Kerry L. Erickson, V. Pharmacia Llc. ( 2024 )


Menu:
  •         IN THE COURT OF APPEALS OF THE STATE OF WASHINGTON
    KERRY L. ERICKSON; MICHELLE M.
    LEAHY; RICHARD A. LEAHY; and                            No. 83287-5-I
    JOYCE E. MARQUARDT;
    DIVISION ONE
    Respondents,
    PUBLISHED OPINION
    v.
    PHARMACIA LLC, a Delaware limited
    liability company, f/k/a Pharmacia
    Corporation,
    Appellant.
    CHUNG, J. — Three former teachers in Monroe School District filed this
    lawsuit under the Washington product liability act (WPLA) against Pharmacia, the
    successor to Monsanto, the manufacturer of chemicals known as polychlorinated
    biphenyls (PCBs). 1 They claimed their exposure to PCBs in fluorescent light
    ballast capacitors (FLBs) at the Sky Valley Education Center (SVEC) caused
    them serious health impacts. After a seven-week trial, a jury found Pharmacia
    liable and awarded plaintiffs over $185 million in compensatory and punitive
    damages. The trial court denied Pharmacia’s post-trial motion for judgment as a
    1 Along with Monsanto, the complaint also named as defendants Pharmacia LLC, which
    is successor to the part of Monsanto that manufactured PCBs, and Solutia, Inc., to which
    Monsanto’s chemical business was “spun-off.” By the time of trial, Pharmacia was the only
    remaining defendant. We refer to Monsanto when discussing the original entity that manufactured
    the PCBs, but refer to the defendants in this action collectively as Pharmacia.
    No. 83287-5-I/2
    matter of law or, alternatively, for a new trial and for remittitur of the damage
    awards.
    Pharmacia’s appeal alleges numerous errors by the trial court, including
    improper application of Missouri law on the statute of repose and punitive
    damages; failure to instruct the jury on two defenses; erroneous admission of
    expert opinions on exposure, injury, and causation; and erroneous admission of
    injury to non-parties. Pharmacia also challenges the trial court’s determination
    that substantial evidence supported the liability verdicts and the denial of
    remittitur of the punitive damages awards.
    We hold that when WPLA provides the applicable law on liability, WPLA’s
    statute of repose is not subject to a separate choice-of-law analysis because it is
    a claim-defining limitation on liability. We further hold that WPLA’s statute of
    repose does not violate the Washington Constitution’s privileges and immunities
    clause. Based on our conclusion that WPLA’s statute of repose applies and that
    it is constitutional, we remand to the trial court for further proceedings, including
    assessing whether Erickson’s claims are subject to WPLA’s statute of repose.
    We also address additional issues that are likely to recur in a new trial. As
    to choice of law on punitive damages, applying the “most significant relationship”
    test, we conclude that Missouri has the greater interest regarding allowing
    punitive damages for liability claims that are cognizable under its own products
    liability law. Thus, while the court properly may allow the jury to consider punitive
    2
    No. 83287-5-I/3
    damages, such damages cannot be awarded for claims not allowed under
    Missouri law, including, as relevant here, post-sale failure to warn.
    As to Pharmacia’s requested jury instruction related to whether the PCB-
    containing capacitor or the FLB that incorporated the capacitor was the relevant
    product, rather than PCBs, we hold that the court properly declined to give this
    instruction. A WPLA claim does not require proof of product failure. And because
    the Washington Supreme Court has rejected the “sophisticated purchaser”
    defense to product liability, the trial court properly denied that instruction as well.
    Regarding Pharmacia’s challenges to admission of expert testimony, we
    conclude that industrial hygienist Kevin Coghlan’s expert testimony that
    calculated the historical levels of PCBs at SVEC by using methodologies he
    developed based on data from two other studies was novel and should have
    been excluded under Frye. 2 However, the trial court properly allowed Coghlan to
    testify about direct comparisons with data from the study of New York schools.
    The trial court also properly allowed the testimony of neuropsychologist Dr.
    Perrillo regarding causation of plaintiffs’ injuries, as he was qualified as an expert
    and his testimony met the standards of ER 702. The court also properly allowed
    toxicologist Dr. James Dahlgren to testify as an expert on causation, except to
    the extent his testimony relied on any inadmissible testimony by Coghlan about
    historical PCB levels.
    2 Frye v. United States, 
    54 App. D.C. 46
    , 
    293 F. 1013
     (1923).
    3
    No. 83287-5-I/4
    Further, the admission of evidence of nonparty harms through Dr. Perrillo
    and Dr. Mahoney’s expert testimony was not an abuse of discretion. Information
    about other people’s illnesses was not admitted as substantive evidence, but as
    the basis for their opinions, and it was of the type reasonably relied upon by such
    experts. Lay witness testimony was limited to personal observations of others at
    the school being ill, not causation. This lay testimony was also properly admitted.
    Because we conclude that WPLA’s statute of repose applies to claims
    brought under WPLA, rather than Missouri law, which has no statute of repose,
    we must reverse the jury verdict and remand for further proceedings.
    FACTS
    I.     Polychlorinated Biphenyls
    In 1929, the Swann Chemical Company invented a family of chemicals
    called polychlorinated biphenyls (PCBs). PCBs are formed by adding chlorine
    atoms to benzene rings. Adding different numbers of chlorine atoms to various
    locations on the benzene rings creates individual types of PCBs, called
    congeners. Scientists have made 209 individual PCB congeners. Congeners with
    more added chlorine, considered “heavier,” were used as plasticizers to keep
    certain materials, such as caulk and paint, from becoming brittle. The lighter
    chlorinated PCBs were used as dielectric fluid in transformers and capacitors.
    Capacitors are used in small electrical devices, such as fluorescent lights.
    Capacitors in a fluorescent light ballast help start the light and smooth out the
    voltage cycling to reduce flickering. Prior to PCBs, electrical equipment relied on
    4
    No. 83287-5-I/5
    mineral oil that had a potential risk of explosion and fire. PCBs are chemically
    very stable and do not burn. As a result, electrical companies such as General
    Electric (GE) quickly adopted the use of these non-flammable PCBs as the
    electric fluid in transformers and capacitors, in place of mineral oil. PCB use in
    capacitors and transformers led to a considerable reduction in the potential for
    fires. GE obtained a patent for these uses in 1933. The World Almanac
    designated the discovery of PCBs, and its application in electrical equipment, as
    one of the world’s greatest inventions and scientific discoveries.
    II.    Monsanto’s PCB Business
    In 1935, Monsanto Company, with its principal place of business in St.
    Louis, Missouri, purchased Swann and began selling PCBs under the name
    Aroclor. Monsanto commissioned toxicity studies of PCB from the 1930’s to the
    1960’s and shared the results with its customers and potential customers. By
    1943, Monsanto had entered into an agreement with the U.S. Public Health
    Service to include on its bills of sales for PCBs that any products including PCBs
    must include a warning to “avoid repeated contact with the skin and inhalation of
    the fumes and dusts.” However, by 1952, Monsanto bills of sales did not include
    this language, because their customers, manufacturers of products that included
    PCBs, would not like being told they had to put such a warning label on their
    products.
    In late 1966, scientists warned that heavier chlorinated PCBs were
    remaining in the environment rather than biodegrading or metabolizing.
    5
    No. 83287-5-I/6
    Monsanto provided PCB samples for research and notified its major customers.
    By 1969, Monsanto had developed a plan to address the environmental
    concerns, which included warning its customers, ceasing production of these
    heavier chlorinated congeners, and buying back millions of pounds of PCBs for
    safe incineration. Monsanto also provided its distributors with suggested
    warnings to provide their customers in order to distribute the information down
    the supply chain, including electrical customers such as GE and Westinghouse.
    By 1970, Monsanto decided to begin phasing out production of higher
    chlorinated congeners. GE became concerned the company would also cease
    production of the lower chlorinated PCBs used in electrical applications, and in
    early 1970, GE met with Monsanto to advocate for continued manufacture of
    these PCBs because they had no comparable replacement in this application. In
    1973, Monsanto informed the government that it would sell PCBs to the electrical
    industry only for use in closed systems as dielectric fluid in transformers and
    capacitors. In 1975, Monsanto announced it would voluntarily cease all PCB
    production and sales and exited the PCB business in 1977. The Environmental
    Protection Agency (EPA) banned the manufacture of PCBs in 1979. After that
    time, newly manufactured electrical products could not contain PCBs. PCBs have
    since been banned in every country. However, PCB-containing light ballasts and
    transformers remained in operation for decades.
    6
    No. 83287-5-I/7
    III.   Sky Valley Educational Center (SVEC)
    The old Middle School buildings in the Monroe School District (MSD) were
    built in 1967-68, at a time when PCBs were used extensively in caulking and
    FLBs. As a result, 95 percent of the FLBs contained PCBs. During the 1990s and
    early 2000s, the EPA and other public health entities sent information to school
    districts about FLBs containing PCBs. As early as 2000, MSD was aware of
    PCB-containing FLBs in the old Middle School buildings. At that time, MSD
    learned it needed to be checking FLBs to determine which contained PCBs. MSD
    knew to remove and dispose of leaking FLBs as hazardous waste. MSD also
    knew that PCBs could pose a potential hazard to people in the buildings, and
    special care was required to remove leaking FLBs in order to prevent exposure.
    In 2007, an assessment of the old Middle School buildings reported that
    the deterioration of the buildings was the most severe of any school within MSD
    and they were in critical need of upgrades and improvements. Despite these
    findings, MSD moved the K-12 Sky Valley Educational Center (SVEC) into the
    old Middle School buildings in 2011.
    IV.    Reports of Illnesses at SVEC
    After moving to the old Middle School buildings, teachers, including Kerri
    Erickson, Michelle Leahy, and Joyce Marquardt, began experiencing declining
    health. Among their symptoms, these teachers reported headaches, fatigue,
    difficulty with memory, blurred vision, sinus symptoms, and respiratory issues.
    The symptoms improved when the teachers spent time outside of SVEC but
    7
    No. 83287-5-I/8
    returned when they went back to the buildings. They eventually became
    concerned that the SVEC buildings were making them sick. All three teachers left
    their positions at SVEC due to their health concerns.
    V.      Product Liability Lawsuit
    In May 2018, several SVEC teachers and families filed a lawsuit against
    Monsanto Company and its successor corporations for PCB contamination in
    SVEC. 3 They alleged that children and adults were exposed to PCBs at SVEC
    and have “been coping with adverse medical effects, including neurological
    damage, autoimmune and endocrine diseases, and cancers.” The complaint
    alleged both strict product liability and negligence claims that PCBs were not
    reasonably safe as designed and that Monsanto failed to provide warnings when
    PCBs were manufactured and after manufacturing. Since the filing of the original
    lawsuit, 200 other SVEC teachers, students, and family members have filed
    similar claims. A total of 17 lawsuits were filed over 13 months. This is the first of
    the many SVEC PCB lawsuits to reach this court on appeal after trial.
    The plaintiffs in this lawsuit are teachers Erickson, Leahy, and Marquardt. 4
    As pleaded by plaintiffs, the trial court applied the Washington product liability act
    (WPLA) for liability, but on plaintiffs’ motion, applied Missouri law to control the
    statute of repose and punitive damages. After an extensive trial, the jury found
    3 The lawsuit also named MSD, Union High School District, and Snohomish Health
    District as defendants.
    4 The lawsuit also includes Leahy’s husband. For the purposes of this opinion, we refer to
    the plaintiffs together as Erickson or the plaintiffs.
    8
    No. 83287-5-I/9
    that Monsanto supplied a product that was not reasonably safe in construction or
    as designed, was not reasonably safe because adequate warnings or
    instructions were not provided with the product or after manufacturing, and was
    the proximate cause of the plaintiffs’ injuries. Erickson received $15 million in
    compensatory damages; Leahy received $18 million in compensatory damages;
    and Marquardt received $17 million in compensatory damages. 5 Under Missouri
    law, the jury awarded each of the three teachers $45 million in punitive damages.
    The total judgment against Pharmacia amounted to over $185 million.
    After trial, Pharmacia brought unsuccessful motions for judgment as a
    matter of law or, in the alternative, for a new trial, and for remittitur of the damage
    awards. Pharmacia appeals.
    ANALYSIS
    I.     Choice of Law
    A threshold and outcome-determinative question is what state provides
    the applicable law for plaintiffs’ claims. Thus, we address these choice-of-law
    issues first.
    After filing claims under WPLA, Erickson moved for the application of
    Missouri law on the issues of the statute of repose and punitive damages. The
    trial court agreed with the plaintiffs, and Pharmacia appeals these choice-of-law
    decisions. Choice of law is a question of law reviewed de novo. Erwin v. Cotter
    Health Ctrs., 
    161 Wn.2d 676
    , 691, 
    167 P.3d 1112
     (2007).
    5 Leahy’s husband received $150,000 for loss of consortium.
    9
    No. 83287-5-I/10
    Choice of law applies on an issue-by-issue basis. FutureSelect Portfolio
    Mgmt., Inc. v. Tremont Grp. Holdings, Inc., 
    175 Wn. App. 840
    , 856 n.15, 
    309 P.3d 555
     (2013). Under this principle, known as dépeçage, “different issues in a
    single case arising out of a common nucleus of facts may be decided according
    to the substantive law of different states.” 
    Id.
    Before engaging in a choice-of-law analysis, the court must first consider
    whether an actual conflict exists between the laws or interests of Washington and
    another state. Rice v. Dow Chem. Co., 
    124 Wn.2d 205
    , 210, 
    875 P.2d 1213
    (1994). An actual conflict arises when the outcome of an issue differs depending
    on which state’s law applies. Pope Res. LP v. Certain Underwriters at Lloyd’s,
    London, 19 Wn. App. 2d 113, 124, 
    494 P.3d 1076
     (2021). Without an actual
    conflict, the presumptive local law applies. Rice, 
    124 Wn.2d at 210
    .
    If there is a conflict, Restatement (Second) of Conflict of Laws § 6 (Am.
    Law Inst. 1971) identifies factors relevant to the choice of the applicable rule of
    law:
    (a) the needs of the interstate and international systems,
    (b) the relevant policies of the forum,
    (c) the relevant policies of other interested states and the relative
    interests of those states in the determination of the particular
    issue,
    (d) the protection of justified expectations,
    (e) the basic policies underlying the particular field of law,
    (f) certainty, predictability and uniformity of result, and
    (g) ease in the determination and application of the law to be
    applied.
    For tort claims, Washington has adopted the “most significant relationship rule.”
    Johnson v. Spider Staging Corp., 
    87 Wn.2d 577
    , 580, 
    555 P.2d 997
     (1976).
    10
    No. 83287-5-I/11
    Restatement (Second) of Conflict of Laws § 145 establishes the general
    principles 6 for resolving conflicts of law in tort cases using this approach:
    (1) The rights and liabilities of the parties with respect to an issue in
    tort are determined by the local law of the state which, with respect
    to that issue, has the most significant relationship to the occurrence
    and the parties under the principles stated in § 6.
    (2) Contacts to be taken into account in applying the principles of
    § 6 to determine the law applicable to an issue include:
    (a) the place where the injury occurred,
    (b) the place where the conduct causing the injury occurred,
    (c) the domicil, residence, nationality, place of incorporation
    and place of business of the parties, and
    (d) the place where the relationship, if any, between the
    parties is centered.
    These contacts are to be evaluated according to their relative
    importance with respect to the particular issue.
    In applying this test, courts evaluate the contacts quantitatively and qualitatively
    to determine the most significant contacts as they relate to the issue at hand.
    Martin v. Goodyear Tire & Rubber Co., 
    114 Wn. App. 823
    , 830, 
    61 P.3d 1196
    (2003). If the contacts are evenly balanced between the two states, the court
    evaluates the interests and public policies of the concerned states to determine
    which has a greater interest in determination of the particular issue. Zenaida-
    Garcia v. Recovery Sys. Tech., Inc., 
    128 Wn. App. 256
    , 260-61, 
    115 P.3d 1017
    (2005).
    6 Comment a to § 145 notes that there are other subsections that deal with particular
    torts, §§ 146-155, and particular issues in tort cases, §§ 156-174. RESTATEMENT (SECOND) § 145.
    For instance, § 146 addresses personal injuries, and issues addressed include defenses (§ 161)
    and damages (§ 171). See RESTATEMENT (SECOND).
    11
    No. 83287-5-I/12
    A.       Choice of Law as to Statute of Repose
    Because Erickson filed her claims under WPLA, she voluntarily selected
    WPLA as the substantive law for her claim. 7 Pharmacia did not challenge this
    choice. Thus, there is no issue regarding which state’s law applies to determine
    liability. However, Erickson sought to apply Missouri law to the issue of the
    statute of repose.
    The parties do not dispute that Washington and Missouri have an actual
    conflict of law with respect to product liability statutes of repose. Missouri does
    not have a statute of repose in cases of liability for defective products. See Lay v.
    P&G Health Care, Inc., 
    37 S.W.3d 310
    , 321-22 (2000). In contrast, one of the
    salient features of WPLA is its statute of repose. Before conducting an analysis
    of which state has the most significant relationship, however, we must first
    determine whether it is appropriate to treat the statute of repose separately from
    the portions of WPLA imposing liability—in other words, whether dépeçage is
    appropriate here.
    “Statutes of repose provide time limits for bringing an action, but they ‘are
    “of a different nature than statutes of limitation.” ’ ” Bennett v. United States, 2
    Wn.3d 430, 440, 
    539 P.3d 361
     (2023) (quoting 1000 Va. Ltd. P’ship v. Vertecs
    Corp., 
    158 Wn.2d 566
    , 574, 
    146 P.3d 423
     (2006) (quoting Rice, 
    124 Wn.2d at 211
    )). A statute of repose “terminates a right of action after a specific time, even
    7 One of the product liability claims brought by plaintiffs, for post-sale failure to warn, is
    available in Washington but not in Missouri. See RCW 7.72.030(1)(c).
    12
    No. 83287-5-I/13
    if the injury has not yet occurred.” Rice, 
    124 Wn.2d at 212
    . While a statute of
    limitation bars a plaintiff from bringing an already accrued claim after a specified
    period, a statute of repose bars a claim from ever arising. See 
    id. at 211-12
    .
    Erickson points to several Washington cases that have applied the choice-
    of-law analysis to the WPLA statute of repose as support for the premise that the
    law of another state can govern the individual issue of the statute of repose. See
    Rice, 
    124 Wn.2d at 212
    , Martin, 
    114 Wn. App. at 829
    , Zenaida-Garcia, 
    128 Wn. App. at 266
    . Indeed, the Washington Supreme Court has acknowledged that a
    statute of repose is a substantive issue that may raise a conflict of law. 8 
    124 Wn.2d at 212
    .
    However, the courts in Rice, Martin, and Zenaida-Garcia were not asked
    to consider the specific question we are faced with here: whether the court
    should apply a separate choice-of-law analysis to the statute of repose in WPLA
    from the choice of law governing the substantive product liability claim. Instead,
    all three cases identified the conflicting laws as the statutes of repose, and began
    with the premise that the significant relationship test should be applied to the
    conflicting laws. Thus, to that end, the courts in those cases performed the “most
    significant relationship” analysis as set out in Restatement (Second) § 145.
    8 The treatment of statutes of repose differs from that of statutes of limitations. Rice, 
    124 Wn.2d at 211
    . Under Washington’s “borrowing statute,” RCW 4.18.020, limitation periods are
    governed by the law of the state whose substantive law applies to the claim as determined by
    choice-of-law analysis. Rice, 
    124 Wn.2d at 210
    .
    13
    No. 83287-5-I/14
    In Rice, a forest service employee had extensive and routine exposure to
    herbicides while working in Oregon. 
    124 Wn.2d at 207
    . He subsequently moved
    to Washington, where he had only minimal contact with the chemical. 
    Id.
     The
    plaintiff developed leukemia and brought product liability claims against Dow
    Chemical Company in a Washington court. 
    Id.
     The court noted that there was
    some question whether WPLA applied, but the parties had not briefed the issue,
    so it did not need to decide that issue to reach the conclusion that there was a
    conflict because of the differing statutes of repose. 
    Id. at 212
    . Then, the court
    weighed the significant relationships and determined that Oregon law applied to
    the claims and affirmed dismissal of the action due to the running of Oregon’s
    statute of repose. 
    Id. at 216-17
    .
    In another case involving Washington and Oregon statutes of repose,
    Oregon residents were involved in a fatal car accident in Washington after a
    metal ring flew off the wheel assembly of a commercial truck. Martin, 
    114 Wn. App. at 826
    . The truck belonged to an Oregon company hauling rock from an
    Oregon quarry to a destination in Washington. 
    Id. at 827
    . Plaintiff filed claims
    including negligence and product liability claims under Washington law in
    Washington. The tire manufacturer, an Ohio company, claimed the trial court
    should have applied Oregon’s statute of repose, which would have resulted in
    dismissal of the claims. 
    Id. at 828
    . Instead, the court applied Washington’s
    statute of repose, and the case went to trial, resulting in a verdict for the plaintiff.
    On appeal, this court reviewed the significant relationships and determined that
    14
    No. 83287-5-I/15
    the number of contacts with Washington was “not overwhelming,” but the fact
    that the injury occurred in Washington was “qualitatively significant.” 
    Id. at 831
    .
    Moreover, even though the plaintiffs were Oregon residents and an Oregon
    company owned the commercial truck that lost the metal ring, Oregon’s contacts
    did not exceed Washington’s. 
    Id. at 832
    . After considering the policy interests
    behind the states’ statutes of repose, the court concluded that Washington had
    the greater interest: “Washington’s interest in protecting persons from injuries
    from defective products within its borders outweighs Oregon’s interest in
    protecting a manufacturer whose product arrives in Oregon through the stream of
    commerce and subsequently causes injury to a third party in another state.” 
    Id. at 835
    . Thus, the court held that Washington’s statute of repose applied and
    affirmed the judgment for the plaintiff. 
    Id. at 836
    .
    In Zenaida-Garcia, Washington’s interests again resulted in application of
    the Washington statute of repose. There, an Oregon resident died in an industrial
    accident at his workplace in Oregon, and his estate brought product liability
    claims against the manufacturer of the machinery involved. 
    128 Wn. App. at 258
    .
    The manufacturer was a Washington corporation whose principal place of
    business was in Washington and that originally sold the equipment to a
    Washington company. 
    Id. at 261
    . This court determined that Oregon and
    Washington’s contacts were evenly balanced, but Washington had a greater
    policy interest in application of its law. 
    Id. at 263-66
    . As a result, this court
    15
    No. 83287-5-I/16
    reversed summary judgment, finding that Washington’s longer statute of repose
    applied and the claim was not barred. 
    Id. at 266
    .
    Thus, even though these cases involved a “most significant relationship”
    analysis based on a conflict between the statutes of repose, they fail to convince
    us that when there is no dispute as to which substantive product liability law
    applies, any other state’s statute of repose could supplant the claim-defining
    statute of repose in WPLA. 9
    We turn to WPLA itself to ascertain “the range of application intended by
    the legislature.” RESTATEMENT (SECOND) § 6 cmt. b. 10 “Statutes of repose
    eliminate an avenue of redress for injured litigants based on policy
    considerations.” Bennett, 2 Wn.3d at 454. WPLA includes a specific “length of
    time product sellers are subject to liability.” RCW 7.72.060. WPLA’s statute of
    repose establishes that “a product seller shall not be subject to liability to a
    claimant for harm under this chapter if the product seller proves by a
    preponderance of the evidence that the harm was caused after the product’s
    9 It is not apparent from the appellate opinions in Martin or Zenaida-Garcia whether the
    respective trial courts applied WPLA for liability—in Martin, at trial, or in Zenaida-Garcia, on
    remand. In Rice, because the court held that Oregon’s statute of repose applied and barred the
    claim, 
    124 Wn.2d at 217
    , no court needed to address what state’s law applied to determine
    liability.
    10 Regarding the intended range of application of a statute, comment b to § 6 notes that
    “[a] court will rarely find that a question of choice of law is explicitly covered by statute. That is to
    say, a court will rarely be directed by statute to apply the local law of one state, rather than the
    local law of another state, in the decision of a particular issue.” RESTATEMENT (SECOND) § 6 cmt.
    b. Thus, in deciding whether an issue falls within the intended range of application of a particular
    statute, “[t]he court should give a local statute the range of application intended by the legislature
    when these intentions can be ascertained and can constitutionally be given effect.” Id. Further,
    “[p]rovided it is constitutional to do so, the court will apply a local statute in the manner intended
    by the legislature even when the local law of another state would be applicable under usual
    choice-of-law principles.” Id.
    16
    No. 83287-5-I/17
    ‘useful safe life’ had expired.” RCW 7.72.060(1). The statute of repose includes a
    rebuttable presumption that if the harm was caused more than 12 years after
    delivery, the useful safe life has expired. RCW 7.72.060(2).Under WPLA’s statute
    of repose, a manufacturer is not liable for harm that occurs after the useful safe
    life expires. Martin, 
    114 Wn. App. at 828
    .
    The legislature purposefully crafted WPLA, including the statute of repose,
    to address two opposing concerns:
    When Washington’s tort reform act was enacted in 1981, the
    legislature announced its desire to balance two interests: protecting
    Washington industries from excessive litigation, and preserving the
    right of consumers to seek redress for injuries caused by unsafe
    products. The legislature enacted a 12-year statute of repose,
    which at the time was the nation’s longest for product liability . . . .
    Zenaida-Garcia, 
    128 Wn. App. at 264
    . The lengthy statute of repose protects
    consumers but also curtails businesses’ insurance costs for “product sellers and
    manufacturers,” as those costs disincentivize innovation and raise costs for
    consumers. LAWS OF 1981, ch. 27, § 1. Specifically, the legislature intended that
    “retail businesses located primarily in the state of Washington be protected from
    the substantially increasing product liability insurance costs and unwarranted
    exposure to product liability litigation.” Id. The statute of repose “provide[s] some
    certainty in the area of manufacturers’ and sellers’ long term exposure for
    product-related claims.” Morse v. City of Toppenish, 
    46 Wn. App. 60
    , 64-65, 
    729 P.2d 638
     (1986).
    17
    No. 83287-5-I/18
    To this end, the legislature integrated the statute of repose’s limitation on
    liability into WPLA such that it is fundamental to the existence of a claim. 11 A
    plaintiff who cannot satisfy the WPLA statute of repose does not have a WPLA
    claim. Due to its claim-defining nature, WPLA’s statute of repose is inextricably
    linked to the cause of action. Where, as here, the legislature made the statute of
    repose mandatory to the existence of a WPLA claim, we apply it as intended to
    limit liability by extinguishing certain claims. 12
    Considering the statute of repose as part and parcel of the liability claim
    for the purpose of a conflicts-of-law analysis is consistent with the approach in
    the more recent Restatement (Third) of Conflict of Laws § 6.12 (Am. L. Inst.,
    Tentative Draft No. 4, 2023). 13 The Restatement (Third) notes that “[t]he
    Restatement Second indicated that the[] connecting factors [set out in § 145]
    would have more weight for some issues than others, providing that they were ‘to
    be evaluated according to their relative importance with respect to the particular
    11 As our Supreme Court has recently stated,
    [W]e reaffirm the legislature’s broad authority to set time limits for commencing
    an action. We also recognize that when exercising this authority, the legislature
    must weigh competing interests and make difficult choices as a matter of policy.
    We do not seek to constrain the legislature’s authority or to second-guess its
    policy decisions.
    Bennett, 2 Wn.3d at 435.
    12 This same conclusion does not necessarily extend to all defenses. Our opinion here is
    limited to the issue before us, the statute of repose. We reject as overly broad Pharmacia’s
    statement that “choice of law principles do not authorize the selection of a state’s causes of action
    without the accompanying affirmative defenses.” Br. of Resp’ts at 35.
    13 Once a Tentative Draft is approved by the American Law Institute Council and
    membership, it “represents the most current statement of the Institute’s position on the subject
    and may be cited in opinions or briefs . . . until the official text is published.” RESTATEMENT (THIRD)
    OF CONFLICT OF LAWS at ix (AM. L. INST., Tentative Draft No. 4, 2023).
    18
    No. 83287-5-I/19
    issue.’ ” RESTATEMENT (THIRD) § 6.01 (quoting RESTATEMENT (SECOND)
    § 145(2)). 14
    The Restatement (Third) adds separate sections on particular torts and
    issues that require more specific rules “because they involve situations in which
    contacts other than those used in the general rules are relevant, or in which the
    contacts carry different weight.” Id., ch. 6, Topic 2, Introductory Note, at 37. Of
    particular relevance here, “for products liability claims, the place in which the
    product was delivered to the first end user is an additional relevant contact.” Id.
    Thus, the Restatement (Third) added a separate section on choice-of-law
    analysis for product liability claims. RESTATEMENT (THIRD) § 6.11. And a comment
    to § 6.11 entitled “Scope of Section limited to liability” it states, “A statute of
    repose relates to liability, and choice of law with respect to such statutes is
    performed under this Section [on products liability claims].” RESTATEMENT (THIRD)
    § 6.11 cmt. h. Also noteworthy is that neither the Restatement (Second) nor the
    14 Noting the “tentative and incomplete way” its predecessor “distinguished issues for
    which . . . [territorial versus personal] connecting factors are more important and suggested rules
    based on certain distributions of connecting factors,” Restatement (Third) described its aim as
    follows:
    This Restatement draws on the experience of courts applying the
    Restatement Second . . . to extend the insights of the Restatement Second and
    articulate them in the form of rules rather than an open-ended balancing of
    factors.
    This Restatement describes the distinction between the issues for which
    territorial connecting factors are more important and those for which personal
    connecting factors are more important as the distinction between issues relating
    to conduct and issues relating to persons. Most courts and scholars describe the
    distinction as between conduct-regulation issues and loss-allocation issues.
    RESTATEMENT (THIRD) § 6.01 cmt.
    19
    No. 83287-5-I/20
    Restatement (Third) identifies statutes of repose as issues requiring separate
    sections or additional analysis. 15
    For a WPLA claim to exist, the plaintiff must satisfy the statute’s integrated
    statute of repose. The trial court’s conclusion that Missouri law applied to this
    issue was an error of law.
    B.      Constitutionality of WPLA Statute of Repose
    Having determined under choice-of-law principles that WPLA’s statute of
    repose applies, we next consider whether the statute of repose is constitutional
    under the Washington Constitution’s privileges and immunities clause, article I,
    section 12. 16 We hold that the WPLA statute of repose is supported by
    reasonable grounds and survives constitutional scrutiny under article I, section
    12.
    Under the privileges and immunities clause, “No law shall be passed
    granting to any citizen, class of citizens, or corporation other than municipal,
    privileges or immunities which upon the same terms shall not equally belong to
    all citizens, or corporations.” WASH. CONST. art. I, § 12. We apply a two-part test
    to assess constitutional claims under article I, section 12. Schroeder v. Weighall,
    15 By contrast, as previously noted, the Restatement (Third) has added sections on
    distinct issues relating to tort claims, including product liability and punitive damages, and the
    Restatement (Second) likewise includes other subsections addressing particular torts, §§ 146-
    155, and particular issues in tort, §§ 156-174.
    16 After oral argument in this case, the Washington Supreme Court held the state’s
    medical malpractice statute of repose was unconstitutional under the privileges and immunities
    clause. Bennett, 2 Wn.3d at 435. Erickson submitted a statement of additional authority calling
    attention to Bennett and its possible impact on this case. Generally, an appellate court will decide
    a case only on the basis of the issues set forth by the parties’ briefing. RAP 12.1(a). Given the
    centrality of the statute of repose to the disposition of this case, we requested supplemental
    briefing and now consider the issue pursuant to RAP 12.1(b).
    20
    No. 83287-5-I/21
    
