COURT FILE NO.: CV-14-00518698-00CP

DATE: 20221220

SUPERIOR COURT OF JUSTICE – ONTARIO

RE: JENNIFER PRICE, INDIVIDUALLY, and MATTHEW JANZIC, A CHILD BY HIS NATURAL MOTHER and LITIGATION GUARDIAN, JENNIFER PRICE, Plaintiffs

AND:

H. LUNDBECK A/S and LUNDBECK CANADA INC., Defendants

BEFORE: Justice Glustein

COUNSEL: Paul Bates and Anthony Tibbs, for the plaintiffs

Frank J. McLaughlin, Brandon Kain, and Dorothy Charach, for the defendants

HEARD: November 15, 16, and 17, 2022

reasons for decision

NATURE OF THE MOTION AND OVERVIEW

Nature of the motion

[1] The plaintiffs, Jennifer Price (“Price”) and her son Matthew Janzic,[^1] bring a motion for an order certifying this action as a class proceeding pursuant to s. 5 of the Class Proceedings Act, 1992, S.O. 1992, c. 6 (the “CPA”).[^2]

[2] For the reasons that follow, I dismiss the motion for certification. I find that the plaintiffs have failed to satisfy the common issue and preferable procedure requirements under ss. 5(1)(c) and (d).

Overview

[3] The plaintiffs allege that the anti-depressant drug Celexa® is a teratogen that causes congenital malformations through its active ingredient, citalopram.

[4] A teratogen is any agent that can, under reasonable circumstances of exposure, disturb the development of an embryo or fetus and thereby cause congenital malformations. Congenital malformations[^3] are structural or functional anomalies that occur in a fetus in utero, which can in turn either cause the child to be born with one or more birth defects, or cause the pregnancy to spontaneously abort and miscarry.

[5] The plaintiffs bring the action against the defendants H. Lundbeck A/S and Lundbeck Canada Inc., who they allege manufactured, distributed, and marketed Celexa®.

[6] The plaintiffs claim that the defendants (i) knew or ought to have known that Celexa® “is or may be a Teratogen” and (ii) “breached a duty to warn Canadian physicians and patients that Celexa® is or may be teratogenic”.

[7] The proposed class is “[a]ll women who were prescribed and ingested Celexa® while pregnant in Canada and subsequently aborted, delivered, or miscarried children with Congenital Malformations, such children born to such women, and family members who may claim under Family Compensation Legislation”.

[8] The plaintiffs do not seek to certify a general causation[^4] or duty to warn issue linking Celexa® to any particular congenital malformation amongst the hundreds of different congenital malformations that can either arise upon birth or cause a spontaneous abortion.[^5]

[9] Instead, the proposed common issue in the present case is: “From 1999, did the Defendants breach their duty to warn Canadian physicians and patients that Celexa® is or may be teratogenic?” (the “PCI”).

[10] As the plaintiffs state in their factum, the PCI “is really a two-step inquiry: (1) is or may Celexa® (citalopram) be teratogenic?; and if so, (2) did the Defendants breach their duty to warn of that teratogenicity?”[^6] Consequently, there are two proposed common issues which must be certified for the class action to proceed:

(i) the “general causation” proposed common issue of whether Celexa® is or may be teratogenic,[^7] and

(ii) if so, the “duty to warn” proposed common issue of whether the defendants breached their duty to warn Canadian physicians and patients of that teratogenicity.[^8]

[11] The defendants object to the certification of the proposed class action. With respect to the General Causation PCI, they submit that certification should be denied under s. 5(1)(c) because:

(i) The plaintiffs failed to establish a workable methodology to determine whether Celexa® is teratogenic on a class-wide basis.

(ii) Even if there is a workable methodology to determine whether Celexa® is or may be teratogenic on a class-wide basis, such a finding would not advance the litigation, since each class member would still be required to establish general causation based on the specific congenital malformation at issue for each class member and, as such, the “general” teratogenicity of Celexa® would be irrelevant to any claim of a class member.

[12] With respect to the Duty to Warn PCI, the defendants submit that certification should be denied under s. 5(1)(c) because:

(i) If the General Causation PCI cannot be certified as a common issue under s. 5(1)(c), then the Duty to Warn PCI cannot be certified under s. 5(1)(c), since like the General Causation PCI, it leaves each class member to establish general causation for each specific congenital malformation at issue.

(ii) Even if the General Causation PCI could be certified under s. 5(1)(c), there is no basis in fact that the Duty to Warn PCI could meet the s. 5(1)(c) requirement for a common issue, since the duty to warn must be in relation to a specific risk, and not to a risk of “teratogenicity” in general (as proposed by the plaintiffs).

[13] The defendants also submit that even if both the General Causation PCI and the Duty to Warn PCI could be certified as common issues under s. 5(1)(c), the plaintiffs have failed to establish a basis in fact that the proposed class proceeding would be the preferable procedure under s. 5(1)(d). The defendants submit that in this case the need for numerous individual issues “undercuts the goal of judicial economy and could overwhelm the resolution of the common issues, producing an inefficient and unmanageable class proceeding”: Bayens v. Kinross Gold Corporation, 2014 ONCA 901, 327 O.A.C. 156, at para. 129, leave to appeal discontinued, [2015] S.C.C.A. No. 59.

[14] For the reasons set out below, I dismiss this motion for certification. In brief:

(i) I disagree with the defendants’ submission that there is no workable methodology to determine, on a class-wide basis, whether Celexa® is or may be a teratogen.

(ii) However, I agree with the defendants’ remaining submissions and find that (a) the plaintiffs have not established some basis in fact that either the General Causation PCI or Duty to Warn PCI can be certified as a common issue under s. 5(1)(c) and (b) even if the PCI (through its two parts) could be certified under s. 5(1)(c), the plaintiffs have not established some basis in fact that a class action would be the preferable procedure under s. 5(1)(d).[^9]

FACTS

Congenital malformations

[15] All pregnancies have a baseline risk of congenital malformations. Approximately 3-5% of newborns have a “major” malformation, i.e. a malformation that interferes with normal functions and can even threaten life. Approximately 10% of newborns have a “minor” congenital malformation, i.e. a malformation that may be important from a cosmetic standpoint, but does not interfere with life functions.

[16] The vast majority of congenital malformations arise from a single or multiple gene absences or abnormalities, or are idiopathic.[^10] Only a very small percentage of malformations, estimated at 1-3%, are specifically due to known environmental exposures. Many of these include non-chemical factors (e.g., radiation therapy, maternal health conditions such as diabetes and phenylketonuria, intrauterine infections such as chicken pox or rubella, periods of hypotension, excessive heat, maternal surgery, intracytoplasmic sperm injection, or amniocentesis).

[17] Hundreds of different congenital malformations exist.[^11] They can arise in any organ or part of an organ in the developing embryo if it is disrupted or altered during embryological development.

[18] There is no common etiology for these defects. Etiology is “the cause, set of causes, or manner of causation of a disease or condition”: Michael Proffitt, Philip Durkin & Edmund Weiner, eds, Oxford English Dictionary (London: Oxford University Press, 2022) sub verbo “etiology”. Each of the hundreds of congenital malformations has its own unique etiology.

[19] Each congenital malformation also has its own unique epidemiological[^12] data, its own set of risk factors, and its own criteria for screening and diagnosis.

[20] There are no known teratogens capable of causing all congenital malformations.

Teratogens

[21] Several chemical substances are also known to cause congenital malformations (e.g., cigarette smoke and alcohol). At present, there are very few confirmed pharmaceutical teratogens. An example is thalidomide, which is associated with phocomelia, a specific limb malformation, as well as visceral defects affecting the heart, urinary tract, gastrointestinal tract, and genital tract. Some anti-epileptic drugs and isotretinoin (an acne medication) have also been associated with teratogenicity.

Celexa®

[22] Citalopram is a selective serotonin reuptake inhibitor (“SSRI”) used to treat depression. It was allegedly produced, marketed and sold by the defendants[^13] under the brand name Celexa®.[^14] It was introduced in Canada in 1999.

The treatment of depression in women

[23] Depression is a common mental disorder associated with numerous negative health consequences. It is the leading cause of disability for women aged 15-44. This includes pregnant women, approximately 35% of whom have depressive symptoms, and at least 10% of whom have been diagnosed with depression. Depression may also worsen during pregnancy, as Price testified happened to her twice.

[24] Depression is associated with an increased risk of congenital malformations. This in turn increases the risks to the mother of emotional suffering, difficulty performing usual activities, failure to seek prenatal care, inadequate diet, poor weight gain, the use of tobacco, alcohol and other harmful substances, poor quality of life, postpartum depression, and the potential for family conflict, self-harm or suicide.

[25] Depression can cause risk to the fetus or child by increasing the potential for miscarriage, low birth weight, preterm birth, poor obstetrical outcomes, neonatal intensive care unit admissions, impairment in mother-infant bonding, infant sleep difficulties, cognitive, behavioural and emotional problems, and developmental delay.

[26] Depression is identified through diagnostic criteria established by the World Health Organization and American Psychiatric Association. It affects women to varying degrees, and is classified into three severity types:

(i) mild depression: the symptoms are disturbing but manageable, and result in minor impairment in social or occupational functioning,

(ii) moderate depression: the symptoms’ intensity and resulting impairment range between the mild and severe categories, and

(iii) severe depression: the symptoms can include suicidal or psychotic thoughts, with greater intensity when accompanied by certain comorbidities such as anxiety. Severe depression is seriously disturbing and unmanageable, and markedly interferes with social and occupational functioning.

[27] Depression may become more severe or resistant to treatment over time, and the risk of self-harm or suicide is a vital consideration.

[28] Depression is a common and treatable mental disorder, and several therapies for it exist. Potential options include:

(i) evidence-based psychotherapy (cognitive behavioural therapy or interpersonal therapy),

(ii) electroconvulsive therapy,

(iii) antidepressant medications, including:

(a) SSRIs,

(b) serotonin norepinephrine reuptake inhibitors,

(c) monoamine oxidase inhibitors (“MAOIs”), and

(d) tricyclic antidepressants.

[29] During pregnancy, these options are reduced. MAOIs are inappropriate for pregnant women, and there are practical limitations which impede access to evidence-based psychotherapies since they are time-intensive, costly, and not easily accessible. As for electroconvulsive therapy, it is reserved for severe, treatment-resistant depression (including during pregnancy), or depression associated with psychotic symptoms or risk of suicide. Most North American women refuse to consent to it.

[30] Given these factors, many patients choose antidepressant medications. Approximately 8% of U.S. women take non-MAOI antidepressants during pregnancy. Antidepressants are prescribed to pregnant women with severe depression and, in appropriate circumstances, to those with mild or moderate depression. Antidepressants may help patients decrease harmful lifestyle behaviours such as smoking, drinking alcohol, or taking harmful drugs, while making them more likely to sleep better, resume eating and attend prenatal care.

Expert evidence

[31] The motion record is replete with evidence from experts on behalf of both parties, addressing the issues of the alleged teratogenicity of Celexa®, the duty to warn, and the methodology to determine the General Causation PCI.

[32] The plaintiffs retained four medical experts:

(i) Dr. Anick Bérard, PhD, an epidemiologist,

(ii) Dr. David Healy, MD, a psychiatrist and psychopharmacologist,

(iii) Dr. Robert M. Cabrera, PhD, a teratologist and laboratory scientist, and

(iv) Dr. Derelle Mangin, MBChB, a practicing family physician in Canada.[^15]

[33] The defendants retained three medical experts:

(i) Dr. Michael Brunskill Bracken, PhD, an epidemiologist,

(ii) Dr. Jeffrey Brent, MD, PhD, a medical toxicologist, and

(iii) Dr. Donna Stewart, MD, D. Psychiatry, O.C., a licensed medical doctor in Ontario and a certified specialist in Psychiatry at the Royal College of Physicians and Surgeons of Canada.

[34] In brief, the plaintiffs’ expert evidence on the alleged teratogenicity of citalopram includes the following:

(i) Dr. Healy’s view is that SSRIs (including citalopram) are teratogens.

(ii) Dr. Bérard’s opinion is that there is peer-reviewed scientific evidence demonstrating a causal association between exposure to citalopram and congenital malformations, particularly cardiac and neural tube defects.

