OPINION
A seizure disorder unfortunately ended Dr. Jean-Philippe Autrans career as a top-notch research scientist with Procter & Gamble. In the ensuing years, Autran received total-disability benefits under the Procter & Gamble Health and Long-Term Disability Plan (the “Plan”). The Procter & Gamble Disability Committee (the “Committee”) later terminated these benefits after concluding that Autran no longer qualified as totally disabled within the meaning of the Plan. Autran sued to overturn the Committees benefits decision under the Employee Retirement Income Security Act (ERISA). Yet the Plan delegates discretionary authority to the Committee to decide benefits claims, so we must review its decision under the deferential arbitrary-and-capricious test. And the Committee had rational reasons to depart from the earlier total-disability finding. Among other new evidence, a doctor who performed many objective tests on Autran for over six hours found no basis to conclude that he suffered from a debilitating condition. We thus affirm the district courts summary-judgment decision for the Committee.
I
After obtaining graduate degrees in physics and chemistry, Autran eventually took a job as a research scientist with Procter & Gamble. He was a high performer at the company for well over a decade. But radiating chest pain forced him to the emergency room in December 2008. Autran had suffered an aortic aneurysm that required emergency surgery. Although the surgery repaired his aorta, it deprived his brain of oxygen for a lengthy period and put him at risk of developing seizures. Autran returned to his job at Procter & Gamble several months later. Yet the surgery affected his mental acuity, and he struggled to perform at his previously high level.
Autrans seizure risk materialized in 2010. He began to experience seizures that made him act strangely. Autran would become disoriented and fatigued during these episodes, and he would not remember his actions. Over the next two years, the episodes happened more and more frequently. By 2012, a neurologist diagnosed him with a seizure disorder. Frustrated with his increasing mental difficulties, Autran decided that he could no longer work.
In October 2012, Autran applied for disability benefits under the Plan at the encouragement of his Procter & Gamble colleagues. The Plan offers both total-disability and partial-disability benefits. It defines “Total Disability” to cover any “mental or physical condition” that, among other things, “is generally considered totally disabling by the medical profession[.]” Admin. R. (A.R.) 2122. It defines “Partial Disability” to cover any “mental or physical condition” that, among other things, renders participants unable to “perform [the] regular duties” of their “current job” but that allows them to “perform other roles” within or outside the company. A.R. 2121. The Plans administrators decided that Autran qualified as “totally disabled” under these definitions and awarded him total-disability benefits.
To keep receiving these benefits under the Plan, though, Autran needed to regularly submit objective medical evidence proving his continued total disability. For years, Autran and the Plans administrators have disagreed over whether he met this ongoing duty.
Their dispute started in 2014. Unsatisfied with the records from Autrans treating doctors, the Plans administrators asked him to visit two independent doctors for medical evaluations. Both doctors opined that Autran was, at most, partially disabled. Beginning in February 2015, therefore, the administrators told Autran that he could receive only partial-disability benefits. Autran appealed this decision to the Plans Trustees, as he had a right to do under the Plan. During the appeal, two other doctors reviewed Autrans file and agreed that he was not totally disabled. Based on this evidence, the Trustees affirmed the decision to switch Autrans payments to partial-disability benefits.
This change lasted only a matter of months because of Autrans successful efforts to obtain federal disability benefits under the Social Security system. He had applied for those benefits in 2013, as required by the Plan. At first, federal personnel denied his claim. But on the same day that the Trustees affirmed Autrans switch to partial-disability benefits, an administrative law judge found that he met the “disability” test under federal law. About a month later, the Social Security Administration awarded Autran over $72,000 in benefits retroactive to March 2013. When the Plans administrators learned of this award, they notified Autran that he owed the Plan roughly the same amount to offset the Plans overpayments in light of the retroactive Social Security award. They also told him that his 52-week lifetime limit on partial-disability benefits would expire in February 2016. Autran again appealed to the Trustees. Pending this appeal, two doctors conducted further file reviews. Both reviewers concluded that Autran was not totally disabled. One of them, neurologist Gregory Whitman, opined that Autran should be limited to “non-mentally demanding activities” and work only “4 hours per day” if he took another job. A.R. 902. Yet, relying primarily on the Social Security award, the Trustees changed their decision. In July 2016, they found Autran totally disabled under the Plan as of September 2015.
