dissenting.
The majority correctly identifies the “double issue” we face in determining the applicable standard of review in this case: first, whether Massachusetts General Hospital (“the Hospital” or “MGH”) is expressly authorized by the MGH Long Term Disability Plan (“the Plan”) to delegate its authority to determine benefits, and, second, whether MGH in fact made such a delegation to the plan administrator, Partners Healthcare System, Inc. (“Partners”). My colleagues answer “yes” to each of those questions. In so concluding, however, they disregard our precedent requiring a clear statement of the authority to delegate, fail to respect the limits of the pertinent Plan language, and uncritically accept Partners’ declaration of fiduciary authority despite that assertion’s in*296consistency with the terms of the primary plan document.
In my view, the Plan does not give the Hospital authority to delegate and, even if it did, Partners’ assertion of its own authority is insufficient evidence that a proper delegation in fact occurred. Hence, because neither question may be answered affirmatively, the de novo standard of review must be used to evaluate the administrator’s decision denying benefits to Maher. Taking a fresh view of the record, I can only conclude that Maher’s symptoms render her incapable of sedentary work. I would thus vacate the district court’s judgment and remand for entry of judgment in Maher’s favor.
I.
Maher challenged the termination of her benefits under section 502 of the Employee Retirement Income Security Act (“ERISA”), 29 U.S.C. § 1132(a)(1)(B). ERISA does not prescribe a standard of review for such actions. To fill this gap, the Supreme Court has held that a denial of benefits challenged under section 1132(a)(1)(B) should be reviewed de novo “unless the benefit plan gives the administrator or fiduciary discretionary authority to determine eligibility for benefits or to construe the terms of the plan.” Firestone Tire & Rubber Co. v. Bruch, 489 U.S. 101, 115, 109 S.Ct. 948, 103 L.Ed.2d 80 (1989). Where such authority is given, our review is more deferential, id. at 111, 109 S.Ct. 948, and a benefits decision will be upheld unless it is “arbitrary, capricious, or an abuse of discretion,” Cusson v. Liberty Life Assurance Co. of Boston, 592 F.3d 215, 224 (1st Cir.2010) (quoting Gannon v. Metro. Life Ins. Co., 360 F.3d 211, 218 (1st Cir.2004)).
The standard of review inquiry must thus start with the terms of the Plan, through which MGH provided Maher LTD benefits. Article 6.1 of the Plan document states in full:
6.1. General. The processing of claims and calculation of benefits shall be the sole responsibility of the Claims Processor. The Hospital shall have full discretionary authority to administer the Plan, including without limitation the authority to determine eligibility for benefits, construe the terms and conditions of the Plan, and resolve disputes as to the interpretation of the Plan documents. Any person having an interest under the Plan m[a]y request a determination by the Hospital with respect to any matter affecting such person, and any such determination of the Hospital will be final and binding, and shall not be subject to de novo review, or be modified, amended, or set aside by any judicial or administrative authority in the absence of clear and convincing evidence that the Hospital’s action was arbitrary and capricious or without rational basis.
The Plan document defines the “Hospital” to mean The General Hospital Corporation, a Massachusetts charitable organization whose sole member is MGH. Because the above provision grants the Hospital “full discretionary authority to administer the Plan,” including “the authority to determine eligibility for benefits [and] construe the terms and conditions of the Plan,” a benefits decision by MGH would typically be reviewed under the deferential “arbitrary and capricious” standard. Firestone, 489 U.S. at 115, 109 S.Ct. 948; Cusson, 592 F.3d at 224.
The benefits decision at issue here was not, however, made by MGH. Instead, it was made by Partners, the Plan’s administrator. To be sure, Partners and the Hospital are closely related. Partners, a nonprofit corporation, coordinates a healthcare system made up of a number of constituent medical institutions, including MGH, *297Brigham and Women’s Hospital, and various other hospitals and medical organizations. However closely bound together, though, there is no question that Partners and the Hospital are legally distinct entities, and the Plan does not argue that the discretionary authority conferred on the Hospital should be imputed to Partners.7 Cf. Glotzer v. Metro. Life Ins. Co., 1 Fed.Appx. 740, 742 (9th Cir.2001) (holding that, because successor corporation inherited in full the rights and burdens of plan administrator through purchase, discretion delegated to plan administrator was imputed to successor).
Instead, the Plan contends that Partners validly exercised discretionary authority that had been delegated to it by the Hospital. ERISA provides that a named fiduciary may delegate to others its responsibilities under a plan — other than trustee responsibilities — where the plan expressly sets forth procedures for doing so. 29 U.S.C. § 1105(c)(1); Rodriguez-Abreu v. Chase Manhattan Bank, 986 F.2d 580, 584 (1st Cir.1993). To determine whether a valid delegation of authority took place here, we must address the majority’s “double issue”: (1) whether the Plan authorized the Hospital to delegate its authority to Partners and, if so, (2) whether there is adequate evidence of the claimed delegation of authority.
