Kandace WILLIAMS, Plaintiff-Appellant, v. METROPOLITAN LIFE INSURANCE COMPANY, a New York insurance company; AT & T Long Term Disability Plan for Occupational Employees, an Erisa welfare benefit plan; AT & T Integrated Disability Service Center, Defendants-Appellees, 459 F. App'x 719

ORDER AND JUDGMENT TERRENCE L. O’BRIEN, Circuit Judge. In this dispute over long-term disability benefits, Kandace Williams appeals from the district court’s

summary judgment in favor of Metropolitan Life Insurance Company (MetLife), AT & T Long Term Disability Plan for Occupational Employees, and AT & T Integrated Disability Service Center.

Williams contends the claims administrator of her long-term disability plan abused its discretion because (1) its decision was contrary to the evidence of her inability to work; (2) it improperly ignored evidence favorable to her claim; (3) it relied on a flawed employability assessment; (4) it denied her benefits after acknowledging her disability by referring her to a Social Security disability advocate; and (5) it impermissibly terminated her benefits after initially approving them. We reject these contentions and affirm. I.

FACTUAL AND PROCEDURAL BACKGROUND Williams worked as a production assistant for AT & T. During her employment, she participated in AT & T’s self-funded long-term disability (LTD) plan.

She filed for benefits under the plan because she suffered from “chronic spinal impairment” and accompanying severe pain exacerbated by a 2001 motor vehicle accident. (Appellant’s Opening Br. at 4.) Plan participants are told they are considered disabled under the plan when: [Y]ou’re unable to do any job, for any employer, for which: You’re qualified, or You may become reasonably qualified by training, education or experience, other than a job that pays less than 50% of your Eligible Base Pay prior to the commencement of LTD Plan benefits. (Appellant’s App.

Vol. 4 at 1263.) Met-Life, then the claims administrator of the plan, approved her claim for benefits in December 2001. A.Termination of Benefits MetLife later asked independent physician Kathleen Kelley to review Williams’ claim file to determine whether she could work. On April 1, 2005, Kelley concluded Williams could return to work “in a sedentary type capacity.” (Appellant’s App. Vol. 3 at 606.) MetLife informed Williams of its review and sent Kelley’s report to Williams’ treating physicians. It asked these physicians to “submit objective medical evidence to support a continuing disability.” (Appellant’s App.

Vol. 3 at 636.) None of the treating physicians responded to this request. Then, MetLife asked vocational rehabilitation consultant David Rauch to conduct an employability assessment and labor market analysis based on Kelley’s report. Rauch concluded there were several jobs Williams could do. Based on that information, MetLife concluded Williams was no longer disabled and terminated her benefits on June 1, 2005. Williams filed three successive appeals. MetLife affirmed its decision each time. B. First Internal Appeal In June 2005, Williams tendered her first appeal of MetLife’s decision.

She included a letter from Julie Colliton, one of her treating physicians. According to Col-liton, Williams had chronic pain preventing her from sitting for more than seven minutes in one position or lifting, twisting, or carrying anything weighing more than two pounds. Colliton also noted Williams would need “frequent breaks to lie down and rest her back” and employment would exacerbate her condition. (Appellant’s App.

Vol. 3 at 625.) MetLife affirmed its cancellation of Williams’ LTD benefits after its reviewing physicians, Tanya Lump-kins and Philip Marion, concluded Williams was capable of sedentary employment in spite of Colliton’s observations. C. Second Internal Appeal While MetLife was considering Williams’ first appeal, Bruce Lippman, a physician who examined Williams to determine her eligibility for Social Security disability benefits, concluded she “probably is completely disabled.” (Appellant’s App.

Vol. 2 at 562.) Lippman opined that Williams’ condition “probably” precludes her from working, but did not specifically address why Williams could not perform sedentary work or indicate the effectiveness of her pain medicines. His assessment was brief and somewhat conclusory: Due to the patient’s multiple orthopedic problems she is able to sit or stand for only brief periods of time which would preclude her from almost all types of work. She cannot lift, twist or carry anything over a couple of pounds. She needs to take frequent breaks to lie down and rest her back.

She hears fine, speaks articulately, able to handle objects but with all her severe orthopedic disabilities the patient probably is completely disabled. (Appellant’s App. Vol. 2 at 562.) In September 2005, Williams submitted a second appeal with new supporting documents, including Lippman’s report. MetLife submitted the new information to reviewing physicians Lumpkins and Marion. Relying on their unchanged conclusions, it affirmed the revocation of Williams’ LTD benefits. D. Third Internal Appeal In May 2006, Williams appealed a third time. She again provided new documents to support her claim.

