DEBRUYN v. DOUGLAS (2026)

AMENDED OPINION

A Michigan state jury convicted Scott Allen DeBruyn of delivering oxycodone to his friend, causing her death. Following the verdict, DeBruyn filed a motion for a new trial claiming that he received ineffective assistance of counsel. The state courts denied his claims on the merits. He then petitioned for a writ of habeas corpus in federal court, renewing the same claims. The district court denied his petition. Because DeBruyn cannot surmount the requirements in the Antiterrorism and Effective Death Penalty Act of 1996, we affirm.

I.

A. On April 12, 2017, upon waking up in his hotel room after a night of drinking, Scott DeBruyn found his friend Camille Gesiakowski slumped over the edge of her bed, her eyes open. He left the room and walked to the hotel's front desk, where he said to an employee, “I think we're going to need an ambulance.” Jury Trial Tr., R.7-9, PageID 894–95. When medical personnel arrived, they confirmed that Gesiakowski was dead, having experienced a drug overdose.

An ensuing investigation by law enforcement revealed the following. DeBruyn had known Gesiakowski since she was a teenager; she used to date his son. Then, sometime in 2015, DeBruyn began having sex with her. According to one of DeBruyn's trial lawyers, the two may not have had “the healthiest” relationship. Jury Trial Tr., R.7-7, PageID 473. DeBruyn was in his fifties, had no job, and lived at his mom's house. Gesiakowski was in her twenties and struggled with drug addiction—oxycodone in particular. DeBruyn fueled her habit, providing her drugs and telling her they would make her feel better.

Gesiakowski did a stint in county jail from August 25, 2016, until April 7, 2017. She died five days after she was released, again, in a hotel room with DeBruyn. This case centers on what happened in the days leading up to her passing.

Between April 1 and April 6, Gesiakowski had seven phone calls with DeBruyn from jail. During these recorded calls, DeBruyn asked Gesiakowski to “hang out” after she was released, promising her that they would “be partners in crime” “like old times.” Jury Trial Tr., R.7-8, PageID 726, 734, 736. He also told her he would get her the “good stuff” once she returned home, which prompted several discussions about oxycodone. Id. at PageID 730.

During one call, for example, Gesiakowski told DeBruyn, “I wish I ․ could take an oxy right now. I've got the worst headache ever”; “I can't wait to do some oxy”; “I just wish I could take a fricking oxy.” Jury Trial Tr., R.7-9, PageID 1033–34. DeBruyn replied, “I've got some. There's [just] no way I can get them to ya [now].” Id. at PageID 1034. During another call, Gesiakowski told DeBruyn that she could not “wait to take some oxys.” Jury Trial Tr., R.7-8, PageID 730. DeBruyn replied, “Yeah, ․ I've got some.” Id. And in a third call, Gesiakowski said to DeBruyn that she wanted “to take an oxy so bad.” Jury Trial Tr., R.7-9, PageID 1090. DeBruyn replied, “Yeah, I've got some of them ․ we'll hook up [when you get home] and do some partying.” Id.

In the morning on April 7, DeBruyn picked Gesiakowski up from jail. He took her to a Walmart, bought cans of air duster, and then took her to his home. Air duster, it bears noting, contains the chemical difluoroethane. While air duster is commonly used to clean computer keyboards, people can abuse the product by inhaling difluoroethane, which gives off a “euphoric affect.” Jury Trial Tr., R.7-7, PageID 500.

Later in the day, Gesiakowski's sister, Celia, came to believe that her sister had gone home with DeBruyn. So she went to DeBruyn's home to find her. When she arrived, Celia discovered Gesiakowski sleeping. Alcohol, empty pill bottles, and air duster cans covered the ground. Celia shook her sister awake. Gesiakowski woke up “delirious” and asked Celia, “Where am I[?]” Jury Trial Tr., R.7-9, PageID 902–03. Celia rushed her sister home and begged her not to see DeBruyn again.

For the next two nights, Gesiakowski stayed with her family. Yet she exchanged several messages with DeBruyn over Facebook. Gesiakowski asked DeBruyn if he could get her “oxys.” Jury Trial Tr., R.7-8, PageID 754. She even called him a “sugar coating liar” because he had not yet made good on his promise to get her the drug. Id. at PageID 763. DeBruyn reassured Gesiakowski, telling her that his friend, Lona Daniels, was “bringing” him “oxys.” Id. at PageID 754. So, he said, he would soon have “all the stuff,” the “good pills.” Id. at PageID 765.

Gesiakowski reunited with DeBruyn at his home on April 9. The following afternoon, DeBruyn bought 40 Percocet pills from Daniels. Percocet has two ingredients: oxycodone (an opioid) and acetaminophen (Tylenol). Less than one hour after he purchased the Percocet pills, DeBruyn took a picture on his phone of a pill with the inscription “A333.” A333 can refer to a drug (usually, Percocet) containing oxycodone and acetaminophen.

The next day, April 11, DeBruyn and Gesiakowski took a taxi to a Walmart, where they bought several cans of air duster. They then rented a room at a Baymont Inn. The following morning, DeBruyn woke and found Gesiakowski dead. She had oxycodone in her blood and acetaminophen in her urine, the ingredients of Percocet.

When officers arrived on scene, DeBruyn told them what had happened. Gesiakowski inhaled a lot of air duster on April 11. They went to bed at around eight o'clock that night. Throughout the night, Gesiakowski “kept getting up and laying back down.” Jury Trial Tr., R.7-8, PageID 628. She complained of being “hot” and had “turned down” the temperature in the room to “almost zero.” Id. at PageID 634–35. She even stripped off most of her clothes because she was so hot. The last time DeBruyn saw Gesiakowski awake was three o'clock in the morning, on April 12. When asked if Gesiakowski had a history of using any drugs, including “[p]ills,” DeBruyn said, “Not that I know of.” Id. at PageID 631.

