Independent Expert Interpretation of Third-Party Verified Data Does Not Trigger the Confrontation Clause

I. Introduction

This Sixth Circuit decision arises from a years-long scheme in which pharmacists operating five Michigan/Ohio pharmacies allegedly billed Medicare, Medicaid, and other insurers for prescriptions they did not dispense. The defendants—registered pharmacists Raef Hamaed, Tarek Fakhuri, Kindy Ghussin, and Ali Abdelrazzaq—went to trial after co-conspirators pleaded guilty.

The core factual dispute at trial centered on billing-versus-purchasing discrepancies: whether the pharmacies’ wholesaler invoices (medication purchased) aligned with prescription-claims data (medication billed). The government’s key analytical witness was Johanna Sullivan of Qlarant, a CMS contractor tasked with Medicare fraud detection, who performed and presented an “invoice review” comparing wholesaler records to claims data.

On appeal, defendants raised multiple issues. The most consequential—and most precedent-facing—concerned the Confrontation Clause: whether Sullivan’s testimony impermissibly conveyed testimonial hearsay from non-testifying analysts who helped collect/validate data. Additional issues included the right to present a complete defense, “expert” bolstering in front of the jury, conspiracy variance (single vs. multiple conspiracies), evidentiary rulings, Rule 17(c) subpoenas, juror impeachment, sentencing loss calculations, sophisticated-means enhancement, and restitution procedure.

II. Summary of the Opinion

The Sixth Circuit affirmed the convictions, sentences, and restitution orders. Key holdings include:

• No Confrontation Clause violation where an expert testifies to her own analysis and interpretation of data, even if others participated in data collection/verification, and where the defendant cannot identify specific out-of-court testimonial statements introduced for their truth.
• No constitutional right-to-defense violation where the district court excluded (a) “legitimate prescriptions” evidence (not a defense to a partial-fraud theory) and (b) generalized evidence of PBM industry bias lacking a concrete nexus to the defendants’ data.
• Improper “expert” designation before the jury (under Sixth Circuit case law), but no plain-error reversal because defendants failed to show outcome prejudice given limiting instructions and substantial independent evidence.
• No variance: the evidence supported a single overarching conspiracy across pharmacies.
• Sentencing affirmed: loss calculations based on Qlarant’s invoice review were permissible; Ghussin’s sophisticated-means enhancement (record falsification to defeat audits) was upheld.
• Restitution affirmed: any procedural misstep was harmless under MVRA procedures because the loss dispute was aired and the defendant had opportunities to challenge the amount.

III. Analysis

A. Precedents Cited (and How They Shaped the Rulings)

1. Confrontation Clause framework and “expert surrogate” limits

The court grounded its Confrontation Clause analysis in the Supreme Court’s modern confrontation line:

• Crawford v. Washington (testimonial statements generally inadmissible unless unavailable + prior cross-examination).
• Melendez-Diaz v. Massachusetts (the prosecution cannot substitute live testimony with certificates/affidavits; the Clause applies to testimonial forensic reporting).
• Bullcoming v. New Mexico (rejecting “surrogate analyst” testimony where the testifying witness did not perform/observe the critical testing).
• Smith v. Arizona (reaffirming that the defendant has the right to cross-examine the person who made the testimonial statement; experts cannot be “mouthpieces” for testimonial hearsay).
• Davis v. Washington, Ohio v. Clark, and Michigan v. Bryant (primary-purpose test defining what is “testimonial”).

The Sixth Circuit also invoked foundational confrontation principles from Delaware v. Van Arsdall, Davis v. Alaska, and Pointer v. Texas, emphasizing cross-examination’s truth-testing function.

2. Evidentiary/defense-rights doctrine

• Holmes v. South Carolina and United States v. Scheffer (right to present a defense exists, but is constrained by legitimate evidentiary rules; exclusions must not be arbitrary or disproportionate).
• Taylor v. Illinois and Sixth Circuit applications like United States v. Reynolds (trial courts may exclude confusing or marginally relevant defense evidence under Rule 403 without violating the Constitution).
• United States v. Dimora, United States v. Ifediba, United States v. Betro, and United States v. Daulton (noncriminal conduct is not a defense where the government’s theory is not “ceaseless” criminality).

These precedents supported the court’s central point: where the indictment and proof involve some fraudulent claims rather than “everything was fraudulent,” evidence that a defendant also did lawful business tends to confuse rather than negate elements.

3. “Expert” bolstering and plain-error prejudice

• United States v. Johnson and United States v. Maya (warning against tendering/declaring “expert” status before the jury because it may lend judicial endorsement).
• United States v. Campbell (describing preferred practice: lay the foundation and elicit opinions, rather than ceremonially “qualifying” the witness before the jury).
• Greer v. United States and Molina-Martinez v. United States (plain-error framework and prejudice requirement).

The panel held the district court’s “expert” references were plainly improper under Sixth Circuit precedent, but—consistent with Greer v. United States—refused reversal absent a reasonable probability of a different outcome.