In particular, her appeal included an April 18, 2006 letter from Colliton. According to the letter, during the period Williams was under her care (June 1, 2005 through October 25, 2005), Colliton believed Williams was “disabled from any gainful employment.” (Appellant’s App. Vol. 2 at 450.) Colliton did not explain her conclusion or attempt to reconcile her view with the opinions of MetLife’s reviewing physicians. The appeal also included an earlier letter from Colliton exhorting MetLife to reinstate Williams’ benefits.

There, Colliton stated the denial of benefits was adversely affecting Williams’ condition, Williams was compliant with the prescribed treatments, and Williams had “never demonstrated exaggerated illness behavior.” (Appellant’s App. Vol. 2 at 501.) Williams also submitted detailed treatment notes of several other physicians helping her to manage her pain. The reviewing physicians again considered her case in light of the new information and again affirmed their earlier opinions. MetLife, in turn, affirmed its revocation of benefits for a third time.

In June 2007, the Social Security Administration (SSA) approved Williams for Social Security disability benefits. By that time, AT & T Integrated Disability Service Center (AT &

T) had taken over claims administration for the plan. Williams requested it reconsider the cancellation of her benefits — a fourth review — in light of the SSA’s decision. It refused. E. The District Court Proceedings Exercising her rights under the Employee Retirement Income Security Act of 1974 (ERISA), Williams filed a complaint in district court claiming she was entitled to LTD plan benefits. The parties moved for

summary judgment. The district court expressed concern that MetLife placed undue emphasis on the reports of its reviewing physicians, but nonetheless concluded it had adequately considered Williams’ reports of pain and the other relevant evidence in the record. It granted the appel-lees’ motion for

summary judgment and denied Williams’ motion for

summary judgment. II. DISCUSSION Williams contends MetLife abused its discretion in terminating her LTD benefits because (1) its decision was contrary to the evidence supporting her inability to work; (2) it ignored material evidence favorable to Williams’ claim; (3) it relied on a flawed employability assessment; (4) it revoked her benefits after acknowledging her disability by referring her to a Social Security disability advocate; and (5) it impermissi-bly revoked her benefits after initially approving them. A.

Standard of Review Where, as Williams concedes is the case here, an ERISA benefit plan gives the administrator discretionary authority to determine eligibility for benefits or to construe the terms of the plan, we review the administrator’s decision for abuse of discretion. Holcomb, 578 F.3d at 1192; see Metro. Life Ins. Co. v. Glenn, 554 U.S. 105, 111, 128 S.Ct. 2343, 171 L.Ed.2d 299 (2008). Under this standard, we uphold the administrator’s decision as long as it has a “reasoned basis.” Graham, 589 F.3d at 1357 (quotation marks omitted).

The administrator’s decision need not be the only logical decision or even the best decision; rather, “our review inquires whether the administrator’s decision resides somewhere on a continuum of reasonableness— even if on the low end.” Adamson v. Unum Life Ins. Co. of Am., 455 F.3d 1209, 1212 (10th Cir.2006) (quotation marks omitted). An administrator abuses its discretion when its decision is not supported by substantial evidence. Graham, 589 F.3d at 1357. Substantial evidence is evidence a reasonable mind would accept as adequate to support a conclusion. Id. at 1358.

It is “more than a scintilla [of evidence], but less than a preponderance.” Id. (quotation marks omitted). “Substantiality of the evidence is based upon the record as a whole.” Caldwell v. Life Ins. Co. of N. Am., 287 F.3d 1276, 1282 (10th Cir.2002). To determine whether the evidence in support of the administrator’s decision is substantial, we consider whether any information in the record undercuts the administrator’s conclusion. Id. “We give less deference if the administrator fails to gather or examine relevant evidence.” Id. B.

Substantial Credible Evidence Williams contends the record supports her inability to work in any occupation. She argues MetLife failed to give a full and fair review of the record. In particular, she argues MetLife improperly ignored her evidence of pain, her use of potent pain medication, the opinions of her treating physicians, the report of the independent physician who examined her for Social Security benefits, and her own description of her condition. We conclude the record supports MetLife’s decision and demonstrates that the reviewers considered those factors.

And, on this record, MetLife reasonably credited the views of its reviewing physicians over those of her treating physicians. The substantial supporting evidence in the administrative record includes the reports of Dr. Kelley and vocational consultant David Rauch written for the initial evaluation, and the reports of Dr. Lump-kins and Dr. Marion written in response to Williams’ three internal appeals. 1. The 2005 Kelley and Rauch Reports First, the record contains the April 1, 2005, report of Dr. Kelley.

