Plaintiff-appellant Susan Hennen worked as a sales specialist for NCR Corporation from 2010 to May 2012, when she sought treatment for a back injury. As an employee, Hennen was covered by long-term disability insurance under a group policy provided by defendant-appellee Metropolitan Life Insurance Company ("MetLife"). When physical therapy and surgery failed to resolve her injury, Hennen applied for long-term disability benefits under the insurance plan.
Acting as plan administrator, MetLife agreed that Hennen was disabled and paid benefits for two years. The plan has a two-year limit, however, for neuromusculoskeletal disorders. That limit is subject to several exceptions, one of which applies to cases of radiculopathy. After paying for two years, MetLife terminated Hennen's benefits, finding that the two-year limit applied. Hennen believes that she is entitled *535to continued benefits because she has radiculopathy. She sued under the Employee Retirement Income Security Act of 1974, 29 U.S.C. § 1001 et seq. (ERISA), arguing that MetLife's determination that she did not have radiculopathy was arbitrary and capricious. The district court granted summary judgment for MetLife, and Hennen appeals. We reverse and remand. MetLife acted arbitrarily when it discounted the opinions of four doctors who diagnosed Hennen with radiculopathy in favor of the opinion of one physician who ultimately disagreed, but only while recommending additional testing that MetLife declined to pursue.
I. Factual & Procedural Background
A. The Employee Benefit Plan
The parties agree that from 2012 through 2014, Hennen qualified for disability coverage under the NCR employee benefit plan. But the plan limits coverage for certain conditions to two years. As relevant here, the plan limits coverage for any disability due to neuromusculoskeletal and soft tissue disorders, "including, but not limited to, any disease or disorder of the spine or extremities and their surrounding soft tissue." This limit applies to Hennen-who was suffering from spinal injury-unless she has "objective evidence" that she suffers from one of six exceptions.
Hennen claims she has radiculopathy, one of the exceptions. The plan defines radiculopathy as: "Disease of the peripheral nerve roots supported by objective clinical findings of nerve pathology." At the end of Hennen's initial two years of coverage, MetLife concluded that she did not have objective evidence of radiculopathy. Hennen disputes this finding.
B. Hennen's Medical History
Hennen has a history of lower back problems. She had her first two back surgeries in 2003 and 2008, which included a surgery fusing three vertebrae in her lower back. She was able to resume her normal routine after the 2008 surgery with the help of a prescription painkiller. Then in February 2012, Hennen suffered a new back injury. She sought treatment from Dr. Shana Margolis, a specialist in physical medicine and rehabilitation. Hennen reported pain radiating down her legs. Dr. Margolis diagnosed her with L3-L4 disc herniation, myofascial and neuropathic pain, and bilateral lumbar radiculitis. Dr. Margolis treated Hennen's pain with physical therapy and pain management techniques.1
Still in pain months later, in May 2012 Hennen enrolled in a month-long pain management program. Dr. Randy Calisoff examined her and noted that she presented "with an exacerbation of low back pain as well as bilateral posterior leg pain running from the low back to the ankles," and that her pain "has remained flared with MRI revealing a new herniation" of a spinal disc. He noted her diagnosis as L3-L4 disc herniation, lumbar myofascial pain syndrome, and lumbar radiculitis. In the pain management program, Hennen continued physical and occupational therapy. She also underwent relaxation therapy and pain psychotherapy, and she was prescribed oral pain medications. In August 2012, Dr. Margolis cleared Hennen to return to work.2
*536Hennen claims that the more conservative treatments did not relieve her pain, so she sought another opinion from an orthopedic surgeon, Dr. Frank Phillips. He recommended surgery and operated on Hennen's L3-L4 disc herniation on September 24, 2012. In his operative report, Dr. Phillips noted that "the nerves were free of compression and mobile" at the end of the surgery. He then informed MetLife that Hennen was under his care and required eight weeks off from work. MetLife approved Hennen's long-term disability benefits effective November 12, 2012 and warned her of the two-year limit on coverage for neuromusculoskeletal disorders.
