Arndt v. Smith, [1997] 2 SCR 539

Arndt v. Smith, [1997] 2 S.C.R. 539

Margaret Smith                                                                                 Appellant

v.

Carole Arndt and Dennis Jackson                                                    Respondent

Indexed as:  Arndt v. Smith

File No.:  24943.

1997:  January 29; 1997:  June 26.

Present:  Lamer C.J. and La Forest, L’Heureux‑Dubé, Sopinka, Gonthier, Cory, McLachlin, Iacobucci and Major JJ.

on appeal from the court of appeal for british columbia

Negligence ‑‑ Medical malpractice ‑‑ Causation ‑‑ Mother contracting chickenpox while pregnant ‑‑ Child born with disabilities resulting from mother’s chickenpox ‑‑ Mother suing doctor for costs associated with raising child ‑‑ Whether loss caused by doctor’s failure to advise properly of risk ‑‑ Appropriate test of causation.

A sued her physician S for costs associated with rearing her daughter, who was congenitally injured by chickenpox A had contracted during her pregnancy.  She contended that had S properly advised her of the risk of injury to her fetus, she would have terminated the pregnancy and avoided the costs she now incurs.  S contended that A would not have terminated the pregnancy even if she had been fully advised, and therefore asserted that the loss claimed was not caused by the failure to advise of risk.  The trial judge dismissed A’s claim.  Evaluating her testimony at trial that she would have had an abortion against the fact that she desired a child, that she was sceptical of “mainstream” medical intervention, that an abortion in the second trimester held increased risks and that an abortion would have required the approval of a committee on health grounds, the trial judge concluded that A would not, on a balance of probabilities, have aborted the pregnancy.  Also supportive of the trial judge’s conclusion was evidence that the risk of serious injury to the fetus was very small and medical advisers would have recommended against an abortion.  The Court of Appeal held that the trial judge had applied the wrong test and directed a new trial.

Held (Sopinka and Iacobucci JJ. dissenting):  The appeal should be allowed.

A’s claims for economic loss are not barred by s. 3(1)(a) of the British Columbia Limitation Act, which provides a two‑year limitation period in actions for “damages in respect of injury to person or property, including economic loss arising from the injury”, because they do not relate to personal injury either to A or to her daughter.

Per Lamer C.J. and La Forest, L’Heureux‑Dubé, Gonthier, Cory and Major JJ.:  When determining whether the loss claimed by A was caused by S’s failure to advise of the risk the court should adopt the modified objective test set out in Reibl v. Hughes.  The test enunciated relies on a combination of objective and subjective factors in order to determine whether the failure to disclose actually caused the harm of which the plaintiff complains.  It requires that the court consider what the reasonable patient in the plaintiff’s circumstances would have done if faced with the same situation.  The trier of fact must take into consideration any “particular concerns” of the patient and any “special considerations affecting the particular patient” in determining whether the patient would have refused treatment if given all the information about the possible risks.  The “reasonable person” who sets the standard for the objective test must be taken to possess the patient’s reasonable beliefs, fears, desires and expectations.  While evidence of reasonable fears and concerns can thus be taken into account, purely subjective fears which are not related to the material risks should not be considered.  The modified objective test strikes a reasonable balance, which cannot be obtained through either a purely objective or a purely subjective approach.  A purely subjective approach fails to take into account the inherent unreliability of the self‑serving assertion of a plaintiff, while the purely objective standard might result in undue emphasis being placed on the medical evidence, essentially resulting in a test which defers completely to medical wisdom.

It is appropriate to infer from the evidence in this case that a reasonable person in the plaintiff’s position would not have decided to terminate her pregnancy in the face of the very small increased risk to the fetus posed by her exposure to the virus which causes chickenpox.  While A did make a very general inquiry concerning the risks associated with maternal chickenpox, there was nothing to indicate to the doctor that she had a particular concern in this regard.  Further, factors such as A’s desire for children and her suspicion of the mainstream medical profession can be taken into consideration when determining what a reasonable person in her position would have done if informed of the risks.  As found by the trial judge, the failure to disclose some of the risks to the fetus associated with maternal chickenpox did not affect A’s decision to continue the pregnancy to term.  It follows that the failure to disclose did not cause the financial losses for which she is seeking compensation.

Per McLachlin J.:  This case raises the hypothetical question of what the plaintiff would have done had the physician discharged her duty to disclose.  General tort principles suggest that this question is a purely factual inquiry to be answered by reference to all the evidence.  This may include evidence from the plaintiff at trial as to what she would have done, but it also includes relevant evidence of her situation, circumstances and mind‑set at the time the decision would have been made.  The trial judge must look at all the evidence and determine whether the plaintiff would have taken the suggested course on a balance of probabilities.  The approach suggested by the fundamental principles of tort law is subjective, in that it requires consideration of what the plaintiff at bar would have done.  However, it incorporates elements of objectivity; the plaintiff’s subjective belief at trial that she would have followed a certain course stands to be tested by her circumstances and attitudes at the time the decision would have been made as well as the medical advice she would have received at the time.

The trial judge applied the right test in this case and did not err in dismissing the action on the ground that A had failed to establish that S’s failure to advise her of the risk to her fetus from chickenpox caused the loss associated with the rearing of her disabled daughter.  While he paid lip service to the reasonable person test, he resolved the issue of causation by asking what the plaintiff herself would have done in all the circumstances.  He considered a number of external circumstances, including the fact that an abortion in the second trimester would have been riskier than an earlier abortion and the need to obtain the approval of a committee for any abortion.  He went on to consider factors going to A’s state of mind, including the fact that she very much wanted to have a child, that she had a natural skepticism of mainstream medicine and had so informed S and that she stated she did not want an ultrasound scan of the developing fetus.  The fact that the medical profession would not have recommended an abortion in these circumstances also supports the trial judge’s conclusion that A would not have chosen an abortion had she been advised of the very small increased risk to the fetus posed by chickenpox.

Per Sopinka and Iacobucci JJ. (dissenting):  As found by McLachlin J., the appropriate test of causation in the present circumstances is not to ask what the “reasonable person” would have done in A’s position, but is to ask what A herself would have done had she been fully apprised of the risks to the fetus resulting from her chickenpox.  The trial judge did not apply the appropriate test, however, but applied an objective test.  He repeatedly referred to the pivotal role played in his causation analysis by the “reasonable and prudent expectant mother” and did not consider A’s evidence  that she would have aborted the fetus if fully informed of the relevant risks because of his view of the objective nature of the causation test.  At best, the trial judge’s reasons are unclear on the question of which test for causation was applied.  This lack of clarity itself would warrant a new trial.  In the further alternative, even if the trial judge did purport to apply a nominally subjective test, the failure to consider A’s testimony would constitute an error of law justifying a new trial. 

Cases Cited

By Cory J.

Applied:  Reibl v. Hughes, [1980] 2 S.C.R. 880; referred to:  Hollis v. Dow Corning Corp., [1995] 4 S.C.R. 634.

By McLachlin J.

Considered:  Reibl v. Hughes, [1980] 2 S.C.R. 880; referred to:  Norberg v. Wynrib, [1992] 2 S.C.R. 226; Hollis v. Dow Corning Corp., [1995] 4 S.C.R. 634; Laferrière v. Lawson, [1991] 1 S.C.R. 541; Chatterton v. Gerson, [1981] 1 Q.B. 432; Hills v. Potter, [1983] 3 All E.R. 716; Sidaway v. Bethlem Royal Hospital Governors, [1985] 1 All E.R. 643; Ellis v. Wallsend District Hospital (1989), 17 N.S.W.L.R. 553; Canterbury v. Spence, 464 F.2d 772 (1972); Young v. Northern Territory of Australia (1992), 107 F.L.R. 264; Bernard v. Char, 903 P.2d 667 (1995); Scott v. Bradford, 606 P.2d 554 (1979); Smith v. Reisig, 686 P.2d 285 (1984); Arena v. Gingrich, 733 P.2d 75 (1987); Cobbs v. Grant, 502 P.2d 1 (1972); Martin v. Insurance Corp. of British Columbia (1979), 13 B.C.L.R. 163.

