United States v. Merigrace Orillo

ROVNER, Circuit Judge. On April 20, 2012, Merigrace Orillo pled guilty in a written plea agreement to one count of healthcare fraud pursuant to 18 U.S.C. §§ 1347 and 2, and one count of paying kickbacks to physicians for patient referrals under a federal health care program in violation of 42 U.S.C. § 1320a-7b and 18 U.S.C. § 2. She was sentenced to 20 months’ imprisonment. The district court determined that the loss amount for the healthcare fraud count was $744,481 and ordered her to pay that amount in restitution. On appeal, Orillo challenges the loss calculation that led to that sentence and restitution determination, and asserts that she has not waived a challenge to the sentence on the kickback count. Orillo, her husband, and a third person co-owned a business known as Chalice Health Services, Inc. (“Chalice”), which was a home health care provider. Orillo and her husband, a doctor, managed Chalice and supervised its daily operations. Chalice provided nurses, nurse aides, physical therapists, and occupational therapists to care for patients in the patients’ homes. Beginning in October 2004, Chalice was an enrolled provider with Medicare and could seek reimbursement of home health care through that program. Medicare restricted payments for home health care to those services that were medically necessary, which included only services required because of disease, disability, infirmity or impairment, to a homebound person. Prior to submitting a claim, the home health care provider, such as the treating nurse, with Chalice was required to complete a Comprehensive Adult Nursing Assessment with Outcome and Assessment Information Set (“OASIS”) form on the patient. The OASIS form was utilized to establish whether the patient was home-bound, the severity of the patient’s symptoms, and the reimbursement rate for Chalice. If a patient required services beyond a 60-day period, the Medicare program required Chalice to submit a Recertifieation/Follow-Up Assessment form (“Re-certification”) completed by that home health care provider to determine the patient’s continued eligibility for such services. Once the home care provider completed those forms, Chalice was required to enter that information into a software program that would determine the rate of reimbursement that Medicare would provide. By manipulating the information on those forms, Chalice could therefore impact the amount provided by Medicare. Orillo supervised the submission of those claims to Medicare, requiring that all forms be sent to her office and personally reviewing the OASIS and Recertification forms. In the plea agreement, she admitted that she falsified those forms by altering the codes and information on the forms that had been completed by the Chalice nurses to make the patient’s condition appear worse and the health care needs greater than the actuality. Those alterations caused the *243 Medicare software program to generate different reimbursement rates, which increased the reimbursement amounts paid to Chalice — a sequence also known as up-coding. Orillo acknowledged in the plea agreement that she made those alterations in two ways — by marking the changes on the forms submitted by the Chalice nurses, often forging their initials next to the alterations, and by replacing entire pages with falsified pages manufactured by Orillo herself. Orillo also admitted that she aided her husband in paying kickbacks to a Chicago doctor in return for referrals of Medicare patients. As to the healthcare fraud count, Orillo conceded that the loss to Medicare caused by her health care fraud scheme exceeded $400,000, and agreed to the entry of a $500,000 forfeiture judgment. The amount of restitution was left to the district court’s determination. Orillo now appeals the calculation of loss and the restitution amount as to that healthcare fraud count. The district court adopted the amount of $744,481 determined by the Probation Officer in the Pre-Sentence Investigation Report. That amount was determined after an extensive statistical analysis undertaken by Eric Vasiloff of Trust Solutions, LLC, which was a government contractor charged with auditing Chalice’s medical records to determine the loss to the Medicare program. Brian Cody, a registered nurse assigned to medical review for Trust Solutions, conducted a review of a random sample of the medical records for claims submitted by Chalice between January 1, 2007 and March 31, 2010. He reviewed records for 177 episodes of care delivered by Chalice to homebound Medicare patients out of 3,400 total episodes attributable to Chalice during that time period. Cody testified that he examined the records to determine whether the items on the OASIS forms were consistent with supporting documentation such as the plan of care, nurse’s notes, or therapy notes. Where the OASIS items were internally inconsistent or where they were inconsistent with that supporting documentation, Cody would determine what OASIS items were properly supported by those records and input the correct information. He then generated the correct Medicare codes and the corresponding reimbursement rates using those proper OASIS items. As a result of that analysis, Cody uncovered overpayments totaling $47,444 for those 177 episodes of care. He found no evidence of underpayments in that sample. In the spreadsheet that he compiled in his review, Cody separated the overpayments into two categories: those related to OASIS items on which alterations had been writteri, and those attributable to OASIS items in which no alterations were apparent but which were nevertheless inconsistent with the underlying documentation. At the outset, it is important to note what Orillo does not challenge in this appeal. She does not contest the reliability of the extrapolation