dissenting:
I respectfully dissent. In my view, the district court recognized and applied the settled principles that govern this controversy, and was proceeding with admirable dispatch and thoroughness to decide the case. The majority has preempted the district court’s exercise of discretion, impeached without record support the good faith of the plan administrator, excluded from consideration virtually all the record facts that bear upon the question of medical coverage, and decided the case in a way that illustrates why judicial deference is properly owed to doctors and plan administrators in matters of medical insurance.
I write at some length in order to demonstrate that although the result in this case is wrong, the governing principles are left intact. Indeed, this case shows what happens when the patterns of deference mandated by precedent are suspended, even for supposedly unique facts and circumstances, and medical decisions are made by judges who (I say with respect) have no more medical expertise than I do. The result is that the majority deems arbitrary and capricious a refusal to pay for a treatment the supposed efficacy of which rests on a study acknowledged to be a scientific fraud.
A
Nickolas Zervos, who suffers from metastatic (Stage IV) breast cancer, seeks coverage for a medical treatment that involves a single cycle of high-dose chemotherapy followed by an autologous blood stem cell transplant (“HDCT”). The largest and most well-respected Phase III1 randomized clinical study comparing HDCT with conventional-dose chemotherapy for the treatment of metastatic (Stage IV) breast cancer determined that (1) HDCT provides no survival benefit over conventional chemotherapy, (2) HDCT actually produces a survival disadvantage for patients (such as Mr. Zervos) who are over age 42, and (3) HDCT is significantly more toxic than conventional chemotherapy. See Edward A. Stadtmauer et al., Conventional-Dose Chemotherapy Compared with High-Dose Chemotherapy Plus Autolo-gous Hematopoietic Stem-Cell Transplantation for Metastatic Breast Cancer, 342 New Eng. J. Med. 1069 (2000) (reporting the results of the National Cancer Institute-sponsored study) (the “Stadtmauer study”). Multiple other Phase III (i.e., randomized) studies have confirmed these finding.
As the majority opinion concedes (mar jority op. at 640, 641, n. 5), the only Phase III study to find a survival benefit for HDCT — the “Bezwoda” study — has been exposed as a fraud.
In March 2000 the American Society of Clinical Oncology warned: “Given the lack of persuasive data demonstrating superior effectiveness of this very toxic therapy, ... [HDCT] for breast cancer should only be performed in the context of a high quality clinical trial.” See Highr-Dose Chemotherapy with Bone Marrow Transplant for Breast Cancer Patients, at http://www.asco.org/people/nr/html/genpr/ m — 0300bmtposition.htm (March 2000) (emphasis added). ECRI, a non-profit health services research institution that studies healthcare technology and healthcare risk, has similarly cautioned that use of HDCT should be confined to National Cancer Institute-sponsored randomized (Phase III) controlled clinical trials. Medicare no longer provides coverage of *650HDCT for the treatment of breast cancer. According to testimony at trial, the treatment is no longer offered at Memorial Sloan Kettering Hospital and Mount Sinai Medical Center in New York and the Dana-Farber Cancer Institute and Brigham & Women’s Hospital in Boston — four of the world’s most prestigious cancer institutes.
The Columbia Journalism Review recently cited favorable news coverage about HDCT as a horrible example of how lazy reporters can be gulled by pseudo-science and quackery:
Throughout the 1990s, some 30,000 women underwent [HDCT].... There was no proof that the $50,000-to-$100,000 procedure arrested the disease. In fact, two years ago, results from four clinical trials gave the definitive answer: it did not work.
Trudy Lieberman, Covering Medical Technology, Columbia Journalism Review (September/October 2001), at http://www.cjr.Org/year/01/5/lieberman.asp.
B
The legal principles governing this case are settled and can be stated briefly.
