PER CURIAM:
Darlene McDonald, an office manager, suffers from degenerative disc disease in her spine. Following a surgical procedure intended to alleviate her back pain, she ceased working and applied for long-term disability benefits under her employer’s ERISA plan with Hartford Life Group Insurance Company. After reviewing McDonald’s medical records and interviewing her treating physicians, Hartford denied benefits, finding that she was capable of performing sedentary work and therefore [601]*601did not meet the plan’s definition of “disabled.” McDonald brought two administrative appeals, both of which Hartford denied. McDonald brought suit, alleging Hartford abused its discretion by denying her claim. The district court granted summary judgment for Hartford, and McDonald appeals. We affirm.
1. BACKGROUND
A. McDonald’s Job Requirements & Long Term Disability Policy
Darlene McDonald (McDonald) worked as an office manager, business manager, and property manager at the law firm of Bruno & Bruno, LLC (B & B) starting in February 2002. Her job required frequent computer use, prolonged sitting, standing, and walking. She also had to frequently bend, stoop, and reach while filing employee information. Any lifting required was normally under ten pounds, but occasionally it might exceed this amount.
B & B offered long term disability insurance coverage for its employees through CNA Group Life Assurance Company (now known as Hartford Life Group Insurance Company, or Hartford). Under B & B’s policy, an individual qualifies for long term disability benefits if the individual is sick or injured during a 90-day elimination period (beginning on the date of the onset of disability) and for another 24 months following the end of the elimination period. The policy refers to this initial time period (the 90-day elimination period plus 24 months) as the “Occupation Qualifier” period, but it is also known in the insurance industry as the “Own Occupation” period. To receive benefits under the policy during the Own Occupation period, the individual must continuously meet the definition of “disabled” for the entire period. The policy defines “disabled” as “[ijnjury or [s]ickness caus[ing] physical or mental impairment to such a degree of severity that [the individual is] 1) continuously unable to perform the [m]aterial and [substantial [d]uties of [her] [r]egular [occupation; 2) and not [gainfully [e]mployed.”1 The policy is governed by the Employee Retirement Income Security Act of 1974 (ERISA), 29 U.S.C. § 1001, et seq., and gives Hartford discretionary authority over the interpretation of the policy and eligibility decisions.
B. McDonald’s Injury and Treatment
In July 2003, McDonald began experiencing headaches and back pain. She suffered a herniated disc in a car accident in June 2004, which exacerbated her pre-ex-isting symptoms. After the accident, she began seeing Dr. Evalina Burger, an orthopedic surgeon. In July 2004, Dr. Burger determined that McDonald suffered from degenerative disc disease2 in both [602]*602the lumbar and cervical regions of her spine, but she did not think that surgery would help at the time. After a November 2004 appointment, Dr. Burger recommended that McDonald avoid “prolonged sitting in front of [a] computer” and limit any lifting. Eventually, Dr. Burger recommended cervical fusion surgery between vertebrae 4-5, 5-6, and 6-7.
Dr. Burger referred McDonald to Dr. John Steck, a neurosurgeon, who concurred in recommending surgery after reviewing McDonald’s MRIs. Dr. Steck performed surgery on McDonald’s lumbar spine on December 28, 2004; McDonald continued working full-time at B & B until December 27, 2004. At the time of the surgery, it was unclear how successful the procedure would be, particularly because the degenerative disc disease affected more than just the lumbar spine. Following the surgery, McDonald reported improvement in her pain but MRIs continued to show evidence of degenerative disc disease. In February 2005, six weeks after the surgery, Dr. Steck noted that McDonald was doing very well and gave her permission to return to work on a part-time basis (four hours a day, three days a week). Dr. Steck’s final chart note for McDonald was dated July 6, 2005, and shows that her MRI reflects “cervical and lumbar spondylosis with arthritic and degenerative changes in the cervicothoracic spine.” McDonald attempted to return to work for a period of time, but soon found it too painful. McDonald applied for long-term disability benefits in June 2005 and her final day of work at B & B was October 12, 2005.
Beginning in October 2004, McDonald started treatment with Dr. Paul Hubbell, a pain management specialist, who determined in a February 13, 2006, letter to Hartford that McDonald “may be able to perform part-time work” but could not return to work on a full-time basis as a result of her “significant arthritic complaints in the cervical spine which cause reflex [sic] significant muscle spasms, headaches, and limitation of position.” Dr. Hubbell also noted in the February 2006 letter that McDonald’s subjective complaints of pain were supported by objective findings of facet pathology and disc pathology, but he also recommended that she receive additional pain therapy, which he predicted might “significantly improve her physical activity capabilities.”3 McDonald reported that her back pain improved some after her surgery, but it has since fluctuated; she still complains of headaches and muscle spasms. Because of her pain, she requires help with her personal grooming and with maintaining her household.
