Arruda v. Zurich American Insurance Co. ( 2020 )


Menu:
  •           United States Court of Appeals
    For the First Circuit
    No. 19-1247
    DENISE ARRUDA,
    Plaintiff, Appellee,
    v.
    ZURICH AMERICAN INSURANCE COMPANY,
    Defendant, Appellant,
    NSTAR ELECTRIC AND GAS BASIC ACCIDENT INSURANCE PLAN,
    Defendant.
    APPEAL FROM THE UNITED STATES DISTRICT COURT
    FOR THE DISTRICT OF MASSACHUSETTS
    [Hon. Douglas P. Woodlock, U.S. District Judge]
    Before
    Lynch, Stahl, and Lipez,
    Circuit Judges.
    Kristyn M. Kelley, with whom Allen N. David, Jane A. Horne,
    and Peabody & Arnold LLP were on brief, for appellant.
    Mala M. Rafik, with whom Sarah E. Burns and Rosenfeld & Rafik,
    P.C. were on brief, for appellee.
    February 24, 2020
    LYNCH, Circuit Judge.        Zurich American Insurance Company
    ("Zurich") denied the claim of Denise Arruda ("Arruda") for death
    benefits following the death of her husband Mr. Joseph Arruda in
    a 2014 car accident.          Zurich concluded, after reviewing the
    extensive record, that his death was not independent of all other
    causes and that it was caused or contributed to by his pre-existing
    health conditions.        As such, Zurich concluded the death was not
    within   the   coverage    clause   of    the   policy   and   was   within   an
    exclusion to the policy.
    Arruda sued under 29 U.S.C. § 1132(a)(1)(B), alleging
    that Zurich violated ERISA by unlawfully denying the insurance
    benefits.      Each party moved for summary judgment.            The district
    court entered summary judgment in Arruda's favor, holding that
    Zurich's decision was arbitrary and capricious, reasoning that the
    denial was not supported by substantial evidence. Zurich appealed.
    We reverse the district court, holding that Zurich's decision to
    deny the claim was supported by substantial evidence.                We direct
    entry of summary judgment for Zurich.
    I.
    A.   The Accident
    In May 2014, Mr. Arruda was 57 years old, employed as a
    sales executive by Northeast Utilities/NStar Electric and Gas, and
    covered under his employer's Basic Accident Policy (the "Policy")
    - 2 -
    issued by Zurich for accidental death or injury.                         He designated
    his wife as the beneficiary for any death benefits.
    On   the    morning       of    May    22,   2014,    Mr.   Arruda    drove
    westbound on Route 9, a four-lane road in Hadley, Massachusetts,
    on his way to a work event at the University of Massachusetts in
    Amherst.     At 9:39 a.m. his car crossed all lanes of traffic,
    collided with a car traveling eastbound, then hit the curb, rolled
    over, and landed on its wheels on the opposite side of the road.
    Police and fire department officials, including paramedics, from
    Hadley and Amherst arrived within ten minutes.                         Mr. Arruda was
    briefly alive following the accident, but quickly succumbed to his
    multiple injuries and was pronounced dead at the scene.
    Arruda timely filed for accidental death benefits on
    June 3, 2014.
    B.   The Terms of the Contract
    Under      Section    XII       of     the   Policy      (General    Policy
    Conditions), Zurich has "the discretionary authority to determine
    eligibility for benefits and to construe the terms of the plan."
    Under   Section      V    (Benefits),        the   Policy    states    that
    Zurich will pay benefits "[i]f an Insured suffers a loss of life
    as a result of a Covered Injury."                     As defined in Section III
    (Definitions), a Covered Injury is "an Injury directly caused by
    accidental    means      which    is    independent        of   all    other     causes."
    (Emphasis omitted).
    - 3 -
    Under Section VII (General Exclusions), the Policy does
    not cover losses that are subject to one or more exclusions:
    A loss will not be a Covered Loss if it is
    caused by, contributed to, or results from
    . . . illness or disease, regardless of how
    contracted, medical or surgical treatment of
    illness or disease; or complications following
    the surgical treatment of illness or disease
    . . . [or] being under the influence of any
    prescription drug, narcotic, or hallucinogen,
    unless such prescription drug, narcotic, or
    hallucinogen was prescribed by a physician and
    taken in accordance with the prescribed
    dosage.
    (Emphasis omitted).
    C.   Information Which Zurich Reviewed
    In response to Arruda's claim, Zurich hired CS Claims
    Group, Inc. ("CS Claims") to investigate and collect all records
    relevant to the claim.       CS Claims assembled Mr. Arruda's pre-
    accident medical records from his primary care doctor, various
    specialists, two hospitals, and his pharmacy.        Zurich later had
    these records examined by independent experts, including by a
    forensic pathologist, Mark L. Taff, M.D.     Dr. Taff concluded that
    these medical records revealed that Mr. Arruda had suffered from
    twenty-seven medical conditions from 2004 until his death.         As
    catalogued by Dr. Taff, the conditions evident from Mr. Arruda's
    medical   records     included,   among   others:   obesity,   chronic
    sinusitis, hypertension, a variant of hypertrophic cardiomyopathy
    (heart enlargement associated with arrhythmias and heart failure),
    - 4 -
    primary hyperaldosteronism, hypokalemia, a sedentary lifestyle,
    depression,     anxiety,   dyslipidemia,   diverticulosis,      insomnia,
    fatigue,     paresthesia   (tingling   sensation   in   the   peripheral
    nerves), a history of myalgias (muscle pain and weakness) and of
    bronchitis, kidney stones, and syncope (fainting spells).
    The records also showed that in mid-January 2014, about
    four months earlier than the accident, Mr. Arruda had an episode
    in which he felt weak, vomited, and fainted.       As a result, within
    a few days of the incident he underwent surgery and had an
    implantable    cardioverter   defibrillator   ("ICD")   placed    in   his
    chest.     The ICD monitored his heart rate and rhythm and could
    administer electric shocks to restore normal heart rhythm if
    necessary.
    Andrew W. Sexton, D.O., an employee of the Commonwealth
    of Massachusetts' Office of the Chief Medical Examiner, issued a
    death certificate on May 22, 2014 saying the cause and manner of
    Mr. Arruda's death were pending.       Dr. Sexton also did the autopsy
    on May 23, 2014.      Dr. Sexton apparently finalized the autopsy
    report on June 12, 2014 and concluded:
    CAUSE OF DEATH: Hypertensive Heart Disease.
    Contributory Factors: Upper Cervical       Spine
    Fracture due to Blunt Impact.
    MANNER OF DEATH: Accident (Driver Involved in
    a Motor Vehicle Collision with Rollover)
    - 5 -
    These       conclusions   apparently    did    not   include   toxicology   and
    cardiac findings done after that date, although the report made
    reference to their existence.1          However, no amended autopsy report
    was ever found.
    Dr. Taff later summarized the significant findings of
    the autopsy report as follows:
    1. Hypertensive      cardiovascular     disease
    associated with cardiomegaly (an enlarged
    heart weighing 530g; normal hearts usually
    weigh no more than 420g), biventricular
    hypertrophy (thickened right and left
    ventricles), mild, non-occlusive (less
    than      30%       luminal       narrowing)
    arteriosclerotic triple coronary artery
    disease,    moderate    atherosclerosis   of
    abdominal aorta, multifocal interstitial
    myocardial fibrosis (abnormal scarring of
    heart muscle) and an intact functioning
    cardiac     pacemaker/ICD      defibrillator
    implant.
    2. Mild pulmonary edema (wet lungs due to an
    abnormal increase of fluid).
    3. Multiple blunt force impact injuries of the
    head (multiple scalp bruises distributed
    about   the    head   and   eyelids),   neck
    (fractured     1st     cervical    vertebra;
    dislocated 3rd and 4th cervical vertebra
    associated with a grossly normal appearing
    cervical spinal cord), torso (multiple (10)
    bilateral anterior rib and upper sternum
    (breast plate) fractures) and upper and
    lower extremities (multiple soft tissue
    bruises).