B. Michigan charged DeBruyn with delivering Gesiakowski a controlled substance, causing her death. Mich. Comp. Laws Ann. § 750.317a. To convict on this charge, the State had to show two things: One, that DeBruyn gave Gesiakowski oxycodone (a controlled substance); and two, that the oxycodone was a “substantial factor” in her death. People v. DeBruyn, No. 352274, 2022 WL 981281, at *5 (Mich. Ct. App. Mar. 31, 2022) (per curiam).

At trial, the State sought to show that DeBruyn gave Gesiakowski oxycodone in the form of Percocet. In the days leading up to, and after, her April 7 release from jail, Gesiakowski asked DeBruyn for oxycodone several times. Then, on April 10, DeBruyn purchased 40 Percocet pills, which contains oxycodone and acetaminophen. Less than an hour later, DeBruyn took a picture on his phone of a pill with the inscription “A333,” which can refer to oxycodone and acetaminophen. The next day, April 11, DeBruyn and Gesiakowski went to a hotel room for the night. Gesiakowski was found dead the following morning. Medical personnel confirmed that she died from a drug overdose. She had oxycodone in her blood and acetaminophen in her urine, the two ingredients of Percocet. Taken together, these uncontradicted facts, the State argued, showed that DeBruyn gave oxycodone to Gesiakowski, who died after consuming the drug. As the State described these events during closing arguments, “[t]here is simply no other explanation as to how [Gesiakowski] could have gotten her hands on any [oxycodone] independent of Scott DeBruyn” and his Percocet. Jury Trial Tr., R.7-11, PageID 1545.

Next, the State presented expert witness testimony to demonstrate that the oxycodone found in Gesiakowski's blood was a substantial factor in her death. Gesiakowski died from a “mixed-drug” overdose caused by four drugs/chemicals in her blood: (1) oxycodone (an opioid), (2) tramadol (an opioid), (3) fluoxetine (Prozac), and (4) difluoroethane (a chemical). Was the oxycodone a substantial factor in her death? Yes, two experts explained. To begin with, Gesiakowski probably had little, if any, tolerance to oxycodone at the time of her death, as she had not had access to the drug from August 2016 (when she went to jail) until April 10 (the day DeBruyn bought her Percocet pills). Gesiakowski also had 180 nanograms per milliliter of oxycodone in her blood, an amount approximately five times the therapeutic range. From this record, the experts said, oxycodone was easily a substantial factor in Gesiakowski's death; in fact, the drug alone could have killed Gesiakowski by causing respiratory failure. The prosecution summed up the expert testimony this way during closing arguments: “[W]e know ․ that oxycodone was a contributing cause and a substantial cause and could have, in and of itself, all alone, killed her.” Jury Trial Tr., R.7-11, PageID 1544.

DeBruyn's lawyers countered that the State had failed to prove beyond a reasonable doubt that oxycodone was a substantial factor in Gesiakowski's death. Counsel presented this defense via cross-examination of the State's experts. They began by getting one of the experts to acknowledge that it is “difficult” (if not impossible) “to tease out, exactly” the extent to which oxycodone contributed to Gesiakowski's death. Jury Trial Tr., R.7-7, PageID 532. They then presented evidence that Gesiakowski could have died from something unrelated to oxycodone. Specifically, counsel suggested that Gesiakowski could have died from either serotonin syndrome (i.e., high body temperature due to the combined effect of the tramadol and fluoxetine in her blood) or an arrhythmia (i.e., a fatal heart problem caused by inhaling too much difluoroethane). DeBruyn's lawyers, in short, argued that (1) nothing in the record conclusively proved that oxycodone was a substantial factor in Gesiakowski's death and (2) Gesiakowski could have died from something unrelated to oxycodone. “Causation,” DeBruyn's lawyers emphasized, “is a very important thing” that the State failed to prove “beyond a reasonable doubt.” Jury Trial Tr., 7-11, PageID 1559.

DeBruyn's defense, however, failed to convince his jury, which convicted him of delivering Gesiakowski a controlled substance causing her death. The trial court later sentenced him to at least 198 months (16.5 years) in prison.

C. DeBruyn appealed his conviction and requested a Ginther hearing—the Michigan vehicle for developing facts to support an ineffective assistance claim. See People v. Ginther, 390 Mich. 436, 212 N.W.2d 922, 925 (1973). The state court of appeals granted his request and remanded the case to the trial court.

The trial court held a two-day Ginther hearing. DeBruyn presented two expert witnesses and examined the State's trial experts. Based on the evidence presented at the hearing, DeBruyn constructed two ineffective assistance of counsel claims, which he later raised in a motion for a new trial.

First, DeBruyn claimed, his lawyers performed ineffectively by failing to investigate before trial what he calls “an acetaminophen-based defense.” Appellant Br. 31. At its core, this defense maintains that Gesiakowski did not consume oxycodone in the form of Percocet, as the State had alleged at trial, because she had no acetaminophen (an ingredient of Percocet) in her blood at the time of her death. The State responded that the absence of acetaminophen in Gesiakowski's blood was consistent with the prosecution's theory of the case: Gesiakowski consumed Percocet on April 10/11 and excreted the drug's acetaminophen into her urine by the time of her April 12 death.

Second, DeBruyn claimed, his lawyers performed ineffectively by failing to call an expert at trial to enhance his defense that oxycodone was not a substantial factor in Gesiakowski's death. In response, the State argued that DeBruyn's trial did not require the use of experts by the defense; rather, the State said, DeBruyn's lawyers were able to effectively present their arguments through cross-examination of the State's experts.

The trial court rejected DeBruyn's claims. The Michigan Court of Appeals affirmed. DeBruyn, 2022 WL 981281, at *10. And the Michigan Supreme Court denied leave to appeal. People v. DeBruyn, 510 Mich. 877, 978 N.W.2d 836 (2022) (mem.). Falling short in state court, DeBruyn petitioned for a writ of habeas corpus in federal court. The district court denied his petition. DeBruyn v. Douglas, No. 23-10219, 2024 WL 3236306, at *12 (E.D. Mich. June 28, 2024). That decision prompted this appeal.

II.

