Thomas BUFFONGE, Plaintiff, Appellant, v. The PRUDENTIAL INSURANCE CO. OF AMERICA; Getronics Wang Co., LLC, Defendants, Appellees, 426 F.3d 20

LYNCH, Circuit Judge. We have not previously addressed the issue raised in this ERISA appeal, which comes to us from the district court’s entry of

summary judgment concluding that a claims administrator’s denial of long-term disability benefits was not arbitrary and capricious. The plaintiff appeals and asserts that his ERISA rights were violated because the administrator relied on material mischaracterizations of the medical record. The claims administrator, Prudential Insurance Company of America, and eo-ap-pellee Getronies Wang Co., LLC, the employer, stated that they based their denial of long-term disability benefits to former Wang employee Thomas Buffonge on four pieces of evidence. The denial rested heavily, but not exclusively, on the opinion of Dr.

Jonathan Rutchik, a consultant hired by Prudential who did not examine Buffonge but who reviewed Buffonge’s medical history and concluded that there was a “consensus” that Buffonge could perform sedentary duties, such as keyboard work. In fact, the medical records showed that no such consensus existed, and, in other ways, Dr. Rutchik’s report materially mischaracterized Buffonge’s medical history. Prudential also rested its benefits denial on the report of another physician, but that report supported the grant, not the denial, of benefits.

Further, Buffonge’s medical records, on which Prudential also purportedly relied, demonstrated the two points made above. We conclude that the process used was materially tainted, and the taint was sufficiently prejudicial, so as to render the process arbitrary. We remand to the district court to remand to the claims administrator for a new review of Buffonge’s claim. We emphasize that we do not reach the issue of whether Buffonge was disabled. I. We recount the record which was before the claims administrator. See Liston v. Unum Corp.

Officer Severance Plan, 330 F.3d 19, 24 (1st Cir.2003) (“Where ... review is under the arbitrariness standard, the ordinary question is whether the administrator’s action on the record before him was unreasonable.”). A. Buffonge’s Medical History Buffonge began working for Wang in 1983. At that time, he became a participant in Wang’s employee benefits plan, which included a long-term disability benefits program; Wang contracted with Prudential to provide the program, and Prudential serves as the program’s claims administrator.

The plan provides that an employee is eligible for long-term disability benefits after he or she has been “totally disabled” for 26 weeks. “Total disability” exists when the following conditions are met:

(1) Due to Sickness or accidental injury, both of these are true: (

a) You are not able to perform, for wage or profit, the material and substantial duties of your occupation. (

b) After the Initial Duration of a period of Total Disability, you are not able to perform for wage or profit the material and substantial duties of any job for which you are reasonably fitted by your education, training or experience ....

(2) You are not working at any job for wage or profit.

(3) You are under the regular care of a Doctor. In 1993, Buffonge was a “field logistics coordinator” at Wang. The job involved helping track and ship computer parts to different offices across the country; it required some lifting as well as computer and other desk work. In June 1993, Buf-fonge injured his neck and back while moving computer parts at the office. His discomfort worsened as the weeks passed, and, by July, Buffonge was reporting severe pain in his chest, radiating into his left arm and elbow. On July 28, 1993, examining physician Dr.

Wendell Pierce diagnosed “cervical disc disease with left radiculitis” and wrote that Buffonge was “in a great degree of discomfort.” Buffonge took leave from work on temporary disability from August 1993 until April 1994. He then returned to work for one week, but suffered pain while driving to and from the office, and so took leave again until early November 1994. Buf-fonge was treated repeatedly during this period for continuing pain in his neck, back, and chest. In August 1994, Dr. Stephen Lipson examined Buffonge on referral from Dr. James Bayley, one of Buf-fonge’s regular physicians. Dr.

Lipson diagnosed Buffonge as suffering from at least one and possibly two herniated discs; he described Buffonge’s pain as “worse with sitting, standing, and lifting, better with walking and better with bed-rest.” Buffonge also saw Dr. David Duhme, a specialist in internal medicine at Harvard University Health Services, who had served as Buffonge’s main treating physician since soon after his initial injury. In a report dated July 20, 1994, Dr. Duhme stated that he “ha[d] been seeing [Buf-fonge] frequently” and had referred him for orthopedic evaluation.