    4. Obesity (5'11"/216 lbs.).
    5. Benign prostatic hypertrophy (BPH) due to
    an enlarged prostate gland.
    1 Like the district court, we decline "to read much into
    this discrepancy as such." The latter two reports are part of the
    record before Zurich and must be considered when assessing whether
    Zurich had substantial evidence to support its decision.
    - 6 -
    6. Hepatomegaly (enlarged soft liver weighing
    2,050g; normal liver weight is up to about
    1,700g).
    7. Diffuse light purple congestion of face,
    lips and mouth associated with petechial
    (pinpoint) hemorrhages of right and left
    lower conjunctiva (eyes) and lips.
    During the autopsy, the ICD was surgically removed and sent to
    Boston Scientific, the manufacturer, for analysis.
    Mindy J. Hull, M.D., also of the Massachusetts Medical
    Examiner's Office, completed a cardiac pathology report on January
    12, 2015.    The report found "mild coronary artery disease" and
    "focal interstitial fibrosis of [the] lateral left ventricle."   It
    did not mention any evidence of an acute cardiac event.2
    In conjunction with the Massachusetts Medical Examiner,
    the Town of Hadley, Massachusetts, on June 9, 2014 issued a death
    certificate with the same primary cause of death as in the autopsy
    report, "hypertensive heart disease."
    Various reports written by first responders to the scene
    of the car accident were part of the record.     A report completed
    by paramedics from the Amherst Fire Department on the day of the
    accident described the paramedics' efforts to save Mr. Arruda's
    2    A blood toxicology report was completed on July 30, 2014
    by the Massachusetts State Police's Forensic Services Group. It
    showed that Mr. Arruda's blood had 17 ng/ml of Delta-9 THC (the
    primary active ingredient in marijuana) and more than 40 ng/ml of
    Delta-9 Carboxy THC, its inactive metabolite.        While Zurich
    independently found marijuana to be a contributing cause to the
    death, we have no need to reach the issue and do not further
    discuss the marijuana evidence or the parties' disputes about it.
    - 7 -
    life and listed in the "Impressions" section "Primary: Cardiac
    Arrest" and "Secondary: Motor Vehicle Accident[,] Trauma."                 An
    Accident Report from the Hadley Police Department completed the
    day after the accident described basic information about the
    trajectory of the crash and recorded the contact information of
    six witnesses.
    The   Massachusetts   State    Police    completed    an   ACISS
    Homicide/Death Report on August 25, 2014.        It included information
    the police gathered from the witnesses, including that Mr. Arruda
    was briefly alive following the accident and was suffering from
    multiple injuries, including an obvious neck injury.              Before the
    paramedics arrived, he "went into breathing distress and started
    to seize" before losing consciousness.          Based on the interviews
    and preliminary autopsy reports, the State Police concluded that
    Mr. Arruda "experienced some type of medical episode while driving
    his vehicle."
    The   Massachusetts    State    Police    also   completed     a
    "Collision     Analysis    and     Reconstruction      Section     Collision
    Reconstruction Report" on February 28, 2015. The officer who wrote
    the report ruled out various causes for the accident, including
    poor road conditions, mechanical failure, engineering design flaws
    in the road, speeding, and other drivers' error.              He concluded
    that Mr. Arruda "had suffered a catastrophic medical event which
    caused him to be unable to control his vehicle."
    - 8 -
    Zurich   initially   turned    this    material   over    to   two
    independent medical doctors for review, and later to a third
    independent expert.     The first was William W. Angell, M.D., whose
    credentials are not in the record.            Dr. Angell submitted his
    opinion on July 6, 2015 in a short, two-paragraph statement which
    was not on official letterhead.     Dr. Angell stated: "[I]t would be
    my opinion that Mr. Arruda experienced a cardiac event at the time
    of the accident which resulted in his death and that the death was
    not independent of an underlying medical condition as indicated in
    the autopsy report."     He did not further explain what he meant by
    a cardiac event.     He also did not explain his reasoning for this
    conclusion but did state he had reviewed the file documents,
    including   the   medical   records,      police   reports,   and     Medical
    Examiner reports.     Later in the claims process, Zurich tried to
    locate Dr. Angell but was not able to do so.
    The second independent medical review for Zurich was
    completed on November 30, 2015 by Michael D. Bell, M.D., a board-
    certified specialist in both Anatomic and Clinical Pathology and
    Forensic Pathology, licensed in New York and Florida.                Dr. Bell
    reviewed all of the medical and non-medical documentation.             He was
    asked specific questions and answered them as follows:
    1. Did the deceased die from an accidental
    bodily injury, independent of all other
    causes? If so, please list all injuries
    sustained.
    - 9 -
    The crash and his death were caused by his
    heart   disease,   whether   it   be  due   to
    hypertension or a variant of [hypertrophic
    cardiomyopathy].     However, based on the
    autopsy results, the decedent's C1 left
    posterior arch fracture was C3-C4 dislocation
    with soft tissue hemorrhage at the injury
    sites would be a contributory cause of death.
    He had a C1 left posterior arch fracture and
    C3-C4 dislocation with soft tissue hemorrhage
    at the neck injury sites. He did not have a
    visible spinal cord injury.      While he had
    multiple scalp bruising, he did not have a
    skull fracture or cerebral, cerebellar or
    brainstem injury.    He had bruising of his
    right arm, left hand, and both legs. The rib
    fractures and chest bruising was believed to
    be caused by resuscitative chest compressions.
    2. Was the death caused by, contributed to or
    the result of illness or disease?      If so,
    please    list    all    medical    conditions
    contributing to death.
    The crash and his death were caused by his
    heart   disease,   whether   it  be   due   to
    hypertension or a variant of [hypertrophic
    cardiomyopathy].    He has been treated for
    hypertension since at least 2008 and it has
    been difficult to control.    The most likely
    mechanism of his crash and death is a
    ventricular arrhythmia secondary to his heart
    disease.    He also has hyperaldosteronism,
    which made controlling his blood pressure
    difficult.   However, the decedent's C1 left
    posterior arch fracture and C3-C4 dislocation
    with soft tissue hemorrhage at the injury
    sites would be a contributory cause of death.
    Based on all of this information, Zurich denied Arruda's
    claim in a letter dated December 8, 2015.    Zurich relied on two
    different Policy clauses in its denial: the coverage grant was not
    triggered because the death was not "independent of all other
    causes" and the death was excluded from coverage because it was
    - 10 -
    "caused by, contributed to, or results from" an "illness or
    disease."      The letter specifically highlighted the independent
    medical reviewers' conclusions and the cause of death recorded on
    the death certificate as determined by the Medical Examiner.
    Arruda timely appealed Zurich's determination on January
    29, 2016.      As part of her appeal letter, she submitted a logbook
    from Boston Scientific that recorded the information Mr. Arruda's
    ICD captured about his heart's condition in the months leading up
    to the accident.3      The logbook has three references to the date of
    Mr. Arruda's death, May 22, 2014.         The first is that at 8:23 a.m.
    on May 22, 2014, seventy-five minutes before the accident, the
    logbook has an entry for a successful "rhythm ID update."                  The
    second is an "alert" from 2:24 p.m., approximately four and a half
    hours after Mr. Arruda's death, saying "Ventricular Tachy mode set
    to value other than Monitor+Therapy." The third is that the report
    says it was "created" on May 22, 2014.          The logbook has no record
    of the cessation of Mr. Arruda's heart occasioned by his death.
    Arruda   did    not   submit   anything   to   Zurich   explaining   how   to
    interpret the logbook, including anything to explain what "rhythm
    ID update" means or the significance of seventy-five minutes
    between that reading and his death.
    3    She also submitted a transcript of a workers'
    compensation hearing and a resulting settlement agreement under
    which the employer agreed to accept liability for Mr. Arruda's
    death and pay Arruda a lump sum settlement amount.