After reviewing Williams’ file, she specifically noted Williams had “chronic pain symptoms involving abdominal discomfort and back pain,” but concluded Williams could: [R]eturn to a work position with restrictions and limitations in a sedentary type capacity. No bending or stooping is recommended with lifting/carrying of up to 10-lbs occasionally. There are no limitations on upper extremity use. Alternate sitting/standing. She may work an 8-hour day with no heavy pushing or pulling. (Appellant’s App. Vol. 3 at 606.) Kelley’s report contained an explicit analysis of the medical history in Williams’ file.

The record also contains the May 18, 2005, assessment of vocational rehabilitation consultant David Rauch. Using Kelley’s description of Williams’ physical limitations as well as Williams’ employment qualifications, Rauch concluded Williams could perform the duties of a “Customer Complaint Clerk,” “Civil-Service Clerk,” “Travel Clerk,” “Scheduler,” or “Telephone Solicitor.” (Appellant’s App. Vol. 3 at 632-33.) 2.

The July 2005 Lumpkins and Marion Reports Next, the record includes the July 2005 reports of two physicians, Tanya Lump-kins and Philip Jordan Marion, who reviewed Williams’ file after she appealed from the cancellation of her LTD benefits. Both reports describe the documentation they reviewed from Williams’ file and particularly discuss the letter in which Colli-ton opined that Williams was not able to work. Lumpkins’ seven-page report includes a detailed list of the records she reviewed, a thorough rheumatology assessment, and answers to MetLife’s specific questions.

The rheumatology assessment recounts Williams’ difficulties with back pain. It reviews the 2001 physical examinations of physician Rick Schwettmann, the 2001-02 medical progress notes of chiropractor Catherine Bohks, the 2002-04 records from spine surgeon Jeffery Kleiner, and the 2005 records from Julie Colliton. It then describes reviewing physician Mark Moyer’s initial 2002 analysis of Williams’ file and Kelley’s subsequent 2005 review. The assessment acknowledges both Williams’ reports of back pain and her treating physicians’ observations of back pain on physical examination.

It notes she had been diagnosed with “thoracic radiculopathy believed to be secondary to a protruded disc,” “abdominal pain ... secondary to intraabdominal visceral pain, generated by the radicular pain, or lower thoracic, and lumbar facet joint pain,” and “fibromyalgia [with] significant pathology from an orthopedic and neurologic standpoint involving her thoracic and lumbar spine.” (Id. at 571, 573.) It also noted she experienced “recalcitrant pain” and had achieved “no significant improvement” from “numerous blocks and injections.” (Id.) It further acknowledged that Williams underwent “thoracic spinal fusion” surgery and “left-sided endoscopic T9-T10 foraminal mierodiscectomy” (Id. at 572.) and had limited range of movement.

The assessment notes Williams’ medical file includes MRI and CT scans of her spine performed at various times. Lumpkins, who is board certified in rheumatology and internal medicine, concluded “[t]he record supports the complaint of chronic pain, and there is objective documentation of arthritic spine problem.” (Id. at 574.) In her view, the potent medicines Williams used to manage her pain would cause “the potential for drowsiness and decreased cognitive function, and as such, with chronic use would limit Ms.

Williams from working at unprotected height, driving a company vehicle, working with heavy machinery or with safety sensitive material.” (Id. at 574.) Lumpkins also noted Williams had “restrictions in range of motion of the lumbar spine” and would “have difficulty with prolonged static position,” but concluded these restrictions were not sufficient “to preclude her from performing a sedentary position.” (Id. at 575.) Marion, who is board certified in physical medicine and rehabilitation as well as pain management, also wrote a detailed report.

Like Lumpkins’ report, Marion’s report reviews the medical file information from Williams’ treating physicians, Schwettmann, Kleiner, and Colliton, as well as her chiropractor, Bohks.

Marion concluded, “[t]he history, physical examination and testing supports the diagnosis of her treating physician.” (Id. at 589.) He also concluded Williams’ “objective spinal impairment is significant and would result in restricting her to a sedentary duty occupation with the ability to stand and stretch for one to two minutes after sitting continuously for more than one hour.” (Id.) In his opinion, Williams was not capable of “significant activities such as climbing, stooping, kneeling, bending, crouching, or crawling.” (Id.) He also noted her “[m]ax-imal lifting should be no more than 10 pounds” and her pain medicines would prevent her from “working at unprotected heights, working with safety sensitive materials/machinery, or driving a company motor vehicle.” (Id. at 589-90.) 3.