He stated that Buffonge was “frustrated at his repeated failures to return to work without aggravating the pain.” Buffonge’s personnel records show that he had been placed on modified duty perhaps as early as the summer of 1993: thereafter, while he still had the mostly desk-bound duties of a “field logistics coordinator,” those duties were altered to avoid regular lifting and other physical labor. Despite these accommodations, Buffonge’s November 1994 return to Wang lasted less than one month. On December 5, 1994, Buffonge visited Dr. Duhme for a followup examination. Dr.

Duhme’s report stated that Buffonge “tried to return to work but now is out of work again due to neck pain radiating to the [left] arm with numbness in the [left] fingers.... [H]e would stand all day at work because the pain was worse sitting.” He diagnosed Buffonge as suffering from both cervical and lumbar disc disease. In January 1995, Buffonge attempted another return to work, but his attendance during the months that followed was sporadic at best: he was unable to work a full week, and even when he was at work he reported an intermittent inability to concentrate due to episodes of severe pain.

In September 1995, Buffonge was examined by Dr. John F. Duff at the request of Wang’s lawyers; Duff reported that Buf-fonge “has headaches ... and he says that he has trouble driving and by the end of the day, his low back gets sore and he moves from different positions but apparently continues to have the complaints.” Buffonge also told Dr. Duff that he “has chest pains when he sits too long and now gets low back pain.” Dr. Duff concluded that despite his pain, Buffonge should have been able to perform a full day’s work so long as he shifted positions during the work day and avoided heavy lifting.

By October 6, 1995, when he saw Dr. Duhme again, Buffonge was reporting serious pain, especially late in the day and late in the week. Dr. Duhme examined him that day; in the diagnosis

section of his report he again wrote “cervical disc disease” and “lumbar disc disease,” and this time added that the latter was “becoming chronic pain.” Dr. Duhme concluded that Buffonge needed to take some time off of work. Later in October, Buf-fonge tried to go to work once again, but he experienced such intense pain that he “had to get down on his hands and knees” to control it.

He visited Duhme’s office that afternoon; Duhme examined Buf-fonge and observed in his report that “[w]hen he straightens up [from a bent-over position] his whole right side of the back is bulging with muscle spasm.” He diagnosed “[e]xacerbation of lumbar inter-vertebral disc disease with objective signs of pain and spasm.” On November 9, 1995, Buffonge again visited Dr. Duhme, who reported that Buf-fonge “[ujsually ... goes into work, tries to work, and may have to stop and lie down. He does not think he can carry on.” Dr.

Duhme assessed him as suffering from “[i]nereasing pain from cervical disc disease” and gave him a letter authorizing “a medical leave of absence for a few weeks.” It is not clear precisely when Buffonge’s leave began; however, the record indicates that he missed work for five consecutive weeks in December 1995 and January 1996 while receiving treatment. On January 18, 1996, a report by Dr. Frederick Mansfield' — a spinal surgeon who examined Buffonge on referral from Dr. Duhme — noted that Buffonge had “developed worse pain in his head and neck” two months earlier. Dr. Mansfield’s report diagnosed.

Buffonge with “[cjervical and lumbar disc degeneration” and stated that “[a]ny exercise or increased bending or twisting increases his pain.” Buffonge returned to work yet again in late January 1996. He had been attending physical therapy for some time as part of his treatment, and on February 5, 1996, he was discharged from the physical therapy program; the therapist wrote in his discharge statement that he was “pain-free and [symptom]-free when last seen.” On February 15, however, Buffonge visited Dr. Duhme again and reported that he had been suffering “severe pain.” Dr.

