AEDPA Deference Survives Loper Bright: “Double Deference” Governs Strickland Claims in § 2254 Habeas Review

1. Introduction

In Scott DeBruyn v. Adam Douglas (6th Cir. Mar. 5, 2026), the Sixth Circuit affirmed the denial of federal habeas relief to a Michigan prisoner convicted of delivering oxycodone (allegedly via Percocet) causing death under Mich. Comp. Laws Ann. § 750.317a. The petitioner, Scott Allen DeBruyn, argued that he received ineffective assistance of counsel at trial. The respondent was Adam Douglas, Warden.

The appeal raised two core post-conviction issues: (1) whether trial counsel were ineffective for failing to investigate an “acetaminophen-based defense” (i.e., arguing the decedent did not ingest Percocet because acetaminophen was absent from her blood), and (2) whether counsel were ineffective for failing to call a defense expert to contest causation (whether oxycodone was a “substantial factor” in death). Overlaying those merits, DeBruyn also launched a structural challenge to AEDPA, arguing that 28 U.S.C. § 2254(d) is unconstitutional and/or invalid after Loper Bright Enterprises v. Raimondo.

2. Summary of the Opinion

Judge Readler, writing for the panel, held that DeBruyn could not overcome AEDPA’s limits on federal habeas relief where the Michigan courts had adjudicated his ineffective-assistance claims on the merits. Applying the “double deference” that results from combining Strickland v. Washington with 28 U.S.C. § 2254(d), the court concluded:

• Failure to investigate acetaminophen theory: the state courts reasonably found no deficient performance and no prejudice.
• Failure to call a defense expert: the state courts reasonably found cross-examination was a permissible strategy and that the proposed expert opinions were speculative or cumulative, hence no prejudice.
• AEDPA challenge: circuit precedent (notably Sanders v. Plappert) foreclosed arguments that § 2254(d) violates Article III or is undermined by Loper Bright Enterprises v. Raimondo.

The Sixth Circuit therefore affirmed the district court’s denial of the writ.

3. Analysis

3.1 Precedents Cited (and How They Shaped the Outcome)

A. Ineffective assistance framework

• Strickland v. Washington, 466 U.S. 668 (1984)
  The governing rule: a petitioner must show (1) deficient performance (objectively unreasonable under prevailing professional norms) and (2) prejudice (a substantial likelihood of a different outcome). The panel repeatedly emphasized Strickland’s “strong presumption” of reasonableness and its tolerance for strategic choices.
• Harrington v. Richter, 562 U.S. 86 (2011)
  Central to both claims. First, it tightened the prejudice lens (“substantial,” not merely “conceivable,” likelihood of a different result). Second, it rejected any notion that defense counsel must always counter the prosecution’s experts with “equal and opposite” experts, underscoring that cross-examination can suffice. The opinion repeatedly used Richter to validate cross-examination as a reasonable strategy and to characterize DeBruyn’s alternative-causation theories as “theoretical possibility” when untethered to case-specific proof.
• Premo v. Moore, 562 U.S. 115 (2011) and Cullen v. Pinholster, 563 U.S. 170 (2011)
  These reinforced deference to counsel’s predictive judgments about how trial will unfold, and the requirement that reviewing courts “affirmatively entertain” plausible reasons for counsel’s choices.
• Knowles v. Mirzayance, 556 U.S. 111 (2009)
  Used to highlight that Strickland is a general standard, giving state courts wide latitude when evaluating strategy—latitude that becomes even broader under AEDPA.
• Dunn v. Reeves, 141 S. Ct. 2405 (2021) (per curiam) and Burt v. Titlow, 571 U.S. 12 (2013)
  Cited to frame the “no competent lawyer” boundary: federal habeas relief is inappropriate unless counsel’s approach is beyond what any competent attorney could choose.
• Hinton v. Alabama, 571 U.S. 263 (2014) (per curiam)
  DeBruyn invoked Hinton to argue expert testimony was essential. The panel distinguished it as a case about a non-strategic mistake (a misunderstanding about funding), not a strategic choice to rely on cross-examination.

B. AEDPA and “fairminded disagreement” constraints

• 28 U.S.C. § 2254(d)
  The controlling statutory limitation: where state courts decided the claim on the merits, relief lies only if the state decision was contrary to or an unreasonable application of clearly established Supreme Court law, or was based on an unreasonable determination of facts.
• White v. Woodall, 572 U.S. 415 (2014) and Brown v. Davenport, 142 S. Ct. 1510 (2022)
  These were used to insist that “clearly established” law must be a Supreme Court holding (not dicta), and must be framed at an appropriate level of generality.
• Shinn v. Kayer, 141 S. Ct. 517 (2020) (per curiam) and Sexton v. Beaudreaux, 585 U.S. 961 (2018) (per curiam)
  The “fairminded disagreement” standard: the state court must be not just wrong, but unreasonably so—beyond the possibility that fairminded jurists could disagree.
• Shoop v. Twyford, 142 S. Ct. 2037 (2022) and 28 U.S.C. § 2254(e)(1)
  The panel stressed the high bar for factual challenges: state factual findings are presumed correct and rebuttable only by clear and convincing evidence; “unreasonable” is a substantially higher threshold than mere error.
• Rayner v. Mills, 685 F.3d 631 (6th Cir. 2012)
  Cited for the procedural point that AEDPA deference applies where the state court adjudicated the ineffective-assistance claim on the merits.