The December 2005 Lumpkins and Marion Reports In her second appeal, Williams provided MetLife with the report of the physician, Bruce Lippman, who examined her for Social Security disability benefits. Lipp-man’s report concluded

(1) Williams’ condition would “preclude her from almost all types of work” and

(2) Williams “probably is completely disabled.” (Appellant’s App. Vol. 2 at 562 (emphasis added).) Lump-kins and Marion again reviewed Williams’ file and considered the new information. Lumpkins’ report explicitly acknowledged and discussed Lippman’s evaluation, which did not change her opinion.

She agreed “[t]he issue is [Williams’] pain,” but concluded appropriate treatment would still allow Williams to perform the duties of a sedentary job. (Id. at 555.) Likewise, Marion’s report concluded Lippman’s report contained no new information regarding Williams’ “functional status,” and, therefore, he maintained Williams could perform the duties of a light or sedentary job. (Id. at 552.) 4. The June 2006 Lumpkins and Marion Reports Finally, Lumpkins and Marion reviewed Williams’ file a third time after Williams provided MetLife with additional medical documentation.

Williams’ third appeal included an April 18, 2006 letter from treating physician Colliton. According to the letter, while Williams was under Colliton’s care (June 1, 2005 to October 25, 2005), she was “disabled from any gainful employment” because of her pain. (Id. at 450.) Colliton’s letter did not explain why Williams would be unable to perform the duties of a sedentary job. MetLife’s inquiry had specifically asked both reviewers if the new information changed their opinions, and if so, to explain why. In separate reports dated June 29, 2006, Lumpkins and Marion each reported no change.

Lumpkins again specifically acknowledged Williams had “chronic pain syndrome” and was being “treated aggressively with multiple narcotic pain medications, mood stabilizers, and acupuncture physical therapy and most recently an attempt for a spinal cord stimulator.” (Id. at 448.) She also noted Williams’ medical file did not indicate any problems with cognitive function due to the pain medicines.

Lumpkins therefore maintained there was no information about “any deficit documented in terms of the musculoskeletal functioning that would ... impair her from performing the routine physical requirements of a sedentary position.” (Id,) Marion’s brief report reiterated ‘Williams has well-established and recognized significant objective thoracic spine impairment.

However this impairment would not prevent her from performing the routine duties of a modified light duty or sedentary duty job as recommended in the ... employability assessment.” (Id. at 444.) In explanation, he noted “[t]here remains no objective impairment via neu-rologic examination, radiological assessment, or noted in her description of her functional ability that would preclude her from performing the routine duties of the modified light duty or sedentary duty occupation.” (Id. at 444.) Thus, MetLife’s decision was supported by substantial evidence in the record, and this substantial evidence shows MetLife and its reviewers considered Williams’ pain, use of pain medicine, and the report of the independent physician.

Despite that Williams argues the reviewing physicians’ opinions cannot be reasonably credited because they are biased in favor of denying benefits. She argues this bias is particularly acute here because the physicians who actually examined her concluded she could not work. Courts have acknowledged the bias of reviewing physicians. See Black & Decker Disability Plan v. Nord, 538 U.S. 822, 832, 123 S.Ct. 1965, 155 L.Ed.2d 1034 (2003).

The Supreme Court has credited the “concern that physicians repeatedly retained by benefits plans may have an incentive to make a finding of ‘not disabled’ in order to save their employers money and to preserve their own consulting arrangements.” Id. (quotation marks omitted). It has also recognized examining physicians may be in a better position than reviewing physicians to evaluate a patient’s condition because an examining physician has “a greater opportunity to know and observe the patient....” Id. (quotation marks omitted).

Yet when an examining physician is also a treating physician, the physician may feel sympathy for her patient and thus favor a finding of disability. See id.; see also Stephens v. Heckler, 766 F.2d 284, 289 (7th Cir.1985) (“The patient’s regular physician may want to do a favor for a friend and client, and so the treating physician may too quickly find disability.”). And, in some cases, reviewing physicians “might have the advantages of both impartiality and expertise.” Dixon v. Massunari, 270 F.3d 1171, 1177 (7th Cir.2001).

Here, two physicians who examined Williams, Colliton and Lippman, concluded Williams could not work. Conversely, MetLife’s reviewing physicians, Kelley, Lumpkins, and Marion, concluded Williams could perform the duties of a sedentary job. The reviewing physicians generally agreed with the examining physicians’ diagnoses but disagreed with their conclusions about Williams’ ability to work. Although the record reveals a genuine difference of medical opinion between the reviewing physicians and the examining physicians, the circumstances make Met-Life’s reliance on the reviewing physicians’ reports reasonable.