At follow-up appointments, Dr. Phillips noted that Hennen was struggling to sit for any extended period of time and complained of "bilateral pain in the buttocks and posterior thighs to the level of the knee," which he described as "persistent radicular complaints." Dr. Phillips ordered an MRI to "rule out any recurrent or residual neural compression." He noted that if the MRI identified "no frank compression," then the symptoms likely represented "some residual nerve pain" that should be treated conservatively, without surgery. Hennen had the MRI on December 28, 2012, which showed no nerve compression.
With surgery no longer an option, Hennen sought treatment from Dr. Asokumar Buvanendran, an anesthesiologist who provided ongoing pain management care. Hennen reported leg weakness and pain, which she claimed was worse than her lower back pain. Dr. Buvanendran treated Hennen's symptoms with a series of epidural steroid injections. He diagnosed her with post-laminectomy pain syndrome and lumbar radiculopathy. When the injections failed to improve Hennen's pain, Dr. Buvanendran implanted an epidural spinal cord stimulator, which delivers a low-voltage electrical current to the spinal cord to block pain sensation. The stimulator provided Hennen relief for a few weeks, but then she again reported recurrent leg weakness and tripping. After dislodging the device in a fall, Hennen had multiple surgeries to fix ongoing issues with it.3
In early 2014, Hennen consulted another orthopedic surgeon, Dr. Shane Nho, about left hip pain she experienced after a fall. An MRI revealed a partial muscle tear and problems with her hip joint, but no nerve compression. Dr. Nho treated the hip injuries with a steroid injection and then surgery. He advised that Hennen could not work for four months following surgery and prescribed physical therapy. Hennen began physical therapy for her hip in June 2014, but she continued to struggle with nerve pain radiating down her leg.
C. MetLife's Disability Determination
With the two-year limit looming, in July 2014 MetLife contacted Hennen's doctors for information about her condition. Dr. Buvanendran responded that Hennen was unable to work due to post-laminectomy pain syndrome and lumbar radiculopathy. Dr. Nho advised MetLife that Hennen's *537hip was structurally sound and that he deferred to Dr. Buvanendran on her back symptoms.
MetLife contacted Hennen to explain that her medical condition fell within the plan's neuromusculoskeletal limit. Around the same time, a nurse-consultant for MetLife reviewed Hennen's file and noted a lack of current MRI or electromyography results in the file. MetLife explained to Hennen that lumbar radiculopathy is an exception to the neuromusculoskeletal limitation, but that Hennen needed additional documentation of a diagnosis. In response, Hennen had another MRI of her lower spine on September 24. Dr. Buvanendran did not forward the results to MetLife right away.
On October 13, 2014, MetLife wrote Hennen to reiterate that her benefits were scheduled to end on November 11, quoting the plan's neuromusculoskeletal limit. MetLife advised Hennen that she could appeal this decision and, if she did, should provide "Office visit notes to support a non-limited disability," "Objective exam findings," "Current test results (MRI, CT, EMG)," "Current restrictions and limitations," and "Current treatment plan." A few days later, Dr. Buvanendran sent MetLife a letter emphasizing that he diagnosed Hennen with post-laminectomy syndrome and lumbar radiculopathy. Dr. Buvanendran described Hennen's continued disabling pain and limited functionality, and he listed the history of her medical treatments. He also faxed the previous month's MRI to MetLife, which showed a new annular fissure but no herniation or spinal stenosis-that is, a new tear in one of Hennen's spinal discs, but no nerve compression.
MetLife consulted with Dr. David Peters, a family medicine physician, who reviewed the MRI and opined that it did not show compression that would support a diagnosis of lumbar radiculopathy. MetLife upheld its determination that Hennen did not qualify for the radiculopathy exception for extended disability benefits.