Statutes and Regulations Cited

Limitation Act, R.S.B.C. 1979, c. 236, s. 3(1)(a).

Authors Cited

Fleming, John G.  The Law of Torts, 6th ed.  Sydney:  Law Book Co., 1983.

Fontigny, Nadine.  “When Yes Really Means Yes:  The Law of Informed Consent in Canada Revisited” (1996), 4 Health L. Rev. 17.

“Informed Consent ‑‑ A Proposed Standard for Medical Disclosure” (1973), 48 N.Y.U.  L. Rev. 548.

Osborne, Philip H.  Annotation to Arndt v. Smith (1995), 25 C.C.L.T. (2d) 264.

Osborne, Philip H.  “Causation and the Emerging Canadian Doctrine of Informed Consent to Medical Treatment” (1985), 33 C.C.L.T. 131.

Robertson, Gerald.  “Informed Consent Ten Years Later:  The Impact of Reibl v. Hughes” (1991), 70 Can. Bar Rev. 423.

APPEAL from a judgment of the British Columbia Court of Appeal (1995), 6 B.C.L.R. (3d) 201, 126 D.L.R. (4th) 705, [1995] 7 W.W.R. 378, 61 B.C.A.C. 57, 100 W.A.C. 57, 25 C.C.L.T. (2d) 262, reversing a decision of the British Columbia Supreme Court (1994), 93 B.C.L.R. (2d) 220, [1994] 8 W.W.R. 568, 21 C.C.L.T. (2d) 66, dismissing the respondents’ claim against the appellant physician.  Appeal allowed, Sopinka and Iacobucci JJ. dissenting.

Christopher E. Hinkson, Q.C., and Andrew F. Wilkinson, for the appellant.

Nathan H. Smith, for the respondents.

The judgment of Lamer C.J. and La Forest, L’Heureux-Dubé, Gonthier, Cory and Major JJ. was delivered by

//Cory J.//

1                                   Cory J. -- I have had the benefit of reading the cogent reasons of McLachlin J. I agree with her analysis regarding the Limitation Act issue and with her disposition of this appeal.  However, I come to a different conclusion as to the test which should be applied when determining whether the loss claimed by the plaintiff was caused by the doctor’s failure to advise of the risk.

2                                   The starting point for this question must be Reibl v. Hughes, [1980] 2 S.C.R. 880, which set out the basic principles for assessing causation in cases involving allegations of negligence by doctors.  Reibl involved an action by a patient against a surgeon for failing to warn him of the risk of paralysis associated with  the elective surgery performed by that surgeon.  One of the defences raised was that even if the surgeon had disclosed all of the risks of the procedure, the plaintiff would nonetheless have gone ahead with the operation.  In other words, the physician disputed whether his negligent failure to disclose had, in fact, caused the plaintiff’s loss.

3                                   The question presented to the Court was how to determine whether the patient would have actually chosen to decline the surgery if he had been properly informed of the risks.  In trying to craft the appropriate test, Laskin C.J. for a unanimous Court quoted with approval an article from the New York University Law Review, entitled “Informed Consent ‑‑ A Proposed Standard for Medical Disclosure” (1973), 48 N.Y.U. L. Rev. 548.  The article distinguished between a subjective test, which asks whether the particular patient would have foregone treatment if properly informed, and an objective test, which asks whether the average prudent person in the patient’s position would have foregone treatment if informed of all material risks.  The authors preferred the objective test, since the subjective standard suffered from what they deemed to be a “gross defect”:  “[I]t depends on the plaintiff’s testimony as to his state of mind, thereby exposing the physician to the patient’s hindsight and bitterness” (p. 550).

4                                   Laskin C.J. shared the authors’ concerns about the subjective test, and rejected the pure subjective approach to causation.  He explained at p. 898 that the plaintiff’s testimony as to what he or she would have done, had the doctor given an adequate warning, is of little value:

It could hardly be expected that the patient who is suing would admit that he would have agreed to have the surgery, even knowing all the accompanying risks.  His suit would indicate that, having suffered serious disablement because of the surgery, he is convinced that he would not have permitted it if there had been proper disclosure of the risks, balanced by the risks of refusing the surgery.  Yet, to apply a subjective test to causation would, correlatively, put a premium on hindsight, even more of a premium than would be put on medical evidence in assessing causation by an objective standard.

In other words, the plaintiff would always testify that the failure to warn was the determining factor in his or her decision to take the harmful course of action.  Accordingly the subjective test would necessarily cause the trier of fact to place too much weight on inherently unreliable testimony.

5                                   While an objective test would prevent an inappropriate emphasis being placed on the plaintiff’s testimony, Laskin C.J. thought that a purely objective test also presented problems.  At p. 898, he discussed his paramount concern with an approach based on the actions of a hypothetical reasonable person:

. . . a vexing problem raised by the objective standard is whether causation could ever be established if the surgeon has recommended surgery which is warranted by the patient’s condition.  Can it be said that a reasonable person in the patient’s position, to whom proper disclosure of attendant risks has been made, would decide against the surgery, that is, against the surgeon’s recommendation that it be undergone?  The objective standard of what a reasonable person in the patient’s position would do would seem to put a premium on the surgeon’s assessment of the relative need for the surgery and on supporting medical evidence of that need.  Could it be reasonably refused?

In short, the purely objective standard might result in undue emphasis being placed on the medical evidence, essentially resulting in a test which defers completely to medical wisdom.

6                                   To balance the two problems, Laskin C.J. opted for a modified objective test for causation, which he set out at length at pp. 898‑900 :

I think it is the safer course on the issue of causation to consider objectively how far the balance in the risks of surgery or no surgery is in favour of undergoing surgery.  The failure of proper disclosure pro and con becomes therefore very material.  And so too are any special considerations affecting the particular patient.  For example, the patient may have asked specific questions which were either brushed aside or were not fully answered or were answered wrongly.  In the present case, the anticipation of a full pension would be a special consideration, and, while it would have to be viewed objectively, it emerges from the patient’s particular circumstances.  So too, other aspects of the objective standard would have to be geared to what the average prudent person, the reasonable person in the patient’s particular position, would agree to or not agree to, if all material and special risks of going ahead with the surgery or foregoing it were made known to him.  Far from making the patient’s own testimony irrelevant, it is essential to his case that he put his own position forward.

The adoption of an objective standard does not mean that the issue of causation is completely in the hands of the surgeon. Merely because medical evidence establishes the reasonableness of a recommended operation does not mean that a reasonable person in the patient’s position would necessarily agree to it, if proper disclosure had been made of the risks attendant upon it, balanced by those against it.  The patient’s particular situation and the degree to which the risks of surgery or no surgery are balanced would reduce the force, on an objective appraisal, of the surgeon’s recommendation.  Admittedly, if the risk of foregoing the surgery would be considerably graver to a patient than the risks attendant upon it, the objective standard would favour exoneration of the surgeon who has not made the required disclosure.  Since liability rests only in negligence, in a failure to disclose material risks, the issue of causation would be in the patient’s hands on a subjective test, and would, if his evidence was accepted, result inevitably in liability unless, of course, there was a finding that there was no breach of the duty of disclosure.  In my view, therefore, the objective standard is the preferable one on the issue of causation.

In saying that the test is based on the decision that a reasonable person in the patient’s position would have made, I should make it clear that the patient’s particular concerns must also be reasonably based; otherwise, there would be more subjectivity than would be warranted under an objective test.  Thus, for example, fears which are not related to the material risks which should have been but were not disclosed would not be causative factors.  However, economic considerations could reasonably go to causation where, for example, the loss of an eye as a result of non‑disclosure of a material risk brings about the loss of a job for which good eyesight is required.  In short, although account must be taken of a patient’s particular position, a position which will vary with the patient, it must be objectively assessed in terms of reasonableness.

