Berg v. UNUM Life Insurance Company of America ( 2023 )

UNITED STATES DISTRICT COURT EASTERN DISTRICT OF MICHIGAN SOUTHERN DIVISION PAULA BERG, 2:21-CV-11737-TGB-DRG Plaintiff, HON. TERRENCE G. BERG vs. ORDER RESOLVING CROSS UNUM LIFE INSURANCE MOTIONS FOR JUDGMENT COMPANY OF AMERICA, ON THE RECORD (ECF NOS. 14, 15) Defendant. In this disability insurance coverage dispute, all parties agree that Dr. Paula Berg1 is disabled from her occupation as a practicing anesthe- siologist and a hospital’s Director of Cardiothoracic Anesthesia, and that she is eligible for disability benefits. The question is for how long— according to the terms of a long-term disability plan issued by Unum Life Insurance Company of America (“Unum”) and governed by the Em- ployee Retirement Income Security Act (“ERISA”). Unum maintains that a psychological condition caused Dr. Berg’s disability, and thus her claim is subject to a 12-month cap on benefits. Dr. Berg contends that her disability is due to a physical, rather than psychological, condition and that she is therefore eligible for benefits paid up to the 48-month 1 Of no known relation to the undersigned. cap that applies to ordinary disability claims made by 63-year-old claimants. To prevail, Unum has the burden of proving that Dr. Berg’s disability arises from a psychological condition. For the reasons ex- plained below, Unum has not met that burden. Accordingly, Plaintiff’s Cross-Motion for Judgment on the Record is GRANTED. Defendant’s Motion for Summary Judgment is DENIED, and this case is DIS- MISSED with prejudice. I. BACKGROUND a. The Plan’s relevant terms Dr. Berg participated in a long term disability plan issued by Un- um (“the Plan”). Administrative Record, ECF No. 11-2, PageID.717. The plan, which is governed by 29 U.S.C. §§ 1001 et seq., defines “disability” as follows: You are disabled when Unum determines that due to your sickness or injury: During the first 12 months of disability, you are unable to perform the material and substantial duties of your regular occupation and after the first 12 months of disability, you are unable to perform the material and substantial duties of your regular occupation and you are not working in your regular occupation or any other occupa- tion. Id. at PageID.733. The Plan also imposes a “lifetime cumulative maximum benefit period” of twelve months for “all disabilities due to mental illness.” Id. at PageID.742. The Plan defines mental illness as a psychiatric or psychological condition classified in the Diagnostic and Statistical Manual of Men- tal Health Disorders (DSM), published by the American Psychiatric Association, most current as of the start of a disability. Such disorders in- clude, but are not limited to, psychotic, emotional or behavioral disorders, or disorders relatable to stress. Id. at PageID.754. b. Dr. Berg’s pre-claim medical history Plaintiff Dr. Paula Berg is a 67-year-old former physician. She worked as an attending anesthesiologist at Genesys Regional Medical Center for approximately 15 years until leaving work after being diag- nosed with breast cancer in January, 2019. As Director of Cardiothorac- ic Anesthesia and a practicing anesthesiologist, Berg had significant and challenging clinical and administrative responsibilities. ECF No. 12-3, PageID.2005. In September 2018—about three months before she was diagnosed with cancer—Berg began therapy with Lauri Keller, LMSW, a social worker. ECF No. 12-2, PageID.1775. Keller diagnosed Berg with Gener- alized Anxiety Disorder. Id. at PageID.1776. The “main theme” of Berg’s twice-monthly sessions with Keller was Berg’s interpersonal relation- ships, particularly with family members. Id. at PageID.1776-85. Keller’s notes at each session describe Berg’s “cognitive functioning” as “Orient- ed/Alert” and “functional status” as “Intact.” Id. On November 28, 2018, shortly before her cancer diagnosis, Keller remarked that Berg’s func- tioning was “reasonably good with only short lived and expectable reac- tions to everyday stressful events.” Id. at PageID.1785. c. Dr. Berg is diagnosed with cancer and stops working On January 3, 2019, Dr. Berg stopped working on medical advice due to an “invasive ductal carcinoma” in her right breast. ECF No. 11-3, PageID.1151. Berg underwent surgery to remove her cancer a month later. Id. Surgery was followed by radiation therapy beginning in May, and hormone therapy thereafter. ECF No. 12-3, PageID.2005. As part of her hormone treatment, Berg was prescribed anastrozole, an “aroma- tase inhibitor.” Id.2 Throughout her cancer treatment, Berg continued her regular therapy sessions. Keller’s contemporaneous notes continue to focus en- tirely on interpersonal relationships until January 9, 2019, when Berg reported that she had been diagnosed with cancer. ECF No. 12-3, Page- ID.1795. Keller remarked that Berg’s “concentration and focus” were impaired because she was “pre-occupied with [her cancer] diagnosis.” Id. Keller’s notes from that visit again describe Berg as “Oriented/Alert” with “Intact” functional status and an “Appropriate” affect. Id. Keller entered substantially the same observations at later appointments in 2 Aromatase inhibitors block the activity of an enzyme the body uses to make estrogen, which lowers a patient’s estrogen levels. They can pre- vent the growth of cancer tissue and keep cancer from returning. See aromatase inhibitor, National Cancer Institute Dictionary of Cancer Terms, (last visited Mar. 21, 2023) [https://perma.cc/N5PK-EAPJ]. January. By February, 1, 2019, Keller’s notes do not reflect any difficul- ties with “concentration and focus,” and appear once again to focus pri- marily on Berg’s interpersonal relationships. Id. at PageID.1801-5. Notes from visits in March reflect a concern on Berg’s part that her job requires “deep concentration and focus,” areas in which Berg re- ported having difficulty because of preoccupation with her cancer treatment. Id. at PageID.1807-11. Keller noted that Berg was “Unable to work . . . due to the affects [sic] and issues related to having cancer.” Id. at PageID.1811. Berg also reported being “very fatigued” from her cancer treatment. Id. at PageID.1815. At a July 3, 2019 visit, Keller ap- pears to have diagnosed Berg with “Adjustment Disorder, With mixed anxiety and depressed mood.” Id. at PageID.1817. At that visit, Berg al- so reported mood swings caused by anastrozole. Id. d. Dr. Berg’s claim history In March, 2019, Dr. Berg’s counsel filed claims for disability bene- fits under the Plan and two individual disability plans also adminis- tered by Unum.3 See ECF No. 11-1, PageID.456. On April 4, 2019, Un- um approved Dr. Berg’s claims under all three policies. ECF No. 11-3, PageID.1150. 