FILED
United States Court of Appeals
UNITED STATES COURT OF APPEALS Tenth Circuit
FOR THE TENTH CIRCUIT July 15, 2019
_________________________________
Elisabeth A. Shumaker
Clerk of Court
MARY D.,
Plaintiff - Appellant,
v. No. 17-4195
(D.C. No. 1:16-CV-00124-DB)
ANTHEM BLUE CROSS BLUE SHIELD; (D. Utah)
ANTHEM UM SERVICES, INC.;
CREDIT SUISSE SECURITIES (USA)
LLC GROUP HEALTH CARE PLAN,
Defendants - Appellees.
_________________________________
ORDER AND JUDGMENT*
_________________________________
Before LUCERO, McHUGH, and MORITZ, Circuit Judges.
_________________________________
Asserting her rights under the Employee Retirement Income Security Act
(ERISA) of 1974, 29 U.S.C. §§ 1001–1461, plaintiff Mary D. (M.D.) seeks recovery
of residential-treatment benefits for her son, A.D., from Credit Suisse Securities
(USA) LLC Group Health Care Plan (the Plan). The district court reviewed the Plan’s
denial of benefits under an arbitrary-and-capricious standard and granted summary
judgment in favor of the Plan and its claims administrator, Anthem UM Services, a
*
This order and judgment isn’t binding precedent, except under the doctrines
of law of the case, res judicata, and collateral estoppel. But it may be cited for its
persuasive value. Fed. R. App. P. 32.1; 10th Cir. R. 32.1.
subsidiary fully owned by Anthem Blue Cross and Blue Shield (collectively,
Anthem). For the reasons discussed below, we affirm.
Background
Because this case involves the denial of healthcare benefits—specifically, the
denial of coverage for residential treatment of A.D.’s psychiatric disorder—we begin
by explaining the terms of the Plan and some of A.D.’s medical history. Further,
because this case comes to us after several levels of administrative review and
because the standard of review is at issue, we also detail what occurred during the
administrative review process and before the district court.
I. The Plan
When the events relevant to this appeal transpired, M.D. worked for Credit
Suisse Securities. Credit Suisse sponsored the Plan, which is a self-funded employee
welfare benefits plan. M.D. elected to participate in the Plan and named her
dependent son, A.D., as a beneficiary.
The Credit Suisse Benefits Committee (the Benefits Committee) has “sole and
complete discretionary authority” to administer the Plan. App. vol. 2C, 1264
(sealed).1 But it delegated part of that authority to Anthem. As a result, Anthem has
certain administrative responsibilities under the Plan: it determines who is eligible to
1
Via our September 14, 2018 and October 30, 2018 orders, we provisionally
granted M.D.’s September 11, 2018 motion to seal volumes 2 through 2K of the
appendix. To the extent we quote from these sealed volumes, we have determined the
quoted material either appears in the parties’ briefs, which the parties did not file
under seal, or isn’t sensitive. Cf. Fed. R. Civ. P. 5.2(h) (stating party waives privacy
protection for own information by filing not under seal).
2
participate in the Plan, decides whether an individual will receive benefits, and
interprets the terms of the Plan. Nevertheless, the Benefits Committee has the
“ultimate responsibility” of administering the Plan. App. vol. 2B, 1056 (sealed).
As relevant here, the Plan covers certain treatments for psychiatric disorders.
But to be covered, any treatment—including treatment for such psychiatric
disorders—must be medically necessary. And the Plan provides distinct
medical-necessity criteria for different levels of psychiatric care.2 One category of
medical-necessity criteria considers the severity of the relevant illness, evaluating the
“condition and circumstances” of the individual seeking coverage. App. vol. 2, 461
(sealed). And the severity-of-illness criteria for residential treatment requires the
individual seeking coverage to demonstrate, among other things, the following:
1. The [individual] is manifesting symptoms and behaviors [that]
represent a deterioration from [his or her] usual status and include
either self[-]injurious or risk[-]taking behaviors that risk serious
harm and cannot be managed outside of a 24[-]hour structured
setting or other appropriate outpatient setting; AND
2. The social environment is characterized by temporary stressors or
limitations that would undermine treatment that could potentially be
improved with treatment while the [individual] is in the residential
facility; AND
3. There should be a reasonable expectation that the illness,
condition[,] or level of functioning will be stabilized and improved
2
The two levels of care relevant to the issues on appeal are residential
treatment and acute inpatient treatment. Residential treatment is “specialized
treatment that occurs in a residential[-]treatment center. Residential treatment is 24
hours per day and requires a minimum of one physician visit per week in a facility[-]
based setting.” App. vol. 2, 467 (sealed). Acute inpatient treatment, on the other
hand, is “treatment in a hospital psychiatric unit that includes 24-hour nursing and
daily active treatment under the direction of a psychiatrist.” Id. (sealed).
3
and that a short[-]term subacute residential[-]treatment service will
have a likely benefit on the behaviors/symptoms that required this
level of care, and that the [individual] will be able to return to
outpatient treatment.
Id. at 463 (sealed).
The Plan also recommends that before receiving any particular treatment, the
individual should request a pretreatment review from Anthem to “make sure the
charges are medically necessary.” App. vol. 2C, 1324 (sealed). But the Plan doesn’t
penalize the individual for failing to request pretreatment review. Indeed, the Plan
expressly allows the individual to seek medical-necessity review after treatment.
Nevertheless, the Plan cautions that if the individual seeks a retrospective review, he
or she “run[s] the risk of reduced or denied benefits if the claims administrator finds
that the care” received “is not medically necessary.” Id. (sealed). If the individual
disagrees with such a finding, the Plan provides level-one and level-two appeals.
Anthem conducts level-one appeals. The Benefits Committee conducts level-two
appeals.