These words are as persuasive today as they were when they were written.  The test enunciated relies on a combination of objective and subjective factors in order to determine whether the failure to disclose actually caused the harm of which the plaintiff complains.  It requires that the court consider what the reasonable patient in the circumstances of the plaintiff would have done if faced with the same situation.  The trier of fact must take into consideration any “particular concerns” of the patient and any “special considerations affecting the particular patient” in determining whether the patient would have refused treatment if given all the information about the possible risks.

7                                   This Court recently had occasion to reconsider the modified objective test in Hollis v. Dow Corning Corp., [1995] 4 S.C.R. 634.  La Forest J. for the majority held that a subjective test was appropriate for an action against a manufacturer of breast implants when determining whether the failure to warn of the risks associated with the implants caused the harm.  However, he specifically supported the continuing application of the modified objective standard in negligence actions between a doctor and patient.  He emphasized at p. 675 the unique policy concerns associated with the doctor‑patient relationship, which justify the modification of the usual approach to causation followed in other tortious actions.

. . . the duty of the doctor is to give the best medical advice and service he or she can give to a particular patient in a specific context.  It is by no means coterminous with that of the manufacturer of products used in rendering that service.  The manufacturer, on the other hand, can be expected to act in a more self‑interested manner.  In the case of a manufacturer, therefore, there is a greater likelihood that the value of a product will be overemphasized and the risk underemphasized.  It is, therefore, highly desirable from a policy perspective to hold the manufacturer to a strict standard of warning consumers of dangerous side effects to these products.

To elaborate on the distinction between manufacturers and doctors, I believe it is important to note that negligence actions against members of the medical profession based on a failure to warn will inevitably be hypothetical, because they are based on constructing what would have happened if the patient had been fully informed of the risks of a procedure.  This introduces a degree of uncertainty into the analysis.  Often, this uncertainty will be increased by the difficulty of determining the extent of a doctor’s obligation to inform in a case where, based upon his or her professional knowledge and experience, the doctor believed that the risk was too insignificant to warrant advising the patient of it.  On the other hand, pharmaceutical manufacturers have no reason not to provide the medical profession at least, if not the public generally, with all available information concerning the medication they put on the market.  It follows that it is eminently sensible to apply a more flexible standard of causation to doctors than to manufacturers.

8                                   Sopinka J. (McLachlin J. concurring) wrote a dissenting opinion in Hollis, but he joined with the majority in endorsing the modified objective test from Reibl.  He also elaborated upon the advantages of this approach to causation.  A subjective approach, he explained, fails to take into account the inherent unreliability of the self‑serving assertion of a plaintiff.  The plaintiff may honestly believe that he would not have consented to a procedure if all the risks were disclosed.  However, this is only the plaintiff’s opinion about what he would have done in a situation which never arose.  As such, the opinion may be honestly held and given, but rejected by the trier of fact.  As Sopinka J. stated at pp. 688‑89:

In evaluating the opinion, the trier of fact must discount its probity not only by reason of its self‑serving nature, but also by reason of the fact that it is likely to be coloured by the trauma occasioned by the failed procedure.  For this reason, the most reliable approach in determining what would in fact have occurred is to test the plaintiff’s assertion by reference to objective evidence as to what a reasonable person would have done. [Emphasis in original.]

At p. 690, Sopinka J. stated his preference for the modified objective standard:

The Reibl approach is a more reliable method of determining what [the plaintiff’s] choice would have been.  The subjective test places too much of a premium on the plaintiff’s present belief as to what it would have been.

In the end, therefore, both the majority and minority judgments explicitly endorsed the continued application of the modified objective test from Reibl for negligence actions by a patient against a doctor.  The decision in Hollis is a very strong and recent affirmation of the Reibl test and should not be lightly disregarded.

9                                   Some of the criticisms directed at the Reibl test may stem from confusion as to what Laskin C.J. intended in his adoption of a modified objective test.  The uncertainty surrounds the basic premise that the test depends upon the actions of a reasonable person in the plaintiff’s circumstances.  Which aspects of the plaintiff’s personal circumstances should be attributed to the reasonable person?  There is no doubt that objectively ascertainable circumstances, such as a plaintiff’s age, income, marital status, and other factors, should be taken into consideration.  However, Laskin C.J. didn’t stop there.  He went on and stated that “special considerations” affecting the particular patient should be considered, as should any “specific questions” asked of the physician by the patient. In my view this means that the “reasonable person” who sets the standard for the objective test must be taken to possess the patient’s  reasonable beliefs, fears, desires and expectations.  Further, the patient’s expectations and concerns will usually be revealed by the questions posed. Certainly, they will indicate the specific concerns of the particular patient at the time consent was given to a proposed course of treatment.  The questions, by revealing the patient’s concerns, will provide an indication of the patient’s state of mind, which can be relevant in considering and applying the modified objective test.

10                               An example may serve to illustrate this.  Imagine a patient considering plastic surgery on his nose.  During a pre‑operative consultation, the patient asks if the surgery will affect his sense of smell.  The physician fails to fairly and adequately explain the attendant risks to this sensory function and does not mention that a certain percentage of patients suffer a permanent loss of a small fraction of their ability to smell.  After the surgery, the patient can no longer smell with the same acuity food that is cooking.  Under Laskin C.J.’s test in Reibl, the patient’s question about the risks to his sense of smell are clearly relevant.  The question posed suggests that the patient had a special concern about losing the sense of smell.  This is not an unreasonable concern.  The loss of a keen sensory perception of smell which is so closely related to the sense of taste is crucial to both those who artistically prepare and those who have a particular appreciation for finely prepared food.  This special fear of the loss of a keen sense of smell could be considered by the trier of fact in determining whether the reasonable person with the particular expressed concern of the plaintiff would have consented to the proposed course of treatment if all the risks had been disclosed.

11                               As another example, let us consider a patient who asks his doctor about a proposed procedure, and particularly poses questions as to whether there might be any effect on his hearing, without advising the doctor of his particular passion for the singing of operatic sopranos.  If the doctor fails to inform the patient of the possibility that the procedure could limit his ability to hear in the upper ranges, the Reibl test would allow the trier of fact to consider the questions posed by the patient in determining whether he would have consented to the proposed treatment if he had been properly informed of all the risks.  Again, the questions asked by this patient may act as an indication of his own reasonable fears and concerns, which are appropriate modifiers of the hypothetical reasonable person.

12                               As further evidence that the patient’s state of mind is relevant to the Reibl test, Laskin C.J. goes on at pp. 899‑900 to caution that the trier of fact may only take into account those particular concerns of the patient which are reasonable:

[T]he patient’s particular concerns must also be reasonably based. . . .  Thus, for example, fears which are not related to the material risks which should have been but were not disclosed would not be causative factors.

Clearly, evidence of reasonable fears and concerns can be taken into consideration and this is evidence which could go to establishing the plaintiff’s subjective state of mind.  Therefore, it is apparent that Laskin C.J. intended that the reasonable subjective beliefs of the patient should be attributed to the hypothetical reasonable person used to set  the objective standard in order to properly reflect the circumstances of the plaintiff.

13                               If the patient’s fears and beliefs were not considered when assessing how the “reasonable person in the patient’s position” would have responded had all risks of a procedure been disclosed, absurd verdicts could be produced.  For example, let us suppose that a plaintiff brought an action based on her doctor’s failure to disclose that there was a very significant risk of her giving birth to a disabled child, that the risk was material and the only issue was causation.  If the plaintiff’s beliefs are not to be considered, the trier of fact could conclude that a reasonable person in the position of the plaintiff would have chosen to terminate the pregnancy and find in favour of the patient even if the plaintiff was so resolutely and unalterably opposed to abortion that she would never have terminated the pregnancy.  The failure to disclose would not have been the actual cause of the harm.  Despite this, under the purely objective standard, the plaintiff could recover.  This example demonstrates why it is important to include some subjective aspects in the assessment of what the reasonable person in the position of the plaintiff would have done if all the risks had been disclosed.