    - 11 -
    On August 24, 2016, Arruda supplemented her appeal with
    an independent medical review from Elizabeth A. Laposata, M.D.,
    dated August 5, 2016, the first of two reports Dr. Laposata
    submitted in support of her claim.         Dr. Laposata is with Forensic
    Pathology & Legal Medicine, Inc., of Providence, Rhode Island.
    She is the former Chief Medical Examiner for the State of Rhode
    Island and a Fellow of both the College of American Pathologists
    and the American Society for Clinical Pathology.
    In her first August 5, 2016 report, Dr. Laposata's main
    conclusion was that Mr. Arruda did not experience "a natural death
    at the wheel" with a resulting collision.           The purpose of this
    conclusion is unclear.      Zurich's denial of benefits made no such
    assertion.     Neither Dr. Angell nor Dr. Bell had stated that Mr.
    Arruda had experienced a natural death at the wheel.         Indeed, Dr.
    Bell expressly acknowledged that a severely injured Mr. Arruda was
    alive when found after the accident.
    Dr.   Laposata's   report   also   criticized   the   Medical
    Examiner's conclusions as "incorrect" and inconsistent with the
    death being "accidental," as the Medical Examiner's report had
    concluded.     She opined that "Mr. Arruda's correct cause of death
    is neck injuries due to blunt force trauma in the circumstance of
    a motor vehicle . . . collision with rollover." As to the question
    of what had caused Mr. Arruda to crash, she stated: "The exact
    reason Mr. Arruda traveled across several traffic lanes and into
    - 12 -
    the other vehicle is unclear."        She did note that "[o]nly seconds
    of distraction or inattention to driving would be needed for his
    car to move out of his lane of travel and into the far lane and
    impact the second vehicle."       She did not opine on whether Mr.
    Arruda's     pre-existing   medical    conditions   either   "caused   or
    contributed to" the crash.
    Dr. Laposata commented on the logbook in her August 5,
    2016 report.     She wrote that since the ICD "showed no abnormal
    heart rhythms recorded prior to the collision," the accident was
    not caused by "incapacitation by heart disease."        She did not say
    explicitly that the absence of data showed that no abnormal heart
    rhythm had occurred between 8:23 a.m. and the later time of the
    accident.    Nor did she explain the absence of a recording in the
    logbook of the cessation of the heartbeat at death.          Arruda never
    submitted to Zurich any materials on proper interpretation of the
    logbook entries, or lack of entries.
    In response to Arruda's appeal, Zurich sought a third
    independent medical review.      It obtained a report dated January
    16, 2017, apparently through a company named ExamWorks, from Dr.
    Taff.   Dr. Taff is a forensic pathologist and clinical associate
    professor of pathology at Mount Sinai School of Medicine in New
    York City.     He had over thirty years' experience as a practicing
    board-certified pathologist and had investigated dozens of fatal
    motor vehicle accidents.     He stated that the opinions he gave "are
    - 13 -
    to a reasonable degree of forensic medical certainty" and were
    based on his over thirty years of experience in the field.
    In reaching his conclusions, Dr. Taff stated he had
    reviewed and analyzed:
    the 450-page file containing the following
    documentary evidence: 1) Massachusetts Police
    Investigative/Motor Vehicle Crash reports;
    2) Joseph Arruda's (JA) autopsy, toxicology,
    histology    (microscopic    examination    of
    tissues),   cardiac    pathology   and   death
    certificate reports; 3) medical expert reports
    prepared by Drs. Elizabeth Laposata, Michael
    Bell and William Angell; 4) pre-mortem medical
    records of Joseph Arruda dated 2004 - 2014;
    5) news clips regarding the fatal motor
    vehicle   collision;    and   6)   testimonial
    transcripts of multiple witnesses.
    In his January 16, 2017 report to Zurich, Dr. Taff ruled out
    several possible causes of the accident.        Although Mr. Arruda had
    suffered from depression and anxiety, Dr. Taff ruled out suicide
    as a cause.     He stated the State Police investigation did not
    reveal   any   vehicle   or   environmental   factors   that   would   have
    contributed to the crash.       He noted that "[t]he issue of texting
    while driving was not addressed in the police final reconstruction
    report."
    In response to the question "Was the accident caused by,
    contributed to or resulted from an illness or disease (cardiac
    event/heart disease)?", Dr. Taff answered:
    The accident was caused by several possible
    pre-existing illnesses or diseases, singly or
    in   combination,   including:   a)   cardiac
    - 14 -
    arrhythmia resulting from pre-existing heart
    disease (hypertensive cardiovascular disease
    or a variant of hypertrophic cardiomyopathy);
    b) an adverse drug reaction for medications
    prescribed for pre-existing illness or heart
    disease; c) prescribed heart medication-
    related      blood      pressure      problems;
    d) electrolyte    imbalance    (e.g.    cardiac
    arrhythmias related to low blood potassium
    levels due to primary hyperaldoasteronism)
    [sic]; e) muscle weakness related to low blood
    potassium     levels     due     to     primary
    hyperaldoasteronism [sic]; f) complications
    of undiagnosed sleep apnea resulting in
    falling   asleep    behind   the   wheel;   and
    g) temporary     or    intermittent     cardiac
    pacemaker failure.
    Before giving the conclusion, he explained the basis for it:
    Although JA died from multiple bodily injuries
    sustained in a motor vehicle collision with
    several rollovers, it is uncertain why he
    suddenly and inexplicably veered off the
    westbound side of Rte 9 into oncoming traffic
    on the eastbound side.     Based on JA's past
    medical history, there are several possible
    human factors, singly or in combination, that
    triggered the pre-impact phase of the motor
    vehicle collision, including a) long-standing
    heart disease (hypertension and variant of
    hypertrophic cardiomyopathy); b) medication-
    related problems for treatment of JA's pre-
    existing pathological conditions (sudden drop
    or increase in blood pressure); c) recent
    implantation    of    a   cardiac    pacemaker;
    d) hypokalemia (low blood potassium levels
    most likely due to pre-existing primary
    hyperaldoasteronism [sic] contributing to
    muscle weakness or a cardiac arrhythmia);
    e) chronic insomnia (falling asleep behind the
    wheel of a car); and f) breathing problems
    (e.g. chronic sinusitis and heavy snoring).
    Although JA was never diagnosed with sleep
    apnea,    several     of    his    pre-existing
    pathological conditions are known to cause
    irregular    sleeping    patterns,    breathing
    - 15 -
    difficulties, chronic fatigue and obesity.
    Based on the circumstances, there is a good
    chance that JA fell asleep behind the wheel.
    The above pre-existing medical conditions,
    singly or in combination, could have set off
    an acute medical crisis that resulted in JA's
    sudden incapacitation behind the wheel of his
    vehicle. According to several reports, post-
    mortem analysis of JA's implantable ICD device
    showed   no   evidence   of   an   ante-mortem
    arrhythmia. Based on the scene findings and
    eyewitness accounts, JA was still alive for a
    brief period of time after the collision and
    rollovers. There is no way to scientifically
    prove   which   human   factor(s)/pre-existing
    medical condition(s) occurred during the pre-
    collision phase of the accident that resulted
    in fatal bodily injuries.
    As this language makes clear, he did consider the analysis of the
    implanted ICD device in the logbook in reaching his conclusion.
    In an addendum to her appeal, also considered by Zurich,
    Arruda replied to Dr. Taff's report with a supplemental report
    from Dr. Laposata dated April 14, 2017.   It is this addendum which
    is now at the core of Arruda's argument.       The second Laposata
    report stated:
    There is no medical or scientific evidence to
    support a conclusion that Mr. Arruda's death
    due to injuries sustained in that motor
    vehicle accident was "caused by, contributed
    to, or results [sic] from illness or disease."
    The Insurance Company misrepresents the
    finding by Dr. Taff. Dr. Taff puts forward
    "several possible human factors" noting Mr.