C. Reliance on Sixth Circuit and other non-Supreme Court authorities (and their limits)

• Richey v. Bradshaw, 498 F.3d 344 (6th Cir. 2007) and Kendrick v. Parris, 989 F.3d 459 (6th Cir. 2021)
  DeBruyn cited Richey for a failure-to-investigate theory. The panel, following Kendrick, treated Richey as effectively overtaken because it predates and conflicts with later Supreme Court AEDPA guidance (especially Harrington) and used an outdated mode of review. This is doctrinally important: the Sixth Circuit signaled that litigants should not rely on older circuit AEDPA applications that do not faithfully implement the Supreme Court’s later “double deference” approach.
• Stermer v. Warren, 959 F.3d 704 (6th Cir. 2020); Glebe v. Frost, 574 U.S. 21 (2014) (per curiam); and Marshall v. Rodgers, 569 U.S. 58 (2013) (per curiam)
  DeBruyn attempted to derive a specific “must-retain-an-expert” rule from Stermer. The panel rejected that move because circuit decisions cannot create “clearly established Federal law” for § 2254(d)(1) purposes, and because lower courts cannot sharpen general Supreme Court principles into specific rules the Supreme Court itself has not announced.
• Clardy v. Pounds, 126 F.4th 1201 (6th Cir. 2025)
  Used to reinforce that Hinton is about non-strategic legal error (mistake about law/funding), not a strategic choice.

D. AEDPA’s constitutionality after Loper Bright

• Sanders v. Plappert, --- F.4th ---, 2026 WL 593932 (6th Cir. Mar. 3, 2026)
  The key, near-contemporaneous circuit authority rejecting the argument that § 2254(d) violates Article III or is undermined by Loper Bright. The panel treated Sanders as binding and dispositive on these structural challenges.
• Bowling v. Parker, 882 F. Supp. 2d 891 (E.D. Ky. 2012) (Thapar, J.)
  Cited approvingly for the proposition that Article III does not prevent Congress from allowing state courts to adjudicate constitutional claims and that § 2254(d) does not transfer federal judicial power to state courts.
• Loper Bright Enterprises v. Raimondo, 144 S. Ct. 2244 (2024)
  The panel characterized Loper Bright as a federal administrative-law decision about courts exercising independent judgment under the APA when reviewing federal agency statutory authority. It held (via Sanders) that Loper Bright does not disturb Congress’s ability to limit the conditions under which federal courts may grant habeas relief to state prisoners.
• Miles v. Floyd, No. 24-1096, 2025 WL 902800 (6th Cir. Mar. 25, 2025)
  Reinforced that similar post-Loper Bright attacks on AEDPA have been rejected in the circuit.

E. State-law building blocks in the background

• People v. Ginther, 212 N.W.2d 922 (Mich. 1973)
  Provided the procedural mechanism for developing an ineffective-assistance record in Michigan through a “Ginther hearing.”
• People v. DeBruyn, No. 352274, 2022 WL 981281 (Mich. Ct. App. Mar. 31, 2022) (per curiam) and People v. DeBruyn, 978 N.W.2d 836 (Mich. 2022) (mem.)
  These were the merits adjudications triggering AEDPA deference and supplying the state courts’ reasoning that the Sixth Circuit measured against § 2254(d).

3.2 Legal Reasoning

A. The “double deference” structure controlled everything

The opinion is a textbook example of layered restraint in federal habeas:

1. First layer (Strickland deference): courts presume counsel acted reasonably and avoid hindsight second-guessing of strategy.
2. Second layer (AEDPA deference): even if a federal judge might disagree, relief is barred unless the state court’s application of Strickland was not just wrong but objectively unreasonable beyond fairminded disagreement.

This framing drove the outcome on both performance and prejudice.

B. Claim 1: failure to investigate an “acetaminophen-based defense”

Performance. DeBruyn argued that competent counsel would have investigated whether the lack of acetaminophen in blood undermined the State’s Percocet-delivery theory. The Sixth Circuit held the Michigan Court of Appeals reasonably treated counsel’s strategic focus—contesting whether oxycodone was a “substantial factor” in a mixed-drug death—as a permissible choice. Given the record evidence (repeated requests for oxycodone, the purchase of 40 Percocet pills, the “A333” pill photo, and the presence of oxycodone in blood plus acetaminophen in urine), the panel found it at least reasonable to proceed on the assumption that Percocet was ingested and to concentrate on causation.

Prejudice. The court accepted as reasonable the state-court view that the blood/urine pattern could fit the prosecution’s timeline because acetaminophen can metabolize and clear more quickly than oxycodone. Because the state court credited testimony supporting that explanation, DeBruyn could not show an unreasonable factual determination or a substantial likelihood that presenting the acetaminophen theory would have changed the verdict.