14                               Laskin C.J. carefully noted that purely subjective fears which are not related to the material risks should not be taken into account in applying the modified objective test.  In other words, fears which are idiosyncratic, which do not relate directly to the material risks of a proposed treatment and which would often be unknown to a physician, cannot be considered. This is what ensures that the objective standard truly is based on the actions of a “reasonable person”.  It means that a doctor will not be held responsible for damages attributable to a plaintiff’s idiosyncrasies.  It ensures that a plaintiff would not be able to successfully prove causation simply by demonstrating an irrational fear which, had the physician disclosed all the risks, would have convinced the plaintiff to forego medical treatment.  For example, if a doctor failed to tell the patient that one of the risks of a procedure was an allergic reaction which could cause a temporary red rash on the skin, and the patient had an irrational belief that a rash is a highly significant and dangerous sign of evil spirits in the body, the patient could not successfully prove causation by demonstrating that he would not have proceeded with the treatment on the basis of this irrational fear.

15                               Reibl is a very significant and leading authority.  It marks the rejection of the paternalistic approach to determining how much information should be given to patients.  It emphasizes the patient’s right to know and ensures that patients will have the benefit of a high standard of disclosure.  At the same time, its modified objective test for causation ensures that our medical system will have some protection in the face of liability claims from patients influenced by unreasonable fears and beliefs, while still accommodating all the reasonable individual concerns and circumstances of plaintiffs.  The test is flexible enough to enable a court to take into account a wide range of the personal circumstances of the plaintiff, and at the same time to recognize that physicians should not be held responsible when the idiosyncratic beliefs of their patients might have prompted unpredictable and unreasonable treatment decisions.

16                               The Reibl test has had the desired effect of ensuring that patients have all the requisite information to make an informed decision regarding the medical procedure they are contemplating.  Members of the medical and legal professions are familiar with its requirements.  It strikes a reasonable balance, which cannot be obtained through either a purely objective or a purely subjective approach.  A purely subjective test could serve as an incitement for a disappointed patient to bring an action.  The plaintiff will invariably state with all the confidence of hindsight and with all the enthusiasm of one contemplating an award of damages that consent would never have been given if the disclosure required by an idiosyncratic belief had been made.  This would create an unfairness that cannot be accepted.  It would bring inequitable  and unnecessary pressure to bear upon the overburdened medical profession.  On the other hand, a purely objective test which would set the standard by a reasonable person without the reasonable fears, concerns and circumstances of the particular plaintiff would unduly favour the medical profession.

17                               It has been said that a subjective test, despite its dangers, is the most logical.  Yet pure logic cannot achieve the fairness attained by the application of Reibl test.  It is said that there is nothing to distinguish between the subjective test and the modified objective test.  If that were the case there could be no grounds for complaint from those who favour the subjective test.  Yet in my view there is a very real distinction.  The modified objective test serves to eliminate from consideration the honestly held but idiosyncratic and unreasonable or irrational beliefs of patients.  The Reibl test is fair and has recently been approved in the Hollis case.  No useful purpose would be served by changing it.  Indeed, to do so may unnecessarily add to the high cost of providing medical care.  In short, I see no reason to abandon the modified objective test to causation set down  in Reibl, a test which asks whether a reasonable person in the circumstances of the plaintiff would have consented to the proposed treatment if all the risks had been disclosed.

18                               Turning now to this appeal, it is appropriate to infer from the evidence that a reasonable person in the plaintiff’s position would not have decided to terminate her pregnancy in the face of the very small increased risk to the fetus posed by her exposure to the virus which causes chickenpox.  Ms. Arndt did make a very general inquiry concerning the risks associated with maternal chickenpox.  However, it should not be forgotten that the risk was indeed very small.  In the absence of a specific and clearly expressed concern, there was nothing to indicate to the doctor that she had a particular concern in this regard.  It follows that there was nothing disclosed by Ms. Arndt’s question which could be used by the trier of fact as an indication of a particular fear regarding the possibility of giving birth to a disabled child which should be attributed to the hypothetical reasonable person in the patient’s situation.  Further, factors such as the plaintiff’s desire for children and her suspicion of the mainstream medical profession can be taken into consideration when determining what a reasonable person in the plaintiff’s position would have done if informed of the risks.  It is not necessary to assess the relative importance these beliefs would have in the determination of the question of causation.  It is sufficient to observe that all these are factors indicating the state of mind of the plaintiff at the time she would have had to make the decision, and therefore may be properly considered by the trier of fact.  I agree with the trial judge that the failure to disclose some of the risks to the fetus associated with maternal chickenpox did not affect the plaintiff’s decision to continue the pregnancy to term.  It follows that the failure to disclose did not cause the financial losses for which the plaintiff is seeking compensation.

19                               I would allow this appeal, set aside the judgment of the Court of Appeal  and reinstate the judgment of the trial judge.  The defendant should have her costs of the proceedings in this Court and the courts below.

The following are the reasons delivered by

//Sopinka and Iacobucci JJ.//

20                               Sopinka and Iacobucci JJ. (dissenting) -- We have read the reasons of our colleague McLachlin J.    We agree with her analysis of the Limitation Act issue and also agree with much of her analysis on causation.  In particular, we agree that the appropriate test of causation in the present circumstances is not to ask what the “reasonable person” would have done in the position of the plaintiff Ms. Arndt, but rather the appropriate test is to ask what Ms. Arndt herself would have done had she been fully apprised of the risks to the fetus resulting from her chickenpox.  We respectfully disagree, however, with McLachlin J.’s proposed disposition of the present appeal.

21                               McLachlin J. concludes that the trial judge in the instant case applied the appropriate test of causation by asking himself whether the particular plaintiff at bar would have carried the fetus to term had she been fully apprised of the risks.  She states at para. 34 that while the trial judge “allud[ed] to a ‘reasonable person’ test, [he] asked himself whether, on all the evidence, the plaintiff would have chosen to abort her pregnancy had she been advised of the risk of injury to her fetus from her chickenpox.”  She further states at para. 69 that the trial judge simply “paid lip service to the reasonable person test”.  We cannot agree with her conclusion in this regard. 

22                               In our view, the trial judge applied an objective test.  In the alternative, at best the reasons for judgment are unclear as to which test was applied.  In the further alternative, even if the trial judge did purport to apply a nominally subjective test, in doing so, he wrongly disregarded the testimony of the plaintiff, on the basis that it could play no role in his causation analysis.  The presence of any one of these alternatives necessitates a new trial.

23                               Certain portions of the judgment are consistent with the application of a subjective test.  For example, as McLachlin J. points out, the trial judge stated ((1994), 93 B.C.L.R. (2d) 220, at p. 223):

In the case at bar, I find Ms. Arndt would have chosen in foresight to carry Miranda to full term had she been fully cognizant of all the risks and the chance of such risks occurring.

24                               However, in our opinion, other parts of the trial judge’s reasons indicate that the causation test applied was not subjective, as McLachlin J. maintains, but rather objective.  On more than one occasion, the trial judge emphasized the need to determine causation according to an objective standard.  For example, he said (at p. 222):

If a physician fails to warn his or her patient of all material risks, the issue of whether that patient would have requested a therapeutic abortion had she been advised of those risks must be determined objectively by the trier of fact after taking into account the patient and her particular circumstances. [Emphasis added.]

And, later in his reasons (at p. 230):

[W]here a pregnant woman has not been fully warned of all risks faced by her fetus, and then makes an uninformed consent regarding abortion, the law dictates that trial judges must determine, on an objective basis, whether or not a “reasonable and prudent” expectant mother would have undergone a therapeutic abortion had she been adequately warned. [Emphasis added.]