    Arruda's medical conditions but concludes
    "There is no way to scientifically prove which
    human      factor(s)/pre-existing      medical
    conditions occurred during the pre-collision
    phase . . ."    There is no evidence in the
    material examined that demonstrates to a
    - 16 -
    reasonable degree of medical certainty that
    any of Mr. Arruda's medical conditions caused
    or contributed to the accident.            The
    interrogation of his cardiac defibrillator
    gives definitive proof that no cardiac
    arrhythmia or event preceded the accident.
    Additionally, Mr. Arruda never received a
    doctor's restriction that would limit his
    ability to operate a motor vehicle safely.
    Trooper Sanford speculates that Mr. Arruda
    "suffered a catastrophic medical event." He
    is clearly not qualified to make such a
    medical determination. Finally, the autopsy
    ruled out any other disease processes that
    would cause physical incapacitation at the
    wheel.
    It is a serious error to conclude that
    the mere existence of medical diagnoses and
    speculation as to what might happen given
    these conditions equates with proof that a
    medical event did occur prior to the accident.
    Dr. Taff concludes that Mr. Arruda died from
    a broken neck, spinal cord injury and
    positional   asphyxia,   all   injuries   that
    occurred due to the motor vehicle accident.
    Mr. Arruda died from accidental bodily injury,
    independent of all other causes.
    (Alteration in original).
    Zurich's appeals committee upheld the denial of benefits
    to   Arruda    on    May   11,   2017,    identifying   the   same   two   Policy
    provisions     and    specifically       stating   reliance   on   the   accident
    reconstruction report, the Commonwealth's autopsy report and death
    certificate, and Zurich's three independent medical reviews.                   It
    did not say it relied on the logbook.                   It acknowledged Dr.
    Laposata's differing opinion.              The appeals committee stated that
    Arruda's claim would be denied because Mr. Arruda's death was not
    - 17 -
    "independent of all other causes" and was "caused by or resulted
    from" his pre-existing medical conditions.4
    D.   Summary Judgment Reasoning of the District Court
    The District Court concluded that Zurich's denial of
    benefits was arbitrary and capricious.      It provided two different
    reasons for finding the denial arbitrary and capricious. The first
    was that it understood Zurich to have concluded that Mr. Arruda's
    "cause of death was heart disease."          But, it reasoned, that
    conclusion was contradicted by Drs. Taff and Laposata and that
    Drs. Bell and Sexton "cite no evidence to support the conclusion
    that heart disease was the cause of death, other than the fact
    that Mr. Arruda had a history of heart disease."      The second reason
    was that it understood Zurich to have concluded only that "Mr.
    Arruda's   preexisting   illness   caused   the   accident,"   (emphasis
    added), which then caused his death.    The court relied on language
    in Dr. Taff's opinion that he could not identify "which human
    4    The issue of which party has the burden of proof once an
    exclusion is invoked, given that both coverage and exclusions are
    at issue, is immaterial here as our conclusion would hold
    regardless. See Glista v. Unum Life Ins. Co., 
    378 F.3d 113
    , 131
    (1st Cir. 2004) ("[T]raditional insurance law places the burden on
    the insurer to prove the applicability of exclusions such as the
    Pre-Ex Clause."). Regardless, under the arbitrary and capricious
    standard, "the issue is only whether there is substantial evidence
    in the record to support the administrator's determination."
    Arruda v. Zurich Am. Ins. Co., 
    366 F. Supp. 3d 175
    , 182 n.1 (D. Mass.
    2019). Zurich's decision is supported by substantial evidence as
    to both the Policy exclusions and the definition of a covered loss
    for coverage purposes.
    - 18 -
    factor(s)/pre-existing medical condition(s) occurred during the
    pre-collision phase of the accident that resulted in fatal bodily
    injuries."      (Emphasis added).      In the district court's view, the
    record "does not provide evidence beyond the mere existence of
    pre-existing illness."        It agreed with Zurich that the logbook
    evidence was inconclusive and that it was not the basis for
    Zurich's denial.
    The district court did not specifically focus on the
    Policy's "contributed to" language or the insurer's reliance in
    its denials on this language in referring to both the Policy and
    the medical evidence.         Nor did the court focus on the reasons
    stated in the denial letter.        Zurich's May 11, 2017 denial letter
    says   that    there   was   medical   evidence   that   the   accident   was
    "contributed to" by pre-existing medical conditions or "was caused
    by or resulted from illness [and] disease."         In the letter, Zurich
    cited Dr. Taff's conclusion that "Mr. Arruda died as the result of
    accidental bodily injuries but they were contributed to by multiple
    pre-existing illnesses or diseases."
    This timely appeal followed.
    II.
    A.     Standard of Review
    We review a district court's grant of summary judgment
    de novo.      D & H Therapy Assocs., LLC v. Boston Mut. Life Ins. Co.,
    
    640 F.3d 27
    , 34 (1st Cir. 2011).
    - 19 -
    Where, as here, the plan administrator is explicitly
    given discretionary authority by the terms of the Policy, we ask
    whether its decision is arbitrary and capricious or an abuse of
    discretion.    See Firestone Tire & Rubber Co. v. Bruch, 
    489 U.S. 101
    , 111 (1989); Doe v. Standard Ins. Co., 
    852 F.3d 118
    , 123 (1st
    Cir. 2017).      That is, we must defer where the "decision is
    reasonable and supported by substantial evidence on the record as
    a whole."     McDonough v. Aetna Life Ins. Co., 
    783 F.3d 374
    , 379
    (1st Cir. 2015).    "Substantial evidence" is "evidence reasonably
    sufficient to support a conclusion."        Doyle v. Paul Revere Life
    Ins. Co., 
    144 F.3d 181
    , 184 (1st Cir. 1998).         Indeed, in Doyle,
    this court cited to an administrative law case that used the
    sufficiency of the evidence standard in administrative law for
    guidance on how to determine what arbitrary and capricious means
    in the ERISA review context.      
    Id. (citing Associated
    Fisheries of
    Me., Inc. v. Daley, 
    127 F.3d 104
    , 109 (1st Cir. 1997)).         Moreover,
    "[s]ufficiency   . . .   does   not   disappear   merely   by   reason   of
    contradictory evidence."    
    Id. The job
    of a court is not to decide
    the "best reading" of the policy, O'Shea v. UPS Ret. Plan, 
    837 F.3d 67
    , 73 (1st Cir. 2016), but rather, to evaluate whether
    Zurich's conclusion was "reasonable."        Colby v. Union Sec. Ins.
    Co. for Merrimack Anesthesia Assocs. Long Term Disability Plan,
    
    705 F.3d 58
    , 62 (1st Cir. 2013).
    - 20 -
    Under this deferential standard, we hold that Zurich's
    decision was reasonable, supported by substantial evidence, and
    not arbitrary and capricious or an abuse of discretion.
    B.    Pre-Existing Medical Conditions as a Contributing Cause of
    Death
    The descriptions in the record before Zurich of the
    causes that contributed to Mr. Arruda's death were all consistent
    that his crash was caused, at least in part, or was contributed to
    by his pre-existing medical conditions.               Taking all of these
    materials and medical opinions "as a whole," 
    McDonough, 783 F.3d at 379
    ,   Zurich's      conclusion   is   not   undermined   because   Dr.
    Laposata's opinion differed.           "[T]he existence of contradictory
    evidence does not, in itself, make the administrator's decision
    arbitrary."      Vlass v. Raytheon Emps. Disability Tr., 
    244 F.3d 27
    ,
    30 (1st Cir. 2001).
    In    fact,     Dr.   Laposata's      first   report   was   not
    inconsistent with Zurich's ultimate conclusion that Mr. Arruda's
    death was not "independent of all other causes."            She only stated
    that "Mr. Arruda was alive at the time of the crash" and did not
    die "a natural death at the wheel."         But that he was alive shortly
    after the crash was never at issue.
    The thrust of Dr. Laposata's second report was her
    assertion that it was impossible to tell with "a reasonable degree
    of medical certainty" that Mr. Arruda's pre-existing pathologies
    - 21 -
    contributed to his having the accident which resulted in his death.
