16-3111 5
proving the loss amount by a preponderance of the evidence,
but a reasonable estimate will suffice. United States v.
Schroeder, 536 F.3d 746, 752 (7th Cir. 2008).
The judge determined that the government’s method for
calculating loss was both “supported by the evidence and …
quite compelling.” He noted that Doctor at Home employed
a routine set of procedures in its scheme to defraud the
government, most of which were illegal in themselves, and
that Gumila personally orchestrated those procedures.
Gumila attacks each category of loss individually, but she
also makes a general argument that the loss calculation
should be limited to the illicit Medicare payments associated
with the eight patients listed in the indictment.
The generalized argument requires little comment. The
judge’s task was to estimate total loss, and to do so he was
permitted to approximate by scaling up the evidence “to
reflect the scope of the loss involved.” United States v. Natour,
700 F.3d 962, 978 (7th Cir. 2012). The eight specific patients
listed in the indictment were merely representative of the
thousands of patients for whom Doctor at Home submitted
fraudulent claims that were subsequently paid by the government.
The judge was not required to limit the loss calculation
solely to those eight patients when evidence
established a far more sweeping overall fraudulent scheme.
See United States v. Sutton, 582 F.3d 781, 784 (7th Cir. 2009).
Gumila’s specific challenges to the separate categories of
loss fare no better. We take each one in turn.
1. Losses Attributable to Home Visits
Medicare will pay for home visits only if there is a documented
medical necessity for that type of care in lieu of an
6 No. 16-3111
office or outpatient visit. The government presented evidence
at trial that a vast number of the home visits performed
by Doctor at Home never qualified for Medicare
payment in the first place.1 Doctor at Home also regularly
and fraudulently upcoded its home visits.
To submit a home-visit invoice to Medicare, Doctor at
Home first had to code the visit. Home visits receive one of
four different billing codes based on the severity of the
medical problem addressed during the visit, the complexity
of the medical decision reached during the visit, the type of
care provided during the visit, and the length of time for the
visit. The more complex or demanding the visit, the larger
the bill. According to Medicare regulations, high-coded
visits are justified when the medical examination is “detailed”
or “comprehensive,” the medical decision-making is
moderately to highly complex, and the problems presented
by the patient are moderately to highly severe. Additionally,
the normal period of time spent with a patient for a highcoded
visit should be about 40–60 minutes. On the other
hand, visits warrant one of two lower billing codes when the
examination is “problem-focused,” the medical decisionmaking
is straightforward and not complex, the problem
1 For example, a doctor employed by Doctor at Home testified that 60%
of her home visits did not qualify for Medicare reimbursement. The
company’s medical director stated that “with each passing day, Doctor
[a]t Home is committing fraud by seeing patients who can drive, go to a
[primary care physician,] or walk out of the house unassisted.” A
number of patients testified that they left their home regularly even
though Doctor at Home claimed them as homebound, and emails
established that patients attempted unsuccessfully to remove themselves
from Doctor at Home’s rolls of homebound patients.
No. 16-3111 7
presented is low to moderately severe, and the amount of
time spent with the patient is about 15–25 minutes.
Doctor at Home billed nearly every home visit at the two
highest codes. But several employees testified at trial that the
vast majority of these visits, which were nearly all scheduled
to occur regularly on a monthly basis, were simple checkups,
not sick visits, and thus did not qualify for billing at
those rates. A memo drafted at Gumila’s command coached
Doctor at Home employees to cajole patients into maintaining
their regular schedule of home visits whenever they
tried to cancel them. Witnesses also testified that the home
visits were routine in nature and that if the visits qualified
for Medicare payment at all, they should have been coded at
the lowest level. And emails showed Gumila knew that at
least one doctor routinely paid only brief visits to the patients,
sometimes not even speaking to them during the visit,
and performed no examination beyond listening to their
hearts and lungs. Nonetheless, nearly all of that doctor’s
visits were billed at the highest code. Other emails showed
that Gumila instructed employees to use only the two highest
codes when billing the visits.