The strongest evidence favoring Williams’ disability is the opinion of the physician, Lippman, who examined Williams for Social Security disability benefits. Unlike the reviewing physicians, Lippman physically examined Williams. And, unlike Colliton, Lippman was not a treating physician who had a reason to favor Williams. Nevertheless, Lippman’s opinion did not indicate Williams was disabled from any job.

On the contrary, Lippman’s report was tentative on this point, only saying Williams’ condition would “preclude her from almost all types of work” and Williams “probably is completely disabled.” (Id. at 562 (emphasis added).) Nor does Lippmaris brief opinion give any reason why, with the proper pain medicines, Williams could not perform the duties of a sedentary job. The persuasiveness of the other strong evidence favoring Williams’ disability— Colliton’s opinion — is likewise undermined by both Colliton’s failure to explain why

(1) Williams, if properly medicated, could not perform the duties of a sedentary job and

(2) Colliton disagreed with the reviewing physicians’ assessments. Moreover, because she actually treated Williams, Colli-ton may have been biased in favor of finding Williams disabled. See Nord, 538 U.S. at 832, 123 S.Ct. 1965. By contrast, Lumpkins’ and Marion’s reports are detailed, unequivocal, and responsive to the medical opinions Williams provided. During each of Williams’ internal appeals, MetLife forwarded the medical file information to these physicians, asked if the information changed their opinions, and asked them to explain why.

Each time these physicians affirmed their awareness of Williams’ pain, but concluded with the proper pain treatment and restrictions, Williams could perform the duties of a sedentary job. In the end, MetLife had to judge the reviewing physicians’ opinions against the opinions of Colliton and Lippman. In making its decision, it unquestionably had to compare the possible biases of the reviewing physicians against Colliton’s possible bias.

Because the reviewing physicians made their findings with the full awareness of the evidence Williams had presented, were unequivocal in their conclusions, and lacked the possible biases of the treating physicians, MetLife’s decision to credit the reviewing physicians’ conclusions was reasonable. See Corry, 499 F.3d at 401 (“[T]he job of weighing valid, conflicting professional medical opinions is not the job of the courts; that job has been given to the administrators of ERISA plans.”).

Although Williams cites a variety of bias-related information from depositions, testimony, and judicial opinions in other cases, we cannot consider this extra-record information. Our duty is to determine whether MetLife abused its discretion based on the evidence before it when it made its decision. See Sandoval v. Aetna Life & Cas. Ins. Co., 967 F.2d 377, 381 (10th Cir.1992); see also Holcomb, 578 F.3d at 1192 (“Our review is limited to the administrative record — the materials compiled by the administrator in the course of making his decision.” (quotation marks omitted)).

For this reason, we do not ordinarily consider information outside the administrative record. See Nance v. Sun Life Assurance Co. of Canada, 294 F.3d 1263, 1269 (10th Cir.2002). As we have explained: A primary goal of ERISA was to provide a method for workers and beneficiaries to resolve disputes over benefits inexpensively and expeditiously. Permitting or requiring district courts to consider evidence .... that was not presented to the plan administrator would seriously impair the achievement of that goal.

If district courts heard evidence not presented to plan administrators, employees and their beneficiaries would receive less protection than Congress intended. Sandoval, 967 F.2d at 380 (quotation marks omitted). In some cases, a district court may allow extra-record evidence to flesh out claims of bias when the employer funding the plan also makes claims decisions. Murphy v. Deloitte & Touche Grp. Ins. Plan, 619 F.3d 1151, 1162 (10th Cir.2010); Jewell v. Life Ins. Co. of N. Am., 508 F.3d 1303, 1309 (10th Cir.2007). Here, however, while AT & T funded the plan, it employed MetLife to evaluate claims.

Thus, any conflict of interest is more attenuated than in other ERISA cases, and the rationale for considering extra-record evidence of bias here is diminished. But see Glenn, 554 U.S. at 112, 114, 128 S.Ct. 2343 (noting the conflict may persist even when the employer delegates duties for claims administration). In addition, both Murphy and Jewell acknowledge extra-record evidence is not appropriate in every case involving bias or a conflict of interest.

In Murphy, we reasoned one purpose of allowing extra-record evidence was to give the claimant “access to the information necessary to establish the seriousness of the conflict.” 619 F.3d at 1157-58. Likewise, in Jewell, we acknowledged certain exceptional circumstances warrant the admission of extra-record evidence, including, as pertinent here, situations “where the payor and the administrator are the same entity and the court is concerned about impartiality.” 508 F.3d at 1309.