3 Dr. Berg’s claims under the two individual disability plans were the subject of a separate lawsuit that has since been resolved. See Compl., ECF No. 1-1, Berg v. Provident Life and Accident Insurance Company, 22-cv-11486 (E.D. Mich. Jun. 30, 2022) 1. Unum denies Dr. Berg’s long term disability claims Unum paid benefits to Dr. Berg for about a year until, in April, 2020, Unum reviewed Berg’s claims. Dr. Joseph Antaki, an Unum- affiliated medical consultant, consulted Berg’s treating physicians. Dr. Li Ding, Berg’s oncologist, explained that although her cancer was in remission, the anastrozole Berg was prescribed to prevent a cancer re- currence was affecting her cognition. ECF No. 12-3, PageID.1897-98. Dr. Ding further stated that anastrozole was known to cause memory and concentration issues. Id. at PageID.1898. Although according to Dr. Ding those effects are “usually mild,” and in Berg’s case did not warrant changing to a potentially less effective medicine, it was Dr. Ding’s opin- ion that even “mild” cognitive effects would prevent Berg from practic- ing medicine as an anesthesiologist. Id. Dr. Sean Smith, another treating physician, explained that alt- hough Berg’s cancer was in remission, he “would still support re- strictions/limitations” on her work “due to cognitive dysfunction espe- cially given the nature of her occupation.” ECF No. 12-3, PageID.1925 Dr. Smith further indicated that he planned to assess Berg’s cognitive and memory issues in greater depth, but was prevented from doing so by restrictions on patient contact imposed by COVID-19. Id. Dr. Sherrie Tefend, Berg’s primary care doctor, opined that Berg was still disabled and indicated that Berg had not tried any alternatives to anastrozole because anastrozole offered her the “best likelihood of remission.” Id. at PageID.1915-16. After communicating with Doctors Ding, Smith, and Tefend, Dr. Antaki issued a report. Id. at PageID.1931-33. He concluded that Berg’s claimed limitations were not supported. He pointed to Dr. Ding’s state- ments that aromatase inhibitors like anastrozole are “generally . . . well tolerated,” and that a different medication could be tried if anastrozole posed problems. Id. at PageID.1932. Dr. Antaki also highlighted that, at a visit in January, 2020, Dr. Ding observed that Berg was experiencing only “mild” side effect from the medications and had an “ECOG perfor- mance score of ‘0.’”4 Id. at PageID.1888, 1932. A second report from that visit—described by Berg as a corrected report—reflects an ECOG score of “1” which characterizes Berg’s functional ability as “no physically strenuous activity, but ambulatory and able to carry out light or seden- tary work.” Id. at PageID.1888, 2030; Pl’s. Mot. for Summ. J., ECF No. 14, PageID.2678. Later that month, Dr. Herbert Dean, an oncologist and Unum medical consultant, reviewed Berg’s file and Dr. Antaki’s report and agreed that Berg was not disabled by cancer or her cancer treatments. Id. at PageID.1937-39. Dr. Dean noted that Berg’s cancer was in remis- 4 An ECOG score describes a patient’s ability to care for themselves and conduct daily activities. A score of zero indicates that a patient is “fully active” and able to carry on pre-disease activities without restriction. sion, and concluded that another aromatase inhibitor could be tried if anastrozole was interfering with her functioning. Id. at PageID.1938- 39. Dr. Dean also opined that the cognitive deficits Berg complained of could be explained by a behavioral health condition, but that she had not undergone “formal neurocognitive testing” or a “Mini-Mental Status Exam.” Id. at PageID.1939. In his report, Dr. Dean also stated that Berg had returned to work after completing her cancer treatment, which was not accurate Id. at PageID.1938, 1947. On April 28, 2020, Unum denied Berg’s long term disability claims. Id. at PageID.1953-61. The letter, which relied primarily on the reports of Doctors Antaki and Dean, explained that Unum had conclud- ed that Berg’s symptoms no longer prevented her from returning to work as an anesthesiologist. Id. at PageID.1954. 2. Dr. Berg appeals Unum’s decision On July 5, Dr. Berg appealed Unum’s benefits decision. Id. at PageID.2004. With her appeal, Berg included a report by Dr. Nicolette Gabel, a psychologist who had examined her and administered cognitive tests. Id. at PageID.2010-14. The tests revealed “executive dysfunction and problems with sustained vigilance [] consistent with [Berg’s] self- reported difficulties.” Id. at PageID.2013. Dr. Gabel also concluded that Berg’s self-reported symptoms and answers to standard surveys “did not indicate significant depression or anxiety.” Id. Gabel noted that Berg reported a “sudden change when starting aromatase inhibitor therapy.” Id. at PageID.2011. Ultimately, Dr. Gabel concluded that the causes of Berg’s cogni- tive changes were “not entirely clear.” Id. at PageID.2013. Dr. Gabel opined—based on her review of studies on the cognitive effects of aro- matase inhibitors and cancer more generally—that “cancer [itself] con- tributed to at least a small to moderate degree,” though cancer “would not be expected to account for the severe degree of presumed changes in executive functioning” that Berg demonstrated. Id. According to Dr. Gabel, other contributing factors were likely “sleep disruption” and the sustained use of lorazepam, a benzodiazepine-class drug Berg took for insomnia. Id. Dr. Gabel’s suggested treatments included cognitive be- havioral therapy for insomnia, the possible