Where, as here, an insurance plan gives its administrator broad discretion to construe the terms of the plan and to determine whether a claimant is entitled to payment of benefits, a court may reverse the administrator’s decision only if it is arbitrary and capricious. See Firestone Tire & Rubber Co. v. Bruch, 489 U.S. 101, 115, 109 S.Ct. 948, 103 L.Ed.2d 80 (1989); Zuckerbrod v. Phoenix Mut. Life Ins. Co., 78 F.3d 46, 49 (1996). A decision is arbitrary and capricious if it is “without reason, unsupported by substantial evidence or erroneous as a matter of law.” Pagan v. NYNEX Pension Plan, 52 F.3d 438, 442 (2d Cir.1995) (internal quotations marks omitted). Substantial evidence is defined as “such evidence that a reasonable mind might accept as adequate to support the conclusion reached by the [administrator and] ... requires more than a scintilla but less than a preponderance.” Miller v. United Welfare Fund, 72 F.3d 1066, 1072 (2d Cir.1995) (internal quotations marks omitted). A court is not “not free to substitute [its] own judgment for that of the [plan administrator] as if [it] were considering the issue of eligibility anew.” Pagan, 52 F.3d at 442. While we review de novo the district court’s assessment of the plan administrator’s decision, in so doing we must review the administrator’s decision itself under the arbitrary and capricious standard. See Miller, 72 F.3d at 1070-71. And if the defect is insufficiency, the appropriate remedy is a remand:
[I]f upon review a district court concludes that the [administrator’s] decision was arbitrary and capricious, it must remand to the [administrator] with instructions to consider additional evidence unless no new evidence could produce a reasonable conclusion permitting denial of the claim....
Id. at 1071.
The district court’s findings of fact may not be set aside unless they are clearly erroneous. See Zuckerbrod, 78 F.3d at 49. We review a district court’s fashioning of relief — including the decision to remand— under an abuse of discretion standard. See Kinek v. Paramount Communications, Inc., 22 F.3d 503, 508 (2d Cir.1994). “A district court ‘abuses’ or ‘exceeds’ the discretion accorded to it when (1) its decision rests on an error of law (such as application of the wrong legal principle) or a clearly erroneous factual finding, or (2) its decision — though not necessarily the product of a legal error or a clearly erroneous factual finding — cannot be located within the range of permissible decisions.” Zervos v. Verizon New York, Inc., 252 F.3d 163, 169 (2d Cir.2001) (“Zervos II”).
*651C
Under the terms of the plan covering Zervos, Empire excludes coverage of procedures that are “Experimental/Investiga-tional,” which is defined as “services or supplies which are not of proven benefit for the diagnosis or treatment of the Covered Person’s condition, or are not generally recognized by the medical community as effective or appropriate for that condition, as determined by. the Claims Administrator.” In this case, Empire considers the relevant medical community to be “the general community of academic oncologists.” The term “proven benefit,” which is not further defined in the policy, is construed by Empire to mean that what must be proven is a positive net health effect — i.e., “the efficacy [of the treatment] balanced by the toxicity and whether that net positive health effect compares favorably with other standard treatments.”
For much of the 1990s, Empire provided coverage for HDCT as a treatment for breast cancer. At a meeting of Empire’s Technology Assessment Committee (“TAC”) in March 2000, however, Empire determined that HDCT as a treatment for breast cancer should be re-categorized as Experimental/Investigational under the terms of the policy. Empire principally based this decision on the various studies (listed above) that found [i] no survival benefit for HDCT over conventional chemotherapy and [ii] significantly greater toxicity.
Empire’s decision to reclassify HDCT as Experimental/Investigational for the treatment of breast cancer is unchallenged in this case: the majority concedes that “[t]he studies before the TAC reasonably could be read to establish that HDCT did not provide a survival benefit over long term conventional chemotherapy for patients with metastatic breast cancer ... [and] also demonstrated that ... it had more side effects and more serious side effects.” Majority Op. at 18-19.
Six months after this change in policy, Zervos applied for coverage of HDCT for the treatment of his metastatic breast cancer. Consistent with the policy it had adopted, Empire’s responded to Zervos by denying coverage: “Autologous Peripheral Stem Cell Harvest/Transplant is Experimental and Investigational in Breast Cancer Stage IV.”
At the request of Zervos’s treating physician, Empire submitted its decision to an external review. Empire refers external reviews to an unaffiliated organization, the Medical Care Ombudsmen Program (“MCOP”), which selects reviewers on a rotating basis according to the reviewer’s specialty, clinical experience, and availability. Empire plays no role in selecting the reviewer.