(iii) Dr. Cabrera’s opinion is that, based on the animal and human evidence, and applying the Bradford Hill criteria,[^16] citalopram is a teratogen.

(iv) The Canadian Pediatric Society Guideline says that “SSRI use in pregnancy has been linked to congenital malformations” and there is “some evidence” that the risk of spontaneous abortion is increased from SSRI exposure.

(v) Dr. Cabrera’s evidence is that studies have shown that vasoconstriction produces a variety of different congenital malformations, which are “all a result of modifying serotonin”.

(vi) The evidence of Dr. Healy and Dr. Bérard is that serotonin is essential for fetal development and changes in serotonin can cause congenital malformation.

[35] Based on the above evidence, the plaintiffs’ experts, Dr. Healy and Dr. Mangin, opine that there is sufficient evidence to warrant a warning that Celexa® is or may be teratogenic.

[36] In brief, the defendants’ evidence that citalopram is not teratogenic includes the following:

(i) Numerous clinical guidelines (including the 2016 CANMAT[^17] guidelines) provide that it is appropriate to use citalopram to treat pregnant women suffering from depression.

(ii) No regulatory agency, medical society or teratology organization has declared citalopram unsafe for use in pregnant women.

(iii) To date, no peer-reviewed academic paper has identified citalopram as a teratogen.

(iv) In 2005, both Health Canada and the U.S. Food and Drug Administration (the “FDA”) required that warnings be added to the product monograph of another SSRI, paroxetine (Paxil®), regarding an increased risk of major and cardiovascular congenital malformations. Health Canada also issued a Safety Alert about this in 2005. Neither Health Canada nor the FDA applied the paroxetine findings to citalopram or required the addition of any warnings about the risks of cardiovascular congenital malformations to the Celexa® product monographs.

The Celexa® Product Monograph

[37] Before approving the introduction of a new drug to the market, Health Canada requires the manufacturer to undergo an extensive review process of its new drug submission and the proposed product monograph (“PM”). This process is undertaken by physicians, pharmacologists, and other scientists to determine whether the safety profile of the medicine is acceptable based on the risk-benefit analysis. If the drug is approved, Health Canada continues to conduct post-marketing surveillance of it, including monitoring its safety.[^18]

[38] The PM for Celexa® states that citalopram is indicated for “symptomatic relief of depressive illness”, a broad term in which the range of severity of the condition and how it presents can vary greatly. The indication encompasses:

(i) major depressive disorder,

(ii) persistent depressive disorder, and

(iii) unspecified depressive disorder with the specifiers anxious distress, mixed features, melancholic features, atypical features, mood congruent psychotic features, peripartum onset, seasonal pattern, or premenstrual dysphoric disorder.

[39] Physicians also appropriately prescribe citalopram for the following non-approved or “off-label” uses: (i) anxiety, (ii) post-traumatic stress disorder, (iii) obsessive compulsive disorder, and (iv) eating disorders.

[40] The usual dose range for citalopram is 20-40 mg. The dose depends on the illness type and severity and other individual factors, including the patient’s past treatment, other drugs the patient may be taking and the presence of drug side effects. For instance, treatment of obsessive-compulsive disorder requires higher doses than depression.

[41] Because it is not considered ethical to prospectively test drugs on pregnant women, none of the randomized clinical trials that supported the citalopram new drug submission (nor those for the other SSRIs) were able to test for congenital malformations. For this reason, the first Celexa® PM in Canada (dated January 28, 1999) warned patients that the safety of citalopram in pregnant women had not been established. It included the warning in the Precautions section of the PM, using language that has subsequently been reiterated in Part I of the current PM:

PREGNANCY AND NURSING MOTHERS

The safety of Celexa during pregnancy and lactation has not been established. Therefore Celexa should not be used during pregnancy, unless in the opinion of the physician, the expected benefits to the patient markedly outweigh the possible hazards to the fetus.

[42] Further, the “Adverse Reactions” section of the 1999 PM (now in Part I of the current PM) listed “spontaneous abortion/fetal death” among the events observed during the post-marketing evaluation of Celexa®.

[43] In addition, the “Toxicology” section of the 1999 PM (now in Part II of the current PM) discussed the reproductive studies for citalopram in animals. It noted that teratogenic effects were seen in rats, but only at toxic doses that far exceeded human therapeutic doses, and that no teratogenic effects were seen at the highest dose that could be assessed in rabbits.

[44] Finally, the “Information for the Patient” section of the 1999 PM (now in Part III of the current PM) warned pregnant women not to take Celexa® at all:

What you should tell your doctor before taking Celexa:

If you are pregnant or thinking of becoming pregnant, or if you are breastfeeding.

When not to use Celexa:

You should not be taking Celexa if you are pregnant or breastfeeding. [Emphasis in original text.]

[45] In 2011, further information was added to the PM regarding the animal reproductive studies, including the following statement in the new (combined) “Warnings and Precautions” section in Part I:

In animal reproductive studies citalopram has been shown to have adverse effects on embryo/fetal and postnatal development, including teratogenic effects when administered at doses greater than human therapeutic doses. There are no adequate and well controlled studies in pregnant women: therefore citalopram should be used in pregnancy only if the potential benefit to the patient justifies the potential risk to the fetus.

[46] The “Toxicology” section in Part II of the 2011 PM added that all doses in which these adverse teratogenic effects occurred in animals well exceeded the standard therapeutic human dose for humans in rats, and were not confirmed in a second rat study, nor in rabbit studies.

Dr. Bérard’s evidence on a methodology to determine whether Celexa® is or may be a teratogen (the General Causation PCI)

[47] Since the issue of whether Celexa® is or may be a teratogen requires expert evidence, the plaintiffs led evidence from Dr. Bérard as to the methodology she would apply to determine the issue, which the plaintiffs submit would provide a class-wide answer.

[48] Dr. Bérard did not take the approach that she would examine a particular congenital malformation (e.g. clubfoot) to reach the conclusion that Celexa® is or may be a teratogen based on the teratogenicity related to that particular congenital malformation. Instead, Dr. Bérard relied on a three-step process that did not require analysis of a particular congenital malformation:

(i) The first step is to look at the risk of overall congenital malformations in the population. This can be done by reviewing 122 “citalopram-specific endpoints” from studies, some directly on citalopram and others which indirectly relate to citalopram.

(ii) The second step is to group the overall data by organ system.

(iii) The third step is to extrapolate findings in an organ system group to specific defects within the group if they occur at similar embryological stages.

[49] Consequently, under Dr. Bérard’s approach, a common answer could be provided to the General Causation PCI, such that there would be a class-wide answer at a common issues trial as to whether Celexa® is or may be a teratogen, without the need to study any particular congenital malformation.

[50] On cross-examination, Dr. Bérard was challenged on several aspects of her analysis. In particular, the defendants raised numerous objections, based on their expert evidence and on cross-examination of the plaintiffs’ experts. The defendants submit that:

(i) Dr. Bérard’s use of 11 organ systems to review 122 citalopram-specific endpoints fails to consider the evidence that there is no common etiological factor among all congenital anomalies with many different causes, including prenatal, genetic and environmental, which could be wholly unrelated to the use of citalopram.

(ii) Significant differences among congenital malformations exist even within organ groups so the determination of whether citalopram causes any specific congenital anomaly must be based on citalopram exposures and that specific anomaly.

(iii) There is no established and admitted biological mechanism by which citalopram is alleged to cause congenital malformations. Vasoconstriction (a physiological mechanism) is only one of four possible mechanisms for congenital malformations (the others being neurological, cellular, and metabolic).

(iv) The time period during pregnancy when the consumption of a teratogen can produce a congenital malformation vary depending on the malformation at issue. Even if citalopram were to be documented as a general cause of a specific malformation or condition, extrapolation to an individual case would require proof that the drug exposure occurred at the relevant time period in the pregnancy.

(v) Dr. Bérard’s methodology does not consider that the teratogenic potential of a medicine may vary by genetics, placing some class members at greater risk than others of citalopram’s alleged potential to cause congenital malformations.

[51] In addition to the above challenges as to the accuracy of any conclusions Dr. Bérard might reach under her proposed approach, the defendants rely on the following objections to submit that Dr. Bérard’s approach would not generate a class-wide result on the General Causation PCI, i.e. whether Celexa® is or may be teratogenic:

(i) Even if Dr. Bérard could establish teratogenicity by organ system groupings, it would still produce separate answers for class members whose alleged congenital malformations fall into the different groupings.

(ii) Even for class members who allege congenital malformations that theoretically fall within the same organ grouping, Dr. Bérard’s proposed methodology would not yield answers that are capable of extrapolation between them. A negative answer when examining overall congenital malformations, or when examining the data by organ grouping, could still “mask” a positive, underlying association for a specific congenital malformations.

(iii) Conversely, even if a court applied Dr. Bérard’s methodology to find there is general causation for a particular organ group based on embryological stages of development, that would not determine whether citalopram can cause specific defects falling within that organ group and a specific embryological stage of development.

(iv) Dr. Bérard uses International Classification of Diseases (“ICD”) codes which are used for medical billing purposes, clinical management, and general health indicators. The defendants rely on Dr. Cabrera’s evidence on cross-examination that “medical coding doesn’t necessarily reflect etiology”.

(v) Dr. Bérard could not use the approach she is proposing since the ICD codes are not organized according to embryological stages of development, which is the third step of her analysis.

(vi) Dr. Bérard’s approach is contrary to the text written by Allen Mitchell (“Mitchell”), upon which she relies to support her methodology. Mitchell states that there is a more appropriate approach than creating organ system categories, which is to create categories that reflect embryologic tissue of origin.

[52] In cross-examination, Dr. Bérard did not accept that any of the above criticisms would prevent her from using her methodology to answer the General Causation PCI, which seeks to determine only if Celexa® is or may be teratogenic. The General Causation PCI does not seek to answer whether the specific congenital malformation relevant to each class member is or may be caused by Celexa®. She maintained her position that the three-step process would be sufficient to determine the General Causation PCI.

[53] Further, Dr. Bérard did not accept the criticisms that her methodology was unworkable or could not be supported. She affirmed the use of ICD codes despite their limitations, stating that all researchers use those codes since that is how the data on congenital malformations is organized. She referred to studies and research that consider the embryologic stage of the origin of the cells, consistent with her methodology.

[54] When challenged on the basis that she allegedly failed to consider factors relied upon by Mitchell, Dr. Bérard acknowledged that Mitchell stated that a more appropriate approach was to create categories that reflect embryologic tissue of origin (instead of by organ system). However, Dr. Bérard stated that:

(i) “Allen Mitchell, myself, everyone else, are doing the groupings by organ system because this is how the coding system that we use are put together”; and

(ii) Mitchell himself organized congenital malformation data by organ system when he published his textbook, even though his text stated that a more appropriate approach would be to create categories that reflect embryologic tissues of origin.

Evidence of representative plaintiff and other mothers

[55] Five mothers provided evidence as to the congenital malformations suffered by their children:[^19]

(i) Pamela Bews of Abbotsford, British Columbia swore an affidavit dated July 6, 2017. While pregnant, she was prescribed Celexa® and her daughter was born with disfigured teeth and severe Ebstein’s Anomaly, a serious heart condition.

(ii) Deanna Gardner of Peterborough, Ontario swore an affidavit dated July 6, 2017. While pregnant, she was prescribed Celexa® and her daughter was born with a cleft foot.

(iii) Lori Lovett of Ottawa, Ontario swore an affidavit dated July 5, 2017. While pregnant, she was prescribed Celexa®. An ultrasound revealed that her unborn child could be suffering from Tetralogy of Fallot and esophageal atresia. After a caesarean section birth, the child was born but died 13 hours after delivery.

(iv) Jennifer Price of Stoney Creek, Ontario swore an affidavit dated July 6, 2017. She was cross-examined. While pregnant, she was prescribed Celexa® and her son was born with several severe birth defects including ventricular septal defect, pulmonary valve stenosis, and absent fifth distal phalanx.