discontinuation of loraze- pam, and “behavioral intervention for cancer related cognitive decline.” Id. As to the latter, Dr. Gabel noted that, due to COVID-19, the “Cancer Rehab Cognitive Disorders Clinic [was] not yet open.” Id. Berg also provided a note from Dr. Tefend indicating that Berg’s oncologist had decided to continue her on anastrozole, the “culprit medi- cation,” because the oncologist “believed that this medication affords [Berg] the least likelihood of developing a recurrence of breast cancer.” Id. at PageID.2033. In August, 2020, two more Unum-affiliated doctors reviewed Berg’s file on appeal. The first, Dr. Chris Bartlett, concluded that Berg was not cognitively disabled by her cancer treatment. Dr. Bartlett then referred the matter to another Unum medical consultant: Dr. Peter Brown, a psychiatrist. Dr. Brown issued a second report, in which he concluded that work restrictions due to any “general medical diagnosis,” including “cognitive side effects of medication” were not supported by Berg’s medical records. ECF No. 13-1, PageID.2146-47. But Dr. Brown agreed that Berg’s symptoms due to a psychological condition were se- vere enough to support restrictions on her ability to work. Id. at Page- ID.2144-47. Unum provided Berg with a copy of Brown’s report, noted that the policy included a 12-month limit on benefits for disabilities due to mental illness, and invited Berg’s response. Id. at PageID.2265. On September 10, Berg responded to Brown’s report. ECF No. 13- 2, PageID.2315-19. Unum referred that response to Dr. Brown, who maintained his original conclusion. Id. at PageID.2322. Dr. Brown, rely- ing heavily on the report of Dr. Gabel, characterized Dr. Gabel’s treat- ment recommendations as “psychiatric treatments for a psychiatric dis- order,” and opined that, although Berg’s file showed cognitive symp- toms, neither he nor Dr. Gabel found “support for the presence of a neu- rocognitive disorder.” Id. In response, Berg submitted a letter from Dr. Gabel. Id. at Page- ID.2333. Dr. Gabel clarified her findings, and was adamant that “it is more likely than not that cancer and related treatment caused” Berg’s cognitive issues, but that other factors— such as insomnia and Berg’s use of benzodiazepines to treat that insomnia—might be exacerbating those problems. Id. at PageID.2336 (emphasis omitted). Dr. Gabel also opined that “psychiatric symptoms” could not “account for the changes” and that there was “no objective evidence” of depression or anxiety con- tributing to Berg’s condition. Id. After reviewing Dr. Gabel’s letter, Dr. Brown agreed that Berg had “significant cognitive symptoms,” and that those symptoms were “cancer related,” but maintained that there was “no reasonable support for an assertion that [Berg’s] cognitive symptoms [were] a physiological result of breast cancer or . . . her treatment” for cancer. Id. at Page- ID.2346. Dr. Brown explained that, in his view, Berg’s symptoms were best understood as the result of a psychiatric condition. Id. at Page- ID.2346-47. Dr. Brown opined that “improved mood but persistent diffi- culties with cognition and/or motivation is the single most common pat- tern” in patients with chronic depression and chronic anxiety, and that Berg’s symptoms were in keeping with that pattern. Id. at PageID.2347. Unum sent a copy of Dr. Brown’s supplemental report to Berg, and again invited her response. Id. at PageID.2352-53. On October 21, 2020, Berg sent Unum a report by Dr. Juan Lopez, a consulting psychiatrist who reviewed her medical history, interviewed her twice, and conducted tests. Id. at PageID.2383. Dr. Lopez concluded that Berg did not meet the diagnostic criteria for generalized anxiety disorder before her cancer diagnosis (contrary to her therapist’s conclu- sion at that time), but had met the criteria for an Adjustment Disorder that had since resolved. He further concluded that Berg’s symptoms both before and after her cancer diagnosis were inconsistent with a di- agnosis of chronic depression or a major depressive disorder. Id. at PageID.2393-94. Dr. Lopez responded to Dr. Brown’s arguments, and ultimately opined that Berg’s symptoms were not the result of a psychi- atric condition. Id. at PageID.2397. Dr. Lopez also noted that Berg had stopped taking lorazepam and had seen sleep improvements, but that her symptoms persisted. Id. at PageID.2396-97. Based on that fact and on his review of medical studies, Dr. Lopez concluded that neither in- somnia nor lorazepam use were responsible for Berg’s cognitive symp- toms. Id. Dr. Brown reviewed Dr. Lopez’s report and opined (1) that Berg’s “disabling symptoms clearly began prior to her cancer treatment,” (2) that her symptoms were “best understood as a residual psychiatric con- dition” in light of the facts that she was in psychotherapy treatment be- fore her cancer, (3) that “persistent cognitive complaints are found in a substantial proportion of breast cancer patients independent of the na- ture of the treatment received,” (4) that Berg’s symptoms were con- sistent with psychiatric impairment, and (5) that Dr. Gabel had recom- mended treatments that seemed directed to psychological causes, such as a reduction in benzodiazepine use and treatments for insomnia. Id. at PageID.2435-36. Dr. Jacqueline Crawford, a neurologist retained by Unum, also re- viewed Berg’s file. She noted that research on the effects of aromatase inhibitors had been “mixed and complex,” and further concluded that aromatase inhibitors were an unlikely culprit. Id. at PageID.2438. This was because Berg was under the care of a behavioral health profession- al and was treated with benzodiazepines months before her breast can- cer diagnosis, and because Dr. Ding indicated that any aromatase inhib- itor side effects “may get better over time.” Id. In November, 2020, Dr. Lopez submitted another rebuttal to the opinions of Doctors Brown and Crawford. ECF No. 13-3, PageID.2490- 93. After interviewing her a second time, Dr. Lopez maintained that Berg was not suffering from any psychiatric conditions that would ex- plain her symptoms, and that she did not meet the diagnostic