Notably, the panel rejected the attempt to constitutionalize a categorical rule that “technical” prosecutions require consultation with a defense expert; it relied on Strickland and Harrington to keep the inquiry case-specific and strategy-sensitive.

C. Claim 2: failure to call a defense expert on causation

Performance. The panel treated the decision to rely on cross-examination—as opposed to presenting a defense expert—as a conventional and often reasonable approach, especially in cases where the defense theme is “reasonable doubt” rather than an affirmative counter-narrative. It leaned heavily on Harrington v. Richter’s admonition that the Constitution does not require a defense expert for every prosecution expert.

Prejudice. The proposed expert testimony (residual opioid tolerance, serotonin syndrome refinements, difluoroethane lethality comparisons, and seizure speculation) was largely deemed speculative, cumulative, or not concretely tied to the decedent’s actual levels and timing in a way likely to alter the jury’s causation finding. The panel emphasized that many of these points established, at most, “theoretical possibilities,” which Harrington treats as inadequate for prejudice.

D. AEDPA survives Article III and Loper Bright challenges

The opinion’s most “precedent-facing” component is its explicit rejection of the argument that AEDPA’s § 2254(d) violates Article III by shifting federal judicial power to state courts, and the related argument that Loper Bright Enterprises v. Raimondo requires independent federal judgment unconstrained by AEDPA. Citing Sanders v. Plappert as controlling, the panel reaffirmed that Congress may define the conditions for federal habeas relief and that nothing in Loper Bright (an APA/federal-agency review decision) displaces AEDPA’s habeas framework.

3.3 Impact

A. Post-Loper Bright habeas litigation in the Sixth Circuit

This decision—together with Sanders v. Plappert (which it follows)—signals that Loper Bright will not become a backdoor tool to loosen § 2254(d) deference. Future petitioners raising Article III or “independent judgment” attacks on AEDPA in the Sixth Circuit should expect swift dismissal absent intervening Supreme Court direction.

B. Reinforcement of strategic latitude in expert-driven causation trials

The court reaffirmed that even in scientifically complex cases (toxicology, mixed-drug overdoses), defense counsel may reasonably choose robust cross-examination over presenting a defense expert, and habeas courts will be reluctant to second-guess that choice—especially under AEDPA.

C. Practical consequences for ineffective-assistance development

The opinion illustrates the evidentiary burden needed to turn a plausible alternative medical theory into habeas prejudice: general propositions (e.g., “residual tolerance exists,” “difluoroethane can be lethal at X in some cases”) are unlikely to matter unless connected to the decedent’s specific levels, timing, and clinical picture in a way that meaningfully undermines the prosecution’s causation proof.

D. Doctrinal cleanup: older AEDPA applications are vulnerable

By treating Richey v. Bradshaw as effectively superseded (through Kendrick v. Parris and the Supreme Court’s later AEDPA cases), the panel discourages reliance on pre-Harrington circuit decisions that applied AEDPA less rigorously than current Supreme Court doctrine demands.

4. Complex Concepts Simplified

• AEDPA (§ 2254(d)) in plain terms: if a state court already decided your federal constitutional claim on the merits, a federal court generally cannot grant habeas relief just because it thinks the state court was wrong. The state court must be unreasonably wrong under Supreme Court holdings, or unreasonably wrong about the facts.
• “Clearly established Federal law”: only Supreme Court holdings count (not lower-court cases, and not Supreme Court dicta).
• “Fairminded disagreement”: if reasonable judges could disagree about the answer, habeas relief is barred under AEDPA.
• Strickland “performance” vs. “prejudice”: performance asks whether counsel’s choices were objectively unreasonable; prejudice asks whether there is a substantial likelihood the result would have changed without the error.
• Ginther hearing: Michigan’s post-trial fact-finding hearing designed to develop evidence for ineffective-assistance claims (often including testimony from trial counsel and experts).
• Why “acetaminophen in urine but not blood” mattered: the prosecution argued Percocet was ingested earlier, and acetaminophen can clear from blood faster than oxycodone, leaving traces in urine even if absent in blood at death.
• “Mixed-drug overdose” causation: where multiple substances are present, the legal question may be whether the charged drug was a “substantial factor” in death, even if other substances contributed.

5. Conclusion

Scott DeBruyn v. Adam Douglas reinforces two intertwined principles in modern habeas review. First, AEDPA’s § 2254(d) deference remains firmly intact in the Sixth Circuit, including after Loper Bright Enterprises v. Raimondo, and Article III-based attacks on AEDPA are foreclosed by binding precedent such as Sanders v. Plappert. Second, on ineffective-assistance claims, the opinion exemplifies “double deference” in action: strategic decisions to focus on causation rather than an alternative ingestion theory, and to rely on cross-examination rather than a defense expert, will rarely be disturbed on federal habeas absent a clearly unreasonable state-court decision and a concrete, non-speculative showing of prejudice.