Empire submits to MCOP a series of questions the purpose of which is to elicit the reviewer’s expert opinion about the current state of the medical literature and about whether the treatment is generally recognized as effective or appropriate for a given patient. The reviewer is not expected to be an expert on the contract or on how the language of the contract applies. Empire then examines the answers provided by the reviewer within the context of the policy language and makes a coverage decision. Before the review process begins, the treating physician is asked to submit any literature or information he or she would like considered on the appeal.
MCOP assigned Zervos’s case to Dr. Thomas R. Spitzer, director of the bone marrow transplant program at Massachusetts General Hospital and deputy chief of its hematology-oncology unit. Asked, “Does the scientific evidence support a result of improvement in health outcome?” Dr. Spitzer answered, “No”; and after out*652lining the various studies mentioned above, he explained: “[T]he conclusions of [the Stadtmauer study] have been widely accepted by the Oncology community, and many have advocated abandonment of this therapy as ‘standard of care’ for metastatic breast cancer.” Asked, “Is this a well designed study for which the patient is a good candidate?” Dr. Spitzer answered, “Unable to respond”; and then added that, while the “proposed research protocol is a well designed phase II study evaluating [HDCT] for ... metastatic breast cancer[,][t]he protocol title refers to [HDCT] support in women. ... Thus, it is not clear whether Mr. Zervos is eligible for the study that was enclosed for my review.” (emphasis in original) Asked, “Is the recommended treatment plan the best treatment plan available for this patient at this time?” Dr. Spitzer answered, “No”; he then explained: “there are no convincing data to show that [HDCT] is a superior approach to conventional chemotherapy alone for male patients with metastatic breast cancer.”
The only answer given by Dr. Spitzer that is arguably supportive of HDCT is in response to the question “Is the technology as beneficial as any of the established alternatives?” Dr. Spitzer explained that “[HDCT] appears to be as effective as prolonged combination chemotherapy for metastatic breast cancer. A possible benefit of [HDCT] is the avoidance of a protracted course of continuous combination chemotherapy.” (The majority wholly relies on this one question and answer, ignoring all of Dr. Spitzer’s other responses.)
Empire considered Dr. Spitzer’s answers in light of its view that a “proven benefit” is a “positive net health effect” and that the relevant medical community for this case is “the general community of academic oncologists.” Empire again determined that Zervos’s claim should be denied as Experimental/Investigational, and so advised Zervos by letter dated October 24, 2000, which quoted Dr. Spitzer’s finding that “[t]here are no convincing data to show that [HDCT] is superior to conventional chemotherapy alone for males with metastatic breast cancer.”
Zervos commenced this litigation on January 29, 2001, seeking a preliminary injunction ordering Empire to cover the cost of HDCT. After the suit was filed, Verizon (Zervos’s employer and the payor on the policy) asked Empire to submit Zervos’ case for a second review. Where (as here) ERISA applies, Empire always complies when an employer makes such a request.
This time, MCOP designated Dr. Raymond Weiss, a clinical professor of medicine and breast cancer specialist at the Lombardi Cancer Center of Georgetown University. After outlining the various studies alluded to above, Dr. Weiss stated that “[m]ost oncologist are of the opinion at this time that transplant therapy ... is of unproven efficacy” and that those who remain convinced of its efficacy are “in a distinct minority among oncologists.” Moreover, Dr. Weiss observed, the clinical study in which Zervos wishes to participate is identified “as a study for women, not men. (emphasis in original) Entering a single male who is ineligible for such a study would be considered inappropriate science by most research-oriented clinicians.” Dr. Weiss’s conclusion was categorical:
There is absolutely no justification for treating such a patient under this protocol as part of a scientific study, and even more so, there is no scientific justification for treating him with such therapy at all, within or outside any clinical trial. I could perhaps be convinced of the scientific validity of such treatment if he were to be entered on a trial for which *653he is eligible, one using an innovative therapy, and one focused on stage IV disease. However, this is not the case on all three points.
Finally, asked whether the clinical study is “a well designed study for which this patient is a good candidate,” Dr. Weiss answered, “No,” and added:
[Zervos] is not eligible for a study in which women are to be treated in the manner proposed, and there is no evidence that [HDCT] provides a net benefit for women with metastatic breast cancer either. The protocol provided with the records of this patient is a phase II “catch-all” protocol ... for women with any stage of breast cancer .... There is nothing innovative about this protocol, and I do not consider it a “well designed study.”
