SUMMARY ORDER
Plaintiff-Appellant Richard B. Durgin (“Durgin”) appeals from a September 24, 2008, 2008 WL 4416285, judgment of the United States District Court for the District of Vermont (Murtha, J.), which granted the motion of Defendant-Appellee *539Blue Cross and Blue Shield of Vermont (“BCBS”) for summary judgment. Dur-gin’s claim, brought pursuant to the Employee Retirement Income Security Act of 1974, 29 U.S.C. § 1132(a)(1)(B), arises from the denial by BCBS of coverage for a “standing component” on his motorized wheelchair under BCBS’s “Vermont Freedom Plan” (the “Plan”). We assume the parties’ familiarity with the facts of the case, its procedural history, and the scope of the issues on appeal.
We review the decision of the District Court to grant summary judgment de novo. McCauley v. First Unum Life Ins. Co., 551 F.3d 126, 130 (2d Cir.2008). Because the Plan gives BCBS discretionary authority to determine eligibility benefits, we will overturn BCBS’s decision only if it was “arbitrary and capricious,” id. at 132 — that is, if it was “without reason, unsupported by substantial evidence or erroneous as a matter of law.” Kinstler v. First Reliance Standard Life Ins. Co., 181 F.3d 243, 249 (2d Cir.1999) (internal quotation marks omitted). While a plan administrator need not “accord special weight to the opinions of a claimant’s physician,” it also “may not arbitrarily refuse to credit a claimant’s reliable evidence, including the opinions of a treating physician.” Black & Decker Disability Plan v. Nord, 538 U.S. 822, 834, 123 S.Ct. 1965, 155 L.Ed.2d 1034 (2003); see also Demirovic v. Bldg. Serv. 32 B-J Pension Fund, 467 F.3d 208, 212 (2d Cir.2006). If BCBS “impose[d] a standard not required by the plan’s provisions, or interpreted] the plan in a manner inconsistent with its plain words, its actions may well be found to be arbitrary and capricious.” McCauley, 551 F.3d at 133 (internal quotation marks omitted).
In denying Durgin’s claim for coverage, BCBS relied on two grounds: (a) that “[t]here are no peer reviewed clinically controlled studies available showing that the stander improves net health outcomes,” and (b) that there is “no evidence showing the benefit of the standing feature and that it will help restore or maintain [Durgin’s] health.” J.A. 76 (emphasis added).1 Neither of these grounds can survive review.
BCBS first stated that Durgin did not show that the standing component was “medically necessary” because there were no “peer reviewed clinically controlled studies ” showing “improvefd] net health outcomes.” Id. (emphasis added). But the Plan does not contain any requirement that a service be supported by “peer reviewed clinically controlled studies” before BCBS will provide coverage, and such a requirement is impossible to square with the lower standard that the Plan establishes for “Medical and Scientific Evidence.” BCBS’s atextual requirement therefore “impose[d] a standard not required by the plan’s provisions,” McCau-ley, 551 F.3d at 133 (internal quotation marks omitted), and accordingly was arbitrary and capricious.
BCBS’s second ground is factually incorrect: Durgin put forward ten articles from medical journals providing varying degrees of support for the medical benefits of the *540standing component. He also proffered his treating physician’s statement that the standing component had led to “a marked decrease in spasticity, as well as an overall improvement in maintaining his skin integrity,” had “very positively impacted his history of decubitus ulcers,” and had “helped [him] maintain bone density and has prevented osteoporosis.” J.A. 68, 74. While this evidence might ultimately be deemed inadequate to require BCBS to insure the standing component (a question we need not and do not reach), it cannot be said that “no evidence” showed the medical benefits that Durgin alleged. J.A. 76 (emphasis added). BCBS’s second ground thus “arbitrarily refuse[d] to credit a claimant’s reliable evidence,” Black & Decker, 538 U.S. at 834, 123 S.Ct. 1965, and cannot support BCBS’s denial of the claim.