A. Was Delegation Authorized?
A plan fiduciary may delegate its fiduciary responsibilities to a third party only where the plan “expressly provide[s] for procedures” for such a delegation. 29 U.S.C. § 1105(c)(1). As explained in Wallace v. Johnson & Johnson, 585 F.3d 11, 15 (1st Cir.2009), our court has placed little emphasis on the statute’s reference to delegation “procedures.” That is, we do not require that a plan establish any procedures governing delegation beyond a basic grant of the authority to delegate itself. See id. (“Our own cases treat delegation ‘authority’ and delegation ‘procedures’ as more or less the same thing....”). However, the statute’s reference to “expressly providing]” delegation procedures means that such a grant of authority to delegate must be clearly stated. This is consonant both with the requirement that the initial grant of discretionary authority by a plan be unambiguous, see Terry v. Bayer Corp., 145 F.3d 28, 37 (1st Cir.1998); Rodriguez-Abreu, 986 F.2d at 583, and with our case law examining plan language granting authority to delegate.
In Wallace, we found a valid delegation of discretionary authority where it was “clear” that the benefits plan at issue “purport[ed] to allow delegation.” 585 F.3d at 14. The relevant language of the plan stated that the fiduciary “may ‘[djelegate its authority established’ by the Plan, ‘designate persons to assist in carrying out fiduciary duties,’ ‘allocate responsibility for the operation and administration’ of the Plan, and ‘[ajppoint persons or committees to assist it to perform its duties’ under the Plan.” Id. at 14-15. We reached the same conclusion in Terry, where the plan stated that “ ‘[t]he Company may appoint one or more individuals to act on its behalf, in which case every reference herein made to the Company shall be deemed to mean or *298include the individuals as to matters within their jurisdiction.’ ” 145 F.3d at 37-38. In each of these cases, the plan language was unambiguous and unqualified in its grant of authority to delegate.
Here, in contrast, the Plan cannot be read to authorize a delegation of plenary administrative authority over the Plan to Partners. The most relevant provision of the Plan document, and the one emphasized by the Plan,8 reads as follows:
6.3. Employment of Agents. The Hospital may employ agents, including but not limited to, a Claims Processor, accountants, attorneys or actuaries to perform such services and duties in connection with the administration of the Plan as it may direct.... The Hospital shall be fully protected in acting upon the advice of any such agent, in whole or in part....
Unlike the plan language in Terry and Wallace, the provision here authorizes no complete devolution of the discretionary authority granted to the Hospital under Article 6.1 of the Plan. Rather, Article 6.3 allows the Hospital to “employ” agents, often referred to as service providers in ERISA parlance, whose activities it may “direct,” to assist with the day-to-day operation of the Plan and to advise the Hospital in carrying out the fiduciary duties assigned to it in Article 6.1. In Terry, in the course of determining whether a third-party claims processor could be sued as a fiduciary under ERISA for its role in denying the claimant benefits, we drew a sharp distinction between these types of “third party service providers],” on the one hand, and the plan administrator and fiduciaries, on the other. 145 F.3d at 35-36.
The reference to a “Claims Processor” as one of the agents the Hospital may employ reflects the limited nature of the delegation authorized by Article 6.3, and, critically, that it does not include the ultimate authority to “determine eligibility for benefits or to construe the terms of the plan” — the authority that triggers deferential review. Firestone, 489 U.S. at 115, 109 S.Ct. 948. A claims processor generally makes the initial determination on a benefits claim and may handle the first level of appeals, but another entity is usually responsible for making a final determination on appeal.9 In the typical instance, we have thus noted that “an entity which merely processes claims ‘is not a fiduciary because such person does not have discretionary authority or discretionary control respecting management of the plan.’ ” Terry, 145 F.3d at 35-36 (quoting 29 C.F.R. § 2509.75-8, D-2 (1997)). Such is plainly the case under the Plan. Article 6.1 sets the Hospital’s authority side-by-side with the Claims Processor’s. It makes clear that, though “[t]he processing of claims and calculation of benefits shall be the sole responsibility of the Claims Processor,” it is the Hospital that has “full *299discretionary authority to administer the Plan, including without limitation the authority to determine eligibility for benefits, construe the terms and conditions of the Plan, and resolve disputes as to the interpretation of the Plan documents.”
Accordingly, I read Article 6.8 to authorize delegation of administrative functions short of the Hospital’s core authority to “determine eligibility for benefits” and “construe the terms and conditions of the Plan.” 10 At a minimum, if this provision were intended to permit the delegation of the totality of the Hospital’s discretionary authority, it lacks the requisite clarity. Cf. Terry, 145 F.3d at 37 (“[T]he grant of discretionary authority must be clear.”). I thus conclude that Article 6.3 did not authorize the Hospital’s delegation of discretionary authority to Partners. However, as discussed below, even if the Plan permitted such delegation, the evidence that a delegation took place is inadequate.