    179 Wn.2d 566
    , 572-73, 
    316 P.3d 482
     (2014). First, we determine whether the
    law grants a “privilege” or “immunity” that implicates a fundamental right of state
    citizenship. Id. at 573. If the answer is “yes,” then we consider whether there is a
    “reasonable ground” for granting the privilege or immunity. Id.
    Washington courts have long recognized “privileges and immunities” to
    include the right to pursue common law causes of action in court. Id.; see also
    Bennett, 2 Wn.3d at 444. Where a cause of action derives from the common law,
    the ability to pursue it is a privilege of state citizenship. Schroeder, 179 Wn.2d at
    573. “A law limiting the pursuit of common law claims against certain defendants
    therefore grants those defendants an article I, section 12 ‘immunity.’ ” Id. That
    only certain defendants are entitled to such immunity, as occurs with a statute of
    repose, triggers the reasonable grounds analysis. Bennett, 2 Wn.3d at 446.
    Here, while the product liability claims are defined by statute, they are
    nonetheless “rooted in the common law tradition,” as prior to WPLA’s enactment,
    what are now WPLA claims were common law causes of action. As Erickson
    notes, RCW 7.72.010(4) defines “[p]roduct liability claims” to include “any claim
    or action previously based on” common law causes of action. Pharmacia
    concedes that WPLA’s statute of repose triggers the reasonable ground test for
    the same reason as in Bennett.
    As WPLA’s statute of repose confers an article I, section 12 immunity
    consisting of a fundamental right of state citizenship, next, we consider whether
    there is a “reasonable ground” for granting the privilege or immunity. See
    21
    No. 83287-5-I/22
    Schroeder, 179 Wn.2d at 573. The reasonable ground test “requires careful
    consideration of the legislative purposes underlying the challenged statute.”
    Bennett, 2 Wn.3d at 447. The court must “scrutinize the legislative distinction to
    determine whether it in fact serves the legislature’s stated goal.” Schroeder, 179
    Wn.2d at 574. “[T]here must be a nexus between the legislature’s stated purpose
    and the challenged statute, which cannot rest solely on hypothesized facts.”
    Bennett, 2 Wn.3d at 449.
    In Bennett, the court explained the legislature’s three stated reasons for
    the medical malpractice statute of repose—reduction of medical malpractice
    insurance, protection from stale claims, and balancing the interests of injured
    plaintiffs and the health care industry—and considered whether they were
    “reasonable grounds” for granting the immunity. 2 Wn.3d at 448-49. On the issue
    of malpractice insurance, the legislature asserted that “ ‘to the extent that the
    eight-year statute of repose has an effect on medical malpractice insurance, that
    effect will tend to reduce rather than increase the cost of malpractice insurance.’ ”
    Id. at 449 (quoting LAWS OF 2006, ch. 8, § 301). According to the court, this
    statement of purpose was insufficient because “the legislature did not assert that
    the statute of repose would, in fact, decrease the cost of medical malpractice,”
    and the court cannot hypothesize that it would have this effect. Id. (emphasis
    added).
    With regard to the medical malpractice statute, the legislature also
    reasoned the statute of repose “ ‘will provide protection against claims, however
    22
    No. 83287-5-I/23
    few, that are stale, based on untrustworthy evidence, or that place undue
    burdens on defendants.’ ” Id. at 449-50 (quoting LAWS OF 2006, ch. 8, § 301). The
    court noted that “the eight-year statute of repose in RCW 4.16.350(3) is not
    addressed to stale claims generally, in light of the explicit exemptions and tolling
    provisions” in the statute and concluded that “the statute of repose does not in
    fact serve the legislature’s stated rationale of preventing stale claims generally.”
    Bennett, 2 Wn.3d at 450.
    Finally, the legislature asserted, “ ‘an eight-year statute of repose is a
    reasonable time period in light of the need to balance the interests of injured
    plaintiffs and the health care industry.’ ” Id. at 451 (quoting LAWS OF 2006, ch. 8,
    § 301). The court disagreed that such a legislative compromise, by itself, was
    sufficient to provide reasonable grounds for the statute of repose. Id. at 451-52.
    The Bennett court’s analysis recognized “the legislature’s authority to set time
    limits for commencement of civil actions” and the “competing policy interests at
    stake,” but concluded that the medical malpractice statute of repose was
    unconstitutional because none of the legislature’s stated reasons satisfied the
    reasonable ground test. Id. at 452.
    To apply the reasonable ground test in this case, we look to the legislative
    purposes underlying the WPLA statute of repose. WPLA’s preamble identifies the
    legislature’s intent to protect product sellers and manufacturers from the high
    cost of product liability insurance:
    Of particular concern is the area of tort law known as product
    liability law. Sharply rising premiums for product liability insurance
    23
    No. 83287-5-I/24
    have increased the cost of consumer and industrial goods. These
    increases in premiums have resulted in disincentives to industrial
    innovation and the development of new products. High product
    liability premiums may encourage product sellers and
    manufacturers to go without liability insurance or pass the high cost
    of insurance on to the consuming public in general.
    It is the intent of the legislature to treat the consuming public, the
    product seller, the product manufacturer, and the product liability
    insurer in a balanced fashion in order to deal with these problems.
    It is the intent of the legislature that the right of the consumer to
    recover for injuries sustained as a result of an unsafe product not
    be unduly impaired. It is further the intent of the legislature that
    retail businesses located primarily in the state of Washington be
    protected from the substantially increasing product liability
    insurance costs and unwarranted exposure to product liability
    litigation.
    LAWS OF 1981, ch. 27, § 1.
    Erickson argues that, like in Bennett, the legislature failed to include any
    finding that the statute of repose would achieve the goal of reducing premiums.
    However, the WPLA statute of repose was crafted after the Washington State
    Senate Select Committee on Tort and Product Liability Reform (Select
    Committee) conducted hearings and research to identify the key issues and
    solutions for the needs of the various actors involved in product liability cases.
    The Select Committee began its work by gathering information to
    determine the parameters of the product liability insurance issue. S. SELECT
    COMM. ON TORT & PRODUCT LIABILITY REFORM, FINAL REPORT at 4 (Wash. Jan.
    1981) [https://perma.cc/R3XP-CRC8]. Statistics from the Insurance Services
    Office, an independent insurance industry statistical and rate making
    organization, showed that over 97 percent of product-related incidents occurred
    24
    No. 83287-5-I/25
    within six years of purchase and 83.5 percent of bodily injuries occurred within 10
    years of manufacturing. Id. at 19. For the Select Committee, this relatively short
    time frame “raise[d] questions regarding the need and effectiveness of a statute
    of repose.” Id. at 9.
    Erickson points to this finding to suggest a parallel between the statute of
    repose at issue in Bennett and WPLA’s statute of repose, noting that unlike the
    legislature’s general conclusion that the medical malpractice statute of repose
    would “tend” to reduce premiums, the Select Committee made no finding that the
    WPLA statute of repose will achieve a reduction in insurance costs. But Erickson
    ignores the substance of the Select Committee’s findings as to the impact on
    rates due to insurance industry concerns. The findings illustrate a specific link,
    whereas in the medical malpractice statute, a similar link between the
    legislature’s stated concern about insurance costs and the statute of repose is
    absent, as discussed in Bennett. Regarding product liability, the Select
    Committee found that “there did not appear to be a severe problem regarding the
    availability of product liability insurance in Washington.” Id. at 14. “Rather, the
    problem was one of affordability.” Id. According to the report, “[o]f greatest
    concern to product insurers is the length of time a product seller is subject to
    liability for harm resulting from a product defect, and they contend that the
    potential ‘long tail’ of exposure is the primary factor influencing rate-setting.” Id. at
    19.
    25
    No. 83287-5-I/26
    While acknowledging the data, the Select Committee noted “an insurer’s
    perception of potential claims, whether substantiated or not, very likely is
    reflected in rates.” Id. Litigation expenses “add appreciably to the cost of product
    liability insurance,” with defense costs amounting to 35 percent of bodily injury
    payments and 48 percent of property damage payments. Id. at 9. Moreover,
    product manufacturers “have often expressed concern over the possibility of the
    imposition of liability based upon an injury caused by an old product” due to both
    the difficulty in defending such claims and the high cost of insurance for claims-
    made policies. Id. at 42. “[T]his open-ended situation also affects insurance rates
    since most product liability insurance is written on a claims-made basis which
    means that the liability insurer at the time a claim is made is liable regardless of
    the date of manufacture of the product. Id. Product liability premiums, therefore,
    must take into account the possibility of claims on products manufactured many
    years ago.” Id. The Select Committee’s findings indicate that, whether founded or
    not, insurers had serious concerns about open-ended liability that leads to high
    insurance rates.
    To bring a measure of certainty to the length of exposure but “reflecting its
    concern in preserving those claims based upon product use which is reasonable
    in light of its unique characteristics,” the Select Committee recommended
    following the UPLA concept of the “useful safe life,” rather than “a complete bar
    after an arbitrary time period.” Id. at 19. The Select Committee acknowledged
    that establishing an absolute cutoff date would provide certainty for
    26
    No. 83287-5-I/27
    manufacturers, but “the length of the statute of repose may not bear any relation
    to the useful life of the product.” Id. at 42. The Select Committee’s solution was to
    tie the statute of repose to the useful safe life of the product. The statute of
    repose includes a presumption of 12 years as the useful safe life, which the
    claimant can overcome with a preponderance of the evidence. In order to
    “mitigate the harshness” the 12-year presumption imposes on claimants, the
    Select Committee recommended the “preponderance of the evidence” standard,
    rather than the higher “clear and convincing” standard in UPLA. Id. at 19-20. “The
    use of a rebuttable twelve-year presumption of usefulness should create a
    degree of certainty in the law without depriving the claimant of the ability to
    demonstrate that, in fact, the product was still in a useful condition at the time of
    the injury.” Id. at 43.
    Erickson contends that the legislature’s desire to balance the interests of
    different stakeholders was not considered an adequate justification for
    immunizing certain defendants in Bennett and should not provide reasonable
    grounds for the WPLA statute of repose. However, unlike the general statements
    of legislative compromise discussed in Bennett, the Select Committee provided
    specific explanations for the ways in which its research drove the
    recommendations that balance the competing interests. The comprehensive
    research and reporting by the Select Committee established that product liability
    insurers’ concern about “long-tail” claims—whether substantiated or not—
    contributed to insurance costs. The insurers wanted a degree of certainty for the
    27
    No. 83287-5-I/28
    length of their liability, and the Select Committee recommended a statute of
    repose to provide a measure of this desired certainty. The Select Committee also
    acknowledged the impact of the statute of repose on claimants and made the
    presumptive timeline long enough for most claims and with a lower
    “preponderance” evidentiary standard to rebut the presumption.
    Finally, Erickson argues that WPLA’s statute of repose contains
    exceptions that, as in Bennett, run counter to the purpose of the statute of
    repose. The exceptions in RCW 7.72.060 allow for claims beyond a product’s
    useful safe life if (1) the seller warranted a longer period of time, (2) the seller
    intentionally misrepresents facts or conceals information about the product and
    the conduct was the proximate cause of claimant’s harm, and (3) the harm was
    caused by exposure to a defective product occurred within the safe useful life of
    the product even though the harm did not manifest until later. RCW
    7.72.060(1)(b). Therefore, these “exceptions” preserve a claimant’s ability to
    pursue a product liability claim in certain circumstances, expanding WPLA’s
    coverage beyond a products useful safe life, generally where the defendant’s
    behavior has warranted such liability.
    Moreover, for the purposes of the privileges and immunities clause, “the
    question is not whether the statute’s exceptions are reasonable. Instead, the
    question is whether barring those remaining claims, which are subject to the
    statute of repose, in fact serves the legislature’s stated rationale.” Bennett, 2
    Wn.3d at 451. As the Select Committee’s research and recommendations show,
    28
    No. 83287-5-I/29
    precluding lawsuits relating to products that are beyond their safe useful life
    furthers the research-supported goal of reducing the insurance industry’s
    concerns about “long-tail” claims. By contrast, in Bennett, due to the exceptions,
    the medical malpractice statute of repose “does not in fact serve the legislature’s
    stated rationale of preventing stale claims generally.” 2 Wn.3d at 451.
    Unlike the cursory legislative reasoning for the medical malpractice statute
    of repose described in Bennett, the Select Committee supplied data and analysis
    that establish reasonable grounds for the WPLA statute of repose. Therefore, the
    WPLA statute of repose does not violate the privileges and immunities clause of
    the Washington Constitution.
    We hold that WPLA’s statute of repose, not Missouri law, applies to
    Erickson’s claims in this case. Because there may be factual issues relating to
    the application of WPLA’s statute of repose that have not been fully litigated, we
    reverse and remand to the trial court for proceedings.
    C.     Choice of Law as to Punitive Damages
    Pharmacia also contends the trial court erred by allowing the jury to
    consider and award punitive damages under Missouri law. Before we engage in
    a choice-of-law analysis, we first consider whether an actual conflict exists. Rice,
    