The reinstatement of Autrans total-disability benefits was also short lived. Records from his then-treating physicians left the Plans administrators unconvinced that he remained disabled. They scheduled two more independent examinations for him. When these appointments fell through (for reasons the parties dispute), the administrators terminated Autrans benefits in April 2017. Autran appealed a third time. Meanwhile, Procter & Gamble amended the Plan by transferring the duty to make final benefits decisions from the Trustees to the Committee (the current defendant).
The Committee ordered a supplemental file review from Dr. Whitman and rescheduled the two in-person examinations. Dr. Whitman found insufficient new evidence in Autrans file to depart from his prior opinion that Autran could work part time. Dr. Malcolm Spica conducted a lengthy neuropsychological evaluation of Autran. His series of tests revealed “no consistent neurocognitive weaknesses” and “performances well within normal limits[.]” A.R. 1737–38. Dr. Keith Caruso conducted an independent medical examination. He diagnosed Autran as meeting the criteria for depression and anxiety but opined that neither condition rendered Autran totally disabled.
In March 2018, the Committee affirmed in part and reversed in part the administrators’ decision to end Autrans benefits as of April 2017. It found that Autran did not meet the Plans total-disability definition. It decided, however, that Autran remained partially disabled. So it awarded him his remaining 19 weeks of partial-disability benefits.
Autran sued the Committee under ERISA seeking the more expansive total-disability benefits. See 29 U.S.C. § 1132(a)(1)(B). He tragically died in an accident during this lawsuit. Autrans son became the administrator of his estate and was substituted as the plaintiff. For simplicity, we will continue to refer to the plaintiff as Autran. The district court granted summary judgment to the Committee. Autran now appeals.
II
This cases procedural posture requires us to identify two standards of review—one for the Committees decision and the other for the district courts. Under ERISA, courts start with a presumption rooted in the common law of trusts that they will review a plan administrators denial of benefits de novo. See Firestone Tire & Rubber Co. v. Bruch, 489 U.S. 101, 112–15, 109 S.Ct. 948, 103 L.Ed.2d 80 (1989). Yet plans often delegate discretionary power to administrators to interpret plan terms and apply those terms to a participants circumstances. See id. at 115, 109 S.Ct. 948. For these plans, courts will review an administrators benefits denial under an arbitrary-and-capricious standard. See Frazier v. Life Ins. Co. of N. Am., 725 F.3d 560, 566 (6th Cir. 2013). In Autrans case, the parties agree that the Plan gives this discretionary power to the Committee, so we need not independently determine whether it does so. See McClain v. Eaton Corp. Disability Plan, 740 F.3d 1059, 1064 (6th Cir. 2014). We thus will review the Committees finding that Autran was not totally disabled under the arbitrary-and-capricious test.
That leaves the standard of review for the district courts decision. It held that the Committee did not act arbitrarily or capriciously. We review such a holding de novo. See Davis v. Hartford Life & Accident Ins. Co., 980 F.3d 541, 547 (6th Cir. 2020).
A
Under arbitrary-and-capricious review, courts give great deference to plan administrators. See McClain, 740 F.3d at 1064–65. We must uphold an administrators benefits decision as long as “it is the result of a deliberate, principled reasoning process and ․ supported by substantial evidence.” Davis, 980 F.3d at 547 (citation omitted). This test has two components—one substantive, the other procedural.
Substantively, plan administrators may reach only those conclusions that are supported by substantial evidence in the administrative record. See id. at 549; Schwalm v. Guardian Life Ins. Co. of Am., 626 F.3d 299, 308 (6th Cir. 2010). An administrators decision will pass muster under this substantial-evidence test if a rational person could conclude that the evidence was “adequate” to justify the decision. Davis, 980 F.3d at 549 (citation omitted). And when the record contains evidence that could support a rational decision in either direction (that a plan participant was or was not disabled), the administrators choice between this conflicting evidence cannot be considered arbitrary on substantive grounds. See, e.g., Jackson v. Blue Cross Blue Shield of Mich. Long Term Disability Program, 761 F. Appx 539, 544 (6th Cir. 2019); Schwalm, 626 F.3d at 308; Hurse v. Hartford Life & Accident Ins. Co., 77 F. Appx 310, 318 (6th Cir. 2003).