B. Did a Delegation Occur?
The Plan asserts that evidence of a delegation of authority to Partners can be found in a 2005 document entitled “Partllers Healthcare System, Inc. Health and Welfare Plan Document” (the “Partners Plan Document”), which the Plan characterizes as a Summary Plan Description (“SPD”). The document, published under Partners’ name, purports to be “a wraparound plan document that contains the definitions, participation and administration provisions of the various Partners health and welfare plans,” which includes the Plan at issue here as well as over forty others, “and incorporates by reference the various Benefit Contracts associated with the Plan to form a complete plan document.” The language supposedly effecting the delegation, located in Section II, Article IX of the Partners Plan Document, simply states, “Partners acts as the Plan Administrator for ERISA purposes of the Plans.” The Partners Plan Document further provides that, as Plan Administrator, Partners “has the discretion to determine all matters relating to eligibility, coverage and benefits under each Plan” and “has the full power to interpret each Plan and is responsible for the operation of each Plan.”
Our cases involving delegation of fiduciary authority under ERISA do not directly *300address what sort of evidence will suffice to show that delegation has been effected. To some extent, the answer in an individual case may depend upon whether the ERISA plan at issue sets forth particular procedures to canalize the delegation process. However, often, as here, the plan will not specify a mechanism for delegation. In these cases, our precedents hint at some basic guidelines for determining whether the evidence before the court establishes the delegation of fiduciary authority.
First, it is not enough merely to show that the putative delegate is carrying out discretionary functions of plan administration. In Rodriguez-Abreu, a senior executive of the defendant employer conducted a review of the denial of plaintiffs benefits and corresponded with the plaintiff regarding his eligibility. 986 F.2d at 582, 584. The defendant argued that these circumstances alone were sufficient to show that the plan fiduciaries had delegated their discretion to the executive, triggering a deferential standard of review. We disagreed. Noting that there was “no expression of intent that [the executive] act as the delegate of the Fiduciaries” in the plan documents or elsewhere in the record, we found no valid delegation of authority and thus employed a de novo standard of review. Id. at 584.
Second, evidence of the delegation may be provided by a written instrument other than the plan documents themselves. In Terry, the benefit plan at issue conferred discretionary authority to administer the plan to the Bayer Corporation, and Bayer in turn delegated that authority to an internal “Benefit Administration Committee.” 145 F.3d at 37-38. The only evidence of this delegation was an internal organizational document (the “Administrative Procedures for the Benefit Administration Committee”), which, among other things, explained that the Committee had been formed to “ ‘act on behalf of [Bayer] by assisting [Bayer] in fulfilling its administrative duties which are set forth in the employee benefit plans.’ ” Id. at 35, 38. We found this “clear and direct delegation — by written instrument — from the Plan Administrator to the Benefit Committee” sufficient to justify a deferential standard of review. Id. at 38; see also Wallace, 585 F.3d at 14-15 (holding extra-plan written instrument transferring discretionary authority to administer plan to third party sufficient evidence of valid delegation to trigger review under the “arbitrary and capricious” standard).
Here, we have a written document — the Partners Plan Document — that purportedly gives to Partners the discretionary authority to administer the Plan. However, the relevant terms in the Partners Plan Document represent nothing more than the bare assertion by a third party of discretionary authority over administration of the Plan. The document states simply that “Partners acts as the Plan Administrator” for all plans listed among its pages,11 and that Partners “has the discretion to determine all matters relating to eligibility, coverage and benefits under each Plan” and “has the full power to interpret each Plan and is responsible for the operation of each Plan.” This is at best corroborative evidence of a transfer of authority from the Hospital to Partners, but insufficient on its own to establish the delegation. The Plan has offered no evi*301dence of an agreement between the Hospital and Partners, a corporate resolution by the Hospital, or other such documentation that would indicate that the Hospital affirmatively granted discretionary authority over the Plan to Partners.
In arguing over the effect of the Partners Plan Document, the parties vigorously dispute whether it properly qualifies as an SPD for the Plan.12 This is a question that need not be resolved here, as nothing turns on it. Even if the Partners Plan Document qualified as an SPD, there would be a direct conflict with the primary Plan document that must be resolved in the Plan document’s favor. Each document purports to grant full discretionary authority over the Plan, with the power to make “final and binding” decisions, to a different entity. Where there is a conflict between the plan and the SPD, the plan language will generally control, except in situations where the beneficiary relied to her detriment on the SPD. See Ringwald v. Prudential Ins. Co. of Am., 609 F.3d 946, 948-49 (8th Cir.2010) (disregarding grant of discretionary authority that appeared only in SPD); Schwartz v. Prudential Ins. Co. of Am., 450 F.3d 697, 699-700 (7th Cir.2006) (same, and explaining that SPD controls if participant detrimentally relied on it); cf. Mauser v. Raytheon Co. Pension Plan for Salaried Emps., 239 F.3d 51, 54-55 (1st Cir.2001) (where SPD conflicts with a plan’s terms, the SPD will control if the claimant demonstrates “significant or reasonable reliance” on the SPD). Thus, to the extent that the Partners Plan Document qualifies as an SPD, the grant of discretionary authority in the Plan document must still be credited over the conflicting grant in the putative SPD.13
In sum, the Plan by its terms did not authorize the Hospital to delegate its authority, and, even if the Plan had permitted such delegation, there is insufficient evidence that the Hospital in fact delegated its authority to Partners. Partners’ denial of Maher’s benefits must therefore be reviewed de novo.