DeBruyn renews the same ineffective assistance claims here: His lawyers wrongly failed to investigate an acetaminophen-based defense before trial, and wrongly failed to call an expert witness at trial. To prevail in this habeas setting, DeBruyn faces a steep climb. Start with the challenge in proving a Sixth Amendment violation. DeBruyn must show that his lawyers performed so incompetently that they in effect deprived him of his right “to have the Assistance of Counsel for his defen[s]e.” U.S. Const. amend. VI. Making that showing requires DeBruyn to prove both that his lawyers performed deficiently, and that their performance prejudiced the outcome of his trial. Strickland v. Washington, 466 U.S. 668, 687, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984). To demonstrate deficient performance, DeBruyn must show that his lawyers made objectively unreasonable decisions, considered under all the circumstances. Id. at 688, 104 S.Ct. 2052. This is a high standard. We start with “a strong presumption” that his lawyers gave “reasonable professional assistance.” Id. at 689, 104 S.Ct. 2052. We then engage in a “highly deferential” review of their actual performance, asking only whether they made objectively reasonable decisions. Id. If DeBruyn can satisfy the performance element, he must also show that he suffered prejudice as a result of his lawyers’ performance. To do so, he must prove a “substantial,” (not “just [a] conceivable,”) likelihood that, but for his lawyers’ performance, he would not have been convicted at trial. Harrington v. Richter, 562 U.S. 86, 112, 131 S.Ct. 770, 178 L.Ed.2d 624 (2011). In short, Strickland demands that DeBruyn show that he suffered prejudice as a result of his counsel's deficient performance.

Yet DeBruyn's burden is increased further still, for he must satisfy these Strickland elements against the backdrop of the Antiterrorism and Effective Death Penalty Act of 1996 (AEDPA). Through AEDPA, Congress imposed strict limitations on federal courts considering habeas petitions from state prisoners. AEDPA constrains our review of “any claim that was adjudicated on the merits in state court.” 28 U.S.C. § 2254(d). Here, the state courts denied DeBruyn's claims on the merits, so AEDPA applies. As a result, to surmount AEDPA, DeBruyn must fit his claims into one of the statute's narrow paths to federal habeas relief. He can try to show that the state court's adjudication of the relevant claim “resulted in a decision that was contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court.” 28 U.S.C. § 2254(d)(1). Or he can try to show that the state court's decision was “based on an unreasonable determination of the facts.” Id. § 2254(d)(2).

Before reaching those issues, we pause to consider a preliminary matter. DeBruyn argues that AEDPA's standard of review violates Article III of the Constitution because it transfers federal judicial power to state courts. DeBruyn also contends that AEDPA's standard of review violates the Supreme Court's decision in Loper Bright Enterprises v. Raimondo, 603 U.S. 369, 144 S. Ct. 2244, 219 L.Ed.2d 832 (2024). But binding cases in our Circuit have already considered these arguments and rejected them. Section 2254(d) limits the power of federal courts to issue the writ of habeas corpus to state prisoners. The statute instructs federal courts that they may issue the writ to prisoners held under a state court judgment only in limited circumstances: relevant here, when a state court adjudication of a claim on the merits resulted in a decision that was contrary to, or involved an unreasonable application of, clearly established Supreme Court precedent, or was based on an unreasonable determination of the facts. 28 U.S.C. § 2254(d). Article III does not prevent Congress from allowing state courts to adjudicate constitutional claims. Sanders v. Plappert, ––– F.4th ––––, –––– – ––––, 2026 WL 593932, at *8–9 (6th Cir. Mar. 3, 2026); Bowling v. Parker, 882 F. Supp. 2d 891, 898–900 (E.D. Ky. 2012) (Thapar, J.). Nor does § 2254(d) authorize state courts to perform any judicial function, let alone a federal one. See Sanders, ––– F.4th at –––– – ––––, 2026 WL 593932, at *7–8. For these reasons, it is difficult to see how AEDPA transfers federal judicial power to state courts.

Loper Bright does not change matters. There, the Supreme Court held that the Administrative Procedure Act, a federal law governing the basic contours of federal agency action, requires federal courts to exercise independent judgement in determining whether a federal agency has acted within its statutory authority. Id. at 2273. We will not dwell on DeBruyn's argument that Loper Bright requires us to strike down AEDPA's standard of review. We have held that it clearly does not. Sanders, ––– F.4th at –––– – ––––, ––––, 2026 WL 593932, at *7–8, 10 (explaining why AEDPA is constitutional after Loper Bright); see also Miles v. Floyd, No. 24-1096, 2025 WL 902800, at *3 (6th Cir. Mar. 25, 2025) (making similar arguments). We will instead repeat what was said above: Neither Loper Bright nor DeBruyn's argument about transferring federal judicial power to state courts prohibit Congress from setting standards for when federal courts may issue the writ of habeas corpus to state prisoners.

A. Turn now to the merits, starting with DeBruyn's claim that his attorneys performed ineffectively by failing to investigate an “acetaminophen-based defense” before trial. Appellant Br. 31. The state court of appeals denied this claim both for lack of deficient performance and prejudice, so AEDPA's highly deferential standard of review applies. Rayner v. Mills, 685 F.3d 631, 638 (6th Cir. 2012). And DeBruyn's claim fails under AEDPA, as he has not shown that the state court decision unreasonably applied Supreme Court precedent or turned on an unreasonable determination of the facts.

Performance. The state court of appeals held that DeBruyn's lawyers were not constitutionally ineffective in foregoing any investigation into an acetaminophen-based defense before trial. The court reasoned that counsel's “determination of trial strategy”—including “decisions about which defenses and arguments to present” at trial—“did not fall below an objective standard of reasonableness.” DeBruyn, 2022 WL 981281, at *5, *7. Disagreeing with that conclusion, DeBruyn contends that the state court's decision “involved an unreasonable application of[ ] clearly established Federal law,” entitling him to relief under AEDPA. 28 U.S.C. § 2254(d)(1). To succeed on that claim, DeBruyn must follow a well-trodden path. He must initially identify a legal principle “clearly established” by a decision of the Supreme Court. For a legal principle to be “clearly established,” it must originate from a Supreme Court holding, and not from its dicta. White v. Woodall, 572 U.S. 415, 419, 134 S.Ct. 1697, 188 L.Ed.2d 698 (2014). And in identifying the “clearly established” legal principle on which he relies, DeBruyn must take care to describe the principle at an appropriate level of specificity. Brown v. Davenport, 596 U.S. 118, 142 S. Ct. 1510, 1525, 212 L.Ed.2d 463 (2022).