4. Single vs. multiple conspiracies (variance doctrine)

• Stirone v. United States and United States v. Miller (grand jury limits; charges cannot be broadened).
• United States v. Hughes (factors for single conspiracy: common goal, nature of scheme, overlap of participants).
• United States v. Beals (divisible business relationships do not compel multiple conspiracies finding).
• United States v. Siefert, United States v. Guerrero, and United States v. Robinson (variance standards; view evidence in light most favorable to government).

Applying these, the panel found a unified profit-driven fraud scheme with overlapping ownership, shared personnel, common methods (non-reversals, forged signatures, targeted drug types), and pooled profit-sharing.

5. Other procedural/evidentiary anchors

• Delaware v. Fensterer (Confrontation Clause guarantees an opportunity for effective cross-examination, not cross-examination in whatever manner the defense wishes).
• United States v. Nixon and United States v. Llanez-Garcia (Rule 17(c) is not a discovery fishing expedition; requires relevance, specificity, necessity, and good faith).
• United States v. Reed and United States v. Freeman (Rule 701 permits lay opinion rationally based on perception; personal knowledge/participation can justify interpretation of communications).
• United States v. Kaplan (distinguished: lay interpretation lacked rational basis where witness had no personal experience with defendant).
• McDonald v. Pless and United States v. Gonzales (juror impeachment limits; post-verdict juror statements about confusion do not undo unanimity).

B. Legal Reasoning

1. The Confrontation Clause holding: requiring a “statement,” not merely “teamwork”

The panel’s reasoning turns on a disciplined two-step: (1) identify the out-of-court statement allegedly introduced, and (2) determine whether it was testimonial hearsay offered for its truth (drawing directly from Smith v. Arizona and Crawford v. Washington).

The defendants could not point to a concrete testimonial statement from a non-testifying Qlarant analyst that Sullivan repeated to the jury. Sullivan described Qlarant’s typical processes and testified to the work she performed—reviewing timeframes, verifying data population into the SAS system, performing “fine-tooth comb” quality checks, interpreting SAS results, and preparing trial charts.

The court treated Sullivan’s cross-examination remark that records “had already been verified” as insufficient to establish the transmission of testimonial hearsay. Critically, the panel characterized defendants’ theory as an “inferential leap”: the fact that others performed verification does not mean Sullivan introduced their out-of-court assertions for their truth.

The opinion also stresses a practical evidentiary boundary drawn from Melendez-Diaz v. Massachusetts: the Confrontation Clause does not require calling “everyone who laid hands on the evidence” or every person who participated in chain-of-custody-like steps; disputes about missing invoices or completeness typically affect weight rather than admissibility.

2. The defense-rights holding: relevance and Rule 403 as constitutional gatekeepers

On “legitimate prescriptions,” the court adopted the logic of United States v. Ifediba: where the government does not allege every transaction is fraudulent, proof of lawful transactions does not negate guilt and risks confusing the jury. The Sixth Circuit framed it as a theory-of-the-case question: legitimacy evidence becomes meaningful only if the indictment alleges “ceaseless criminal conduct” (invoking United States v. Daulton).

On PBM bias evidence, the court treated the offered proof as speculative and untethered to the case’s concrete data. The defense theory required the jury to assume PBMs collectively manipulated claims data to “frame” independent pharmacies—an attenuated chain the court found Rule 403 could properly exclude. Even while acknowledging the general relevance of bias evidence (cf. Delaware v. Van Arsdall), it held this showing was “marginally relevant at best.”

3. “Expert” references: error recognized, reversal denied for lack of prejudice

The panel reinforced Sixth Circuit caution that juries should not hear ceremonial “qualification” declarations (United States v. Johnson; United States v. Campbell). It found error “plain,” but—under Greer v. United States and Molina-Martinez v. United States—required a showing of outcome prejudice.

Two features defeated prejudice: (1) limiting instructions reminding jurors they need not accept Sullivan’s opinions and should assign weight independently, and (2) extensive independent evidence (co-conspirator testimony and communications) supporting convictions beyond Sullivan’s testimony.

4. Variance: an “overarching conspiracy” supported by common profit motive and overlapping participation

Applying the United States v. Hughes factors, the panel emphasized:

• Common goal: increasing profits through billed-but-undispensed prescriptions; profit-sharing checks tied participants together.
• Common methods: non-reversal of claims, forged signatures for audits, targeted high-value medications, and (at one pharmacy) brand/generic substitution fraud.
• Overlap: shared ownership and roles across pharmacies, shared personnel, and preexisting relationships.

Ownership and roles could be subdivided, but United States v. Beals foreclosed treating that divisibility alone as multiple conspiracies.

5. Sentencing and restitution: deference to reasonable loss estimates and MVRA procedures

For loss, the court reaffirmed that district courts need only a “reasonable estimate” supported by a preponderance, not mathematical precision (citing United States v. Wala and United States v. Siefert). Defendants’ objections were largely speculative (e.g., hypothesized missing invoices without production). The panel emphasized the challenger’s “heavy burden” to show the calculation fell outside permissible computations (United States v. Wala).