    But Zurich could reasonably rely on Dr. Taff's opinion "to a
    reasonable degree of forensic medical certainty" that that is
    exactly what happened.        That Dr. Taff was reluctant to conclude
    further exactly which of the many pre-existing pathologies, singly
    or in combination with others, provided the precise contribution
    does not negate his ultimate conclusion.             Rather, it reinforces
    the care with which he analyzed the data before reaching his
    conclusion.      That care is also evidenced by his exclusion of two
    pathologies as contributions.
    Nor was Zurich obligated to accept Arruda's view that
    the medical opinions on which Zurich relied were nothing more than
    speculation because they did not "provide evidence beyond the mere
    existence   of    pre-existing    illness."     Dr.    Taff's   report,   in
    particular, carefully rules out other possible causes of the
    accident, gives a detailed account of Mr. Arruda's medical history,
    acknowledges potentially conflicting evidence, and comes to a
    reasoned conclusion.
    Arruda offers no support for her contention that Dr.
    Taff needed to determine the precise mechanism or mechanisms by
    which Mr. Arruda's pre-existing conditions contributed to Mr.
    Arruda's car suddenly veering across multiple lanes of traffic and
    his fatal car accident.       It is sufficient that Dr. Taff reached a
    firm   conclusion    to   a   reasonable    degree   of   forensic   medical
    - 22 -
    certainty,     which   was   self-evidently    reasoned,    that   some
    manifestation(s) of Mr. Arruda's pre-existing conditions caused
    him to have the accident that killed him.      As is evident from the
    passages of Dr. Taff's report excerpted above, Dr. Taff showed a
    strong familiarity with the facts of the case and drew reasoned
    conclusions by applying his medical expertise.
    Arruda and her expert criticize Dr. Taff's report, in
    particular, as engaging in speculation because of his use of
    language such as "mostly likely," "a good chance," and "could
    have."   In leveling this criticism, they would have us ignore his
    conclusions given "to a reasonable degree of forensic medical
    certainty."    Zurich could reasonably rely on that earlier language
    and conclude it did not undercut the conclusion.           According to
    common dictionary definitions, "likely" establishes a probability.
    Likely, Black's Law Dictionary (10th ed. 2009) ("Apparently true
    or real; probable . . . [s]howing a strong tendency; reasonably
    expected");      Likely,     Merriam-Webster    Online      Dictionary,
    https://www.merriam-webster.com/dictionary/likely        (last   visited
    Feb. 19, 2020) ("having a high probability of occurring or being
    true: very probable"); see also Glista v. Unum Life Ins. Co., 
    378 F.3d 113
    , 127 (1st Cir. 2004) (citing a dictionary definition of
    "treatment" while interpreting a policy clause in an ERISA case).
    We have said that the arbitrary and capricious standard
    has some "bite," 
    McDonough, 783 F.3d at 379
    , but that does not
    - 23 -
    mean that an insurer cannot rely on a doctor's conclusion because
    another doctor found his language not sufficiently precise.
    We address our differences with the dissent.5                   The
    dissent relies heavily on the ICD logbook, but in doing so it
    misstates how Zurich used the logbook and what the logbook showed.
    Zurich did not rely on a particular interpretation of the logbook
    to deny Arruda's claim and it does not rely on one now to support
    its appeal.   It is also untrue that the proper interpretation of
    the logbook is undisputed.
    Zurich never rested on the logbook to support its denial.
    Indeed, Arruda's opening brief to this court argued that because
    Zurich had not relied on the logbook to deny benefits it could not
    later use the logbook entries to support its denial because Zurich
    had not done so earlier.      In its reply brief, Zurich argued that
    it had not waived its right to argue that the arrhythmia logbook
    was   inconclusive   and   repeated   that    it   did   not   rely   on   the
    inconclusive logbook in denying benefits.
    Zurich has explained why it did not rely on the logbook
    to support its denial of her appeal.         The proper interpretation of
    the logbook, which contains many technical medical terms and
    5   The dissent mischaracterizes Zurich's reasons for
    denial. Zurich did not conclude that Arruda's claim was denied
    because of "the mere existence of [Mr. Arruda's] pre-existing
    illness." Neither Zurich nor any of its doctors so represented.
    - 24 -
    abbreviations, is contested. As the district court correctly held,
    "the logbook does not bear all the weight Mrs. Arruda seeks to
    place on it."       Arruda maintains that the logbook must mean that
    the ICD recorded any and all heart irregularities in real time up
    through all events associated with the accident. Zurich reasonably
    interpreted the logbook as inconclusive, and that view is supported
    by the record.      The logbook did not record anything after the last
    "rhythm    ID    update"   seventy-five       minutes   before     the   accident,
    including by the fact that the logbook failed to record the
    stopping    of    Mr.    Arruda's    heartbeat    on    his    death,    therefore
    providing evidence it was not working properly.
    The dissent, nonetheless, takes the position that Zurich
    was   compelled     to   accept     Dr.    Laposata's   understanding      of   the
    logbook.    That is wrong for multiple reasons.               That reading is not
    unrebutted in the record. We have already pointed out deficiencies
    in Dr. Laposata's opinion.          The ICD captured only events which it
    was programmed to capture.           There is no evidence anywhere in the
    record as to how the device was programmed.
    Separately,      two      of     Zurich's    independent       medical
    reviewers, Drs. Bell and Taff, both considered the ICD evidence
    and concluded that his death was caused or contributed to by
    illness or disease, even assuming favorably to Arruda that the ICD
    continued to record accurately.            The dissent misses the point when
    it insists the only possible pre-existing medical condition which
    - 25 -
    could have contributed to the event was a cardiac arrhythmia or
    other cardiac event preceding the accident.       Dr. Taff's opinion
    lists at least seven different possible medical conditions that,
    singly or in combination, caused or contributed to Mr. Arruda's
    death.   One of those was "heart disease," a broader term than
    "heart attack" or "heart arrhythmia."     Another was a "temporary or
    intermittent cardiac pacemaker failure."      The other pre-existing
    conditions Dr. Taff specified were independent of heart attack or
    arrhythmia.    Dr. Taff did not have to provide further explanation
    for how those conclusions are compatible with the logbook because
    there is no evidence the ICD captured all seven of the possible
    pre-existing causes set forth by Dr. Taff, and from the nature of
    the device, it is clear that it could not.
    At most, Dr. Laposata's view, summarized in her addendum
    report, was that the ICD gives "proof that no cardiac arrhythmia
    or event proceeded the accident."       She did not say that it gave
    proof that no pre-existing condition at all contributed to the
    accident.   Indeed, Zurich was entitled to consider, in finding the
    logbook inconclusive, Dr. Laposata's earlier view that the ICD
    showed no episodes of "sustained ventricular tachycardia and no
    defibrillation discharges" and her expressed view that whatever
    caused the accident could have occurred within the time frame of
    mere seconds.    (Emphasis added).
    - 26 -
    C.   Zurich Was Not, In the Face of Medical Evidence to the
    Contrary, Required to Accept Claimant's Evidence
    Beyond this assessment of why the evidence supports the
    denial,       Arruda's   premise     is    that   judges   may   find   insurers'
    decisions as to benefits to be arbitrary even after the insurer
    relied on several independent experts and a record such as this.6
    Such a premise is in considerable tension with the standard of
    review we use, which requires deference to the insurer's decision
    under       both   Supreme   Court   and    our   circuit's   precedent.7    See
    
    Firestone, 489 U.S. at 111
    ; see, e.g., Terry v. Bayer Corp., 
    145 F.3d 28
    , 37 (1st Cir. 1998). Zurich's interpretation of the Policy
    is "by no means unreasonable and so must prevail."                 Dutkewych v.
    Standard Ins. Co., 
    781 F.3d 623
    , 636 (1st Cir. 2015) (quoting
    Wallace v. Johnson & Johnson, 
    585 F.3d 11
    , 15 (1st Cir. 2009)).