25                               The most telling argument against the view that the trial judge applied a subjective test is his treatment of the evidence of the plaintiff.  Under the subjective test, the function of the objective evidence is to test the reliability of the plaintiff’s assertion as to what her conduct would have been if properly warned.  The plaintiff’s evidence is crucial in this regard.  While the plaintiff’s evidence must be evaluated by reference to its reasonableness, in the final analysis it is the plaintiff’s evidence that is either accepted or rejected.  In the instant case, having adverted to the plaintiff’s testimony that she would have aborted the fetus if fully informed of the relevant risks, the trial judge goes on to imply that this testimony is not relevant.  He said (at p. 235):

Given my finding of Dr. Smith’s negligence, I must now determine if Ms. Arndt would have aborted her fetus had she received sufficient information.  Ms. Arndt testified that she would have aborted her pregnancy if she had been told of “any risk of abnormalities at all”.  I must, however, determine this issue objectively, replacing the actual plaintiff with the “reasonable and prudent” expectant mother. [Emphasis added.]

26                               Thus, the trial judge repeatedly referred to the pivotal role played in his causation analysis by the “reasonable and prudent expectant mother” and did not consider the plaintiff’s evidence because of his view of the objective nature of the causation test.  Given these clear statements by the trial judge, we cannot agree with McLachlin J.’s conclusion that, in fact, the trial judge applied a subjective test.  On the contrary, we would conclude that the trial judge purported in substance to apply an objective test.  At best, the trial judge’s reasons are unclear on the question of which test for causation was applied.  This lack of clarity itself would warrant a new trial.

27                               Even if it is held that the trial judge did apply a subjective test, the failure to consider the testimony of the plaintiff would constitute an error of law.  As set out above, while disregard for the plaintiff’s testimony may be consistent with an objective test for causation, it is not consistent with the subjective test, a test which seeks to ascertain how this plaintiff would have acted if fully informed of the relevant risks.  Of course, a trial judge need not accept the plaintiff’s evidence, but he or she must at least consider it in determining whether this particular plaintiff would have acted differently had the negligence not occurred.  Therefore, in our view, even if the trial judge applied a subjective test, he erred in refusing to consider Ms. Arndt’s testimony.

28                               To illustrate the importance of testimony in this context, suppose a hypothetical plaintiff is negligently ill-informed of the risks of a particular medical treatment and the treatment goes awry, causing injury to the plaintiff.  Suppose further that there is evidence that almost every patient in the plaintiff’s situation would not take this treatment if fully informed.  While other evidence suggests that the plaintiff should succeed, if the plaintiff testifies that she would have gone ahead with the treatment even if informed and the trial judge accepts this testimony, the plaintiff cannot succeed.  Conversely, if there is evidence indicating that almost every patient would have taken the treatment despite the risks, but the plaintiff testifies that she would not have done so and the trial judge accepts this testimony, the plaintiff should succeed.  Testimonial evidence, while tested against other available evidence, may be pivotal in each case. 

29                               In our view, the trial judge in the present case, by refusing to consider the plaintiff’s testimony, misdirected himself and the judgment he reached therefore cannot stand.  As an additional reason for this conclusion, the trial judge considered factors such as the plaintiff’s preference for homeopathic medicine and the fact that the pregnancy was planned which, in our view, are of dubious assistance in determining what the plaintiff would have done had she been fully informed.  It is not at all inconsistent either with a planned pregnancy or with a preference for homeopathic medicine for the plaintiff to claim that she would have had an abortion had she been fully informed of the risks.  By basing his conclusion in part on these factors, in our view the trial judge cast further doubt on his finding on causation.

30                               Because of the flaws in the trial judge’s consideration of causation, in our view his  judgment cannot stand.  We would dismiss the appeal and order a new trial on causation, applying the test set out in McLachlin J.’s reasons.

The following are the reasons delivered by

//McLachlin J.//

McLachlin J. --

                                                                        I.

31                               The plaintiff Arndt sues for costs associated with rearing her daughter, who was congenitally injured by chickenpox.  The plaintiff contends that had her physician properly advised her of the risk of injury to her fetus, she would have aborted the pregnancy and avoided the costs she now incurs.  The defendant Smith contends that the plaintiff  would not have aborted the pregnancy even if she had been fully advised, and therefore asserts that the loss claimed was not caused by the failure to advise of risk.  The issue on appeal is what legal principles a trial judge, facing an issue such as this, should apply in determining whether the loss claimed was caused by the failure to advise of the risk.

32                               I conclude that the test for causation is what, on a balance of probabilities,  the particular plaintiff at bar would have done having regard to all the evidence bearing on the issue, including the medical recommendations she would have received at the time.

                                                                       II.

33                               It is a fundamental rule of tort law that the plaintiff must prove two things.  First, the plaintiff must prove that the defendant breached a duty owed to the plaintiff.  Second, the plaintiff must prove that the breach caused the loss for which the plaintiff claims damages.  The trial judge found the plaintiff had satisfied the first requirement  of proving breach, but had failed to establish that she would have had an abortion had she been properly advised.  He dismissed her claim on the ground that she had failed to establish that the breach had caused the loss:  (1994), 93 B.C.L.R. (2d) 220, [1994] 8 W.W.R. 568, 21 C.C.L.T. (2d) 66 (hereinafter cited to B.C.L.R.).

34                               The trial judge, although alluding to a “reasonable person” test, asked himself whether, on all the evidence, the plaintiff would have chosen to abort her pregnancy had she been advised of the risk of injury to her fetus from her chickenpox.  Evaluating her testimony at trial that she would have had an abortion against the fact that she desired a child, that she was sceptical of “mainstream” medical intervention, that an abortion in the second trimester held increased risks and that an abortion would have required the approval of a committee on health grounds, the trial judge concluded that the plaintiff would not, on a balance of probabilities, have aborted the pregnancy.  Also supportive of the trial judge’s conclusion was evidence that the risk of serious injury to the fetus was very small and medical advisors would have recommended against an abortion.

35                               The Court of Appeal held that the trial judge applied the wrong test and directed a new trial: (1995), 6 B.C.L.R. (3d) 201, 126 D.L.R. (4th) 705, [1995] 7 W.W.R. 378, 61 B.C.A.C. 57, 100 W.A.C. 57, 25 C.C.L.T. (2d) 262.  Lambert and Wood JJ.A. held that the law required the judge to determine not what this plaintiff would have done, but what a hypothetical reasonable person in her position would have done.  Lambert J.A. went on to suggest that the physician’s breach should be considered not on the basis of tort, but as a breach of a fiduciary duty to disclose depriving the plaintiff of a choice.  Hollinrake J.A. agreed that the test is the reasonable person test, but suggested that the trial judge could take into account the plaintiff’s state of mind prior to the breach.  He agreed with Wood J.A. that the trial judge had taken matters not in evidence into account and ordered a new trial on this ground.

                                                                      III.

36                               It is common ground that the plaintiff bears the burden of establishing, on a balance of probabilities, that the breach caused the loss.  The question is what the plaintiff must prove to establish that the breach caused the loss.  Is the issue what she would have done had she received the required information?  This is sometimes called the subjective approach.  Or is the issue what a hypothetical reasonable person would have done?  This is referred to as the objective approach.  Or can the debate surrounding what would have happened had the proper advice been given be avoided by holding that the plaintiff is entitled to sue her physician for depriving her of the right to choose, as Lambert J.A. suggested?

37                               I deal first with the suggestion that the debate can be avoided by treating the loss as the loss of the right to choose whether to have an abortion or not.  The suggestion that loss of choice is in itself compensable is related to the suggestion that failure to advise of risk of medical intervention negatives the patient’s consent, making the physician’s intervention tortious battery.  This Court unanimously rejected this approach in Reibl v. Hughes, [1980] 2 S.C.R. 880, per Laskin C.J., at pp. 891-92:

I can appreciate the temptation to say that the genuineness of consent to medical treatment depends on proper disclosure of the risks which it entails, but in my view, unless there has been misrepresentation or fraud to secure consent to the treatment, a failure to disclose the attendant risks, however serious, should go to negligence rather than to battery.  Although such a failure relates to an informed choice of submitting to or refusing recommended and appropriate treatment, it arises as the breach of an anterior duty of due care, comparable in legal obligation to the duty of due care in carrying out the particular treatment to which the patient has consented.  It is not a test of the validity of the consent. [Emphasis added.]