D. Hennen's Administrative Appeal
Hennen appealed MetLife's disability determination through its administrative review process. Through counsel, she submitted a letter challenging MetLife's conclusion that her medical records did not document radiculopathy or neurological deficits. She also submitted the results of an electromyogram (EMG) by Dr. Joseph Kipta, a certified neurologist and clinical neurophysiology fellow, on June 8, 2015. An EMG is a diagnostic procedure that can reveal nerve dysfunction and problems with the nerve-muscle connection. Dr. Kipta recorded some nerve-related abnormalities on the EMG and concluded that it confirmed radiculopathy in four nerve roots, though he also wrote that he could not rule out a different nerve disorder. Dr. Kipta also conducted a physical examination, which revealed that Hennen had normal strength reflexes in her legs but diminished nerve sensation. Dr. Kipta concluded that the EMG and nerve sensation abnormalities supported a diagnosis of radiculopathy. Dr. Rabia Malik, a board-certified neurologist and neurophysiologist and assistant professor of neurology, supervised Dr. Kipta's EMG and agreed with his findings.
MetLife's medical director, Dr. Dupe Adewumni, reviewed Hennen's appeal. Dr. Adewumni agreed with Dr. Kipta and Dr. Malik that the EMG supported a diagnosis of lumbar radiculopathy. He also reasoned that although the EMG was not conducted until June 2015, it was reasonable to conclude that Hennen had radiculopathy on November 11, the end of the initial two-year coverage period.
Concluding that Hennen satisfied the radiculopathy exception, MetLife turned to *538assessing her functionality to determine whether her condition made her disabled under the plan. MetLife consulted with Dr. Neil McPhee-whose expertise is physical medicine, rehabilitation, and pain medicine-to assess her functionality. MetLife asked him two questions. The first was whether Hennen's medical file supported functional limitations due to a physical condition as of November 12, 2014, and if so what those functional limitations were. The second question was whether clinical evidence supported limitations or side effects due to medications.
Despite the limited scope of these questions, Dr. McPhee addressed and disagreed with the finding that Hennen had radiculopathy. He opined that the June 2015 EMG "was negative for active radiculopathy with no abnormal spontaneous or insertional activity recorded in any of the muscles examined." He criticized Dr. Kipta's and Dr. Malik's findings and wrote that Dr. Kipta "should have performed needle examination in corresponding right lower extremity muscles ... before coming to a conclusion of polyradiculopathies involving four nerve roots," which Dr. McPhee found unlikely because the MRI did not reveal ongoing compression of any nerves. Dr. McPhee also criticized Hennen's self-reported pain levels as implausible and inconsistent. He found her doctor's notes on muscle weakness inconsistent, too. Dr. McPhee then answered the questions MetLife had asked. He found that Hennen's physical condition-which he summarized as "longstanding chronic narcotic dependent pain," surgical history of spinal fusions, and recent annular fissure-limited her ability to work.
MetLife reviewed Dr. McPhee's report and relied on his assessment of the EMG to reject medical director Dr. Adewumni's conclusion and to decide that Hennen did not have radiculopathy that would avoid the two-year limit on benefits. Through counsel, Hennen submitted a formal response challenging Dr. McPhee's conclusion. Hennen asserted that the June 2015 EMG confirmed radiculopathy, that Dr. McPhee's criticisms of the EMG were unsupported, and that Hennen's ongoing disability entitled her to benefits under the policy's terms. Dr. Buvanendran responded that the "EMG study results prove, without any doubt, that the patient suffers from radiculopathy." He also disagreed with Dr. McPhee's opinion that Hennen's self-reported pain levels were inconsistent or implausible, explaining that different activities and treatments caused her pain to vary from day to day.
In response to this challenge to his opinion, Dr. McPhee prepared an addendum in response. He said he continued to believe that Hennen's EMG and MRI did not show radiculopathy, but also clarified:
it is still my opinion that additional electrodiagnostic testing would be helpful. Similarly, consideration should be given to an independent medical examination by a physical medicine and rehabilitation specialist or neurologist whose training and practice includes electromyography which can be used as an extension of the clinical examination if needed to further assess the issue of possible radiculopathy.
Dr. McPhee concluded that, at most, Hennen "may have lumbar radiculitis with a past history of nerve compression prior to corrective surgery rather than lumbar radiculopathy based on clear cut examination findings, imaging, and/or electrodiagnostic findings."
MetLife did not order an independent medical examination or additional electrodiagnostic testing, as Dr. McPhee recommended. Nor did MetLife explain why additional observation or testing was unnecessary to resolve Hennen's appeal. Instead, *539the next day, MetLife upheld its decision that the two-year neuromusculoskeletal limit applied and that Hennen did not satisfy the exception for radiculopathy supported by objective evidence.