Duhme assessed Buffonge as having “chronic neck pain due to cervical disc disease” and reported that Buffonge was undergoing acupuncture and would be in for follow-up evaluations. Six days later, on February 21,1996, Buffonge left work because of the pain; it would prove to be his last day on the job. On June 5, 1996, Buffonge applied to Prudential for long-term disability benefits. In the “Attending Physician”

section of the application, Dr. Duhme reported his diagnosis that Buffonge suffered from “chronic neck, shoulder and chest pain aggravated by work,” and that “Chronic Intervertebral Disc disease” resulted in Buffonge’s “long term disability.” He reported that Buffonge had tried many times to return to work and failed because “pain returns whenever he returns to work for more than 2 wks.” Asked whether Buffonge could work part-time or in another type of job, Dr. Duhme wrote, “don’t know but probably not.” Asked' what work duties Buffonge could perform, he wrote, “none.” Thereafter, Dr.

Duhme saw Buffonge several times during the spring and early summer. He reported each time that Buf-fonge was in pain and was still disabled from working. On June 27, 1996, another doctor examined Buffonge at Prudential’s request. Dr. William Kermond found that Buffonge suffered from disc herniation and that physical examination substantiated Buffonge’s complaints of neck, arm, and chest pain. Dr. Kermond also stated that Buffonge’s injuries appeared causally related to his 1993 accident at work.

However, he concluded that Buffonge could perform a desk job with phone and computer duties, so long as he did not engage in heavy lifting or frequent overhead work. As the record shows, Buffonge had been placed in exactly such a desk job as early as 1993, and his medical history thereafter was described above. B. The Denial of Benefits and Subsequent Appeals Prudential examined Buffonge’s medical records and denied his claim for long-term disability benefits on September 20, 1996. It noted that (1) the February 1996 physical therapy discharge

summary had deemed Buffonge pain-free,

(2) Dr. Ker-mond had opined that Buffonge could perform a desk job, and

(3) Wang had made physician-requested accommodations for Buffonge, including placing him on light duty and permitting him to get up and walk around every hour. It therefore concluded that there was “no objective medical evidence in our file to support Total Disability.” Buffonge’s employment was terminated on October 21, 1996. A month later, he administratively appealed Prudential’s benefits decision. Prudential upheld its denial in a letter dated November 26,1996. It stated that Buffonge did not qualify for long-term disability because (1) according to Dr.

Kermond, he could perform a job involving desk work or light lifting as of June 1996, and therefore (2) he had not been “totally disabled” for 26 weeks after he left work in February 1996, as required under the plan’s definition of “total disability.” Buffonge again appealed, and Prudential affirmed its denial again, on the same grounds, in a letter dated June 9, 1997. On July 25, 1997, Buffonge filed his final administrative appeal. This time he submitted additional reports and letters from two physicians — his primary physician, Dr. Duhme, and another examining physician, Dr.

Emilio Jacques — in support of his claim. Dr. Jacques, an orthopedist, examined Buffonge on April 9, June 11, and July 9, 1997, and concluded each time that Buffonge was fully disabled. His report of June 11, submitted with Buffonge’s final appeal, stated that Buffonge had told him “his pain is constant and ... he is only able to maintain activity for one to two hours.” Dr. Jacques concluded that Buf-fonge was “disabled and unable to resume his occupation” and that the disability was causally related to his 1993 work accident. A follow-up examination by Dr.

Jacques, conducted on August 27, 1997, produced a report that Buffonge submitted as a supplement to the appeal. In the August 27 report Dr. Jacques detailed his findings on examination and then reiterated the conclusion he had reached in all his earlier reports, stating: “[I]n my opinion [Buf-fonge] is disabled from any gainful employment at the present time.” He also added: In my opinion he is disabled from repetitive bending, lifting, pushing, pulling and carrying over fifteen to twenty pounds and he is disabled from any repetitive squatting and crawling positions. The letter from Dr.

Duhme, dated July 16, 1997, stated the doctor’s conclusion, based on numerous examinations of Buf-fonge, that Buffonge was “completely disabled by chronic neck pain due to cervical disc disease and low back pain due to lumbar disc disease.” He wrote that Buf-fonge was “unable to perform any job that I can think of,” that “[ejither sustained activity or sustained sitting will aggravate these pains,” and that Buffonge was “unable to devote his attentions to any work.” As part of its handling of Buffonge’s appeal, Prudential hired Dr. Jonathan Rut-chik as a consultant.