DeBruyn characterizes the relevant legal principle as follows: Defense counsel “has a constitutional obligation to reasonably investigate possible defenses” before trial. Appellant Br. 31. That purported principle is helpful only insofar as it goes. True, counsel must undertake a “thorough investigation of law and facts relevant to plausible options” for the defense. Strickland, 466 U.S. at 690, 104 S.Ct. 2052. But Strickland does not require counsel to investigate every conceivable defense. Id. at 690–91, 104 S.Ct. 2052. Instead, counsel's defense strategy must be supported by a “reasonable professional judgment[ ],” measured by “all the circumstances.” Id. at 691, 104 S.Ct. 2052. Where, as here, a petitioner argues that his counsel failed to investigate a defense before trial, we ask only whether counsel made a “reasonable” defense “decision,” one that made “particular investigations unnecessary.” Id. If a counsel's “choices” about defense strategy are reasonable “based on predictions of how ․ trial [might] proceed,” we may not second guess those “choices.” Premo v. Moore, 562 U.S. 115, 132, 131 S.Ct. 733, 178 L.Ed.2d 649 (2011). Through it all, we must “affirmatively entertain the range of possible reasons” defense counsel “may have had for proceeding as [she] did.” Cullen v. Pinholster, 563 U.S. 170, 196, 131 S.Ct. 1388, 179 L.Ed.2d 557 (2011) (citation modified).

On top of that, AEDPA requires DeBruyn to show that the state court unreasonably applied these principles to the facts underlying his claim. 28 U.S.C. § 2254(d)(1). To do that, he must demonstrate that the state court decision was “so obviously wrong that its error lies beyond any possibility for fairminded disagreement.” Shinn v. Kayer, 592 U.S. 111, 141 S. Ct. 517, 523, 208 L.Ed.2d 353 (2020) (per curiam) (citation modified). So if “disagreement is possible,” DeBruyn's “claim must be denied.” Sexton v. Beaudreaux, 585 U.S. 961, 965, 138 S.Ct. 2555, 201 L.Ed.2d 986 (2018) (per curiam). In other words, DeBruyn must show that “every fairminded jurist would agree that” his lawyers performed deficiently. Dunn v. Reeves, 594 U.S. 731, 141 S. Ct. 2405, 2411, 210 L.Ed.2d 812 (2021) (per curiam) (citation modified). In making that assessment, we must remember that, under AEDPA, “the more general the rule, the more leeway state courts have” when applying the rule to a defendant's specific facts. Sexton, 585 U.S. at 968, 138 S.Ct. 2555 (citation modified). And as Strickland’s performance standard is generally defined, the state court had wide “latitude” in holding that DeBruyn's lawyers performed competently. Knowles v. Mirzayance, 556 U.S. 111, 123, 129 S.Ct. 1411, 173 L.Ed.2d 251 (2009).

DeBruyn's claim fails. A fairminded jurist could conclude that his trial attorneys, based on a reasonable prediction of how trial might proceed, pursued a reasonable defense strategy, one that made any investigation into an acetaminophen-based defense unnecessary. That defense would have challenged a key premise of the State's case, namely, that Gesiakowski took oxycodone in the form of Percocet. Gesiakowski, the theory goes, could not have consumed Percocet because she had no acetaminophen in her blood. Yet DeBruyn's lawyers had good reason to presume that Gesiakowski consumed Percocet. The uncontradicted facts show why. Before and after her release from jail, Gesiakowski repeatedly asked DeBruyn for oxycodone. Each time, DeBruyn reassured Gesiakowski that he would get her the drug. Then, on April 10, DeBruyn purchased 40 Percocet pills. Percocet contains oxycodone and acetaminophen. Less than an hour after the purchase, DeBruyn took a picture on his phone of a pill with the inscription “A333,” which can refer to oxycodone and acetaminophen. The next day, DeBruyn and Gesiakowski went to a motel room for the night. Gesiakowski was found dead the following morning. She had oxycodone in her blood and acetaminophen in her urine—Percocet's two ingredients.

On this record, a fairminded jurist could conclude that DeBruyn's lawyers made a “reasonable [ ]calculation” about “how the trial would proceed”: The jury would find that Gesiakowski consumed oxycodone in the form of Percocet before her death. Harrington, 562 U.S. at 110, 131 S.Ct. 770; Premo, 562 U.S. at 132, 131 S.Ct. 733 (citing Harrington, 562 U.S. at 107–08, 131 S.Ct. 770).

In light of the strong evidence that Gesiakowski took Percocet, a fairminded jurist could conclude that DeBruyn's lawyers “ma[de] a reasonable decision” about how to defend their client. Strickland, 466 U.S. at 691, 104 S.Ct. 2052. DeBruyn's lawyers challenged the second premise of the State's case—that oxycodone was a substantial factor in Gesiakowski's death. From information gathered from the State's experts before trial, DeBruyn's counsel made the following arguments. One, nothing in the record conclusively proved that oxycodone was a substantial factor in Gesiakowski's death. Two, Gesiakowski could have died from either serotonin syndrome (i.e., increased body temperature due to the combined effect of tramadol and fluoxetine in her blood), or an arrhythmia (i.e., an irregular heartbeat caused by inhaling too much difluoroethane). In short, the theory goes, because Gesiakowski's cause of death could have been unrelated to oxycodone, the State had failed to prove beyond a reasonable doubt that oxycodone was a substantial factor in her death. Defense lawyers do not perform incompetently by choosing to train their efforts on a reasonable defense challenging a necessary premise of the prosecution's case. Cf. Cullen, 563 U.S. at 196, 131 S.Ct. 1388 (citation modified) (instructing that we must entertain any possible reason DeBruyn's lawyers may have had for not investigating a potential defense). Thus, DeBruyn has not shown that the state court of appeals unreasonably applied clearly established federal law.

