Esther White, the beneficiary of David White's life-insurance policy, appeals a summary judgment granted in favor of the insurer and plan administrator, Life Insurance Company of North America ("LINA"), on her claim for benefits. LINA had denied benefits on the ground that David's death was caused in part by intoxication or drug abuse. Finding that LINA abused its discretion in denying benefits, we reverse and instruct the district court to enter judgment in favor of White.

I.

In July 2014, David and Esther White were in a horrible car crash in Arkansas. David was driving. As the highway curved right, David inexplicably kept going straight. He thus crossed three lanes of traffic, including the center divider line, and then collided head-on with an oncoming eighteen-wheeler truck. David died as a result. Esther is the beneficiary of his life-insurance policy.

In this section, we briefly state the relevant policy provisions. We then turn to the evidence in the administrative record. After that, we describe the questionable administrative proceedings giving rise to this appeal.

A.

David was insured under two life-insurance policies issued by LINA. LINA both insures the plans and determines entitlement to benefits. Esther, David's widow and the plaintiff here, is the beneficiary of those policies. Relevant to this appeal, both policies contain exclusions if death is caused , at least in part, either by "intoxication" as defined by Arkansas law, or by the "voluntary ingestion" of any "narcotic" or "drug" that is not prescribed.1

*765Under Arkansas law, a driver is "intoxicated" if he is "influenced or affected by the ingestion of alcohol [or] a controlled substance ... to such a degree that the driver's reactions, motor skills, and judgment are substantially altered and the driver, therefore, constitutes a clear and substantial danger of physical injury or death to himself or herself or another person." Ark. Code § 5-65-102(4).

Now for the facts in the administrative record.

B.

The crash occurred on July 26, 2014, at around 4:36 pm: Broad daylight; weather and road conditions clear; no speeding; vehicles were functioning properly.

At the scene of the crash, however, paramedics reported to the police that they smelled alcohol on David's breath. So the Arkansas State Police drew a blood sample and cited David for "Driving While Intoxicated" ("DWI"). On the collision report, however, the police also noted that it was "unknown" whether David was impaired at the time of the accident. Two hours later, at the hospital, another blood sample was taken to test for alcohol. And two hours after that, the hospital collected a urine sample for a drug-screen panel.

The hospital's toxicology analysis indicated that David tested negative for alcohol. The results did, however, reveal the presence, but not the amount, of a variety of controlled substances in David's system. Specifically, the drug screen indicated that David tested positive for amphetamines, cocaine, opiates, benzodiazepine, and cannabinoids. All of the toxicology reports indicated that these positives were only preliminary, non-quantitative results and that further confirmatory testing would be required to determine the level of drugs in David's system. No additional testing was requested by anyone.

On August 1, 2014, a few days after the crash, David died from a stroke. The coroner prepared a death certificate, which stated that the "immediate cause" of David's death was a "massive stroke," and that the "underlying cause[s]" of death were "multiple trauma," "cocaine abuse," and "amphetamine abuse." The death certificate also listed "marijuana abuse" as "other significant conditions contributing to death but not resulting in the underlying cause" of death.

In September 2014, the Arkansas State Crime Laboratory issued its own blood toxicology report. Like the hospital's toxicology reports, the police toxicology report indicated that David tested negative for alcohol but positive for benzodiazepine, cannabinoids, cocaine, and opiates. Also like the hospital's reports, the police toxicology report indicated that these results were only preliminary, non-quantitative results. The toxicology report indicated that if no additional testing was requested, the blood specimen would be destroyed after 90 days. No additional testing was requested within that 90-day window.

We turn now to the administrative proceedings.

C.

In January 2015, while processing Esther's claim for benefits under the life-insurance policies, LINA hired Dr. Fochtman, a toxicologist, to review her claim. Among other things, LINA asked Dr. Fochtman to "comment on any impairment Mr. White would have been experiencing at the time of his crash and to what extent might these impairments affect his driving abilities." So Dr. Fochtman reviewed the documents sent to him by LINA, including the toxicology results, the death certificate, and the collision report.

*766On January 19, Dr. Fochtman sent LINA his report. Relevant here, Dr. Fochtman noted that it was impossible to estimate the level of Mr. White's intoxication, and thus his level of impairment, at the time of the crash:

Since the only blood test done was an alcohol [test] that was negative and no blood tested for the presence of drugs, an estimation of Mr. White's level of impairment cannot be done. The drugs present in his urine only show that he had prior exposure and cannot be used to estimate a level of impairment. Further, the drug screen that was done on Mr. White's urine specimen only provided qualitative positive results. However, in the absence of any other cause of the collision, the drugs in his system could explain his level of impairment that resulted in his crash.