based on established statistical methods, nor the random sample used. She does not contend that Cody improperly determined when the coding was inaccurate in determining overpayments, nor does she challenge the amount of overpayments that were found. In fact, she raises no challenge at all to the determination as to the amount of overpayments. Her sole argument is that the court erred in attributing all of the over-payments to criminal conduct, and that in assessing the loss and restitution amounts, the court should have relied only on over-payments relating to visibly altered items and not on any overpayments related to items not visibly altered. Orillo asserts that the government produced no evidence tying her criminal conduct with the over-payments that resulted from forms in which there were no apparent alterations. *244 According to Orillo, a certain amount of overpayment is natural and expected as an everyday occurrence as a result of human error, such as error in entering the data. She contends that those errors are not related to any fraudulent scheme, and thus the overpayments resulting from them should not be part of either the loss calculation or the restitution award. A loss determination must be based on the conduct of conviction and relevant conduct that is criminal or unlawful, and the government must demonstrate by a preponderance of the evidence that the loss amount is attributable to that criminal or unlawful conduct. United States v. Littrice, 666 F.3d 1053, 1060 (7th Cir.2012). That standard requires only that the fact-finder believe that the, existence of a fact is more probable than its non-existence, and for the purposes of determining the loss amount, a reasonable estimate is sufficient. Id. We review the district court’s finding of loss amount for clear error, and will reverse only if “ ‘based on the entire record, we are left with the definite and firm conviction that a mistake has been committed.’ ” Id., quoting United States v. Severson, 569 F.3d 683, 689 (7th Cir.2009). In other words, in order to reverse for clear error we would need to find that the “district court’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). The standards applicable to restitution awards are slightly different. The amount of restitution is limited to the actual losses caused by the specific conduct underlying the offense, and, like the loss amount, the government must establish that by a preponderance of the evidence. United States v. Kennedy, 726 F.3d 968 (7th Cir.2013). We review the district court’s determination of the restitution amount for abuse of discretion, viewing the evidence in the light most favorable to the government. Id.; United States v. Robers, 698 F.3d 937, 941 (7th Cir.2012). Orillo challenges both the loss and restitution amounts, but her argument is focused almost entirely on the loss amount, with restitution mentioned only a handful of times. Although the restitution analysis differs from that of the loss amount, in this case Orillo raises one challenge applicable equally to both determinations, and therefore we address them together. See United States v. Ali 619 F.3d 713, 720 (7th Cir.2010) (analyzing challenges to the loss calculation, forfeiture order, and restitution award together because the challenges to the three were the same). Orillo does not challenge the calculation of the overpayment amounts; she solely challenges the application of one category of overpayment to that loss and restitution determination. Orillo contends that only overpayments related to visible alterations of OASIS entries should be used in determining the loss amount and the restitution award. Of the 177 claims identified by Cody, 24 files involved overpayments totally unrelated to altered items, generating overpayments of $8,542.73 out of the total overpayment amount of $47,444. In addition, Orillo points out that in a group of medical files involving 61 episodes of care in which both altered and unaltered items resulted in overpayments, in 16 of those files altered and non-altered OASIS items that triggered overpayments appear together on the same page. The 61 files which contained both types of unsupported entries generated $30,995 of the $47,444 overpayment. Orillo contends that the court, in calculating the loss and restitution amount, should have limited itself to only overpayments related to visible alterations, and should not have considered any overpayments related to OASIS entries *245 that were unsupported but not visibly altered. As to those overpayments associated with OASIS entries lacking visible alterations, Orillo asserts that there is no reason to attribute it to wrongdoing as opposed to routine human error. In support of that contention, Orillo points to Cody’s statement that he did not attempt to determine fraud, but was limited to determining only whether the OASIS entries were supported by the documentation. Orillo stated that she had expected Cody to determine instances of upcoding, and that his statement that he did not attempt to determine fraud was inconsistent with that and an indication that the overpayments did not equate with fraudulent conduct. There are numerous problems with that argument. First, it is of no import that Cody testified that he was not attempting to determine whether the entries indicated fraud. Cody testified that he determined whether the OASIS assessment included claims inconsistent with the record, and provided a listing of those instances and the resulting overpayments. That is evidence of upcoding, which is precisely what Orillo acknowledges that Cody was entrusted to unearth. Cody’s testimony was introduced merely to determine the extent to which the supporting medical documentation was inconsistent with the OASIS entries, and the impact those inconsistencies had on the Medicare