    124 Wn.2d at 210
    . Here, a conflict does exist, as Washington law does not
    provide for punitive damages, but Missouri law does. See Barr v. Interbay
    Citizens Bank of Tampa, Fla., 
    96 Wn.2d 692
    , 699, 
    635 P.2d 441
     (1981) (“punitive
    damages are not allowed unless expressly authorized by the legislature”); Poage
    29
    No. 83287-5-I/30
    v. Crane Co., 
    523 S.W.3d 496
    , 515 (Mo. App. 2017) (discussing requirements for
    allowing punitive damages).
    Next, as with the statute of repose, we must determine whether to conduct
    a separate choice-of-law analysis on this issue, distinct from the choice of
    substantive law on product liability. Restatement (Second) § 171 addresses the
    separate issue of damages in tort actions, stating, “The law selected by
    application of the rule of § 145 determines the measure of damages.” Then,
    § 171, comment b elaborates:
    The determination of the state of the applicable law should be
    made in the light of the choice-of-law principles stated in § 6. In
    general, this should be the state which has the dominant interest in
    the determination of the particular issue. The state of conduct and
    injury will not, by reason of these contacts alone, be the state that is
    primarily concerned with the measure of damages in a tort action.
    The local law of this state will, however, be applied unless some
    other state has a greater interest in this issue.
    Comment d specifies that “[t]he law selected by application of the rule of § 145
    determines the right to exemplary damages.” 17 Therefore, the issue of punitive
    17 The more recent Restatement (Third) contains a separate section on punitive
    damages, which is consistent with Restatement (Second) § 171 that the law governing the
    availability of punitive damages for torts is the law selected under the other choice-of-law rules for
    the tort. RESTATEMENT (THIRD) OF CONFLICT OF LAWS, § 6.12 (AM. L. INST., Tentative Draft No. 4,
    2023). The comment to this new § 6.12 explains why this separate section was added:
    A Section specific to punitive damages is necessary . . . because the significance
    of certain connecting factors differs from the ordinary tort case. Because punitive
    damages are designed to punish the tortfeasor rather than to compensate the
    victim, the tortfeasor’s domicile is a more significant contact, and the victim’s
    domicile less significant.
    Id. § 6.12 cmt. By contrast, the Restatement (Second) identifies exemplary damages
    simply as a subset of other damages and states, “The law selected by application of the
    rule of § 145 determines the measure of damages.” RESTATEMENT (SECOND) § 171
    (emphasis added). Using the terminology from the Restatement (Third), “measure of
    damages” is a “loss-allocation” issue—i.e., an “issue relating to persons,” Restatement
    30
    No. 83287-5-I/31
    damages would generally follow the choice-of-law analysis for the substantive
    law on liability—in this case, WPLA. 18 WPLA has no explicit provision on punitive
    damages. 19 Because WPLA has not expressly authorized punitive damages,
    they are not authorized for WPLA claims under Washington law. See Barr, 96
    Wn.2d at 697 (express legislative authorization required for punitive damages).
    However, notwithstanding these statements in Restatement (Second)
    § 171, Washington courts have applied a separate choice-of-law analysis to the
    issue of punitive damages for tort claims, analyzing the factors identified in
    Restatement (Second) § 145 to determine which state has the most significant
    interest. Singh v. Edwards Lifesciences Corp., 
    151 Wn. App. 137
    , 144, 
    210 P.3d 337
     (2009); Barr v. Interbay Citizens Bank of Tampa, Fla., 
    96 Wn.2d 692
    , 697-
    98, 
    635 P.2d 441
     (1981); Kammerer v. W. Gear Corp., 
    96 Wn.2d 416
    , 421-22,
    
    635 P.2d 708
     (1981). In doing so, courts have sometimes engaged in dépeçage
    and, after analyzing the issue of punitive damages, have determined that a
    different state’s law applies to punitive damages than to the underlying tort
    liability claim. In other words, depending on the issue, some contacts may be
    more important in determining which state’s interest is greater.
    (Third) § 6.05, whereas punitive damages relate to “conduct-regulation.” Id. § 6.04. Thus,
    under the Restatement (Third) approach, different factors may be more or less important
    depending on the specific issue, even if related to the same tort claim.
    18 As discussed above, here, there was no need for a choice-of-law analysis analyzing
    the Restatement (Second) § 145 contacts as to the substantive law regarding liability because
    Erickson opted to file a WPLA claim, and Pharmacia did not challenge this choice as to liability.
    19 WPLA evinces an intent not to disturb previous existing law, but this statement is
    limited to the law on product liability, not other aspects of existing law such as punitive damages.
    RCW 7.72.020(1) (“The previous existing applicable law of this state on product liability is
    modified only to the extent set forth in this chapter.”).
    31
    No. 83287-5-I/32
    In Barr, a lender in Florida hired agents in Nevada to travel to Washington
    to repossess the plaintiff’s automobile. 96 Wn.2d at 694-95. The plaintiff filed suit
    for conversion, intentional infliction of emotional distress, and outrageous
    conduct by the defendant bank. Id. at 695. Washington was the state of the injury
    and plaintiff’s domicile. Florida was the bank’s domicile and where the conduct at
    issue occurred. Id. at 698. While Florida had an interest in deterring the conduct,
    its interest would not be furthered when the immediate conduct causing the injury
    was in Nevada and Washington; thus, the court found Florida’s interest
    subordinate to Washington’s. Id. at 699.
    Kammerer, decided the same day as Barr, involved a lawsuit for breach of
    a patent licensing agreement and fraud after a Washington corporation failed to
    pay royalties for manufacturing products based on patents owned by California
    residents. 96 Wn.2d at 418. The defendant had gone to California to negotiate
    the agreement, and the parties had agreed California law would apply. Id. at 423.
    In determining whether to apply California law allowing punitive damages, the
    court concluded, “[w]here the most significant relationships were in California and
    where the conduct and acts as to the fraud and misrepresentation were
    accomplished in California that state has a specific interest to be furthered.” Id.
    California had an obvious interest in protecting its residents from fraud, whereas
    Washington had no interest in protecting a company that committed fraud. Id. at
    422. Washington’s interests were not superior to or inconsistent with California’s
    32
    No. 83287-5-I/33
    interests. Therefore, the application of California law to allow punitive damages
    would further California’s specific interest. Id. at 422-23.
    The court in Singh likewise held that California law on punitive damages
    could apply. 
    151 Wn. App. at 148
    . There, a California company manufactured a
    medical device that malfunctioned during surgery, resulting in irreparable
    damage to plaintiff’s heart and leading to a heart transplant and subsequent
    cancer diagnosis from the anti-rejection medication. 
    Id. at 140-41
    . As a
    Washington resident, plaintiff filed a product liability claim under WPLA, and the
    Washington hospital filed cross-claims against the manufacturer for fraud,
    violation of the Consumer Protection Act, and breach of contract. 
    Id. at 141
    . Both
    the patient and the hospital sought punitive damages under California law. 
    Id.
    Evidence presented at trial demonstrated that the manufacturers knew of the flaw
    and the resulting malfunction but decided against recalling the product or warning
    users. 
    Id.
     The jury found defendant’s conduct to be malicious and awarded
    punitive damages under California law to both the patient and the hospital. 
    Id. at 142
    . This court affirmed the award of punitive damages under California law.
    “Even though Washington has a strong policy against punitive damages, it has
    no interest in protecting companies that commit fraud. Where, as here, an entity
    headquartered in California, committed the conduct in California that resulted in
    the plaintiff’s damages, California had the greater interest in deterring such
    fraudulent activities.” 
    Id. at 140
    .
    33
    No. 83287-5-I/34
    As Singh and Kammerer demonstrate, Washington courts will allow
    punitive damages under the law of another state when that state has an interest
    in punishing or deterring egregious conduct that is greater than any interest
    Washington has in not allowing punitive damages. A key consideration in these
    cases appears to be the location of the conduct that caused the injury and that
    state’s interest in deterring the conduct at issue by the tortfeasor. 20
    Using the “most significant relationship” factors in Restatement (Second)
    § 145, we consider the place of injury, the conduct that caused the injury, the
    parties’ domiciles, and the place where the parties’ relationship, if any, is
    centered. 21 Here, the contacts are evenly split: the place of injury is Washington;
    the conduct that caused the injury is in Missouri; and the parties are domiciled in
    Washington and Missouri, respectively. As in Kammerer and Singh, the state
    where the conduct at issue occurred—here, Missouri—has an interest in
    deterring the unlawful conduct that caused the injury, particularly when the
    20 This weighting of factors accords with Restatement (Third) § 6.12 on punitive
    damages. Like Restatement (Second) § 171 on damages generally, Restatement (Third)
    suggests that the same state’s law selected under the choice-of-law rules applicable to the tort
    should govern the availability of punitive damages. It adds that punitive damages “may not be
    awarded if they are disallowed under the law of two of the following three states: (1) the
    defendant’s domicile; (2) the place of the conduct; (3) the place of injury.” RESTATEMENT (THIRD)
    § 6.12. This guidance is consistent with the results in Singh (defendant in California, conduct in
    California, injury in Washington, California law on punitive damages applied), Kammerer
    (defendant in California, conduct in California, injury in Washington, California law on punitive
    damages applied), and Barr (defendant in Florida, conduct in Washington and Nevada, injury in
    Washington, Washington law on punitive damages applied).
    21 Restatement (Third) § 6.11 adds the place in which the product was delivered to the
    first end user as an additional relevant contact for product liability claims. Here, that state is
    Washington.
    34
    No. 83287-5-I/35
    alleged tortfeasor has its place of business there. Washington does not have an
    interest in protecting Missouri companies that engage in injury-causing conduct.
    In this case, Missouri has the greater interest in deterring or punishing any
    egregious conduct found in this case. Missouri product liability law permits
    punitive damages where clear and convincing evidence proves “ ‘that the
    defendant showed a complete indifference or conscious disregard for the safety
    of others.’ ” Peters v. Gen. Motors Corp., 
    200 S.W.3d 1
    , 24 (Mo. Ct. App. 2006)
    (quoting Letz v. Turbomeca Engine Corp., 
    975 S.W.2d 155
    , 165 (Mo. Ct. App.
    1997)). In Missouri, punitive damages may be submitted to the jury if some
    element of outrageous conduct is demonstrated that shows the defendant acted
    with a “ ‘willful, wanton or malicious culpable state.’ ” Ingham v. Johnson &
    Johnson, 
    608 S.W.3d 663
    , 714 (Mo. Ct. App. 2020) (quoting Poage, 
    523 S.W.3d at 515
    ).
    Thus, to the extent Missouri’s product liability law prohibits the same
    conduct as prohibited by WPLA, plaintiffs may seek punitive damages because
    Missouri has a more significant interest pursuant to an analysis of relevant
    contacts. However, Missouri lacks a cause of action for post-sale failure to warn.
    Therefore, Missouri cannot be said to have an interest in deterring conduct that it
    does not deem unlawful, and punitive damages would not be available for
    Erickson’s post-sale failure to warn claim.
    The jury awarded substantial punitive damages, but due to the general
    nature of the verdict form, we cannot determine which cause of action resulted in
    35
    No. 83287-5-I/36
    the award of punitive damages. 22 Compare Blanks v. Fluor Corp., 
    450 S.W.3d 308
    , 364 (Mo. Ct. App. 2014) (reversing punitive damage award against one
    defendant as jury instructions required jury to consider undifferentiated conduct).
    In future proceedings, special interrogatories are required to establish the
    particular theory of liability supporting punitive damages, limited to claims that
    exist under Missouri law.
    II.    Proposed Jury Instructions
    Although we reverse and remand to the trial court based on our
    conclusion that Washington law, including WPLA’s statute of repose, applies to
    determine liability for plaintiffs’ product liability claims, because additional issues
    raised by Pharmacia are likely to recur in future proceedings in this case, we
    address them in this opinion. Thus, we turn next to Pharmacia’s claims that the
    trial court erred by refusing to give jury instructions related to two defenses.
    “Where substantial evidence supports a party’s theory of the case, trial
    courts are required to instruct the jury on the theory.” Taylor v. Intuitive Surgical,
    Inc., 
    187 Wn.2d 743
    , 767, 
    389 P.3d 517
     (2017). Jury instructions are sufficient if
    they are supported by the evidence, allow each party to argue their theory of the
    case, and, when taken as a whole, properly inform the trier of fact of the
    applicable law. Fergen v. Sestero, 
    182 Wn.2d 794
    , 803, 
    346 P.3d 708
     (2015).
    Whether to give a jury instruction is within the court’s discretion and, therefore,
    22 Given this ruling on punitive damages, we need not address Pharmacia’s arguments
    challenging the amount of punitive damages awarded.
    36
    No. 83287-5-I/37
    reviewed for abuse of that discretion. Id. at 802. “A trial court abuses its
    discretion if its decision is manifestly unreasonable or based on untenable
    grounds or untenable reasons.” In re Marriage of Littlefield, 
    133 Wn.2d 39
    , 46-47,
    