II.
In a de novo review of a benefits decision, “no deference [is given] to administrators’ opinions or conclusions based on the[ ] facts.” Orndorf v. Paul Revere Life Ins. Co., 404 F.3d 510, 518 (1st Cir.2005); see also Richards v. Hewlett-Packard Corp., 592 F.3d 232, 239 (1st Cir.2010) (same). Instead, an independent review of the evidence in the administrative record must be performed, Orndorf, 404 F.3d at 518; Richards, 592 F.3d at 239, and the “guiding principle” is that the plaintiff bears the burden of proving he is disabled, Orndorf, 404 F.3d at 518-19. To meet that burden, the terms of the Plan require Maher to prove a “complete incapacity” preventing her “from performing any and every duty of any occupation or employ*302ment, for which [she] is reasonably qualified by education, training or experience.” The Plan has narrowed the scope of the inquiry somewhat by specifying the occupations for which it believes Maher to be qualified: telephonic triage nurse, nurse case manager, and utilization review nurse.
Maher’s arguments on the merits of the Plan’s benefits determination fall into two categories. First, Maher argues that her use of narcotics renders her incapable of filling any of the above-mentioned jobs for which she is qualified by “education, training, or experience.” In support of this argument she points to state law that purportedly prohibits her from performing any nursing duties while under the influence of narcotics. Alternatively, she submits that, legality aside, the effects of her medications would impair her judgment and interfere with her ability to perform nursing duties. Second, Maher contends that the evidence supports the disability caused, by her chronic pain and nausea, and that Partners wrongly discounted evidence of the manner and extent to which pain affects her ability to function.
I address each of these arguments in turn.
A. Disabling Effects of Narcotics Use
Maher’s medical records confirm that, for nearly a decade, Maher has treated pain with a varying array of powerful narcotics, and she usually has taken multiple drugs at the same time and in significant doses. In a 2007 letter, Dr. Elizabeth Cuevas, Maher’s primary care doctor, noted that her patient is “treated with narcotics that are prescribed at very high doses,” and Dr. Wolfram Goessling, Maher’s treating gastroenterologist, has similarly noted that Maher “takes impressive amounts of narcotics.” Over the course of her treatment, Maher has consulted several pain clinics and periodically changed medications, but her records reveal no success in scaling back the use of narcotic painkillers. According to Dr. Herbert Malinoff, one of Liberty’s consulting doctors, Maher’s record of treatment supports a diagnosis of “chemical dependency relative to opioid and Benzodiazepine use over a long period of time.”
Maher contends that her drug use, as a matter of law, prevents her from working in any job within the field of nursing, and thus precludes employment in the one area for which she is qualified by education, training, or experience. As Maher notes, Massachusetts’ Standards of Conduct for licensed nurses prohibit nurses from practicing “while impaired.” 244 Mass.Code Regs. § 9.03(36). Maher argues that the sedentary nursing roles that Liberty has suggested she would be capable of filling all necessarily involve the “practice” of nursing, defined by statute to include “coordination and management of resources for care delivery.” Mass. Gen. Laws Ann. ch. 112, § 80B. Because she is “therapeutically addicted to narcotic medication,” Maher suggests that she is legally “impaired” and cannot serve in the above positions.14
Though Maher tries to segregate this line of reasoning from her alternative argument that the effects of her medications would interfere with her ability to carry out the duties of a nursing job, they cannot be so easily disentangled. The fact of narcotics use does not, by itself, appear to preclude legal function as a nurse. The governing regulations define “impaired” to mean “the inability to practice nursing *303with reasonable judgment, skill, and safety by reason of alcohol or drug abuse, use of substances, a physical or mental illness or condition, or by any combination of the foregoing.” 244 Mass.Code Regs. § 9.02. Thus, whether Maher is “impaired” turns not on whether she is addicted to narcotics, but on a factual determination of whether her narcotic use is likely to interfere with her “judgment [and] skill.” Id.
The evidence of the effect of Maher’s narcotic use on her ability to concentrate and function is mixed. There is, to be sure, an intuitive logic to the notion that we should be wary of entrusting patient care to a medical professional who takes large amounts of narcotic medications. This is a point that Maher emphasizes in her briefs, and it has some basis in fact. Medical authorities warn that narcotics may cause sedation, “mental clouding,” and impaired psychomotor function.' However, according to a publication of the American Medical Association included in the administrative record, these types of side effects “usually dissipate with continued treatment, normally within a week with regularly scheduled dosing, and studies have demonstrated that most patients on chronic opioid therapy can safely drive ears.” One of Liberty’s consulting physicians, Dr. Robert Millstein, makes the same point in his February 2007 report, though he acknowledges “very few studies have examined the effects of opioids on selective, divided, and sustained attention.”