For Ghussin’s relevant conduct, the panel applied U.S.S.G. § 1B1.3 principles (citing United States v. Donadeo and United States v. Kennedy) to hold Wayne Campus losses foreseeable given ownership interest, profit, and involvement in discussions about its profitability.

For “sophisticated means,” the court relied on the holistic scheme approach (citing United States v. Simmerman, United States v. Yousef, and United States v. Crosgrove) and held falsifying records/signatures to defeat audits fits the enhancement.

For restitution, the panel invoked the MVRA’s procedural flexibility and the Sixth Circuit’s approach in United States v. Vandeberg: even if restitution was entered before the promised briefing, any error was harmless because the parties litigated loss at sentencing and via post-judgment motion, and the defendant failed to provide reliable competing proof. The opinion also reiterated that while Guidelines loss and MVRA restitution may differ in some cases, they align where the loss finding reflects actual loss (citing United States v. Simpson and United States v. Clay).

C. Impact

1. Confrontation Clause: a practical boundary for data-driven fraud prosecutions

The opinion’s most significant forward-looking effect is its clarification that Confrontation Clause challenges to expert analytics in complex fraud cases must identify an actual testimonial statement introduced for its truth—not merely point to collaborative workflows where others collected, input, or “verified” data. This is especially consequential in healthcare fraud, where:

• large datasets are assembled by teams,
• software (here, SAS) performs computations, and
• an expert witness interprets outputs and quality-checks inputs.

The ruling signals that, at least in the Sixth Circuit, the government can often present a single qualified analyst to explain an integrated review—so long as the analyst’s testimony is genuinely her own and does not serve as a conduit for an absent declarant’s testimonial assertions.

2. Trial practice: “expert” qualification in front of juries remains a trap

Although affirming, the panel reaffirmed that tendering a witness “as an expert” in front of the jury is disfavored and may be plain error. The decision thus functions as a warning to district judges and litigants: comply with United States v. Campbell and United States v. Johnson to avoid creating avoidable appellate risk—particularly when credibility of a technical witness is central.

3. Defense strategy: relevance showings must be case-specific, not industry-general

The court’s treatment of PBM bias evidence underscores that generalized criticisms of an industry (even if supported by broader policy debates) will likely be excluded absent a concrete link to the particular data, witnesses, or motives in the case.

4. Sentencing: loss challenges require counterproof, not speculation

On loss calculations and restitution, the decision reinforces a recurring theme in Sixth Circuit sentencing: defendants must produce credible competing estimates or demonstrate specific methodological faults; generalized assertions of missing documents or unreliability rarely move the needle.

IV. Complex Concepts Simplified

• PBM (Pharmacy Benefit Manager): an intermediary that processes prescription claims for insurers (verifies eligibility, approves claims, authorizes payment).
• Invoice review: comparing what a pharmacy billed insurers (claims data) with what it bought from wholesalers (invoices). If billing far exceeds purchasing, it suggests prescriptions were billed but not dispensed.
• Confrontation Clause / “testimonial hearsay”: the prosecution generally cannot introduce an out-of-court statement made for use at trial unless the declarant is available for cross-examination (or is unavailable and was previously cross-examined). Not all out-of-court information triggers the Clause—only testimonial statements offered for their truth.
• “Surrogate analyst” problem: an expert cannot simply repeat another analyst’s testimonial results (as in Bullcoming v. New Mexico and Smith v. Arizona). But an expert may testify to her own analysis of underlying data if she is not acting as a mouthpiece for an absent declarant’s assertions.
• Variance (single vs. multiple conspiracies): a mismatch between the conspiracy charged and the conspiracy proved. Courts look for a common goal, common scheme, and overlap in participants.
• Rule 17(c) subpoena: a tool to obtain specific documents for trial, not a broad discovery device. It requires relevance, specificity, necessity, and good faith (no fishing expeditions).
• Sophisticated means (Guidelines): especially intricate steps to execute or conceal a fraud. Even if the core theft is simple, elaborate concealment (e.g., falsified records to defeat audits) can qualify.
• Guidelines “loss” vs. MVRA restitution: Guidelines loss can sometimes include intended loss; MVRA restitution is based on actual victim loss. When the loss finding is actual loss, the numbers often converge.

V. Conclusion

United States v. Tarek Fakhuri is a data-driven healthcare fraud case that meaningfully clarifies Confrontation Clause limits for expert testimony in complex investigations. The Sixth Circuit’s central takeaway is pragmatic but doctrinally grounded: team-based data handling does not itself create a confrontation violation—the defense must identify a specific testimonial hearsay statement being introduced, and courts will distinguish impermissible “surrogate” testimony from permissible independent expert interpretation.

The decision also reinforces: (1) trial courts may exclude “legitimate business” evidence where it does not negate a partial-fraud prosecution theory, (2) generalized industry-bias theories require a concrete case link, (3) ceremonial “expert” qualification before a jury remains error even if often harmless on plain-error review, and (4) sentencing and restitution disputes require evidence-based counterproposals rather than speculation.