DeBruyn's counterarguments do not change matters. Start with the “clearly established” legal principle he urges—that when a prosecution's case involves a “technical” issue, defense counsel must consult a “relevant” defense expert before trial to prepare possible defenses. Appellant Br. 31, 36. Had his lawyers consulted a relevant defense expert (here, a toxicologist), he says, they would have discovered his preferred acetaminophen-based defense. (For purposes of resolving this argument, we will assume that DeBruyn's lawyers did not consult a defense toxicologist before trial.) DeBruyn's argument fails because he has not identified a Supreme Court holding that clearly establishes his rule. No such precedent plainly establishes that a lawyer must consult a “relevant” defense expert before trial in a “technical” case. To be sure, “[c]riminal cases will arise where the only reasonable ․ defense strategy requires consultation with experts” before trial. Harrington, 562 U.S. at 106, 131 S.Ct. 770. But because there are “countless ways to provide effective assistance in any given case,” counsel's failure to “consult” an “expert” may be ineffective only “in some cases.” Id. at 106, 131 S.Ct. 770 (citation modified). That “formulation is sufficiently general that state courts ․ have wide latitude in applying it.” Id. at 106, 131 S.Ct. 770. In short, the Supreme Court has not adopted the hard and fast legal rule DeBruyn advances. See Dunn, 141 S. Ct. at 2410 (quoting Harrington, 562 U.S. at 104, 106–08, 131 S.Ct. 770).

No more availing is DeBruyn's reliance on Richey v. Bradshaw, 498 F.3d 344 (6th Cir. 2007). There, a state jury convicted Kenneth Richey of murder by arson. Id. at 346. In state court post-conviction proceedings, Richey argued that he received ineffective assistance of trial counsel because his attorney failed to investigate whether the fire was actually caused by arson. Id. at 357. The state courts rejected the claim on the merits. Id. at 358. Richey then filed a habeas petition in federal court, renewing the same claim. We held under AEDPA that the state courts had unreasonably applied Strickland’s performance element. Why? Because “Richey's counsel did not conduct the investigation that a reasonably competent lawyer would have conducted into an available defense—that the fire was not caused by arson.” Id. at 362.

Richey’s AEDPA holding, however, has been overtaken by subsequent Supreme Court decisions. In Kendrick v. Parris, 989 F.3d 459 (6th Cir. 2021), we explained that Richey “predates the Supreme Court's guidance in Harrington, which held that a federal court misapplied AEDPA by finding that an [ineffective assistance claim] had merit under de novo review and then declaring without further explanation that the state court's contrary ruling was unreasonable.” Id. at 476 (citation modified). As “Richey engaged in that type of now-outdated review and barely referenced AEDPA's deferential standards when granting relief,” its holding that relief was warranted under AEDPA is no longer a viable precedent. Id.

In sum, a fairminded jurist could conclude that DeBruyn's lawyers chose to pursue a reasonable defense strategy that made any investigation into an acetaminophen-based defense unnecessary. Thus, DeBruyn cannot show that the state court's performance ruling amounted to an unreasonable application of clearly established Supreme Court precedent.

Prejudice. Even if DeBruyn could show that the state court of appeals unreasonably applied Strickland’s performance element, he also must demonstrate that the court made a similarly egregious application of Strickland’s prejudice element. The state court of appeals held that DeBruyn suffered no prejudice from his lawyers’ failure to investigate an acetaminophen-based defense, DeBruyn, 2022 WL 981281, at *9, so AEDPA applies. Rayner, 685 F.3d at 638. Recognizing as much, DeBruyn paints the state court's decision as an unreasonable determination of the facts and an unreasonable application of clearly established federal law. 28 U.S.C. § 2254(d). We disagree.

To demonstrate that the state court decision was based on an unreasonable determination of the facts, DeBruyn must show that the underlying facts found by the state court were not just debatable or incorrect but “unreasonable—a substantially higher threshold” for obtaining relief. Shoop v. Twyford, 596 U.S. 811, 142 S. Ct. 2037, 2043, 213 L.Ed.2d 318 (2022) (citation modified). Factual findings of the state court are presumed correct and can be rebutted only by “clear and convincing” evidence. 28 U.S.C. § 2254(e)(1). The state court's prejudice ruling turned on the following. The fact that Gesiakowski had no acetaminophen in her blood at the time of her death was consistent with the State's theory of the case: She consumed oxycodone in the form of Percocet on April 10 or 11 and then excreted the Percocet's acetaminophen into her urine by the time of her death on April 12. DeBruyn, 2022 WL 981281, at *7.

The state court of appeals’ factual determination was not unreasonable. The Ginther hearing showed that Percocet has two ingredients: oxycodone (an opioid) and acetaminophen (Tylenol). When a person consumes Percocet, oxycodone and acetaminophen enter the blood and then begin to “metabolize.” Ginther Hr'g, R.7-17, PageID 1961. Through this process, the two are excreted “by the kidney into the urine.” Id. But acetaminophen metabolizes faster than oxycodone, even up to twice as fast in some individuals. Against this backdrop, one of the State's experts testified that the State's theory of the case—Gesiakowski consumed oxycodone in the form of Percocet—was “absolutely” “consistent” with the fact that Gesiakowski had no acetaminophen in her blood because her body could have excreted the acetaminophen into her urine by the time of her death. Id. at PageID 1997. In adopting the expert's conclusion, the state court did not unreasonably interpret the evidence. DeBruyn, 2022 WL 981281, at *7.