Maher’s medical records indicate that she has often complained of the sedation brought on by her regimen of narcotics. However, close scrutiny of these records indicates that many of her reports of somnolence and other drug-related side effects are correlated with a change in medication or an adjustment of dosage. To take one example, Maher saw Dr. Goessling in March 2005 and complained that the morphine she was taking made her “loopy and more sleepy than usual.” Her medical records reveal that she was started on morphine only a couple of days prior, when she was admitted to the hospital for abdominal pain and the pain clinicians recommended that she transition from OxyContin.
Overall, I do not find convincing support in the record for the proposition that Maher’s drug use, by itself, would render her incapable of working in one of the sedentary nursing positions identified by Liberty. There is no clear prohibition in Massachusetts on serving as a nurse while on a medically prescribed regimen of narcotics, so long as the drugs do not impair the nurse’s “judgment, skill, and safety.” Maher’s doctors have not suggested that her use of drugs would interfere with her judgment or ability to make rational decisions. Indeed, Dr. Goessling and Dr. Cuevas both identify excessive sedation to be the major side effect associated with Maher’s medication. I acknowledge that Dr. Goessling has opined that he “cannot imagine that it would be in [Maher’s] or society’s benefit to have her work as a nurse making decisions over other people’s lives,” but he does not point to any specific effect of the drugs that would render Maher incapable of safely managing the limited nursing roles suggested for her by Liberty.
I am not unconcerned by the possibility that narcotic-related sedation would make it difficult for Maher to complete a regular workday. In one instance in 2005, Maher reported that the somnolence associated with a change from OxyContin to Palladone was causing her to take three naps per day. Such considerable sedation would clearly interfere with most if not all possible employment. However, because Maher’s reports of drowsiness and other adverse effects from medication appear *304largely related to alterations in her drug regimen — which conforms to the medical evidence that side effects are most pronounced when first starting a narcotic — I do not see convincing evidence that Maher’s ability to concentrate and exercise judgment would be significantly compromised by a regular and stable program of narcotic medications.
B. Disabling Effects of Pain and Other Symptoms
I reach a different conclusion with respect to the impact of the symptoms of Maher’s maladies — chronic pain, nausea, vomiting, and food intolerance — on her ability to work. After careful review of the record, I find persuasive evidence that Maher’s symptoms would prevent her from performing the duties of the jobs identified by Liberty.
Before evaluating the record on this issue, I note that my concern here lies with the evidence of limitations attributable to Maher’s symptoms, and not with whether the evidence supports Maher’s underlying medical conditions. There is considerable uncertainty regarding the etiology of Maher’s abdominal pain and other complaints, and attempting to resolve a question that has stymied multiple doctors for the past decade is both unnecessary and beyond my expertise.15 The diagnosis makes little difference here. Our court has emphasized before that in dealing with hard-to-diagnose, pain-related conditions, it is not reasonable to expect or require objective evidence supporting the beneficiary’s claimed diagnosis. See Cook v. Liberty Life Assurance Co., 320 F.3d 11, 21 (1st Cir.2003). Our focus instead must be on whether evidence supports an inability to work due to “the physical limitations imposed by the symptoms of such illnesses.... ” Boardman v. Prudential Ins. Co. of Am., 337 F.3d 9, 16 n. 5 (1st Cir.2003).
1. Evidence Relied upon by the Plan
I begin by assessing the evidence that, in the view of the Plan, calls Maher’s limitations into question, the foremost of which is the alleged inconsistency between Maher’s reported capabilities and the level of activity confirmed by surveillance. Liberty conducted surveillance of Maher on nineteen days between October 2002 and October 2006,. portions of which were recorded on video.16 The Plan and its doctors highlighted a handful of examples of increased activity by Maher captured in the surveillance. On one occasion, Maher was observed walking to the front of her property “carrying what appeared to be a flower/plant and a large bucket,” and, four minutes later, walking back with the same bucket. On another, she drove herself a short distance (a four-minute drive) to a local school, where she went inside and returned carrying her son (then close to three years old), whom she placed inside the car before departing for her home. On a third occasion — and the one on which the Plan places the most emphasis — Maher’s husband drove Maher and her son to a local soccer field on a Saturday morning. There, Maher “was periodically observed as she and her husband flew a kite with the young boy,” as she “walked and jogged around the soccer field,” and as she “at one point lifted the small boy and swung *305him around in her arms.” The outing lasted thirty-four minutes.
This surveillance.evidence does not have the significance that the Plan ascribes to it. In the activity questionnaires she submitted to Liberty, Maher consistently reported that the level of activity she can sustain is entirely dependent on her pain, nausea, and level of medication. For example, in her latest questionnaire, dated September 2006, Maher indicated that the amount of time she can tolerate sitting, standing, and riding in or driving a car depends on the presence of pain, nausea, vomiting, and diarrhea. She also noted that she leaves the house during the week two to three times a day (one to two times on weekend days), and that she helps take care of her children and perform small chores when she is able.