criteria for depression or any mood or anxiety disorder. Id. at PageID.2491. He reiterated that Berg had discontinued the use of lorazepam and that her insomnia had largely resolved, but that her cognitive problems persist- ed. Id. He further opined, for the first time, that there was another po- tential reason for Berg’s symptoms: prior to her diagnosis, she had been receiving long-term hormone replacement therapy, but was forced to abruptly discontinue it upon her cancer diagnosis. Id. at PageID.2492- 93. Relying on studies examining the effect of estrogen on cognition, Dr. Lopez concluded that Berg’s symptoms were consistent with the abrupt loss of estrogen caused by discontinuation of her hormone replacement therapy and compounded by use of an aromatase inhibitor—which blocks the production of estrogen. Id. Doctors Brown and Crawford reviewed the letter from Dr. Lopez, and declined to change their conclusions, noting that the cognitive ef- fects of aromatase inhibitors “is a topic of intense research.” Id. at Page- ID.2497-2500. On December 1, 2020, Unum informed Berg’s counsel that it had reviewed the report of Dr. Lopez as well as updated records from Dr. Gabel, but maintained its earlier position. Id. at PageID.2507. Berg then requested that Unum make an immediate decision on her appeal. Id. at PageID.2526. On December 9, 2020, Unum denied Berg’s benefits claim on appeal. Id. at PageID.2531. II. STANDARD OF REVIEW “Ordinarily, a plan administrator’s denial-of-benefits decision is reviewed de novo.” Clemons v. Norton Healthcare Inc. Ret. Plan, 890 F.3d 254, 264 (6th Cir. 2018) (citing Firestone Tire & Rubber Co. v. Bruch, 489 U.S. 101, 115 (1989)). Arbitrary-and-capricious review ap- plies if a plan “gives the administrator or fiduciary discretionary au- thority to determine eligibility for benefits or to construe the terms of the plan,” Id. at 264. The Sixth Circuit has explained that “discretion is the exception, not the rule,” and that the de novo standard applies ab- sent a “clear grant of discretion to determine benefits or interpret the plan.” Wulf v. Quantum Chem. Corp., 26 F.3d 1368, 1373 (6th Cir. 1994) (emphasis in original). For plans delivered to or issued in Michigan, however, discretionary clauses are null and void by operation of Michi- gan law, and de novo review still applies. Mich. Admin. Code R. 500.2202. For reasons that seem more related to advocacy than law, Unum unhelpfully takes inconsistent positions on the appropriate standard of review throughout its motion papers. First, Unum argues that Michi- gan’s ban on discretionary clauses does not apply, because this plan was issued and delivered in Florida. That much is true: the plan was issued to Dr. Berg’s employer in Florida, and states by its terms that it was de- livered in and governed by the laws of Florida. ECF No. 11-2, Page- ID.717. Courts in this district have repeatedly found that out-of-state plans escape Michigan’s ban on discretionary clauses. See, e.g., Mellian v. Hartford Life & Accident Ins. Co., 161 F. Supp. 3d 545, 556 (E.D. Mich. 2016) (collecting cases). Despite having cleared the way for possi- ble application of the more favorable arbitrary and capricious standard, Unum ignores the plan’s language and fails to address whether there is a discretionary clause in the plan’s terms.5 5 Though neither party raised this point, the following plan language could be read as granting Unum discretion: “You are disabled when Unum determines that due to your sickness or injury. . . .” ECF No. 11- 2, PageID.733 (emphasis added). While other circuits appear to require more explicit language, the Sixth Circuit has concluded that similar language can trigger abuse of discretion review. See, e.g., Yeager v. Re- liance Standard Life Ins. Co., 88 F.3d 376, 381 (6th Cir. 1996) (plan granted discretion where it required claimant to submit “satisfactory Instead, it briefly suggests that the standard of review does not matter, then repeatedly argues throughout its papers that the Court should apply de novo review.6 See Def’s. Resp., ECF No. 18, Page- ID.2742 (describing de novo standard); Id. at 2755 (“… where, as here, the Court’s review is de novo”); Def’s. Mot. for Summ. J., ECF No. 15, PageID.2714 (describing only de novo standard in section of motion cap- tioned “Standard of Review”); Def’s. Repl., ECF No. 20, PageID.2786 (“The correct framing is that the Court must review the record evidence proof of Total Disability to us”); cf. Woods v. Prudential Ins. Co. of Am., 528 F.3d 320, 322-23 (4th Cir. 2008) (“when Prudential determines” language did not grant discretion to administrator); Nichols v. Pruden- tial Ins. Co. of Am., 406 F.3d 98, 108-9 (2d Cir. 2005) (same). Yet Sixth Circuit caselaw has not addressed whether the precise language in this plan provides the kind of clear grant of discretion that should trigger arbitrary-and-capricious review. And because Unum is asking the Court to apply de novo review, it fails to address whether or how this lan- guage grants it discretion. Accordingly, the Court will conduct a de novo review. 6 As Plaintiff observes, see Pl’s. Repl., ECF No. 19, PageID.2776, it is unusual for a Plan administrator to vigorously disclaim a standard that would entitle its administrative decision to substantial deference. The reasons for that position become clear in Unum’s Response to Berg’s motion and Reply regarding its own motion: Unum seeks to introduce a new rationale for denying Plaintiff’s claim—one that it never adduced during the nearly two years this matter was under administrative re- view. Unum argues that while new reasons are precluded by arbitrary- and-capricious review, they may be raised when review is de novo. But, as will be explained below, an ERISA plan administrator may not in- troduce new, post-hoc rationales for its decision in litigation, regardless of what standard of review applies. to determine whether it is more likely than not that Dr. Berg’s condition is due to a ‘Mental Illness’ . . . .”). Accordingly, as both parties request the Court to apply the de novo standard, and Defendant has clearly waived any