Even then, however, we concluded “the party offering the extra-record evidence must demonstrate that it could not have been submitted to the plan administrator at the time the challenged decision was made.” Id. (quotation marks omitted). Here, despite the apparent availability of documentary evidence illustrating the bias of the reviewing physicians against claimants, Williams expressed no concerns about this bias to MetLife in any of her appeals.

Had Williams done so, MetLife could have asked another reviewer to examine Williams’ file, engaged an independent consultant to examine Williams, or explained its procedures for guarding against bias in its medical reviews.

See Glenn, 554 U.S. at 117, 128 S.Ct. 2343 (noting conflicts of interests are less significant when “the administrator has taken active steps to reduce potential bias and to promote accuracy, for example, by walling off claims administrators from those interested in firm finances, or by imposing management checks that penalize inaccurate decisionmaking irrespective of whom the inaccuracy benefits”). Thus, we cannot consider Williams’ extra-record evidence. C.

Ignoring Favorable Material Evidence Williams next argues MetLife abused its discretion because it failed to examine material portions of the relevant evidence. In particular, she argues

(1) MetLife and the reviewing physicians improperly ignored letters she and her fiancé wrote describing her condition;

(2) Met-Life improperly ignored Williams’ favorable Social Security disability benefits determination. We are not persuaded. A plan administrator abuses its discretion when it “fail[s] to examine a material portion of the relevant evidence.” See Rekstad v. U.S. Bancorp, 451 F.3d 1114, 1120-21 (10th Cir.2006); see also Rasenack ex rel. Tribolet v. AIG Life Ins. Co., 585 F.3d 1311, 1326 (10th Cir.2009) (concluding the administrator abused its discretion when it inexplicably disregarded highly probative favorable medical evidence and affidavits favorable to claimant).

Here, the letters and the Social Security disability determination were not a material portion of the relevant evidence because they contained either outdated or cumulative information. In her June 17, 2005 letter to MetLife, Williams stated her pain was preventing her from driving or bathing on her own. However, Williams’ file contained extensive later notes from her treating physicians. These notes, as discussed previously, provided ample probative evidence about Williams’ pain and its consequences on her ability to perform the duties of a sedentary job.

In addition, as Williams used a variety of drug combinations to treat her pain, her letter was only a snapshot into the effects of her pain. It had limited probative information about her disability under the later pain management approaches documented in her file. For instance, Colliton’s records showed, as of August 2, 2005 — shortly after Williams wrote her letter — Williams was in “excruciating pain” because she could no longer afford all of her pain medicines. (Appellant’s App.

Vol. 2 at 501.) Yet Colliton’s October 25, 2005, notes showed Williams had obtained a reduced price service allowing her to again purchase all of her pain medicines. MetLife could reasonably assume Williams regained her ability to bathe and drive on her own after she was able to fully manage her pain. And the reviewing physicians explicitly acknowledged Williams would require pain treatment and this would affect her ability to work. Compare Rasenack, 585 F.3d at 1326 (noting the administrator “fail[ed] to even acknowledge Dr. Weintraub’s conclusion that Mr.

Rasenack suffered complete and irreversible hemiplegia”). Likewise, the letter Williams’ fiancé John Carr wrote in support of Williams’ claim was not material evidence favorable to Williams’ claim because it too was devoted to a discussion of Williams’ condition under earlier pain management regimes. Similarly, the Social Security disability determination contained only cumulative information. Because MetLife considered Lippman’s report when it decided Williams’ second appeal, the final Social Security disability determination would not have added any new material medical information to her file.

Nor did MetLife or AT & T have a continuing legal duty, emanating from its fiduciary obligations to plan participants, to review this new evidence. See Nance v. Sun Life Assurance Co. of Canada, 294 F.3d 1263, 1269-70 (10th Cir.2002) (noting we look exclusively to the terms of the plan to determine the extent of the administrator’s duty to review new evidence, as long as the plan’s rules provide adequate “opportunity for an appropriate review of the denial” of the claim). D. Employability Assessment Relying on Spangler v.

Lockheed Martin Energy Systems, Inc., Williams also argues MetLife improperly withheld favorable evidence from its vocational consultant. 313 F.3d 356 (6th Cir.2002). The Spangler court concluded an administrator’s decision to limit a vocational consultant to the sole piece of evidence unfavorable to the claimant’s case was an abuse of discretion. Id. at 362.