Rutchik spent a total of three hours reviewing Buffonge’s records and writing a report; he did not examine Buffonge personally. In his report, dated October 31, 1997, Dr. Rutchik wrote that “a consensus exists about the fact that Mr. Boffonge [sic] can perform work tasks such as sedentary duties and keyboard work.” He therefore concluded that Buffonge could have performed the largely desk-bound duties of his job as a field logistics coordinator at Wang. Dr. Rutchik reached this conclusion by parsing and in part disregarding the findings of physicians who had examined Buf-fonge. For example, he quoted Dr.

Jacques’ assessment that Buffonge was “disabled from repetitive bending, lifting, pushing, pulling and carrying over fifteen to twenty pounds” and followed it with this comment: “Dr. Jacques was specific about the tasks that Mr. Buffonge could not do and thus inferred that there were jobs he could do.” He thus ignored Dr. Jacques’ statement in the same report that Buf-fonge was “disabled from any gainful employment at the present time.” As to Dr.

Duhme’s conclusion, based on multiple physical examinations and years of handling Buffonge’s case, that “[e]ither sustained activity or sustained sitting will aggravate these pains,” Dr. Rutehik wrote: “[t]his letter provided no objective evidence that he has a condition where he could not do sustained sitting or sustained activity.” He did not address Dr. Duhme’s statements in the same letter that Buf-fonge was “unable to perform any job that I can think of’ and “unable to devote his attentions to any work.” Prudential subsequently denied Buf-fonge’s appeal in a letter dated November 25,1997.

It stated: [Biased on a review of [Buffonge’s] medical file, the restrictions provided by Dr. Kermond on June 26, 1996 and by Dr. Jacques on August 27, 1997 and the information provided by Dr. Rutehik, we do not find that Mr. Buffonge’s symptoms impaired him from performing the duties of his occupation for the period February 21, 1996 through August 20, 1996. Therefore, we are upholding our decision to deny Mr. Buffonge’s claim.... Prudential thus identified four bases for its denial:

(1) Buffonge’s full medical file, (2) the Kermond report, (3) the August 27 Jacques report, and (4) the Rutehik report. II. Buffonge filed suit on November 21, 2003 for review of Prudential’s denial pursuant to the Employee Retirement Income Security Act (ERISA), 29 U.S.C. § 1001 et seq. Specifically, Buffonge sought relief under

Section 502, ERISA’s civil enforcement provision, which permits a participant in an ERISA-covered benefit plan to sue “to recover benefits due” and “to enforce .. v rights” under the plan. 29 U.S.C. § 1132(a)(1)(B). Buffonge argued that Wang and Prudential violated the plan’s terms and wrongfully denied him long-term disability coverage; he sought a retroactive award of benefits. The parties agreed that the Wang benefits plan confers discretionary authority to the administrator to determine eligibility for benefits.

As the district court noted, that meant Prudential’s decision to deny Buffonge benefits was subject to deference, to be reversed only if “arbitrary, capricious, or an abuse of discretion.” Gannon v. Metro. Life Ins. Co., 360 F.3d 211, 212-13 (1st Cir.2004) (internal quotation marks omitted). Under the “arbitrary and capricious” framework, as applied to ERISA review, a decision will be upheld “if it is reasoned and supported by substantial evidence.” Id. at 213. On October 20, 2004, Wang and Prudential moved for

summary judgment, arguing that their decision to deny benefits was supported by substantial evidence and therefore was not arbitrary or capricious. Buffonge filed an opposition and cross-motion for

summary judgment, arguing that Dr. Rutchik’s report was skewed, that Prudential relied on it in denying him benefits, and that as a result Prudential did not give adequate weight to the conclusions of his treating physicians. He did not argue that the conclusions of his treating physicians deserved special status— nor could he have, given the Supreme Court’s holding that “courts have no warrant to require administrators automatically to accord special weight to the opinions of a claimant’s physician.” Black & Decker Disability Plan v. Nord, 538 U.S. 822, 834, 123 S.Ct. 1965, 155 L.Ed.2d 1034 (2003). Applying the “arbitrary and capricious” standard, the district court granted

summary judgment for Wang and Prudential. The court began by finding that, contrary to Buffonge’s contention, Dr. Rutchik’s report did not “ignore” or “intentionally misinterpret” the conclusions of Dr. Duhme and Dr. Jacques. This was so for two reasons. First, the court noted, Dr. Rut-chik’s report contained multiple references to and repeated discussion of the other doctors’ reports. Second, the court in essence agreed with Dr. Rutchik’s parsing of the Jacques report.