Once again, Empire sent notice to Zer-vos informing him that his claim was denied, quoting many passages from Dr. Weiss’s report.
On October 22, 2001, the district court found that “Empire’s decision to deny coverage for HDCT for breast cancer was based on substantial evidence and is not arbitrary and capricious.” Zervos v. Verizon New York, Inc., No. 01-CIV-685, 2001 WL 1262941, at *11 (S.D.N.Y. Oct.22, 2001) (“Zervos III”). But the court concluded that Empire’s appeals process was arbitrary and capricious, principally because “none of the questions posed to the reviewers are direct questions that elicit a clear and unequivocal response regarding whether the treatment is of proven benefit or generally recognized by the medical community as effective or appropriate for the condition.” Id. at *13. Nevertheless, the district court held that,
based on the administrative record, this Court cannot find that the treatment is not experimental/investigational as defined in the contract or that there is some other medical or clinical reason that the patient should receive this treatment. This is still a medical determination to be made by a full and fair review by medical experts based on the current state of the medical literature and plaintiff’s current medical condition and suitability for the proposed treatment.
Id. at *15.
The district court therefore remanded for the supplementation of the record that it deemed crucial. Such a remand is the proper step to take under the principles enunciated by this Court in Miller: “[I]f upon review a district court concludes that the [administrator’s] decision was arbitrary and capricious, it must remand to the [administrator] with instructions to consider additional evidence unless no new evidence could produce a reasonable conclusion permitting denial of the claim....” 72 F.3d at 1071. The remand order directed Empire to conduct an immediate and accelerated MCOP review within 48 hours of Zervos’s submission of current relevant information, the review to include “direct questions that elicit a clear and unequivocal response regarding whether the treatment is of proven benefit or generally recognized by the medical community as effective or appropriate ... [as well as] whether there is any other medical or clinical reason that the patient should receive the treatment despite its experimental/investigational nature.” Zervos III, at *15. The district court signaled that “[t]he totality of the evidence presented does not convince this Court that a different determination is necessarily likely or appropriate. However, plaintiff is entitled to a full and fair review.” Id.
The third MCOP review (as ordered by the district court) was completed on October 25, 2001 — three days after the district court issued its order and four days before *654this Court heard oral argument on this expedited appeal. This time, MCOP designated Dr. Joyce O’Shaughnessy, a board certified medical oncologist whose primary areas of expertise include high dose chemotherapy for breast and ovarian cancer. Asked, “Is the proposed treatment of ‘proven benefit’ for the treatment of the patient’s condition?” Dr. O’Shaughnessy answered, “No,” adding:
I believe that the available data permit us to conclude that there [is] no proven clinical benefit associated with [HDCT] compared with standard therapy for patients with metastatic breast cancer. The morbidity and mortality of [HDCT] are greater than that associated with standard therapy. [HDCT] does not represent accepted standard care for patients with metastatic breast cancer and is not medically necessary.
Asked whether “the proposed treatment [is] ‘generally recognized by the medical community as effective or appropriate’ for the patient’s condition,” Dr. O’Shaughnes-sy answered, “No,” adding: “most medical oncologists feel that [HDCT] is ineffective therapy for patients with metastatic breast cancer in general, and for patients like Mr. Zervos, in particular, who have largely chemotherapy-resistant and heavily-pretreated metastatic breast cancer.” She warned: “I believe that [HDCT] could even be harmful for Mr. Zervos because he is unlikely to benefit, i.e., to achieve a durable remission of his disease, and [HDCT] is substantially more toxic than treatment with conventional chemotherapy agents for metastatic breast cancer.” Finally, in response to the question “If your answer to either of the first two questions is ‘no’, is there any other medical or clinical reason that the patient should receive the treatment, despite its ‘experimental/in-vestigational’ nature?” Dr. O’Shaughnes-sy answered: “I believe Mr. Zervos should not receive the proposed [HDCT] because it is highly unlikely to benefit him, and, due to its associated toxicity, may even be harmful.... I do not think there is a medical or clinical reason for him to receive this therapy.”