Even so, the Span-gler court also noted “the ultimate issue in an ERISA denial of benefits case is not whether discrete acts by the plan administrator are arbitrary and capricious but whether its ultimate decision denying benefits was arbitrary and capricious.” Id. The Spangler court thus turned to the medical evidence in the claimant’s file and concluded the record overwhelmingly demonstrated the claimant could not work. Id. Unlike the evidence in the Spangler claimant’s file, the evidence in Williams’ file demonstrated a genuine difference of medical opinion on her ability to work.

Met-Life reasonably concluded Williams could work; thus, Spangler is not applicable here. E. Referral to Social Security Disability Advocate Williams also argues MetLife abused its discretion by cancelling her benefits after referring her to a law firm that could help her pursue Social Security disability benefits. She asserts the “only logical reason MetLife would refer Williams to [the law firm] to pursue her benefit claim is if it agreed she was disabled from any occupation....” (Appellant’s Opening Br. at 49.) We disagree.

It is prudent for employers (or, as in this case, claims administrators who administer a plan on an employer’s behalf) who offer LTD plans to advise every employee who makes a disability claim to also pursue Social Security disability benefits. It may help the employee and it could help to defray the costs of providing disability benefits to employees. See Ladd v. ITT Corp., 148 F.3d 753, 755 (7th Cir.1998). Such prudence does not imply an opinion about the merits of any claimant’s application. Williams’ reliance on Ladd is misplaced.

In Ladd, MetLife “provided [the claimant] with legal representation to assist her with the application.” Id. We see no indication in the record that MetLife provided Williams with representation; it merely referred her to a law firm. Because a referral does not indicate the same level of endorsement of the claimant’s case as providing representation, Ladd is inapplicable here. Therefore, MetLife did not abuse its discretion in denying Williams’ claim after referring her to a Social Security disability advocate. F.

Cessation of Benefits Finally, contrary to Williams’ argument, MetLife did not abuse its discretion when it reversed its prior decision finding Williams disabled. The new information MetLife obtained was sufficient, under the terms of the plan and as a matter of law, for it to reverse its previous decision. The Eighth Circuit’s reasoning in Kecso v. Meredith Corp. is persuasive. 480 F.3d 849 (8th Cir.2007).

Under Kecso, “unless information available to an insurer alters in some significant way, the previous payment of benefits is a circumstance that must weigh against the propriety of an insurer’s decision to discontinue those payments.” Id. at 854 (quotations omitted). Even so, paying benefits does not prevent an administrator from denying benefits when the administrator becomes aware of new information about the claimant’s eligibility for benefits. Id. On the contrary, an administrator breaches its fiduciary duty to plan participants if it pays benefits to unqualified claimants. Id. (citing Tillery v.

Hoffman Enclosures, Inc., 280 F.3d 1192, 1197 (8th Cir.2002)). Here, as in Kecso, the plain language of Williams’ LTD plan vests the claims administrator with the responsibility for determining whether the claimant continues to qualify for LTD benefits. The plan contains no express language indicating a claims administrator’s responsibility for determining whether claimants continue to qualify for benefits can be exercised only when a claimant experiences a change in condition.

Moreover, the information available to MetLife when it cancelled Williams’ benefits changed significantly after the time it initially approved her claim. After paying Williams’ benefits for approximately three and a half years, MetLife obtained a new file review from its reviewing physician and used the review to determine there were sedentary jobs Williams could perform. Were we to adopt a rule preventing administrators from terminating previously granted benefits, the likely consequence of the rule would be to make administrators more wary of granting benefits in the first place.

Indeed, in Kecso, the administrator elected to pay benefits when the claimant’s entitlement to benefits was unclear, and reversed its decision only after it obtained new information demonstrating the claimant’s ability to work. Kecso, 480 F.3d at 854. Had the Kecso administrator known it would never be able to reverse its benefits determination based on new evidence, it would likely never have approved the claimant’s benefits while it investigated “apparent inconsistencies and ambiguities in the medical record.” Id. AFFIRMED.

Williams’ unopposed motion to seal volumes two through five of the appellant’s appendix is GRANTED. This order and judgment is an unpublished decision, not binding precedent. 10th Cir. R. 32.1(A). Citation to unpublished decisions is not prohibited. Fed. R.App. 32.1. It is appropriate as it relates to law of the case, issue preclusion, and claim preclusion. Unpublished decisions may also be cited for their persuasive value. 10th Cir. R. 32.1(A). Citation to an order and judgment must be accompanied by an appropriate parenthetical notation — (unpublished). Id. . See 29 U.S.C. § 1132(a)(1)(B). .