As the majority points out, Liberty’s surveillance is not inconsistent with Maher’s own account of her activities. Over nineteen days of surveillance, there were a number of days in which Maher was confirmed to be at home and never left the residence. On other days, she left the house — either as a passenger or driving herself — to run a limited number of errands, mostly picking up or dropping off her children at school, and once to go to dinner at a restaurant. She was also observed outside on two brief occasions involved in what could generously be described as yard chores: carrying a flower pot and, on another occasion, sitting in the bed of her husband’s truck holding a broom or rake while her husband appeared to be cleaning up. All of this activity is consistent with her description of a low level of activity dependent on the ebb and flow of her symptoms. It would be unusual for a mother of three children to be able to avoid all activity.
With regard to the kite-flying episode, which strays the farthest from Maher’s reported limitations, Maher has indicated that the outing was a “special event” for which she premedicated with morphine. In other circumstances, this explanation might strain credulity. Here, however, the notion that Maher premedicates to prepare for activities that may trigger pain finds support in records that predate the incident.17 In a March 2003 activity questionnaire, for example, Maher noted that she travels by plane only with “pre-medication for pain and nausea from increased cabin pressure on abdomen.” In her September 2006 questionnaire, Maher also noted that her ability to carry out various activities of daily life “always depend[s] on how much pain medicine I use ... to help myself.” Moreover, the entire outing at the athletic fields was very brief, lasting just over half an hour. It would be unfair to read too much into one short episode of increased activity, given the consistency of the larger record of surveillance with Maher’s reported capabilities.
Turning to an evaluation of the medical opinions concerning Maher’s limitations, my conclusion again diverges from that of the Plan and its doctors. Among Maher’s treating doctors, there is thin support for her capacity to return to a sedentary job. The most direct evidence is found in a questionnaire, completed by Dr. Gale Hay-dock, indicating that Maher is “OK to per-
*306form sedentary duties.” However, Dr. Haydock treated Maher only once, in the winter of 2006, when Maher was admitted to the hospital for several days to treat a flare-up of abdominal pain, and thus Dr. Haydock had no opportunity to observe the course of Maher’s symptoms over time.
One could also, as the Plan has, read various statements by Maher’s primary care physician (Dr. Cuevas) to support Maher’s ability to perform sedentary work. Most notably, in a conversation with Dr. Malinoff, Dr. Cuevas stated her agreement with Dr. MalinofPs opinion that, “from a purely internal medicine perspective, there is no identifiable physical exam or anatomic/laboratory abnormality that would prevent [Maher] from working at a very minimum at a sedentary level.” This awkwardly precise statement is technically true and is, undoubtedly, an accurate reflection of Dr. Cuevas’s medical opinion.18 It is also transparently misleading. Maher’s medical records make clear that no doctor has been able to identify a physical or anatomic abnormality that causes her symptoms. However, the absence of a diagnosed medical condition says nothing about the reliability of Maher’s complaints or whether her reported symptoms prevent her from working. On those questions, Dr. Cuevas’s opinion is unequivocal. In a letter dated March 2007, Dr. Cuevas wrote that Maher “remains in significant disability, both from her chronic pain and from the side effects the pain medication cause,” and that she “is unable to reliably perform duties because her pain can become so severe so quickly.”
Lastly, the record also contains opinions from the three doctors retained by the Plan — Dr. Millstein, Dr. Malinoff, and Dr. Dean Hashimoto, Chief of Occupational and Environmental Medicine at MGH— concluding that the available evidence does not support Maher’s claimed inability to work a sedentary job. I find the opinions rendered by these doctors unpersuasive. Each doctor relied to a significant degree on the surveillance records in evaluating Maher’s capabilities, focusing on the episodes of activity detailed above and finding them inconsistent with Maher’s claims. Dr. Malinoff and Dr. Hashimoto also interpreted Dr. Cuevas’s statements as supportive of Maher’s ability to return to work. For the reasons stated, I have reached contrary conclusions based on the same evidence. Additionally, Dr. Hashimoto observed that, even accepting the veracity of Maher’s reported symptoms, there has been no attempt to evaluate “through neuropsych testing, scans, or other available means” the extent to which Maher’s pain and use of narcotics affect her cognition and ability to function. This failing can be attributed, to some degree, to the Plan’s decision not to pursue an independent medical evaluation of Maher to assist in the assessment of her claim.19 *307Nevertheless, as I discuss below, I find the record evidence of Maher’s limitations sufficient even absent the sort of testing suggested by Dr. Hashimoto.
2. The Evidence Supports Maher’s Limitations
Viewing the totality of the medical evidence in the administrative record, I am persuaded that Maher’s symptoms prevent her from reliably performing the duties of a sedentary nursing job. At the fore of that evidence are the opinions of Maher’s treating doctors, Dr. Cuevas and Dr. Goessling. As noted, Dr. Cuevas’s assessment as of March 2007 was that Maher “remains in significant disability” and “is unable to reliably perform duties because her pain can become so severe so quickly.” Similarly, Dr. Goessling, who has followed Maher since the onset of her abdominal symptoms in late 2001, opined in a 2007 letter:
In her current status, Mrs. Maher is barely able to provide for herself and her 3-year-old son during the day. She is not able to stand or walk for prolonged periods of time. She is suffering from constant nausea that is only partially relieved by her ... medication. She has intermittent diarrhea due to malabsorption from lack of pancreatic enzymes followed by constipation caused by her high doses of narcotics medication. On top of her chronic abdominal pain, she has frequent exacerbations, [and] often this pain prohibit[s] her eompletely from taking any solid foods.... [¶] ... [L]et me assure that I do not see any way that my patient would be able to sit or stand for prolonged periodfs] of time let alone do physically or intellectually demanding work.