D
The district court therefore did what we have mandated: “[I]f upon review a district court concludes that the [administrator’s] decision was arbitrary and capricious, it must remand to the [administrator] with instructions to consider additional evidence unless no new evidence could produce a reasonable conclusion permitting denial of the claim.... ” Miller, 72 F.3d at 1071. It is certainly hard going for the majority to decide in this case that “no new evidence could produce a reasonable conclusion permitting denial of the claim”: as it happens, the third review, ordered by the district court and rendered before this Court’s hearing of this appeal, was new evidence (albeit confirmatory of the earlier medical evidence) and did support “a reasonable conclusion permitting denial of the claim.”
The majority opinion deals with this embarrassment by ruling that the only relevant coverage issue was settled by a single answer given by Dr. Spitzer to a single question, and that therefore the record— thus complete — required no supplementation. Majority Op., at 648 (citing Zuckerbrod v. Phoenix Mut. Life Ins. Co., 78 F.3d 46, 51 n. 4 (2d Cir.1996) (finding a remand to the administrator inappropriate “where the difficulty is not that the administrative record was incomplete but that a denial of benefits based on the record was unreasonable”)).
After thus shrinking the administrative record in this case to a few lines, the majority declares that it can “only be read *655to support granting coverage.” Majority Op. at 648.
This ruling is premised on several errors. First, the majority reads the policy to furnish any medical procedure that is “effective” regardless of the toxicity of the treatment. Second, the majority truncates the record to omit everything other than Dr. Spitzer’s single isolated response. Third, the majority rules that the remand order was an abuse of discretion because the district court failed to act with the dispatch required by the medical exigency. Id.
1. The Policy Wording
To repeat, the district court ruled that Empire’s appellate process was arbitrary and capricious principally because the questions posed to the reviewers did not track the wording of the Experimental/In-vestigational exclusion, and therefore (according to the district court) the answers elicited by these questions did not settle the issue of whether HDCT fell under the terms of the exclusion. Accordingly, the district court concluded that the record was incomplete and remanded for its completion.
The majority opinion agrees that the appeals process was arbitrary and capricious (a point I do not consider), but for a different reason. If the majority agreed that the record was incomplete, the majority would be hard pressed to ignore Dr. O’Shaughnessy’s categorical opinion rendered on remand. Instead, the majority finds that the record was in fact complete because the only record evidence that matters is Dr. Spitzer’s answer to a single question: “[HDCT] appears to be as effective as prolonged combination chemotherapy for metastatic breast cancer.” The majority holds that Empire’s appellate process was arbitrary and capricious because it denied coverage notwithstanding that response:
[Empire] required that a treatment be superior to another existing treatment in order to avoid exclusion under the policy’s experimental/investigational language while the language itself requires only that the treatment be effective— not more effective than alternatives— and considered to be appropriate by the relevant medical community. Thus, contrary to basic ERISA principles, [Empire] in effect added additional language to the policy.
Majority Op. at 647 (citing Gallo v. Madera, 136 F.3d 326, 330 (2d Cir.1998) (“Even when trustees of a pension plan are entitled to deference in interpreting the terms of the plan, deference cannot be so broad as to permit them to graft additional requirements onto unambiguous plan definitions.”)).
The majority misreads the policy language and fails to give deference to Empire’s reasonable interpretation of it. The actual wording of the policy requires a treatment to be of “proven benefit” for the claimant’s condition and to be “generally recognized by the medical community as effective or appropriate” for that condition. The term “proven benefit” is not self-reading, and lends itself to interpretation. And Empire as the plan administrator is entitled to deference in interpreting the terms of the plan. See Firestone Tire & Rubber Co. v. Bruch, 489 U.S. 101, 115, 109 S.Ct. 948, 103 L.Ed.2d 80 (1989); Miller, 72 F.3d at 1070.
Empire interprets “proven benefit” as a (proven) positive net health effect — i.e., “the efficacy [of the treatment] balanced by the toxicity and whether that net positive health effect compares favorably with other standard treatments.” (emphasis added) The district court concluded that this is a natural reading of the term:
*656Medically, a treatment cannot be said to be of proven benefit where its toxicity or side effects outweigh the benefits gained from the treatment. Moreover, a treatment cannot be said to be of proven benefit where it merely matches the benefits gained from other treatments, but exposes patients to greater harm and potential death.