Since the Rutchik report was valid, the court reasoned, Prudential had been faced with a garden-variety conflicting-evidence situation, and the case therefore fit within the rule that the existence of medical evidence pointing in two directions does not render arbitrary or capricious a plan administrator’s decision to credit one viewpoint or the other. See, e.g., Leahy v. Raytheon Co., 315 F.3d 11, 19 (1st Cir.2002); Vlass v. Raytheon Employees Disability Trust, 244 F.3d 27, 32 (1st Cir.2001); Gannon, 360 F.3d at 213. From this grant of

summary judgment Buffonge timely appealed. III. Our review of the district court’s grant of

summary judgment on the administrative record is de novo. Glista v. Unum Life Ins. Co. of Am., 378 F.3d 113, 125 (1st Cir.2004). We, like the district court, must defer to the claims administrator’s benefits decision, disturbing it only if it was “arbitrary, capricious, or an abuse of discretion.” Gannon, 360 F.3d at 212-13. Buffonge’s argument focuses most heavily on the Rutchik report. Buffonge argues that Dr. Rutchik distorted and misrepresented the medical record, discounted conclusions by Dr. Duhme and Dr.

Jacques that Buffonge was fully disabled, and falsely claimed a “consensus” that Buffonge could perform sedentary work. He argues that Prudential was intimately familiar with Buffonge’s medical records, and it therefore had to know that Dr. Rutchik was misreading the Jacques and Duhme reports. He says Prudential’s heavy reliance on the Rutchik report, combined with its citation of the Jacques report as support for its decision when the report clearly cut the other way, renders the decision arbitrary and capricious. We agree. A.

The Evidence As previously noted, Prudential cited four pieces of evidence in support of its final decision:

(1) Buffonge’s medical file as a whole, (2) the Kermond report, (3) the August 27 Jacques report, and (4) the Rut-chik report. Of these four, at least two— the Rutchik report and the Jacques report — do not provide reasoned support for Prudential’s decision. We examine them in turn. 1. The Rutchik report Dr. Rutchik’s report suffers from fundamental flaws. As an initial matter, Dr. Rutchik’s conclusion that “a consensus exists” that Buffonge could perform a desk job is far from accurate. Buffonge’s medical record as of October 1997, the date of Rutchik’s report, contained at least three recent reports from Dr.

Jacques concluding just the opposite. Dr. Rutchik also had before him the reports of Dr. Duhme, Buffonge’s longtime treating physician, who had repeatedly examined Buffonge and deemed him fully disabled, and who had stated that very summer that Buf-fonge could not engage in “sustained sitting” or “devote his attentions to any work.” The Duhme reports alone preclude Dr. Rutchik’s finding of “a consensus” that Buffonge could perform desk work. Similarly, his conclusion that there was no “objective” evidence to support Dr. Duhme’s conclusion is contradicted by Dr.

Duhme’s personal observations on examination of Buffonge. Dr. Rutchik’s parsing of Dr. Jacques’ August 27 report is also, we think, not a fair reading. The Jacques report’s explicit statement that Buffonge was “disabled from any gainful employment ” (emphasis added) does not lend itself to the inference that “there were jobs that [Buffonge] could do.” This is especially true considering that Dr. Jacques had filed at least three other reports concluding that Buffonge was fully disabled. The flaws in the Rutchik report are crucial to our analysis for several reasons.

The report’s mischaracterizations are material, and so it does not provide support for Prudential’s decision. Further, the fact that Prudential relied on the Rutchik report at all brings into question the integrity of Prudential’s decision-making process in this case. Prudential had before it all of Buffonge’s records, and it had to be aware that the Rutchik report misconstrued the conclusions of Buffonge’s doctors. 2. The August 27 Jacques report Prudential also relied in part on “the restrictions provided by ... Dr. Jacques on August 27, 1997.” But Dr.