On January 26, White's attorney sent a letter to LINA, requesting "any and all documents you may rely on toward making your decision [on coverage] so [White] can pursue any and all legal remedies she may have afforded to her by law." LINA did not provide Dr. Fochtman's report to White.

On January 29, LINA called the Arkansas police to ask the basis of David's DWI citation. The police told LINA that they could release the results only to White. The police then faxed to LINA an amended collision report, which stated that Mr. White "was positive" for drugs at the time of the accident and, further, changed the "driver impairment" box from "unknown" to "impaired." LINA then asked White to obtain the blood-test results from the Arkansas police.

Two months later, the Arkansas police responded to White's request for the blood-test results, stating that no quantitative tests were performed on Mr. White's blood specimen. White reported this to LINA and also informed LINA that the specimen was likely destroyed.

A week later, LINA issued its initial denial of coverage. LINA's denial applied the exclusions for death caused, at least in part, by "intoxication" or the voluntary ingestion of "narcotics" or "drugs." White appealed through LINA's administrative process; LINA affirmed. Importantly, although the letters of denial relied entirely on the toxicology results, the death certificate, and the amended collision report (which included the DWI citation), neither of LINA's denials of coverage mentioned Dr. Fochtman's report.

In August 2015, White initiated this suit against LINA in federal district court. LINA answered the complaint and filed the administrative record with the court. For the first time, White discovered Dr. Fochtman's report in the documents LINA filed with the court. She then moved to supplement the record with an affidavit of the coroner who prepared Mr. White's death certificate. In the affidavit, the coroner explains that it was his practice, when filling in the blanks on a death certificate, to list every cause of death that was "merely possible," "so that their presence is simply acknowledged of record."

White's motion to supplement the record was denied. The parties then cross-moved for judgment on the record. The magistrate judge issued a recommendation in LINA's favor, holding that there was "substantial evidence" in the record to support LINA's finding that David's intoxication or drug abuse at least partly caused his own death. White submitted objections. The district court entered judgment for LINA, adopting the magistrate's recommendation.

White timely appealed the denial of her motion to supplement the record, as well as the judgment.

*767II.

This Court reviews the district court's grant of summary judgment de novo, applying the same legal standards that controlled the district court's decision. Robinson v. Aetna Life Ins. Co. , 443 F.3d 389, 392 (5th Cir. 2006). "[W]hen an administrator has discretionary authority with respect to the decision at issue, the standard of review should be one of abuse of discretion." Conn. Gen. Life Ins. Co. v. Humble Surgical Hosp., LLC , 878 F.3d 478, 483 (5th Cir. 2017) (quoting Vega v. Nat'l Life Ins. Servs., Inc. , 188 F.3d 287, 295 (5th Cir. 1999) (en banc), overruled on other grounds by Metro. Life Ins. Co. v. Glenn , 554 U.S. 105, 128 S.Ct. 2343, 171 L.Ed.2d 299 (2008) ). Here, the parties stipulated that the policies vested LINA with "discretionary authority to determine the eligibility for benefits and/or construe and interpret the terms of the plan." Accordingly, we review LINA's denial of benefits for abuse of discretion.

III.

"In reviewing the plan administrator's decision, we 'take into account ... several different considerations.' " Schexnayder v. Hartford Life & Accident Ins. Co. , 600 F.3d 465, 469 (5th Cir. 2010) (quoting Glenn , 554 U.S. at 117, 128 S.Ct. 2343 ). Such considerations "are case-specific and must be weighed together before determining whether a plan administrator abused its discretion in denying benefits." Id. Here, four considerations lead us to conclude that LINA abused its discretion in denying benefits: (A) LINA's conflict of interest; (B) LINA's failure to address Dr. Fochtman's report in any of its denials; (C) LINA's withholding of Dr. Fochtman's report; and (D) the closeness of the evidence to support LINA's determination that intoxication or drug abuse caused David's death.

A.