Resisting this conclusion, DeBruyn argues that the state court erred in determining the facts because it adopted the State's interpretation of the evidence. For example, he asserts that the court “ignored” his expert's testimony that Gesiakowski could not have consumed Percocet, because she had no acetaminophen in her blood. Appellant Br. 54–55. We recognize the conflicting expert evidence. That said, DeBruyn has not offered clear and convincing evidence that the state court unreasonably erred in choosing to credit the State's expert testimony over DeBruyn's. For one thing, as already described, the court, after reviewing the Ginther record, determined that Gesiakowski's blood/urine tests were consistent with the State's theory of the case. DeBruyn, 2022 WL 981281, at *7. A state court does not make a clear and convincing factual error when it chooses to credit a permissible view of the evidence. Anderson v. City of Bessemer City, 470 U.S. 564, 574, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985); Hill v. Shoop, 11 F.4th 373, 389 (6th Cir. 2021) (en banc) (“[F]aced with two reasonable interpretations of evidence, we cannot say that the state court's decision to go with one over the other was unreasonable.”).

For another thing, DeBruyn's expert appears to have based his conclusion that Gesiakowski could not have consumed Percocet on his representation that “oxycodone and acetaminophen are metabolized at the same rate.” Ginther Hr'g, R.7-16, PageID 1834–35. Again, ample evidence from the Ginther hearing said otherwise.

For these reasons, a fairminded jurist could conclude that no juror would have been swayed by DeBruyn's acetaminophen-based defense. The prosecution, as the state court of appeals concluded, had a strong case that Gesiakowski consumed Percocet. DeBruyn, 2022 WL 981281, at *7. Remember, Gesiakowski repeatedly asked DeBruyn for oxycodone, and he then purchased 40 Percocet pills for Gesiakowski from his friend. A few days later, Gesiakowski overdosed and had Percocet's two ingredients in her blood and urine. Based on these facts, it was reasonable for the state court to conclude that it was not substantially likely that the jury would have been moved by the argument that Gesiakowski did not take Percocet.

All in all, DeBruyn's failure-to-investigate claim comes up short.

B. We now consider DeBruyn's claim that his lawyers performed ineffectively because they failed to call an expert at trial to support their defense that oxycodone was not a substantial factor in Gesiakowski's death. Here too, the state court of appeals denied this claim for lack of both deficient performance and prejudice, so AEDPA's highly deferential standards (again) apply. Rayner, 685 F.3d at 638. And DeBruyn's claim (again) fails under AEDPA, because he has not shown that the state court decision unreasonably applied Supreme Court precedent or turned on an unreasonable determination of the facts.

Performance. The state court of appeals found no deficiency in DeBruyn's lawyers’ decision to not call an expert witness at trial. Counsel's “strategy to present the defense theories through cross-examination of the prosecution's experts,” the court explained, was not “objectively unreasonable.” DeBruyn, 2022 WL 981281, at *7. DeBruyn counters that the state court's ruling amounted to an “unreasonable application of[ ] clearly established Federal law[ ] as determined by the Supreme Court.” 28 U.S.C. § 2254(d)(1). We disagree.

DeBruyn carries a familiar burden: He must first identify the “clearly established” legal principle on which he relies. On that note, he is correct that “[c]riminal cases will arise where the only reasonable and available defense strategy requires” the “introduction of expert evidence.” Harrington, 562 U.S. at 106, 131 S.Ct. 770. But with “countless ways to provide effective assistance in any given case,” counsel typically has “wide latitude” in deciding whether to call an expert. Id. (citation modified); see also id. at 111, 131 S.Ct. 770 (noting that Strickland “does not enact Newton's third law for the presentation of evidence, requiring for every prosecution expert an equal and opposite expert from the defense”). Indeed, in presenting a defendant's case, cross-examination is “sufficient” in “many instances,” especially where counsel's “overall performance indicates active and capable advocacy.” Id. at 111, 131 S.Ct. 770. And counsel need not be “exemplary” during cross. Dunn, 141 S. Ct. at 2410 (citation modified). Rather, she needs to exercise “such skill and knowledge as will render the trial a reliable adversarial testing process,” considering all the circumstances. Strickland, 466 U.S. at 688, 104 S.Ct. 2052. In the end, so long as counsel did not take “an approach that no competent lawyer would have chosen,” her decision not to hire an expert is reasonable. Dunn, 141 S. Ct. at 2410 (citing Burt v. Titlow, 571 U.S. 12, 23–24, 134 S.Ct. 10, 187 L.Ed.2d 348 (2013)).

Next, DeBruyn must show that the state court unreasonably applied these principles to the specific facts of his claim. 28 U.S.C. § 2254(d)(1). The state court's ruling must be “so lacking in justification that there was an error well understood and comprehended in existing law beyond any possibility for fairminded disagreement.” Harrington, 562 U.S. at 103, 131 S.Ct. 770.

DeBruyn's claim fails twice over. First, a fairminded jurist could conclude that at least one competent lawyer would have chosen to present DeBruyn's defense through cross examination of the State's experts. See Dunn, 141 S. Ct. at 2411. When “defense counsel does not have a solid case, the best strategy can be to say that there is too much doubt about the State's theory for the jury to convict.” Harrington, 562 U.S. at 111, 131 S.Ct. 770. Here, as explained, the prosecution had a strong case that DeBruyn gave Gesiakowski Percocet and that she died after consuming the drug; it was reasonable for the state court to determine that his lawyers decided to use cross-examination to expose flaws in the State's theory and, at the same time, present alternative theories of Gesiakowski's death.

Second, a fairminded jurist could conclude that DeBruyn's lawyers did not perform deficiently in presenting their defense through cross-examination of the State's experts. See Harrington, 562 U.S. at 111, 131 S.Ct. 770. In fact, as the state court of appeals explained, the record “supports” that DeBruyn's lawyers conducted “a thorough and knowledgeable cross-examination of the [State's] experts.” DeBruyn, 2022 WL 981281, at *7. Consider, for example, their serotonin syndrome defense. At trial, DeBruyn's attorney pressed the State's experts on how tramadol and fluoxetine can combine in dangerous ways to produce serotonin syndrome, a “potentially fatal” condition where a person's body cannot absorb serotonin. Jury Trial Tr., R.7-7, PageID 604. Oxycodone plays no role in the syndrome's onset. With increased serotonin in the body, a person can die from “an extremely high temperature.” Id. at PageID 531. Common symptoms of serotonin syndrome include increased body temperature, agitation, and restlessness. For these reasons, DeBruyn's lawyers argued, Gesiakowski could have died from serotonin syndrome. She had tramadol and fluoxetine in her blood at the time of her death. And she had classic symptoms of the syndrome in the hours before her death: She was hot, had turned down the air, had stripped almost naked, and was pacing back and forth.