Jacques’ listing of physical activities Buffonge could not perform did not create the inference ■that any activity not on the list was one Buffonge could perform. Dr. Jacques began the crucial passage by stating that Buffonge was “disabled from any gainful employment at the present time.” In light of that unequivocal statement, the specifics that follow — disabled from bending, lifting, etc. — appear to us illustrative. It would make no sense for Dr.

Jacques to have stated that Buffonge was “disabled from any gainful employment” and then in the very next sentence to have implied that there was in fact gainful employment he could perform. Dr. Jacques had also deemed Buffonge fully disabled in all of his other reports; Prudential’s reading of the exemplars as refuting Dr. Jacques’ consistent conclusion that Buffonge was fully disabled is simply unreasonable. The August 27 report therefore does not support Prudential’s conclusion that Buffonge could perform sedentary work. B.

Analysis In sum, we find that two of Prudential’s four cited pieces of evidence fail to provide reasoned support for its conclusion; we also find that Prudential’s willingness to rely on a report it knew or should have known to be misleading, and its mis-eharacterization of the conclusions of Dr. Jacques, raises concerns about the fairness of its decision-making process. The question is whether these flaws render the decision “arbitrary, capricious, or an abuse of discretion.” Gannon, 360 F.3d at 213. We find that they do.

An administrator’s decision must be “reasoned” to survive “arbitrary and capricious” review, Gannon, 360 F.3d at 213, and we cannot say that a decision relying on multiple pieces of faulty evidence was “reasoned.” See id. at 214-15 (finding that a claims administrator “reasonably relied” on evidence only after confirming that the evidence was “credible” and “reliable”). Further, in giving content to “arbitrary and capricious” review in this context, we find it useful to consider other ERISA provisions. Cf. Global Naps, Inc. v.

Verizon New England, Inc., 396 F.3d 16, 24 (1st Cir.2005) (noting that statutory subsections must be read in light of structure and intent of the larger statute).

Section 503 of ERISA, for example, states in relevant part that “every employee benefit plan shall ... afford a reasonable opportunity to any participant whose claim for benefits has been denied for a full and fair review by the appropriate named fiduciary of the decision denying the claim.” 29 U.S.C. § 1133 (emphasis added). While this language refers to the benefit plan itself (and to procedural safeguards), it also has some substantive content.

One oft-quoted decision has held that the “full and fair review” verbiage “must be construed ... to protect a plan participant from arbitrary or unprincipled decision-making.” Grossmuller v. Int’l Union, UAW, Local 813, 715 F.2d 853, 857 (3d Cir.1983). “Unprincipled decision-making” of the sort forbidden by

Section 503 is also “arbitrary” (i.e. not “reasoned”) within the meaning of our review under

Section 502. This does not mean that every perceived unfairness rises to the level of arbitrariness. We address Wang and Prudential’s two main arguments to the contrary. First, they argue, the decision was based on substantial evidence, and must be upheld, because Prudential considered not just the Rutchik report, but Buffonge’s entire medical record, some of which supports Prudential’s decision. This argument ignores the requirement that the decision must not be arbitrary.

Given the flaws present here, Prudential and Wang’s argument must boil down to a claim that an administrator’s arbitrary analysis should be ignored if the end result — denial of benefits — can nonetheless be saved by a single valid piece of evidence that supports it. The argument would be stronger if the evidence here compelled or virtually compelled a conclusion that Buffonge was not disabled. That is not the case; Buffonge’s claim of disability is a serious one, and he has been prejudiced by the process. We do not know what the administrator would have determined if the process had been less flawed.

Second, Wang and Prudential reiterate the district court’s analysis, citing Vlass for the proposition that “[t]he existence of contradictory evidence does not, in itself, make the administrator’s decision arbitrary.” 244 F.3d at 30. The point is true but not pertinent here. IV. Having concluded that Prudential’s decision fails “arbitrary and capricious” review, we are left to consider the remedy. There is no question that this court has the power to remand to the claims administrator; it also has the power, in appropriate cases, to award benefits to the disability claimant. See Cook v.