    The Supreme Court reminded us in Conkright v. Frommert,
    
    559 U.S. 506
    , 517 (2010), of the importance of giving deference to
    6 Arruda cites Buffonge v. Prudential Insurance Co. of
    America, 
    426 F.3d 20
    (1st Cir. 2005), for the proposition that we
    should carefully scrutinize the medical opinions for the allegedly
    missing causation analysis. We disagree that Buffonge aids her.
    In Buffonge, we held that the insurer's decision was arbitrary and
    capricious because it relied on the opinion of an expert who had
    clearly misrepresented the opinions of other experts, an error
    that should have been obvious to the insurer on any reasonable
    review of the 
    record. 426 F.3d at 28-29
    . No such evidence of
    misrepresentation by any doctor is presented here; indeed, both
    Dr. Taff and Dr. Laposata relied on the same information.
    7    We certainly may not, as the dissent proposes, develop
    our own theories not present in the record, like theorizing that
    Arruda may have fallen asleep because of stress at work, to find
    an insurer's decision arbitrary.
    - 27 -
    claims fiduciaries such as Zurich.          As the Court noted, such
    "[d]eference promotes efficiency by encouraging resolution of
    benefits    disputes   through   internal   administrative   proceedings
    rather than costly litigation," "predictability, as an employer
    can rely on the expertise of the plan administrator rather than
    worry about unexpected and inaccurate plan interpretations that
    might result from de novo judicial review," and "uniformity,
    helping to avoid a patchwork of different interpretations of a
    plan . . . that covers employees in different jurisdictions."            
    Id. We are
    aware that a few other circuits, in reviewing
    whether    something   "contributed   to"   a   covered   loss   under    an
    insurance policy, have chosen to adopt a "substantial factor" test
    to aid their interpretation.      Under the "substantial factor" test,
    "a pre-existing infirmity or disease is not to be considered as a
    cause unless it substantially contributed to the disability or
    loss."     Adkins v. Reliance Standard Life Ins. Co., 
    917 F.2d 794
    ,
    797 (4th Cir. 1990) (emphasis added) (quoting Colonial Life & Acc.
    Ins. Co. v. Weartz, 
    636 S.W.2d 891
    , 894 (Ky. Ct. App. 1982)); see
    also Dixon v. Life Ins. Co. of N. Am., 
    389 F.3d 1179
    , 1184 (11th
    Cir. 2004); McClure v. Life Ins. Co. of N. Am., 
    84 F.3d 1129
    , 1136
    (9th Cir. 1996).8      The standard of review in this case, as all
    8    The Tenth Circuit has adopted a "plain meaning" approach
    instead of a "substantial factor" test.     See Pirkheim v. First
    Unum Life Ins., 
    229 F.3d 1008
    , 1010 (10th Cir. 2000). Again, we
    rely on our own circuit law.
    - 28 -
    parties agree, is for abuse of discretion.                In our view, the
    substantial factor test is in tension with our circuit law on the
    abuse of discretion test.
    Further, as we have said, "our review of whether a plan
    administrator abused its discretion does not require that we
    determine either the 'best reading' of the ERISA plan or how we
    would read the plan de novo."        D & H Therapy Assocs., 
    LLC, 640 F.3d at 35
    .      Our existing circuit law addresses the appropriate
    test for abuse of discretion review issues.
    We also keep in mind the Supreme Court's admonition in
    Conkright that, in passing ERISA, Congress desired "to create a
    system that is not so complex that administrative costs, or
    litigation expenses, unduly discourage employers from offering
    ERISA plans in the first 
    place." 559 U.S. at 517
    (alterations and
    internal quotation marks omitted).
    III.
    Zurich's    determination     that    Mr.   Arruda's   death   was
    caused or contributed to by pre-existing medical conditions was
    supported   by    substantial   evidence    and    was   not   arbitrary   or
    capricious.      We reverse and remand for entry of summary judgment
    for Zurich.      No costs are awarded.
    -Dissenting Opinion Follows-
    - 29 -
    LIPEZ, Circuit Judge, dissenting.               I agree with my
    colleagues on the legal principles that govern our review in this
    case.     We part ways, however, in applying that law to the record
    before us.    Although the majority reasons otherwise, Zurich cannot
    defend its conclusion that Mr. Arruda's heart disease or other
    pre-existing conditions caused or contributed to his car accident
    and death.      As I shall explain, the record inescapably reveals
    that Zurich denied Mrs. Arruda's claim for the reason aptly
    described by the district court: "the mere existence of [Mr.
    Arruda's] pre-existing illness."           Arruda v. Zurich Am. Ins. Co.,
    
    366 F. Supp. 3d 175
    , 186 (D. Mass. 2019).                   That flawed logic
    produces an unjust result.
    Because    Zurich's     decision       is    not     supported   by
    substantial evidence, my colleagues err in reversing the district
    court's     judgment   for   Mrs.   Arruda.     I     therefore    respectfully
    dissent.
    I.
    As the majority notes, following Mr. Arruda's death, his
    ICD   was    removed   and   submitted     to   the      manufacturer,   Boston
    Scientific, for analysis.       The arrhythmia logbook report generated
    by Boston Scientific -- i.e., the record of cardiac "events"
    measured by the ICD -- shows no events after May 20, 2014, two
    days before the car crash.          The report also shows that a "Rhythm
    ID Update" was completed about an hour before the crash, at 8:23
    - 30 -
    a.m. on May 22.      Faced with these facts, Zurich argues on appeal
    that the logbook functions in a particular way:
    The Logbook last updated at 8:23 a.m. on the
    day of the crash.       The fact that the
    defibrillator was intact and working at the
    time of Mr. Arruda's death means that the
    Logbook does not update continuously in real
    time. The Logbook shows that Mr. Arruda did
    not experience a cardiac event before 8:23,
    but it is silent as to what happened in the
    hour leading up to the 9:30 crash. It does
    not even record the alleged seizure observed
    by witnesses after the crash or that Mr.
    Arruda's   heart  stopped   beating  shortly
    thereafter.
    Zurich's assertion that the logbook did not record continuously in
    real time appears to be an attempt to support its suggestion that
    Mr. Arruda experienced a cardiac event at the time of the crash
    that   had   not   yet   been   recorded.    However,   Zurich   offers   no
    evidentiary support for its depiction of how the ICD operated.
    In fact, none of the medical experts describe the ICD as
    functioning in the way that Zurich argues.         Nor do they place any
    significance on the absence from the ICD logbook report of Mr.
    Arruda's seizure or his heart stoppage.            Four medical experts
    rendered opinions about the accident, but only three mention the
    ICD.   And only one, Mrs. Arruda's expert, directly opines on the
    meaning of the logbook report.
    To be specific, one of Zurich's experts, Dr. Bell,
    mentions the ICD itself, but not the logbook report.              Dr. Bell
    notes that "the ICD was normally working and not activated prior
    - 31 -
    to the crash" based on State Trooper William McMillan's paraphrase
    of the autopsy results in an accident report.              He then opines that
    Mr.   Arruda's    "crash    and   his   death     were   caused   by    his   heart
    disease."        Another    Zurich      expert,    Dr.   Taff,    states      that,
    "[a]ccording to several reports, post-mortem analysis of [Mr.
    Arruda]'s implantable ICD device showed no evidence of an ante-
    mortem arrhythmia."        Despite his acknowledgement that there was no
    evidence of an arrhythmia, Dr. Taff lists "cardiac arrhythmia
    resulting from pre-existing heart disease" as one of the "several
    possible    pre-existing     illnesses     or     diseases"   that     caused   the
    accident.
    Mrs. Arruda's expert, Dr. Laposata, authored two reports
    about the accident, the first before Dr. Taff rendered his opinion
    and the second afterwards.           In her initial report, Dr. Laposata
    notes that "interrogation of the internal cardiac defibrillator
    did not show any abnormal heart rhythms prior to the accident."