II. A.D.’s Medical History and Treatment
A.D. was first diagnosed with generalized anxiety disorder at age nine. His
anxiety manifested in various ways: procrastination, perfectionism, rigidity,
academic difficulty, and isolationism. In elementary school, he received outpatient
treatment from a psychologist.
A.D.’s anxiety escalated over time. In 2013, during his freshman year of high
school, A.D. threatened to harm himself with a knife. As a result, he received two
4
weeks of inpatient treatment at a hospital, nearly two weeks of outpatient treatment,
and then some additional inpatient treatment. After the second round of inpatient
treatment, the hospital suggested A.D. be placed in residential treatment. Despite this
recommendation, and with guidance from an education consultant, M.D. enrolled
A.D. in a wilderness-therapy program called Aspiro.3
A.D. stayed at Aspiro for nine and a half weeks. During that time, psychologist
Todd Corelli observed that A.D. had difficulty dealing with his emotions and lacked
useful coping skills. Corelli reported that A.D. was often angry, particularly toward
his parents. Corelli’s observations aligned with A.D.’s father’s statement to Corelli
that it was hard to manage A.D. at home. Corelli also noted that although A.D. had
experienced suicidal thoughts in the past, A.D. reported he hadn’t been suicidal in
three to four months.
At the end of A.D.’s treatment at Aspiro, Corelli reported that A.D. had
benefited from the program. But he recommended sending A.D. to a residential-
treatment program, which could provide therapy and support 24 hours a day, so that
A.D. could further internalize the improvements he’d made. A.D.’s therapist at
Aspiro, Jamie Kaczmarek, echoed Corelli’s recommendation for residential
treatment. She suggested that without the structure and support of a residential-
treatment program, A.D.’s prior difficulties might resurface. Following these
recommendations, M.D. sent A.D. to a residential-treatment center called Daniels
3
The Plan’s terms specify that “[w]ilderness programs” like Aspiro “are not
considered residential[-]treatment programs.” App. vol. 2, 467 (sealed).
5
Academy. She later moved A.D. from Daniels Academy to another residential-
treatment center called WayPoint Academy (WayPoint).
WayPoint admitted A.D. based on A.D.’s anxiety disorder and its disruption to
his daily life. At intake, WayPoint therapist Brett Walker conducted a suicide-risk
screening. He noted that despite previous suicidal ideation, A.D. wasn’t currently
suicidal and hadn’t been suicidal for several months. A.D. denied being suicidal at
least twice more while at WayPoint. He remained at WayPoint for ten months.
III. Administrative Review Process
A. Initial Denial and Level-One Appeal
M.D. didn’t request a pretreatment review for medical necessity before
WayPoint admitted A.D. for treatment; Anthem first processed a claim for benefits
after A.D. had been at WayPoint for several months. Anthem issued an Explanation
of Benefits denying benefits because “EITHER NO PRECERT[IFICATION] WAS
OBTAINED OR ADDITIONAL INFORMATION HAS BEEN REQUESTED.” App.
vol. 2, 318 (sealed).
In response, M.D. invoked the administrative review process, submitting a
request for either a retrospective benefits determination or a level-one appeal. Along
with the request, M.D. submitted invoices from WayPoint and A.D.’s medical
records to allow Anthem to “conduct a retrospective review of all of his claims.” Id.
at 315 (sealed). M.D. also asked Anthem to provide her with “specific detailed
reasons for [any further] denial.” Id. at 316 (sealed).
6
As requested, Anthem conducted a retrospective review, which included a
review by psychiatrist Narsimha Muddasani. Muddasani noted that after numerous
failed attempts to contact WayPoint, he completed his medical-necessity review
based on clinical notes in A.D.’s medical records. He found that A.D. had anxiety
and poor family interaction; didn’t appear actively homicidal or suicidal; wasn’t
psychotic or medically unstable; wasn’t placed on strict suicide watch; and could
have been effectively treated in a lower level of care. Ultimately, after applying the
medical-necessity criteria for residential treatment, Muddasani concluded that such
treatment wasn’t medically necessary.
Based on Muddasani’s review, Anthem denied benefits. Anthem issued an
initial denial letter, stating in relevant part:
The information your provider gave us does not show that [residential
treatment] is medically necessary. You have not caused serious harm to
anyone. You have not harmed yourself to such a degree that has caused
serious medical problems. You have not had recent treatment for this in
a structured outpatient program. You are also likely to benefit from
structured outpatient treatment. We based this decision on the
health[-]plan guidelines (Psychiatric Disorder Treatment—Residential
Treatment Center (RTC) (CG-BEH-03)[)].
App. vol. 2K, 3502 (sealed). Anthem stated M.D. could “request, free of charge,
reasonable access to, and copies of all documents, records, and other information,
including the . . . medical[-]necessity criteria that were used in making this decision.”
Id. at 3503 (sealed).
7
M.D. then sought a level-one appeal.4 In her appeal letter, M.D. provided a
detailed medical history for A.D. She further provided clinical documentation
regarding his treatment history and the recommendations for residential treatment
from Corelli and Kaczmarek. M.D. requested a “full, fair, and thorough” level-one
review and an explanation from Anthem “regarding the amount of weight [it] gave
the clinical evidence . . . when making [its] determinations.” App. vol. 2D, 1519
(sealed).
As part of the level-one appeal, Anthem had another psychiatrist, Rasik Lal,
conduct an additional medical-necessity review. Lal reviewed M.D.’s letters,
Muddasani’s findings, Rudo’s findings, the initial denial letter, and A.D.’s medical
records. Lal found A.D. wasn’t actively suicidal or homicidal when admitted to
WayPoint; didn’t behave in a combative or psychotic manner; didn’t require
continuous monitoring; wasn’t dangerous to himself or others; and didn’t present
physiological problems or instabilities requiring residential treatment. Thus, like
Muddasani before him, Lal concluded that the residential treatment wasn’t medically
necessary.