C. Administrative Claims Process
In June 2005, McDonald filed a claim for disability benefits with Hartford. According to Hartford, the 90-day elimination period began December 28, 2004 — the day McDonald underwent surgery — and ended March 28, 2005. The Own Occupation period lasted from December 28, 2004, to March 28, 2007. Therefore, to qualify for disability benefits, McDonald needed to show that she met the definition of “disabled” for the duration of the Own Occupation period. McDonald submitted documentation from Dr. Steck stating that she [603]*603should not bend, stoop, climb, or lift more than ten pounds. Hartford gathered other medical records, including the claim notes, Dr. Steck’s L[ong] T[erm] Disability] Physician’s Statement and Functional Assessment Tool, clinical notes from Drs. Steck and Hubbell, the discharge summary after the surgery, the operative report, and four MRIs of her lumbar and cervical spine. The Functional Assessment Tool indicates that Dr. Steck did not think that McDonald was capable of performing full time work involving “sitting, standing, and walking for varying periods of time, typing on a computer, some bending, stooping, and reaching, regularly lifting items under [ten pounds] and occasionally lifting items over [ten pounds]” as of July 14, 2005. Hartford interviewed McDonald and determined that she was able to get help at work to avoid bending, stooping, and lifting and that B & B was cooperative, allowing her to get help with tasks.
After compiling McDonald’s medical records, Hartford hired Dr. Bruce LeForce, a physician with Reed Review Services, to review McDonald’s file. Dr. LeForce determined that McDonald was:
capable of sitting up to eight hours per day given an opportunity for frequent breaks and changes in position. She can stand and walk occasionally. She cannot lift or carry more than ten pounds and this can be done only occasionally. She can exert a negligible amount of force continuously. She cannot climb, bend, or stoop. These limitations should be considered permanent.... She can work full-time given the other restrictions and
Letter to Attorneys from Dr. Evalina Burger dated July 12, 2004; A physical examination form dated 7/12/04 which was not complete; A Physical Therapy referral dated 7/12/04; A Spine Assessment Form undated; MRI of the Cervical Spine dated 9/10/04; Clinic note[s] dated 10/4/04 through 11/1/04 from Dr. Evalina Burger; Prescription for medication [including Vi-codin] dates 11/1/04; LTD Physician's Statement; MRI of the Cervical Spine dated 10/19/05; Evaluation by Dr. Paul J. Hub-bell dated 10/26/04; Office notes from Dr. Paul J. Hubbell dated 10/26/04 through 8/16/05; Consent forms and clinic record from [surgery] dated 12/13/04.
limitations.... [T]he objective findings indicate that she is capable of full-time work provided that she is limited to a sedentary type job with no climbing, bending, or stooping.
Dr. LeForce concluded that McDonald’s MRI showed “only some degenerative changes without spinal stenosis or evidence of nerve root impingement.” In a letter dated September 30, 2005, Hartford determined that the information contained in McDonald’s file “fail[ed] to support an inability to perform the material and substantial duties of [her] regular occupation” and therefore she was ineligible for benefits.
McDonald appealed the denial and presented additional information for Hartford’s review.4 Hartford hired Dr. Barry Turner, an orthopedic surgeon employed by University Disability Consortium (UDC), to review the original file and the new documents submitted by McDonald. According to Hartford’s January 26, 2006, letter, Dr. Turner did not review Dr. Le-Force’s conclusions, but independently reviewed the file and came to his own conclusions. Dr. Turner attempted to contact McDonald’s treating physicians. He was unable to speak directly with Dr. Steck, but he spoke with a person in Dr. Steck’s office who indicated that McDonald’s last visit was in October 2005, at which time Dr. Steck’s notes reflected that sedentary-level work would be acceptable, without any limitations as to time.
Dr. Turner reported that the records he reviewed supported the conclusion that [604]*604McDonald suffered “generalized cervical and lumbar spondylosis with arthritic changes consistent with her age,” and had undergone a “satisfactory cervical and lumbar diskectomy [the surgical procedure] with fusions and no complications or resultant radiculopathy, myelopathy, or nerve root compression noted.” Dr. Turner concluded that McDonald’s condition was “secondary to the normal aging process” and found “no evidence of any significant impairment.” He opined that “there is no reason that oral analgesics and anti-inflammatory therapy would not be effective” and “provide[d] no restrictions or limitations that would preclude full-time work activity.” In a letter dated January 26, 2006, Hartford informed McDonald that, following the appeal, it appeared that she “at a minimum retained] the functional capacity for sedentary-type work activity” and therefore did not meet the policy’s definition of “disabled.”