    In her supplemental report, Dr. Laposata responds to Dr. Taff's
    findings with an explicit opinion that "[t]he interrogation of
    [Mr. Arruda's] cardiac defibrillator gives definitive proof that
    no cardiac arrhythmia or event preceded the accident."9                  There is
    no evidence in the record rebutting that statement.
    9The majority criticizes Dr. Laposata for not explicitly
    stating in her first report that "the absence of data show[s] that
    no abnormal heart rhythm had occurred between 8:23 a.m. and the
    later time of the accident."      Supra Section I.C.     But that
    - 32 -
    II.
    Zurich concluded that Mr. Arruda's death is not covered
    under the Policy because it was "caused by, contributed to, or
    result[ed] from . . . illness or disease," i.e., Mr. Arruda's heart
    disease or some other pre-existing condition, and marijuana use.
    There is not substantial evidence in the record to support either
    factor.
    A.   Illness or Disease
    Mr. Arruda's autopsy did not reveal evidence of a heart
    attack or heart failure.    Cf. Dixon v. Life Ins. Co. of N. Am.,
    
    389 F.3d 1179
    , 1181 (11th Cir. 2004) (undisputed cause of driver's
    death following car crash was "heart failure" where autopsy showed
    "complete blockage of one of the main arteries that supplies blood
    to the heart" and "no evidence of external injury"); Vickers v.
    Bos. Mut. Life Ins. Co., 
    135 F.3d 179
    , 180 (1st Cir. 1998)
    (undisputed that fatal car crash was caused by driver's heart
    attack where autopsy showed he had suffered an "acute coronary
    insufficiency"). In an ordinary case, the absence of such physical
    evidence may not be determinative because it does not rule out an
    conclusion is implicit in her statement that interrogation of the
    ICD showed no abnormal heart rhythms prior to the accident. If
    Dr. Laposata understood the logbook report to be inconclusive as
    to what happened after the Rhythm ID Update was recorded, she would
    have said only that the ICD showed no abnormal heart rhythms prior
    to 8:23 a.m. Both of Dr. Laposata's reports reflect her consistent
    opinion that the logbook report shows no evidence of an arrhythmia
    prior to the accident itself.
    - 33 -
    arrhythmia.    But Mr. Arruda had an ICD, the very purpose of which
    was to measure cardiac irregularities.          The ICD logbook report is,
    therefore, a critical piece of medical evidence that bears upon
    the reasonableness of Zurich's decision.10
    Mrs. Arruda submitted the logbook report to Zurich when
    she appealed from its decision denying benefits, and she later
    submitted the two expert reports by Dr. Laposata that discuss the
    report.    Yet Zurich did not mention the logbook report in its
    letter denying Mrs. Arruda's appeal.          Suggesting that somehow this
    disregard is a factor in Zurich's favor, the majority emphasizes
    that Zurich did not rely on the logbook report to deny Mrs.
    Arruda's claim for benefits.       Zurich's choice not to engage with
    a   critical   piece   of   evidence   does    not   weigh   in   its   favor.
    Recognizing the import of this failure, Zurich now argues belatedly
    that the logbook report is "inconclusive," a position that my
    colleagues insist is reasonable.       Supra Section II.B.        I disagree.
    Dr. Laposata is the only medical expert who actually interpreted
    the logbook report, and her unrebutted opinion is that the logbook
    report "gives definitive proof that no cardiac arrhythmia or event
    10
    Although the district court expressed uncertainty about the
    meaning of the "Rhythm ID Update," it concluded that the logbook
    report "underscore[s]" the speculative nature of a conclusion that
    heart disease was the cause of Mr. Arruda's death. Arruda, 366 F.
    Supp. 3d at 185 n.4.
    - 34 -
    preceded the accident."11         If Zurich believed that the logbook did
    not record cardiac irregularities in real time, and therefore it
    had doubts about Dr. Laposata's interpretation, it should have
    challenged her opinion with a second opinion.                      Zurich was not
    entitled,     however,     to      ignore     the     only        medical   expert
    interpretation of the logbook report in the record and now, on
    appeal,    dismiss   the   significance       of    the    logbook    report    with
    conjecture about how it works.
    The absence of any evidence of a heart attack, heart
    failure,    arrhythmia,      or    other    cardiac       event    undermines    the
    reasonableness of Zurich's denial of benefits on that basis.
    Nevertheless, the majority says that this focus on heart disease
    "misses the point," citing to Dr. Taff's list of "possible medical
    conditions that, singly or in combination, caused or contributed
    to Mr. Arruda's death."           Supra Section II.B.           It is enough, the
    majority says, that Dr. Taff reached a "self-evidently reasoned"
    conclusion    that   "some    manifestation(s)        of     Mr.    Arruda's    pre-
    existing conditions" caused the accident.                 
    Id. What is
    a "self-
    11The majority suggests that the opinions of Dr. Bell and Dr.
    Taff rebut Dr. Laposata's conclusion about the significance of the
    logbook report. They do not. Dr. Bell noted only that the ICD
    was "normally working and not activated prior to the crash," and
    Dr. Taff stated that "post-mortem analysis of [Mr. Arruda]'s
    implantable ICD device showed no evidence of an ante-mortem
    arrhythmia." Yet both experts then concluded that Mr. Arruda's
    heart disease contributed in some way to the car crash, without
    explaining how those conclusions are compatible with the absence
    of any cardiac irregularity readings in the logbook.
    - 35 -
    evidently reasoned" conclusion? One that relies on purported logic
    instead of evidence?     One that posits that a man with so many pre-
    existing conditions must have gotten into a sudden and unexplained
    accident because of those conditions?          That "reasoning" is nothing
    more than speculation.
    The   majority    emphasizes     that   Dr.   Taff    rendered   his
    opinion "to a reasonable degree of forensic medical certainty."
    Supra Section II.B.      His use of the phrase "reasonable degree of
    forensic medical certainty," the indispensable ultimate assertion
    in   any   testimony   from    a   medical    expert,     has    no   talismanic
    significance.     Its probative force depends on the quality of the
    evidence    underlying   it.       Here    that     underlying    evidence   is
    strikingly feeble.     Dr. Taff lists a grab-bag of seven "possible"
    causes. Included in the list are "cardiac arrhythmia," even though
    the ICD had not recorded a cardiac event, and "complications of
    undiagnosed sleep apnea resulting in falling asleep behind the
    wheel."    In fact, despite the absence of any medical history of
    sleep apnea (hence Dr. Taff's reference to "undiagnosed sleep
    apnea"), Dr. Taff suggests that Mr. Arruda fell asleep behind the
    wheel:
    Although [Mr. Arruda] was never diagnosed with
    sleep apnea, several of his pre-existing
    pathological conditions are known to cause
    irregular    sleeping   patterns,    breathing
    difficulties, chronic fatigue and obesity.
    Based on the circumstances, there is a good
    - 36 -
    chance that [Mr. Arruda] fell asleep behind
    the wheel.
    This    "good   chance"    conclusion    discomforts     the   majority.     My
    colleagues treat it as an unwelcome and irrelevant gloss on Dr.
    Taff's      obligatory     "reasonable    degree    of    forensic    medical
    certainty" observation.           
    See supra
    Section II.B.       They say that
    Zurich could ignore it in favor of Dr. Taff's more congenial and
    formally correct observation.          But that "good chance" observation
    reveals the speculative nature of Dr. Taff's opinion about the
    relationship     between    Mr.    Arruda's   medical    conditions   and   the
    accident.
    The inescapable fact is that many healthy people fall
    asleep at the wheel while driving, and many sick people fall asleep
    at the wheel while driving for reasons that have nothing to do
    with their illness.        Mr. Arruda left his home in Bristol, Rhode
    Island, around 6:30 a.m. on the day of the accident to drive to
    Amherst, Massachusetts, a distance of about 105 miles,12 for a work
    event.      At the time of the accident, Mr. Arruda was about ten
    minutes from the University of Massachusetts Amherst,13 where the
    12
    Driving Directions from Bristol, RI, to Amherst, MA, Google
    Maps, http://maps.google.com (search for "Amherst, MA"; then click
    "Directions" and enter "Bristol, RI" as the starting point).