4
WayPoint also filed its own level-one appeal with Anthem. As part of that
separate appeal, Anthem requested that psychiatrist Andrew Rudo review A.D.’s
medical records and apply the medical-necessity criteria for residential treatment.
Rudo determined A.D.’s residential treatment at WayPoint wasn’t medically
necessary, so Anthem denied WayPoint’s appeal.
Although the outcome of WayPoint’s separate appeal isn’t before us, we
mention Rudo’s review because a later reviewer and the district court both relied in
part on his assessment.
8
Relying on Lal’s review, Anthem denied M.D.’s level-one appeal. Anthem
stated that in resolving the appeal, it considered all the information that it received in
connection with both the retrospective review and this level-one appeal. It further
stated:
Your doctor wanted you to have residential[-]treatment[-]center care.
The reason we were given for this was that you were at risk for serious
harm without 24[-]hour care. We understand that you would like us to
change our first decision. Now we have new information from the
medical records and appeal letter from your mother. We still do not
think this was medically necessary for you. We believe our first
decision was correct for the following reason: You were not at risk for
serious harm [such] that you needed 24[-]hour care. You could have
been treated with outpatient services. We based this decision on this
health[-]plan guideline. (Psychiatric Disorder Treatment—Residential
Treatment Center (RTC) (CG-BEH-03)).
Id. at 1507 (sealed). Anthem also offered, upon request, “copies of all documents[,]
including the actual benefit provision, guideline, protocol, or other similar criterion
on which this decision was based.” Id. at 1509 (sealed).
B. Level-Two Appeal
After Anthem denied her level-one appeal, M.D. pursued a level-two appeal
before the Benefits Committee. In her appeal letter, M.D. lodged a number of
objections to Anthem’s retrospective review and level-one appeal. M.D. first pointed
out that despite her request, Anthem didn’t specify the weight it gave to the treating
physicians’ opinions or the medical records. She further suggested that Anthem and
its reviewers hadn’t reviewed A.D.’s medical records at all. And she identified
information in those medical records that, according to her, demonstrated that
treatment at WayPoint was indeed medically necessary under the Plan’s residential-
9
treatment criteria. M.D. next complained that the initial denial letter and level-one
appeal denial letter only stated “conclusions” and didn’t make “any specific
references to the medical records.” Id. at 1497 (sealed). She requested that the
Benefits Committee base its review on the residential-treatment criteria. M.D. also
argued that Anthem’s reviewers had denial rates that “demonstrate[d] a lack of
consideration for what is best for the member.” Id. (sealed). In support, she pointed
to a transcript from a news story that alleged Anthem’s reviewers had denial rates
from 95 to 100 percent. M.D. asked the Benefits Committee to “tell [her] the denial
rates of the last three years for [Lal] as well as [Muddasani].” Id. (sealed).
As part of the level-two appeal, the Benefits Committee requested that a
physician from the Medical Review Institute of America (MRIoA) analyze whether
A.D.’s treatment at WayPoint was medically necessary. The MRIoA physician
reviewed the level-two appeal procedures, the residential-treatment criteria,
WayPoint’s claims, M.D.’s letters, A.D.’s medical records, the prior medical
reviewers’ determinations, and Anthem’s letters. The MRIoA physician initially
addressed whether, “[p]er [the American Medical Association] or another
professional standard of care,” A.D.’s “inpatient confinement in a [residential-
treatment center was] supported by” his diagnosis or condition. App. vol. 2K, 3525
(sealed). In answering that question, the MRIoA physician found there was no
evidence that A.D. presented an imminent risk of harm to himself or others; he
demonstrated no psychotic symptoms; and he exhibited no evidence of deteriorating
function. As such, the MRIoA physician determined that “the available
10
documentation [did] not support a severity of symptoms that would require
residential treatment.” Id. at 3526 (sealed).
The MRIoA physician next addressed whether, “[b]ased on medical records,
. . . the patient was at risk for serious harm that constituted [a need for] 24[-]hour
care.” Id. (sealed). As to that question, the MRIoA physician, consistent with Lal and
Muddasani, found that A.D. didn’t demonstrate a risk of harm to himself or others
that warranted 24-hour care. Thus, the MRIoA physician concluded that A.D.’s
clinical presentation didn’t indicate that residential treatment was medically
necessary.
Relying on the MRIoA physician’s report, the information M.D. submitted
with her appeal letter, the medical records that WayPoint submitted to Anthem, and
“the relevant [P]lan provisions,” the Benefits Committee denied M.D.’s level-two
appeal. Id. at 3561 (sealed). The Benefits Committee stated that its “determination
was in agreement with Anthem’s view that [A.D.’s] clinical presentation did not
support the use of residential-treatment care during” his time at WayPoint. Id.
(sealed). The Benefits Committee also informed M.D. that she had “the right to
receive, upon request and without charge, reasonable access to or copies of any
relevant documents, records, or other information relied upon” in the level-two
appeal. Id. (sealed).
IV. District-Court Proceedings
After exhausting the administrative review process, M.D. filed this lawsuit in
district court seeking recovery of benefits. See § 1132(a)(1)(B). The parties filed
11
cross-motions for summary judgment. In ruling on these cross-motions, the district
court first determined that because the Plan gave Anthem and the Benefits Committee
discretionary authority and because “there were no procedural deficiencies . . . that
r[o]se to the level of lessening the standard of review,” the arbitrary-and-capricious
standard of review applied. App. vol. 1, 223. Second, the district court determined
that the opinions of the four reviewing physicians supported the denial of benefits
and that the denial wasn’t arbitrary and capricious. Thus, the district court denied
M.D.’s motion for summary judgment and granted summary judgment to Anthem and
the Plan. M.D. now appeals.