We first consider LINA's conflict of interest. When, as here, the insurer of the plan also determines whether the claimant is entitled to benefits, a conflict of interest arises. Glenn , 554 U.S. at 108, 128 S.Ct. 2343. Indeed, LINA concedes it has a conflict of interest. Thus, our task is determining the weight to give to LINA's conflict of interest in the overall review of the administrator's decision. See ids="3677099" index="12" url="https://cite.case.law/us/554/105/">id. "[C]onflicts are but one factor among many that a reviewing judge must take into account." Schexnayder , 600 F.3d at 470 (quoting Glenn , 554 U.S. at 116, 128 S.Ct. 2343 ). A conflict of interest, such as the one in this case, "should prove more important (perhaps of great importance) where circumstances suggest a higher likelihood that it affected the benefits decision." Glenn , 554 U.S. at 117, 128 S.Ct. 2343. Keeping this consideration in mind, we now turn to evaluate the remaining circumstances of this case.

B.

We are concerned with LINA's failure to address Dr. Fochtman's report in its denials of benefits. White argues that such failure amounts to "procedural unreasonableness." See Schexnayder , 600 F.3d at 469. "[P]rocedural unreasonableness is important in its own right and also 'justifie[s] the court in giving more weight to the conflict.' " Id. at 471 (quoting Glenn , 554 U.S. at 118, 128 S.Ct. 2343 ).

In Schexnayder v. Hartford Life & Accident Insurance Company , the Social Security Administration ("SSA") had determined that Schexnayder was "fully disabled" and unable to perform any work. Id. The SSA had issued disability payments to Schexnayder, who in turn had reimbursed the plan administrator, Hartford, *768for payments Hartford had already made to Schexnayder. Id. at 468-69. When Hartford later terminated Schexnayder's benefits on the ground that Schexnayder was not disabled, however, Hartford's denial did not mention the SSA's determination that Schexnayder was disabled. Id. at 471. "Because Hartford failed to acknowledge an agency determination that was in direct conflict with its own determination, its decision was procedurally unreasonable." Id. (footnote omitted). "We do not require Hartford to give any particular weight to the contrary findings; indeed, Hartford could have simply acknowledged the award and concluded that, based on the medical evidence before it, the evidence supporting denial was more credible. It is the lack of any acknowledgement which leads us to conclude that Hartford's decision was procedurally unreasonable and suggests that it failed to consider all relevant evidence." Id. at 471 n.3. We noted that, "[a]lthough Hartford based its decision on substantial evidence, we must consider other factors under Glenn , such as the conflict of interest and Hartford's treatment of the SSA award." Id. at 470. Thus, "[a]lthough substantial evidence supported Hartford's decision, the method by which it made the decision was unreasonable." Id. at 471. We concluded that Hartford abused its discretion and affirmed the district court's order that Hartford make payments under the plan. See Schexnayder v. CF Indus. Long Term Disability Plan for Its Employees , 553 F.Supp.2d 658, 668 (M.D. La. 2008), affirmed in relevant part by Schexnayder , 600 F.3d at 472.

Here, as in Schexnayder , it is undisputed that LINA did not address Dr. Fochtman's report. LINA argues that Schexnayder is distinguishable because the SSA determination in that case was in direct conflict with the administrator's findings, whereas in this case Dr. Fochtman's report was favorable toward LINA's finding that Mr. White's death was caused by his intoxication or drug abuse. We must disagree.

Dr. Fochtman's report effectively stated that the level of the drugs in Mr. White's system could not be determined, and thus whether the cause of Mr. White's death was due to intoxication or drug abuse could only be speculative. The inability to determine the level of drugs in Mr. White's system was critical to the application of the "intoxication" exclusion because Arkansas defines "intoxicated" as being influenced by alcohol or drugs "to such a degree " that the driver is "a clear and substantial danger" to himself and those around him. See Ark. Code § 5-65-102(4) (emphasis added). The inability to determine whether David was under the influence of alcohol or drugs at the time of the accident does not afford a reasonable conclusion that his death was caused by intoxication or drug abuse.

Thus, LINA should have at least addressed the report in its denials, especially because Dr. Fochtman's report was the only expert opinion in the record. As in Schexnayder , "[i]t is the lack of any acknowledgement which leads us to conclude that [LINA's] decision was procedurally unreasonable and suggests that it failed to consider all relevant evidence." See Schexnayder , 600 F.3d at 471 n.3. Accordingly, even if "substantial evidence supported [LINA's] decision, the method by which it made the decision was unreasonable." See ids="4223169" index="26" url="https://cite.case.law/f3d/600/465/#p469">id. at 471. That said, we proceed further in evaluating the appropriate action to take in this case.