McDonald again requested that Hartford reconsider its decision, and she submitted additional documentation, including letters from McDonald regarding her pain; a job description for her position at B & B; Dr. Hubbell’s February 2006 letter regarding her subjective pain; and a MRI of her cervical spine from October 2005, which she had previously submitted. Hartford hired another UDC orthopedic surgeon, Dr. Robert Pick, to conduct the second review. Dr. Pick reviewed all of the previous evidence submitted by McDonald.5 Dr. Pick spoke with Dr. Burger, who reported only that her last visit with McDonald was in November 2004 and she did not have access to McDonald’s records, as they had been affected by Hurricane Katrina. Dr. Pick also spoke with Dr. Hub-bell, who stated that “the only objective finding is the MRI,” and when asked if McDonald could engage in sedentary work, replied: “Instead of guessing, get a Functional Capacity Evaluation.” Dr. Pick called Dr. Steck and spoke with him after hours when he did not have access to McDonald’s records; Dr. Steck reported from general memory that McDonald could engage in sedentary work for eight hours a day.
From his review of the documents and his conversations with the treating physicians, Dr. Pick concluded that “there is no documentation or substantation [sic] that Ms. McDonald cannot engage in at least the sedentary — light work category on full-time basis.” He observed that the file indicated McDonald had “degenerative arthritic changes of the cervical and lumbar spine”; her surgery had been successful; and she had “satisfactory postoperative progress and recovery with improvement in [her] preoperative symptoms.” He noted that “no specific intervention has been recommended other than conservative care” and “the case file does not document any substantive objective orthopedie/mus-culoskeletal findings that would prevent Ms. McDonald from engaging in full-time work activities in at least the sedentary— light work category....” He concluded that McDonald’s file did not “establish a complete impairment from gainful employment.” Based on Dr. Pick’s review and conclusions, Hartford sent McDonald a letter on April 26, 2006, declining to change the prior decision to deny benefits. The letter stated that it was “final and binding” and that McDonald had “exhausted all [administrative remedies.”
[605]*605
D. Litigation
McDonald filed suit against B & B 6 and Hartford on June 12, 2006, under ERISA. McDonald alleged that Hartford failed to pay benefits under the insurance policy as required by 29 U.S.C. § 1132(a)(1)(B) and that Hartford breached its fiduciary duties under 29 U.S.C. § 1104(a). McDonald and Hartford both filed motions for summary judgment. The district court determined that the administrative record was incomplete because Hartford had relied upon Dr. Pick’s discussion with Dr. Steck after hours, when Dr. Steck was at the health club and did not have access to McDonald’s records. On March 28, 2008, the district court remanded to Hartford for further clarification of Dr. Steek’s position, with instructions that the remand was “for the limited purpose of interviewing and/or deposing Dr. Steck to clarify his position as regards [McDonald’s] limitations.”
Following another interview with Dr. Steck, Hartford noted that his opinion was that “McDonald’s ongoing complaints of pain are subjective in nature” and he could not determine at this stage whether she could work. In a November 21, 2008, letter to McDonald’s attorney, Hartford determined that any new information from the interview of Dr. Steck was “not compelling based on the totality of the facts presented” and declined to change its prior decision. McDonald re-urged her motion for summary judgment, as did Hartford.
On October 27, 2007, before the district court issued its remand order, McDonald received a Declaration of Disability from the Social Security Administration (SSA award). McDonald did not inform the district court of the award at the time, but she did forward the award, with the accompanying letter of reasons, to Hartford shortly after the case was remanded and requested that Hartford consider the award. In its November 21, 2008, letter, Hartford did not mention the award and the record does not show if Hartford included the award in its investigation of McDonald’s claim. The first time that McDonald informed the district court of the SSA award was in her re-urged motion for summary judgment.
The district court granted summary judgment for Hartford. In its final order on April 29, 2009, the district court declined to give controlling weight to the opinions of McDonald’s treating physicians and found that “Hartford conducted a thorough and reasonable analysis of McDonald’s disability, basing its initial decision and two appeal decisions on the opinions of three separate, independent physicians.” The district court noted that McDonald admits her job is “primarily sedentary” and that she “gets help filing to avoid bending and stooping.” Furthermore, the district court pointed out that “both her treating physicians and Hartford’s physicians have agreed she should be able to perform [sedentary work].” The district court rejected McDonald’s argument that Hartford was bound to consider her SSA award — as Hartford had already reached its final decision more than one year prior to the award, it could not have considered the additional information, particularly in light of the district court’s specific instructions limiting the scope of the inquiry on remand. McDonald timely appealed.
II. DISCUSSION
McDonald raises two main points on appeal. She argues that the district court improperly refused to allow her to supplement the administrative record with her SSA award, and she contends that the district court erred in finding that Hart[606]*606ford’s decision to deny benefits was supported by substantial evidence.
A. Supplementation of Administrative Record
McDonald argues that the district court abused its discretion by declining to permit McDonald to enter her SSA award into the administrative record. The SSA award was issued October 23, 2007, more than one year after Hartford issued its final decision in April 2006. McDonald argues that the district court’s remand for further investigation re-opened the administrative process, making it proper to supplement the administrative record with the SSA award. Hartford argues that, given the limited purpose of the remand, there was no obligation for Hartford to consider new evidence outside of the district court’s instructions and the district court acted within its discretion in declining to consider the SSA award. The determination of whether evidence should be included in the administrative record is an evidentiary decision, and we review for abuse of discretion. Corry v. Liberty Life Assurance Co. of Boston, 499 F.3d 389, 398 n. 12 (5th Cir.2007).