While there is some evidence that Dr. Goessling did not actively treat Maher in 2006 and 2007, he saw her repeatedly in the preceding years and she appears to have consulted with him prior to his writing the letter quoted above.20
These opinions echo Maher’s own assessment of her limitations. In an affidavit, Maher stated that she “cannot be counted on to do anything,” because her symptoms come on unpredictably and leave her in “excruciating pain” that “is so bad that it sucks the wind out of [her].” Though pain is subjective and thus difficult to reliably document, her characterization appears to be borne out by the record. From late 2001, she has consistently complained of intermittent and severe abdominal pain. Her complaints have been credible enough to convince the numerous doctors who have seen her that she needs serious narcotics to relieve her pain and allow her to function. While we might suspect drug-seeking tendencies in such circumstances, the record does not reveal such tendencies. An early note from Dr. Goessling indicates that Maher was “quite reluctant to take pain medications,” and there are multiple indications in later records of her desire to move off of the painkillers.21 Even with her regular *308regimen of heavy narcotics, Maher’s abdominal pain has repeatedly brought her to the emergency room, where she was admitted on at least two occasions for multiple-day stays to manage her pain.
Maher’s record of treatment thus bespeaks significant and debilitating pain. Given the number of medical professionals who have examined her and found her distress genuine, I have no reason to question the reality of this pain. As the Seventh Circuit noted in Carradine v. Barnhart, 360 F.3d 751 (7th Cir.2004):
What is significant is the improbability that [the claimant] would have undergone the pain-treatment procedures that she did, which included ... heavy doses of strong drugs ..., merely in order to strengthen the credibility of her complaints of pain and so increase her chances of obtaining disability benefits; likewise the improbability that she is a good enough actress to fool a host of doctors and emergency-room personnel into thinking she suffers extreme pain; and the (perhaps lesser) improbability that this host of medical workers would prescribe drugs and other treatment for her if they thought she were faking her symptoms. Such an inference would amount to an accusation that the medical workers who treated [the claimant] were behaving unprofessionally.
Id. at 755 (internal citation omitted). I therefore credit Maher’s reports of abdominal pain, and note as well that her gastrointestinal and food intolerance symptoms — which are more readily verified— find support in numerous records.
I similarly find the evidence sufficient to corroborate Maher’s claims that these symptoms would interfere with her ability to work. Maher’s recurring acute attacks of abdominal pain would, at a minimum, result in frequent absences from work, which would be prohibitively disruptive of any attempt to maintain regular employment. Surveillance also suggests that her background level of symptoms is sufficient to keep her housebound with some frequency, or to permit only limited levels of activity. Though Maher may occasionally run errands, contribute to household chores, or even recreate with her family for short periods of time, there is a sharp “difference between a person’s being able to engage in sporadic physical activities and her being able to work eight hours a day five consecutive days of the week.” Id. On balance, I conclude that the evidence demonstrates that Maher cannot reliably perform the duties of a full-time sedentary nursing job.
III.
Based on a close, de novo review of the administrative record, I am persuaded that Maher’s abdominal pain and related symptoms effectively prevent her “from performing any and every duty of any occupation or employment, for which [she] is reasonably qualified by education, training or experience.” Hence, I believe we must go beyond vacating the district court’s grant of summary judgment in favor of the Plan, and remand for entry of judgment in Maher’s favor. I therefore dissent.
. To the extent the majority is suggesting that the Plan’s argument for deferential review is strengthened by this alternative theory — i.e., that MGH and Partners are affiliated entities between whom delegation is unnecessary— the suggestion is gratuitous given that the Plan did not offer the theory. Moreover, if the Plan, a sophisticated litigant, thought such a view of the entities’ relationship was viable, it presumably would have made the claim to avoid relying solely on an approach that even the majority considers justified "perhaps only by a modest margin.”
. The Plan contains another potentially relevant provision, Article 6.4, which provides authority for the Hospital to "delegate to the Claims Processor responsibility for certifying the information and amount necessary for the proper payment of claims from the Fund under the provisions of the Plan.” The Plan does not contend, however, that the Hospital’s authority to delegate its discretionary authority to Partners arises from this provision. This is because the delegation authorized in Article 6.4 is a limited one relating to claims processing duties, and, as discussed below, the authority that may be delegated to the Claims Processor under the Plan falls short of the sort of administrative authority that will trigger deferential review.
. Here, for instance, Liberty (acting as claims processor) made the initial benefits determination and processed Maher's first appeal, but, upon Maher's second and final appeal, Partners made the ultimate determination to deny.