    940 P.2d 1362
     (1997).
    A.     Relevant Product
    Pharmacia sought an instruction that would have allowed the jury to find
    that the PCB-containing capacitor or the FLB that incorporated the capacitor,
    rather than PCBs, was the “relevant product” that caused harm by failing to
    contain the PCBs. The trial court denied the requested instruction, ruling, “There
    is no question that PCBs are the relevant product as a matter of law and there is
    no factual issue to submit to the jury.” Erickson argues the trial court correctly
    held, as a matter of law, that PCBs are the relevant product. We agree with
    Erickson.
    To determine what is the relevant product, we look to how WPLA defines
    and uses the term. First, WPLA defines when a manufacturer is liable:
    [a] product manufacturer is subject to liability to a claimant if the
    claimant’s harm was proximately caused by the negligence of the
    manufacturer in that the product was not reasonably safe as
    designed or not reasonably safe because adequate warnings or
    instructions were not provided.
    RCW 7.72.030(1).
    The term “relevant product” also appears in several other definitions that
    limit who may be held liable under WPLA. “ ‘Manufacturer’ includes a product
    seller who designs, produces, makes, fabricates, constructs, or remanufactures
    37
    No. 83287-5-I/38
    the relevant product or component part of a product before its sale to a user or
    consumer.” RCW 7.72.010(2) (emphasis added). “ ‘Product seller’ means any
    person or entity that is engaged in the business of selling products, whether the
    sale is for resale, or for use or consumption. The term includes a manufacturer,
    wholesaler, distributor, or retailer of the relevant product.” RCW 7.72.010(1)
    (emphasis added).
    The statute then defines “relevant product” as “that product or its
    component part or parts, which gave rise to the product liability claim.” RCW
    7.72.010(3). Further, WPLA specifies the kinds of actions that can cause harm
    for which a claim can be brought, i.e., the actions that trigger liability: A “ ‘product
    liability claim’ includes any claim or action brought for harm caused by the
    manufacture, production, making, construction, fabrication, design, formula,
    preparation, assembly, installation, testing, warnings, instructions, marketing,
    packaging, storage or labeling of the relevant product.” RCW 7.72.010(4)
    (emphasis added).
    Thus, the “relevant product” is a limiting factor as to whether a particular
    defendant may be held liable for the alleged harm as the manufacturer or product
    seller. Here, plaintiffs’ product liability claims are for harms suffered as a result of
    the toxic nature of PCBs, which they claim were “not reasonably safe” in
    construction or design and did not have proper warnings.
    Pharmacia mischaracterizes the cases discussing “relevant product.” First,
    it suggests that a product must fail in order to be a relevant product, citing
    38
    No. 83287-5-I/39
    Sepulveda-Esquivel v. Cent. Machine Works, Inc., 
    120 Wn. App. 12
    , 
    84 P.3d 895
    (2004). There, the plaintiff suffered injuries when a large piece of equipment fell
    on him. Id. at 15. The crane that moved the equipment had a hook, and the
    equipment was secured in the hook by a device called a “mouse.” Id. at 16-17.
    One of the defendants had forged the hook and the other had supplied it, but
    neither had a connection to the mouse. Id. Pharmacia points to the court’s
    statement that “since the hook did not fail, the hook is not a ‘relevant product’
    which would give rise to the product liability claim.” Id. at 13. But the key holding
    was not that a product must fail to be a “relevant product”; rather, the case turned
    on whether the defendants had a qualifying relationship to the product that is
    alleged to have caused the harm. As the court explained,
    [N]either Central nor Ulven made, supplied, or sold the finished,
    completed hook assembly with the mouse. Neither Central nor
    Ulven was asked to design, forge, make, or sell an interior, locking
    device on its hook. Because there was no defect in the hook itself,
    Ulven is without fault; Central is without fault because it did not
    design the hook and merely provided the hook according to the
    purchaser’s specifications, including the lack of a closing device.
    Id. at 19. The company that manufactured the mouse was not named as a party
    to the lawsuit. Id. at 15. The hook was not the relevant product, so the
    defendants were not manufacturers or sellers of the relevant product. Therefore,
    the defendants were not liable for the injuries. Id. at 19. Sepulveda-Esquivel does
    not alter the statutory language to require that a “relevant product” must “fail.”
    Pharmacia also cites O’Connell v. MacNeil Wash Sys. Ltd., 2 Wn. App. 2d
    238, 246, 
    409 P.3d 1107
     (2017), as support for the proposition that to prevail, a
    39
    No. 83287-5-I/40
    WPLA plaintiff must show that the “relevant product” failed. But neither O’Connell
    nor WPLA supports this proposition. Rather, in accord with RCW 7.72.030(1),
    O’Connell states, “a plaintiff must show that (1) a manufacturer’s product (2) was
    not reasonably safe because of its design or because of lack of adequate
    warnings or instructions, which (3) caused harm to the plaintiff.” O’Connell, 2 Wn.
    App. 2d at 246 (citing Ayers v. Johnson & Johnson Baby Prods. Co. , 
    117 Wn.2d 747
    , 752, 
    818 P.2d 1337
     (1991); Thongchoom v. Graco Children’s Prods., Inc.,
    
    117 Wn. App. 299
    , 304, 
    71 P.3d 214
     (2003)). “Failure” of a product is not an
    element of a WPLA claim.
    Rather, as Erickson notes, a defendant may be liable even when “the
    relevant product [is] functioning as intended, yet the overall design of said
    product can be defective due to its lack of incorporated safety components or
    warnings about dangers related to its design.” O’Connell, 2 Wn. App. 2d at 248
    n.3. But this does not change the elements of a WPLA claim to require “failure.”
    Pharmacia additionally argues that a plaintiff’s complaint “does not dictate
    the ‘relevant product,’ ” and “the plaintiff must actually prove what the ‘relevant
    product’ is.” It is certainly a correct statement that the plaintiff must prove the
    elements of its WPLA claim, and that a product manufacturer cannot be liable
    unless the plaintiff establishes “the claimant’s harm was proximately caused by
    the negligence of the manufacturer in that the product was not reasonably safe
    as designed or not reasonably safe because adequate warnings or instructions
    were not provided.” RCW 7.72.030(1). But contrary to Pharmacia s assertion, it is
    40
    No. 83287-5-I/41
    the plaintiff who determines what is alleged to be the relevant product that
    caused harm, as well as whom it seeks to hold liable under WPLA. Indeed, as
    Sepulveda-Esquivel demonstrates, this control over the claim may be at the
    plaintiff’s peril, if it has not sued the correct party. Thus, a defendant in a WPLA
    case may defend by establishing that the product it manufactured was not the
    cause of the harm. Whether other additional entities not named by the plaintiffs
    as defendants in their complaint could also be liable for the products they
    manufactured is not at issue here, similar to Sepulveda-Esquivel.
    WPLA does not require that the plaintiff prove that the “relevant product”
    failed. Even without the requested “relevant product” instruction, Pharmacia was
    able to argue this theory of the case. See Fergen, 
    182 Wn.2d at 803
     (jury
    instructions must allow each party to argue their theory of the case). Pharmacia
    could, and did, argue that blame was attributable to the FLBs and capacitors
    rather than PCBs themselves. Because the trial court’s refusal was not made on
    untenable grounds, we conclude that it did not abuse its discretion in declining to
    issue Pharmacia’s requested “relevant product” instruction.
    B.     Raw Materials Supplier for Sophisticated Purchasers
    Pharmacia contends the trial court erroneously refused to instruct the jury
    on its theory that it was a “supplier of raw materials to a sophisticated purchaser.”
    41
    No. 83287-5-I/42
    Erickson argues that Washington has declined to adopt this rule outside of the
    pharmaceutical context. Erickson is correct.
    Pharmacia argues that a supplier has no duty to warn an ultimate user
    where the supplier insures the buyer was knowledgeable about the dangers and
    could warn the ultimate users, or the buyer was already aware of the dangers. In
    support, Pharmacia cites Zamora v. Mobil Corp., 
    104 Wn.2d 199
    , 204-05, 
    704 P.2d 584
     (1985), in which a house fire killed several children due to a propane
    leak that went undetected because the gas was not properly odorized. 
    Id. at 201
    .
    The appellants brought a wrongful death suit against several defendants,
    including Mobil, who marketed the propane but did not physically handle, modify,
    alter, transport, or refine it. 
    Id. at 202
    . Relying on a Kansas propane case,
    Zamora stated that a propane seller can delegate to the retailer its duty to warn
    the ultimate consumer. 
    Id. at 205
    . “[T]he distributor had only a duty to insure that
    the retailer was knowledgeable regarding the dangers and was able to warn the
    ultimate buyers.” 
    Id.
     Additionally, “the bulk seller has no duty to warn the retailer
    of the dangers of a product if that retailer is already aware ‘through common
    knowledge or learning’ of a specific hazard.” 
    Id. at 205
     (quoting Lancaster Silo &
    Block Co. v. N. Propane Gas Co., 
    75 A.D.2d 55
    , 
    427 N.Y.S.2d 1009
     (1980)). As
    a result, Mobil had no duty to warn the propane retailer of the obvious danger
    posed by the gas, and could rely on the retailer to warn the ultimate customers.
    Zamora, 
    104 Wn.2d at 205
    .
    42
    No. 83287-5-I/43
    At first blush, the facts of Zamora appear similar to this case, as Monsanto
    did not sell PCBs directly to Erickson, but instead was a manufacturer and bulk
    seller of PCBs to large sophisticated industrial entities like GE, who then used
    the PCBs to manufacture FLBs. These electrical equipment manufacturers knew
    the hazards of PCBs.
    However, in a case subsequent to Zamora, the Washington Supreme
    Court explicitly “reject[ed] this ‘sophisticated purchaser’ approach to apparent
    manufacturer liability as inconsistent with Washington law.” Rublee v. Carrier
    Corp., 
    192 Wn.2d 190
    , 207-08, 
    428 P.3d 1207
     (2018). The court explained that
    Washington does not differentiate between types of users or consumers in
    product liability cases, using the sophisticated purchaser defense as an example
    of this uniform approach to product liability. Id. at 208. “Washington courts have
    uniformly rejected a sophisticated user defense, under which product distributors
    are not required to instruct or warn sophisticated users about certain risks
    because such users are presumably already aware of the risks due to their
    expertise or sophistication.” Id. The only exceptions arise in the context of
    consumers of pharmaceuticals or medical devices, where the “learned
    intermediary” doctrine applies, and manufacturers of medical products obtainable
    only through physicians fulfill their duty to warn by giving adequate warning to the
    physicians who prescribe the product. Id. at 208-09.
    When Pharmacia requested a sophisticated purchaser jury instruction, the
    trial court declined because, “there isn’t a sophisticated purchaser thing in
    43
    No. 83287-5-I/44
    Washington law other than the learned intermediary when we’re talking about
    pharmaceuticals.” The court’s denial of the requested instruction reflects the law
    as stated in Rublee. A sophisticated purchaser instruction would have
    misinformed the jury by providing the law on a theory that does not exist in
    Washington. The trial court did not abuse its discretion by denying the instruction
    because it correctly applied the law and, thus, was not on untenable grounds.
    III.   Expert Testimony
    Both Pharmacia and Erickson proffered testimony from multiple experts.
    Pharmacia challenges the admission of testimony from three experts under Frye
    and/or ER 702: Kevin Coghlan, Dr. Richard Perrillo, and Dr. James Dahlgren.
    To admit expert testimony under ER 702, the trial court must determine
    that the witness qualifies as an expert and the testimony will assist the trier of
    fact. Lakey v. Puget Sound Energy, Inc., 
    176 Wn.2d 909
    , 918, 
    296 P.3d 860
    (2013). “The courts serve the gatekeeping function of keeping out ‘unreliable,
    untested, or junk science.’ ” L.M. v. Hamilton, 
    193 Wn.2d 113
    , 127, 
    436 P.3d 803
    (2019) (quoting Anderson v. Akzo Nobel Coatings, Inc., 
    172 Wn.2d 593
    , 606, 
    260 P.3d 857
     (2011)).
    To that end, to be admissible, expert testimony based on novel scientific
    evidence must satisfy both the Frye test and ER 702. Lakey, 
    176 Wn.2d at 918
    .
    “Frye and ER 702 work together to regulate expert testimony: Frye excludes
    testimony based on novel scientific methodology until a scientific consensus
    44
    No. 83287-5-I/45
    decides the methodology is reliable; ER 702 excludes testimony where the
    expert fails to adhere to that reliable methodology.” 
    Id. at 918-19
    .
    “Only after novel scientific evidence is found admissible under Frye does
    the court turn to whether it is admissible under ER 702.” Anderson, 
    172 Wn.2d at 603
    . ER 702 allows expert testimony that is helpful to the jury. Unreliable
    testimony fails the helpfulness requirement of ER 702. Lakey, 
    176 Wn.2d at 920
    .
    We review de novo a trial court’s decision on whether to exclude evidence
    under Frye. Anderson, 
    172 Wn.2d at 600
    . We review a trial court’s decision on
    the admissibility of expert testimony under ER 702 for abuse of discretion. L.M.,
    