    13
    Driving Directions from 73 Russell Street, Hadley, MA, to
    the   University   of   Massachusetts   Amherst,   Google   Maps,
    http://maps.google.com (search for "University of Massachusetts
    Amherst" and click on the first result; then click "Directions"
    and enter "73 Russell Street, Hadley, MA" as the starting point).
    - 37 -
    event was being held.      Perhaps he had a sleepless night because he
    was worried about getting to the event on time.              Even if Dr. Taff
    is correct that Mr. Arruda fell asleep at the wheel (a speculative
    conclusion in itself), there is as good a chance that he fell
    asleep because of work anxiety as there is that he fell asleep
    because of undiagnosed sleep apnea.
    My colleagues suggest that the parties' dispute comes
    down to a battle of the experts between Dr. Taff and Dr. Laposata.
    
    See supra
    Section II.B.         But that is not so.        Indeed, on perhaps
    the most essential point, the opinions of Dr. Taff and Dr. Laposata
    are not in conflict.       Dr. Taff acknowledges that "[t]here is no
    way to scientifically prove which human factor(s)/pre-existing
    medical condition(s) occurred during the pre-collision phase of
    the accident."     Dr. Laposata likewise observes that "[t]here is no
    medical or scientific data to conclude that the accident was caused
    or   contributed     to    by     Mr.    Arruda's     pre-existing        medical
    conditions."       The    two     experts     diverge,    however,   in     their
    willingness to speculate about what happened despite the lack of
    supportive medical evidence.
    Dr. Laposata does not purport to know what occurred prior
    to   the   accident.       Like    Dr.      Taff,   she   rules   out     several
    possibilities, including a heart attack or other "acute natural
    event incompatible with life" -- because the autopsy revealed no
    evidence of such an event -- and "incapacitation by heart disease"
    - 38 -
    -- because the ICD logbook report "showed no abnormal heart rhythms
    recorded prior to the collision."       But she asserts that "[i]t is
    a serious error to conclude that the mere existence of medical
    diagnoses and speculation as to what might happen given these
    conditions equates with proof that a medical event did occur prior
    to the accident."   I agree.
    I recognize that Zurich does rely on other records, in
    addition to Dr. Taff's report, to support the determination that
    heart disease caused or contributed to Mr. Arruda's crash: the
    autopsy report and death certificate prepared by Dr. Sexton, the
    Massachusetts Collision Reconstruction Report completed by Trooper
    Sanford, and the two other medical expert reports written by Dr.
    Bell and Dr. Angell.   Although this list gives the appearance of
    substantiality, the appearance does not survive scrutiny.
    The front page of Dr. Sexton's autopsy report reads, in
    relevant part, as follows:
    CAUSE OF DEATH: Hypertensive Heart Disease.
    Contributory Factors: Upper Cervical        Spine
    Fracture due to Blunt Impact.
    MANNER OF DEATH: Accident (Driver Involved in
    a Motor Vehicle Collision with Rollover)
    The death certificate also states that the immediate cause of death
    was hypertensive heart disease.14 But, as the district court noted,
    14 The copy of the death certificate reproduced in the
    administrative record is illegible. Zurich, however, stated in
    its letter denying Mrs. Arruda's claim for benefits that "[t]he
    - 39 -
    "Dr. Sexton's report was based solely on an examination of Mr.
    Arruda, and did not include any examination of his defibrillator
    device."    
    Arruda, 366 F. Supp. 3d at 180
    .          In addition, Dr. Taff
    points out "discrepancies" in Dr. Sexton's preparation of the
    autopsy report which "suggest that Dr. Sexton never took the
    . . . cardiac findings into consideration before finalizing his
    opinions about [Mr. Arruda]'s cause and manner of death."                  Dr.
    Sexton's cause of death determination, which was reached without
    consideration of all of the relevant medical evidence, is therefore
    unreliable.
    Trooper Sanford states in his accident report that Mr.
    Arruda suffered from some kind of medical event that caused the
    crash.    That opinion is baseless.     As the district court observed,
    "[t]he record does not indicate Trooper Sanford has meaningful
    medical training in this area."              
    Id. at 185.
           Indeed, Zurich
    appropriately concedes that the "State Police are not medical
    experts    and   their   opinions    could     not   be   the    basis   for   a
    determination that heart disease was the cause of death."
    Dr. Bell opines in his medical expert report that Mr.
    Arruda's
    crash and his death were caused by his heart
    disease, whether it be due to hypertension or
    a variant of [hypertrophic cardiomyopathy].
    However, based on the autopsy results, [Mr.
    Death Certificate stated that the immediate cause of death was
    Hypertensive Heart Disease."
    - 40 -
    Arruda's] C1 left posterior arch fracture and
    C3-C4 dislocation with soft tissue hemorrhage
    at the injury sites would be a contributory
    cause of death.
    He does not explain how or why he concludes that Mr. Arruda's heart
    disease caused the car crash and Mr. Arruda's death.    It appears,
    however, that he relied on the flawed autopsy report.
    Finally, the district court correctly found that Dr.
    Angell's report is "unreliable" because his "credentials are not
    contained in the record, and Zurich could not even identify [him]."
    
    Id. In addition,
    his brief conclusory opinion provides no basis
    for his findings.
    In sum, the record lacks substantial medical evidence
    that bridges the gap between Mr. Arruda's pre-existing conditions,
    which he had been living with for years, and the cause of the fatal
    car accident.   Without more, Zurich's decision amounts to a denial
    of benefits based on the mere existence of Mr. Arruda's pre-
    existing conditions.    But it is not enough to reason that an
    indisputably sick man must have had the fatal car accident because
    of his sickness. Zurich's denial of benefits based on Mr. Arruda's
    medical conditions, singly or in combination, is not "reasonable
    and supported by substantial evidence on the record as a whole."
    See McDonough v. Aetna Life Ins. Co., 
    783 F.3d 374
    , 379 (1st Cir.
    2015).
    - 41 -
    B.   Marijuana
    Zurich's decision to rely on the narcotics exclusion is
    unreasonable for similar reasons.                   Dr. Taff's assertion that the
    marijuana in Mr. Arruda's system alone "would have impaired his
    ability   to      operate      his   motor    vehicle"        is    undermined         by   his
    acknowledgement       that      "[r]esponses        to   marijuana            vary   from   one
    person to another and precise and predictable behavioral and
    physiological reactions to the drug cannot be rendered."                                 As the
    district court correctly observed, "[t]here is no evidence in the
    record regarding how the marijuana in Mr. Arruda's system may or
    may not have impaired his driving and caused the car accident."
    
    Arruda, 366 F. Supp. 3d at 187
    .                   Notably, the majority does not
    even   attempt      to    defend      Zurich's       reliance        on       the    narcotics
    exclusion.
    III.
    In    rejecting         the   decision      of        the    district        court
    overturning Zurich's denial of benefits, the majority questions
    the "premise" that "judges may find insurers' decisions as to
    benefits to be arbitrary even after the insurer relied on several
    independent experts and a record such as this," observing that
    "[s]uch a premise is in considerable tension" with the abuse of
    discretion standard of review.               Supra Section II.C.                 There is no
    such tension here.             We have said many times that a standard of
    deference      does      not    negate      our     obligation           to     ensure      that
    - 42 -
    "substantial evidence" underlies the decisions of insurance plan
    administrators.   The district court met that obligation and so
    should we.   Quantity is not a proxy for substance.   Here, when the
    450 or so pages15 of documentation reviewed by Zurich are fairly
    examined, they are devoid of the substantial evidence required by
    law to support Zurich's denial of benefits.           I respectfully
    dissent.
    15 Dr. Taff noted that he reviewed a "450-page file" of
    documentary evidence when he prepared his report.
    - 43 -