Analysis
M.D. asserts the district court erred by applying the wrong standard of review
when it assessed Anthem and the Plan’s decision to deny benefits. Specifically, M.D.
alleges that rather than deferring to Anthem and the Plan’s decision by asking
whether it was arbitrary and capricious, the district court should have reviewed
Anthem and the Plan’s decision de novo.
For the reasons discussed below, we reject this argument and hold that the
district court did not err in reviewing Anthem and the Plan’s decision under the
deferential arbitrary-and-capricious standard. See infra Section I. And we further
conclude that even assuming the district court did err in applying this more
deferential standard of review, that error was harmless; Anthem and the Plan’s
decision to deny benefits survives even de novo review. See infra Section II.
12
I. The Applicable Standard of Review
M.D. contends that the district court erred in applying a deferential standard of
review to Anthem and the Plan’s decision to deny benefits. “We review de novo the
‘district court’s determination of the proper standard to apply in its review of an
ERISA plan administrator’s decision . . . .’” Rasenack ex rel. Tribolet v. AIG Life Ins.
Co., 585 F.3d 1311, 1315 (10th Cir. 2009) (alteration in original) (quoting DeGrado
v. Jefferson Pilot Fin. Ins. Co., 451 F.3d 1161, 1167 (10th Cir. 2006)).
A district court applies de novo review in a benefits-eligibility case “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 (1989) (emphasis added). Where the plan gives the
fiduciary or administrator discretionary authority, the district court “employ[s] a
deferential standard of review, asking only whether the denial of benefits was
arbitrary and capricious.” Eugene S. v. Horizon Blue Cross Blue Shield of N.J., 663
F.3d 1124, 1130 (10th Cir. 2011) (quoting LaAsmar v. Phelps Dodge Corp. Life,
Accidental Death & Dismemberment & Dependent Life Ins. Plan, 605 F.3d 789, 796
(10th Cir. 2010)). But even where the plan affords such discretionary authority to the
fiduciary or administrator, deferential review isn’t guaranteed: in the face of
procedural irregularities in the administrative review process, a district court will
instead review the benefits denial de novo. See LaAsmar, 605 F.3d at 797; Gilbertson
v. Allied Signal, Inc., 328 F.3d 625, 631 (10th Cir. 2003).
13
In this case, M.D. concedes that the Plan’s terms give the Benefits Committee
and Anthem complete discretion to administer it. Thus, the district court would have
ordinarily reviewed Anthem and the Benefits Committee’s decision for arbitrariness.
Firestone, 489 U.S. at 115. Nevertheless, M.D. argues that four procedural
irregularities in the administrative review process should have triggered de novo
review. Specifically, she asserts that (1) Anthem shifted its basis for denying the
claim; (2) Anthem and the Benefits Committee failed to engage in a meaningful
dialogue with her; (3) the Benefits Committee failed to respond to her request for
specific information; and (4) Anthem and the Benefits Committee used the incorrect
medical-necessity criteria. We address each assertion in turn.
First, M.D. alleges that Anthem based its denial on “shifting” and
“inconsistent” reasons. Aplt. Br. 34. In support, she points out that Anthem initially
indicated in the Explanation of Benefits (EOB) that it denied coverage either because
M.D. failed to obtain precertification or because Anthem had requested but hadn’t yet
received certain additional information. Yet Anthem’s later denials, M.D. points out,
turned on the lack of medical necessity.
But M.D. doesn’t explain how these two responses are inconsistent. Critically,
the denial in the EOB didn’t turn solely on the absence of precertification;5 instead, it
pointed out that in the absence of precertification, Anthem needed additional
5
Indeed, precertification couldn’t be the sole basis for the denial in the EOB;
the Plan’s terms specifically state that there is no penalty for failing to obtain
precertification.
14
information to process the claim. That is, the EOB simply sought additional
information that would allow Anthem to conduct a retrospective review of whether
the treatment was medically necessary and therefore covered under the Plan’s terms.
Once Anthem conducted that retrospective review using the additional information
that M.D. provided—as M.D. specifically asked it to do—it concluded that
residential treatment wasn’t medically necessary. And after that review, Anthem
consistently denied benefits based on medical necessity. Accordingly, Anthem’s
reasons for denying coverage didn’t shift over time, and we reject M.D.’s first
procedural-irregularity argument.
M.D. next argues that Anthem and the Benefits Committee failed to engage in
a “meaningful dialogue” with her during the administrative process. Aplt. Br. 38. The
meaningful-dialogue requirement stems from subsections (g) and (h) of 29 C.F.R.
§ 2560.503-1.6 See Metzger v. UNUM Life Ins. Co. of Am., 476 F.3d 1161, 1168 n.4
(10th Cir. 2007) (stating that requirements in § 2560.503-1(g) and (h) “enable
claimants to submit informed responses to the adverse decision and to engage in
meaningful dialogue with the plan administrator”); Gilbertson, 328 F.3d at 635
(stating that ERISA and its regulations contemplate “meaningful dialogue” between
plan administrators and claimants (quoting Booton v. Lockheed Med. Benefit Plan,
6
Technically, subsection (g) applies only to notifications of benefits and
subsection (h) applies only to appeal denials. See Hancock v. Metro. Life Ins. Co.,
590 F.3d 1141, 1153 (10th Cir. 2009). But in this case, the Plan’s terms make most of
the provisions of subsection (g) applicable to notifications of benefits and appeal
denials, rendering this technical distinction irrelevant.
15
110 F.3d 1461, 1463 (9th Cir. 1997))). Subsection (g), as incorporated by the Plan,
requires in part that any notice of denial must (1) provide the specific reason for the
adverse determination, (2) reference the specific provision warranting denial, and
(3) for medical-necessity denials, explain the scientific or clinical judgment
supporting the determination. Here, the initial denial letter, the level-one appeal
denial letter, and the level-two appeal denial letter met these requirements: they cited
lack of medical necessity as the specific reason for each denial; they referenced the
residential-treatment criteria that governed the medical-necessity determination; and
they provided clinical judgment supporting each denial. Thus, none of the letters
violate the requirements of subsection (g).