argument to the contrary, the Court will apply the de novo of review standard. Under the de novo standard, no deference or presumption of cor- rectness is afforded to the administrator’s decision, and the Court in- stead endeavors to determine whether the administrator made the “cor- rect decision” based only on the record before the administrator. Perry v. Simplicity Engineering, 900 F.2d 963, 966 (6th Cir. 1990). The plaintiff bears the burden of proving, by a preponderance of evidence, that he or she is disabled. Javery v. Lucent Techs., Inc. Long Term Disability Plan, 741 F.3d 686, 700-01 (6th Cir. 2014). The Court must give a “fresh look” at the record, giving “proper weight to each expert’s opinion in accord- ance with supporting medical tests and underlying objective findings.” Id. at 700. Under the plan, Berg bears the burden of proving that she is enti- tled to benefits. See ECF No. 11-2, PageID.722 (“Proof of your claim, provided at your expense, must show…”); Seiser v. UNUM Provident Corp., 135 F. App’x 794, 797 (6th Cir. 2005). But Unum bears the bur- den of proving that the mental illness duration cap applies. See McCar- tha v. Nat’l City Corp., 419 F.3d 437, 443 (6th Cir. 2005) (“An ERISA plan, not the participant, has the burden of proving an exclusion applies to deny benefits.”); Okuno v. Reliance Standard Life Ins. Co., 836 F.3d 600, 609 (6th Cir. 2016) (insurer bore burden to prove that “Mental and Nervous Disorders Limitation” provision limited benefits to twelve months). III. ANALYSIS Unum does not appear to dispute that Dr. Berg suffers from cogni- tive symptoms and a decline in function, and that her decline prevents her from working as an anesthesiologist given the demanding responsi- bilities of that job. This makes good sense; to practice medicine, a physi- cian must be mentally fit and without cognitive impairment.7 Fatigue, concentration problems, and medication side effects can all render a physician unfit to practice. See, e.g., Chamness v. Liberty Life Assurance Co. of Bos., 234 F. Supp. 3d 885, 894-96 (W.D. Mich. 2017) (claimant was disabled and prevented from working as a physician where claim- ant’s treating doctor indicated claimant was limited from “practice of medicine [due] to impaired concentration and fatigue” as a result of de- pression and anxiety). 7 See Code of Medical Ethics Opinion 9.3.1, American Medical Associa- tion (“When physician health or wellness is compromised, so may the safety and effectiveness of the medical care provided. To preserve the quality of their performance, physicians have a responsibility to main- tain their health and wellness, . . . To fulfill this responsibility individ- ually, physicians should . . . [t]ake appropriate action when their health or wellness is compromised, including [e]ngaging in honest assessment of their ability to continue practicing safely.”). What determines the outcome if this case, then, is whether Unum has shown that Dr. Berg’s disability resulted from a psychological con- dition. If so, Berg’s disability falls within the policy’s “mental illness” limitation, and she has already exhausted the policy’s twelve-month lifetime cap on benefits for such conditions.8 If Unum has not borne its burden in showing that Berg’s disability resulted from a psychological condition, Berg is entitled to disability benefits up to the general dura- tion limit of the policy so long as she continues to satisfy whatever other conditions the policy may impose. In Okuno v. Reliance Standard Life Ins. Co., 836 F.3d 600, 607-8 (6th Cir. 2016), the Sixth Circuit considered the application of “mental illness” limitations in long-term disability insurance plans. Adopting the position adhered to by every federal appeals court to consider the issue, the Sixth Circuit held that the proper inquiry is whether the mental disability is a but-for cause of the plaintiff’s total disability.9 Id. 8 The policy defines “mental illness” as any “psychiatric or psychologi- cal condition classified” in the Diagnostic and Statistical Manual of Mental Health Disorders (DSM) most current at the time of the disabil- ity claim. The applicable edition is the DSM-5, which was published in 2013. 9 In Okuno, the Sixth Circuit considered a mental illness limitation that applied to disabilities “caused by or contributed to by” mental illness. The clause here applies to disabilities “due to” mental illness. Neither party argues that this policy’s “due to mental illness” language should be construed differently from the Sixth Circuit’s interpretation of the similar phrase “caused by or contributed to by mental illness,” so the As explained above, Okuno holds that Unum bears the burden on this issue. The Court therefore must determine whether Unum has shown by a preponderance of evidence that but for the existence of a psychiat- ric or psychological condition, Dr. Berg would be able to work and would not be disabled. First, as a threshold matter, the Court must also de- termine whether Unum can introduce a new rationale for its adminis- trative decision that it did not raise in the administrative process.. a. Whether Unum may introduce a new rationale for de- nial at this stage In its motion for summary judgment, Unum asserts for the first time that the DSM-5 classifies a particular condition that would appear to encompass Dr. Berg’s symptoms: “Substance/Medication-Induced Ma- jor or Mild Neurocognitive Disorder.” Unum therefore argues that even if Dr. Berg’s symptoms are the result of anastrozole or some other can- cer treatment rather than anxiety or depression, her disability still falls within a disorder classified by the DSM-5. Berg points out that Unum never mentioned such a rationale in its benefits decision, and argues that she would be prejudiced if Unum Court will apply the same “but-for” test. See also Kamerer v. Unum Life Ins. Co. of Am., 334 F. Supp. 3d 411, 414, 428 (D. Mass. 2018) (applying but-for test where mental illness exclusion included “due to” language). were allowed to assert a basis for denial that it did not previously raise.10 Pl’s. Resp., ECF No. 17, PageID.2729. Unum concedes