(v) Katherine Stymiest of Prince George, British Columbia swore an affidavit dated July 4, 2017. While pregnant, she was prescribed Celexa®. An ultrasound showed that her unborn child had a recessed jaw. The unborn child was subsequently diagnosed to have Treacher Collins Syndrome, a congenital disorder characterized by craniofacial deformities involving the ears, eyes, cheekbones, and jawbone. The child was born with severe micrognathia (undersized lower jaw), no ears, malformed or under-formed cheek bones, a severe cleft palate, and a hole between his throat and esophagus. The child died three hours after birth.

PROCEDURAL HISTORY

[56] The plaintiffs commenced this putative class action on December 22, 2014. The causes of action asserted in it are negligent failure to warn, strict or negligent liability to fetuses in utero, and statutory derivative claims (consisting of dependants’ relief claims on behalf of family members, and subrogated claims on behalf of provincial health insurers).

[57] The sole common issue the plaintiffs propose is the PCI, i.e., whether the defendants failed to warn class members that Celexa® is or may be a teratogen, which the plaintiffs acknowledge (in their current factum) requires a two-step inquiry to show a basis in fact for certification of both the General Causation PCI and the Duty to Warn PCI.

[58] The PCI is the product of a previous decision by Perell J., in which he denied certification of the class action (the “First Certification Decision”),[^20] together with a decision by the Divisional Court in which the First Certification Decision was set aside and the certification motion was ordered to be reheard by this court (the “Appeal Decision”).[^21]

[59] In their factum for the first certification motion, the plaintiffs proposed two common issues, the first of which asked “[i]s citalopram or may citalopram be teratogenic?”, and the second of which asked “[i]f so, did the Defendants breach a duty to warn Canadian physicians and patients that citalopram is or may be teratogenic?”

[60] However, after receiving Lundbeck’s factum, which outlined why the general causation issue could not be meaningfully resolved on a class-wide basis given the many different congenital malformations it involved, the plaintiffs abandoned the first issue and asked that a class action be certified based on the proposed duty to warn common issue they propose again on the present motion. The Divisional Court noted in the Appeal Decision, at para. 27:

Counsel below grappled with the fact that they claimed that many different types of birth defects were allegedly caused by citalopram. They were apparently concerned that having to deal with different kinds of birth defects suffered by different plaintiffs or groups of plaintiffs could undermine the submission that the duty to warn presents an issue that is common to every class member. Therefore, counsel apparently conceded before the motion judge that both general and specific causation were issues for individual hearings rather than the common issue hearing…. [Emphasis added.]

[61] In view of this development, Perell J. concluded that the plaintiffs had made the “strategic choice” to seek certification based solely on a failure to warn common issue, without any common issue related to general causation. He proceeded to deny certification upon concluding that the failure to warn issue did not meet the commonality requirement in s. 5(1)(c), and that even if it did, a class action would not be the preferable procedure as required by s. 5(1)(d).

[62] On appeal, the Divisional Court set aside the First Certification Decision. The court held that Perell J. erred in finding that the plaintiffs did not propose a common issue involving general causation. The Divisional Court held, at para. 31, that the failure to warn issue which the plaintiffs put forward “always contained an aspect of causation despite counsel’s concession.” The court stated, at paras. 28-29:

The common issue proposed by the plaintiffs before and after counsel’s concession was “did the Defendants breach a duty to warn Canadian physicians and patients that Citalopram is or may be a teratogen”. That is, the common issue proposed included not just the duty to warn but the duty to warn that the drug can cause birth defects…

Despite counsel's apparent concession, the common issue that counsel continued to propose to the motion judge did include a general causation issue that may indeed be common to the entire class. [Italics in original; bold added.]

[63] Given this conclusion, the Divisional Court held that Perell J. erred in finding that the plaintiffs failed to satisfy ss. 5(1)(c) or (d), because he (i) never considered these criteria on the basis that the proposed common issue contained the general causation aspect outlined above, and (ii) made findings under them that were derivative of his error about the withdrawal of a general causation issue.

[64] However, the Divisional Court did not accept the plaintiffs’ submission that the court should certify the class action. The Divisional Court did not determine whether the “duty to warn causation issue” (i.e., the General Causation PCI now before this court) satisfied the conditions for certification under ss. 5(1)(c) or (d). Instead, the Divisional Court set aside the First Certification Decision, and ordered that the certification motion be reheard by a different judge, based on the proposed common issue with the general causation aspect it identified. The court held, at para. 49:

While an appellate court has the jurisdiction to substitute its decision for that of a motion judge, it would not be appropriate to do so here. We have found that the common issue advanced by the plaintiffs includes at least one aspect of general causation that may be common across the proposed class, and that this aspect may bear upon other causation issues in the case. However, … there has yet to be consideration of whether the common issue as proposed satisfies each aspect of the applicable test. We are not in a position to make a decision at first instance as to the effect of the proposed common issue with its remaining general causation aspect and perhaps group general causation as proposed by Mr. Bates before us. As this discussion should make clear, these are decisions to be made at first instance, by the judge who would then be responsible for moving forward with the litigation if the class is certified. The certification motion must be heard afresh. [Emphasis added.]

[65] On November 15, 16, and 17, 2022, I heard this matter “afresh” as directed by the Divisional Court.

ANALYSIS

[66] I first review the scope of the hearing, given the procedural history set out above.

[67] I then consider the applicable law on the test for certification and the “some basis in fact” test.

[68] Then, I consider each of the objections raised by the defendants under ss. 5(1)(c) and (d).

[69] Finally, I consider the ancillary issues under ss. 5(1)(a) and (b), which are relevant only if the requirements under ss. 5(1(c) and (d) are satisfied by the plaintiffs (which I do not find).

The scope of the hearing

[70] The effect of the Appeal Decision is that this court must decide the certification motion anew, making its own evidentiary and legal findings. The First Certification Decision cannot be res judicata. Stinson J. stated this principle in Zesta Engineering Ltd. v. Cloutier, 2010 ONSC 5810, 86 C.C.E.L. (3d) 1, at para. 245, var’d on other grounds, 2014 ONCA 762, 20 C.C.E.L. (4th) 168:

As to Wright J.’s decision, his judgment was set aside by the Court of Appeal, save in relation to Cloutier's admitted liability for payment of the quantum of the commissions received by him. Once an appellate court sets aside a judgment below, the first decision ‘disappears altogether’. [Citations omitted; emphasis added.]

[71] As for the Appeal Decision, there is “no estoppel where a court of appeal grants a new trial of an entire case without restriction”, although if the “court of appeal in granting a new trial decides a substantive question in the litigation”, then “[t]hat question … is taken to have been conclusively determined between the parties for purposes of the litigation”: MacKinnon v. National Money Mart Company, 2009 BCCA 103, 89 B.C.L.R. (4th) 1, at para. 78.

[72] The only “substantive question” determined by the Divisional Court was that the common issue advanced by the plaintiffs included an aspect of general causation: see the Appeal Decision, at para. 49.

[73] The Divisional Court expressly declined to decide whether the certification criteria in the CPA were satisfied, instead requiring that the court that would hear the certification motion consider the general causation issue it identified: see the Appeal Decision, at paras. 47-49. This court must therefore review the record and determine if the certification test under the CPA is met.

The applicable law on the test for certification and the “some basis in fact” test

The test for certification

[74] In the First Certification Decision, Perell J. set out the general principles for the court to apply on a certification motion. The Divisional Court did not find any error with his analysis of the law. I adopt his concise summary, at paras. 80-81:

The court has no discretion and is required to certify an action as a class proceeding when the following five-part test in s. 5 of the Class Proceedings Act, 1992 is met: (1) the pleadings disclose a cause of action; (2) there is an identifiable class of two or more persons that would be represented by the representative plaintiff; (3) the claims of the class members raise common issues; (4) a class proceeding would be the preferable procedure for the resolution of the common issues; and (5) there is a representative plaintiff who: (a) would fairly and adequately represent the interests of the class; (b) has produced a plan for the proceeding that sets out a workable method of advancing the proceeding on behalf of the class and of notifying class members of the proceeding, and (c) does not have, on the common issues for the class, an interest in conflict with the interests of other class members.

For an action to be certified as a class proceeding, there must be a cause of action shared by an identifiable class from which common issues arise that can be resolved in a fair, efficient, and manageable way that will advance the proceeding and achieve access to justice, judicial economy, and the modification of behaviour of wrongdoers. On a certification motion, the question is not whether the plaintiff's claims are likely to succeed on the merits, but whether the claims can appropriately be prosecuted as a class proceeding. The test for certification is to be applied in a purposive and generous manner, to give effect to the goals of class actions; namely: (1) providing access to justice for litigants; (2) encouraging behaviour modification; and (3) promoting the efficient use of judicial resources. [Citations omitted.]

The “some basis in fact test

[75] In the First Certification Decision, Perell J. reviewed the applicable case law and principles for the “some basis in fact” test, which requires the plaintiff to lead evidence that the requirements of ss. 5(1)(b) to (e) are met. I adopt his summary at paras. 82-85, as it concisely sets out the relevant legal principles:

The purpose of a certification motion is to determine how the litigation is to proceed and not to address the merits of the plaintiff's claim; there is to be no preliminary review of the merits of the claim. However, the plaintiff must show “some basis in fact” for each of the certification criteria other than the requirement that the pleadings disclose a cause of action. In the context of the common issues criterion, the "some basis in fact" standard involves a two-step requirement that: (1) the proposed common issue actually exists; and (2) the proposed issue can be answered in common across the entire class.

The “some basis in fact” test sets a low evidentiary standard for plaintiffs, and a court should not resolve conflicting facts and evidence at the certification stage or opine on the strengths of the plaintiff's case. In particular, there must be a basis in the evidence to establish the existence of common issues. To establish commonality, evidence that the alleged misconduct actually occurred is not required; rather, the necessary evidence goes only to establishing whether the questions are common to all the class members.

The representative plaintiff must come forward with sufficient evidence to support certification, and the opposing party may respond with evidence of its own to challenge certification. Certification will be denied if there is an insufficient evidentiary basis for the facts on which the claims of the class members depend.

On a certification motion, evidence directed at the merits may be admissible if it also bears on the requirements for certification but, in such cases, the issues are not decided on the basis of a balance of probabilities, but rather on the much less stringent test of some basis in fact. The evidence on a motion for certification must meet the usual standards for admissibility. While evidence on a certification motion must meet the usual standards for admissibility, the weighing and testing of the evidence is not meant to be extensive, and if the expert evidence is admissible, the scrutiny of it is modest. In a class proceeding, the close scrutiny of the evidence of experts should be reserved for the trial judge. [Citations omitted.]

[76] While the “some basis in fact” test is a “low evidentiary standard”, the court has an important gate-keeping function. In Pro-Sys Consultants Ltd. v. Microsoft Corporation, 2013 SCC 57, [2013] 3 S.C.R. 477, the court held, at paras. 103-04:

[I]t is worth reaffirming the importance of certification as a meaningful screening device. The standard for assessing evidence at certification does not give rise to “a determination of the merits of the proceeding” … nor does it involve such a superficial level of analysis into the sufficiency of the evidence that it would amount to nothing more than symbolic scrutiny…

There must be sufficient facts to satisfy the applications judge that the conditions for certification have been met to a degree that should allow the matter to proceed on a class basis without foundering at the merits stage by reason of the requirements of … the CPA not having been met. [Emphasis added.]

[77] The gate-keeping function ensures fairness to both plaintiffs and defendants. In Krishnan v. Jamieson Laboratories Inc., 2021 BCSC 1396, Branch J. held, at para. 41:

The court has an important gate-keeping role requiring it to screen proposed claims to ensure they are suitable for class action treatment. In Thorburn v. British Columbia, 2012 BCSC 1585, appeal dismissed 2013 BCCA 480, the court stated:

The goal of the CPA is to be fair to both plaintiffs and defendants… “it is imperative to have a scrupulous and effective screening process, so that the court does not sacrifice the ultimate goal of a just determination between the parties on the altar of expediency.” [Ellipsis in original; emphasis added.]