When conducting abuse of discretion review of a denial of benefits based on an administrative record, we have generally required that the scope of review be limited to facts known to the plan administrator at the time of the benefits decision. S. Farm Bureau Life Ins. Co. v. Moore, 993 F.2d 98, 102 (5th Cir.1993). However, we have recognized certain limited exceptions to this rule. See Vega v. Nat’l Life Ins. Servs., Inc., 188 F.3d 287, 299 (5th Cir.1999) (en banc) (“To date, th[e] exceptions have been related to either interpreting the plan or explaining medical terms and procedures relating to the claim.”).7 These exceptions have been judged on a case-by-case basis, and we have declined to adopt any per se rules in this area. Cf. Duhon v. Texaco, Inc., 15 F.3d 1302, 1309 (5th Cir.1994).
When compiling the administrative record, the plan administrator must identify what evidence constitutes the administrative record, and the claimant must have “a reasonable opportunity to contest whether that record is complete.” Estate of Bratton v. Nat’l Union Fire Ins. Co. of Pittsburgh, 215 F.3d 516, 521 (5th Cir.2000). While the administrative record is generally limited to “relevant information made available to the administrator prior to the complainant’s filing of a lawsuit and in a manner that gives the administrator a fair opportunity to consider it,” we have attempted to avoid abuse or mistake by allowing “the claimant’s lawyer [to] add additional evidence to the administrative record simply by submitting it to the administrator in a manner that gives the administrator a fair opportunity to consider it.” Id. at 521 & n. 5. We have been clear, however, that “the district court is precluded from receiving evidence to resolve disputed material facts — i.e., a fact the administrator relied on to resolve the merits of the claim itself.” Vega, 188 F.3d at 299. Had the district court not remanded to Hartford for further investi[607]*607gation of Dr. Steck’s opinion, the question of whether the SSA award should be included in the administrative record would be clear: the administrative record closed when McDonald filed suit in June 2006. Cf. Moore, 993 F.2d at 102. The SSA award does not fall into the two acknowledged exceptions: evidence interpreting the plan or explaining medical terms and procedures.
The timing of the remand order complicates the analysis somewhat, for Estate of Bratton suggests that McDonald had the opportunity to offer additional information to Hartford, so long as the submission of new information was conducted “in a manner that gives the administrator a fair opportunity to consider it.” Estate of Bratton, 215 F.3d at 521 n. 5. Assuming without deciding that McDonald could have supplemented the record on remand, she missed her opportunity. The SSA award was issued on October 23, 2007, five months before the district court remanded the case to Hartford on March 28, 2008. Yet McDonald did not bring the SSA award to the district court’s attention until after the conclusion of the remand period, despite at least one opportunity to do so. For example, Hartford moved for reconsideration of the remand order, which McDonald opposed in writing — without mention of the SSA award. McDonald was aware of the district court’s specific instructions on remand but did not ask the district court to expand the scope of the remand to include consideration of the award. The first mention of the SSA award in McDonald’s filings is in her January 23, 2009, reurged motion for summary judgment. In light of the district court’s very specific instructions limiting the scope of the remand and McDonald’s failure to submit the award for consideration at an appropriate time, the district court did not abuse its discretion in refusing to consider the SSA award and in approving Hartford’s decision not to consider the award.
B. Substantial Evidence
We typically follow a two-step process to determine if an ERISA plan administrator has abused its discretion, asking first if the plan administrator’s determination was legally correct; if it is not, we proceed to the second question of whether the decision was an abuse of discretion. Holland v. Int’l Paper Co. Ret. Plan, 576 F.3d 240, 246 n. 2 (5th Cir.2009). However, this process is not rigid; “we may skip the first step if we can more readily determine that the decision was not an abuse of discretion.” Id. Here, we may proceed directly to the second step. McDonald raises several points that she contends either justify a less deferential standard of review or show abuse of discretion by Hartford.
We review a grant of summary judgment in an ERISA case de novo, applying the same standard as the district court. Wade v. Hewlett—Packard Dev. Co. LP Short Term Disability Plan, 493 F.3d 533, 537 (5th Cir.2007). The Supreme Court requires that a denial of benefits be reviewed under a de novo standard “unless the benefit plan gives the administrator or fiduciary discretionary authority to determine eligibility for benefits.” Firestone Tire & Rubber Co. v. Bruch, 489 U.S. 101, 115, 109 S.Ct. 948, 103 L.Ed.2d 80 (1989). Abuse of discretion is the proper standard for review of “determinations made pursuant to a plan that gives the administrator discretionary authority to determine eligibility,” as this plan does. Sweatman v. Commercial Union Ins. Co., 39 F.3d 594, 598 (5th Cir.1994). Where, as here, a challenge to a denial of benefits does not involve the interpretation of plan terms but disputes whether an individual’s conditions qualify as a disability, the inquiry involves factual determinations; therefore, abuse of discretion is the proper standard. See Wade, 493 F.3d at 540.