that it is introducing this rationale for the first time, and that an ERISA administrator may not raise new reasons for denial at least when a court reviews a plan administrator’s decision un- der the arbitrary and capricious standard. But, Unum argues, that principle has no application when a court conducts a de novo review. Under de novo review, Unum says, the court’s task is to weigh the evi- dence and to determine whether benefits were correctly denied or granted—so the Court may consider reasoning that was not previously presented in the administrative process. Unum appropriately concedes that new reasons for denial cannot be raised when arbitrary and capricious review applies. Though the Sixth Circuit does not appear to have addressed the issue, decisions of other Circuits have unanimously rejected attempts to raise new reasons for denial once litigation has begun. See Glista v. Unum Life Ins. Co. of 10 Addressing the merits of Unum’s argument, Plaintiff argues that the plan’s definition of mental illness should not be read to include this condition, because not every condition mentioned in the DSM-5 is “clas- sified” as a mental illness. ECF No. 17, PageID.2729. But Plaintiff does not offer any evidence that “medication-induced major or mild neu- rocognitive disorder” is one of the conditions that is mentioned in the DSM-5 without actually being classified as a mental disorder. And in any event, the Court need not reach the merits of Unum’s newly raised rationale. Am., 378 F.3d 113, 130 (1st Cir. 2004) (“Unum violated ERISA and its regulations by relying on a reason in court that had not been articulated to the claimant during its internal review.”); Spradley v. Owens-Illinois Hourly Emps. Welfare Ben. Plan, 686 F.3d 1135, 1140-41 (10th Cir. 2012) (explaining that a plan administrator may not “treat the adminis- trative process as a trial run and offer a post hoc rationale in district court”) (citation and quotation marks omitted); Robinson v. Aetna Life Ins. Co., 443 F.3d 389, 395 n. 4 (5th Cir. 2006) (holding that a court must review the “actual basis for the administrator's denial of benefits, not its post-hoc rationalization”) (cleaned up and citation omitted). Contrary to Unum’s position, these principles are not limited to the arbitrary-and-capricious-review context. They apply with equal force when a court’s review is de novo. As the Ninth Circuit recently ex- plained, the rule that “a district court cannot adopt post-hoc rationaliza- tions” not presented to a claimant during the administrative process “protects the same procedural fairness concerns” outlined in the ERISA statute, regardless of what standard of review applies. Collier v. Lincoln Life Assurance Co. of Bos., 53 F.4th 1180, 1186-88 (9th Cir. 2022) (ex- plaining that plan administrator “sandbagged” claimant by introducing new rationales at stage where she could not meaningfully respond, even though review was de novo); see also Wolf v. Life Ins. Co. of N. Am., 46 F.4th 979, 985 (9th Cir. 2022). In the Sixth Circuit, as in the Ninth Cir- cuit, a reviewing district court’s task on de novo review is to “evaluate[] the plan administrator’s reasons for denying benefits without giving deference to its conclusions or opinions.” Collier, 53 F.4th at 1188 (em- phasis added); Perry v. Simplicity Eng’g, a Div. of Lukens Gen. Indus., Inc., 900 F.2d 963, 966 (6th Cir. 1990) (“In the ERISA context, the role of the reviewing federal court is to determine whether the administrator or fiduciary made a correct decision, applying a de novo standard.”). The Court’s role is to examine the administrator’s decision and reasoning in light of only the evidence in the record. Allowing a plan administrator to belatedly raise new reasons for denying a claim, regardless of the type of review to be applied, would not only contravene the instructions of the Sixth Circuit that district courts should not assume the role of “substitute plan administrators.” Perry, 900 F.2d at 966. It would also, as the Ninth Circuit observed, frustrate ERISA’s fundamental goals: to promote the consistent treat- ment of claims for benefits, enable non-adversarial resolution of claims, and minimize the cost of claim settlement for all parties. Collier, 53 F.4th at 1188-89. Accordingly, the Court will not consider Unum’s new- ly presented rationale for its decision. b. Whether Unum made the correct decision based on the record The etiology or causation of Dr. Berg’s disabling symptoms is a close and difficult question. The record leaves little doubt that her con- dition is complicated and its causes difficult to determine. Though all parties agree that her symptoms prevent her from working as an anes- thesiologist, both Berg and Unum raise credible arguments about the causes of those symptoms. However, the Court concludes that Unum has not borne the bur- den of showing by a preponderance of evidence that Berg’s symptoms are the result of a psychological condition for three reasons. First, Berg has introduced ample evidence that she suffers cognitive impairments, a point on which Unum agrees. And Berg has proffered credible evi- dence, in the form of opinions from her treating doctors, that these prob- lems result from the drug anastrozole and her cancer itself. Second, the medical opinions offered by Unum do not persuasively respond to alter- nate explanations offered by Berg’s treating doctors or, as will be ex- plained below, are less persuasive for other reasons. And third, where the conclusions of Unum’s file reviewers and Berg’s treating doctors conflict, the opinions of her doctors are entitled to more weight. 