38                               For the same reasons, I would reject the alternative approach of fiduciary obligation proposed by the respondent.  As with battery, the effect would be to replace the factual analysis of standard of care and causation appropriate to negligence actions with a choice-based analysis that makes recovery virtually automatic upon proof of failure to provide relevant information.  I see no reason to depart from the approach which considers the failure of a physician to advise of medical risks under the law of negligence relating to duty of care, absent special circumstances like fraudulent misrepresentation or abuse of power for an unprofessional end:  see Reibl and  Norberg v. Wynrib, [1992] 2 S.C.R. 226.  Such conduct is neither alleged nor proven in the case at bar.

                                                                      IV.

39                               Applying the law of negligence, is the proper test what the particular plaintiff before the court would have done had she been fully informed, or what a hypothetical reasonable person would have done?

40                               The fundamental principles of negligence law suggest that the test is what the particular plaintiff before the court would have done.  Breach established, the question in a negligence action is whether the breach caused loss to the plaintiff.  This is a factual, not a hypothetical, inquiry.  In cases of negligent action or misfeasance, the matter is clear.  If a plaintiff breaks her leg as a result of being struck by a negligently driven automobile, the question is not whether a reasonable person so struck would have broken her leg; it is whether she, the particular plaintiff at bar, in fact broke her leg.  There is no reason in principle why the inquiry should be different where the claim is based on the defendant’s failure to act or non-feasance,  raising the question of what the plaintiff would have done in a hypothetical state of affairs.  The same principles apply in both cases, although the factual inquiry in cases of a breach by way of failure to act may raise more difficulties in application: J. G. Fleming, The Law of Torts (6th ed. 1983), at p. 171.

41                               The problem has frequently arisen in cases where the issue is whether failure to comply with a statutory requirement caused the plaintiff’s loss.  For example, if a plaintiff who is not wearing a helmet receives a head injury that a helmet would have prevented and sues the defendant for breach of legal duty to provide a helmet, the question arises whether the plaintiff would have worn the helmet had it been provided.  The court’s inquiry, as Fleming points out, remains “factual in the sense of dependent on the evidence available” (p. 172).  Fleming notes that cases of misfeasance or failure to act offer a “certain latitude” which “may occasionally be exploited by judge or jury to introduce policy views into what is ordinarily regarded as a purely factual issue”.  At one time, in England, it was thought that in such a case the burden of disproving causation should shift to the defendant.  A purely objective test, based on a hypothetical reasonable plaintiff, provides another example of the intrusion of policy into the purely factual inquiry of whether the breach caused the loss.  However, Fleming asserts, such deviations have been “denounced  in favour of the general rule that, regardless of the type of action, the ultimate burden always rests on the plaintiff to prove that the particular precaution would more probably than not have averted the injury.  Even in considering hypothetical human reactions ... the proper test is not what the plaintiff as a reasonable man should, but what the evidence  suggests he would, have done” (p. 172) (emphasis added).

42                               The physician’s failure to advise constitutes a failure to take an action required by law.  A finding of breach is a finding that the physician should have done something which he or she negligently failed to do.  This, like the case of the employee injured as a result of the absence of a helmet required by law, raises the hypothetical question of what the plaintiff would have done had the physician discharged his or her duty.  General tort principles suggest that this question is a purely factual inquiry to be answered by reference to all the evidence.  This evidence may include evidence from the plaintiff at trial as to what she would have done.  But it also includes relevant evidence of her situation, circumstances and mind-set at the time the decision would have been made.  The trial judge must look at all the evidence and determine whether the plaintiff would have taken the suggested course on a balance of probabilities.  One way of expressing this is to say that the plaintiff’s hindsight assertion at trial of what she would have done is tested or evaluated by reference to the evidence as to her circumstances and beliefs at the time the decision would have been made.  These circumstances include the medical advice she would have received at the time which might have influenced her decision.  In this way, the plaintiff’s subjective evidence as to what she would have done is evaluated by reference to the reasonableness of the competing courses of action.  As Sopinka J. (dissenting, but not on this ground) put it in Hollis v. Dow Corning Corp., [1995] 4 S.C.R. 634, at p. 689: “the most reliable approach in determining what would in fact have occurred is to test the plaintiff’s assertion by reference to objective evidence as to what a reasonable person would have done” (emphasis in original).

43                               This approach accords with the decision of this Court in Laferrière v. Lawson, [1991] 1 S.C.R. 541, which held (at p. 609) that causation “must be established on the balance of probabilities, taking into account all the evidence: factual, statistical and that which the judge is entitled to presume”.  It is consistent with the view there expressed that “[s]tatistical evidence may be helpful as indicative but is not determinative”, and that  “where statistical evidence does not indicate causation on the balance of probabilities, causation in law may nonetheless exist where evidence in the case supports such a finding”.  While Laferrière arose in the context of the civil law of Quebec, Gonthier J., speaking for a majority of the Court, made extensive reference to common law  jurisdictions, suggesting that the principles discussed may be equally applicable in other provinces.

44                               The approach suggested by the fundamental principles of tort law is subjective, in that it requires consideration of what the plaintiff at bar would have done.  However, it incorporates elements of objectivity; the plaintiff’s subjective belief at trial that she would have followed a certain course stands to be tested by her circumstances and attitudes at the time the decision would have been made as well as the medical advice she would have received at the time. 

                                                                       V.

45                               In discharging its task of assessing the best approach to the issue of causation in cases of medical non-disclosure, it behoves this Court to have regard to the test used in other common law jurisdictions as well as the views of scholars who have considered the matter.  I turn first to the law in other jurisdictions.

46                               In England, as the extracts from Fleming, supra, suggest, the factual test of what the plaintiff would have done on a balance of probabilities applies.  Judges consistently ask what the particular plaintiff before the court would have done: see, e.g. Chatterton v. Gerson, [1981] 1 Q.B. 432; Hills v. Potter, [1983] 3 All E.R. 716 (Q.B.D.).  The issue did not so much as elicit mention in the leading decision on informed consent to medical intervention: Sidaway v. Bethlem Royal Hospital Governors, [1985] 1 All E.R. 643 (H.L.).  It may be taken as settled that the so-called subjective approach applies.

47                               In Australia, too, courts have adopted the subjective approach based on what the particular plaintiff at bar would have done in all the circumstances.  In Ellis v. Wallsend District Hospital (1989), 17 N.S.W.L.R. 553 (C.A.), Samuels J.A., for the majority, discussed the Canadian decision in Reibl and the American decision in Canterbury v. Spence, 464 F.2d 772 (D.C. Cir. 1972), and held (at p. 581):

I do not myself find these objections to the subjective test persuasive.  I respectfully agree with Cox J in Gover (at 566) when he said: “. . . At any rate the basic causation principle governing actions in negligence plainly supports, in my opinion, the subjective test.”. . . .

It is, of course, true that a patient’s evidence about what he or she would have done if told of certain risks may be coloured by the fact that the risks did in fact eventuate; but it is open to a court to disbelieve evidence found to be tainted by hindsight: Manderson, “Following Doctors’ Orders: Informed Consent in Australia” (1988) 62 ALJ 430 at 434.  Obviously, in endeavouring to ascertain what the plaintiff’s response would have been to adequate information had it been conveyed at the appropriate time, a court will be greatly assisted by evidence of the plaintiff’s temperament, the course of any prior treatment for the same or a like condition, the  nature of the relationship between patient and doctor including pre-eminently, so far as it can be established, the degree of trust reposed in the doctor by the patient.  The extent to which the procedure was elective or imposed by circumstantial exigency and the nature and degree of the risk involved will all be matters of considerable importance: see Robertson, “Informed Consent to Medical Treatment” (1981), 97 LQR 102 at 122.