Procedurally, plan administrators must engage in reasoned decisionmaking. See Davis, 980 F.3d at 547. We have asked a variety of questions when deciding whether an administrator has done so, depending on the issues relevant to the appeal. To name a few: Did the administrator consider all the evidence or overlook evidence that cut the other way? Compare Balmert v. Reliance Standard Life Ins. Co., 601 F.3d 497, 504 (6th Cir. 2010), with Evans v. UnumProvident Corp., 434 F.3d 866, 877–79 (6th Cir. 2006). If the administrator departed from its earlier benefits ruling, did it adequately explain the change? Compare Morris v. Am. Elec. Power Long-Term Disability Plan, 399 F. Appx 978, 984–85 (6th Cir. 2010), with Kramer v. Paul Revere Life Ins. Co., 571 F.3d 499, 507–08 (6th Cir. 2009). If the Social Security Administration found that the participant was disabled under federal law, did the administrator consider this ruling? Compare Hurse, 77 F. Appx at 316–18, with Bennett v. Kemper Natl Servs., Inc., 514 F.3d 547, 553–54 (6th Cir. 2008). If the administrator credited certain doctors over others, did the credited doctors undertake a mere “file” review or conduct a thorough in-person evaluation? Compare Huffaker v. Metro. Life Ins. Co., 271 F. Appx 493, 502–03 (6th Cir. 2008), with Shaw v. AT&T Umbrella Benefit Plan No. 1, 795 F.3d 538, 550 (6th Cir. 2015). And did the administrator have a conflict of interest that affected its decision? Compare Judge v. Metro. Life Ins. Co., 710 F.3d 651, 663–64 (6th Cir. 2013), with Calvert v. Firstar Fin., Inc., 409 F.3d 286, 292–93, 297 (6th Cir. 2005). None of the potentially relevant factors is dispositive in its own right; we must weigh them all when deciding whether the administrators ultimate conclusion resulted from a rational process. See Metro. Life Ins. Co. v. Glenn, 554 U.S. 105, 117–18, 128 S.Ct. 2343, 171 L.Ed.2d 299 (2008); Schwalm, 626 F.3d at 308.
The Committees decision in Autrans case satisfies these deferential substantive and procedural standards. Because we must address whether the Committee reasonably found that Autran was not totally disabled within the meaning of the Plan (not in some abstract sense), we start with basics about the Plans key terms. See, e.g., Huffaker, 271 F. Appx at 502. The Plan defines “Total Disability” to require a participant to have a “mental or physical condition” that is considered “totally disabling by the medical profession[.]” A.R. 2122. It adds that a total-disability finding typically requires serious limitations: “Usually, Total Disability involves a condition of such severity as to require care in a hospital or restriction to the immediate confines of the home.” Id. And it places the burden on participants to establish through “objective medical evidence” that they satisfy the demanding total-disability definition. A.R. 2125, 2128; A.R. 2493.
As a substantive matter, the record contained more than adequate evidence for the Committee to conclude that Autran did not meet this definition. See Davis, 980 F.3d at 549. Three pieces of evidence illustrate the point. Start with Dr. Spicas neuropsychological evaluation. Spica spent over 6 hours interviewing Autran and conducting some 17 tests. The objective testing showed substantial malingering on Autrans part. Yet Autran still performed “well within normal limits,” a performance that was “not consistent with a debilitating disorder.” A.R. 1737–38. Or consider Dr. Carusos independent medical examination. Caruso interviewed Autran for over three hours and found “no evidence of significant cognitive deficits[.]” A.R. 1791. Although Autran met the criteria for depression and anxiety, Caruso opined that these ailments did not render Autran totally disabled (and certainly did not require him to be hospitalized or confined to his home). A.R. 1792. To the contrary, a “clearly very intelligent” Autran could undertake most of lifes daily activities without aid. Id. He managed his finances, cooked, exercised (including biking and lifting weights), and engaged in hobbies (including painting, writing, poetry, and reading). A.R. 1788, 1792–93. Autran also admitted that he could still work in some capacity; he was presently working on “translating a book [from French to English] and writing a screenplay.” A.R. 1792. Dr. Whitmans supplemental file review pointed in the same direction. He suggested that Autran had provided no new evidence for Whitman to depart from his prior opinion that Autran could work part time. A.R. 1692–93.
As a procedural matter, the Committee engaged in reasoned decisionmaking. See Davis, 980 F.3d at 547. Its opinion made clear that it had considered “all the records” for Autrans claim; it did not ignore evidence that went against its decision. A.R. 1816–17. In addition, the Committee relied primarily on high-quality evaluations by Drs. Spica and Caruso; it did not rest its conclusion on file reviews. Id. The Committee also confronted the contrary disability finding by the Social Security Administration, explaining that the federal agency follows disability standards different from those in the Plan. A.R. 1818. And even though the Trustees had previously found that Autran was totally disabled, they did not have the updated medical evaluations on which the Committee relied. A.R. 1816.