[608]*608Under an abuse of discretion standard, “[i]f the plan fiduciary’s decision is supported by substantial evidence and is not arbitrary or capricious, it must prevail.” Ellis v. Liberty Life Assurance Co. of Boston, 394 F.3d 262, 273 (5th Cir.2004). “ ‘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.’” Corry, 499 F.3d at 398 (quoting Ellis, 394 F.3d at 273). If a decision is made “without a rational connection between the known facts and the decision or between the found facts and the decision,” the decision is arbitrary. Bellaire Gen. Hosp. v. Blue Cross Blue Shield of Mich., 97 F.3d 822, 828 (5th Cir.1996). “An administrator’s decision to deny benefits must be ‘based on evidence, even if disputable, that clearly supports the basis for its denial.’ ” Lain v. UNUM Life Ins. Co. of Am., 279 F.3d 337, 342 (5th Cir.2002) (quoting Vega, 188 F.3d at 299). If the administrator’s decision to deny a claim is supported by “ ‘some concrete evidence in the administrative record,’” the administrator did not abuse discretion. Id. (quoting Vega, 188 F.3d at 302). The reviewing court may not substitute its judgment for that of the plan administrator. Wade, 493 F.3d at 541.
1. Stnctural Conflict of Interest
McDonald complains that Hartford’s dual role as insurer and plan administrator gave rise to a conflict of interest that justifies de novo review, rather than abuse of discretion; alternatively, McDonald argues that Hartford’s structural conflict suffices to show abuse of discretion. The Supreme Court recognized in Metropolitan Life Insurance Co. v. Glenn that when “a plan administrator both evaluates claims for benefits and pays benefits claims,” it creates a structural conflict of interest. 554 U.S. 105, 128 S.Ct. at 2348. “If a benefit plan gives discretion to an administrator or fiduciary who is operating under a conflict of interest, that conflict must be weighed as a factor in determining whether there is an abuse of discretion.” Firestone, 489 U.S. at 115, 109 S.Ct. 948 (internal modification and quotation marks omitted). “If the administrator has a conflict of interest, we weigh the conflict of interest as a factor in determining whether there is an abuse of discretion in the benefits denial, meaning we take account of several different considerations of which conflict of interest is one.” Holland, 576 F.3d at 247 (internal quotation marks omitted). If claimants do not present evidence of the degree of the conflict, the court will generally find that any conflict is “not a significant factor.” Id. at 249 (finding that where claimant “adduced no evidence ... that [administrator’s structural] conflict affected its benefits decision or that it had a history of abuses of discretion,” any conflict was insignificant in abuse of discretion analysis).
Here, the district court considered Hartford’s “mere technically dual role” and determined that abuse of discretion was the proper standard. An examination of the record bears out the district court’s determination: McDonald has not pointed to any specific evidence of a history of abuses of discretion or of how Hartford’s structural conflict of interest may have affected its benefits decision in this particular case. In addition, the record does not show that McDonald attempted to conduct discovery on any potential conflicts of interest. Rather, the record indicates that Hartford conducted nothing less than a “careful investigation” of McDonald’s claim. See Corry, 499 F.3d at 398 n. 11 (finding that any potential structural conflict of interest did not adversely affect plan administrator’s handling of claim where administrator spent over two and a half years reviewing the claim and hired three specialists who gave “clear and unequivocal” opin[609]*609ions). Here, Hartford engaged three outside specialists to review McDonald’s file; McDonald submitted additional documentation twice; and Hartford entertained two appeals. McDonald has failed to put forth any evidence of the degree of the alleged conflict, and she has also failed to show that any structural conflict impacted Hartford’s decision in her case. Therefore, to the extent that Hartford’s dual role as both insurer and plan administrator may create a conflict, that conflict is not a significant factor that would justify a change in the standard of review. The district court did not err in applying an abuse of discretion standard to Hartford’s decision or in concluding that any alleged conflict did not support McDonald’s claim that Hartford abused discretion in denying benefits.