[78] Consequently, the court must still undertake a comprehensive review of the relevant evidence to determine whether there is some basis in fact to meet the requirements under ss. 5(1)(b) to (e): Bowman v. Ontario, 2022 ONCA 477, 83 C.C.L.T. (4th) 235, at para. 101. Each case must be decided on its own facts, based on the adequacy of the record before the court: Hollick v. Toronto (City), 2001 SCC 68, [2001] 3 S.C.R. 158, at para. 23; Microsoft, at para. 104.

[79] I now review the objections raised by the defendants.

The objections under s. 5(1)(c)

[80] Before considering each of the objections under s. 5(1)(c), I first review the applicable law for the court to consider when determining whether a proposed issue is a common issue appropriate for certification. I then consider the particular objections.

The applicable law

(i) The importance of the commonality requirement

[81] The commonality requirement has been described as “‘[t]he central notion of a class proceeding’” and “is based on the notion that ‘individuals who have litigation concerns “in common” ought to be able to resolve those common concerns in one central proceeding rather than through an inefficient multitude of repetitive proceedings’”: Microsoft, at para. 106, citing M.A. Eizenga et al., Class Action Law and Practice (loose-leaf), at p. 3-34.6.

(ii) The test to establish commonality

[82] The Divisional Court in the Appeal Decision held, at para. 22, that “the motion judge properly instructed himself on the general principles applicable to the identification of common issues”. I adopt the concise and thorough analysis of Perell J. at paras. 104-11 of the First Certification Decision, which I set out below:

The third criterion for certification is the common issues criterion. For an issue to be a common issue, it must be a substantial ingredient of each class member's claim and its resolution must be necessary to the resolution of each class member's claim. The underlying foundation of a common issue is whether its resolution will avoid duplication of fact-finding or legal analysis of an issue that is a substantial ingredient of each class member's claim and thereby facilitate judicial economy and access to justice. In Pro-Sys Consultants Ltd. v. Microsoft Corporation, the Supreme Court of Canada describes the commonality requirement as the central notion of a class proceeding which is that individuals who have litigation concerns in common ought to be able to resolve those common concerns in one central proceeding rather than through an inefficient multitude of repetitive proceedings.

All members of the class must benefit from the successful prosecution of the action, although not necessarily to the same extent. The answer to a question raised by a common issue for the plaintiff must be capable of extrapolation, in the same manner, to each member of the class.

An issue is not a common issue if its resolution is dependent upon individual findings of fact that would have to be made for each class member. Common issues cannot be dependent upon findings which will have to be made at individual trials, nor can they be based on assumptions that circumvent the necessity for individual inquiries.

Commonality is a substantive fact that exists on the evidentiary record or it does not, and commonality is not to be semantically manufactured by overgeneralizing; i.e., by framing the issue in general terms that will ultimately break down into issues to be resolved by individual inquiries for each class member. In Rumley v. British Columbia, Chief Justice McLachlin stated that an issue would not satisfy the common issues test if it was framed in overly broad terms; she stated:

[...] It would not serve the ends of either fairness or efficiency to certify an action on the basis of issues that are common only when stated in the most general terms. Inevitably such an action would ultimately break down into individual proceedings. That the suit had initially been certified as a class action could only make the proceeding less fair and less efficient.

However, the commonality requirement does not mean that an identical answer is necessary for all the members of the class, or even that the answer must benefit each of them to the same extent; it is enough that the answer to the question does not give rise to conflicting interests among the members; success for one member must not result in failure for another.

The common issue criterion presents a low bar. An issue can be a common issue even if it makes up a very limited aspect of the liability question and even though many individual issues remain to be decided after its resolution. Even a significant level of individuality does not preclude a finding of commonality. A common issue need not dispose of the litigation; it is sufficient if it is an issue of fact or law common to all claims and its resolution will advance the litigation.

As already noted above, in the context of the common issues criterion, the some basis in fact standard involves a two-step requirement that: (1) the proposed common issue actually exists; and (2) the proposed issue can be answered in common across the entire class.

Where questions relating to causation or damages are proposed as common issues, the plaintiff must demonstrate with supporting evidence that there is a workable methodology for determining such issues on a class-wide basis. [Citations omitted; ellipsis in original; emphasis added.]

(iii) The first step of the two-step requirement

[83] The first step of the two-step inquiry, i.e. to establish that “the proposed common issue actually exists”, requires that the plaintiff “adduce some evidence demonstrating that there is a ‘colourable claim’, or a rational connection between the class members as defined, and the proposed common issues”: Kuiper v. Cook (Canada) Inc., 2020 ONSC 128, 149 O.R. (3d) 521 (Div. Ct.), at para. 27; and Hollick, at paras. 19-20.

[84] The evidence before the court on a certification motion “must raise a legitimate possibility that the question or questions could be answered in [the plaintiff’s] favour”, in the sense that there is “some factual basis − in the form of admissible evidence − to support the allegation” to which the common issues relate: Simpson v. Facebook, Inc., 2022 ONSC 1284, 160 O.R. (3d) 629 (Div. Ct.), at para. 27, citing Williams v. Canon Canada Inc., 2012 ONSC 3692, 294 O.A.C. 251 (Div. Ct.), at para. 23.

[85] The defendants do not dispute that this first step has been met in the present case. The argument at the hearing and in the defendants’ factum addressed the second step I review below.

(iv) The second step of the two-step requirement

[86] The second step of the two-step inquiry, i.e. that “the proposed issue can be answered in common across the entire class” requires the plaintiff to establish some basis in fact that there is a common issue which will advance the litigation.

[87] If causation or damages issues are proposed as common issues, the plaintiff must establish some basis in fact that there is a workable methodology for determining such issues on a class-wide basis.

[88] I address each of these issues below.

(a) A proposed common issue must advance the litigation

[89] An issue stated in general terms, even if it results in a finding common to the class, will not be appropriate as a “common issue” to support certification if it only provides “context” and does not yield concrete answers to real claims that would advance the litigation in a meaningful way: Dennis v. Ontario Lottery and Gaming Corp., 2013 ONCA 501, 116 O.R. (3d) 321, at paras. 58-59, 66, 68-70, leave to appeal refused, [2013] S.C.C.A. No. 373; Fehr v. Sun Life Assurance Company of Canada, 2018 ONCA 718, 84 C.C.L.I. (5th) 124, at paras. 56, 59-60, leave to appeal refused, 2019 37480 (S.C.C.).

[90] In Bennett v. Hydro One Inc., 2018 ONSC 7741 (Div. Ct.), the court held that the proposed common issue was not appropriate for certification, even though the issue as to whether the defendants’ billing system was poorly designed and implemented could give rise to a common answer. The court held, at para. 40: “[T]here is no saving compared to individual trials. Alternatively expressed, no issue is off the table at individual trials because it has been resolved already at the common trial” (emphasis added).

[91] In Loveless v. OLG, 2011 ONSC 4744, 340 D.L.R. (4th) 120, Strathy J. (as he then was) held, at para. 64, that a proposed common issue of “[h]ave retailers wrongfully deprived purchasers of OLG tickets of their winnings?” was a “soft pitch” which did “nothing to advance the claims of class members to first base”. He held, at para. 64:

A “yes” answer will do nothing to move any class member closer to the resolution of his or her claim. Every single class member with a claim will have to demonstrate that he or she was actually deprived of their winnings by a retailer. The answer to a general question asking whether some retailers have defrauded some ticket purchasers is utterly meaningless. [Emphasis added.]

[92] Strathy J. further stated in Loveless, at paras. 62 and 67, respectively:

(i) “[t]he generality and breadth of the common issues is one of several reasons why, in my view, this action is not capable of proceeding as a class action”; and

(ii) the proposed common issue was “so general that it is of no value in advancing the resolution of class members’ claim” (emphasis added).

(b) The test to establish a workable methodology

[93] An expert methodology presented by a plaintiff seeking certification of a proposed common issue must be “sufficiently credible or plausible to establish some basis in fact for the commonality requirement”: Microsoft, at para. 118. In the same paragraph, the court held that:

This means that the methodology must offer a realistic prospect of establishing loss on a class-wide basis so that, if the overcharge is eventually established at the trial of the common issues, there is a means by which to demonstrate that it is common to the class (i.e. that passing on has occurred). The methodology cannot be purely theoretical or hypothetical, but must be grounded in the facts of the particular case in question. There must be some evidence of the availability of the data to which the methodology is to be applied. [Emphasis added.]

[94] It is inappropriate to weigh expert evidence on a certification motion. In Hague v. Liberty Mutual Insurance Co. (2004), 2004 95289 (ON SC), 21 C.C.L.I. (4th) 264 (Ont. S.C.J.), Nordheimer J. (as he then was), held, at para. 75:

[I]t is inappropriate on a certification motion to engage in an evaluation of the strength or weaknesses of a given party’s evidence, especially expert evidence. That is properly the function of a trial judge. Other than being satisfied that there is “some evidence” to support a party’s assertions, the certification judge should not engage in a weighing of competing evidence. To do so would not only embark on a preliminary merits review, it would also ignore the recognized reality that a motion is generally an unsuitable forum in which to make such evaluations.

[95] I now address each objection under s. 5(1)(c). I first consider the objections to the General Causation PCI and then consider the objections to the Duty to Warn PCI.

Objections to the General Causation PCI

(i) Objection 1: Is there a workable methodology to establish the General Causation PCI?

[96] The General Causation PCI does not seek a linkage to a particular congenital malformation, but instead asks the court to answer the general question of whether Celexa® is or may be a teratogen.

[97] Consequently, the issue of whether Dr. Bérard’s methodology is workable on a class-wide basis to answer the General Causation PCI must focus on the particular text of the General Causation PCI, and how Dr. Bérard proposes to answer that question for the class.

[98] I first address the terms of the General Causation PCI proposed by the plaintiffs in the present case.

[99] I then consider Dr. Bérard’s proposed methodology to determine whether it is plausible or credible.

(a) The terms of the General Causation PCI

[100] The plaintiffs do not assert that Celexa® caused any particular congenital malformation, nor even a particular type of congenital malformation. The plaintiffs acknowledge that Celexa® cannot cause all congenital malformations.

[101] Instead, the plaintiffs rely on the General Causation PCI, in which the general causation issue is whether Celexa® is, or may be, a teratogen. Consequently, I only consider whether Dr. Bérard has proposed a workable methodology to determine, on a class-wide basis, the teratogenicity of Celexa®.

(b) Is Dr. Bérard’s methodology workable to answer the General Causation PCI on a class-wide basis?

[102] For the reasons that follow, I agree with the plaintiffs that Dr. Bérard’s proposed methodology to establish the general teratogenicity of Celexa® is workable.

[103] The scope of Dr. Bérard’s retainer is summarized in the plaintiffs’ factum, in which they confirm that “teratogenicity” is “the actual common issue”. Dr. Bérard states that “[w]hen congenital malformations are examined together, a clearer picture emerges as to citalopram’s teratogenicity than when malformations are examined separately”.

[104] Similarly, in her report, Dr. Bérard describes her mandate as “analyzing the teratogenicity of citalopram”, based on her methodology which she describes as follows:

[F]indings for an organ system group (for example – heart defects overall) can be extrapolated to specific defects within the group if they occur at similar embryological stages. This is the case for all congenital malformations. Serotonin reuptake inhibition produced by citalopram use during organogenesis can cause multiple defects. The fact that up to 40% of pregnancies spontaneously abort limits the ability of researchers to study rare, specific defects on livebirths, hence the importance of studying groupings of defects. When analyzing the teratogenicity of citalopram, it materially advances a causal inquiry to analyze a variety of congenital malformations as a group and to later apply that to specific defects. [Emphasis added.]

[105] As discussed at para. 48 above, Dr. Bérard’s methodology assesses the teratogenicity of Celexa® based on a three-step approach. With respect to the first step (an overall analysis), Dr. Bérard relies on peer-reviewed scientific evidence that she asserts establishes a causal association between exposure to citalopram and congenital malformations.

[106] Dr. Cabrera also provided evidence of studies that establish an association between citalopram and congenital malformations. Dr. Bérard, Dr. Healy and Dr. Cabrera all rely on evidence that the effects of changes in serotonin can cause congenital malformations.

[107] The above evidence establishes a plausible basis that the court could accept the above evidence as supportive of the first (“overall”) step in Dr. Bérard’s proposed methodology.