E. This Lawsuit
Hennen sued MetLife in the Northern District of Illinois, seeking ERISA plan benefits under 29 U.S.C. § 1132(a)(1)(B). Hennen and MetLife agreed to limit discovery to the administrative record and to exchange cross-motions for summary judgment. The district court granted summary judgment for MetLife, reasoning that MetLife reasonably interpreted the plan to require proof of "active radiculopathy" in November 2014 and that Hennen had failed to offer evidence of active radiculopathy at that time. The court also found that MetLife had reasonably decided to credit Dr. McPhee's opinion over the opinions of the other doctors. The district court entered judgment in favor of MetLife.
II. Analysis
A. Standard of Review
We review de novo the district court's grant of summary judgment. Tompkins v. Central Laborers' Pension Fund , 712 F.3d 995, 999 (7th Cir. 2013), citing Edwards v. Briggs & Stratton Ret. Plan , 639 F.3d 355, 359 (7th Cir. 2011). Where, as here, the plan grants the administrator discretionary authority to determine benefits, we review the administrator's decision under the arbitrary-and-capricious standard. Id. , citing Firestone Tire & Rubber Co. v. Bruch , 489 U.S. 101, 115, 109 S.Ct. 948, 103 L.Ed.2d 80 (1989). This standard is deferential but " not a rubber stamp," and "we will not uphold a termination when there is an absence of reasoning in the record to support it." Holmstrom v. Metropolitan Life Ins. Co. , 615 F.3d 758, 766 (7th Cir. 2010), quoting Hackett v. Xerox Corp. Long-Term Disability Income Plan , 315 F.3d 771, 774-75 (7th Cir. 2003).
We will uphold the administrator's decision "as long as (1) it is possible to offer a reasoned explanation, based on the evidence, for a particular outcome, (2) the decision is based on a reasonable explanation of relevant plan documents, or (3) the administrator has based its decision on a consideration of the relevant factors that encompass the important aspects of the problem." Tompkins , 712 F.3d at 999, quoting Hess v. Hartford Life & Accident Ins. Co. , 274 F.3d 456 (7th Cir. 2001). "In conducting this review, we remain cognizant of the conflict of interest that exists when the administrator has both the discretionary authority to determine eligibility for benefits and the obligation to pay benefits when due." Jenkins v. Price Waterhouse Long Term Disability Plan , 564 F.3d 856, 861 (7th Cir. 2009), citing Metropolitan Life Ins. Co. v. Glenn , 554 U.S. 105, 108, 128 S.Ct. 2343, 171 L.Ed.2d 299 (2008). That is the situation here, and the conflict of interest is "weighed as a 'factor in determining whether there is an abuse of discretion.' " Glenn , 554 U.S. at 115, 128 S.Ct. 2343, quoting Firestone , 489 U.S. at 115, 109 S.Ct. 948.
B. Disability Determination
Hennen has shown that MetLife's decision to terminate her benefits was arbitrary and capricious. MetLife acted arbitrarily when it credited Dr. McPhee's opinion over the opinions of four other doctors, including Hennen's treating physician, two neurologists with clinical training in electrodiagnostic testing, and MetLife's own medical director. The arbitrary character is highlighted by MetLife's choice not to follow Dr. McPhee's ultimate recommendation, when his opinion was challenged, to order an independent medical evaluation *540and additional electrodiagnostic testing. For these reasons, we agree with Hennen that MetLife acted arbitrarily and that a remand to MetLife is necessary.
MetLife terminated Hennen's benefits because it found that she lacked "objective evidence" of active radiculopathy. To reach this conclusion, MetLife relied on Dr. McPhee's opinion based on his review of Hennen's files without examining her. He concluded that Hennen's EMG was "negative for active radiculopathy."