2. Possible Bias by Reviewing Physicians
McDonald argues that because the three reviewing physicians are employed by agencies that contract with Hartford, the physicians were biased in favor of Hartford; therefore, Hartford abused its discretion in relying on their opinions. McDonald points to several federal district court cases that note a potential conflict of interest on the part of UDC, the organization that employs Drs. Turner and Pick, because of its “significant and ongoing relationship” with Hartford. See, e.g., Caplan v. CNA Fin. Corp., 544 F.Supp.2d 984, 991-92 (N.D.Cal.2008) (reviewing Hartford’s denial of claim with “skepticism” because structural conflict of interest accompanied by “reliance on UDC, a company which Hartford knows benefits financially from doing repeat business with it”). McDonald notes that Drs. Turner and Pick have been criticized by district courts for deficient reviews in similar cases. Hicklin v. Hartford Life & Accident Ins. Co., No. CV06-4543, 2007 WL 4729856, at *7-8, *11 (C.D.Cal. Dec. 12, 2007) (criticizing Hartford for “ignoring] the obvious, comb[ing] the record and [taking] selective evidence out of context as a pretext to deny” a claim, detailing misstatements and omissions by Dr. Turner, and describing Dr. Pick’s review as “deficient”). Hartford, in turn, cites several district court cases affirming Hartford’s claim decisions that relied on opinions of UDC physicians. See, e.g., Singley v. Hartford Life & Accident Ins. Co., 497 F.Supp.2d 807, 812 n. 9 (S.D.Miss.2007) (upholding Hartford’s denial of claim even though Hartford used reviewing physicians from UDC, including Dr. Turner); Dowdy v. Hartford Life & Accident Ins. Co., 458 F.Supp.2d 289, 296 n. 9 (S.D.Miss.2006) (same).
While the Fifth Circuit has yet to discuss in great detail the impact of potential physician bias on the ERISA standard of review, we have briefly dismissed similar arguments in the past. For example, in Sweatman v. Commercial Union Insurance Co., 39 F.3d at 601 n. 14, we considered and rejected the argument that reviewing physicians were biased, based solely on their employment with a contracting agency. In that case, the physicians reviewed twenty to thirty files per month for a contracting agency, but the claimant pointed to no evidence to show that the physicians were financially dependent upon the agency or the plan administrator. Id. We noted that “the only way for [plan administrators] to meet [the claimant’s] standard for impartiality would be to seek physicians willing to volunteer their time to review the medical files of disability claimants.” Id. The Seventh Circuit recently examined a similar issue when a claimant argued that de novo review — rather than abuse of discretion— was warranted because the plan administrator used in-house doctors for its file reviews. Davis v. Unum Life Ins. Co. of Am., 444 F.3d 569, 575-76 (7th Cir.2006). [610]*610The Seventh Circuit held that absent evidence of “any specific incentive [for the in-house doctors] to derail [a] claim,” such as giving the doctors “some specific stake in the outcome of [a] case,” the theoretical argument that “in-house doctors have an inherent conflict in every case” is insufficient to change the standard of review. Id.
Here, McDonald does not appear to have pursued discovery on this issue, nor has she presented the type of specific evidence of bias that would show abuse of discretion or justify a change in the standard of review. While she points to the conclusions of district courts and cites evidence presented in other cases, the record before us contains no documentary or testimonial evidence regarding the financial relationship between Hartford, UDC, and the individual physicians. Furthermore, McDonald does not present any evidence showing an incentive for the doctors to undermine her case in particular. McDonald’s attempts to provide specific evidence of bias fail to rise past the level of conclusory allegations; the district court did not err in finding that Hartford did not abuse its discretion on this point.
3. Lack of Physical Examination
McDonald argues that Hartford abused its discretion by failing to order a new physical examination, or Functional Capacity Examination (FCE), as Drs. Hubbell and Steck recommended during the administrative review process. However, the policy places the burden on the claimant to provide proof of loss — at the claimant’s own expense — including “[objective medical findings which support [the] disability. Objective medical findings include but are not limited to tests, procedures, or clinical examinations standardly [sic] accepted in the practice of medicine, for [the] disabling condition(s).” In addition, “the burden is not solely on the administrator to generate evidence relevant to deciding the claim.” Gooden v. Provident Life & Accident Ins. Co., 250 F.3d 329, 335 (5th Cir.2001) (modification and internal quotation marks omitted); Gothard v. Metro. Life Ins. Co, 491 F.3d 246, 249 & n. 7 (5th Cir.2007) (declining to require a physical examination prior to denial of claim and citing cases in support); cf. Holland, 576 F.3d at 250 (declining to require administrator to consult with vocational expert, and quoting Duhon, 15 F.3d at 1309, for proposition that a “reviewing court [may] decide, on a case-by-case basis, whether under the particular facts the plan administrator abused his discretion by not obtaining the opinion of a vocational rehabilitation expert”). McDonald fails to show abuse of discretion on this point.
L Lack of Deference to Treating Physicians’ Opinions
McDonald contends that Hartford abused its discretion by improperly discounting the opinions of her treating physicians. However, the Supreme Court has explicitly disapproved of a “treating physician” rule in the ERISA context and held that “plan administrators are not obliged to accord special deference to the opinions of treating physicians.” Black & Decker Disability Plan v. Nord, 538 U.S. 822, 825, 123 S.Ct. 1965, 155 L.Ed.2d 1034 (2003). Administrators do not bear “a heightened burden of explanation ... when they reject a treating physician’s opinion.” Id. at 830, 123 S.Ct. 1965. “So long as the [p]lan [a]dministrator’s decision is rationally related to the evidence, we do not require the [p]lan [administrator to credit a particular area of expertise when deciding on an applicant’s prognosis.” Holland, 576 F.3d at 249, 250 (quoting language from Black & Decker, 538 U.S. at 834, 123 S.Ct. 1965, that courts may not “impose on plan administrators a discrete burden of explanation when they credit reliable evidence [611]*611that conflicts with a treating physician’s evaluation”).