We next turn to subsection (h). That subsection requires every plan to provide
claimants “a reasonable opportunity to appeal[,] . . . under which there will be a full
and fair review of the claim and the adverse benefit determination.” § 2560.503-
1(h)(1). Full and fair reviews must, among other things, “take[] into account all
comments, documents, records, and other information submitted by the claimant
relating to the claim,” and provide “reasonable access to, and copies of, all
documents, records, and other information relevant to the claimant’s claim for
benefits.” § 2560.503-1(h)(2)(iii), (iv). Additionally, in deciding an appeal based on
medical necessity, the plan must “consult with a health[]care professional who has
appropriate training and experience in the field of medicine involved in the medical
judgment.” § 2560.503-1(h)(3)(iii).
16
Here, both the level-one and level-two appeal denial letters show that Anthem
and the Benefits Committee provided a full and fair review. For instance, Anthem
consulted Lal during the level-one appeal, and the Benefits Committee consulted an
MRIoA physician during the level-two appeal. Anthem, the Benefits Committee, and
the medical reviewers all stated they considered the letters and records that M.D.
submitted. And although M.D. asserts that she requested responses to the materials
and arguments she submitted, she doesn’t cite any authority—nor are we aware of
any—that required Anthem and the Benefits Committee to affirmatively respond to
these submissions. Instead, subsection (h) merely required Anthem and the Benefits
Committee to “take[]” these materials and arguments “into account.”7 § 2560.503-
1(h)(2)(iv). Moreover, Anthem and the Benefits Committee stated in both appeal
denial letters that M.D. could have access to the records they relied on in making the
benefits determination free of charge. As such, M.D. hasn’t demonstrated that
Anthem or the Benefits Committee violated subsections (g) or (h) or failed to engage
7
M.D. cites to Gaither v. Aetna Life Insurance Co., 394 F.3d 792, 802 (10th
Cir. 2004), for the proposition that the Tenth Circuit requires “more” than what she
characterizes as a “reassuring pat[] on the head” that the plan administrator
“considered all of the materials” she submitted. Aplt. Br. 39. But Gaither—a case
involving the denial of long-term disability benefits, rather than the denial of medical
benefits—is inapposite. 394 F.3d at 795. Compare § 2560.503-1(g)(1)(v) (requiring
denial of medical benefits to include “explanation of the scientific or clinical
judgment for the determination”), with § 2560.503-1(g)(1)(vii) (requiring denial of
disability benefits to include both “explanation of the scientific or clinical judgment
for the determination” and basis for disagreeing with or not following “views
presented by the claimant to the plan of health[]care professionals treating the
claimant and vocational professionals who evaluated the claimant”).
17
in a meaningful dialogue at the notification-of-benefits or appeal stages. So we find
no procedural irregularity on this basis.
In a related procedural-irregularity argument, M.D. next asserts that the
Benefits Committee failed to provide her with certain information during her level-
two appeal—specifically, “the denial rates of the last three years” for Lal and
Muddasani. App. vol. 2D, 1497 (sealed). In support, M.D. points out that the
regulations entitle a claimant to “reasonable access to, and copies of, all documents,
records, and other information relevant to the claimant’s claim for benefits.”
§ 2560.503-1(h)(2)(iii) (emphasis added). Information is relevant to a claim for
benefits if it “[d]emonstrates [the administrator’s] compliance with the administrative
processes and safeguards required pursuant to paragraph (b)(5).” § 2560.503-
1(m)(8)(iii). And paragraph (b)(5), in turn, requires the administrator to put
procedures in place to ensure both that “benefit claim determinations are made in
accordance with governing plan documents” and that the plan is “applied consistently
with respect to similarly situated claimants.” § 2560.503-1(b)(5).
According to M.D., the denial rates are relevant to her claim for benefits
because they could demonstrate whether Anthem complied with procedures that
ensure the Plan is properly and consistently applied. To begin, she offers a
conditional reason for why the denial rates may show Anthem didn’t properly apply
the Plan’s terms: they “may” show high denial rates for the reviewers. Aplt. Br. 38.
In making this argument, M.D. seems to be conflating the broad concept of relevance
under the Federal Rules of Evidence with the narrow regulatory definition of
18
relevance in subsection (m)(8)(iii). But these are distinctly different standards of
relevance. Compare Fed. R. Evid. 401 (stating that evidence is relevant if “it has any
tendency to make a fact more or less probable than it would be without the
evidence”), with § 2560.503-1(m)(8)(iii) (stating that information is relevant if it
“[d]emonstrates [the administrator’s] compliance with the administrative processes
and safeguards”). And M.D. provides no authority to support her assertion that the
denial rates are relevant under the narrower regulatory provision merely because they
might show Anthem’s lack of compliance with safeguards and procedures. Nor have
we found any. The regulation states that relevant information is that which in fact
demonstrates compliance, and M.D. fails to explain how the denial rates meet that
standard. We therefore reject M.D.’s argument that the denial rates are relevant to
show proper application of the Plan.
Next, M.D. argues that if the denial rates for Lal and Muddasani are
substantially different, the disparity may show inconsistent application of the Plan to
similarly situated claimants. As an initial matter, this argument suffers from the same
infirmity as M.D.’s prior point—it relies on an overbroad definition of relevance.
Further, as Anthem and the Plan point out, Lal’s and Muddasani’s denial rates
wouldn’t, standing alone, demonstrate that Anthem applied the Plan’s terms
inconsistently to similarly situated claimants. That’s because the denial rates
wouldn’t shed any light on whether the denied claimants were in fact similarly
situated: for instance, denial rates wouldn’t show whether the denied claimants were
the same age as A.D., whether they were diagnosed with generalized anxiety
19
disorder, or whether they were seeking coverage for residential treatment. Thus, we
also reject M.D.’s assertion that the denial rates are relevant to show inconsistent
application of the Plan among similarly situated claimants.