    193 Wn.2d at 134
    . A trial court abuses its discretion by “issuing manifestly
    unreasonable rulings or rulings based on untenable grounds.” Lakey, 
    176 Wn.2d at 919
    . In order to conclude a trial court abused its discretion, a reviewing court
    must be convinced that no reasonable person would take the view adopted by
    the trial court. L.M., 
    193 Wn.2d at 134
    . We will not disturb a trial court’s ruling if
    the basis for admission is fairly debatable. 
    Id.
     Moreover, a reviewing court will not
    hold a trial court abused its discretion simply because it would have decided the
    issue differently. 
    Id.
    A.      Kevin Coghlan
    Erickson offered testimony from industrial hygienist Kevin Coghlan, M.S.,
    C.I.H., to prove plaintiffs’ exposure to harmful levels of PCBs and furans during
    45
    No. 83287-5-I/46
    their tenure at SVEC. 23 Coghlan provided a conservative estimate of PCB air
    levels “several hundred to several thousand nanograms per cubic meter,”
    ranging from approximately 100 or 200 ng/m3 to over 2,500 ng/m3. While the
    range was broad, Coghlan opined that the plaintiffs had exposures that exceeded
    the NIOSH 24-recommended workplace exposure level of 1,000 ng/m3.
    Coghlan conceded that no testing for PCBs occurred during the time
    frame that the plaintiffs worked at SVEC. As a result, Coghlan estimated
    historical PCB levels to show that plaintiffs had been exposed while at SVEC. He
    testified, “without actual air data, we had to rely on some other techniques to
    reconstruct those exposures.” Coghlan used three approaches to arrive at an
    estimated exposure range, each of which Pharmacia challenges: back-
    calculation from carpet samples, reliance on remediation calculations in six New
    York schools, and a direct comparison to the New York schools in the study. We
    discuss each approach in turn.
    1.      Back-Calculation from Carpet Samples
    One of Coghlan’s challenged methods involved extrapolating historical air
    levels based on carpet samples obtained from SVEC by teachers in December
    2015. The teachers double-bagged the samples in Ziploc bags and stored them
    23 Pharmacia moved to exclude Coghlan’s estimates of historical air levels under Frye
    and ER 702. The trial court initially granted the motion, noting, “it’s not generally accepted in the
    scientific community yet. It’s never been tested. It’s never been peer reviewed.” Plaintiffs moved
    for reconsideration, which the trial court granted. According to the court, Coghlan’s testimony did
    not use a novel scientific theory and “[t]he Frye objection appears to go to the manner in which
    Coghlan applied the theory.”
    24 National Institute for Occupational Safety and Health.
    46
    No. 83287-5-I/47
    in a paper bag in the basement of a teacher’s house. In 2019, Coghlan had the
    carpet samples tested for PCBs and used those results to estimate the historical
    level of PCBs in the air at SVEC.
    Coghlan based his carpet back-calculation on an EPA study of “PCB
    transport from primary sources to building materials.” Building materials,
    furniture, and other indoor features passively “pick up” PCBs from the
    contaminated air through processes known as advection and diffusion. The
    “source-sink dynamics” approach investigates the mechanisms by which PCBs
    transfer or migrate from a primary source to other secondary materials, or
    “sinks.”
    In the study on which Coghlan based his methodology, the “Guo Study,” 25
    the EPA conducted a controlled chamber experiment to determine the rate at
    which certain building materials and furniture adsorbed 26 airborne PCBs. The
    specific objectives of the study were:
    (1) conduct laboratory experiments to study the transport of PCBs
    through material/air partitioning (i.e., from the air to interior surfaces
    and settled dust) and through material/source partitioning (i.e., from
    primary sources to settled dust); (2) to identify mathematical tools
    that can be used to rank the strengths of PCB sinks and to predict
    their behavior; and (3) to estimate the key parameters required as
    inputs to the mathematical tools, such as sorption capacity, partition
    coefficients, and diffusion coefficients.
    25 Zhishi Guo, et al., EPA, Laboratory Study of Polychlorinated Biphenyl (PCB)
    Contamination and Mitigation in Buildings, Part 2. Transport from Primary Sources to Building
    Materials and Settled Dust, Jan. 2012.
    26 Coghlan’s report describes adsorption as “the adhesion of atoms, ions or molecules
    from a gas, liquid or dissolved solid to a surface of the material.”
    47
    No. 83287-5-I/48
    The methodology of the Guo Study involved placing samples of different
    materials in a chamber and exposing them to a single source of PCBs in the form
    of a field caulk sample. The tested materials included two types of carpet. The
    PCB concentrations in the test chamber air were monitored, and the material
    samples were removed at different times to test their PCB content. These
    samples were used to calculate the concentration of adsorbed PCBs for the
    various materials and to estimate the partition and diffusion coefficients.
    Coghlan used the Guo Study’s calculated rate of accumulation of PCBs in
    carpet from the chamber testing to extrapolate the historical PCB levels in the air
    based on the carpet samples from SVEC. To do so, Coghlan first measured the
    PCB levels in the SVEC carpet samples, then applied the “multiplication property
    of equality,” to rearrange the Guo Study’s mathematical equation for the rate of
    adsorption of airborne PCBs to carpet. Using this new equation, Coghlan
    calculated the historical air levels of PCBs from the known value in the SVEC
    carpet samples. Through this “back-calculation,” Coghlan estimated air values at
    a range of 340 to 9,500 ng/m3 for SVEC during the time the teachers worked at
    the school.
    The parties dispute whether these results are admissible under Frye and
    ER 702. When faced with expert testimony based on novel science, the court
    applies the Frye test, considering (1) whether the underlying theory is generally
    accepted in the scientific community, and (2) whether there are techniques,
    experiments, or studies using the theory that are capable of producing reliable
    48
    No. 83287-5-I/49
    results and are generally accepted in the scientific community. Anderson, 
    172 Wn.2d at 603
    . “To determine whether a consensus of scientific opinion has been
    achieved, the reviewing court examines expert testimony, scientific writings that
    have been subject to peer review and publication, secondary legal sources, and
    legal authority from other jurisdictions.” Eakins v. Huber, 
    154 Wn. App. 592
    , 599,
    
    225 P.3d 1041
     (2010). The scientific community need not be unanimous; the
    court should exclude the expert opinion only if there is a significant dispute
    among qualified scientists. L.M., 
    193 Wn.2d at 128
    .
    Erickson contends that Coghlan’s estimated levels reflect generally
    accepted science of source-sink dynamics as set out in a published and peer-
    reviewed EPA study. Coghlan claimed he “simply re-arranged the equation using
    an uncontroversial and widely-accepted algebra technique to solve for the levels
    of PCBs in the air that were required to produce the levels of PCBs measured in
    the carpet samples, relying on the empirically-derived relationship between PCBs
    in the air and carpet measured by the EPA.”
    According to Erickson, Pharmacia’s complaint is that Coghlan misapplied
    source-sink dynamics by reversing the EPA study’s model. It is true that the
    application of accepted techniques to reach novel conclusions does not implicate
    Frye. Lakey, 
    176 Wn.2d at 919
    . But Erickson mischaracterizes Pharmacia’s
    concern. Pharmacia does not contest the general acceptance of the science of
    source-sink dynamics. Instead, it alleges that the novel science is Coghlan’s
    49
    No. 83287-5-I/50
    method of calculating PCB levels in the air from a known adsorbent source
    material.
    Indeed, several experts testified that the method Coghlan used to arrive at
    his estimates was not generally accepted or valid. John Woodyard, PhD, an
    environmental engineer with an expertise in investigating the presence of PCBs
    and furans, stated that Coghlan’s methods were “unprecedented.” According to
    Woodyard, Coghlan took the PCB results from the carpet “and basically assumed
    that all of the PCBs in that carpet sample came from the air, and then theorized,
    based on some calculations, what the previous air levels had been that would
    have allowed the PCBs to reach that level in the carpet.” Woodyard testified that
    using the carpet to back-calculate air levels was not scientifically valid and had
    never been done before. He said, “you always would rely on real data. You can’t
    back-calculate like that.”
    Environmental scientist Russell Keenan, PhD, who specialized in
    toxicology and risk assessment, provided similar testimony critiquing Coghlan’s
    methodology. He emphasized that “[t]he cardinal rule in risk assessment is to use
    real data.” He further explained, “[y]ou cannot use data from unverified models. A
    model would . . . have to be vetted and peer reviewed before it could be used,
    and even then . . . it is not as high up on the hierarchy as real data are.” But in
    this case, Coghlan used estimated data based on unverified models. Keenan
    testified that Coghlan’s “methodology was a theoretical approach, which has
    never been subjected to peer review, never been subjected to validation, and just
    50
    No. 83287-5-I/51
    doesn’t make sense, when one has real data.” In Keenan’s experience, the EPA
    would not accept an unvalidated model such as the carpet back-calculations.
    Environmental health scientist Shannon Gaffney, PhD, agreed, stating in a
    report, “Mr. Coghlan chooses to use the chamber study calculation in a way that
    is unprecedented, and to our knowledge, a method that has never been proven
    to be valid for real-world scenarios.”
    Coghlan admitted that he did not know of any peer-reviewed literature
    where a scientist took concentrations of PCBs in carpet and from those, then
    estimated levels of PCB in the air, nor was he aware of any scientists who had
    used his method. He also agreed that nobody, including the EPA, had conducted
    the reverse experiment—i.e., used a piece of PCB-contaminated carpet in a
    chamber and then measured the resulting PCB concentrations in the air—in a
    manner that would support his calculations.
    Moreover, the Guo Study underpinning Coghlan’s methodology provided a
    warning on the use of its source-sink data. The study notes that it was limited
    only to laboratory testing. It explicitly states,
    Because of time constraints, this study tested only a small number
    of sink materials (20 building and furniture materials and two types
    of dust). The number of tests conducted was also rather small.
    There are many types of building and furniture materials, and there
    are many brands and varieties of each type, all of which have
    different physical and chemical properties. Thus, care should be
    taken when applying the test results to seemingly similar materials
    in real-world situations.
    Coghlan’s methodology ignored these stated limitations and purported to
    calculate results in a “real-world situation” without the “stable source of PCBs” as
    51
    No. 83287-5-I/52
    in the study. Coghlan acknowledged that the “rough estimate” values from the
    Guo Study had not been validated in real world situations outside of the
    experimental chamber. In a section entitled “Method Limitations,” the Guo Study
    stated “the results should be treated as rough estimates. To solve the
    fundamental problem, the partition and diffusion coefficients should be
    determined independently.” Coghlan admitted he did not independently
    determine the partition and diffusion coefficients with respect to the carpet.
    He also conceded that in performing these calculations, he assumed what he
    was trying to prove, that the PCBs in the carpet came from the air.
    Lake Chelan Shores Homeowners Ass’n v. St. Paul Fire & Marine
    Insurance Co. is instructive. 
    176 Wn. App. 168
    , 
    313 P.3d 408
     (2013). There, in
    order to determine whether an insurance policy applied, a condominium
    association relied on expert testimony to identify the first date that hidden decay
    reached a collapse condition. 
    Id. at 177
    . The expert used a mathematical formula
    to back-date and trace the progression of wood rot. 
    Id.
     The condominium
    association argued that Frye did not apply because the science of wood decay
    was not new or novel, “but instead is well known and well established and,
    further, that the mathematical equation used by its experts relies on that
    accepted science to draw conclusions about when the rot caused certain
    buildings to reach a state of collapse.” 
    Id. at 180
    . The court noted the issue was
    not “the science of wood decay,” or the conclusions reached; rather, the question
    was whether the methodology of the formula backdating the decay process was
    52
    No. 83287-5-I/53
    generally accepted. 
    Id. at 180-81
    . Both the theory and the methodology used to
    implement the theory must be generally accepted in the scientific community. 
    Id. at 175
    . But there, the equation “did not come from any scientific literature”; the
    expert described his calculations as “ ‘educated guesses’ and was unable to
    identify any other person or literature stating his formula is a proper equation for
    estimating rot progression.” 
    Id. at 177
    . Nor did any witness testify that the
    scientific community generally accepted the equation the expert used to “work[]
    backward from present rot conditions.” 
    Id. at 178
    .
    In this case, as in Lake Chelan, there is no dispute that the underlying
    theory—there, wood decay, and here, “source-sink dynamics”—is generally
    accepted in the scientific community. However, the methodology used by the
    experts in both cases is not. The Guo Study was an effort to “fill some of the data
    gaps associated with the characterization of PCB sinks in contaminated
    buildings,” not to apply or develop a replicable methodology based on the theory
    of source-sink dynamics to determine the amount of PCBs in the air at the outset.
    Thus, based on the other experts’ opinions as well as his own admissions,
    Coghlan’s methodology of using data from the Guo Study to determine historical
    PCB levels in the air, particularly when it is unknown if the PCB source was
    stable, does not enjoy the same general acceptance as the theory of source-sink
    dynamics. Coghlan did not merely apply a generally accepted methodology;
    rather, like the expert in Lake Chelan, he developed a novel equation to “work
    53
    No. 83287-5-I/54
    backward” from present condition of carpet samples that were obtained from the
    “real-world” and not maintained or controlled in a laboratory setting.
    The opinions from multiple experts 27 consistently state that Coghlan’s
    methodology of back-calculating historical PCB air levels from SVEC carpet
    samples was novel and not generally accepted in the scientific community.
    Coghlan himself acknowledged he did not know of any scientists who had used
    his method of using concentrations of PCBs in carpet to estimate levels of PCBs
    in the air, much less any peer-reviewed literature validating this method. It is
    precisely in such a situation as this that the trial court must “serve the
    gatekeeping function of keeping out ‘unreliable, untested, or junk science.’ ” L.M.,
    