Ordinarily, a plan administrator is free to choose among different medical opinions so long as the administrator provides a rational explanation that has support in the record. Becker v. Chrysler LLC Health Care Benefits Plan , 691 F.3d 879, 889 (7th Cir. 2012) ; see also Black & Decker Disability Plan v. Nord , 538 U.S. 822, 831, 123 S.Ct. 1965, 155 L.Ed.2d 1034 (2003). Also, under ERISA there is no presumption against file-reviewing physicians or in favor of examining physicians. Nord , 538 U.S. at 831, 123 S.Ct. 1965 (finding no "heightened burden of explanation on administrators when they reject a treating physician's opinion"); Leger v. Tribune Co. Long Term Disability Ben. Plan , 557 F.3d 823, 832 (7th Cir. 2009) (rejecting presumption against file-reviewing doctors). The fact that MetLife credited a file-reviewing physician over competing opinions does not itself make MetLife's decision arbitrary.
That said, MetLife engaged in arbitrary decision-making in this case. To start, every physician who examined Hennen after her 2012 spine operation concluded that she had radiculopathy : Dr. Buvanendran, her treating physician; Dr. Kipta, who performed the EMG; and Dr. Malik, who oversaw the EMG. Doctors Margolis and Calisoff also recorded nerve-related symptoms and diagnosed Hennen with radiculitis (meaning inflammation of the nerve roots ) before Hennen's 2012 surgery to fix a herniated disc. Dr. Adewumni, MetLife's medical director, reviewed Hennen's file and concluded that she had objective evidence of radiculopathy.
Those doctors' opinions had substantial medical support. Hennen's medical file contains at least five doctors' clinical observations-from Drs. Margolis, Calisoff, Buvanendran, Kipta, and Malik-of muscle weakness and sensory loss, two symptoms of nerve root damage. Those observations both pre-date and post-date Hennen's 2012 surgery to relieve nerve compression. In addition, the EMG showed several abnormalities that Dr. Kipta, Dr. Malik, Dr. Buvanendran, and Dr. Adewumni thought were consistent with radiculopathy. Dr. McPhee was the only doctor who believed that the abnormalities were too few to show radiculopathy. (Dr. Peters agreed with Dr. McPhee that a 2014 MRI did not show radiculopathy, but he did not examine Hennen or consider evidence beyond the MRI itself.)
Faced with these various diagnoses of radiculopathy, MetLife chose to credit Dr. McPhee's opinion that Hennen did not have objective evidence of radiculopathy. But MetLife never asked Dr. McPhee to diagnose Hennen or to make that finding. In fact, MetLife referred Hennen's case file to Dr. McPhee after its own medical director, Dr. Adewumni, concluded that she actually met the plan's radiculopathy exception. Once Dr. Adewumni made that decision, MetLife asked Dr. McPhee only to assess Hennen's functional limitations given her physical condition and any side effects from medication. Dr. McPhee took it upon himself to assert that Hennen did not have radiculopathy-or at least, not objective evidence of it-at all.
It's not just that all the examining doctors disagreed with Dr. McPhee on the *541key issue. Another indication of arbitrary decision-making was MetLife's failure to heed Dr. McPhee's recommendation to seek more electrodiagnostic testing and an independent medical evaluation. When Hennen and Dr. Buvanendran challenged his opinion, Dr. McPhee responded "that additional electrodiagnostic testing would be helpful" and that "consideration should be given to an independent medical examination" to "further assess the issue of possible radiculopathy."
MetLife chose not to follow up on Dr. McPhee's advice. Instead, MetLife treated his original opinion as definitive and immediately sent Hennen a letter affirming the denial of her benefits. The letter asserted that MetLife found Dr. McPhee's opinion "more compelling" than other doctors'. MetLife did not address Dr. McPhee's recommendation for additional testing and examination to settle the dispute between his view and the views of all the doctors who had examined her. "MetLife's reliance on the opinions of its reviewing doctor[ ] here is all the more arbitrary in light of the fact that it ignored the key final recommendation" of that doctor for further testing to resolve the dispute more reliably. Holmstrom , 615 F.3d at 775.