See also Young v. Northern Territory of Australia (1992), 107 F.L.R. 264 (S.C.N.T.).

48                               In the United States, the issue before us has been much debated. Many states purport to apply the objective test of the reasonable person.  However, they may introduce so many subjective factors into the so-called objective test that it becomes virtually indistinguishable from a subjective test.  For example, the Supreme Court of Hawaii in Bernard v. Char, 903 P.2d 667 (1995), purports to adopt a purely objective test that views the question of causation as what a reasonable person in the patient’s position would have done had risk information been disclosed.  However, the Court goes on to permit reference to the “individual fears and beliefs . . . of idiosyncratic patients” (p. 674), casting doubt on how reasonable the hypothetical reasonable person may in fact be.  In the end, it seems, one looks at the individual before the court and evaluates what he or she would have done by reference to all the evidence, including evidence of what a reasonable person would have done as indicated by the medical advice that he or she would have received at the time.

49                               Other states expressly apply a subjective test: see Scott v. Bradford, 606 P.2d 554 (Okla. 1979); Smith v. Reisig, 686 P.2d 285 (Okla. 1984); Arena v. Gingrich, 733 P.2d 75 (Or. Ct. App. 1987).  In the latter case, the court pointed out (at pp. 78-79) that the approach of looking at what the particular plaintiff would have done is most consistent with the aims of the fact-finding process and does not preclude evaluation of the plaintiff’s subjective testimony for reasonability by objective factors, since those factors would have a bearing on what the plaintiff would have decided:

Stated succinctly, the objective causation test espoused by defendant and the courts on whose decisions he relies makes no sense to us.  The only rationale the court offered for the test in Cobbs was that a plaintiff’s testimony about the subjective fact is likely to be colored by the events that occurred after he consented and will, ipso facto, be highly adverse to the defendant.  Although that is true, it is not unique.  Many kinds of information relevant to the determination of liability in tort cases are peculiarly within the plaintiff’s knowledge.  Juries are aware that a party’s testimony which favors his own cause must be weighed carefully and that credibility determinations must be made.  We do not think that it is necessary to present the jury with a hypothetical question rather than the real one in order to protect the factfinding process, nor do we think that presenting hypothetical questions is consistent with what the factfinding process exists to achieve. 

....Evidence and arguments about whether other patients-- hypothetical or real-- would have consented under similar circumstances can assist the factfinder in evaluating the plaintiff’s credibility and in exercising its common sense.  The question that the jury must answer is whether the plaintiff would in fact have withheld consent if apprised of the undisclosed information, but the jury may consider, in deciding that question and in weighing the plaintiff’s credibility, the likelihood that a competent and prudent person would make the decision which the plaintiff claims that he would have made.  Although a patient’s decision to give or refuse consent need not be an objectively reasonable one, the trier of fact is entitled to consider its perception of the reasonableness of a decision as one of the factors bearing on what the plaintiff in the given case would have decided.  [Emphasis in original.]

50                               Canadian academic commentary largely supports the so-called subjective approach to causation in medical negligence cases.

51                               P. H. Osborne, Annotation to Arndt v. Smith (1995), 25 C.C.L.T. (2d) 264, argues that the concern that a subjective test depends too much on the testimony of the plaintiff may be overstated if not illusory.  He states (at p. 267):

Causation is not proved solely by the plaintiff’s testimony.  The court must make an assessment of credibility which would appear not to be easier nor more difficult than in other contexts.  There may also be relevant evidence from third parties to assist the trier of fact.

Osborne suggests that a doctrine that protects only reasonable “choosers” is too harsh and may fail to protect individual plaintiffs.  In “Causation and the Emerging Canadian Doctrine of Informed Consent to Medical Treatment” (1985), 33 C.C.L.T. 131, Osborne strongly advocates the subjective approach.  He indicates that since Reibl, there has been diversity in the manner in which the test has been applied.  The diversity is found in the range of personal factors and circumstances to be incorporated into the reasonable person test.  In some cases, the courts examine little more than the available medical data.  In others, a limited number of personal factors are considered.  In still others, the courts have taken into account a wide range of subjective factors in applying the modified objective approach.  Finally, some courts have applied both a subjective and a modified objective test.

52                               Osborne argues that if any concern is to be shown for a patient’s self determination, a subjective test is mandated.  Under the objective approach, the patient’s right to self determination may be denied because the reasonable person’s decision may be different from that of the individual patient.  He adds that under a subjective test, the plaintiff’s testimony is not conclusive.  “The question is not what the plaintiff says he would have done but what the Judge or jury on the basis of all the evidence believes that the plaintiff would have done” (p. 141).

53                               G. Robertson, “Informed Consent Ten Years Later: The Impact of Reibl v. Hughes” (1991), 70 Can. Bar Rev. 423, suggests that “[i]n principle this causal link should be assessed subjectively, by considering whether this particular patient (that is, the plaintiff) would have declined the treatment if the risks had been disclosed” (p. 426) (emphasis in original). He notes that since Reibl, which many courts have interpreted as imposing a strict objective test, causation has proved to be a formidable obstacle for plaintiffs.

54                               N. Fontigny, “When Yes Really Means Yes: The Law of Informed Consent in Canada Revisited”(1996), 4 Health L. Rev. 17, although opining that framing the duty of disclosure in fiduciary terms is simple and logical, suggests that consideration be given to a subjective test such as that here proposed.  She states at p. 22:

... the Supreme Court could consider returning to a subjective test of causation.  Authors, such as P. H. Osborne, have argued that there is nothing inherently unworkable or unfair in a subjective test, since the question is not what the patient would have done, but what the Judge or jury, on the basis of all the evidence, believes that the plaintiff would have done.  The evidence of the plaintiff would be important, but it would be evaluated.  The ultimate determination would only be made after a careful consideration of the plaintiff’s credibility; in the light of the plaintiff’s personal circumstances and personality, the nature of the illness, the nature of the proposed treatment, the relationship between the plaintiff and his doctor, and other relevant evidence.  It is therefore possible to argue that the Supreme Court should change the modified subjective test, and adopt a subjective test, and then frame this test in such a way so as to not be unfairly prejudicial to the health care providers.

55                               I conclude that while views diverge, the preponderance of authority in other common law jurisdictions as well as academic commentary support a test which asks what the particular plaintiff would have done in all the circumstances, but accepts that the reasonableness of the one choice over another, as reflected in the medical advice the plaintiff would have received, is an important factor bearing on that decision.

                                                                      VI.

56                               This brings us to the arguments against the position that the trial judge’s task is to determine what the particular plaintiff at bar would have done had she been properly advised of the risk.  The most formidable is the submission that this Court rejected such an approach in Reibl, supra.

57                               Reibl concerned a patient who had suffered a stroke and paralysis as a result of an endarterectomy performed to reduce the risk of stroke in the future.  His physician had failed to advise him of a risk of stroke or death during or in the aftermath of surgery.  He told him only that his risk of having a stroke was greater without the surgery than with it.  The patient argued that had he been informed of the risk of stroke consequent on surgery, he would have postponed the surgery for a year and one-half until a lifetime retirement pension from his employment vested.  He also asserted that he would have opted for a shorter, normal life rather than a longer life as a cripple.  The issue before the Court was how to assess whether the patient would in fact have foregone the surgery had he been properly informed.

58                               The Court, per Laskin C.J., began by rejecting the argument that the real wrong was being deprived of the right of choice or consent, giving rise to the tort of battery.  Laskin C.J. stated at pp. 891-92:

I can appreciate the temptation to say that the genuineness of consent to medical treatment depends on proper disclosure of the risks which it entails, but in my view, unless there has been misrepresentation or fraud to secure consent to the treatment, a failure to disclose the attendant risks, however serious, should go to negligence rather than to battery.  Although such a failure relates to an informed choice of submitting to or refusing recommended and appropriate treatment, it arises as the breach of an anterior duty of due care, comparable in legal obligation to the duty of due care in carrying out the particular treatment to which the patient has consented.  It is not a test of the validity of the consent.  [Emphasis added.]