All told, thorough medical opinions gave the Committee a firm foundation to conclude that Autran did not, in the Plans words, suffer from a “mental or physical condition” that the “medical profession” would consider “totally disabling.” A.R. 2122. And the Committee adhered to a rational process when reaching that conclusion.
B
Autran responds with six procedural reasons why the Committee erred. (He also argues that the Committee lacked substantial evidence for its decision, but his substantive challenge merely incorporates (and so turns on) his procedural claims.) Specifically, Autran argues that the Committee: (1) wrongly departed from the Trustees’ prior decision; (2) cherry-picked helpful information and ignored harmful information; (3) failed to adequately distinguish the Social Security Administrations decision; (4) failed to identify a job that Autran could perform; (5) ignored the side effects of his medications; and (6) acted with an improper conflict of interest. We will consider each objection in turn.
1. Did the Committee arbitrarily change course? Autran initially argues that the Committee applied the Plans total-disability definition inconsistently and that he experienced no material health change between July 2016 (when the Trustees found that he was totally disabled) and March 2018 (when the Committee found that he was not). He is mistaken.
To be sure, Autran correctly notes that a plan administrator acts arbitrarily if it haphazardly departs from a prior decision with no rational explanation for the change. See McCollum v. Life Ins. Co. of N. Am., 495 F. Appx 694, 704 (6th Cir. 2012); Kramer, 571 F.3d at 507–08. Yet administrators are not rigidly locked into their initial benefits decision, especially when new medical evidence comes to light or the participants health changes. See Rose v. Hartford Fin. Servs. Grp., Inc., 268 F. Appx 444, 454 (6th Cir. 2008); see also Davis, 980 F.3d at 548–49; Saunders v. Procter & Gamble Health & Long-Term Disability Benefit Plan, 659 F. Appx 272, 277 (6th Cir. 2016). Such an “illogical” rule would harm participants by making administrators wary of awarding benefits at the outset for fear that they could not stop the flow of benefits if new evidence shows that their initial decision was wrong. Rose, 268 F. Appx at 454. To make a benefits change, then, an administrator need only identify a “rational” reason for it. Morris, 399 F. Appx at 984.
The Committee met this low bar. It had new information before it in 2018 that the Trustees lacked in 2016. A.R. 1816–17. When the Committee initially reviewed Autrans case file without the evaluations from Drs. Spica and Caruso, it realized that it needed to have these evaluations before it could reach a proper decision. A.R. 1816. So it arranged for their scheduling. Id. And these new evaluations “at the beginning of 2018” unambiguously failed to show any significant neurocognitive weaknesses in Autran. A.R. 1817. The Committee acted perfectly rationally by relying on this new evidence to make the change. See Morris, 399 F. Appx at 984.
Autran counters that the Committees change was still arbitrary because the new evidence did not show an improvement in his condition between 2016 and 2018. He is wrong on the facts and the law. Factually, the Committee cited evidence that pointed to better mental functioning. A.R. 1816. Dr. Spica, for example, noted that Autrans test results showed improvement from prior results in 2014. A.R. 1738. Autran also told Spica that “his seizure disorder” was currently “well-controlled” and that he had “not had a seizure” for many months. A.R. 1731. In his supplemental file review, Dr. Whitman likewise suggested that “there may have even been some improvement” in Autrans condition since his last file review. A.R. 1692.
Legally, the Committee did not need to rest its change from the Trustees’ earlier decision on the ground that Autrans condition had improved. See Davis, 980 F.3d at 548. It is also rational for a plan administrator to make a change based on, say, better evidence showing the extent of the participants condition, Morris, 399 F. Appx at 984, or an acknowledgement that the administrators original decision was simply mistaken, Saunders, 659 F. Appx at 277. And here, the new evidence unambiguously showed that Autran was not totally disabled.