Liberty Life Assurance Co., 320 F.3d 11, 24 (1st Cir.2003) (“Once a court finds that an administrator has acted arbitrarily and capriciously ... the court can either remand the case to the administrator for a renewed evaluation of the claimant’s case, or it can award a retroactive reinstatement of benefits.”). The question, then, is which of these remedies is more appropriate here. Some courts addressing similar fact patterns have leaned toward categorical rules. See, e.g., Quinn v.

Blue Cross & Blue Shield Ass’n, 161 F.3d 472, 477 (7th Cir.1998) (holding that retroactive award is usually proper when claimant had benefits and lost them because of arbitrary conduct and/or when no evidence supported denial, while remand is appropriate where decision-maker “fails to make adequate findings or fails to provide an adequate reasoning”).

This court, while acknowledging that framework, see Cook, 320 F.3d at 24, has taken a more flexible approach: it has held that “the variety of situations is so great” in ERISA review that the court must have “considerable discretion” to craft a remedy after finding a mistake in the denial of benefits. Id.; see also Glista, 378 F.3d at 131-32 (concluding that where the administrator attempts to rely on a new rationale on appeal, the court’s response and the remedy should be resolved on a case-by-case basis, in light of the facts and equities).

Here, the problem is not that Buffonge was denied benefits to which he was clearly entitled; the evidence does not compel such an outcome. The problem is with the integrity of Prudential’s decision-making process. The appropriate response is to let Buffonge have the benefit of an untainted process.

See Quinn, 161 F.3d at 478 (finding that where administrator’s decision was faulty but it was not “clear-cut that it was unreasonable” to deny benefits, remand was appropriate because a benefits award “might provide [claimant] with an economic windfall should she be determined not disabled upon a proper reconsideration”); cf. DiGregorio v. Hartford Comprehensive Employee Benefit Serv. Co., 423 F.3d 6, 10 (1st Cir.2005) (stating that in an ERISA

Section 503 proceeding, remand is appropriate where administrator’s procedural failings prejudiced claimant). V. Under the facts of this case, we reverse the grant of

summary judgment to Prudential and Wang. We remand to the district court with instructions to enter judgment for Buffonge on his claim that the denial of benefits was arbitrary and, as the appropriate remedy, to remand the case to the plan administrator for proceedings consistent with this opinion. In our view, that includes the plan administrator taking new evidence should any party wish to submit the same. Costs are awarded to Buffonge. . At that time, the employer was Wang Laboratories Inc. Getronies Wang Co., LLC has since acquired the business in relevant part.

For the sake of simplicity, we refer to the employer as "Wang.” . Several of Prudential's letters denying Buf-fonge benefits stated that he had been placed on "light duty” after his initial injury; one stated that he was "answering the telephones” at a desk job. These accommodations are confirmed in an August 1993 report by Dr.

Duhme stating that Buffonge had told him he "now can do [his job] without any lifting.” Buffonge’s job involved "occasional bending, squatting, or reaching” and "very infrequent lifting of packages that are less than 10 pounds,” according to the 1997 letter in which Prudential denied Buffonge's final administrative appeal. . This document is dated May 16, 1996, not June 5.-However, from the context it appears to be part of Buffonge's application for long-term benefits, filled out by Dr. Duhme days in advance of the application’s submission. . The record is quite murky as to how and why this occurred.

The only available information (besides the fact of the termination) is contained in an April 16, 1996 report by Dr. Duhme, which states: "[Buffonge’s] job at work has been eliminated. He is now out of work on short term disability, but when he returns to work, he will be given 30 days to find a new job in the company or be terminated....’’ We infer that after months passed and Buffonge did not return to seek a new position, the company terminated his employment. Whatever the details of the termination, however, they do not affect our disposition of the case. . Dr.