Zenos III, 2001 WL 1262941, at *8. Dr. Weiss concurred at his deposition: “if the outcome, the overall survival is the same, yet the treatment is more toxic, you don’t want to use the more toxic therapy because there is no gain for the cost.” Zer-vos’s own expert witness also agrees: “if the costs in terms of higher morbidity and mortality rates are much greater, but the benefits are only shown to be slightly greater ..., then it’s totally appropriate not to give that treatment over standard therapy.” [Zervos Ex. F at 417-18] HDCT is a treatment so invasive and verges so closely on killing the patient that the netting out of risks is no minor miscalculation.2 In medicine, as in virtually everything else, a calculation of benefit entails the netting out of the pluses and minuses. Otherwise, the concept of benefit becomes absurd: if medical benefit were a matter of pure efficacy, regardless of detriment, amputation would be deemed a beneficial therapy for athlete’s foot.
Empire therefore is not “grafting] additional requirements onto unambiguous plan definitions.” Gallo, 136 F.3d at 330. This case is distinguishable from Gallo, on which the majority relies. In Gallo, the insurance company attempted to add a “break-in-service” clause to the eligibility section of a pension plan; the insurance company did not claim to be interpreting a provision of that section, but instead wanted to “import” the clause from another unrelated section. Gallo, 136 F.3d at 328.
2. Truncation of the Record
The majority’s key ruling is that the administrative record in this case can “only be read to support granting coverage.” Majority Op. at 648. In defense of this ruling, the majority has excluded from “the record” all of what Dr. Weiss and Dr. O’Shaughnessy had to say, has ignored (or *657misinterpreted) most of what Dr. Spitzer had to say, and has seemingly forgotten that Empire’s decision to end funding for HDCT was supported by extensive medical learning and is unchallenged.
The majority excludes the second and third MCOP reviews from the administrative record on the ground that Empire procured these additional reviews to “bolster” its legal position rather than to provide a fuller review of Zervos’s claim. Majority Op. at 647. There simply is no factual support for the idea that, if either review had recommended HDCT for Zervos’s condition, Empire would have disregarded this recommendation and denied the claim. It is undisputed that the second MCOP review (by Dr. Weiss) was conducted because Verizon asked for it, and Empire testified that it always complies with such requests from the claimant’s employer; and the third MCOP review (by Dr. O’Shaughnessy) was conducted pursuant to the order of the district court. These second opinions were thus conducted at the behest of the employer (who would pay the bill) and the district court (which could order the bill paid).
These reviews were therefore part of an ongoing review of Zervos’s claim that would ultimately determine whether or not his claim would be paid. See Zuckerbrod, 78 F.3d at 50 n. 2 (upholding a district court’s inclusion in the administrative record of a doctor’s review received after the insurance company denied the claim, because “it was ... not clear error for the district court to treat the insurer’s ultimate decision as having been rendered after [the doctor’s] review”). In any event, I do not see how Empire can be taxed with improperly bolstering its case with additional opinions by eminent, neutral doctors selected by an unaffiliated intermediary.
Finally, the majority’s key finding (that the record only supports coverage) ignores the body of evidence accumulated when Empire changed its policy in March 2000 and decided as a matter of policy to deny coverage of HDCT — a decision that the majority concedes was perfectly reasonable. Majority Op. at 647-48. And it ignores as well the opinion of all the medical experts in this case — including Zervos’s own physician as well as his expert witness — that HDCT is significantly more toxic than conventional chemotherapy, i.e., that it has significantly higher morbidity and mortality rates. Given this uncontested view, it is not enough for HDCT to be as effective as conventional chemotherapy for the treatment of metastatic breast cancer; in order to compete (“compare favorably”) with conventional chemotherapy in terms of benefit, HDCT must be more effective. This is why Dr. Spitzer was able to say, without self-contradiction, [i] that HDCT is as effective as conventional chemotherapy for the treatment of metastatic breast cancer and [ii] that HDCT is not the best treatment available to Zervos. And that is how Dr. Spitzer characterized his views in the deposition he gave in this case: “the reason that the treatment is not well adopted by the oncology community anymore is that there isn’t an advantage, and given potential disadvantage of the treatment in terms of morbidity, ... it probably does not have a substantial place in the treatment of [metastatic breast cancer].” [Empire Ex. M at 20-21]
This case is a far cry from Zuckerbrod, upon which the majority relies. 78 F.3d at 51 n. 4 (finding a remand to the administrator inappropriate “where the difficulty is not that the administrative record was incomplete, but that a denial of benefits based on the record was unreasonable”). In Zuckerbrod, we granted an injunction mandating coverage when two reports *658upon which the insurance company claimed to rely stated that it was absolutely essential that claimant receive the full-time private duty nursing care that the insurer refused to fund. Zuckerbrod, 78 F.3d at 49-50. No one, including Dr. Spitzer, has suggested here that it is absolutely essential that Zervos receive HDCT.