Autran next argues that the Committee has not adhered to a consistent interpretation of the Plans total-disability definition. In this case, the Committee argued that the definition lacks an inability-to-work (or “occupational”) element and that the definition turns on whether medical professionals would deem a condition totally disabling. In another case before this court, by contrast, the Committee relied on a participants ability to work as the basis to find that he was not totally disabled. See Lloyd v. Procter & Gamble Disability Benefit Plan, Plan #501, 2021 WL 4026683, at *6–8 (6th Cir. Sept. 3, 2021). These facially inconsistent positions, however, are rationally reconcilable. As we explained in Lloyd, the Committee could reasonably find “that the medical profession would include a work-capacity component among the relevant criteria” in deciding whether a condition is totally disabling. Id. at *7. The Committee took the same approach here. It did not treat an inability to work as a required element for a total-disability finding; it found that Autrans ability to work was evidence that he was not totally disabled. And the Committee could rationally conclude that this ability-to-work evidence reinforced its conclusion that Autran did not meet the Plans rather stringent total-disability definition.
2. Did the Committee arbitrarily fail to consider the whole record? Autran also claims that the Committee considered only the new evidence from Drs. Spica, Caruso, and Whitman, and ignored his treating physicians’ contrary findings. He is mistaken here too.
Autran again correctly identifies the background law: administrators may not selectively review the administrative record by picking out the opinions of the doctors that support their decisions while ignoring the opinions of a participants treating doctors that do not. See Shaw, 795 F.3d at 548–49. Administrators instead must consider all opinions on both sides of a disputed disability question. See Evans, 434 F.3d at 876, 879. That said, administrators need not give “special weight” to a treating physicians opinions over the opinions of others. See Black & Decker Disability Plan v. Nord, 538 U.S. 822, 834, 123 S.Ct. 1965, 155 L.Ed.2d 1034 (2003); Balmert, 601 F.3d at 504. Rather, as long as administrators offer a “reasoned explanation” for crediting independent physicians over treating physicians, we must respect their choice. Davis, 980 F.3d at 548 (quoting McDonald v. W.-S. Life Ins. Co., 347 F.3d 161, 169 (6th Cir. 2003)). And we have repeatedly found such a reasoned explanation when objective medical evidence did not support the treating doctors’ opinions. See Castor v. AT&T Umbrella Benefit Plan No. 3, 728 F. Appx 457, 466 (6th Cir. 2018); Judge, 710 F.3d at 660–61; Oody v. Kimberly-Clark Corp. Pension Plan, 215 F. Appx 447, 452 (6th Cir. 2007); Yeager v. Reliance Standard Life Ins. Co., 88 F.3d 376, 382 (6th Cir. 1996).
For a second time, the Committee met this low bar. As an initial matter, it did not “totally ignore” the records from Autrans treating physicians—Dr. Dinu Nodit, a neurologist, and Dr. Candice Drewry, a primary-care physician. Balmert, 601 F.3d at 504. It expressly cited them. A.R. 1816. Admittedly, Dr. Nodit provided a letter expressing his belief that Autran was not “able to return to work even part-time” because Autran could not focus for more than a couple hours or drive. A.R. 1152. Yet the Committee had a “reasoned” basis for crediting the opinions of the independent physicians over Dr. Nodits opinion. Davis, 980 F.3d at 548 (citation omitted). The Plan requires Autran to rely on “objective” evidence—that is, evidence “that can be observed or verified by someone other than” himself. The American Heritage Dictionary of the English Language 1215 (5th ed. 2011); see A.R. 2128. And, as the Committee explained, Dr. Nodits opinion was contradicted by the objective findings from his own examinations of Autran. A.R. 1816. Those examinations showed that Autran had largely normal characteristics, including normal “[m]ental [s]tatus,” “attention span and concentration,” “[o]rientation ․ to person, place, and time,” and “basic gait[.]” A.R. 1271. Dr. Drewrys evaluations likewise showed largely normal findings. A.R. 1074, 1130. Autran himself also told Dr. Spica that he generally “drives without difficulty[.]” A.R. 1732. In short, the objective evidence backed up the opinions of the independent physicians more than they did Dr. Nodits opinion. Cf. Judge, 710 F.3d at 660–61; Balmert, 601 F.3d at 504.
Autran insists that the Committee overlooked other portions of his physicians’ records. He, for example, points to a list of symptoms in Dr. Nodits examination notes, including “difficulty walking, memory loss, imbalance or falling, confusion, syncope, and dizziness[.]” A.R. 1270. Autran leaves out a critical element: The symptoms were prefaced with the qualifying phrase “[p]atient reports.” Id. That is, the records listed subjective symptoms perceived by Autran, not objective evidence perceived by others. The Committee could rationally accept the findings of the independent physicians (which were supported by a slew of objective tests) over his own doctors’ opinions (which were supported by subjective self-assessments). See Castor, 728 F. Appx at 466.