payments received by Chalice, not to reach a legal conclusion as to the import of that evidence. The government properly did not ask Cody to draw any conclusions as to whether the errors constituted fraud, which is a legal determination. As noted, Orillo does not contest Cody’s qualifications to make the OASIS assessments, nor does she dispute the validity of Cody’s conclusions as to which OASIS entries were supported in the record and which were contradicted by it. She argues, however, that because Cody failed to testify that the overpayments resulted from fraud, the court could not rely on the overpayments themselves as evidence of criminal' conduct in determining the loss and restitution amounts. She argues that the court should have limited its loss and restitution calculations to overpayments that resulted from visible alterations on the OASIS forms. That argument rests on the flawed premise that the only evidence tying the overpayments to Orillo’s criminal conduct was the testimony by Cody and the visible alterations. Orillo’s own plea agreement provided evidence explicitly linking the overpayments to her conduct. Orillo admitted in her plea that she altered the OASIS and Recertification forms to make the patients’ conditions appear worse and the health care needs greater than the actuality, thus resulting in overpayments. Significantly, she further admitted that she made those alterations in two ways— by marking the changes on the forms submitted by the Chalice nurses, and by replacing entire pages with falsified pages manufactured by Orillo herself, which would therefore not necessarily contain alterations visible to a reviewer. Orillo’s contention that the court should be limited to considering only visible alterations ignores the second part of that admission, which is that she manufactured overpayments by falsifying entire pages. Orillo even recognizes that problem at one point, asserting that the government has no evidence linking the overpayments from OASIS items not visibly altered to her conduct other than her own admission that she substituted pages in some OASIS forms. A defendant’s own admission is, of course, evidence enough of the matter admitted. Orillo admitted that the scheme involved not only OASIS forms that con *246 tained visible alterations but also included forms on which the changes would not be visible because the entire form was fabricated. That provided an adequate basis to link the overpayments to her conduct even where those overpayments related to OASIS forms on which no alterations were visible. Orillo attempts to escape that consequence, however, by arguing that for some forms, overpayments were attributable to both altered parts of the OASIS forms, and to other items on those OASIS forms that were not visibly altered. She asserts that documents containing both altered and unaltered entries inconsistent with the underlying records should not be considered. The existence of both types of unsupported entries on one document is not, however, inconsistent with the scheme alleged. The claim was that Orillo altered OASIS entries to obtain a desired payment from Medicare. Therefore, even if substituting pages, Orillo would have had a reason to further make alterations if the amounts resulting from those initial entry changes were not sufficient to achieve the desired overpayment. Orillo relies on United States v. Schroeder, 536 F.3d 746 (7th Cir.2008), to argue that the unaltered entries cannot be considered because the district court is required to establish that an erroneous payment is attributable to the defendant’s fraud and not to mistake or fraud by others. In Schroeder, the defendant pled guilty to tax preparer fraud, and the district court in calculating the loss attributable to the defendant included all tax returns by Schroeder’s clients in which the clients could not justify the deductions on the forms. Id. at 749-50. Although we remanded the case for reconsideration of the loss amount, that determination was based upon problems with the court’s sentencing hearing as a whole and its allocation of the burden of proof, rather than a rejection of that method of loss calculation. As we explained in United States v. Littrice, 666 F.3d 1053, 1062 (7th Cir.2012), the sentencing in Schroeder was “flawed from the outset” because the district court announced its loss finding at the start of the sentencing hearing before the defendant had an opportunity to present evidence and.repeatedly confused the government’s burden of proof with the determination of the evidence’s admissibility. Schroeder, 536 F.3d at 752. The case was remanded on that basis rather than on a determination that the approach to determining the loss amount was fatally defective. In fact, Littrice upheld a loss determination in a tax preparer fraud case that similarly relied on the improper deductions in the tax forms submitted. In Lit-trice, we noted that under the clearly erroneous standard of review, a district court’s calculation must be upheld unless the defendant shows that it was not only inaccurate but outside the realm of permissible computations. 