Together, these facts show that MetLife arbitrarily and capriciously terminated Hennen's benefits. As a fiduciary, MetLife owed Hennen a duty to execute faithfully the terms of the plan and "to see that those entitled to benefits receive them." Gaither v. Aetna Life Ins. Co. , 394 F.3d 792, 807-08 (10th Cir. 2004) ; see also Tompkins , 712 F.3d at 1001 ("[A] benefits determination by a plan administrator is a fiduciary act, one in which the administrator owes a special duty of loyalty to the plan beneficiaries."). Here, MetLife took an extra step for its own benefit when it referred Hennen's file to Dr. McPhee for review. But when Dr. McPhee recommended that MetLife take an additional step for Hennen's benefit-to confirm whether his lone opinion that she did not suffer from radiculopathy was accurate-MetLife declined to take that step. That was arbitrary and capricious.
Hennen raises two more points that concern us and that need to be addressed on remand. First, Hennen argues that MetLife unreasonably interpreted the radiculopathy exception to require ongoing compression of a nerve root when that is only one potential cause of radiculopathy. As Hennen points out, the plan does not define radiculopathy as nerve root disorders resulting from ongoing compression. And at oral argument, MetLife agreed that nerves can remain damaged after compression is relieved by surgery.4
Second, Hennen argues that neither MRIs nor EMGs are conclusive for radiculopathy. Hennen cites articles saying, for example, that "radiculopathies may occur without structural findings on MRI, and likewise, without EMG findings." Timothy Dillingham, How to Evaluate Patients with Suspected Radiculopathy , AANEM Basics with the Experts, 9 (2013). MetLife does not point us toward any medical opinions to the contrary. MetLife responds only that it reasonably concluded that the MRI did not confirm radiculopathy, that the causes of inflammation to nerve roots are unclear, and that EMGs are rarely falsely positive for radiculopathy.
MRI and EMG findings could be relevant-even highly relevant-in diagnosing *542radiculopathy, but MetLife's unpersuasive responses in this appeal are troubling. Although it is reasonable for MetLife to require objective support for a diagnosis of radiculopathy, it would be unreasonable to discount clinical observations of Hennen's treating physicians in favor of testing that is inconclusive for the condition. This issue needs further attention on remand.
As we have often noted, diseases like radiculopathy present problems for insurers and insured alike because they involve pain that can be difficult to confirm through objective evidence. See Holmstrom , 615 F.3d at 769. Hennen has consistently reported radiating, debilitating pain that sounds like it is caused by nerve-root injury. She has had three surgeries to address spinal problems and has surgically fused vertebrae from the S1 level up to the L3 level. Her most recent MRI indicated more spinal degeneration at the L2-L3 disc, though it did not reveal compression that would affect a nerve at that level. She was entitled to continued long-term disability benefits if she can show that her pain was real and was caused by diseased nerve roots, which is difficult to do.
Hennen's reported pain, of course, is subjective rather than objective evidence. MetLife is "understandably concerned about the possibility of malingering and exaggeration." Holmstrom , 615 F.3d at 775. MetLife must also distinguish between legitimate neuromusculoskeletal disorders, which can be difficult to diagnose, and drug addiction and drug-seeking behaviors. Id.
Dr. McPhee does not accuse Hennen outright of feigning pain, but he suggests that her pain is implausible because it has varied in severity and her recorded vital signs were inconsistent with the severity of the pain she reported. These could turn out to be valid critiques, but critical facts prevent us from upholding MetLife's decision based on this line of Dr. McPhee's reasoning. Dr. McPhee never examined Hennen, so he was not in the position to determine whether she was reporting her pain reliably, exaggerating, or a mix of both. Perhaps recognizing the shortcomings of a file review, Dr. McPhee pointed out the inconsistencies and recommended an independent medical examination to address them. Without explanation, MetLife never followed up on that recommendation but just terminated her benefits. As we said above, this was an arbitrary abuse of MetLife's discretion and a violation of the fiduciary duty it owed Hennen as a plan beneficiary. See Tompkins , 712 F.3d at 1001, quoting Raybourne v. Cigna Life Ins. Co. of N.Y. , 700 F.3d 1076, 1081-82 (7th Cir. 2012) ; see also Gaither , 394 F.3d at 807-808 ("While a fiduciary has a duty to protect the plan's assets against spurious claims, it also has a duty to see that those entitled to benefits receive them.").