. In relying on Pettaway v. Teachers Insurance and Annuity Ass’n of America, 644 F.3d 427 (D.C.Cir.2011), for the proposition that decisional authority to determine benefits can be inferred from plan language, the majority fails to take seriously the "clear statement” standard in our case law. Whether or not the language in Pettaway would meet that standard, the language here does not. The Pettaway court described the relevant plan provisions as follows:
As stated in the Plan Document, the "Academy shall be the Plan Administrator and the ‘Named Fiduciary'" with the "absolute power, authority and discretion to administer the [Academy] Plan.” Plan Document at 3.1, 3.2. "All interpretations of the Plan, and questions concerning its administration and application, shall be determined” by the Academy, which has the authority to "appoint such accountants, counsel, specialists, and other persons as it deems necessary or desirable in connection with the administration of the Plan.” [Plan Document] at 3.2.
Pettaway, 644 F.3d at 434. Thus, unlike the MGH Plan, delegation under the plan in Pettaway was not limited to "agents” whom the administrator would “direct.” See supra pp. 297-98. Instead, the Pettaway administrator had broad authority to delegate administrative responsibility to others whenever such a transfer of duties was "deem[ed] necessary or desirable.”
The limited language in the MGH Plan renders irrelevant the majority's observation that lawyers, accountants and actuaries may be assigned fiduciary duties. The issue here is not whether such professionals are considered fiduciaries, but whether this Plan clearly endorsed delegation of the authority to make final benefits determinations.
. As Maher points out, the Partners Plan Document never accurately refers to the Plan. In a table detailing the various subsidiary plans to which it applies, the Partners Plan Document lists an "MGH LTD Insurance Plan.” The Plan is not, in fact, an insurance plan, as it is funded through a trust, and thus the reference to an “LTD Insurance Plan” is erroneous.
. Maher argues at some length that the Partners Plan Document fails to include all categories of information required by statute to be present in an SPD. See 29 U.S.C. § 1022(b). The Plan counters that the Partners Plan Document explicitly states that it is only one of several documents that “collectively constitute the summary plan description for each Plan,” implying that those additional documents supply the missing information.
. The majority fails to confront the inconsistency between the documents and, indeed, cites the Partners Plan Document as evidence that the Hospital was authorized to delegate its discretionary authority. But a bare assertion of authority by the putative delegee is even less meaningful as proof that delegation was authorized than it is as evidence that a delegation in fact occurred. In effect, the majority assumes that Partners’ assertion that it had authority is evidence that it was given that authority. Such circular reasoning is both illogical and unpersuasive.
. Because I dispose of Maher's argument based on the definition of "impaired,” I do not address her contention that the sedentary nursing roles identified by Liberty involve the "practice” of nursing under Massachusetts law.
. Dr. Goessling, a trained gastroenterologist and Associate Professor of Medicine at Harvard Medical School, continues to believe that Maher’s abdominal symptoms are most likely caused by chronic pancreatitis. The Plan's doctors disagree, but concede that the record supports a diagnosis of either "chronic pain syndrome with abdominal focus," per Dr. Malinoff, or fibromyalgia, per Dr. Millstein.
. This included six days of surveillance in 2002, three days in 2005, and ten days in 2006.
. The district court concluded that, assuming that Maher’s outing with her family can be explained by premedication, "it is a reasonable inference that she could also premedicate to perform a sedentary job.” I cannot agree. Maher takes a large amount of narcotics daily to address her background pain. The fact that she can, on top of this background dosage, take additional pain medications to ward off pain during the occasional short episodes of increased activity does not suggest that it would be feasible for her to regularly take extra medication to make it through an eight-hour workday.
. Following their conversation, Dr. Malinoff mailed a letter to Dr. Cuevas that recapitulated the substance of their conversation and asked that she sign to verify its accuracy. She did so and returned the letter to Dr. Malinoff.
. Indeed, I find the Plan's failure to conduct an independent medical examination somewhat troubling. There is no requirement that a plan administrator arrange for a medical examination prior to terminating benefits, see Orndorf, 404 F.3d at 526, but here the circumstances certainly should have suggested its utility. As early as 2004, Dr. Millstein counseled that "[i]f it is felt to be important to ascertain whether impairment due to abdominal pain exists, I would suggest consideration of functional assessment by some alternative means.” After her benefits had been denied, Maher even offered to make herself available for a physical examination by a doctor of the Plan’s choice. The record reflects that the Plan's administrators internally discussed the possibility of an independent medical exami*307nation in September 2007, but declined to pursue one due, in part, to concern for slowing down the process.
. The letter, addressed to Maher's case manager at Liberty, begins, “I would like to update you on [Maher’s] overall condition, especially in light of the recent denial letter for her benefits that she received,” implicitly suggesting that Dr. Goessling had current knowledge of Maher's condition at the time of writing.
. Still, the record is mixed as to the sincerity of Maher's desire to discontinue narcotic use, as she has twice started treatment with a pain clinic and then failed to follow up. She ascribes her reluctance to continue treatment at the clinics to interpersonal conflict with the doctors at one clinic, and a feeling that the type of program offered by the other clinic *308was not appropriate for her. Her lack of follow-through in this regard does not diminish the overall force of the evidence of her pain.