In her fourth and final procedural-irregularity argument, M.D. alleges that the
medical reviewers for Anthem and the Benefits Committee erroneously used the
medical-necessity criteria for acute inpatient care rather than the criteria for
residential treatment. For acute inpatient care—which is a higher degree of care than
residential treatment—the criteria are more onerous. In particular, the acute-
inpatient-care criteria require, among other things, “[i]mminent suicidal risk or
danger to others.” App. vol. 2, 461 (sealed). But for residential treatment, the criteria
require only “self[-]injurious or risk[-]taking behaviors that risk serious harm.” Id. at
463 (sealed).
In support for her position that the reviewers relied on the former rather than
the latter, M.D. points to the reviewers’ findings on A.D.’s risk for suicide and
homicide, alleging that those findings mirror the criteria for acute inpatient care, not
residential treatment.
As an initial matter, we note that Lal, Muddasani, and Anthem all specified the
residential-treatment criteria, not the acute-inpatient-care criteria, as the basis for
their determination. What’s more, none of the reviewers even mentioned the acute-
inpatient-care criteria. Neither did Anthem or the Benefits Committee. We see no
reason to think these entities identified one set of criteria but applied another.
Further, as the district court explained, the criteria for acute inpatient care and
20
residential treatment partially overlap. That is, although Lal, Muddasani, and the
MRIoA reviewer implicitly addressed the acute-inpatient-care criteria by finding that
A.D. wasn’t suicidal or at imminent risk of self-harm when WayPoint admitted him,
these findings were also relevant to determining whether, for purposes of applying
the residential-treatment criteria, he was engaging in self-injurious or risk-taking
behavior that risked serious harm. For instance, a finding that A.D. had repeatedly
attempted suicide would necessarily compel the conclusion that he was engaged in
self-injurious behavior. Likewise, a finding that A.D. had recently attempted
homicide would necessarily compel the conclusion that he was engaged in behavior
that risked serious harm to others. Thus, the reviewers’ reliance on their findings
about the existence or nonexistence of A.D.’s homicidal and suicidal ideations
doesn’t indicate that they used the incorrect medical-necessity standard.
Moreover, the issue of whether residential treatment is medically necessary
turns on more than just risk of harm; it also requires a finding that the individual’s
disorder can’t be properly “managed outside of a 24[-]hour structured setting or other
appropriate outpatient setting.” App. vol. 2, 463 (sealed). And Lal, Muddasani, and
the MRIoA physician found A.D. failed to meet this criterion because he didn’t need
continuous monitoring and treatment. Likewise, Anthem mentioned A.D.’s failure to
satisfy this criterion in the initial denial letter and in the level-one appeal denial
letter. Notably, this consideration—whether the disorder can be managed in
outpatient care—isn’t part of the acute-inpatient-care criteria. Thus, we reject M.D.’s
21
argument that Anthem and the Benefits Committee procedurally erred by applying
the incorrect medical-necessity criteria.
In short, M.D. fails to identify any procedural irregularities in the
administrative review process.8 Accordingly, the district court did not err in applying
the arbitrary-and-capricious standard of review.
II. Denial of Residential-Treatment Benefits
Alternatively, even assuming we agreed with M.D. and concluded that the
district court erred in applying the arbitrary-and-capricious standard of review, that
conclusion wouldn’t necessarily entitle M.D. to relief. Instead, this conclusion would
only permit us to reverse the district court’s decision outright if M.D. could show that
the district court’s error prejudiced her—i.e., that it can be reasonably concluded the
district court would have reversed the denial of benefits had it reviewed the denial
decision de novo. See Fed. R. Civ. P. 61 (“At every stage of the proceeding, the court
must disregard all errors and defects that do not affect any party’s substantial
rights.”); Shinseki v. Sanders, 556 U.S. 396, 410 (2009) (noting that in civil cases,
“the party seeking reversal normally must explain why the erroneous ruling caused
8
Because we find no procedural irregularities in the administrative review
process, we need not address M.D.’s argument that the substantial-compliance
doctrine no longer applies under the 2002 version of the ERISA procedural
regulations. See Gilbertson, 328 F.3d at 634–35 (defining substantial-compliance
doctrine as court’s willingness to “overlook administrators’ failure to meet certain
procedural requirements” when administrator “substantially complied with the
regulations”); Rasensack, 585 F.3d at 1316 (“The 2002 amendments have . . . called
into question the continuing validity of the substantial[-]compliance [doctrine]
. . . .”).
22
harm”). For the reasons discussed below, we conclude that M.D. cannot make this
showing here; even under de novo review, we see no indication that Anthem and the
Plan wrongly denied benefits. Thus, even assuming the district court erred in failing
to review Anthem and the Plan’s decision de novo, that error was harmless.
In performing this prejudice analysis, we review Anthem and the Plan’s
decision to deny benefits, “as opposed to reviewing the district court’s ruling.”
Holcomb v. Unum Life Ins. Co. of Am., 578 F.3d 1187, 1192 (10th Cir. 2009). In so
doing, “[o]ur review is ‘limited to the administrative record—the materials compiled
by the administrator in the course of making [its] decision.’” Id. (quoting Fought v.
UNUM Life Ins. Co. of Am., 379 F.3d 997, 1003 (10th Cir. 2004)).
As stated above, A.D.’s residential treatment was only covered if it was
medically necessary. And under the Plan’s terms, residential treatment for a
psychiatric disorder is only medically necessary if the individual meets three criteria.
First, the individual must manifest deterioration from his or her usual status and
demonstrate self-injurious or risk-taking behaviors that risk serious harm and cannot
be managed outside of a 24-hour structured setting. We refer to these three
requirements collectively as the injury-risk criterion and separately as the
deterioration requirement, the behaviors requirement, and the 24-hour requirement.