1. Berg has introduced evidence that her impairment is caused by cancer and her cancer treatments Berg’s treating doctors and Dr. Lopez all agree that her disabling symptoms were caused by her cancer and cancer treatment. Berg re- ported that her cognitive problems began shortly after she began taking anastrozole. ECF No. 12-3, PageID.1824. Dr. Ding acknowledged that anastrozole was known to cause cognitive side effects, and that even mild cognitive dysfunction would prevent Berg from working as an an- esthesiologist. Id. at PageID.1898. Dr. Tefend opined that, despite its possible side effects, anastrozole offered Berg the best chance of keeping her cancer in remission. After conducting cognitive tests, Dr. Gabel concluded that it was more likely than not “that cancer and related treatment” caused Berg’s symptoms, and that a psychiatric condition did not account for her cog- nitive changes. Dr Gabel also identified Berg’s insomnia and benzodiaz- epine use as potential contributing factors. And Dr. Lopez, after inter- viewing Berg and conducting tests and an examination of his own, opined that no psychological condition explained her symptoms. In- stead, he concluded that her problems were caused by the discontinua- tion of supplemental estrogen treatment and initiation of anastrozole, an estrogen blocking medication. Both Doctors Lopez and Gabel also noted that Berg denied experiencing any significant emotional distress, anxiety, or depressed mood, which suggested that neither an anxiety disorder nor depression was responsible for her symptoms. Finally, Dr. Lopez noted that although Berg’s insomnia had significantly improved and she had stopped taking a benzodiazepine medication for sleep by the time of his examination, her cognitive symptoms remained un- changed, serving to rule those factors out as possible causes of her symptoms. 2. The contrary opinions of Unum’s reviewers are not sufficient- ly persuasive to meet Unum’s burden In the administrative process Unum’s reviewers relied on several observations by Berg’s treating doctors. And while those reasons are compelling, when viewed in the context of the entire record, they are in- sufficient to carry Unum’s burden to show that, but for a psychological condition, Berg would not be disabled. The opinions of Unum’s file reviewers—Dr. Brown and Dr. Craw- ford—rested on a few critical points. First, that Berg had been diag- nosed with generalized anxiety disorder before her cancer, and that she began experiencing her cognitive symptoms after she was diagnosed with, but before she began treatment for, cancer—suggesting that her symptoms were an exacerbation of the underlying anxiety disorder. Second, that when Berg denied feeling depressed or anxious after her treatment, she could have been underestimating her symptoms. Third, that no physical condition explained Berg’s, condition, that her core symptoms are common in psychiatric conditions, and that post- treatment psychiatric dysfunction is common in breast cancer survivors. Fourth, that the cognitive effects of anastrozole are still being studied without a clear consensus. And finally, that Dr. Gabel and Dr. Ding stated that neither cancer nor anastrozole would be expected to account for a decline as severe as the one Berg experienced, and that Dr. Gabel recommended what appeared to be psychiatric treatments, not those that would be expected to treat a physiological condition. While compelling, these points are less persuasive when consid- ered in the full factual context. First, although Berg underwent therapy with Ms. Keller prior to her cancer diagnosis, the record does not sup- port Dr. Brown’s conclusion that Berg’s “disabling symptoms clearly be- gan prior to her cancer treatment.” Contemporaneous notes from Berg’s therapy sessions both before and after her diagnosis repeatedly describe Dr. Berg’s “cognitive functioning” as “Oriented/Alert” and “functional status” as “Intact.” ECF No. 12-2, PageID.1776-85 According to the notes, those sessions focused almost entirely on Berg’s relationships with her family. Notes reflect no concern by Berg or her therapist about Berg’s cognitive function until after her cancer diagnosis. And while Berg reported some difficulty focusing on her work after her diagnosis due to preoccupation with thoughts of her cancer, there is no indication at all that these difficulties were as severe as the symptoms she demon- strated post-treatment and that have ultimately prevented her return to work. Second, Unum’s file reviewers make much of Dr. Ding’s statement that the cognitive effects of anastrozole are usually “mild.” See ECF No. 12-3, PageID.1894. But the interpretation they draw from that state- ment alone contradicts Dr Ding’s clearly stated central conclusion: that even anastrozole’s expected “mild” cognitive impairment would prevent Berg from fulfilling the demanding requirements of her full-time anes- thesiology practice. Id. Similarly, Both Dr. Brown and Dr. Crawford based their opinions in large part on Dr. Gabel’s statement that “cancer contributed to at least a small to moderate degree to” Berg’s dysfunction and that other contributing factors were “important targets for treat- ment.” But that statement must be considered in light of Dr. Gabel’s subsequent clarification: that she intended her statement to mean that “it is more likely than not that cancer and related treatment caused de- clines in [Berg’s] executive functioning and vigilance” but that other fac- tors could be exacerbating the issue. ECF No. 13-2, PageID.2336. Third, Dr. Brown’s opinion that Berg may have been underreport- ing depression and anxiety symptoms is squarely rebutted by Dr. Lopez, who interviewed Berg on two occasions for a total of more than two hours, finding her responses credible throughout. ECF No. 13-3, Page- ID.2491. The Sixth Circuit has “repeatedly cautioned” that plan admin- istrators should not make credibility judgments about a patient’s sub- jective complaints without the benefit of a physical examination. Guest- Marcotte v. Life Ins. Co. of N. Am., 730 F. App’x 292, 302 (6th Cir. 2018). That warning applies with even greater force when cognitive symptoms are at issue. Accordingly the opinion of Dr. Brown is entitled to little weight on this point, and certainly less weight than that of Dr. Lopez, examined Berg twice. Fourth, Dr. Lopez’s notes and records from Dr. Abbey Dunn, a sleep specialist, show that, by mid-2020, Berg had been tapered off of the benzodiazepine medication she had been taking to treat her insom- nia, and that she was doing “fairly well” without “much trouble” falling asleep. ECF No. 