The Government Cannot Cut Corners: Appellate Courts Reverse Sentences in Three Recent Cases Because the Government Did Not Adequately Prove Loss

Three appellate courts recently reversed sentences because the government did not adequately present evidence to support the alleged loss calculation underlying the court’s determination of the sentencing guidelines range. In one case, the court found that the government’s extrapolation of the alleged loss amount was not statistically valid. In the other two, the courts found that the government presented insufficient or mathematically flawed evidence to support its claimed loss amount. While the convictions were affirmed in these cases, the sentences were vacated and remanded to the district court for resentencing.

The three cases ranged from illegal dispensing of controlled substances to healthcare fraud to theft of trade secrets, but all three appellate courts concluded that the government did not provide sufficient evidence to support the sentences imposed and could not rely on speculation in determining the alleged loss amount. Together, these cases indicate that appellate courts will hold the government to its burden to present reliable evidence in order to increase a defendant’s sentence based on relevant, but uncharged, conduct and to show that the intended loss is supported with appropriate evidence of reliability.

United States v. Titus (3rd Cir., Aug. 22, 2023)

The defendant doctor was convicted of operating a “pill mill” and unlawfully dispensing controlled substances related to 13 specific prescriptions. During the trial, the government presented evidence from its statistician and medical expert of many allegedly illegal prescriptions beyond those charged in the indictment. The government’s statistician identified 1,142 patients who had received a prescription for controlled drugs from the defendant during his practice’s last two years, from which he pulled a sample of 300 patients. Extrapolating from these 300 files to the total universe of patients, the witness concluded that the defendant had provided (a) 29,323 prescriptions for controlled substances to 948 patients with at least one inconsistent drug test; and (b) 1,552 prescriptions for controlled drugs to 352 patients whom he had already discharged from his practice. From this sample of 300 patients, the government pulled 282 patient files and had its medical expert review the first 24 files. The medical expert found that the defendant had written illegal prescriptions to 18 of the 24 patients.

In calculating the sentencing guidelines, the district court determined the weight of the controlled substances, including amounts extrapolated from the 24 reviewed files, even though the district court acknowledged that this was “not a statistically valid number.” The severity of the sentence increased substantially based on the extrapolation, resulting in a guidelines range of 292 to 365 months’ imprisonment, and the court sentenced the defendant to 240 months.

The Court of Appeals found that, while extrapolation is permissible in sentencing calculations, there must be an adequate factual basis for the extrapolation and the extrapolation must be performed consistent with accepted standards of reliability. Here, however, the government had not shown that the sample size was large enough to be reliably representative of the remaining thousands of prescriptions, nor documented proper extrapolation methods. Further, the district court had extrapolated from the government’s medical expert’s review of 24 files to infer thousands of illegal prescriptions even though it acknowledged that the sample was not statistically valid. In affirming the conviction but reversing the sentence, the appellate court stated:

Though the prosecution bears a heavy burden of proof, we will not let it cut corners. Dr. Patrick Titus wrote thousands of prescriptions for controlled substances. The government properly proved that many of these prescriptions were unlawful, so we will affirm Titus’s conviction. But many other prescriptions were lawful. And the severity of Titus’s sentence depended on how many were not. Rather than review every patient’s file, the government urged the court to extrapolate from a small sample. Yet the government failed to show that doing so would satisfy its burden to prove the drug quantity by a preponderance of the evidence. Because the court sentenced Titus without enough proof, we will vacate his sentence and remand for resentencing.

United States v. Newton (7th Cir., Aug. 7, 2023)

The defendant worked for a home health provider and was convicted of conspiracy to commit healthcare fraud and wire fraud. Specifically, the government charged that at least part of the healthcare company’s operation was a ruse to collect Medicare reimbursements fraudulently. The company would submit Medicare claims for health services, including skilled nursing services, provided to patients who the company represented were confined to their homes. In fact, however, many of these patients were not actually homebound or in need of skilled care, and therefore did not qualify for Medicare reimbursement. The company also doctored service notes either to overcharge for services, bill for services not provided, or alter dates to resubmit claims that had already been rejected by Medicare.