    193 Wn.2d at 127
     (quoting Anderson, 
    172 Wn.2d at 606
    ). Therefore, under Frye,
    the trial court should have excluded Coghlan’s testimony as to specific historical
    PCB levels that he calculated using this novel method of back-calculation derived
    from the Guo Study data.
    2.      Calculations Based on EPA Study of New York Schools
    Coghlan also relied on an EPA study of PCB air levels in six New York
    schools (New York schools study). This study measured air levels of PCBs in the
    schools before and after PCB remediation. The purpose of the study was to help
    improve the understanding of PCBs in school buildings and approaches for
    27 We note that in assessing admissibility under Frye, we are not limited to reviewing the
    evidence from the parties. Ruff v. Dep’t of Labor & Indus., 
    107 Wn. App. 289
    , 300, 
    28 P.3d 1
    (2001), overruled on other grounds by Anderson, 
    172 Wn.2d 593
    . Here, there is no information in
    the record or otherwise supporting the conclusion that Coghlan’s back-calculation methodology
    for determining historical levels of PCBs in the air enjoys general acceptance.
    54
    No. 83287-5-I/55
    mitigating exposures. The study’s research objectives included characterizing
    levels of PCBs in various aspects of the school environment and investigating
    relationships between PCB sources and environmental levels, applying an
    exposure model for estimating children’s exposure to PCBs in schools, and
    providing information to assist in developing risk management practices for
    reducing exposure in schools. To accomplish this,
    [a] limited set of measurement data and information collected using
    a systematic approach at six schools was used to characterize
    primary sources of PCBs, to evaluate whether secondary sources
    were present and their relative importance, to describe PCB
    concentrations in school environmental media, and to prepare
    modeled estimates of exposure and characterize the relative
    importance of different routes of exposure.
    The New York schools study itself acknowledged, “[i]t is not known if these
    results are representative of older schools nationwide, both in terms of the
    presence of PCB-containing materials and components and the environmental
    concentrations measured in and around the school buildings.” Moreover,
    “considerable variability” existed both between and within schools.
    Despite the variability within the New York schools and the unknown
    differences between remediation performed in those schools and SVEC, Coghlan
    employed data from the New York schools study to “back-extrapolate” the PCB
    concentration in the air at SVEC pre-remediation, circa 2011. According to
    Coghlan’s report, the New York schools study “shows the relative impact of
    remediation (as was done at SVEC in 2014-2016) in reducing levels of PCBs in
    the indoor air. Any data collected during or after the remediation work at SVEC
    55
    No. 83287-5-I/56
    would likely be some fraction of the pre-remediation levels in the School.” To
    adjust for the impact of remediation and to estimate pre-remediation levels at
    SVEC, Coghlan used the New York schools data to calculate “a simple ratio” of
    effectiveness of remediation. He determined the median post-remediation levels
    of indoor air for the five New York schools was approximately 28 percent of pre-
    remediation levels, with a range of 10 percent to 72 percent. He calculated that
    the median pre-remediation levels were about 3.5 times higher compared to
    post-remediation levels with a range of effectiveness of 1.4 to 10.
    Coghlan used the lowest and highest factors (1.4 and 10) and applied
    them to adjust the February and May 2016 air levels from the SVEC pod building
    to obtain a pre-remediation estimate. Out of the hundreds of measurements of
    PCB levels taken throughout SVEC in February and May 2016, Coghlan selected
    the highest and lowest readings. He applied the 1.4 effectiveness factor to the
    lowest reading of PCBs detected at SVEC and the factor of 10 to the highest
    measured value. 28 These two data points became the outer limits of the pre-
    remediation exposure band Coghlan calculated for SVEC. This process of
    estimating pre-remediation PCB levels at SVEC was not based on a known
    method applied to the facts of this case, but a method Coghlan derived from an
    unrelated dataset from a single study that did not purport to make any similar
    calculations.
    28 As an example, Coghlan explained a pre-remediation PCB reading of 1,000 ng/m3 after
    remediation with an effectiveness factor of 10 would yield a post-remediation reading of 100
    ng/m3.
    56
    No. 83287-5-I/57
    Pharmacia’s experts strongly disagreed with Coghlan’s reliance on the
    method he developed based on the New York schools study. Dr. Woodyard
    testified “the two datasets, the schools, if you will, have nothing in common.
    [Coghlan] could not demonstrate that they were identical in terms of remediation
    or other conditions.” To derive the pre-remediation numbers at SVEC using the
    high/low remediation factors he had calculated based on the New York study,
    Coghlan had to assume “that what had been done in these schools in New York
    was exactly the same as what was done at Sky Valley, which was not.”
    According to Dr. Woodyard, this method “provides a gross exaggeration of
    historical PCB levels, which is what he was trying to show.” Woodyard further
    opined that Coghlan’s method of calculating pre-remediation PCB levels in the air
    is “totally inappropriate” and not accepted within the scientific community.
    Coghlan used the data from the New York schools to calculate a
    “remediation coefficient” and estimate the adjusted SVEC air levels in a
    completely different setting. 29 In so doing, Coghlan employed a novel method
    that is not generally accepted in the scientific community. The trial court should
    have excluded Coghlan’s testimony premised on this method, as it fails the Frye
    test.
    29 Coghlan said that “schools are remarkably similar in many ways,” and schools built in
    the same era typically have the same style of ventilation. However, there is no evidence of other
    studies using similar methodology to develop a “remediation factor” from other school samples.
    57
    No. 83287-5-I/58
    3.     New York Schools Direct Comparison
    Coghlan’s final approach to estimating historical PCB levels at SVEC was
    a direct comparison, rather than application of any methodology to the SVEC
    data. We evaluate this evidence under ER 702, not under the Frye test. We
    review a trial court’s admission of expert testimony under ER 702 for abuse of
    discretion. L.M., 
    193 Wn.2d at 134
    .
    Coghlan explained how he determined SVEC PCB levels through
    comparison with the New York schools:
    [T]he study provides a range of values for different school
    environments. And so I took -- to be conservative, I took the lowest
    value they reported and the highest value to say that this is the
    possible range or likely possible range of values that could have
    existed at Sky Valley.
    Coghlan suggested that the New York schools were similar enough to SVEC for
    comparison:
    [S]chools are remarkably similar in many ways. If you go in and see
    a school system -- a school has been built, say, pre-1975 or
    something, the architect and the style of the ventilation typically
    uses heat ventilators with exhaust systems or some kind of
    combination thereof, or window AC units, and will use larger
    ventilation systems for things like gyms and locker rooms, and
    things of that nature.
    In response, Pharmacia pointed to the warning in the New York schools
    study itself that it “is not known if these results are representative of older schools
    nationwide.” The study noted that “PCB emissions from materials and light
    ballasts were not directly measured at the six schools. Modeled emission
    estimates and the resulting predictions of indoor air concentrations have
    58
    No. 83287-5-I/59
    considerable uncertainties.” The study also explained, “[e]missions from light
    ballasts are likely to vary depending on the lighting fixture design and the
    condition of the ballast and capacitor.” Dr. Keenan noted that the study’s authors
    warned against extrapolating the results to other school systems and testified
    “that every building is different, every room is different, so it behooves you to use
    real data, which is location specific.”
    But these concerns go to the weight of Coghlan’s testimony that the New
    York schools were directly comparable to SVEC rather than its admissibility. To
    the extent Coghlan did not apply any novel calculations derived from the New
    York schools data, but rather, opined that the New York schools were sufficiently
    similar to SVEC to make a fruitful comparison, this testimony would assist the
    jury in determining pre-remediation PCB levels. Allowing the jury to hear this
    testimony and weigh it against the testimony from Pharmacia’s experts was not
    an abuse of discretion. 30
    B.      Richard Perrillo
    Neuropsychology expert Richard Perrillo, PhD, opined that all three
    plaintiffs had “abnormal brain[s]” caused by “caused by chronic toxic exposure.”
    Dr. Perrillo reached these conclusions after conducting various
    neuropsychological tests on each plaintiff, as well as other teachers and
    30 We likewise reject Pharmacia’s challenges to Coghlan’s testimony as inadmissible
    under ER 702 because he did not defer to the 2016 testing results and because he relied on
    unreliable carpet samples. These concerns about his testing procedures relate to weight, not
    admissibility.
    59
    No. 83287-5-I/60
    students. The trial court admitted the evidence, stating that Pharmacia could call
    their experts to dispute the testimony “by saying this isn’t good science, this isn’t
    the way you do it, these things are not, you know, sufficient incidents, we don’t
    know what the cause is, you know, whatever else they want to argue about it.”
    On appeal, Pharmacia argues that in so ruling, the trial court “effectively
    held” that all Frye challenges should be resolved through cross-examination,
    when it should have excluded the evidence. Specifically, Pharmacia claims (1)
    Dr. Perrillo is unqualified to give a causation opinion and (2) because he invented
    and employed novel and unreliable testing methodologies to show cognitive
    impairment from PCB exposure, his causation testimony is inadmissible under
    Frye, 31 as well as unhelpful to the trier of fact under ER 702.
    1.      Qualifications
    Pharmacia argues Dr. Perrillo was not qualified to opine that plaintiffs
    suffered from cognitive deficits caused by chronic PCB exposure because he is
    not a medical doctor and has no training in medicine or toxicology that would
    allow him to diagnose the case of cognitive symptoms. We agree with Erickson
    31 Erickson claims Pharmacia failed to raise Frye in its trial court briefing challenging
    admission of Dr. Perrillo’s opinion evidence. While its briefing does not discuss Frye, Pharmacia
    raised Frye issues during oral argument on motions in limine. Pharmacia argued that Dr. Perrillo
    created novel composite testing scores that had not been peer-reviewed, validated through
    testing or adopted by any organization. Pharmacia also asserted Dr. Perrillo employed a
    methodology, which relates directly to the Frye analysis. The record shows that Pharmacia
    challenged Dr. Perrillo’s testimony as novel under the Frye test and, therefore, did not waive the
    issue.
    60
    No. 83287-5-I/61
    that despite lacking those specific credentials, Dr. Perrillo was qualified to testify
    about neuropsychological effects of PCB exposure.
    Under ER 702, a witness may qualify as an expert “by knowledge, skill,
    experience, training, or education.” “When determining whether a witness is an
    expert, courts should look beyond academic credentials.” L.M., 
    193 Wn.2d at 135
    . As an example, a nonphysician may qualify as an expert to testify in a
    medical malpractice case. 
    Id.
     In assessing an expert’s qualifications, the court
    must consider whether the expert has sufficient expertise in the relevant
    specialty. 
    Id.
     “[T]rial courts are afforded wide discretion, and trial court expert
    opinion decisions will not be disturbed on appeal absent an abuse of such
    discretion.” Johnston-Forbes v. Matsunaga, 
    181 Wn.2d 346
    , 355, 
    333 P.3d 388
    (2014).
    Here, Dr. Perrillo has significant experience with neuropsychological
    assessments of both abnormal and normal populations, having evaluated
    thousands of patients in the last 34 years. Dr. Perrillo testified that he had worked
    extensively with patients experiencing toxic acquired brain injury, including
    people exposed to toxic solvents, toxic treatment for hepatitis, and silicone breast
    implant ruptures. His current area of expertise involves the clinical and
    neuropsychological effects of toxic exposure on the brain. During the last 20
    years, he has been involved in “beta testing” and standardization of well-known
    neuropsychological tests, and he is currently on the standardization committee
    for upcoming versions of several commonly used tests. The trial court did not err
    61
    No. 83287-5-I/62
    in determining that Dr. Perrillo was qualified to conduct neuropsychological
    testing and testify as to whether plaintiffs’ injuries arose from toxic PCB
    exposure. 32
    2.       Neurological Testing
    Pharmacia claims that Dr. Perrillo failed to apply the generally accepted
    methodologies for neuropsychological testing and relied instead on his own
    unpublished, unreliable “novel composite score methodologies” that are not
    generally accepted. Pharmacia argues Dr. Perrillo’s testimony fails to satisfy Frye
    in two respects: he departed from generally accepted methodologies in applying
    the Test of Premorbid Function (TOPF) that estimates pre-exposure cognitive
    function, and he used novel methodologies to create composite scores of
    plaintiffs’ capabilities. Erickson contends Dr. Perrillo used methods long-accepted
    by the scientific community that do not implicate Frye.
    Dr. Perrillo used the TOPF to determine the pre-exposure function of the
    plaintiffs, conducted a battery of tests to measure current functionality, and
    compared individual scores and composite scores. Pharmacia acknowledges the
    standard TOPF method for identifying neuropsychological deficits requires
    “comparing premorbid level of function to current with lower scores interpreted as
    32 Pharmacia points out that other courts have excluded Dr. Perrillo from testifying in
    similar cases. E.g., Sanderson v. Int’l Flavors and Fragrances, Inc., 
    950 F. Supp. 981
    , 1001 (C.D.
    Cal. 1996) (court determined that Perrillo, as a neuropsychologist, was not qualified to give expert
    testimony on toxic causation). However, the broad abuse of discretion standard means that
    courts can reasonably reach different conclusions about whether an expert is qualified. L.M., 
    193 Wn.2d at 136
    .
    62
    No. 83287-5-I/63
    deficits.” Indeed, Pharmacia’s own expert conducted similar TOPF testing and
    analyzed data for comparison.
    The difference of opinion between Pharmacia’s expert Dr. Schoenberg
    and Dr. Perrillo lies in the interpretation and manipulation of the data and the
    resulting conclusions—not the administration of the tests. Pharmacia
    characterizes Dr. Perrillo’s methodology as novel and not generally accepted
    because he used only plaintiffs’ TOPF actual reading scores as their premorbid
    baseline and then compared the results of unrelated neuropsychological tests for
    30 different cognitive functions, even though they were not correlated to the
    TOPF actual reading score. Pharmacia argues that under the generally accepted
    TOPF methodology, to determine changes in pre- and postmorbid capabilities,
    the examiner must combine the TOPF actual and predicted reading scores with
    demographic variables to estimate five index scores that are psychometrically
    correlated to the Wechsler set of tests, and then compare each of those five
    against the person’s actual Wechsler test scores.
    “The application of accepted techniques to reach novel conclusions does
    not raise Frye concerns.” Lakey, 
    176 Wn.2d at 919
    . Rather, Frye “applies where
    either the theory and technique or the method of arriving at the data relied upon
    is so novel that it is not generally accepted by the relevant scientific community.”
    Anderson, 
    172 Wn.2d at 611
    . Because Pharmacia’s critique involves Dr.
    Perrillo’s novel interpretation of the data, rather than a novel method of “arriving”
    63
    No. 83287-5-I/64
    at the data, Frye does not preclude Dr. Perrillo’s use of TOPF and his diagnostic
    opinion testimony.
    Similarly, Pharmacia challenges Dr. Perrillo’s use of composite scores as
    “wholly novel . . . methodologies that he invented.” Perrillo compared the
    plaintiffs’ TOPF actual reading scores against scores on other cognitive functions
    that were not psychometrically correlated to the reading score, then used the
    baseline comparisons to create composite scores for each plaintiff. According to
    Pharmacia, the use of composite scores is a generally accepted
    neuropsychological methodology only when the methodology has been
    published, peer reviewed, and researched to confirm validity and reliability.
    Erickson counters that the use of composite scores is common in
    neuropsychology, as even Pharmacia’s expert acknowledged, and also points to
    studies involving composite scores for executive function. 33 Additionally, Dr.
    Perrillo testified that composite scores have better predictive validity than single
    scores.
    Moreover, the type of testing that generated the components of the
    composite scores is undisputedly generally accepted. As our Supreme Court has
    stated, “Frye does not require every deduction drawn from generally accepted
    theories to be generally accepted.” Anderson, 
    172 Wn.2d at 611
    . “Other
    33 See, e.g., Br. of Resp’ts at 115 (citing Yana Suchy, Executive Functioning: A
    Comprehensive Guide for Clinical Practice (Oxford Univ. Press 2016)).
    64
    No. 83287-5-I/65
    evidentiary rules—not Frye . . . —bar deductions that are too speculative.” L.M.,
    
    193 Wn.2d at 131
    .
    Because Dr. Perrillo’s opinions were derived from generally accepted
    neuropsychological testing methods, his testimony satisfies Frye. The trial court
    did not err in determining Dr. Perrillo’s opinions using the TOPF and composite
    scores were not precluded by Frye.
    3.     ER 702
    Pharmacia argues that even if they satisfy Frye, Dr. Perrillo’s cognitive
    impairment opinions are unreliable, lack foundation, and do not assist the trier of
    fact because they do not adhere to the generally accepted neuropsychological
    methodologies. The trial court disagreed, concluding that Pharmacia’s arguments
    related to weight rather than admissibility of the evidence. As previously noted,
    “an expert’s errors in applying proper procedures go to the weight, not the
    admissibility, of the evidence unless the error renders the evidence unreliable.”
    Lakey, 
    176 Wn.2d at 920
    .
    To demonstrate the reliability of his testimony, Dr. Perrillo identified
    scientific literature on the effects of PCBs consistent with his findings from
    plaintiffs’ neuropsychological testing. In a declaration supporting Erickson’s
    opposition to Pharmacia’s motion in limine to exclude the non-party injury
    evidence, Dr. Perrillo noted that the pattern of brain injury “is similar to what the
    literature shows one would expect to see in people exposed to toxic substances
    capable of crossing the blood/brain barrier.”
    65
    No. 83287-5-I/66
    To formulate his opinion that plaintiffs’ brain injuries were caused by PCB
    and furan exposure, Dr. Perrillo ruled out other causes at SVEC. In addition, Dr.
    Perrillo considered studies that investigated the neurocognitive injuries that can
    be caused by PCBs. He then concluded PCBs caused the injuries: “[I]t would
    appear that it was, you know, a PCB exposure . . . I ruled out a number of things
    like mold and, you know, lead, other things that could have possibly been there,
    but there is just nothing compelling to say that the other elements were there
    other than that.” He further articulated,
    Well, it’s a matter of elimination. In this particular case there was a
    mold study done, which said that the east pod that the teachers
    were in had no evidence of mold. There were -- there was no
    evidence I saw of lead, you know, there were no serum levels of
    lead, there was no hair analysis, no blood analysis, there was just
    nothing that showed that there was lead there, so that’s out. The
    mold was out, because the lab that had done it, the certified lab,
    said that there was no mold . . . And so, you know, you just left
    them with a few possibilities, and one of the possible and
    probabilities -- one of the probabilities is PCB.
    Dr. Perrillo also relied on his testing of 52 SVEC occupants to reach his
    conclusion that their brain injuries arose from toxic exposure to PCBs. 34 He
    testified that the group had significant impairments and all had the same
    environment, SVEC, in common.
    Pharmacia’s arguments regarding the reliability of this evidence go toward
    the weight it should be afforded, rather than its admission. Given the deferential
    34 We address separately below the arguments that Dr. Perrillo’s testimony relying on
    evidence of brain injury from non-parties who were present at SVEC was inadmissible and
    prejudicial, along with the analysis of other arguments relating to evidence of non-party injuries.
    66
    No. 83287-5-I/67
    abuse of discretion standard for reviewing admission of evidence under ER 702,
    the trial court did not err in admitting Dr. Perrillo’s diagnostic opinions.
    C.     James Dahlgren
    James Dahlgren, MD, is an internal medicine doctor with a subspecialty in
    toxicology. He provided testimony on causation for plaintiffs’ injuries, stating that
    plaintiffs have brain damage and “have been poisoned by PCBs and the furans
    that were present at the Sky Valley school.” Dr. Dahlgren reviewed the medical
    records and interviewed the plaintiff teachers and several other teachers,
    parents, and students who spent time in SVEC, noting they experienced
    headaches, dizziness, brain fog, weakness, and nausea. He explained “each day
    that they had some exposure . . . there would be some effect on the brain.” The
    exposure builds up, causing more symptoms because brain cells damaged by
    PCBs do not recover.
    Dr. Dahlgren also testified about the role furans played in the brain
    injuries: “[F]urans are toxic in a way that is additive or even synergistic with PCBs
    in the immune system and also in the brain.” He concluded that the plaintiffs
    incurred brain damage due to both PCBs and furans. After reviewing reports
    prepared by Dr. Perrillo, Dr. Mahoney, and Coghlan, Dr. Dahlgren opined that the
    PCB levels estimated by Coghlan were sufficient to cause brain injury.
    Pharmacia argues Dr. Dahlgren is not qualified to testify about cognitive
    impairment because he is not a neurologist or neuropsychologist and has never
    administered or interpreted neuropsychological testing. Pharmacia also
    67
    No. 83287-5-I/68
    challenges the admission of Dr. Dahlgren’s testimony due to his reliance on
    opinions from Dr. Perrillo and Coghlan.
    Expert testimony is admissible if the expertise is supported by the
    evidence, the opinion is based on material reasonably relied on in the
    professional community, and the testimony is helpful to the trier of fact. Deep
    Water Brewing, LLC v. Fairway Res. Ltd., 
    152 Wn. App. 229
    , 271, 
    215 P.3d 990
    (2009). “One expert may rely on the opinions of another expert when formulating
    opinions.” Driggs v. Howlett, 
    193 Wn. App. 875
    , 900, 
    371 P.3d 61
     (2016).
    Pharmacia argues Dr. Dahlgren’s testimony was inadmissible because he
    did not perform any of his own cognitive testing and based his finding of
    moderate brain dysfunction on Dr. Perrillo’s cognitive impairment opinions. While
    Dr. Dahlgren interviewed the plaintiffs, he did not perform any of his own testing
    and did not make an independent diagnosis. He conceded that he relied on Dr.
    Perrillo’s neuropsychological testing to diagnose “moderate brain dysfunction” in
    the three plaintiffs. As discussed above, Dr. Perrillo’s expert scientific evidence
    was properly admitted. Therefore, Dr. Dahlgren could properly rely on Dr.
    Perrillo’s admissible expert opinions as a basis for his opinions.
    However, to the extent Dr. Dahlgren’s exposure-based causation opinions
    rely on Coghlan’s inadmissible opinions about PCB levels at SVEC, his opinions
    should have been excluded. When testifying, Dr. Dahlgren opined that the levels
    estimated by Coghlan were sufficient to cause brain injury. In Dr. Dahlgren’s
    report, the sections pertaining to each plaintiff’s exposure to PCBs and furans at
    68
    No. 83287-5-I/69
    SVEC direct the reader to “refer to the report by Kevin Coghlan” or his company,
    EH&E. To the extent Dr. Dahlgren’s causation opinions relied on Coghlan’s PCB
    levels that were calculated using novel methodologies, they were not based on
    material reasonably relied on in the professional community, and, therefore, were
    inadmissible. On the other hand, Dr. Dahlgren’s opinions that did not rely on
    Coghlan’s estimates are admissible. 35
    IV.     Evidence of Nonparty Harms
    The parties contested the introduction of evidence of nonparty injuries of
    SVEC occupants other than plaintiffs. This contested evidence falls into two
    categories: (1) testimony by plaintiffs’ experts Dr. Perrillo and Dr. Pamela
    Mahoney incorporating information from, respectively, neuropsychology testing
    and a “health assessment survey” involving individuals other than the three
    plaintiffs who were connected to SVEC and (2) testimony by nonparty witnesses
    about people connected with SVEC experiencing illness. According to
    Pharmacia, the trial court erred in allowing this type of nonparty evidence in
    plaintiffs’ case-in-chief as substantive evidence of exposure and causation.
    Erickson counters that the expert evidence was admissible, and to the extent
    evidence of nonparty harm was not relevant, it was duplicative of other evidence,
    and, thus, not prejudicial.
    35 Dr. Dahlgren also cited sources other than Coghlan in his causation testimony. For
    example, Dr. Dahlgren testified that the types of symptoms experienced by the plaintiffs would be
    experienced at exposure rates of 13,000 to 260,000 ng/m3 and that the plaintiffs would not have
    experienced any symptoms at rates of less than 1,000 ng/m3. Dr. Dahlgren also testified that an
    ecology study found that PCBs in the amount of one to five nanograms per cubic meter were
    enough to cause people to have serious illness.
    69
    No. 83287-5-I/70
    A.    Expert Opinions Relying on Evidence from Nonparties
    Pharmacia sought to exclude, and Erickson sought to admit Dr.
    Mahoney’s health assessment survey of 164 children and adults who spent time
    at SVEC who experienced similar symptoms and epidemiological analysis based
    on that survey. Pharmacia also sought to exclude Dr. Perrillo’s testimony of
    neuropsychological testing of more than 50 SVEC occupants involved in various
    other lawsuits. The trial court admitted the evidence both pretrial and during
    additional arguments during trial, and post-trial, it denied reconsideration. At the
    time of trial, the court concluded,
    [I]t’s fine for the expert to offer an opinion based upon having
    surveyed a bunch of folks, and without talking about a specific
    person.
    In other words, the expert can have an opinion, say, you
    know, I’ve collected information about what people have
    experienced at this locale, without, you know, going and saying
    that, you know, Suzy Jones got this specific problem and this is the
    kind of damages she had and so on, that that’s different.
    Based on this, the trial court allowed Drs. Perrillo and Mahoney to testify “about
    what they found . . . in doing their analysis and it’s a classic battle of the experts
    stuff.”
    We review evidentiary decisions for abuse of discretion. Gerlach v. Cove
    Apts. LLC, 
    196 Wn.2d 111
    , 119, 
    471 P.3d 181
     (2020). A trial court abuses its
    discretion if no reasonable person would take the view adopted by the trial court.
    State v. Jennings, 
    199 Wn.2d 53
    , 59, 
    502 P.3d 1255
     (2022).
    “Expert medical testimony is necessary to establish causation where the
    nature of the injury involves ‘obscure medical factors which are beyond an
    70
    No. 83287-5-I/71
    ordinary lay person’s knowledge.’ ” Fabrique v. Choice Hotels Int’l, Inc., 
    144 Wn. App. 675
    , 685, 
    183 P.3d 1118
     (2008) (quoting Riggins v. Bechtel Power Corp.,
    