The majority also attempts to reinforce its conclusion by citing Dr. Spitzer’s characterization of the proposed research protocol involved in this case as a “well designed Phase II study” and by alluding to a prior case in which Empire provided coverage for HDCT after the MCOP reviewer characterized the proposed research protocol involved in that case as a “well-designed randomized phase III trial.” However, Dr. Spitzer indicated that the proposed protocol involved in this case is for women, not men; and Dr. Weiss explained why that matters. It is therefore not clear that Zervos is even eligible to participate in the proposed protocol at issue here. As to the prior claim, there are material differences between Zervos’s case and that case. In the prior case, the patient had Stage II breast cancer and was seeking participation in a randomized Phase III clinical trial; here Zervos has a more severe form of breast cancer (Stage IV) and is seeking participation in a lower quality clinical trial (Phase II). As the district court found:
[The] evidence indicates that coverage for the treatment was provided for the female patient and denied for Mr. Zer-vos because of the quality of the clinical study. This Court cannot conclude that Empire acted arbitrarily and capriciously in providing coverage for one patient and denying coverage for another where there are seminal differences in the state of the patients’ disease and the quality of the studies for which they sought coverage.
Zervos III, 2001 WL 1262941, at 12.
3. Time Constraints
The majority opinion finds an abuse of discretion on the ground that the district court ordered a remand in disregard of the exigency presented by Zervos’s condition and the risk that his condition might change in a way that would render him ineligible for HDCT. I think that the district court acted with energy and speed, ordering an immediate expedited review within 48 hours of Zervos’s submission of current relevant information. That expedited review was completed four days before we heard oral argument on this appeal. There is every reason to think that the district court would have ruled on this third review as rapidly as we ruled on this appeal.
Fortunately, the majority opinion expressly casts its ruling in terms of the “exigent and unique circumstances of this case,” majority op. at 646, a limitation that invites fact-based distinctions in any subsequent case, however similar. In short, the majority opinion has not altered the law of this Circuit in any fashion.
. For an explanation of the categorization of clinical studies, see majority op., at 640 n. 4.
. We have previously described HDCT as follows:
The treatment is a procedure by which stem cells are harvested from the bone marrow of the patient’s body and purified of cancer cells. The patient is placed under general anesthesia while the bone marrow is extracted by needle. The bone marrow is then frozen and stored while the patient receives high, and potentially toxic, doses of chemotherapy. In some cases, the chemotherapy is administered in doses which exceed one thousand times the standard dosage for conventional chemotherapy treatment. This high dose chemotherapy kills not only the cancer, but also the patient’s remaining bone marrow which produces white blood cells to protect the body from infection. The bone marrow, which is the most sensitive of all the body tissue, is also the most damaged by chemotherapy. After the chemotherapy is completed, the patient’s stored bone marrow is reinfused intravenously so that it may re-engraft. The bone marrow then multiplies rapidly to replace the marrow destroyed during the high-dose chemotherapy. Given that the bone marrow is the patient’s own tissue, there is little danger of rejection. There are, however, significant dangers associated with administering high-dose chemotherapy without some additional treatment to regenerate the bone marrow. Because the toll on a patient’s white blood cells is significant, the secondary treatment is essential to the patient's chances for survival. Thus, the bone marrow must be quickly reintroduced after high-dose chemotherapy treatment to "rescue” the patient from otherwise almost certain death.
Zervos v. Verizon New York, Inc., 252 F.3d 163, 166 n. 1 (2d Cir.2001) (quoting Sharona Hoffman, A Proposal for Federal Legislation to Address Health Insurance Coverage for Experimental and Investigational Treatments, 78 Or. L.Rev. 203, 211-12 (1999)).