To estimate the losses attributed to these upcoded home
visits, the judge determined the amount that Medicare
would have paid had the visits been billed at a lower code
rate instead of at the top two rates. The difference totaled
$2.375 million. The judge’s calculation generously assumed
that each home visit qualified for some level of reimbursement
from Medicare, even though the evidence at trial
established that a large number of these visits did not qualify
at all. Thus, the calculation for this category was more
conservative than it might have been. See United States v.
8 No. 16-3111
Mikos, 539 F.3d 706, 714 (7th Cir. 2008) (holding that the loss
should not be discounted for value of services rendered
because the services did not qualify for Medicare payment in
the first place). We find no error in this approach.
2. Losses Attributed to Skilled-Nursing Services
Medicare reimburses for skilled-nursing services only if
the patient is homebound and requires such services. A
physician must sign an order requesting the service. Evidence
at trial showed that Doctor at Home trained nonphysician
employees to alter patient charts to make it appear
that the attending physician qualified the patient as homebound
and to delete information indicating that the patient
didn’t need nursing services. One former employee testified
that a Doctor at Home physician would sign a stack of
orders for skilled-nursing services without examining the
patients or even reading the patients’ files. Gumila also
regularly overruled doctors who tried to discharge nonhomebound
patients from nursing services. Finally, Gumila
herself authorized orders for nursing services for patients
whom no Doctor at Home physician had ever seen.
To estimate the loss attributable to the fraudulently billed
skilled-nursing services, the government focused on
Drs. Pauwaa and Newman, both of whom worked with
Doctor at Home during the relevant time period. Medicare
paid approximately $16.6 million for nursing services ordered
by Dr. Pauwaa and approximately $8.2 million for
nursing services ordered by Dr. Newman. Based on a review
of these payments along with other claims data showing that
many of those patients who received nursing services didn’t
qualify as homebound, the judge estimated that 40% of the
payments for services ordered by Dr. Newman and 43% of
No. 16-3111 9
the payments for services ordered by Dr. Pauwaa were
directed to patients who did not qualify as homebound. This
fraudulent billing totaled about $9.35 million.2
Though the judge limited his calculation of skillednursing
services to those requested by Drs. Pauwaa and
Newman, evidence established that many other employees
also fraudulently claimed patients as being homebound. In
other words, the judge’s estimate of loss for this category of
fraudulent billing was again on the conservative side.
Gumila has not identified any clear error in the judge’s
approach to estimating losses attributable to skilled-nursing
3. Losses Attributed to Care-Plan Oversight Services
Care-plan oversight services include physician supervision
of patients requiring complex or multidisciplinary care
and ongoing physician involvement. Medicare pays for
oversight services that take 30 minutes or longer to perform
so long as certain requirements are met, including the requirement
that the patient’s problems are complex enough
to require a doctor’s ongoing involvement in the patient’s
care plan. Evidence at trial established that Doctor at Home
employees in Illinois and the Philippines fabricated forms
claiming Medicare reimbursement for nonexistent oversight
services. Witnesses testified that employees scoured patient
files to find anything that might be passed off as a potentially
covered activity on the Medicare oversight-services form
and attributed time to that service. Moreover, these employ-
2 Dr. Newman also confirmed the amount attributed to him in his own
10 No. 16-3111
ees billed Medicare without ever confirming that a doctor
had spent time performing oversight services for that patient.
The judge estimated that Medicare paid $3.779 million to
Doctor at Home for these fraudulent care-plan oversight
services. The judge determined that all the Medicare invoices
in this category were fraudulent because there was no evidence
that any oversight services qualifying for Medicare
payment were ever performed. Gumila points out that at
least some employees filled out care-plan oversight billing
forms by referring to “services that had been documented as
performed in the charts.” But the notation alone does not
establish that oversight services were in fact performed,
were medically justified, or were accurately calculated
according to the amount of time actually spent on the patient
as required. Indeed, Gumila acknowledges that these employees
did nothing to verify what services (if any) had been
performed by the attending physicians. The judge did not
clearly err in determining that none of the payments for
care-plan oversight services were warranted.