    44 Wn. App. 244
    , 254, 
    722 P.2d 819
     (1986)). An expert witness may base an
    opinion on facts or data that are not admissible in evidence if the facts or data
    are “of a type reasonably relied upon by experts in the particular field in forming
    opinions or inferences upon the subject.” ER 703.
    In forming their expert opinions, both Dr. Perrillo and Dr. Mahoney relied
    on the information about the experience of nonparties who were exposed to
    PCBs while at SVEC. Pharmacia argues that the nonparty evidence was not
    relevant as a matter of law, and therefore, was inadmissible. But an expert’s
    testimony disclosing inadmissible facts or data to explain the expert’s opinion “is
    not proof of them” as substantive evidence. State v. Caril, 23 Wn. App. 2d 416,
    428, 
    515 P.3d 1036
     (2022) (citing Grp. Health Coop. of Puget Sound, Inc. v.
    Dep’t of Revenue, 
    106 Wn.2d 391
    , 399-400, 
    722 P.2d 787
     (1986); State v.
    Wineberg, 
    74 Wn.2d 372
    , 381-82, 
    444 P.2d 787
     (1968)).
    “When a party seeks to introduce otherwise inadmissible facts or data
    through an expert witness who has relied on them, the trial court has discretion to
    determine the extent to which the expert may relate the inadmissible information
    to the trier of fact.” Caril, 23 Wn. App. 2d at 427 (citing ER 705). A trial court
    should weigh the probative and prejudicial value under ER 403 to determine
    whether to allow disclosure of inadmissible underlying facts. 
    Id. at 428
    . This
    “prevent[s] an expert’s opportunity to explain the basis for an opinion from
    71
    No. 83287-5-I/72
    becoming merely ‘a mechanism for admitting otherwise inadmissible evidence’ or
    ‘to avoid the rules for admissibility of evidence.’ ” 
    Id. at 427
     (quoting State v.
    Anderson, 
    44 Wn. App. 644
    , 652, 
    723 P.2d 464
     (1986)). Trial courts have wide
    discretion in balancing the prejudicial and probative values of evidence. Gerlach,
    196 Wn.2d at 120.
    Dr. Mahoney is an epidemiologist. She reviewed government publications
    and scientific literature to identify the signs and symptoms of exposure to PCBs
    and furans. She identified seven categories of illnesses or symptoms commonly
    reported with PCB and furan exposure. Dr. Mahoney then examined health
    assessment surveys that were completed by 164 adults and children who
    formerly spent time at SVEC for complaints related to these categories. 36 Dr.
    Mahoney conducted an epidemiological analysis based on these surveys to
    determine whether SVEC occupants reported a higher incidence of certain
    symptoms after spending time at the school. Dr. Mahoney testified that they
    “found statistical significance” of increased symptoms in adults and children after
    exposure to SVEC. Dr. Mahoney concluded, after exposure to SVEC, the
    proportion of people reporting symptoms increased in every health category
    known to be affected by PCB and furan exposure.
    Dr. Mahoney testified as to her opinions gathered as a result of her
    epidemiological study of the symptoms experienced by 164 adults and children
    36 Dr. Mahoney acknowledged that not all visitors to SVEC had completed questionnaires
    and those who completed them were not a random selection.
    72
    No. 83287-5-I/73
    who had spent time at SVEC. The information provided in the health
    assessments was of a type reasonably relied upon to study the health
    experiences of participants before and after exposure. Any testimony disclosing
    inadmissible facts or data to explain her opinion was not substantive evidence,
    but background on how she arrived at her opinions. Dr. Mahoney did not testify
    about the facts of the health assessments; she testified about her expert opinions
    derived from this information. This testimony is permissible under ER 703, and its
    admission was not an abuse of discretion by the trial court.
    Pharmacia also requested that the trial court exclude Dr. Perrillo’s
    testimony about information gathered about other adults and children who spent
    time at SVEC. Dr. Perrillo, as discussed above, is a neuropsychologist. He
    conducted neuropsychological testing of not only plaintiffs, but 52 people total
    who had been at SVEC. In analyzing the results, he grouped the 17 children into
    two groups, ages 7 to 15 and 16 to 19, and then adults 20 years or older. He
    testified that “you see the same pattern, you know, of acquired brain injury” in the
    adults tested and, overall, those examined showed an “unusual level of brain
    impairment.” He explained, “The take-away is that, you know, you have a group
    of people with diminished cognition. You have a group of people that have
    significant impairments across the board.” Dr. Perrillo then made the express link
    that the commonality among all the individuals was their environment, specifically
    SVEC. “[T]hey all shared a common experience. The kids did, the teachers did,
    and even though there may be individual variables or individual affects, you
    73
    No. 83287-5-I/74
    know, as a group you could see where the trend is -- I mean, it’s clear, it’s very
    clear where the trend is here.” Thus, Drs. Perrillo and Mahoney testified to their
    opinions about medical causation of plaintiffs’ injuries, relying in part on
    information gathered from people exposed to the SVEC environment. The trial
    court did not abuse its discretion in allowing the experts to state the basis for
    their opinions, even to the extent it relied on information gleaned from
    individuals. 37 While that information may have been prejudicial to Pharmacia, it
    was not unfairly so when it was presented as part of the basis for the expert
    opinions rather than as substantive evidence.
    B.       Lay Witness Testimony
    Erickson also sought to admit, and Pharmacia sought to exclude, direct
    lay witness testimony of nonparty injury. Pharmacia argues the evidence is not
    admissible under the constraints on such testimony prescribed in Intalco
    Aluminum v. Department of Labor and Industries, 
    66 Wn. App. 644
    , 
    833 P.2d 390
     (1992).
    In Intalco, the plaintiffs brought workers’ compensation claims for
    occupational diseases sustained as a result of air pollution in an aluminum plant.
    37 By contrast, in Caril, while the evidence supporting an expert’s opinion was relevant,
    the trial court excluded it because the prejudice from the evidence was substantially outweighed
    by the danger of unfair prejudice to both parties, it could mislead the jury, and it could confuse the
    issues. 23 Wn. App. 2d at 428. The trial court allowed the expert psychologist to testify that the
    defendant lacked the capacity to form criminal intent at the time of the murder for which he was
    charged. Id. at 420. However, the court prohibited the expert from testifying to hearsay
    statements from another psychologist’s report that the expert had relied on, because the
    excluded statements concerned the collateral issues of the defendant’s competency to stand trial
    and potential future need for civil commitment. Id. This court reasoned that hearing about that
    information “could confuse the jury or divert [it] from the issues it was charged with deciding.” Id.
    at 429.
    74
    No. 83287-5-I/75
    Id.at 647-48. The employer challenged the admission of testimony from two of
    the plaintiffs’ coworkers who experienced and observed other workers with
    similar symptoms, arguing it was irrelevant and the prejudicial impact outweighed
    the probative value. Id. at 664. This court disagreed, concluding the testimony
    was properly admitted as rebuttal evidence to the employer’s expert witness,
    medical director for a different aluminum manufacturing company, who testified
    that he had never observed similar neurologic symptoms among 3,000
    employees performing similar work. Id. at 665. The coworkers’ testimony was
    “directly relevant to rebut the inferences raised” by this expert testimony, i.e., that
    he had never observed similar symptoms. Id. The trial court also gave a limiting
    instruction that the coworkers’ testimony was not offered to prove that they had
    the same or similar conditions as the claimants, and should only be considered
    on the question of whether others in the aluminum industry had similar
    symptoms. Id. 38 Thus, the trial court limited any prejudice from the testimony. Id.
    Here, the trial court admitted certain nonparty evidence of illness at SVEC,
    explaining as follows: 39
    38 A subsequent unpublished case emphasizes the role of this evidence in rebuttal.
    Shoemake v. Eli Lilly & Co., No. 72716-8-I, slip op. at 9 (Wash. Ct. App. June 13, 2016)
    (unpublished), https://www.courts.wa.gov/opinions/pdf/727168.pdf. While unpublished cases are
    not binding, we may cite and accord them such persuasive value as we deem appropriate. GR
    14.1(a). In Shoemake, a workers’ compensation case, an employee filed a claim alleging that
    fumes and odors from a building remodel injured her. Shoemake, slip op. at 1. The trial court
    excluded the testimony of two lay witnesses who stated they had suffered similar symptoms
    during the remodel because the evidence was not offered as rebuttal, and this court agreed.
    Shoemake, slip op. at 8-9.
    39
    Pretrial, the court stated:
    In this case, I think that it would be unfair and misleading to imply or allow
    inferences that these four plaintiffs were the only ones allegedly affected by
    75
    No. 83287-5-I/76
    [A] lay person is perfectly capable of saying, “Gee, you know, 30
    percent of the people aren’t at work this week,” you know, that’s
    just an observation any of us can make. They don’t know anything -
    - they don’t know what it is, they can’t testify as to causation or
    what they think it is or anything like that, but just that people are
    out, that they are ill is -- you know, they can testify to that.
    At trial, the testimony from the lay witnesses was limited to precisely this type of
    observation. One teacher testified, “a few years into our time at the building, I
    noticed there were quite a few students and teachers who had become ill.”
    Another testified that she noticed more and more teachers and students
    becoming ill: “maybe 2014 I heard a little bit, but 2015 I heard more. And then of
    course by 2016, there was a lot.” Other teachers agreed that they noticed a
    change in the rate of illness at the school. Another witness noted that more
    people were getting sick in the new building than in the old location. Another
    teacher stated that she had previous exposure to mold in other buildings, but the
    symptoms she experienced in SVEC were not similar to mold reactions. All of
    this testimony reflected personal observations, without any opinion about the
    possible reason people in the building were becoming ill.
    Moreover, as Pharmacia acknowledges, the teachers’ testimony was not
    the only source of this information. Erickson admitted an Indoor Air Quality
    Assessment Report that stated carpet had been removed “[d]ue to reported
    health complaints teachers expressed concerns about potential contaminants in
    PCBs. It also rebuts Monsanto’s claim that other workers did not get sick,
    although they were exposed to higher levels. That is a fairness issue as well.
    76
    No. 83287-5-I/77
    the carpets. The teachers were concerned about PCB and pesticides
    contaminants.” The jury heard testimony that the company was hired to perform
    the assessment “prompted primarily by a number of indoor air quality complaints
    by teachers and others in the school right around that time frame.” The jury saw
    an example of one of these indoor air quality reports, filed by Marquardt with
    Monroe Public Schools, noting many symptoms arising during the school year.
    The jury heard the same information through other testimony and
    evidence to which Pharmacia failed to object; thus, the evidence was not
    prejudicial. Additionally, the evidence was probative of the three teachers’
    experiences at SVEC. The trial court did not abuse its discretion in allowing the
    lay witness nonparty observations limited to other people’s illness.
    CONCLUSION
    We conclude that the trial court erred in applying Missouri law, which lacks
    a statute of repose for product liability claims, rather than WPLA’s statute of
    repose, to plaintiffs’ WPLA claims and reverse on that basis. We also hold that
    WPLA’s statute of repose does not violate the Washington Constitution’s
    privileges and immunities clause, article I, section 12. We remand for further
    proceedings consistent with this opinion. 40
    Reverse and remand.
    40 We do not reach Pharmacia’s claims regarding whether substantial evidence
    supported the judgment or for remittitur of the punitive damages award.
    77
    No. 83287-5-I/78
    WE CONCUR:
    78
    Erickson v. Pharmacia LLC, No. 83287-5-I
    DWYER, J. (concurring and dissenting) — An experienced, proven trial
    judge assumed responsibility for the conduct of a sprawling, highly contentious
    civil lawsuit. During the pretrial, jury trial, and posttrial stages of the litigation, the
    judge made dozens, perhaps hundreds, of discretionary rulings and decided
    dozens of questions of law. The jury returned a massive verdict in favor of the
    plaintiffs. After this verdict was reduced to judgment, a highly litigated appeal
    commenced. The majority opinion, filed today, resolves the issues raised.
    I agree with the majority on the two primary legal questions presented and
    on its decision as to several contested trial court evidentiary rulings. However, I
    part ways with the majority on other evidentiary questions. I think the trial court
    got these right. In the remainder of this separate opinion, I more fully explain my
    analysis.
    I
    As to a primary question on appeal—whether the statutory repose period
    in the Washington product liability act 1 (WPLA) applies to the Washington
    product liability act cause of action pleaded in this cause—I agree with the
    majority that it does. There are several reasons for this.
    1 Ch. 7.72 RCW.
    No. 83287-5-I/2
    1. The statutory repose provision was contained in the bill, passed by the
    legislature, that became codified as the WPLA. In short, the provision is an
    integral part of the act.
    2. In this way, the WPLA statutory repose period differs from other
    statutes of repose—in particular the builder’s and contractor’s statute of repose
    set forth in RCW 4.16.310—in that it is not an “overlay” on to a variety of causes
    of action. Instead, the WPLA statutory repose period applies only to WPLA
    claims.
    3. The WPLA repose provision differs greatly from most other such
    provisions in effect in 1981 when it was adopted. It provides for a presumption of
    a 12-year useful life period, which may be rebutted by a plaintiff who proves a
    longer useful life. Unlike RCW 4.16.310, which provides a 6-year period
    applicable to all cases, the WPLA provision was plainly the result of legislative
    give-and-take: a compromise.
    4. The WPLA establishes that a WPLA cause of action cannot arise
    unless it arises within the repose period. Thus, the statutory repose provision is
    part of the very definition of whether a cause of action is extant. Compliance with
    its terms is part-and-parcel of establishing that the cause of action exists.
    Courts have a duty to honor legislative public policy decisions and to give
    effect to legislative determinations. The majority does so by its resolution of this
    question.
    2
    No. 83287-5-I/3
    II
    As to another primary question, controlling appellate authority establishes
    that punitive damages can be awardable as part of a WPLA claim as the result of
    a choice of law analysis. However, Washington is not a punitive damages state.
    Accordingly, in this case, the punitive damages sought can be recoverable so
    long as they are limited to only those transgressions as to which the Missouri
    legislature has evidenced a public policy need for the availability of such
    damages.
    However, where Missouri substantive law does not prohibit a certain
    action, the Missouri legislature has necessarily not evidenced such a public
    policy need.
    Here the WPLA allowed for recovery on a greater number of tort theories
    than does Missouri law. Only those theories of recovery that offend both
    Washington and Missouri law can support a punitive damages award. But the
    jury was not so instructed and its verdict was not so restricted. I agree with the
    majority that, in this circumstance, reversal is required.
    III
    The majority resolves several challenged evidentiary rulings entered by
    the trial judge. As to those not addressed herein, I agree with the majority’s
    decisions.
    3
    No. 83287-5-I/4
    IV
    The majority today reverses the trial court’s decisions on two discretionary
    evidentiary rulings. For the reasons set forth in the remainder of this opinion, I
    dissent from these holdings.
    The trial court performs an indispensable gatekeeping function in
    determining the admissibility or inadmissibility of evidence. See, e.g., Anderson
    v. Akzo Nobel Coatings, Inc., 
    172 Wn.2d 593
    , 600, 
    260 P.3d 857
     (2011). By
    reviewing admissibility rulings for an abuse of discretion, we recognize that the
    trial court—not the appellate court—is particularly suited for making such
    determinations. 2 When, as here, the trial court has presided over exceptionally
    complex litigation involving innumerable evidentiary rulings, the case for our
    deference is especially salient. An appellate court will never have the full
    evidentiary picture possessed by the trial court, as we are seldom presented with
    the evidence that was excluded by that court to ensure fairness to the parties.
    Thus, when we do not afford proper deference to the trial court’s evidentiary
    rulings, we risk upsetting the delicate balance sought by that court in resolving
    questions of admissibility.
    Moreover, it is the role of neither the trial court, in determining the
    admissibility of evidence, nor the appellate court, in reviewing such discretionary
    rulings, to usurp the jury’s critical fact-finding function. Accordingly, pursuant to
    Frye v. United States, 
    54 App. D.C. 46
    , 
    293 F. 1013
     (1923), only when the
    2 We review de novo a trial court’s exclusion of evidence pursuant to Frye v. United
    States, 
    54 App. D.C. 46
    , 
    293 F. 1013
     (1923). Lakey v. Puget Sound Energy, Inc., 
    176 Wn.2d 909
    , 919, 
    296 P.3d 860
     (2013). We review for an abuse of discretion a trial court’s decision
    regarding the admissibility of expert evidence. Lakey, 
    176 Wn.2d at 919
    .
    4
    No. 83287-5-I/5
    scientific theory or methodology underlying expert testimony is not generally