While the record does contain some evidence indicating that McDonald suffers from some permanent health issues,8 Hartford’s decision to deny benefits also finds support in the record. The fact that Hartford’s support comes from reviewing physicians does not render its decision arbitrary or capricious; even McDonald’s physicians have not expressed consistent opinions regarding the extent of her disability and her ability to perform sedentary work.
For example, during the remand period, when Hartford re-interviewed Dr. Steck— the neurosurgeon who performed McDonald’s surgery — he did not give a clear “yes” or “no” answer when asked about the extent of McDonald’s disability. The interviewer asked: “Is there anything medically really why [McDonald] could not have done [sedentary work] three months post-op?” Dr. Steck replied:
We are dealing basically with subjective complaints but they are based on objective data in that she has degenerative disc disease in the lumbar spine[;] she has documented lumbar disc herniation that was bad enough to require a lumbar decompression fusion, and I have dealt with enough of these patients [to know] that there will be a significant subset who will not return to any type of employment due to complaints of pain and there is no way I can prove that she doesn’t have pain, we could always say well I don’t see why she can’t work,
most people could but not everybody ....
The objective data is that, yes, she did have surgery, the subjective component is that although she is neurologically normal and everything looks just fine, although many people in this situation would be able to work although she tells me that she actually is better from the surgery, she had it, but doctor I just hurt too much to work. Based on that I just can’t sign a letter or do a dictation saying that I think she can work,
(emphasis added). While Dr. Steck expressed concern for McDonald’s subjective complaints of pain and noted that the subjective complaints were based objectively on her degenerative disc disease, he also made the statement that “she is neurologically normal and everything looks just fine.”
When Hartford reached its final decision, it had consistent reports from the three reviewing physicians indicating that McDonald did not meet the definition of “disabled”; it had Dr. Steck’s mixed statement; and it had Dr. Hubbell’s letter dated February 13, 2006, which recommended that McDonald receive additional pain treatment and indicated: “I know that she can’t perform full-time work at this time, but I believe that if her cervical pain were resolved, she would be able to return to at least twice as much work as she is able to do today if not full-time work.” Hartford’s decision to deny benefits has clear support in the administrative record, and the deci[612]*612sion is rationally related to the evidence Hartford had before it at the time of the decision. See Holland, 576 F.3d at 249. “[T]he job of weighing valid, conflicting professional medical opinions is not the job of the courts ... [but rather the job of] the administrators of ERISA plans,” Cor-ry, 499 F.3d at 401, and therefore McDonald’s argument that Hartford failed to give adequate weight to the opinions of her treating physicians must fail. Hartford did not abuse its discretion in adopting the opinions of the reviewing physicians over the treating physicians.
5. Lack of Consideration of Subjective Complaints of Pain
McDonald argues that Hartford abused its discretion by failing to give adequate weight to her consistent complaints of pain. To support her argument, McDonald relies on Audino v. Raytheon Company Short Term Disability Plan, 129 Fed.Appx. 882 (5th Cir.2005) (per curiam), where we reversed a summary judgment in favor of a plan administrator that had denied benefits to a claimant who complained of pain. In Audino, we found an abuse of discretion because the administrator
ignored [the claimant’s] consistent complaints of pain as subjective, either minimized or ignored objective evidence of disability corroborating those complaints, and concluded that the evidence did not show an inability to do her job functions without analyzing the effect that her conditions would have on her ability to perform her specific job requirements.
129 Fed.Appx. at 885. However, in that case the claimant presented specific evidence of misstatements and oversights by the reviewing physicians that the plan administrator relied upon in denying the claim. Id. at 884-85 (noting that one physician misstated objective test results, while another mentioned exam results in a summary of evidence but failed to discuss those results in analysis of whether claimant was disabled).
More applicable than Audino is the case of Corry v. Liberty Life Assurance Co. of Boston, where we addressed in detail whether an administrator’s review adequately considered a claimant’s subjective complaints of pain. 499 F.3d at 399-401. There, the claimant’s experts opined that she was disabled due to fibromyalgia — a diagnosis reached by reliance on the claimant’s subjective reports of pain. Id. at 401. The plan administrator ultimately rejected the claimant’s assertion that she was disabled, relying on the opinions of three outside reviewing physicians. Id. All three reviewing physicians discussed the claimant’s subjective complaints and her previous diagnosis of fibromyalgia in their analyses; yet they each ultimately concluded that no medical evidence existed establishing a disability. Id. In Corry, we concluded that this constituted a “battle of the experts,” where the administrator was “vested with discretion to choose one side over the other”; therefore, we rejected the argument that the administrator “fail[ed] to consider and give proper weight to relevant evidence” of subjective pain. Id.