[108] In the second step of her proposed methodology, Dr. Bérard proposes to consider the “overall” data by grouping organ systems to conduct a more detailed analysis of teratogenicity.

[109] On cross-examination, Dr. Bérard affirmed the use of ICD codes despite their limitations. She stated that all researchers use those codes since that is how the data on congenital malformations is organized.

[110] Consequently, the plaintiffs have established some basis in fact for a plausible methodology to address this second stage of Dr. Bérard’s analysis.

[111] In the third step of her methodology, Dr. Bérard proposes to analyze the evidence from organ systems by embryologic stage of the origin of the cells. She maintains such an analysis is required because different organs may be exposed to the risk of citalopram at different stages of embryonic development.

[112] In response to the defendants’ criticisms that Dr. Bérard offered no methodology to consider the embryologic stage of development, Dr. Bérard referred in cross-examination to studies and research which considered the embryologic stage of the origin of the cells, consistent with her methodology. On cross-examination, Dr. Bérard did not accept that she could not use data as she proposed.

[113] In challenging Dr. Bérard’s proposed methodology, the defendants submit that they are not asking the court to weigh expert evidence and instead are relying solely on the plaintiffs’ evidence to establish that there is no workable methodology. I disagree.

[114] In essence, the defendants ask the court to reject Dr. Bérard’s evidence since the defendants’ experts challenge the validity of the premises underlying her approach. The defendants’ experts opine that the teratogenicity of Celexa® can only be determined by analyzing a particular congenital malformation. The defendants submit that a finding of teratogenicity with respect to one particular congenital malformation could not apply on a class-wide basis.

[115] However, neither Dr. Bérard nor any of the other experts retained by the plaintiffs accept that the concerns raised by the defendants would prevent Dr. Bérard from using her proposed approach. The evidence before the court establishes that Dr. Bérard’s approach is plausible and credible. Any challenges to that approach would have to be considered by a common issues judge.

[116] Under Dr. Bérard’s approach, the court on a common issues trial may be satisfied that Celexa® is or may be teratogenic, without requiring a finding that is based on a specific congenital malformation. If a finding is made that Celexa® is teratogenic, all class members can rely on that finding, without having to establish that issue by their own expert evidence.

[117] For the above reasons, I find that Dr. Bérard’s methodology to establish an answer to the General Causation PCI on a class-wide basis is workable.

(ii) Objection 2: Does the answer to the General Causation PCI advance the litigation?

[118] The defendants submit that there is no basis in fact that the answer to the General Causation PCI could advance the litigation, even if there is a workable methodology to arrive at a class-wide answer. For the reasons that follow, I agree with the defendants.

[119] The key issue under s. 5(1)(c) is whether the proposed common issue advances the litigation or whether it is a “soft pitch” that does not advance the plaintiff to “first base”: Loveless, at para. 64.

[120] In the present case, the plaintiffs do not assert a general causation issue that links Celexa® to a particular congenital malformation.

[121] Instead, the plaintiffs ask the court to certify the General Causation PCI – i.e. whether Celexa® is or may be a teratogen. The answer to that question, even if it could be determined on a class-wide basis applying the methodology proposed by Dr. Bérard, does not advance the case.

[122] To assess whether the General Causation PCI advances the litigation for a class member, I assume for the purposes of the plaintiffs’ submission that a class member would succeed on the General Causation PCI such that the common issues judge finds that Celexa® is or may be teratogenic.

[123] If the common issues trial judge finds that Celexa® is not, and cannot be, teratogenic, then all class members would have no claim. However, the commonality of a proposed common issue in a pharmaceutical duty to warn case cannot be determined solely on the basis of what occurs if general causation is not established.

[124] If the General Causation PCI is decided in the plaintiffs’ favour, no class member is advanced in their case. Liability of the defendants can only exist if each class member can establish general causation that Celexa® could have caused their specific congenital malformation. Nothing is gained by a finding that Celexa® is or may be a teratogen.

[125] In Bartram (Litigation guardian of) v. GlaxoSmithKline Inc., 2012 BCSC 1804, aff’d 2013 BCCA 462, relied upon by the plaintiffs, the court refused to certify the initial proposed common issue of “Did Paxil cause or increase the likelihood of birth defects”: at para. 35. That question is similar to that proposed in the present case.

[126] Instead, the court in Bartram narrowed that proposed common issue to whether Paxil caused cardiovascular congenital malformations, matching the same language in the proposed class definition in that case. N.H. Smith J. held, at para. 35: “I would, however, narrow the question to whether Paxil causes or increases the likelihood of cardiovascular birth defects. That is the type of defect alleged in the case of the proposed representative plaintiff and is the only type referred to in the proposed class definition.”

[127] In Miller v. Merck Frosst Canada Ltd., 2013 BCSC 544, aff’d, 2015 BCCA 353, leave to appeal refused, 2016 20439 (S.C.C.), relied upon by the plaintiffs, the general causation issue before the court, based on the expert’s methodology, was whether the drugs could cause sexual dysfunction that continued after using the medication. The court narrowed the proposed general causation issue from the initial version, at para. 144, of “[c]an ingesting Propecia or Proscar cause side effects that continue after ceasing to take Propecia or Proscar”, and modified it so that general causation was linked to the specific risk of sexual dysfunction. Punnett J. held, at para. 175: “This issue must be narrowed to read: ‘Can ingesting Propecia or Proscar cause sexual dysfunction which persists after ceasing to take Propecia or Proscar?’”

[128] It was on the basis of the narrowed general causation issue that the court in Miller found that a common issue could advance the case and could be established on a class-wide basis by a plausible methodology.

[129] In the present case, the option of narrowing the General Causation PCI, as was done in Miller and Bartram, is not available to the court. The plaintiffs led no evidence of a methodology to establish, on a class-wide basis, that Celexa® can cause any particular congenital malformation.

[130] The client information form in the plaintiffs’ litigation plan proposes 11 different congenital malformations as possible options for class members, as well as leaving space for class members to add up to three “other” congenital malformations which could be from any of the hundreds of potential congenital malformations. There is no evidence that would enable the court to limit the proposed class definition or the General Causation PCI.

[131] The plaintiffs’ position simply “seizes on the superficial commonality” that citalopram is or may be a teratogen, using a “common label [to] conve[y] a false impression of commonality”: McCracken v. Canadian National Railway Company, 2012 ONCA 445, 111 O.R. (3d) 745, at para. 128.

[132] As the test is stated in McCracken, “the plaintiff has not shown that any significant element of his claim is capable of common proof”: at para. 128.

[133] The proposed class members cannot avoid proof of general causation for the particular congenital malformation even if Celexa® is or may be teratogenic. If a class member cannot establish that Celexa® could have caused the particular congenital malformation, the defendants cannot be liable for any damage.

[134] In these circumstances, “[r]ather than providing an effective procedural tool to advance the resolution of the claims of the proposed class members, a class proceeding would amount to little more than a general commission of inquiry” into whether citalopram is a “teratogen”, prompting the same concern as raised by Sharpe J.A. in Dennis, at para. 59, in relation to “the prevention of problem gambling”.

[135] The plaintiffs submit that a common issues judge could provide a “nuanced” approach at trial, since not all class members have to be affected in the same way by the answer to a common issue. In this regard, the plaintiffs rely on the decision in Rumley v. British Columbia, 2001 SCC 69, [2001] 3 S.C.R. 184, at para. 32. However, a proposed common issue that is so general as to not advance the claim cannot be the basis of certification, even if a common answer can be provided. McLachlin C.J.C held, at para. 29:

There is clearly something to the appellant’s argument that a court should avoid framing commonality between class members in overly broad terms. As I discussed in Western Canadian Shopping Centres, supra, at para. 39, the guiding question should be the practical one of “whether allowing the suit to proceed as a representative one will avoid duplication of fact-finding or legal analysis”. It would not serve the ends of either fairness or efficiency to certify an action on the basis of issues that are common only when stated in the most general terms. Inevitably such an action would ultimately break down into individual proceedings. That the suit had initially been certified as a class action could only make the proceeding less fair and less efficient. [Emphasis added.]

[136] On the facts of the present case, the General Causation PCI offers “no saving compared to individual trials” since general causation is not “off the table at individual trials because it has been resolved already at the common trial”: Bennett, at para. 40.

[137] For the above reasons, I accept the defendants’ objection and find that the General Causation PCI does not raise a common issue suitable for certification.

Objections to The Duty to Warn PCI

[138] The defendants submit that the Duty to Warn PCI is not a common issue appropriate for certification under s. 5(1)(c) because:

(i) If the General Causation PCI does not raise a common issue under s. 5(1)(c), then the Duty to Warn PCI necessarily fails to meet the s. 5(1)(c) test.

(ii) In any event, the Duty to Warn PCI cannot stand on its own terms since it necessarily requires a warning with respect to a specific risk, and cannot be established on the basis of a failure to warn of teratogenicity in general.

[139] For the reasons that follow, I agree with both objections raised by the defendants. I address each of these issues below.

(i) Objection 1: If the General Causation PCI does not raise a common issue under s. 5(1)(c), then the Duty to Warn PCI necessarily fails to meet the s. 5(1)(c) test

[140] General causation has been described as the “first step” and “central question” in pharmaceutical class actions: Wise v. Abbott Laboratories Ltd., 2016 ONSC 7275, 34 C.C.L.T. (4th) 25, at para. 340; Miller, at paras. 42, 57, 62.

[141] It is only if general causation is made out that “the state of any manufacturer’s knowledge of the risks of causing that harm become material”, making “a consideration of causation… central to procedural screening criteria for class proceedings founded in negligence”: Harrington v. Dow Corning Corp., 2000 BCCA 605, 82 B.C.L.R. (3d) 1, at paras. 56-58, leave to appeal refused, [2001] S.C.C.A. No. 21.

[142] The plaintiffs submit that the defendants had a duty to warn that Celexa® is or may be teratogenic, since some class members may establish at individual trials that they would not have taken Celexa® had they been warned of its alleged teratogenicity.

[143] However, as with the General Causation PCI, even if the Duty to Warn PCI could impose a general teratogenicity warning as proposed by the plaintiffs (which I do not find as discussed below), each class member would still be required to establish general causation on an individual level for each particular congenital malformation at issue.

[144] Consequently, there would be no “saving” compared to individual trials, and the Duty to Warn PCI would be a “soft pitch” that does not advance the class to “first base”, even if the claimed duty to warn of general teratogenicity under the Duty to Warn PCI is accepted by the court.

[145] On this basis, I find that the Duty to Warn PCI fails to establish a common issue under s. 5(1)(c) since general causation cannot be established without individual trials based on the particular congenital malformation.

(ii) Objection 2: In any event, the Duty to Warn PCI cannot stand on its own terms since it necessarily requires a warning with respect to a specific risk, and cannot be established on the basis of a failure to warn of teratogenicity in general

[146] I first review the positions of the parties and then review the issues before the court.

(a) The positions of the parties

[147] The plaintiffs submit that a duty to warn applies to a general risk of harm. The plaintiffs submit that a duty to warn could be based solely on the defendants being required to advise women using Celexa® that it “is, or may be teratogenic”, as set out in the Duty to Warn PCI, because some women may have decided not to use Celexa® during pregnancy if they had been advised that it could cause unidentified congenital malformations.

[148] The defendants submit that the duty to warn applies only to specific risks, whether identified narrowly (e.g. cardiac arrythmia) or more broadly (e.g. drowsiness). The defendants submit that the case law does not impose a duty to warn of general “harm” or general “teratogenicity”, which would not disclose what specific risk is at issue.

[149] For the reasons that follow, I accept the defendants’ position.

(b) Analysis

[150] The leading case on the duty to warn is Hollis v. Dow Corning Corp., 1995 55 (SCC), [1995] 4 S.C.R. 634. I agree with the defendants that the inquiries required under Hollis establish that the duty to warn can only arise with respect to specific risks. A duty to warn cannot be imposed to require a general warning of potential harm or teratogenicity (without any reference to the specific risk).

[151] First, the court in Hollis held that a duty to warn exists only for “material risks associated with the use of a medical product”: at para. 57.