Second, the individual’s social environment must have temporary stressors or
limitations that could undermine treatment (the social-environment criterion). Third,
the individual must reasonably be expected to both stabilize and improve from short-
term residential treatment and then return to outpatient treatment. We refer to these
23
two requirements collectively as the reasonable-expectation criterion and separately
as the stabilization requirement and the return-to-outpatient-care requirement. M.D.
argues that she can meet her burden to establish that A.D. met these criteria and that
Anthem and the Plan therefore wrongly denied her claim for benefits. See Rasenack,
585 F.3d at 1324.9
1. The Injury-Risk Criterion
In attempting to show A.D. satisfied the behaviors requirement (i.e., that he
displayed self-injurious or risk-taking behaviors that posed a risk of serious harm),
M.D. relies exclusively on the opinions of treating medical providers Corelli and
Kaczmarek. Corelli and Kaczmarek recommended that A.D. obtain residential
treatment after he completed Aspiro’s wilderness-therapy program. But ERISA
doesn’t require an administrator to defer to a treating physician’s opinion. See Black
& Decker Disability Plan v. Nord, 538 U.S. 822, 825 (2003) (“We hold that [under
ERISA,] plan administrators are not obliged to accord special deference to the
opinions of treating physicians.”). Instead, ERISA merely prohibits administrators
from “arbitrarily refus[ing] to credit . . . the opinions of a treating physician.” Id. at
834. Thus, Anthem and the Benefits Committee weren’t required to defer to Corelli
or Kaczmarek in the face of other credible medical evidence. And they possessed
9
M.D. also faults Anthem and the Benefits Committee for failing to discuss all
three of these criteria. But the residential-treatment criteria are conjunctive. That is,
the absence of any one criterion would preclude a finding of medical necessity. Thus,
because Anthem and the Benefits Committee found that A.D. failed to satisfy the
injury-risk criterion, they were not required to address the social-environment
criterion or the reasonable-expectation criterion.
24
other credible evidence in this case, in the opinions from Lal, Muddasani, and the
MRIoA reviewer. See Blair v. Alcatel-Lucent Long Term Disability Plan, 688 F.
App’x 568, 576 (10th Cir. 2017) (unpublished) (finding plan administrator wasn’t
required to defer to treating physician where credible evidence from file reviewer and
primary-care physician supported denial of coverage).
Undeterred, M.D. also argues that because A.D.’s condition involves mental
health, “[i]t is especially improper to ignore the findings and conclusions of a
patient’s treating physician.” Aplt. Br. 45. In support, she relies on numerous Sixth
Circuit cases that question the reliability of file reviews in benefits disputes related to
the treatment of psychiatric disorders. See, e.g., Okuno v. Reliance Standard Life Ins.
Co., 836 F.3d 600, 611–12 (6th Cir. 2016) (“File reviews are particularly
‘questionable as a basis’ for an administrator’s determination to deny benefits where
the claim, as here, involves a mental[-]illness component.” (quoting Javery v. Lucent
Techs., Inc. Long Term Disability Plan for Mgmt. or LBA Emps., 741 F.3d 686, 702
(6th Cir. 2014))); Smith v. Bayer Corp. Long Term Disability Plan, 275 F. App’x
495, 505–09 (6th Cir. 2008) (unpublished) (“Courts discount the opinions of
psychiatrists who have never seen the patient for obvious reasons.” (quoting Sheehan
v. Metro. Life Ins. Co., 368 F. Supp. 2d 228, 255 (S.D.N.Y. 2005))). But we have
uncovered no similar caselaw in our circuit, and M.D. points to none. In fact, we
have expressly rejected the argument that administrators must defer to treating
physicians in the mental-health context. See Eugene S., 663 F.3d at 1134–35
(rejecting appellant’s argument that opinion of treating physician is entitled to special
25
weight in ERISA benefits dispute related to mental-health treatment). Thus, we aren’t
persuaded that we must depart from the general rule that we need not “accord special
deference to the opinions of treating physicians.” Nord, 538 U.S. at 825; see also
Eugene S., 663 F.3d at 1135 (declining plaintiff’s “invitation to announce a
‘treating[-]physician rule’ for ERISA claims relating to mental[-]health care”).
Further, even if we were to credit the treating medical providers’ opinions,
those opinions don’t necessarily contradict the conclusion that A.D. wasn’t engaging
in self-injurious or risk-taking behaviors when WayPoint admitted him. At Aspiro,
Corelli noted that although A.D. had suicidal thoughts in the past, A.D. hadn’t been
suicidal in three or four months. Likewise, once A.D. entered WayPoint, Walker
conducted a suicide-risk screening and determined that A.D. hadn’t thought of
suicide in the previous few months and wasn’t currently suicidal or threatening self-
harm. And A.D. denied being suicidal at least twice more while at WayPoint. The
only real evidence of self-injurious or risk-taking behavior was A.D.’s suicide threat
in February 2013—but that occurred approximately six months before he entered
WayPoint. Further, and critically, although Corelli and Kaczmarek recommended
residential treatment after Aspiro, they did so based on their belief that A.D. needed
to internalize the improvements he made at Aspiro, not because he was engaged in
self-injurious or risk-taking behavior. Accordingly, even assuming Anthem and the
Benefits Committee should have deferred to the treating providers, we see no
26
indication that they would have reached a different medical-necessity conclusion had
they done so.10
Moreover, contrary to M.D.’s argument that residential treatment was
medically necessary because A.D. was engaging in self-injurious and risk-taking
behaviors, considerable credible evidence supports the opposite conclusion.