DeBruyn's counterarguments come up short. He draws a purported “clearly established” legal principle from Stermer v. Warren, 959 F.3d 704 (6th Cir. 2020), namely, that when the State offers expert testimony, a lawyer acts incompetently by failing to “retain someone to help respond to the state's expert testimony.” Id. at 739. But our precedent, it bears emphasizing, does not “constitute clearly established Federal law, as determined by the Supreme Court.” Glebe v. Frost, 574 U.S. 21, 24, 135 S.Ct. 429, 190 L.Ed.2d 317 (2014) (per curiam) (citation modified). Nor may we use our decisions to “sharpen a general principle” from the Supreme Court “into a specific legal rule” the Supreme Court has not clearly established. Marshall v. Rodgers, 569 U.S. 58, 64, 133 S.Ct. 1446, 185 L.Ed.2d 540 (2013) (per curiam). Equally true, because Stermer expressly refused to base its decision on the expert-witness claim at issue, any discussion there regarding a lawyer's failure to present an expert was dicta. Kendrick, 989 F.3d at 476. And, at all events, the Supreme Court has rejected the precise legal rule DeBruyn derives from Stermer. Harrington, for instance, confirms that defense lawyers have “wide latitude” in deciding whether to call an expert at trial. Harrington, 562 U.S. at 111, 131 S.Ct. 770 (quoting Strickland, 466 U.S. at 689, 104 S.Ct. 2052).

Not so fast, DeBruyn responds. Hinton v. Alabama, 571 U.S. 263, 134 S.Ct. 1081, 188 L.Ed.2d 1 (2014) (per curiam), he says, held that “ ‘the introduction of expert evidence’ is essential when the state's case turn[s] on ․ expert testimony.” Appellant Br. 43 (quoting Hinton, 571 U.S. at 273, 134 S.Ct. 1081). Hinton’s holding, however, is not so sweeping. There, the Supreme Court deemed it unreasonable for an attorney not to seek funding for an expert witness where that failure was based not on a strategic choice but on a mistaken belief about the availability of funds. Hinton, 571 U.S. at 273, 134 S.Ct. 1081. But unlike Hinton, DeBruyn's lawyers made a strategic decision to present their defense through cross-examination of the State's experts. Cf. Clardy v. Pounds, 126 F.4th 1201, 1211 (6th Cir. 2025) (holding that counsel did not commit legal error when the failure to procure an expert was not based on a mistake of state law).

Even then, says DeBruyn, defense attorneys perform incompetently whenever they use cross-examination to “affirmatively present[ ] defense theories.” Appellant Br. 41. No Supreme Court precedent clearly establishes this rule, however. In fact, as already explained, the Supreme Court's holdings cut the other way. See Harrington, 562 U.S. at 111, 131 S.Ct. 770; see also Dunn, 141 S. Ct. at 2411 (discussing the manifold ways in which counsel may choose to present a theory).

Lastly, DeBruyn argues that his lawyers performed incompetently because his “jury” “never” heard “evidence” that Gesiakowski could have died from something “other than oxycodone.” Appellant Br. 41. The trial transcript indicates otherwise. The jury heard plenty of evidence that Gesiakowski could have died from something unrelated to oxycodone—serotonin syndrome, as one example.

In sum, the state court of appeals reasonably determined that DeBruyn's lawyers performed competently in presenting their defense through cross-examination of the prosecution's experts.

Prejudice. In state court (and here), DeBruyn argued that expert testimony would have enhanced his arguments on four medical topics, each of which would have strengthened his defense that oxycodone did not play a substantial role in Gesiakowski's death. After reviewing the proposed testimony, the state court of appeals concluded that the jury likely would not have been swayed, meaning no prejudice occurred. Because DeBruyn's proposed expert testimony was “largely speculative,” the court reasoned, “there would” not “have been a reasonable probability of a different outcome” had his lawyers presented the expert “at trial.” DeBruyn, 2022 WL 981281, at *9. DeBruyn paints the state court's decision as both an unreasonable determination of the facts as well as an unreasonable application of clearly established law. Again, we disagree.

Topic one: opioid residual tolerance and re-tolerance. DeBruyn first focuses on his experts’ proposed testimony about opioid residual tolerance and re-tolerance. At the Ginther hearing, DeBruyn's experts said that opioid users can have “some degree of residual tolerance” to oxycodone even if they do not use the drug for some period of time. Ginther Hr'g, R.7-16, PageID 1896. They added that former opioid users can redevelop a tolerance to oxycodone “very quickly.” Ginther Hr'g, R.7-17, PageID 2017. But beyond those general statements, the experts did not opine on Gesiakowski's specific tolerance. Nonetheless, this testimony, according to DeBruyn, likely would have changed the outcome of his trial because it “could have precluded oxycodone from being a substantial factor in” Gesiakowski's death. Reply Br. 19. Why? Because, he says, the jury could have found that Gesiakowski had a residual tolerance (or developed a tolerance) to oxycodone by the time of her death on April 12.

To begin, the state court did not commit an unreasonable mistake in finding that “Gesiakowski's potential for opioid tolerance was not” adequately “supported.” DeBruyn, 2022 WL 981281, at *7. Nothing in the record shows that Gesiakowski had a residual tolerance to 180 nanograms of oxycodone (the amount in her blood) by the time of her death. Nor does the record reveal that Gesiakowski had developed a tolerance to that amount of oxycodone over the five days between her release from jail and her death.

Based on this finding, it was reasonable for the state court to conclude that it was not substantially likely that testimony about opioid residual tolerance and re-tolerance would have changed the jury's verdict. A fairminded jurist could conclude that the “theoretical possibility” that Gesiakowski had been tolerant enough to the 180 nanograms of oxycodone in her blood would not convince any juror that oxycodone was not a substantial factor in her death. Harrington, 562 U.S. at 112, 131 S.Ct. 770.