C. Remedy
"In a case where the plan administrator did not afford adequate procedures in its initial denial of benefits, the appropriate remedy respecting the status quo and correcting for the defective procedures is to provide the claimant with the procedures that she sought in the first place." Hackett , 315 F.3d at 776, citing Wolfe v. J.C. Penney Co., Inc. , 710 F.2d 388, 394 (7th Cir. 1983). The fact that MetLife acted arbitrarily "does not mean that the claimant is automatically entitled to benefits." Id. A remand to MetLife is necessary here so that it can reassess Hennen's claim consistent with this opinion. Also, MetLife has not yet determined Hennen's degree of disability. That determination will be necessary if on remand MetLife finds that Hennen has satisfied the radiculopathy exception to the neuromusculoskeletal limit.
REVERSED and REMANDED.
Easterbrook, Circuit Judge, dissenting *543MetLife's decision must stand unless arbitrary and capricious. That's an exceptionally difficult standard for any plaintiff to meet-and impossible when the plan's decision has the reasoned support of a physician. My colleagues' contrary conclusion boils down to the view that the medical majority should rule. They believe that more physicians found radiculopathy (which would qualify Hennen for long-term benefits) than found radiculitis (which would not). A welfare-benefit plan might adopt a majority-rules position, but this plan did not. As long as a decision has rational support in the record, it must stand. Becker v. Chrysler LLC Health Care Benefits Plan , 691 F.3d 879, 885 (7th Cir. 2012) ; Pokratz v. Jones Dairy Farm , 771 F.2d 206, 208-09 (7th Cir. 1985). Dr. McPhee's conclusion cannot be described as irrational (nor do my colleagues so describe it). That requires decision for MetLife.
This case is nothing like Holmstrom v. Metropolitan Life Insurance Co. , 615 F.3d 758 (7th Cir. 2010). The primary physician who found Holmstrom not disabled retracted that opinion when presented with additional evidence, but the plan adhered to its decision even though the rug had been pulled out from under it. Dr. McPhee, by contrast, did not retract his analysis when confronted with a challenge. Instead he reaffirmed it. That's the opposite of what happened in Holmstrom .
My colleagues emphasize that, while adhering to his view, Dr. McPhee also recommended additional tests. But they don't explain why this required MetLife to conduct them. A plan's language might require more testing whenever a physician so recommends, but this plan does not. That leaves MetLife with discretion. Recommendations of additional testing are common in the medical profession, often (and perhaps here) just to protect a decision from criticism. Defensive medicine is a tendency often decried as needlessly driving up the cost of medical care. Sooner or later the marginal value of testing is negative: the additional information is worth less than the cost. Hennen had been tested over and over by quite a few physicians. I can't see why it was irrational for MetLife to call a halt, and again my colleagues do not describe this as an irrational decision. Indeed, Hennen herself may have seen that there was little point in more tests; after all, she could have had them on her own initiative but chose not to do so. MetLife's conclusion that enough is enough must be respected.
The majority portrays Dr. McPhee as an outlier in finding the absence of radiculopathy, yet Drs. Margolis and Calisoff joined Dr. McPhee in diagnosing Hennen with radiculitis rather than radiculopathy. The majority's assertion that "Dr. McPhee was the only doctor who believed that the abnormalities were too few to show radiculopathy" (slip op. at 14) sits uncomfortably beside its acknowledgment that Dr. Peters reached the same conclusion.
My colleagues implicitly assume that personal examination enables a physician to separate radiculitis from radiculopathy, but they do not cite any medical support for the view that a hands-on examination is necessary or even helpful. As far as I can see a diagnosis depends on an accurate interpretation of tests plus checks for the presence of diagnostic clues that are recorded in examining physicians' notes. Dr. McPhee delivered a 36-page report explaining why Hennen's medical record does not support a finding of radiculopathy ; this report critiques the approach of physicians who concluded otherwise. ERISA does not authorize the federal judiciary to substitute its medical judgment *544for the one accepted by the plan's administrator.
Given that Drs. McPhee, Margolis, and Calisoff all found that Hennen suffers from radiculitis rather than radiculopathy, this should have been an easy case.