Muddasani determined A.D. didn’t appear actively homicidal or suicidal; wasn’t
psychotic or medically unstable; wasn’t on strict suicide watch; and could have been
treated with a lower level of care. Lal found A.D. wasn’t actively suicidal or
homicidal when WayPoint admitted him; didn’t behave in a combative or psychotic
manner; didn’t require continuous monitoring; wasn’t dangerous to himself or others;
and didn’t present physiological problems or instabilities requiring residential
treatment care. And the MRIoA reviewer stated A.D. didn’t demonstrate a risk of
harm to himself or others warranting 24-hour care. Also, each reviewer
independently concluded A.D. could have been treated with a lower level of care. In
turn, based on the reviewers’ findings, Anthem and the Benefits Committee
concluded that A.D. wasn’t engaging in self-injurious or risk-taking behaviors. In
10
In M.D.’s appeal letters, she included additional excerpts from A.D.’s
medical records that, according to M.D., support a finding that residential treatment
was medically necessary. She doesn’t explicitly identify many of these excerpts in
her briefing. But we have reviewed the excerpts in her appeal letters and find nothing
that would indicate A.D. was displaying self-injurious or risk-taking behaviors that
couldn’t be managed in an outpatient setting. And even assuming such evidence
exists elsewhere in the record, “[n]othing in ERISA requires plan administrators”—or
reviewing courts, for that matter—“to go fishing for evidence” that a claimant
doesn’t “br[ing] to their attention.” Rasenack, 585 F.3d at 1325 (quoting Gaither, 394
F.3d at 804).
27
sum, then, regardless of whether we credit the treating physicians or the medical
reviewers, M.D. hasn’t demonstrated that A.D. met this criterion when he entered
WayPoint.
Nor has M.D. shown that A.D. met the other part of this criterion—i.e., that he
“manifest[ed] symptoms and behaviors [that] represent[ed] a deterioration from [his]
usual status”—prior to entering WayPoint. App. vol. 2, 463 (sealed). On this
deterioration requirement, M.D. urges us to consider A.D.’s condition not just at the
time he entered WayPoint, but “from a more broad[-]spectrum perspective.” Reply
Br. 22. M.D. doesn’t cite any authority for her broad-spectrum point. Nor have we
found any. Moreover, the record simply doesn’t show that A.D.’s condition was
deteriorating at the time he entered WayPoint or in the preceding months. Instead, the
record shows that A.D. had been improving. Prior to attending WayPoint, A.D.
attended Aspiro and Daniels Academy. Corelli found that A.D. significantly
benefited from his time at Aspiro. Kaczmarek likewise noted A.D.’s improvements
after Aspiro. Thus, although A.D.’s condition may have deteriorated before attending
Aspiro, as M.D. posits, we cannot say that his condition deteriorated before he
entered WayPoint.
In sum, we find that the record presents no evidence that A.D. was
deteriorating or engaging in self-injurious or risk-taking behavior that couldn’t be
managed except in a 24-hour structured setting. He therefore fails to satisfy the
injury-risk criterion. And because the three medical-necessity criteria for residential
treatment are conjunctive, we reject M.D.’s argument that A.D.’s residential
28
treatment was medically necessary. But in the interest of completeness, we briefly
address M.D.’s arguments on the remaining two criteria.
2. The Social-Environment Criterion
M.D. next argues that “it is evident that [A.D.] was subjected to temporary
stressors or limitations at home that made removal from that social environment
necessary.” Aplt. Br. 55. Notably, she cites nothing in the record to support this
assertion. And although our independent review of the record indicates this may have
been true before A.D. attended Aspiro, we see no indication this was still the case by
the time WayPoint admitted him. Instead, Corelli indicated that A.D. had benefited
from Aspiro. Likewise, Kaczmarek noted that upon his discharge from Aspiro, A.D.
had stabilized. And because A.D. never went home between leaving Aspiro and
entering WayPoint, there’s simply no way to know whether his home would have
been a temporary stressor or limitation that would have undermined his treatment.
Under these circumstances, we conclude that M.D. has failed to demonstrate that
when A.D. entered WayPoint, his family remained a “temporary stressor or
limitation[]” that would undermine his treatment. App. vol. 2, 463 (sealed); see also
Rasenack, 585 F.3d at 1324.
3. The Reasonable-Expectation Criterion
In addressing this third criterion, M.D. points to an absence in the record of
any finding that A.D. failed to satisfy this criterion, rather than to the presence of any
29
record evidence that might support a contrary finding.11 But the absence of such a
finding isn’t relevant; it’s M.D.’s burden to show that A.D.’s residential treatment
was medically necessary, not the administrator’s burden to show they determined it
wasn’t. See Rasenack, 585 F.3d at 1324. And M.D. fails to satisfy that burden here.
Conclusion
For the reasons outlined above, we conclude that the district court didn’t err
when it reviewed the benefits denial for arbitrariness. Alternatively, even assuming
the district court should have reviewed the denial decision de novo, we conclude that
the district court’s error in failing to do so was harmless; the denial decision survives
even de novo review because A.D.’s residential treatment wasn’t medically
necessary.12 Accordingly, we affirm the district court’s order entering summary
judgment in favor of Anthem and the Benefits Committee. As a final matter, we grant
M.D.’s motion to seal.
Entered for the Court
Nancy L. Moritz
Circuit Judge
11
M.D. insists she “provided” such “evidence” to Anthem during the review
process. Rep. Br. 23. But she neither identifies the specific evidence she provided nor
directs our attention to the location in the record where that evidence appears. Cf.
Fed. R. App. P. 28(a)(8)(A) (requiring appellant to provide “citations to . . . parts of
the record” relied on).
12
Likewise, because M.D.’s wrongful-denial argument fails even under de
novo review, we need not separately address whether, as she asserts, the decision to
deny benefits was arbitrary and capricious. See Nance v. Sun Life Assurance Co. of
Can., 294 F.3d 1263, 1266 (10th Cir. 2002) (finding coverage decision that survived
“stringent” de novo review necessarily couldn’t be arbitrary and capricious).
30