Topic two: serotonin syndrome. DeBruyn next maintains that expert testimony would have enhanced his trial defense that Gesiakowski died from serotonin syndrome. But a fairminded jurist could conclude that expert testimony about serotonin syndrome would not have likely changed the outcome of his trial, as it would have been cumulative of evidence already presented by his lawyers. See Cullen, 563 U.S. at 200, 131 S.Ct. 1388.

Through cross-examination, DeBruyn's lawyers made the case that Gesiakowski could have died from serotonin syndrome. Tramadol and fluoxetine, they said, can combine in dangerous ways and produce serotonin syndrome, a “potentially fatal” condition where a person's body cannot absorb serotonin. Jury Trial Tr., R.7-7, PageID 604. In that instance, a person can die from “an extremely high temperature, with increased body temperature, agitation, and restlessness as the common symptoms. Id. at PageID 531. And Gesiakowski, DeBruyn's lawyers argued, showed the signs of serotonin syndrome. Her behavior in the hotel room suggested she was hot and agitated—hallmark symptoms of serotonin syndrome. Coupled with the tramadol and fluoxetine in her blood at the time of her death, these facts pointed toward serotonin syndrome as the cause of death.

What more information could DeBruyn's expert witness have told the jury about serotonin syndrome? Something important, DeBruyn says. He points out that one of the prosecution's experts wrongly testified at trial that oxycodone can lead to serotonin syndrome. That, he says, left the jury with the mistaken impression that, even if Gesiakowski died from serotonin syndrome, oxycodone still played a substantial role in her death. The problem for DeBruyn, however, is that his lawyers did challenge the State's expert on this point at trial. After the State's expert testified that oxycodone can lead to increased serotonin levels, DeBruyn's lawyers confronted him with two international studies concluding that oxycodone “does not affect serotonin levels.” Jury Trial Tr., R.7-7, PageID 578. Faced with these studies, the State's expert conceded that if the studies were true, he would “certainly ․ change” his “opinion.” Id. at PageID 579. But even when apprised of those studies, the jury still found DeBruyn guilty. All in all, because DeBruyn's proposed expert testimony on serotonin syndrome would have been cumulative, a fairminded jurist could conclude it would not have likely changed the outcome of his trial. See Cullen, 563 U.S. at 200, 131 S.Ct. 1388.

Topic three: difluoroethane. DeBruyn next maintains that an expert would have enhanced his argument at trial that Gesiakowski died from difluoroethane. At trial, recall, the defense argued that Gesiakowski could have died from a heart problem caused by inhaling too much difluoroethane. One of the State's experts, however, testified that Gesiakowski had only 2.2 micrograms per milliliter of difluoroethane in her blood at the time of her autopsy. That is a small amount, he said, especially considering that people have died from difluoroethane with “concentrations [of the chemical around] 29 micrograms.” Jury Trial Tr., R.7-7, PageID 512. The defense countered this point at trial by arguing that Gesiakowski could have had a much higher concentration of difluoroethane in her blood at the time of her death (even 29 micrograms) because the chemical is volatile and easily escapes from the blood.

Turn, then, to the Ginther hearing. There, DeBruyn presented evidence that 5.3 micrograms per milliliter of difluoroethane can cause someone to die from a lethal heart problem. This evidence, DeBruyn argues, “would have made it significantly more plausible that” Gesiakowski “died from huffing [difluoroethane] alone.” Reply Br. 20.

Maybe so. But DeBruyn's argument fails under AEDPA. At trial, the State's experts said that it is impossible to know with certainty how much difluoroethane Gesiakowski had in her blood when she died. She could have had 2.2 micrograms (the amount in her blood at the time of her autopsy). Or she could have had a higher amount. Thus, because DeBruyn cannot show that Gesiakowski had more than 5.3 micrograms of difluoroethane in her blood when she died, the amount he presented as potentially lethal, a fairminded jurist could find that his new evidence would not have moved the jury. After all, it established nothing more than a “theoretical possibility” that Gesiakowski died from difluoroethane. Harrington, 562 U.S. at 112, 131 S.Ct. 770.

Topic four: seizure. DeBruyn, lastly, emphasizes his experts’ testimony that Gesiakowski could have died from a seizure. At the Ginther hearing, DeBruyn's experts pointed out that Gesiakowski had been prescribed anti-seizure medication (Keppra) in jail and that she had no such medication in her blood at the time of her death. And, they added, if a person “receiving anti-seizure treatment for a seizure condition[ ] abruptly” stops taking her medication, she could be prone to having a seizure. Ginther Hr'g, R.7-17, PageID 2069. DeBruyn takes this testimony and runs with it. Because Gesiakowski “abruptly stopp[ed]” taking anti-seizure medication, and because she had a seizure condition, DeBruyn says, she had an “increased risk of a fatal seizure,” allowing the jury to find that she died from a seizure, not oxycodone. Reply Br. 22.

A fairminded jurist, however, could find that DeBruyn's testimony about anti-seizure medication established “nothing more than a theoretical possibility” that Gesiakowski died from a seizure. Harrington, 562 U.S. at 112, 131 S.Ct. 770. That theory, in turn, rests on several factual assumptions. One is that Gesiakowski had abruptly stopped taking her medication, a point for which DeBruyn offered no supporting evidence. Yet it is similarly possible that she had stopped taking the medication well before her release from jail. Another is that Gesiakowski had been prescribed anti-seizure medication in jail for a seizure condition. But what if she had been prescribed that medication, as one of the State's experts pointed out at the Ginther hearing, to help her cope with opioid withdrawal?

In sum, a fairminded jurist could conclude that the expert medical testimony DeBruyn offered at his Ginther hearing would not have likely moved his jury. Thus, he cannot show that the state court of appeals unreasonably applied Strickland’s prejudice prong.

* * * * * *

We affirm the district court's judgment and deny the writ of habeas corpus.

READLER, Circuit Judge.