3. Did the Committee wrongly depart from the Social Security decision? Autran next says that the Committee arbitrarily refused to follow the Social Security Administrations decision finding him disabled under federal law. He cites our cases holding that an administrator cannot ignore a federal disability finding if the plan requires a participant to apply for Social Security benefits and if such a federal award helps the plan (by reducing the amount that it owes the participant). See, e.g., Bennett, 514 F.3d at 553–54. Autran proved some of these factors here. The Plan requires participants to apply for Social Security benefits. It also benefitted financially from Autrans federal award because administrators made him return over $70,000 in disability payments that they had distributed before that retroactive award. But Autran failed to prove the most important factor: The Committee did not ignore the federal disability finding. See OBryan v. Consol Energy, Inc., 477 F. Appx 306, 308 (6th Cir. 2012) (per curiam).
To the contrary, the Committee “considered” the federal finding and gave rational reasons for rejecting it. A.R. 1818. For one thing, the Committee uses different “criteria” to “determine” a participants total disability than the criteria used by the Social Security Administration to determine a persons disability under federal law. Id. Federal law strictly follows an occupational model that ties an applicants disability to the applicants ability to work. See 42 U.S.C. § 423(d). The Plan, by contrast, asks whether the medical community would view a condition as totally disabling. A.R. 2122. While the Plans administrators have relied on ability-to-work evidence when deciding whether a participant meets this medical-community test, that test is not strictly occupational in the same way as the federal test. See Hurse, 77 F. Appx at 317–18.
For another thing, at the time of Autrans favorable Social Security decision, the Social Security Administration adhered to a “treating physician rule” that gave more weight to the opinions of a persons treating physicians than to the opinions of other doctors. A.R. 1818; see 20 C.F.R. § 416.927(c)(2); but cf. 20 C.F.R. § 416.920c(a) (eliminating rule for claims filed on or after March 27, 2017). Under the Plan, by contrast, the Committee did not need to give more weight to Autrans treating physicians. See Black & Decker, 538 U.S. at 831–34, 123 S.Ct. 1965; Whitaker v. Hartford Life & Accident Ins. Co., 404 F.3d 947, 949 (6th Cir. 2005).
For a third, the Committee relied on evidence that the Social Security Administration did not have before it when finding that Autran was disabled. A.R. 1816–17. Drs. Spica and Caruso, for example, conducted their in-person evaluations after the federal decision. See OBryan, 477 F. Appx at 308. The Committee thus rationally explained its departure from that decision.
4. Did the Committee need to identify a job that Autran could perform? Autran next suggests that our cases required the Committee to identify a specific job that he could perform in order to find that he was not totally disabled. He misreads our law. Unlike the Plan, other ERISA plans often directly tie a participants total disability to the participants inability to work—that is, these plans follow a disability test more like the federal occupational test. See, e.g., Holden v. Unum Life Ins. Co. of Am., 2021 WL 2836624, at *4 (6th Cir. July 8, 2021); Judge, 710 F.3d at 654; Douglas v. Gen. Dynamics Long Term Disability Plan, 43 F. Appx 864, 866 & n.2 (6th Cir. 2002). Even for these plans, however, we have rejected a legal rule requiring administrators to introduce vocational evidence identifying jobs that participants can perform. See Judge, 710 F.3d at 662–63. So it would make no sense for us to require the Committee to identify such jobs given that the Plans terms place less weight on a participants ability to work.
Neither of the cases on which Autran relies adopts a contrary rule. See Brooking v. Hartford Life & Accident Ins. Co., 167 F. Appx 544, 548–49 (6th Cir. 2006); McDonald, 347 F.3d at 170–72. McDonald overturned a plan administrators decision on the evidentiary ground that the administrator lacked sufficient evidence, not the legal ground that it needed to identify a job that the participant could perform. See 347 F.3d at 170–71. We reasoned that the administrator arbitrarily relied on a doctors speculative claim that the participant might one day return to work because the “overwhelming evidence” showed the contrary. Id. at 171. In Brooking, the administrator had identified specific jobs that the participant could perform, but its conclusion that she could perform those jobs conflicted with the objective evidence. 167 F. Appx at 548–49. Nowhere did Brooking suggest that an administrator must identify such a job if the objective medical evidence showed that the participant did not have a health condition that prevented work. See id.