C.

Thus, we consider LINA's actual withholding of Dr. Fochtman's report from Esther White. White argues that LINA's *769failure to provide her with Dr. Fochtman's report, despite her written request for "any and all documents" LINA would "rely on toward making [its] decision," violated ERISA's procedural rules and thus denied White a "full and fair review."

Section 1133 of ERISA provides: "In accordance with regulations of the Secretary, every employee benefit plan shall ... afford a reasonable opportunity to any participant whose claim for benefits has been denied for a full and fair review by the appropriate named fiduciary of the decision denying the claim." 29 U.S.C. § 1133. The relevant regulations provide that a "full and fair review" requires providing the claimant, upon her request, with documents "relevant" to her claim for benefits. See 29 C.F.R. § 2560.503-1(h). The regulations also define "relevant" in this context: any document "submitted, considered, or generated in the course of making the benefit determination, without regard to whether such document, record, or other information was relied upon in making the benefit determination." Id. § 2560.503-1(m)(8).

Here, Dr. Fochtman's report was clearly "relevant" to White's claim, and White made a request for such documents. Thus, by not providing the report to White, LINA violated § 1133. Indeed, LINA does not contest that, technically, it violated ERISA's procedural regulations.

LINA reasons that it "substantially complied" with ERISA's procedural requirements. "Challenges to ERISA procedures are evaluated under the substantial compliance standard. This means that technical compliance with ERISA procedures will be excused so long as the purposes of section 1133 have been fulfilled." Robinson v. Aetna Life Ins. Co. , 443 F.3d 389, 392-93 (5th Cir. 2006) (quotations and citations omitted). LINA contends that the purpose of § 1133 is to "afford the beneficiary an explanation of the denial of benefits that is adequate to ensure meaningful review of that denial." See Wade v. Hewlett-Packard Dev. Co. LP Short Term Disability Plan , 493 F.3d 533, 539 (5th Cir. 2007) (quoting Schneider v. Sentry Grp. Long Term Disability Plan , 422 F.3d 621, 627-28 (7th Cir. 2005) ), abrogated on other grounds by Hardt v. Reliance Standard Life Ins. Co. , 560 U.S. 242, 130 S.Ct. 2149, 176 L.Ed.2d 998 (2010). Specifically, LINA argues that it did not have to turn over Dr. Fochtman's report when requested because the report would not have helped explain why LINA denied benefits. After all, LINA argues, White was already in possession of the documents addressed in Dr. Fochtman's report, including the non-quantitative toxicology results, the death certificate, and the collision reports. Moreover, according to LINA, the purpose of § 1133 has been fulfilled because Dr. Fochtman's report was favorable to LINA's position, not White's.

LINA's explanation is unsatisfactory. It is true that White was in possession of the toxicology results and other documents addressed in Dr. Fochtman's report. But that report also expressed the opinion that, without a quantitative drug test, "an estimation of Mr. White's level of impairment cannot be done." Thus, contrary to LINA's assertions, Dr. Fochtman's report undermines LINA's position that intoxication or drug abuse was the cause of David's death. Further, this opinion comes from the report of LINA's very own expert. As such, without Dr. Fochtman's report, during the administrative process White was unable to meaningfully challenge LINA's finding that David's death was caused by intoxication or drug abuse. We therefore conclude that LINA did not substantially comply with ERISA's procedural requirements and, consequently, denied White a "full *770and fair review."2

D.

Continuing our full review of the administrative decision, we next consider whether LINA's denial of benefits, on the merits, is supported by "substantial evidence." "Substantial evidence is 'more than a scintilla, less than a preponderance, and is such relevant evidence as a reasonable mind might accept as adequate to support a conclusion.' " Dutka ex rel. Estate of T.M. v. AIG Life Ins. Co. , 573 F.3d 210, 213 (5th Cir. 2009) (quoting Ellis v. Liberty Life Assurance Co. , 394 F.3d 262, 273 (5th Cir. 2004) ). "A plan administrator abuses its discretion where the decision is not based on evidence, even if disputable, that clearly supports the basis for its denial." McCorkle v. Metro. Life Ins. Co. , 757 F.3d 452, 457 (5th Cir. 2014) (quoting Holland v. Int'l Paper Co. Ret. Plan , 576 F.3d 240, 246 (5th Cir. 2009) ). Given the deferential standard afforded to LINA under the policy, White bears the burden to prove that substantial evidence does not support LINA's position. George v. Reliance Standard Life Ins. Co. , 776 F.3d 349, 352 (5th Cir. 2015).