In this context, the abuse of discretion and the arbitrary and capricious standards of review are “interchangeable.” Weber v. GB Group Life Assurance Co., 541 F.3d 1002, 1010 n. 10 (10th Cir.2008). Our review is de novo, see Graham v. Hartford Life & Accident. Ins. Co., 589 F.3d 1345, 1357 (10th Cir.2009), and we critically review the plan administrator’s denial of benefits. Holcomb v. Unum Life Ins. Co. of Am., 578 F.3d 1187, 1192 (10th Cir.2009). Nevertheless, the district court’s analysis may well inform the debate. .

Colliton's conclusion is somewhat surprising after, on April 4, 2005, she noted Williams’ medicines "cover[ed] her pain really quite nicely.” (Appellant’s App.

Vol. 2 at 513.) On the other hand, Colliton’s August 2, 2005, letter indicated Williams was having difficulty affording her pain medicines, was in "excruciating pain,” and had to use crutches to move about. (Id. at 501.) At that time, Colliton noted Williams was using a "Dura-gesic patch” and "Avinza.” (Id.) In her October 25, 2005, note, Colliton wrote Williams was again able to access pain medicines at "decreased cost” through Bristlecone Hospice Services. (Id. at 498.) At that time, Colli-ton noted Williams was using the "Duragesic patch” and "MS Contin” as well as "Dilau-did,” "Oxycodone,” and "Cymbalta.” (Id. at 498.) .

In our review to determine whether the record contains substantial evidence to support the administrator’s decision, it is not our role to weigh or evaluate the medical evidence in the record. Cf. Gardner v. Bishop, 362 F.2d 917, 920 (10th Cir.1966) (reviewing for substantial evidence the denial of Social Security disability benefits); see also Corry v. Liberty Life Assurance Co. of Boston, 499 F.3d 389, 401 (5th Cir.2007) (noting it is the administrator's job to resolve conflicts among medical opinions). .

Another physician, Kleiner, who examined Williams on February 22, 2002, found her able to work four hours per day. But because Williams’ pain situation appears to have changed significantly as her pain management approaches changed, Kleiner’s opinion is not particularly helpful to our analysis. .

Williams briefly argues Lumpkins' opinions are "invalid” because Lumpkins was not provided with Colliton's July 6, 2005 opinion letter. (Appellant's Opening Br. at 38.) While the record is unclear on this point— Lumpkins’ report notes she had reviewed Col-liton's records from March 8, 2005 to April 18, 2006 — any omission was not material because MetLife explicitly considered the letter, and Colliton's position was clear from her other letters and notes. See Rekstad v. U.S.

Bancorp, 451 F.3d 1114, 1120-21 (10th Cir.2006) (concluding the administrator abuses its discretion when it does not "examine a material portion of the relevant evidence”) (emphasis added). . Although Williams criticizes the reviewing physicians for conflating the ability to perform “activities of daily life” and the ability to work, MetLife could reasonably infer the requirements of a sedentary job would overlap considerably with the requirements of “activities of daily life.” Social Security disability determinations often consider a claimant’s ability to perform such activities for this reason.

See, e.g., Calhoun v. Barnhart, 85 Fed. Appx. 678, 682, 684-85 (10th Cir.2003) (unpublished); Brock v. Astrue, 244 Fed.Appx. 175, 178-79 (10th Cir.2007) (unpublished). . Unlike the plan in this case, the plan in Jewell did not reserve discretion to the plan administrator. Jewell, 508 F.3d at 1308. Therefore, judicial review of the Jewell administrator's decision was under the de novo standard rather than the abuse of discretion standard as it is here. Id.

Nevertheless, the party seeking to introduce extra-record evidence must "demonstrate that it could not have been submitted to the plan administrator at the time the challenged decision was made” in a case involving de novo review. Id. at 1309 (quotation marks omitted). The party seeking to introduce extra-record evidence in an "abuse of discretion” case must meet at least the same burden. .

The primary rationale for the Ladd court’s reversal of the administrator’s denial was that ”[n]o one who examined Ladd, including the doctor ... selected by [the administrator] to examine her, believed that she was capable of working.” 148 F.3d at 755. The Ladd court acknowledged the administrator’s support for the claimant’s Social Security disability application and subsequent denial of benefits under the plan might not "provide an independent basis for rejecting [the administrator’s] evaluation.” Id. at 756.

Rather, the administrator's reversal of position merely ”cast[] additional doubt on the adequacy of [the administrator's] evaluation of Ladd’s claim.” Id.