666 F.3d at 1060. The government in Littrice demonstrated a pattern of submitting tax returns containing false deductions for business and educational expenses and charitable deductions. We held that the district court did not err in attributing the underpayments to the defendant in all cases in which the taxpayer failed to contest the audit. Id. at 1061. Although Littrice argued, as had Schroeder, that some of those errors on the tax forms could have been the result of mistake or fraud by the taxpayer rather than the defendant, that did not prevent the district court from including those amounts in the loss calculation. We emphasized in Littrice that the defendant had been provided an opportunity to analyze the government’s evidence, and that Littrice failed to meet his burden to draw the facts of the PSR sufficiently into question, presenting only “un *247 likely” and “implausible” justifications to the district court to explain the large quantity of materially false returns. Id. at 1062-63. This case presents a similar scenario. The government has identified a pattern of a large quantity of improper OASIS coding to achieve overpayments, and Orillo has acknowledged that she reviewed all OASIS forms and altered the forms to obtain overpayments. The district court has based its loss calculation on those payments to Chalice that were not supported by the medical records and therefore not reflective of the care provided. Unlike Littrice and Schroeder in which the taxpayers had an independent incentive to inflate the numbers, Orillo cannot even argue that other persons may have been attempting to defraud the government; she argues only that the overpayments could have been attributable to mistakes. She has not singled out any claims or records to illustrate the possible mistake, instead relying on speculation. As in Lit-trice, that is insufficient to call into question the calculations by the district court. See also United States v. Austin, 54 F.3d 394, 402 (7th Cir.1995) (rejecting factually unsupported speculation in a challenge to the loss amount). In fact, the speculation that the overpayments were the result of human error is belied by the numbers. The numbers found by Cody are themselves strong evidence connecting the overpayments on all forms to Orillo’s fraudulent conduct, and thus support the district court’s conclusion that those amounts had been established by a preponderance of the evidence. Orillo’s sole claim was that the overpayments not attributed to visible alterations could have resulted from human error. The absence of any such “human errors” that resulted in underpayments, however, is a rather strong indication that the errors were intentional rather than random and accidental. In the sample of 177 episodes of care, Cody found $47,444 in overpayments related to the unsupported entries, and $0 in underpayments. Those errors are skewed entirely in Chalice’s favor, with no errors at all that resulted in a loss to Chalice. The skewness of that error rate would itself be a basis to deduce intentional wrongdoing, particularly given Orillo’s failure to identify any evidence that a different distribution of errors is the norm in this situation. In short, there is no reason at all to expect inadvertent human error to result in a monetary benefit to Chalice without exception, and correspondingly, the evidence that the entries solely benefitted Chalice and never shortchanged it indicates that the errors were intentional rather than inadvertent. Orillo’s all-or-nothing argument fails in light of those numbers, because there is absolutely no reason to believe that errors of that magnitude are attributable to human error. And of course, she has admitted that at least some of those errors were part of her fraudulent scheme, in her admission that she substituted whole pages on OASIS forms. The conclusion that the overpayments were related to the fraud, not to human error, is rendered even more likely when considering Orillo’s admission that she caused all OASIS and Recertification forms to be forwarded to her and that she “personally reviewed” those forms. That further provides a basis for the court to determine, by a preponderance of the evidence, that the overpayments were attributable to Orillo’s conduct and a basis to determine the loss and restitution amount. Finally, the court limited the danger of overcounting by using the most conservative estimate of the loss in making its ruling. Of the figures determined by Trust Solutions, the district court used the lower limit on the range of possible loss amounts that generated a 95% confidence *248 level, which was the most conservative estimate of loss reached in the Trust Solutions analysis. In conclusion, Orillo’s sole argument that the loss and restitution amount should be limited to only those stemming from visible alterations is without any support, and would ignore criminal conduct which she admitted in the plea agreement. As that is the only contention raised by Orillo with respect to the loss and restitution determination, her argument is without merit. The only other argument raised by Orillo is the curious contention that she did not waive the right to contest the appropriate sentence for her Count II conviction. Orillo does not, however, actually raise any challenge to the sentence on that count before this court on appeal. Nor did she raise any challenge to that sentence determination in the district court. In fact, the offense level of Count II first is mentioned in the district court’s order on Orillo’s motion for release pending appeal. The government, in arguing that there was no substantial probability that Orillo would obtain a reduction in prison time as a result of the appeal, argued that even if Orillo succeeded in her challenge to Count I, Orillo would not achieve a reduced sentence because of the adjusted offense level for Count II. The district court noted that Orillo had not advanced any argument challenging that calculation, and in a footnote stated that the issue would likely be considered waived by this court if presented on appeal because Orillo had specifically identified the issue in the district court and had chosen not to address it. Orillo has still not raised any challenge to the calculation of the adjusted offense level for Count II before this court. There is, in short, no issue raised for which we must determine whether waiver applies. Accordingly, there is nothing more for this court to address. The decision of the district court is AFFIRMED.