13-2, PageID.2468-77. This suggests that insomnia and benzodiazepine use were not responsible for Berg’s symptoms. Finally, none of Unum’s file reviewers seriously engaged with Dr. Lopez’s conclusion that Berg’s symptoms were potentially caused by the compound effects of a sudden halt in Berg’s supplemental estrogen and use of anastrozole, an estrogen-blocking medication. In response to that argument, Doctors Brown and Crawford merely reiterated their earlier conclusions that the effects of anastrozole are a subject of intense re- search. Unum contends that there is “no record of Berg’s estrogen use, dosage, duration, frequency, or when it was allegedly discontinued.” Def’s. Mot. for Summ. J., ECF No. 15, PageID.2713. But that argument is belied by the record. Berg’s pharmacy records show that, for at least five years—until January, 2019 when she was diagnosed with cancer— she regularly filled a prescription for an estrogen-containing medication used to treat menopause symptoms.11 ECF No. 11-3, PageID.1009-12. 11 The records also show dosing information: one milligram of nore- thindrone acetate and five micrograms of ethinyl estradiol once daily. And records from Berg’s appointments with Dr. Gabel and other physi- cians show that by mid-2020, she was no longer taking any supple- mental estrogen medication. See, e.g., ECF No. 13-2, PageID.2460, 2466, 2473. In response to Dr. Lopez’s point, Dr. Brown simply reiterated his earlier conclusion that treatment records with Dr. Gabel show that Berg was undergoing “cognitive behavioral therapy,” and that Dr. Lopez’s arguments did not change Dr. Brown’s conclusions. ECF No. 13- 3, PageID.2497. Dr. Crawford, too, merely repeated her conclusion that “cognition in the setting of aromatase inhibitors is a topic of intense re- search,” and relied on the points in her earlier opinion. Id. at Page- ID.2499. Neither physician rebutted nor even addressed Dr. Lopez’s core point: that the estrogen-blocking anastrozole compounded the ef- fects of Berg’s sudden cessation of supplemental estrogen. 3. The opinions of Plaintiff’s treating doctors are entitled to more weight than Unum’s file reviewers It is significant that Doctors Brown, Crawford, Antaki, and Bart- lett never met with, interviewed, or examined Berg. And only Dr. An- taki appears to have spoken to or consulted any of Berg’s treating phy- sicians. Of course, a plan administrator is not required to examine a claimant, and there is “nothing inherently improper” in relying on the opinions of doctors who conducted file reviews only. Okuno, 836 F.3d at 610 (quoting Shaw v. AT & T Umbrella Ben. Plan No. 1, 795 F.3d 538, 550 (6th Cir. 2015)). But file reviews are particularly “questionable” where a claim involves, as this one does, a mental component or symp- toms that are subjectively manifested, such as those of which Dr. Berg complains. Javery v. Lucent Techs., Inc. Long Term Disability Plan for Mgmt. or LBA Emps., 741 F.3d 686, 702 (6th Cir. 2014). “Courts dis- count the opinions of psychiatrists who have never seen the patient for obvious reasons”: a psychiatrist typically treats or evaluates a patient’s subjective symptoms, which “depends on interviewing the patient and spending time with the patient[.]” Smith v. Bayer Corp. Long Term Dis- ability Plan, 275 F. App’x 495, 508 (6th Cir. 2008) (quoting Sheehan v. Met. Life Ins. Co., 368 F.Supp.2d 228, 254-55 (S.D.N.Y.2005). Applying those principles, the opinions of Doctors Gabel, Ding, Te- fend, and Lopez are entitled to substantial weight, because all inter- viewed and examined Berg on multiple occasions. They examined Berg’s mental state and conducted cognitive tests. By contrast, Doctors Brown, Dean, Antaki, Crawford, and Bartlett relied entirely on file reviews to reject the conclusions of Berg’s treating physicians. Accordingly, their opinions are entitled to less weight. Dr. Dean’s report also included a significant factual error—his statement that Dr. Berg returned to work after her surgery—which calls into question the reliability of his other conclusions and thoroughness of his review. Ultimately, the cognitive difficulties that prevent Dr. Berg from practicing her profession may well have multifarious and complicated causes. And it could be that her symptoms were caused or exacerbated in part by a psychological condition. But evaluating the record afresh, as this court must do, Unum has not carried its burden of showing that, absent any psychological condition, Dr. Berg would not suffer from the cognitive problems that preclude her from working as a doctor. Every doctor who has directly examined Dr. Berg has concluded that her cognitive symptoms prevent her from working. And her treat- ing oncologist and psychologist both determined that her symptoms were not caused by a mental condition. So did Dr. Lopez, another psy- chiatric consultant who directly examined Berg twice. The Court finds that the opinions of those doctors are persuasive and supported by the available record evidence. IV. CONCLUSION For the foregoing reasons, Plaintiff’s Motion for Judgment on the Record is GRANTED. Defendant’s Motion for Judgment on the Record is DENIED. Unum is directed to pay Plaintiff under the terms of the plan. Plaintiff is entitled to receive back payments for long-term disabil- ity benefits beginning on the day Plaintiff’s benefits were terminated which appears to be January 30, 2021, see Pl’s. Mot. for Summ. J., ECF No. 14, PageID.2687, and continuing so long as Plaintiff continues to meet the requirements of the plan to receive such benefits. IT IS FURTHER ORDERED that if Plaintiff wishes to request an award of attorney fees, such request, along with documentation in support, must be filed within 21 days of the day this Order issues, and that Defendant may respond within 14 days thereafter. As all of the outstanding claims in this matter have been fully re- solved, this case is hereby DISMISSED with prejudice. The Court will retain jurisdiction to consider any request for reasonable attorney’s fees. IT IS SO ORDERED. Dated: March 23, 2023 s/Terrence G. Berg TERRENCE G. BERG UNITED STATES DISTRICT JUDGE