The district court determined that the loss amount attributable to the criminal conduct was $6.3 million, or 90% of the total amount the home health provider billed to Medicare during the course of the conspiracy, and that the guidelines range was 70 to 87 months in prison. The court sentenced the defendant to a below-guidelines term of 56 months in prison and to pay $6.3 million in restitution.  The Court of Appeals held that the trial court had improperly attributed 90% of the home healthcare company’s Medicare billings to fraud without a sufficient factual basis to support the finding.

Based on the testimony of one of the company’s employees, the district court deducted 10% from the total Medicare billings to account for the possibility that some of the charges were legitimate. However, there were other employees who did not connect the vast majority of the company’s work to the fraud. Based on these facts, the appellate court vacated the sentence and remanded for re-sentencing, stating, “Only speculation can support the district court’s starting presumption that 90% of Medicare reimbursements to Care Specialists were fraudulent, regardless of which nurse performed the care. The calculation of the loss amount attributable to Newton was, therefore, clearly erroneous.”

United States v. You, (6th Cir., July 11, 2023)

The defendant was charged and convicted of stealing trade secrets from her former employers related to making coatings for soda cans without the chemical bisphenol-A (“BPA”). While working as a chemist for Coca-Cola, the defendant had access to formulas for making BPA-free chemical coatings that Coca-Cola had received, subject to non-disclosure agreements, from companies that wanted to enter into agreements with Coca-Cola. During this time, the defendant started a company in China to make a BPA-free chemical coating. Prior to being laid off from her employment at Coca-Cola, she transferred the BPA-free formulas to her personal email and to a USB drive. She later did the same thing at her next employer, who discovered what she did and reported her to the FBI.

The defendant was ultimately charged with conspiracy to commit theft of trade secrets, possessing stolen trade secrets, wire fraud, and economic espionage, and convicted at trial. The district court sentenced her to 168 months of imprisonment. In determining the loss amount, the district court calculated the anticipated profits that the defendant hoped to earn by using the stolen information to capture the Chinese market for BPA-free can coatings, finding that, in a market dominated by a few multinational companies, anticipated profits were the “most reasonable measure of loss.” The district court estimated that the defendant anticipated earning $17.4 million in sales each year, and multiplied potential sales of $17.4 million per year by seven years, resulting in an intended lost total of $121.8 million. The district court acknowledged that its calculation rested on several “assumptions,” including that the Chinese market for BPA-free coatings would remain fixed, that the defendant would absorb all sales in China from Chinese can makers, and that Chinese can makers were now buying BPA-free coatings in China from the foreign companies.

While agreeing that “loss” under the sentencing guidelines can include intended loss, the Court of Appeals found that the district court calculated the intended loss incorrectly, and that the underlying basis for the estimate was flawed. First, the trial court used estimates for the size of the BPA-free chemical market that were included in the defendant’s business projections that it previously rejected. Second, and more importantly, the trial court equated sales with profits when determining the loss amount. It did not account for operating expenses, manufacturing costs, and research and development costs, among other expenses. According to the appellate court, “In deciding to calculate anticipated profits, the district court needed to pick some profit margin — arguably, any margin between 0% and 100%. Its failure to do so was clear error.”


Under the sentencing guidelines, a federal trial court makes determinations about the potential loss amount or the volume of drugs at issue when calculating a defendant’s sentence. In doing so, the court is authorized to make reasoned estimates and extrapolations. But, the government must present sufficient and reliable information to support the court’s guidelines calculations. As the United States Court of Appeals for the Third Circuit said, the government cannot cut corners. The government must use a statistically valid sampling methodology when attempting to extrapolate loss and must support its loss calculations with reliable, and not speculative, evidence. Appellate courts will closely scrutinize the basis for drug volumes or loss amounts in sentencing and vacate sentences that are based on improperly calculated loss amounts.