[152] For a risk to be material (and, as such, to require disclosure), the cause of the injury must be known. In Knickle v. Rayner (1991), 1991 2752 (PE SCAD), 88 Nfld. & P.E.I.R. 214 (P.E.I.C.A.), at p. 22, the court held:

It is the opinion of the court that it was necessary for the trial judge to decide the cause of the injury before he decided the issue of informed consent. The trial judge must know what particular risk he is dealing with before he can decide whether the risk involved was material as to require disclosure. [Emphasis added.]

[153] The court in Knickle, at p. 23, adopted a passage from Esson J.A. in Schinz v. Dickinson (1985), 1984 400 (BC CA), 59 B.C.L.R. 351 (C.A.) at p. 68 in which he held that it was not “any risk, however slight its magnitude, or however remote the chance that it would come to pass, [that] must be disclosed”.

[154] The necessity that the risk be material to require disclosure supports the defendants’ position that a duty to warn generally of congenital malformations cannot be encompassed in a duty to warn.

[155] There are hundreds of congenital malformations, some of which might be material and others which may not. The patient receiving a general warning of teratogenicity from a doctor would have no basis to assess the specific risk and make an informed decision as to whether to take Celexa® to help protect both the mother and the baby from potential risks arising from depression, or to choose not to do so. This informed consent is the raison d’être of the duty to warn, and it could not be effected under the Duty to Warn PCI.

[156] Second, the court in Hollis held that the duty to warn can only arise for material risks “of which [the manufacturer] has knowledge or ought to have knowledge”: at para. 20. Reasonable foreseeability is required.

[157] Reasonable foreseeability of harm in tort law requires the court to ask whether the manufacturer “ought reasonably to have contemplated the type of harm the plaintiff suffered”: Rankin (Rankin’s Garage & Sales) v. J.J., 2018 SCC 19, [2018] 1 S.C.R. 587, at para. 21. Reasonable foreseeability is assessed at the date when the plaintiff used the allegedly dangerous product: Hollis, at para. 37.

[158] With respect to the duty to warn, the court held in Bow Valley Husky (Bermuda) v. Saint John Shipbuilding Ltd., 1997 307 (SCC), [1997] 3 S.C.R. 1210, at para. 61, per McLachlin J. (as she then was) (dissenting on other grounds):

Where a duty to warn is alleged, the issue is not reliance (there being nothing to rely upon), but whether the defendants ought reasonably to have foreseen that the plaintiffs might suffer loss as a result of use of the product about which the warning should have been made. … [Emphasis added.]

[159] “Teratogenicity” is not the harm suffered, but instead a term that relates to the possibility of hundreds of congenital malformations, only some of which (if any) might have been reasonably foreseeable. If it were found that the defendants ought to have known that citalopram is a teratogen because it can cause clubfoot, it may not have been the case that the defendants ought to have known (or known at the same time) that citalopram is a teratogen because it can cause craniosynostosis. The etiologies and evolving epidemiological data differ for each congenital malformation.

[160] The reasonable foreseeability of any particular congenital malformation would require analysis of the etiology for each such malformation, which will vary. Not every congenital malformation would be, or could have been, reasonably foreseeable based on the etiology and epidemiological studies available at a particular time. Consequently, there could be no commonality under the proposed terms of the Duty to Warn PCI.

[161] Third, the court in Hollis held that even where a duty to warn exists, “[t]he required explicitness of the warning will, of course, vary with the danger”: at para. 22. Accordingly, for a court to determine whether the warning is sufficient, the court would have to consider the specific risk at issue.

[162] Similarly, the Court of Appeal held in Buchan v. Ortho Pharmaceutical (Canada) Ltd. (1986), 1986 114 (ON CA), 54 O.R. (2d) 92 (C.A.), at p. 30:

Once a duty to warn is recognized, it is manifest that the warning must be adequate. It should be communicated clearly and understandably in a manner calculated to inform the user of the nature of the risk and the extent of the danger; it should be in terms commensurate with the gravity of the potential hazard, and it should not be neutralized or negated by collateral efforts on the part of the manufacturer. The nature and extent of any given warning will depend on what is reasonable having regard to all the facts and circumstances relevant to the product in question. [Emphasis added.]

[163] Product monographs contain different warnings with varying levels of explicitness. As stated in Batten (SCJ), at paras. 85-86:

A Product Monograph consists of three distinct parts; that is: (1) Part I: Health Professional Information, which provides information to healthcare professionals for the prescribing, dispensing, and administering of the medication; (2) Part II: Scientific Information, which provides highly detailed scientific research information such as the drug’s chemical composition, toxicology and data from clinical trials; and (3): Part III: Consumer Information, which provides information for the consumer about the medication, how to use it and what are the side effects.

Within a Product Monograph, any particular risk can be profiled with greater or lesser prominence. The most prominent warnings take the form of “boxed warnings,” found in Part I and repeated in Part III that highlight serious risks that are clinically significant or life-threatening. Other serious risks will be presented within the WARNINGS AND PRECAUTIONS section in Parts I and III. [Emphasis added.]

[164] A general statement of a risk of harm or a risk of congenital malformations cannot provide the court with the ability to assess the gravity or probability of the risk, and as such the explicitness of the product monograph warning cannot be considered.

[165] In the present case, the gravity and probability of congenital malformation vary significantly across the hundreds that can arise. The disclosure required for the risk of a virtually imperceptible cosmetic abnormality (assuming it is material and reasonably foreseeable) is much different from congenital malformations such as neural tube defects, which can cause significant damage.

[166] In addition to the above factors in Hollis, other relevant principles from case law on the duty to warn support the defendants’ position.

[167] The plaintiffs submit that, as long as a class member might not have taken Celexa® if she was warned that it was teratogenic, a duty to warn can be established regardless of whether Celexa® could have caused the congenital malformation. I do not agree.

[168] Where a duty to warn about one defect is breached, it cannot result in liability to class members who allege a different defect about which there was no failure to warn (or which, if disclosed, would not have changed their decision to use the product), regardless of whether those class members would not have used the product if the duty to warn about the first defect had been met: Knickle, at pp. 19-23.

[169] The duty to warn of general teratogenicity proposed by the plaintiffs would prevent the defendants from defending each claim on the basis that the failure to warn could not have caused the congenital malformation actually suffered. The case law does not support the plaintiffs’ position. For example, if the defendants breached a duty to warn about the risk of clubfoot, then a class member who alleges craniosynostosis would still need to establish (i) an independent failure to warn about craniosynostosis, and (ii) the class member would not have used citalopram had a proper warning of craniosynostosis (as opposed to clubfoot) been given.

[170] Consequently, a duty to warn of general teratogenicity cannot impose liability if a class member would not have used a pharmaceutical product after they were told it was or might be teratogenic. The Duty to Warn PCI does not raise a common issue. Each class member would be required to establish damage based on the specific congenital malformation, requiring proof of all aspects of the case, starting with general causation.

[171] In Adam v. Ledesma-Cadhit, 2021 ONCA 828, leave to appeal refused, [2022] S.C.C.A. No. 13, Brown J.A. reviewed the applicable case law and held that the duty to warn arises with respect to “specific” risks or dangers. He held, at para. 19:

The general principles governing the duty to warn by manufacturers of medical products are well known, not in dispute, and were summarized by the Supreme Court in Hollis v. Dow Corning Corp., 1995 55 (SCC), [1995] 4 S.C.R. 634, at paras. 20 to 29:

(iii) All warnings must be reasonably communicated and must clearly describe any specific dangers that arise from the ordinary use of the product;

(iv) The nature and scope of the manufacturer's duty to warn varies with the level of danger associated with the ordinary use of the product. Where there are significant dangers, it will rarely be sufficient for manufacturers to give general warnings concerning those dangers. Instead, the warnings

must be sufficiently detailed to give the consumer a full indication of each of the specific dangers arising from the use of the product;

… [Emphasis added.]

[172] In Adam, at para. 20, Brown J.A. summarized the law on the duty to warn, reiterating that it required the disclosure of a specific risk. He held:

In Hollis, the Supreme Court identified the overarching question to be answered as whether the manufacturer owed the patient a duty to warn of a specific risk. The Supreme Court broke that overarching question down into two sub-questions:

(i) Did the manufacturer have a duty to warn recipients of the medical product directly or could it satisfy its duty by warning a learned intermediary, such as a physician?

(ii) If the manufacturer could properly discharge its duty by warning the physician, did it adequately warn the physician of the specific risk in light of its state of knowledge at that time? [Emphasis added.]

[173] The approach in Adam is fully consistent with Hollis and supports the defendants’ position.

[174] The plaintiffs rely on a single paragraph from Hollis in which the court was not considering the requirements for the duty to warn, but was instead addressing the “learned intermediary” rule, which allows a manufacturer to warn a learned intermediary such as a doctor instead of warning the patient directly. In that analysis, the court stated, at para. 31, that:

I conclude that a manufacturer in Dow’s position can discharge its duty to the ultimate consumer by giving the treating surgeon clear, complete and current information concerning any general and specific risks that arise from the ordinary use of the product. [Emphasis added.]

[175] However, the above passage in Hollis does not support a duty to warn based on a generalized and unidentified risk of harm. The requirements of the tort, as set out in Hollis and Adam, impose a duty to warn of a specific risk. While some specific risks may be described more broadly than others (e.g. drowsiness), the requirements in Hollis and the other relevant principles that establish a duty to warn support the defendants’ position.

[176] In summary, all the requirements to establish a duty to warn require that a specific risk (whether in specific terms or described more broadly) be disclosed. The Duty to Warn PCI is based on the General Causation PCI and proposes only that the duty to warn arises by a failure to warn that Celexa® is or may be teratogenic. As such, the Duty to Warn PCI cannot stand on its own terms. It is incompatible with the case law and principles discussed above.

[177] For the above reasons, I accept the defendants’ objection.

(iii) Conclusion

[178] The plaintiffs have not established a basis in fact that either the General Causation PCI or the Duty to Warn PCI is a common issue which can be certified under s. 5(1)(c).

[179] For the above reasons, I would not certify the PCI as a common issue under s. 5(1)(c) and I would dismiss the motion on that basis alone.

Preferable procedure

[180] Given my finding on s. 5(1)(c), it is not necessary to consider whether the proposed class action is the preferable procedure under s. 5(1)(d).

[181] However, for the reasons that follow, I find that even if the PCI was found to be a suitable common issue for certification, the proposed class action is not the preferable procedure.

[182] It is settled law that each of the requirements under s. 5(1) must be met for certification. Consequently, certification can be refused under s. 5(1)(d), even if the test under s. 5(1)(c) is met: Microsoft, at para. 139; Hollick, at paras. 27-36.

The applicable law

[183] I first review the applicable law on the preferable procedure requirement. I then apply that law to the record on this motion.

[184] I rely again on the summary of the applicable principles on the s. 5(1)(d) requirement as set out by Perell J. in the First Certification Decision, at paras. 145-46:

To satisfy the preferable procedure criterion, the proposed representative plaintiff must show some basis in fact that the proposed class action would: (a) be a fair, efficient and manageable method of advancing the claim; (b) be preferable to any other reasonably available means of resolving the class members' claims; and (c) facilitate the three principal goals of class proceedings; namely: judicial economy, behaviour modification, and access to justice.

In considering the preferable procedure criterion, the court should consider: (a) the nature of the proposed common issue(s) and their importance in relation to the claim as a whole; (b) the individual issues which would remain after determination of the common issue(s); (c) the factors listed in the Act; (d) the complexity and manageability of the proposed action as a whole; (e) alternative procedures for dealing with the claims asserted; (f) the extent to which certification furthers the objectives underlying the Act; and (g) the rights of the plaintiff(s) and defendant(s). [Citations omitted.]

[185] If a class action will not be fair, efficient and manageable, then it is immaterial that no other method of resolving the class members’ claims is preferable. As Strathy C.J.O. held in Amyotrophic Lateral Sclerosis Society of Essex County v. Windsor (City), 2015 ONCA 572, 387 D.L.R. (4th) 603, at para. 62:

Although recent case law, particularly AIC Limited v. Fischer, has focused on the second question – the comparison to alternative procedures – the proposed class proceeding must nevertheless be "fair, efficient and manageable" in order for it to be certified. As stated by Winkler J., as he then was, in Caputo, at para. 62:

[I]t is not enough for the plaintiffs to establish that there is no other procedure which is preferable to a class proceeding. The court must also be satisfied that a class proceeding would be fair, efficient and manageable. Both parts of the test must be considered in the context of the three goals of the CPA, judicial economy, access to justice and behavioural modification of tortfeasors. [Emphasis added.]