One particular case is instructive here. In Dutka v. AIG , the plaintiff-beneficiaries sought to recover benefits under ERISA. 573 F.3d at 212. The policy in Dutka contained an exclusion virtually identical to the exclusions at issue in this case. See ids="3673271" index="42" url="https://cite.case.law/f3d/573/210/#p213">id. Flying at low altitude, the decedent pilot failed to maintain adequate airspeed, resulting in a stall from which there was no time to recover; he crashed, killing himself and his two passengers. Id. As in this case, in Dutka the plaintiffs challenged the administrator's determination that the decedent was "intoxicated" at the time of the crash and that the intoxication "caused" the crash. Id. at 213-14. Further bearing similarity to this case, the toxicology reports in Dutka disclosed "the presence of chemicals in the decedent's body consistent with the use of multiple drugs around the time of the accident," but also a "therapeutic dose" of a narcotic at the time. Id. at 214. We held that, with "good visual meteorological conditions" and "no evidence of mechanical failure," the pilot's failure to maintain airspeed at such a low altitude was "a fundamental piloting error" that, combined with the otherwise inconclusive toxicology analysis, made it reasonable to conclude that the accident was caused by the pilot's being under the influence of drugs. Id.

LINA argues that the similarities to Dutka are obvious. Here, toxicology reports indicated that David had been exposed to a variety of controlled substances. With good weather conditions and no evidence of mechanical failure, crossing over the center dividing line into oncoming traffic is a fundamental driving error. Combined with the inconclusive toxicology results, LINA argues that, under Dutka , it was reasonable to conclude David's death resulted from the drugs detected in his system.

White, however, argues that Dutka is distinguishable. White argues that the pilot's failure to maintain adequate speed at low altitude for a long enough time to cause a stall was a fundamental piloting error in Dutka ; here, on the other hand, the brief failure to negotiate a right-hand turn on a highway is not a comparable driving error. More persuasively, White *771argues that Dutka is distinguishable because in Dutka the expert gave the opinion that the pilot was impaired due to his "therapeutic dose" of narcotics. Indeed, the expert in that case gave the opinion that the decedent was under the influence of painkillers "at the time of the crash" and that he had "recently" used alcohol and cocaine. Id. at 213. Here, in contrast, the only expert in this case, Dr. Fochtman, gave the opinion that no level of intoxication could be determined from the available evidence.

This evidence presents a close call. In the light of Dr. Fochtman's report, the only clear evidence supporting LINA's application of the exclusion is that David went straight while the road curved right.3 Unlike the fundamental piloting error of failing to maintain adequate air speed while flying at a low altitude in Dutka , the mere failure to negotiate a right-hand turn on a highway is not, on its own, sufficient to support a rational conclusion that the driver was intoxicated while driving. See ids="3673271" index="48" url="https://cite.case.law/f3d/573/210/#p213">id. at 214. Indeed, unlike Dutka , the record here contains some non-drug-related reasons that the unusual accident occurred. See ids="3673271" index="49" url="https://cite.case.law/f3d/573/210/#p213">id. at 212. That said, there is still the unexplained evidence of controlled substances in David's system. This evidence, however, is only equivocally connected to the accident's cause.

In short, although the administrator's decision is supported by relevant evidence, we cannot deny that the evidence is close. Thus, in accordance with the Supreme Court's instruction in Glenn , and taking into account all facets of this case, we conclude that LINA's conflict of interest "affected the benefits decision," and, accordingly, we may not uphold its decision. See Glenn , 554 U.S. at 117, 128 S.Ct. 2343 ; see also Schexnayder , 600 F.3d at 471 (holding that even if "substantial evidence" supported the administrator's denial of coverage, the conflicted administrator's failure to address evidence in the record contrary to its denial was an abuse of discretion).

IV.

In sum, taking into account LINA's conflict of interest, its procedural unreasonableness, its denial of a full and fair review, and the counter-balanced nature of the evidence, we hold that LINA abused its discretion in denying benefits. See Glenn , 554 U.S. at 117, 128 S.Ct. 2343. Accordingly, we reverse the district court's judgment. We remand with instructions to enter judgment in favor of White and for such further proceedings as may be necessary and fully consistent with this opinion.

REVERSED and REMANDED.4