Here, Hartford and its reviewing physicians clearly “considered, evaluated, and addressed” McDonald’s subjective complaints of pain; however, the reviewing physicians still reached the conclusion that McDonald’s administrative record did not contain objective medical evidence of disability. Corry, 499 F.3d at 401. The denial letters indicate that Hartford considered her subjective complaints. In the first denial letter, Hartford acknowledged McDonald’s continuing neck pain and her “difficulties with pain” but concluded that no neurological abnormalities were present and that the evidence did not “support a functional impairment that would preclude [613]*613[her] from performing the material and substantial duties of [her] regular occupation on a full-time basis.” When denying her first appeal, Hartford informed McDonald that:
we considered your self-reported symptoms and to what extent the findings on physical examination and testing results confirm the symptoms. We also considered the impact the findings would have as far as your ability to function on a daily basis and how it would continuously affect your ability to perform your regular occupational work activity on a full-time basis.
In its final denial letter, Hartford mentioned that it had considered letters from McDonald herself and from Dr. Hubbell detailing her subjective complaints. In addition, the administrative record contains notes from an interview with McDonald, detailing the impact of her pain on her daily life at work and at home.
The reviewing physicians also clearly considered and addressed McDonald’s subjective complaints. Dr. LeForce, the initial reviewing physician, noted “complaints of neck and low back pain,” and “continued neck pain.” On the first administrative appeal, Dr. Turner discussed McDonald’s reports that she suffers radiating low back pain, aggravated by sitting and walking and partially relieved by rest, as well as chronic neck pain. Finally, Dr. Pick considered evidence that McDonald suffered from “chronic pain” and clinical notes from her treating physicians regarding her subjective complaints, including: “a history of an insult with severe low back pain and radiating right pain and recent onset numbness in her upper extremities”; “lumbar pain as well as leg pain”; “[s]he states the pain is getting worse”; “[s]he still has some low back pain, some buttock pain, some chronic neck pain, but all those better than preop.” All three reviewing physicians considered McDonald’s subjective complaints but ultimately concluded that these subjective complaints were insufficient to support a finding of disability.
Unlike in Audino, the reviewing physicians did not ignore McDonald’s complaints but included them in their analyses. McDonald argues that the reviewing physicians “mischaracterized the results of her MRIs,” but even Dr. Steck, her treating physician, stated that McDonald was “neurologically normal.” Any difference of opinion between the reviewing and treating physicians on the interpretation of her MRIs falls into Hartford’s area of discretion; McDonald does not point to any affirmative misstatements of objective test results of the kind presented in Audino.
While Hartford’s conclusions conflict with Dr. Hubbell’s evaluation of McDonald’s condition, Dr. Steck’s final interview with Hartford on remand contains language that supports the conclusion that the record did not contain objective medical evidence of disability. Hartford has discretion in this battle of experts, and in the absence of evidence that Hartford failed to consider McDonald’s complaints of pain, Hartford was within its discretion to accept the opinions of its three qualified medical experts. Hartford’s decision was neither arbitrary nor capricious on this point.
6. Insufficient Evidence to Support Denial of Claim
Finally, and more generally, McDonald complains that Hartford “cherry-picked” quotes and facts out of the administrative record to support its decision to deny her claim for benefits. However, under Fifth Circuit law, Hartford has discretion under the plan to investigate the claim and draw the conclusions it deems proper. “The law requires only that substantial evidence support a plan fiduciary’s decisions, including those to deny or to terminate benefits, not that substantial ev-[614]*614idenee (or, for that matter, even a preponderance) exists to support the employee’s claim of disability.” Ellis, 394 F.3d at 273.
Here, Hartford solicited the medical opinions of three separate physicians. Drs. Turner and Pick are both board certified orthopedic surgeons, “specialists and qualified experts in [a] field[] specifically related” to McDonald’s symptoms; at least two of the three physicians reached their conclusions independently.9 Corry, 499 F.3d at 402. In Corry, we found it “indisputable that the medical opinions of [the plan administrator’s] three consulting physicians ... constitute substantial evidence supporting [the disability decision].” Id. In addition to the opinions of the three reviewing physicians, the administrative record contains other evidence in support of Hartford’s decision: in particular, Dr. Steck’s statement that McDonald was “neurologically normal and everything looks just fine.” Furthermore, McDonald admitted that she gets help to fulfill the duties of her job and that her employer has been cooperative and flexible. Hartford’s decision does not need to be correct; it simply must not be arbitrary. Cf. Gothard, 491 F.3d at 250 (“MetLife’s decision may not be correct, but we cannot say that it was arbitrary.”). On the administrative record, Hartford’s decision to deny her claim was supported by substantial evidence and there was no abuse of discretion.
III. CONCLUSION
For the above reasons, we AFFIRM.
Pursuant to 5th Cir. R. 47.5, the court has determined that this opinion should not be published and is not precedent except under the limited circumstances set forth in 5th Cir. R. 47.5.4.