5. Did the Committee wrongly ignore the side effects of Autrans medications? Autran next argues that the Committee overlooked that he was taking medications with side effects. True, if an administrator ignored a treating physicians records objectively demonstrating that a participants medications have caused debilitating side effects, we have found that the administrator acted arbitrarily. See Godmar v. Hewlett-Packard Co., 631 F. Appx 397, 405–06 (6th Cir. 2015); Smith v. Contl Cas. Co., 450 F.3d 253, 265 (6th Cir. 2006). If, by contrast, other individuals could not verify the side effects (that is, if they were not rooted in objective evidence), those effects have provided no basis to overturn an administrators decision. See Fenwick v. Hartford Life & Accident Ins. Co., 841 F. Appx 847, 854 (6th Cir. 2021).
Autrans evidence falls on the wrong side of this line. He relies on his own subjective assessments that his medications were causing side effects. He, for example, sent an administrator a link to the page on rxlist.com for one medication, alleging that he had virtually all the listed side effects. But he does not point to any objective evidence reinforcing his claim that his medications were causing these effects. See Rothe v. Duke Energy Long Term Disability Plan, 688 F. Appx 316, 320 (6th Cir. 2017); Schwalm, 626 F.3d at 313.
Autrans failure to identify objective evidence distinguishes his case from those on which he relies. In these other cases, physician records confirmed that a patient had been experiencing the health impacts of the prescribed drugs. See Godmar, 631 F. Appx at 405–06; Smith, 450 F.3d at 264. But the records in this case do not confirm Autrans subjective assessments. In these other cases, moreover, many factors influenced the courts conclusion that the plan administrator had acted arbitrarily; the court did not rest on the side effects of the medications alone. Take, for example, Godmar. There, the administrators failure even to mention the participants medications was just one factor among many other shortcomings that rendered the decision arbitrary and capricious. 631 F. Appx at 406–07; see also Smith, 450 F.3d at 261–65. That is not the case here. The Committee relied on objective evidence from multiple sources to terminate Autrans total-disability benefits, and its failure to expressly discuss Autrans subjective complaints of side effects cannot suffice to render its decision arbitrary. Cf. Fenwick, 841 F. Appx at 854.
6. Did the Committees conflict of interest affect its decision? Autran lastly argues that the Committee acted with a conflict of interest when finding that he was not totally disabled. The Supreme Court has recognized that an inherent conflict of interest exists when a defendant plays the “dual role” of both deciding a plan participants eligibility for benefits and paying out those benefits from its own coffers. See Glenn, 554 U.S. at 108, 128 S.Ct. 2343. And here, the Committee agrees that it had such an inherent conflict when deciding claims and paying benefits on behalf of the Plan.
The Supreme Court has held that this type of conflict must be a “factor” to consider when deciding whether an administrator arbitrarily denied benefits. See id. at 115–19, 128 S.Ct. 2343. But we have since emphasized that “conclusory allegations of bias” based on this (relatively common) inherent conflict do not deserve much weight. See Judge, 710 F.3d at 664 (citation omitted). A participant instead should attempt to uncover evidence suggesting that the conflict materialized in a concrete way to influence the administrators decisional process. See Frazier, 725 F.3d at 570; Cooper v. Life Ins. Co. of N. Am., 486 F.3d 157, 165 (6th Cir. 2007). This type of concrete evidence can take the form of, for instance, internal emails suggesting bias against granting a claim or repeated reliance on a suspect doctor with a history of questionable medical opinions. See Holden, 2021 WL 2836624, at *17 n.21 (citing Elliott v. Metro. Life Ins. Co., 473 F.3d 613, 620 (6th Cir. 2006); Evans, 434 F.3d at 880).
Autran has pointed to no such evidence. To the contrary, the Committee receives training to ensure that it does not take Procter & Gambles financial interests into account when resolving claims, and administrators ensure that the independent doctors who evaluate claims receive the same pay no matter their disability findings. The administrators in Autrans case also sought to use a new doctor with each new evaluation, which further minimized the risk that the Committee would over-rely on any one (potentially biased) source. The Committees inherent conflict of interest thus did nothing to undermine its conclusion in this case. See Judge, 710 F.3d at 664.
* * *
Because Autran has failed to prove that the Committee acted arbitrarily, we affirm.
MURPHY, Circuit Judge.