Jacques used phrases ranging from "disabled" to "totally and permanently disabled.” . It is unclear from the record whether Dr. Duhme examined Buffonge on July 16 or, if not, how recently he had last examined him. It is clear, however, that he was Buffonge's treating physician and had examined him on numerous occasions from 1993 through 1996. . Buffonge did not identify in his complaint which subsection of

Section 502 he was relying on. However, given his request for retroactive benefits, subsection (a)(1)(

B) is most directly applicable. . At oral argument he requested a remand to the plan administrator as an alternative remedy- . It wrote: "The obvious implication of [Dr. Jacques' statement listing activities Buffonge could not perform] is that plaintiff could perform work so long as it did not involve those specific, prohibited activities. Plaintiff attempts to diffuse Dr.

Jacques' statement by arguing that it was superceded by the Doctor's eventual conclusion that Buffonge is 'disabled’ but that construction is unpersuasive . because placing emphasis upon only the latter statement would transform the former, more specific, statement into meaningless surplus-age.” . "[I]n an ERISA case where review is based only on the administrative record before the plan administrator and is an ultimate conclusion as to disability to be drawn from the facts,

summary judgment is simply a vehicle for deciding the issue.” Orndorf v. Paul Revere Life Ins. Co., 404 F.3d 510, 517 (1st Cir.2005). ."This means the non-moving party is not entitled to the usual inferences in its favor.” Id. .This same standard has been described in other language. This court has said, for example, that to pass "arbitrary and capricious” review in the ERISA context a decision must be "reasonable and supported by substantial evidence.” Glista, 378 F.3d at 126.

It also has said an administrator’s decision is arbitrary and capricious if "the insurer’s eligibility determination was unreasonable in light of the information available to it,” Pari-Fasano v. ITT Hartford Life & Accident Ins. Co., 230 F.3d 415, 419 (1st Cir.2000), and that "the arbitrary and capricious standard is functionally equivalent to the abuse of discretion standard,” Wright v. R.R. Donnelley & Sons Co. Group Benefits Plan, 402 F.3d 67, 74 n. 3 (1st Cir.2005). . The district court agreed with Prudential’s reading of the Jacques report.

However, we owe no deference to the district court in this setting, see Glista, 378 F.3d at 125, and we simply disagree with its analysis. In describing tire Jacques report, the district court wrote that Dr. Jacques first listed Buffonge's specific disabilities and then wrote that he was "disabled.” In that context, it wrote, crediting only the latter statement "would transform the former, more specific, statement into meaningless surplusage.” But in fact, Dr.

Jacques wrote not simply that Jacques was "disabled,” but "disabled from any gainful employment." Further, he made this statement first and followed it with a list of specifics, not the other way around. The passage is most reasonably read as a categorical statement of total disability followed by a list of examples. . As the Seventh Circuit said in a similar context, "[w]e emphasize ... that we neither determine whether [Buffonge] is disabled, nor whether [Prudential's] decision was incorrect, rather only that [Prudential] denied [Buffonge’s] benefits in an arbitrary and capricious manner.” Quinn v.

Blue Cross & Blue Shield Ass'n, 161 F.3d 472, 476 (7th Cir.1998). . This is true despite the fact that 29 U.S.C. § 1132(a)(1)(B), the subsection most directly relevant to Buffonge's claim, does not explicitly authorize administrative remand as a remedy. Numerous decisions by this court and others have ordered, or approved in theory, remand to the administrator where arbitrariness is found in the course of 29 U.S.C. § 1132(a)(1)(

B) review, see, e.g., Cook, 320 F.3d at 24; Quinn, 161 F.3d at 477, and we have seen none holding that remand is impermissible. .We cannot say on this record that Prudential intentionally set up a biased process. In any event ERISA is not a punitive statute. See Hotz v. Blue Cross & Blue Shield of Mass., 292 F.3d 57, 61 (1st Cir.2002) (stating that ERISA does not permit punitive damages for benefit claims (citing Pilot Life Ins. Co. v. Dedeaux, 481 U.S. 41, 53-54, 107 S.Ct. 1549, 95 L.Ed.2d 39 (1987), abrogated in part on other grounds by Ky. Ass’n of Health Plans, Inc. v. Miller, 538 U.S. 329, 123 S.Ct. 1471, 155 L.Ed.2d 468 (2003))).