Gumila’s final challenge is that the judge’s overall calculation
did not take into account the fair market value of the
services rendered. She argues that the patients received some
value from the doctors’ and nurses’ visits, which must be
reflected in any discounting of Medicare payments received
for those services. As we’ve noted, however, the judge did
account for the fair value of services actually rendered, but
only when the record arguably supported it. For medical
home visits, for example, the judge calculated the loss by
taking the difference between the amount that Doctor at
Home overbilled Medicare and the billing rates that more
No. 16-3111 11
accurately reflected the types of home visits that Doctor at
Home actually performed. For home nursing services, the
judge relied on evidence from two doctors that approximately
40% of the services ordered didn’t qualify for Medicare at
all. But for the care-plan oversight services, Gumila was
unable to establish that any of the services had even been
performed, let alone that they qualified for Medicare reimbursement.
B. Substantive Unreasonableness
Gumila next argues that her 72-month prison term is
substantively unreasonable. This is a steep uphill climb. Our
review is deferential, for abuse of discretion only. United
States v. Annoreno, 713 F.3d 352, 356–57 (7th Cir. 2013). A
sentence within a properly calculated guidelines range is
presumptively reasonable, United States v. Mykytiuk, 415 F.3d
606, 608 (7th Cir. 2005), but more to the point here, we have
“never deemed a below-range sentence to be unreasonably
high,” United States v. Wallace, 531 F.3d 504, 507 (7th Cir.
2008). Gumila’s 72-month sentence is less than half the low
end of the guidelines range of 151 to 188 months. She has
given us no good reason to overturn her sentence as unreasonably
C. Supervised Release Procedural Error
The judge imposed a 24-month term of supervised release
and 18 conditions of supervision as recommended in
the PSR. Gumila argues that the judge committed procedural
error by failing to adequately explain the length of the term
and the conditions of supervised release.
This argument is waived. The PSR gave Gumila written
notice of the proposed term and conditions of supervised
12 No. 16-3111
release (and the justifications for each condition) well in
advance of the sentencing hearing. The judge directed her to
respond in writing with any objections to the report. She did
so, but her memorandum challenged only the loss calculation
and the PSR’s suggested evaluation of the § 3553(a)
factors in relation to the recommended prison sentence. She
did not object to any of the supervised-release conditions or
the term of supervised release.
The sentencing hearing is the “main event,” and when
the court gives advance notice of the proposed term and
conditions of supervised release, the parties can “prepare
and identify issues they wish to address.” United States v.
Lewis, 823 F.3d 1075, 1083 (7th Cir. 2016). Advance notice
permits the defendant to “present an informed response” at
the hearing. United States v. Kappes, 782 F.3d 828, 843 (7th Cir.
2015). Here the PSR gave Gumila all the notice she needed to
make an informed objection to the proposed term and
conditions of supervised release. She did not do so. We’ve
held that a defendant’s “failure to object in th[ese] circumstances
can amount to waiver.” United States v. Gabriel,
831 F.3d 811, 814 (7th Cir. 2016) (citing Lewis, 823 F.3d at
1083–84); see also United States v. Bloch, 825 F.3d 862, 873 (7th
Cir. 2016) (stating waiver exists when there was no “lack of
notice or surprise at the conditions the district court planned
Gumila’s written response to the PSR challenged several
factors that would bear on the prison term and restitution
(e.g., the loss calculation), but she lodged no objection to the
proposed term or conditions of supervised release. That’s a
waiver, as we’ve recently held in a materially identical case.
No. 16-3111 13
See United States v. Ranjel, 872 F.3d 815, 821–22 (7th Cir.