Alternatively, BCBS sought to write off the various medical articles submitted by Durgin by asserting that they were not published in peer-reviewed journals and that the standing component was accordingly “experimental” or “investigational” under the Plan.2 This too was factually erroneous; as post-argument briefing revealed, a significant number of the articles Durgin proffered did in fact appear in peer-reviewed journals.
When we determine a denial of benefits to be arbitrary and capricious, our usual practice is to “remand to the [administrator] with instructions to consider additional evidence unless no new evidence could produce a reasonable conclusion permitting denial of the claim or remand would otherwise be a ‘useless formality.’ ” Miller v. United Welfare Fund, 72 F.3d 1066, 1071 (2d Cir.1995). Whether the articles and other evidence proffered by Durgin are sufficient to bring the standing option within generally accepted practice parameters, when the Plan is construed in a reasonable manner, is a question to be considered in the first instance by BCBS.3 *541Accordingly, we VACATE the judgment of the District Court and REMAND the case for further proceedings consistent with this order.
. In its initial denial of Durgin’s request and in response to his first administrative appeal, BCBS stated that the standing option was “intended primarily for convenience or comfort beyond what is necessary to meet the member’s legitimate medical needs.” J.A. 31. BCBS does not advance this argument on appeal, and has thereby waived it. See, e.g., Design Strategy, Inc. v. Davis, 469 F.3d 284, 300 (2d Cir.2006). In any event, BCBS provided no explanation of how the standing option fit into the “convenience or comfort" category, which appears to be a poor fit for the standing component. See J.A. 58 (identifying, as examples of such equipment, “decorative items, unique materials (e.g. magnesium wheelchair wheels, lights, extra batteries, etc.)”).
. Durgin contends that this argument should not be considered in the federal courts because it was not relied upon during the claims process. Because Appellees' rationale fails even if it was adequately stated, however, we need not consider this issue.
. In its post-argument letter brief, BCBS argues that the content of the articles is insufficient to establish medical necessity. Because this claim was raised at such a late date in the proceeding, and is not based on the administrative record, we will not consider it. See Elec. Inspectors, Inc. v. Vill. of E. Hills, 320 F.3d 110, 123 n. 8 (2d Cir.2003) (argument not raised until supplemental brief abandoned); Miller, 72 F.3d at 1071 ("[D]is-trict courts are required to limit their review to the administrative record ....”); cf. Univ. Hosps. v. Emerson Elec. Co., 202 F.3d 839, 848-49 n. 7 (6th Cir.2000) ("[I]t strikes us as problematic to, on one hand, recognize an administrator’s discretion to interpret a plan by applying a deferential 'arbitrary and capricious' standard of review, yet, on the other hand, allow the administrator to ‘shore up’ a decision after-the-fact by testifying as to the ‘true’ basis for the decision after the matter is in litigation, possible deficiencies in the decision are identified, and an attorney is consulted to defend the decision by developing creative post hoc arguments that can survive deferential review.”).
Nor does this argument render remand futile. It is by no means clear that the criticisms that BCBS levels against the substance of the six peer-reviewed articles even support BCBS’s conclusion, let alone justify a denial of coverage. BCBS acknowledges that one of the articles "recommends standing devices for persons with spinal cord injury,” J.A. 99, which if anything appears to strengthen Dur-gin’s contention that the standing option is "consistent with generally accepted practice parameters,” as the Plan requires, J.A. 115. Its primary complaint regarding four other articles is that they lack clinical evidence. But, for the reasons given above, this is insufficient to support denial of the claim. As to the final peer-reviewed article, BCBS merely notes that it concludes that controlled studies are needed, and says nothing about any other recommendations or conclusions that that article makes.
*541Nothing in this order generally, or in this footnote in particular, should be read as precluding Durgin from presenting additional evidence to BCBS on remand if indeed additional relevant evidence has come into being since this claim was last before BCBS’s internal review panel.