[186] In Bayens, the court held that a class action is not the preferable procedure if “the need for numerous individual inquiries undercuts the goal of judicial economy and could overwhelm the resolution of the common issues, producing an inefficient and unmanageable class proceeding”: at para. 129.

[187] The plaintiffs submit in their factum that “the Ontario legislature only required that the class action preferably resolve the common issues”, and that “[t]he focus is not on resolving the litigation, nor the individual issues”. The plaintiffs submit that “[t]he presence of individual issues supports, rather than detracts, from certification”.

[188] However, the plaintiffs’ submission is contrary to the case law. In Hollick, McLachlin C.J.C. held, at paras. 29-30, 32, that s. 5(1)(d) requires a court to consider the common issues in the context of the individual ones that will remain:

The Act itself, of course, requires only that a class action be the preferable procedure for “the resolution of the common issues”… and not that a class action be the preferable procedure for the resolution of the class members’ claims. I would not place undue weight, however, on the fact that the Act uses the phrase “resolution of the common issues” rather than “resolution of class members’ claims”. …

The question of preferability… must take into account the importance of the common issues in relation to the claims as a whole. … There must be a consideration of the common issues in context. As the Chair of the Attorney General’s Advisory Committee put it, the preferability requirement asks that the class representative “demonstrate that, given all of the circumstances of the particular claim, [a class action] would be preferable to other methods of resolving these claims and, in particular, that it would be preferable to the use of individual proceedings”… M.G. Cochrane, Class Actions: A Guide to the Class Proceedings Act, 1992 (1993), at p. 27.

I am not persuaded that the class action would be the preferable means of resolving the class members’ claims. Turning first to the issue of judicial economy, I note that any common issue here is negligible in relation to the individual issues. While each of the class members must, in order to recover, establish that the Keele Valley landfill emitted physical or noise pollution, there is no reason to think that any pollution was distributed evenly across the geographical area or time period specified in the class definition. … Once the common issue is seen in the context of the entire claim, it becomes difficult to say that the resolution of the common issue will significantly advance the action. [Emphasis added.]

[189] I now apply the above law to the present case.

Application of the law to the present case

(i) The class action cannot be the preferable procedure if there is no common issue suitable for certification under s. 5(1)(c)

[190] In the First Certification Decision, Perell J. held, at para. 153: “It is axiomatic that if the common issues criterion is not satisfied, the preferable procedure criterion is not satisfied. Therefore, in the immediate case, I can quickly conclude that the preferable procedure criterion is not satisfied”.

[191] I adopt this conclusion. A class action cannot be the preferable procedure if there is no common issue that can advance the case on behalf of the class members.

(ii) The class action would not be the preferable procedure even if the PCI was a common issue under s. 5(1)(c)

[192] If the PCI were certified, the only issues which would be determined in common would be whether (i) Celexa® is or may be a teratogen (the General Causation PCI) and (ii) the defendants breached their duty to warn of that teratogenicity (the Duty to Warn PCI).

[193] Even if those issues were found to advance the litigation (or suffice as a legal basis for a duty to warn), any such advance would be peripheral and minor. The individual issues would overwhelm any benefit from a common determination of teratogenicity and a concomitant duty to warn of such general teratogenicity.

[194] General causation of each specific congenital malformation would remain an issue for each class member.

[195] An overwhelming number of other individual issues would remain (including general causation), such as the issues of whether:[^22]

(i) the class member suffered an injury;

(ii) the injury was caused “in fact” by Lundbeck, i.e., it would not have occurred “but for” Lundbeck’s alleged failure, which requires determining both:

(a) “general causation” (whether citalopram is capable of causing the class member’s injury), at least for class members who allege that citalopram caused defects that are not part of a finding of teratogenicity in the common issues trial; and

(b) “specific causation” (whether citalopram in fact caused the class member’s injury);

(iii) the injury was caused “in law” by Lundbeck, i.e., the actual injury suffered is not too remote but the reasonably foreseeable result of the alleged failure;

(iv) for failure to warn, the class member was not independently aware of the risk, and their physician was not warned of the risk and failed to pass it on;

(v) also for failure to warn, there was “decision causation”, i.e., that if the class member had been hypothetically warned of the risk they suffered they would have subjectively refrained from taking citalopram;

(vi) again for failure to warn, there was “learned intermediary causation”, i.e., that if the class member’s physician(s) had been hypothetically warned of the risks they would not have engaged in extraneous conduct which would have made the warning irrelevant;

(vii) for failure to contraindicate, the risk of the injury the class member suffered outweighed the benefits of citalopram use for them;

(viii) for the family law claimants, each element of the provincial or territorial dependant’s relief statute under which they claim;

(ix) for all claims, the effect of the multiple defences which Lundbeck will assert against them, e.g., contributory negligence, failure to mitigate, limitation periods, etc.;

(x) again for all claims, the governing laws applicable to each cause of action based on the relevant statutory or common law choice of law rules; and

(xi) that general and special damages can be assessed, along with any amounts paid by class members (or their insurers) to Lundbeck for any wrongdoing.

[196] The plaintiffs rely on cases such as Miller and Bartram where the general causation issue was capable of advancing the litigation because each class member had suffered a common specific injury and the general causation question sought to answer whether the drug could cause that specific injury. In those cases, as I discuss above, both the general causation and duty to warn issues could advance the case (assuming that a workable methodology was established for causation).

[197] In cases such as Miller and Bartram, a class action was the preferable procedure since there could be a fair, efficient, and manageable way of advancing the claim such that each class member would avoid an individual action on the complex elements (including expert evidence) required to establish general causation and the duty to warn. In such cases, the preferable procedure test was met despite the presence of significant additional individual issues similar to those summarized at para. 195 above.

[198] However, there is no basis in fact for such a finding in the present case. Even if the General Causation PCI and Duty to Warn PCI are certified, no liability can be established unless each class member brings their own proceeding to address general causation.

[199] The plaintiffs submit that “groups” of class members might have general causation determined at a common issues trial, but offer no methodology to do so and the litigation plan does not set out a mechanism to do so. No common issue is proposed for such a determination.

[200] Given that each of the hundreds of malformations has its own unique etiology, epidemiological data, set of risk factors, and criteria for screening and diagnosis, the nature of the individual inquiries is particularly complex and unique. There are no known teratogens capable of causing all congenital malformations. The plaintiffs did not pursue a general causation issue linked to a particular congenital malformation or type of congenital malformation.

[201] In her affidavit, Dr. Stewart reviewed the evidence from the five women who swore affidavits in support of the motion. That evidence establishes a level of variability much greater than that which arose in the cases relied upon by the plaintiffs which addressed general causation with respect to a particular harm or type of harm. That variability between the five affiants would increase exponentially given the plaintiffs’ estimate that the class size consists of 5,723 women, not including “an additional number of in utero deaths (perhaps 20%)”.

[202] In her review of the evidence of the five affiants, Dr. Stewart found numerous personal risk factors for malformations and extensive variability in their claims. Dr. Stewart stated in her affidavit:

[T]he specific congenital malformations experienced by the affected children are not a signature pattern or ones that do not otherwise occur in the general population. Consideration of the cause of these congenital malformations in these children would require consideration of the following facts:

• First Trimester Incidents: X-rays, infections, trauma, smoking, alcohol and other drugs are known teratogens in the first-trimester. One of the patients experienced several potentially teratogenic complications during the first trimester of her index pregnancy, including x-ray imaging, pneumonia and a fall resulting a [sic] herniated disk (Ms. Gardner). Another patient had a mammogram early in her first trimester (Ms. Lovett). It is unclear from the medical records whether any of the class member affiants smoked or consumed other drugs during their index pregnancies. If they did, this would be another potential cause of the anomalies.

• Genetic Causes: Molecular testing in one patient confirms a genetic origin of malformation (Ms. Stymiest).

• Generic Citalopram: The records for two patients indicate they were prescribed generic citalopram rather than Celexa (Ms. Gardner and Ms. Bews), and the records for two of the other patients are inconclusive on whether they were prescribed Celexa or generic citalopram (Ms. Price and Ms. Stymiest).

[203] Consequently, the failure of the PCI to address the “first step” of general causation tips the balance of individual factors against a finding of preferable procedure.

[204] There are more factors that raise the complexity of individual issues.

[205] The proposed class definition includes women who spontaneously aborted a fetus with a congenital malformation. All such class members must establish the fact of a congenital malformation, since “the causes of action advanced are on behalf of ‘persons’ who had malformations from Celexa® exposure, whether they survived or died in utero” (as submitted by the plaintiffs in their factum).

[206] Dr. Bracken’s uncontested evidence is that “spontaneous abortions frequently occur for many reasons” and “the risk factors for spontaneous abortion include many which are more prevalent among depressed women: e.g., maternal smoking, alcohol use and stress”. Consequently, the issue of general causation for these potential class members raises even more complexity on individual issues.

[207] Further, to address the uniquely individual nature of the causation inquiry, the defendants would require, from each class member (as summarized in the defendants’ factum):

(i) personal and family medical records, together with pharmaceutical records,

(ii) evidence of the class member’s lifestyle and personal circumstances before, during and after citalopram treatment,

(iii) expert medical evidence regarding: (a) the prevalence, nature, etiology and risk factors for the malformations at issue, (b) the causal relationship between the malformations alleged and the gestational exposure stage and genetic profile of the patient, and (c) the causal relationship between the malformations alleged and the class member’s personal risk factors,

(iv) evidence of the class member and their prescribing physician(s) regarding: (a) when they began taking citalopram, (b) who they purchased it from, (c) the dosage and duration of their citalopram treatment, (d) their efficacy and tolerability responses to it, (e) the degree to which they followed the product monograph recommendations, (f) the details of any other pharmaceutical treatments prescribed to the class member, (g) their condition(s) for which citalopram was prescribed, (h) any informed consent discussions they held; and (i) what they would have done if “adequately” warned, and

(v) evidence of: (a) the class member’s pharmacist(s) regarding any information they may have given the class member, and (b) any other citalopram information reviewed by the class member (including the Consumer Information section of the PM, the Internet, etc.), their physician(s) and any other third parties who communicated such information to the class member (e.g., family members).

[208] The plaintiffs filed a litigation plan with the court entitled “Third Workable Method”. The only reference to the resolution of the individual issues is the following:

If the common issue is resolved in favour of the Class, in whole or in part, the Representative Plaintiff will propose an Individual Issues Resolution Procedure to the Defendants.

Class members will have the choice during the Individual Issues Resolution Procedure to retain counsel of their choosing. While some may choose to retain MLG to represent them in this aspect of the proceeding, they will not be required to do so. This is consistent with prior class action administration of individual issues in the settlement context.

A component of the Individual Issues Resolution Procedure shall be the preparation and submission of a Claims Form (Schedule “A” infra), which will provide the Defendants with critical information regarding the nature and scope of the claim being asserted, as well as medical and personal histories and identification and contact coordinates of relevant medical practitioners.

If the Defendants do not agree to that approach, the Court will determine the method in accordance with s. 25 of the CPA.

[209] Given the vast number of complex and idiosyncratic individual issues as discussed above, the proposed “claims form” procedure is inadequate for the extensive trial processes required for individual class members. The defects in the litigation plan underscore why a class action is not the preferable procedure.

[210] In this case, there is no reasonable plan to address individual issues. The plaintiffs suggest that test cases might be available if teratogenicity is addressed by organ group, embryonic stage, or individual congenital malformation. However, such an option assumes that general causation can be established for particular congenital malformations, when no such evidence is before the court.

[211] The present case lacks a workable way forward. It will break down into potentially thousands of individual trials, all with every liability and damages issue to be proven, including general causation of the particular congenital malformation, since the class does not purport to certify such an issue.