    accepted in the relevant scientific community is that evidence excluded. See,
    e.g., Anderson, 
    172 Wn.2d at 603
    . Otherwise, so long as the evidence is offered
    by a qualified witness and is helpful to the jury, the question is one of weight, not
    of admissibility. See ER 702. Such questions are properly presented to the jury
    for resolution.
    Here, the majority reverses the trial court’s rulings admitting into evidence
    the expert opinion of industrial hygienist Kevin Coghlan, M.S., C.I.H., who
    testified at trial that the plaintiffs were exposed to unsafe levels of polychlorinated
    biphenyls (PCBs) while teaching at the Sky Valley Education Center (Sky Valley).
    Coghlan’s testimony was premised on his application of three methods used to
    assess the plaintiffs’ PCB exposure levels. In admitting his testimony, the trial
    court determined that Monsanto’s objections to the evidence “are not to the
    scientific theory but to how Mr. Coghlan applied that theory in this particular
    instance.” Such objections, the court ruled, are pertinent to the credibility of the
    expert’s opinion, not to the admissibility of the testimony. In my view, the trial
    court got this right.
    Among the approaches employed by Coghlan to assess the plaintiffs’ PCB
    exposure levels is the so-called “carpet-based methodology.” 3 In applying this
    approach, Coghlan measured the PCB levels in carpet samples obtained from
    Sky Valley in order to reconstruct the historical levels of PCBs in the air. To do
    so, he relied on a study conducted by the Environmental Protection Agency
    3 See Br. of Appellant at 72.
    5
    No. 83287-5-I/6
    (EPA) that modeled how PCBs in the air are deposited onto various building
    materials. Coghlan, however, was calculating estimated PCB air levels based on
    the PCB levels in the carpet samples, while the EPA study analyzed how PCBs
    transferred from the air onto various materials, including carpet. Accordingly, as
    Monsanto recognizes, Coghlan’s approach relied on a mathematical principle
    that involved, “essentially, the rearrangement of a simple algebraic equation”
    obtained from the EPA study. 4 However, according to Monsanto, this “back-
    calculation” is a novel scientific methodology that, pursuant to Frye, should have
    compelled the exclusion of Coghlan’s testimony.
    The trial court disagreed:
    Mr. Coghlan’s estimate of PCB air contamination based on a
    carpet sample containing PCBs was admissible. Mr. Coghlan
    explained at trial that these calculations were based on established
    EPA studies and scientific formulas. He testified that industrial
    hygienists recognize, based on source-sink dynamics, that it is
    possible to use adsorbent material containing PCBs to estimate
    ranges of probable PCB air concentrations in the environment [from
    which] the material was taken. In a controlled test, the EPA
    exposed multiple building materials, including two types of carpet,
    to concentrations of PCBs in the air and determined a partition
    coefficient, which is a ratio of how much PCBs are in the air versus
    how much are in the material—in this case, the carpet. Mr.
    Coghlan used that established formula to determine an estimated
    air concentration of PCBs from the known quantity of PCBs in the
    carpet. This approach is analogous to the method used to estimate
    PCB levels from air samples, which also involves measuring the
    levels of PCBs in an adsorbent material after exposing it to air.
    Thus, the scientific principles behind Mr. Coghlan’s analysis were
    well-established, and his methodology satisfies Frye.
    [Monsanto’s] objections are not to the scientific theory but to
    how Mr. Coghlan applied that theory is this particular instance.
    However, “[a] Frye objection is not appropriate and should be
    overruled if the only objection is the manner in which the theory and
    4 Br. of Appellant at 78.
    6
    No. 83287-5-I/7
    method of analysis were applied in the present case.” Tegland, 5D
    Wash. Prac., § 702:11. “The issue is whether the expert’s
    methodology is generally accepted as being capable of producing
    an accurate result, not whether the expert employed the
    methodology correctly.” Tegland, 5B Wash. Prac., Evid. Law and
    Prac. § 702.19 (6th ed. 2020). “An objection that an expert
    employed the methodology in an improper or unscientific manner
    goes only to the credibility of the expert’s opinion, not the
    admissibility of the expert’s testimony.” Id. So it is here, and Mr.
    Coghlan’s testimony was properly admitted.
    I agree.
    Monsanto, assigning error to the trial court’s ruling, contends that “[t]here
    is no ‘established formula’ that works in reverse, thus rendering Coghlan’s
    calculation novel, not generally accepted, unreliable, and violative of Frye and
    ER 702.” 5 For instance, Monsanto contends that Coghlan’s testimony was
    inadmissible because the EPA study on which he relied was conducted in a
    controlled environment, “which cannot be replicated in the real world.” 6 Thus,
    Monsanto argues, the source of the PCBs in the carpet samples could have been
    due to “tracking” rather than unsafe levels in the air. 7 Monsanto contends that,
    because Coghlan was not able to implement the same controls as those in the
    EPA study, his opinions were rendered “mere speculation.”8
    However, each of these critiques, as the trial court concluded, are
    premised not on the underlying theory or methodology employed, but on
    Coghlan’s application of that methodology. The scientific theory underlying
    Coghlan’s testimony is that of “source-sink dynamics,” which investigates “the
    5 Br. of Appellant at 76.
    6 Br. of Appellant at 77.
    7 Br. of Appellant at 81.
    8 Br. of Appellant at 82.
    7
    No. 83287-5-I/8
    mechanisms by which PCBs may transfer or migrate from primary sources to
    other building materials or ‘sinks.’” As Monsanto indicated, Coghlan’s
    methodology consisted of the “rearrangement of a simple algebraic equation”
    obtained from a peer-reviewed EPA study. 9 Monsanto nowhere asserts that
    either the theory of “source-sink dynamics” or Coghlan’s use of the EPA study’s
    mathematical formula are novel. Rather, Monsanto argues that Coghlan
    incorrectly applied the generally accepted science. “While Frye governs the
    admissibility of novel scientific testimony, the application of accepted techniques
    to reach novel conclusions does not raise Frye concerns.” Lakey v. Puget Sound
    Energy, Inc., 
    176 Wn.2d 909
    , 919, 
    296 P.3d 860
     (2013). So long as the
    evidence is reliable, and, thus, admissible pursuant to ER 702, “an expert’s
    errors in applying proper procedures go to the weight, not the admissibility, of the
    evidence.” Lakey, 
    176 Wn.2d at 920
    .
    Monsanto also takes issue with Coghlan’s so-called “invent[ion of] two
    estimation methods” in employing a study of PCB levels in New York schools. 10
    According to Monsanto, Coghlan’s use of the New York schools study was
    “novel” because those schools might not be comparable to Sky Valley.
    Moreover, Monsanto asserts, Coghlan employed a “novel methodology” by using
    two data points to estimate PCB levels, rather than using an average of all of the
    available datasets. 11 Again, Monsanto’s assertions are not that the theory or
    methodology employed constitute novel science; rather, Monsanto’s critiques are
    9 Br. of Appellant at 78.
    10 Br. of Appellant at 87-88.
    11 Br. of Appellant at 92.
    8
    No. 83287-5-I/9
    to Coghlan’s application of that science. Again, this goes to the weight, rather
    than the admissibility, of the evidence. Lakey, 
    176 Wn.2d at 920
    . In my view,
    the trial court correctly concluded that Frye is not implicated. See Anderson, 
    172 Wn.2d at 611
     (“The Frye test is implicated only where the opinion offered is
    based upon novel science.”).
    Nor, in my view, did the trial court abuse its considerable discretion by
    rejecting Monsanto’s contention that Coghlan’s testimony was unreliable and,
    thus, inadmissible pursuant to ER 702. See Lakey, 
    176 Wn.2d at 918-19
     (“Frye
    excludes testimony based on novel scientific methodology until a scientific
    consensus decides the methodology is reliable; ER 702 excludes testimony
    where the expert fails to adhere to that reliable methodology.”). Contrary to
    Monsanto’s assertion, Coghlan’s testimony was not rendered “mere speculation”
    because the source of the PCBs in the carpet samples could not be determined
    with complete certainty. 12 Nor was Coghlan’s opinion unreliable because he was
    unable to implement the same controlled conditions implemented in an EPA
    laboratory experiment. If this were so, it is unlikely that any scientific expert
    testimony could be properly presented to a jury. However, pursuant to ER 702,
    expert testimony is generally admissible “if it will be helpful to the jury in
    understanding matters outside the competence of ordinary lay persons.”
    Anderson, 
    172 Wn.2d at 600
    . Because the trial court did not act in a manifestly
    unreasonable manner in admitting Coghlan’s testimony, I would not disturb this
    discretionary ruling on appeal.
    12 Br. of Appellant at 82.
    9
    No. 83287-5-I/10
    Monsanto additionally asserts that the trial court erred by “invok[ing] a lack
    of evidence as a reason to permit unreliable evidence that is not generally
    accepted within the scientific community.” 13 This the trial court did not do.
    Rather, as its ruling demonstrates, the court determined that Monsanto’s
    objections regarding imperfections in the proffered evidence go to the weight, not
    the admissibility, of that evidence. The court explained that
    obviously determining [PCB] concentrations from carpet samples is
    not the ideal way to do it. The problem is, . . . of course, we don’t
    have any valid air samples, because although . . . there were air
    samples taken prior to remediation, there were all kinds of reasons
    to be doubtful about the validity of those air samples . . . .
    So . . . we’re dealing with a situation [in] which we don’t have
    the ideal kind of evidence you’d like to have, and while I think that
    there’s certainly lots of things that the defense can point to as
    reasons to be concerned about the result that Mr. Coghlan arrived
    at, I do think it’s a valid test . . . .
    It’s an application of a test that the EPA does. Yes, it’s done
    a bit differently because of the fact that he’s trying to sort of
    reconstruct events that he doesn’t have control over, but I think that
    goes to the weight that the jury should give to this rather than its
    admissibility.
    Where, as here, proffered expert testimony has been properly admitted by
    the trial court, that evidence may be further “tested by the adversarial process
    within the crucible of cross-examination.” Anderson, 
    172 Wn.2d at 607
    . Such
    was the case here. The jury was properly presented with Coghlan’s testimony.
    Monsanto was then “permitted to present other challenging evidence.”
    Anderson, 
    172 Wn.2d at 607
    . By ruling in the plaintiffs’ favor, the jury
    demonstrated that it afforded weight to that testimony.
    13 Br. of Appellant at 96.
    10
    No. 83287-5-I/11
    Pertinent, too, to the questions presented herein, is the purpose of the
    Frye test and the functional difference between the courtroom and the laboratory.
    As our Supreme Court has astutely recognized:
    Frye envisioned an evolutionary process with novel scientific
    techniques passing through an “experimental” stage during which
    they would be scrutinized by the scientific community until they
    arrive at a “demonstrable” stage. Frye, 54 App. D.C. at 47.
    However, science never stops evolving and the process is
    unending. Each scientific inquiry becomes more detailed and
    nuanced. As one commentator has noted, there is a “difference
    between the quest for truth in the courtroom and in the laboratory.
    Law must resolve disputes finally and quickly, whereas science
    may consider a multitude of hypotheses indefinitely.” Loevinger,
    [Science as Evidence, 35 JURIMETRICS J.] 153, 177 [(1995)].
    Anderson, 
    172 Wn.2d at 607
    .
    In my view, no error occurred here. Accordingly, I dissent from the
    majority’s determination that the trial court erroneously admitted Coghlan’s
    testimony. 14
    V
    While this case was pending, well after oral argument was held, our
    Supreme Court filed its decision in Bennett v. United States, 2 Wn.3d 430, 
    539 P.3d 361
     (2023). Almost immediately, respondents requested an opportunity to
    brief the impact, if any, of that decision on the issues in this case. The panel
    correctly authorized the parties to submit supplemental briefing on the effect, if
    any, of the Bennett decision on this case.
    14 The majority additionally concludes that “to the extent Dr. Dahlgren’s exposure-based
    causation opinions rely on Coghlan’s opinions about PCB levels at [Sky Valley] and selectively
    considered the results of PCB testing in determining causation, his opinions should have been
    excluded.” Majority at 57. Because I believe that Coghlan’s testimony was properly admitted, I
    additionally conclude that Dr. Dahlgren’s opinion should not be excluded on this basis.
    11
    No. 83287-5-I/12
    I agree with the majority’s conclusion that Bennett does not mandate
    invalidation of the WPLA statute of repose.
    In the course of the supplemental briefing, the question of the severability
    of the WPLA repose provision was raised and discussed. Given the basis for the
    majority’s decision, it was unnecessary for that opinion to weigh in on the matter.
    I concede that it is similarly unnecessary for me to arrive at a definitive
    answer in this opinion. Indeed, given the lack of necessity, I have not
    endeavored to do all of the legislative history detective work necessary to reach a
    sound conclusion. Instead, I discuss this issue today so as to forewarn future
    litigants of a possible conclusion that may not seem immediately apparent.
    The existence of a severability clause is not dispositive on the question of
    severability. Instead, a severability clause is best thought of as an indication of
    legislative intent that may be rebutted. See Mt. Hood Beverage Co. v.
    Constellation Brands, Inc., 
    149 Wn.2d 98
    , 118, 
    63 P.3d 779
     (2003). As has been
    explained,
    the various provisions of a legislative enactment are not severable
    if
    the constitutional and unconstitutional provisions are so
    connected—that it could not be believed that the
    legislature would have passed one without the other; or
    where the part eliminated is so intimately connected with
    the balance of the act as to make it useless to
    accomplish the purposes of the legislature.
    Leonard v. City of Spokane, 
    127 Wn.2d 194
    , 201, 
    897 P.2d 358
     (1995) (quoting
    Hall v. Niemer, 
    97 Wn.2d 574
    , 582, 
    649 P.2d 98
     (1982)).
    12
    No. 83287-5-I/13
    Stated somewhat differently, a court may “not sever an unconstitutional
    clause . . . if to do so would broaden the statute’s application, because we cannot
    presume the legislature meant it to be applied to persons it specifically excluded.”
    Mt. Hood Beverage Co., 
    149 Wn.2d at 118
    . Thus, if “‘by striking out the proviso
    the remainder of the statute would have a broader scope either as to subject or
    territory, then the whole act is invalid.’” Mt. Hood Beverage Co., 
    149 Wn.2d at 118
     (quoting 16A AM.JUR.2d Constitutional Law § 218, at 109 (1998)).
    The concern is always the legislature’s intent. Thus, we fairly recently
    explained why severability was not available therein: “striking only those discrete
    provisions challenged by Pasado’s would bring about a result that our legislature
    ‘never contemplated nor intended to accomplish.’” Pasado’s Safe Haven v.
    State, 
    162 Wn. App. 746
    , 755, 
    259 P.3d 280
     (2011) (quoting Jensen v.
    Henneford, 
    185 Wash. 209
    , 223, 
    53 P.2d 607
     (1936)).
    Two simple truths lead to competing assumptions about whether resort to
    legislative history will prove fruitful in a case in which legislative intent is
    paramount in deciding this severability dispute. First, 1981 (the year of
    enactment) was a most unusual year in Washington legislative politics. Second,
    Washington’s legislature during that era had a notoriously incomplete and spotty
    approach to memorializing its legislative history. So it may be that even the most
    diligent legislative history detective will be able to unearth suitable
    memorialization of the relevant legislative history to prove a case.
    Nationally, the November 1980 general election was historically
    significant. Democrat Jimmy Carter was defeated in his reelection bid by
    13
    No. 83287-5-I/14
    Republican Ronald Reagan. Similarly, Republican election success resulted in
    the GOP gaining the majority in the United States Senate.
    Similar results obtained in Washington. Democratic Governor Dixy Lee
    Ray failed in her reelection bid and was replaced by Republican John Spellman.
    The GOP won a majority in the State House of Representatives. Then, three
    weeks into the 1981 legislative session, Democratic Senator Peter von
    Reichbauer switched his party allegiance to the GOP. This gave the
    Republicans a 25-24 numerical majority in the Senate. For the first time in more
    than two decades the Republicans controlled both houses of the legislature. And
    for the first time in a generation, the GOP dominated both houses of the
    legislature and the governor’s mansion.
    It was in the midst of all of this that the 1981 legislature passed the WPLA.
    In section I of this opinion, I set forth the several iterations of how the
    WPLA was a product of legislative give-and-take. It is unique in several ways
    that manifest this. Of primary interest to the business community—in large part
    the base supporters of the new legislative Republican majority—was the repose
    provision. It defies belief to assume that the GOP would have allowed the WPLA
    to become law in the absence of such a provision.
    Yet, I concede the spotty nature of the memorialization of Washington
    legislative history at this time. Thus, although I am confident of what the real
    answer is, I am uncertain if proof of that answer is presently ascertainable.
    So, in the end, this portion of this separate opinion is more of an
    admonition or an alert. Beware to any litigant who seeks to make the answer to
    14
    No. 83287-5-I/15
    the question herein posed a dispositive question. Before striding down that dark
    alley of legislative history, take heed of that which may lurk around the corner.
    15
    

Document Info

Docket Number: 83287-5

Filed Date: 5/1/2024

Precedential Status: Precedential

Modified Date: 5/1/2024