[212] For the above reasons, I find that the plaintiffs have not established a basis in fact that a class action would be the preferable procedure. As such, I dismiss the motion on the basis that the plaintiffs have failed to satisfy s. 5(1)(d). Consequently, I would not certify the class action even if the plaintiffs could satisfy s. 5(1)(c), which I do not find as I set out above.

Ancillary issues

[213] In their factum, the defendants raised ancillary issues that arise only if the court certifies the proposed class action. Since I dismiss the certification motion, I do not need to address these ancillary issues.

[214] However, if it is found that the proposed class action can be certified under s. 5(1), I briefly address the ancillary issues below.

Ancillary issue 1: Does the statement of claim disclose a cause of action against H. Lundbeck A/S?

[215] The defendants submit that the proposed class action cannot be certified against H. Lundbeck A/S since the claim discloses no cause of action. I do not agree.

[216] The settled law under s. 5(1)(a) was summarized by Perell J. in the First Certification Decision, at para. 87:

In a proposed class proceeding, in determining whether the pleading discloses a cause of action, no evidence is admissible, and the material facts pleaded are accepted as true, unless patently ridiculous or incapable of proof. The pleading is read generously, and it will be unsatisfactory only if it is plain, obvious, and beyond a reasonable doubt that the plaintiff cannot succeed. [Citations omitted.]

[217] Perell J. noted that the defendants challenged the inclusion of H. Lundbeck A/S. Perell J. stated that he did not need to consider the issue since he was dismissing the motion on other grounds: at para. 90.

[218] In the claim, the defendants are both H. Lundbeck A/S and Lundbeck Canada Inc. The plaintiffs define “Lundbeck” as “the defendants and their global predecessors”. The plaintiffs plead, at paras. 8 and 9 of the claim:

H. Lundbeck A/S and Lundbeck Canada Inc. are respectively incorporated under Danish and Canadian laws.

The Defendants are jointly, severally, and vicariously liable.

(a) From January 28th, 1999 as a global partnership centered in Denmark, Lundbeck manufactured and distributed Celexa® throughout the world, including in Canada.

(i) Lundbeck’s business was and is inextricably connected with the Defendants’.

(ii) Lundbeck had and has a common plan to manufacture and distribute Celexa® throughout the world, including in Canada, for profit.

(b) They are joint tortfeasors.

[219] The above pleading alleges that both defendants had a common plan to manufacture and distribute Celexa® throughout the world, including in Canada. The claim asserts that the two entities “knew or ought to have known that Celexa® is or may be a Teratogen”, based on knowledge both defendants received, as well as from scientific literature and epidemiological studies. The plaintiffs plead that the defendants (i) collectively owed the plaintiffs a duty of care to warn physicians or mothers that Celexa® is a teratogen and (ii) breached that duty by failing to provide such a warning.

[220] The defendants rely on the decisions in Martin and Norman v. Thunder Bay Regional Health Sciences Centre, 2015 ONSC 3252, in which the courts held that no cause of action was disclosed against the international parent corporation. However, the allegations in the present statement of claim can be distinguished from those in either Martin or Norman.

[221] In Martin, the plaintiffs pleaded that the Canadian defendant was “involved in and/or responsible for the sales, distribution and marketing of Seroquel in Canada”: at para. 111. As such, the plaintiffs’ broad allegation that “the Defendants designed, developed, tested, manufactured, distributed, marketed and sold Seroquel in Canada” (at para. 114) was inconsistent. In such circumstances, the court found that no cause of action was pleaded against the foreign parent corporation in a proposed class action since the pleadings (i) were inconsistent, (ii) lacked clarity as to each defendant’s role, and (iii) lumped the defendants together under the enterprise approach: at paras. 117-28.

[222] Similarly, in Norman, the court found no cause of action pleaded in a proposed class action against Sanofi Canada and Sanofi Pharma since the plaintiff had not pleaded any specific acts against them and instead “lumped [the defendants] together as one defined entity”: at paras. 56, 77, and 78.

[223] In the present case, however, the plaintiffs do not raise any inconsistent pleadings. It is possible (if the facts are accepted as true) that both corporate defendants engaged in a “global partnership centered in Denmark” in which they “manufactured and distributed Celexa® throughout the world, including in Canada”, with knowledge that Celexa® is or may be a teratogen and without warning physicians and women of that risk. On that basis, both defendants could be liable.

[224] For the above reasons, I find that the pleadings disclose a cause of action against H. Lundbeck A/S.

Ancillary issue 2: The direct claims brought by the children

[225] The defendants submit that the direct claims brought by children born with congenital malformations disclose no cause of action under s. 5(1)(a) and as such cannot be certified. Consequently, the defendants submit that the class definition under s. 5(1)(b) must be modified to exclude children as separate members of the class with direct claims.[^23] For the reasons that follow, I agree.

[226] The plaintiffs plead that the defendants owed a direct duty of care to children born with congenital malformations. This direct claim is based on the plaintiffs’ pleading that:

(i) The defendants are “strictly liable” for the children’s congenital malformations; and

(ii) The defendants “alternatively owed a duty of care” to the children for their congenital malformations.

[227] Both claims by the children are based on the plaintiffs’ allegation that the children “could not give informed consent to accept or reject Celexa®”.

[228] For the reasons that follow, neither the strict liability nor negligence claims brought directly by the children can stand.

(i) Strict liability

[229] The claim for strict liability to fetuses in utero must be struck as it is settled law in Canada that there is no strict liability in tort for defective products: Goodridge v. Pfizer Canada Inc., 2010 ONSC 1095, 101 O.R. (3d) 202, at paras. 102-04; Schick v. Boehringer Ingelheim (Canada) Ltd., 2011 ONSC 1942, at paras. 20-21; Palmer, at paras. 211-12.

(ii) Direct claims in negligence

[230] It is also settled law that direct claims by children for damage arising from lack of informed consent of the mother either before or after conception cannot be brought.

[231] To the extent the plaintiffs assert liability in negligence or statute to fetuses who die in utero, such claims must be struck as it is settled law that no common law or statutory duty is owed to them: Winnipeg Child and Family Services (Northwest Area) v. G. (D.F.), 1997 336 (SCC), [1997] 3 S.C.R. 925, at paras. 11, 13, 15-17, 22, 47; Musselman v. 875667 Ontario Inc. (c.o.b. Cities Bistro), 2010 ONSC 3177, 93 C.L.R. (3d) 58, at paras. 258-260, aff’d 2012 ONCA 41.

[232] Further, there is no recognized duty of care to a future child where, as here, the allegation is, at its core, a failure to provide the mother with fulsome information so that she could make an informed decision about whether to take the drug, regardless of whether the failure to provide informed consent arose before or after conception: Joshi v. Chada, 2022 ONSC 4910, at paras. 158-159, 166, citing Paxton v. Ramji, 2008 ONCA 697, 299 D.L.R. (4th) 614, at para. 76, leave to appeal refused, [2008] S.C.C.A. No. 508.

[233] The plaintiffs submit that the decision in Paxton was decided per incuriam, and that the law in Ontario is not settled. In Joshi, the court addressed similar arguments and held that the appellate law on this issue was settled. The same principles apply in the present case.

[234] Consequently, if the class action were to be certified, I would exclude any direct claims by children under s. 5(1)(a), and I would modify the class definition under s. 5(1)(b) to remove children as direct claimants. The children would still be able to bring indirect claims as “family members who may claim under Family Compensation Legislation”.

ORDER AND COSTS

[235] For the above reasons, I dismiss the motion for certification. Neither party sought costs of the motion. As such, I order no costs of the motion.

GLUSTEIN J.

Date: 20221220

[^1]: Price brings the action in her individual capacity. She also acts as litigation guardian for her son in his claim.

[^2]: Unless otherwise stated, all statutory references are to the CPA.

[^3]: “Congenital malformations” are also called congenital disorders, congenital anomalies, or birth defects. See Congenital Anomalies, online: World Health Organization <www.who.int/health-topics/congenital-anomalies#tab=tab_1>. In these Reasons, I use the term “congenital malformation” as both parties did so in their written and oral submissions.

[^4]: I adopt the definition of “general causation” and “specific causation” as stated by Perell J. in Palmer v. Teva Canada Ltd., 2022 ONSC 4690, at para. 65: general causation is defined as “whether or not an agent has the capacity to cause a disease or medical condition” while specific causation is defined as “whether or not an agent did cause a disease or medical condition to be suffered by a specific person”.

[^5]: At the hearing, the parties agreed that there were “hundreds and hundreds” of congenital malformations. Other health organizations refer to “thousands” of congenital malformations. See e.g. “Why do we have World Birth Defects Day” (March 1, 2017), online: March of Dimes <www.marchofdimes.org/find-support/blog/why-do-we-have-world-birth-defects-day; Glossary of Birth Defects Terms, online: Texas Health and Human Services </www.dshs.texas.gov/birthdefects/glossary.shtm>. The precise number is irrelevant to these Reasons. I refer to the number as being in the “hundreds” based on the submissions of counsel.

[^6]: Teratogenicity is the ability of an agent to have a teratogenic effect.

[^7]: I refer to this proposed common issue as the General Causation PCI.

[^8]: I refer to this proposed common issue as the Duty to Warn PCI.

[^9]: The defendants also raise some ancillary objections that would only be relevant if the court certified the proposed class action. Those objections relate to the cause of action pleaded against H. Lundbeck A/S and the direct claim brought by children for congenital malformations. I will address these issues briefly at the end of my Reasons, even though I dismiss the motion for certification.

[^10]: Idiopathic congenital malformations are those of uncertain or unknown cause.

[^11]: See footnote 5 above.

[^12]: Epidemiology is defined as “the study (scientific, systematic, and data-driven) of the distribution (frequency, pattern) and determinants (causes, risk factors) of health-related states and events (not just diseases) in specified populations”. See “What is Epidemiology?” (last reviewed June 17, 2016), online: Centers for Disease Control and Prevention <www.cdc.gov/careerpaths/k12teacherroadmap/epidemiology.html#:~:text=By%20definition%2C%20epidemiology%20is%20the,state%2C%20country%2C%20global>.

[^13]: There is no dispute that the defendant Lundbeck Canada Inc. is a proper defendant in relation to the claim. The defendants submit that there is no cause of action pleaded against the parent corporation, H. Lundbeck A/S, and ask the court to deny certification against H. Lundbeck A/S under s. 5(1)(a). While I do not accept the defendants’ position on that issue (as I discuss later in these Reasons), it is not a determinative issue since I deny certification on the basis that the plaintiffs have not met the test for certification under s. 5(1)(c) and (d).

[^14]: The proposed class action is brought only in relation to Celexa®, allegedly sold by the defendants, and not in relation to the production, marketing and sale of citalopram by generic pharmaceutical companies.

[^15]: The plaintiffs also filed an expert report from a U.S. attorney, Noble K. McIntyre, which was not relied upon by the parties at this hearing.

[^16]: The Bradford Hill criteria are a set of “viewpoints” or considerations that can be applied to evaluate epidemiologic evidence to determine if causation can be deduced.

[^17]: Canadian Network for Mood and Anxiety Treatments

[^18]: The process is discussed in detail in Martin v. Astrazeneca Pharmaceuticals PLC, 2012 ONSC 2744, at paras. 77, 80, 82, 84-85, 88-90, aff’d 2013 ONSC 1169 (Div. Ct.) and in Batten v. Boehringer Ingelheim (Canada) Ltd., 2017 ONSC 53, at paras. 81-90, aff’d 2017 ONSC 6098 (Div. Ct.), leave to appeal to Ont. C.A. refused, M48535 (28 February 2018).

[^19]: I adopt verbatim the summary of this evidence as set out by Perell J. at para. 6 of the First Certification Decision (as that term is defined at para. 58 below).

[^20]: See Price v. H. Lundbeck A/S, 2018 ONSC 4333.

[^21]: See Price v. H. Lundbeck A/S, 2020 ONSC 913 (Div. Ct.).

[^22]: I adopt the summary of these individual issues as set out verbatim by the defendants in their factum (footnotes omitted).

[^23]: The defendants acknowledge that the indirect claims by children under Family Compensation Legislation disclose a cause of action.