59                               The Court went on to approach the problem on the assumption that a plaintiff suing for negligent non-disclosure will always assert that he or she would have acted differently had he or she been fully informed.  Laskin C.J. wrote (at p. 898):

It could hardly be expected that the patient who is suing would admit that he would have agreed to have the surgery, even knowing all the accompanying risks.  His suit would indicate that, having suffered serious disablement because of the surgery, he is convinced that he would not have permitted it if there had been proper disclosure of the risks, balanced by the risks of refusing the surgery.

60                               Laskin C.J. dismissed a purely subjective approach on the ground that “it depends on the plaintiff’s testimony as to his state of mind, thereby exposing the physician to the patient’s hindsight and bitterness” (quoting at p. 898 from “Informed Consent -- A Proposed Standard for Medical Disclosure” (1973), 48 N.Y.U. L. Rev. 548, at p. 550).  Laskin C.J. observed that “the issue of causation would be in the patient’s hands on a subjective test, and would, if his evidence was accepted, result inevitably in liability unless, of course, there was a finding that there was no breach of the duty of disclosure” (p. 899).  For this reason, Laskin C.J. rejected a purely subjective approach based on what the plaintiff at trial asserted he would have done.

61                               At the same time, Laskin C.J. was alive to the problems presented by a purely objective approach based on what a hypothetical reasonable person would have done.  He noted (at p. 898):

... a vexing problem raised by the objective standard is whether causation could ever be established if the surgeon has recommended surgery which is warranted by the patient’s condition.  Can it be said that a reasonable person in the patient’s position, to whom proper disclosure of attendant risks has been made, would decide against the surgery, that is, against the surgeon’s recommendation that it be undergone?  The objective standard of what a reasonable person in the patient’s position would do would seem to put a premium on the surgeon’s assessment of the relative need for the surgery and on supporting medical evidence of that need.  Could it be reasonably refused?

62                               Having rejected both a purely subjective and a purely objective approach, Laskin, C.J. crafted a test that fell between the two, the “modified objective” test.  While approaching the issue of what the patient would have done objectively, the judge should take into account “special considerations affecting the particular patient” (p. 898).  This would, in his view, avoid the problem of leaving the matter of causation entirely in the surgeon’s hands:  “The patient’s particular situation and the degree to which the risks of surgery or no surgery are balanced would reduce the force, on an objective appraisal, of the surgeon’s recommendation” (p. 899).  At the same time, consideration of all factors that might have “reasonably” affected the decision avoids exclusive reliance on the plaintiff’s assertion at trial.

63                               While “the patient’s particular concerns” at the time should be considered (pp. 899-900) (for example, the judge might consider specific questions which the patient may have asked, evincing specific concerns (p. 899)), they must be “reasonably based” to avoid excessive subjectivity.  Thus “fears which are not related to the [undisclosed] material risks . . . would not be causative factors”.  Summing up, Laskin C.J. stated (at p. 900):

In short, although account must be taken of a patient’s particular position, a position which will vary with the patient, it must be objectively assessed in terms of reasonableness.

64                               There is little profit in debating whether the test Laskin C.J. had in mind  should be labelled objective or subjective.  Suffice it to say it contains elements of both the subjective and objective and has been read in different ways.  Two assertions can, however, be ventured.  First, the Court was concerned to ensure that the plaintiff’s particular concerns and circumstances be considered.  To hold otherwise would be to virtually place the outcome of the causation inquiry in the hands of the physician.  Second, the Court was concerned to ensure that the plaintiff’s subjective assertion of what she would have done had she been properly advised be tested “in terms of reasonableness”.  To hold otherwise would give undue weight to the plaintiff’s hindsight assertion that she would have acted in a way that supports her claim for damages.  The approach suggested above -- that causation is a question of fact for the trial judge to determine on all the evidence including the plaintiff’s assertion at trial examined in the light of her circumstances, mind-set and the medical advice she would have received at the time -- satisfies both these concerns.

65                               This brings us to the ultimate objection to the proposed test: that it treats the plaintiff unfairly by diminishing her right to choose.  As noted earlier, absent extraordinary conduct such as fraud or abuse of power, the inquiry into the damages that flow from failure to advise of medical risks is governed by the law of negligence.  The issue is not the plaintiff’s right to choose, as it would be for the tort of battery or an action for fraud, but whether as a factual matter the negligent act caused the loss.

66                               Applying the standard appropriate to the law of negligence, the question  is whether the proposed test is fair to both plaintiff and defendant.  In my view, it is.  It avoids the trap of determining the issue of causation either solely by the physician’s opinion of what a reasonable person would have done or solely by the plaintiff’s unilateral assertion at trial that he or she would have acted differently.  Instead, it places the issue in the hands of the trial judge for determination on all the evidence.  It takes into account the plaintiff’s right of choice, rather than presuming that choice on the basis of a hypothetical reasonable person.  And it permits serious consideration of the plaintiff’s evidence as to what that choice would have been.  As the California Supreme Court  stated in Cobbs v. Grant, 502 P.2d 1 (1972), at pp. 11-12, cited by Sopinka J. in Hollis, supra, at p. 689, “[t]he patient-plaintiff may testify on this subject but the issue extends beyond his credibility”.  At the same time, it is fair to the physician, who may introduce evidence of what the reasonable patient would have done as it bears on the choice the particular patient at bar would have made.

67                               The proposed test also meets the need, eloquently voiced by Wood J.A. in the British Columbia Court of Appeal, for a test which accommodates cases where the decision involves a “delicate balancing of overlapping personal, ethical, and medical considerations which can lead to more than one ‘reasonable’ choice” (p. 225 B.C.L.R.).  The objective test, based on the hypothetical reasonable person, depreciates the plaintiff’s personal choice in such situations and deprives her testimony of any weight.  The test of what the particular plaintiff at bar would have done, determined in the light of all the circumstances and his or her own values and attitudes, avoids these problems.  Nor does the fact that there was more than one reasonable choice present a problem.  The question is what choice this plaintiff would have made on all the evidence and on a balance of probabilities.

                                                                      VII.

68                               In my view, the trial judge applied the right test and did not err in dismissing the action on the ground that the plaintiff failed to establish that the physician’s failure to advise her of the risk to her fetus from chickenpox caused the loss associated with the rearing of her disabled daughter. 

69                               Although the trial judge paid lip service to the reasonable person test, he resolved the issue of causation by asking what the plaintiff herself would have done in all the circumstances.  My colleagues Sopinka and Iacobucci JJ. suggest that the trial judge did not apply the subjective test because he did not evaluate the plaintiff’s evidence.  I respectfully disagree.  The trial judge considered the plaintiff’s evidence in its entirety but concluded that “Ms. Arndt’s testimony [was] coloured, probably not by an intent to mislead the court, but rather by the tremendous stress that her daughter’s wretched health [had] inflicted upon her” (p. 233).  Because her evidence was coloured, the trial judge tested the plaintiff’s assertion that she would have had an abortion had she been told of “any risk of abnormalities at all” (p. 235) against other circumstances existing at the time that the choice would have been made.  He considered the fact that an abortion in the second trimester would have been riskier than an earlier abortion and the need to obtain the approval of a committee for any abortion.  He went on to consider factors going to the plaintiff’s state of mind, including the fact that she very much wanted to have a child, that she had a natural skepticism of mainstream medicine and had so informed Dr. Smith, and that she stated she did not want an ultrasound scan of the developing fetus.  For all these reasons he concluded “that Ms. Arndt would have chosen in foresight to carry Miranda to full term had she been fully apprised of the nature and probability of risks her fetus faced” (p. 236).

70                               The trial judge’s approach conforms to the test proposed in these reasons.  He asked the right question: what would this plaintiff have decided in all the circumstances?  He then evaluated her assertion at trial that she would have had an abortion by reference to her external circumstances and her values and state of mind as evidenced by her conduct and words at the time the decision fell to be made.  In short, he proceeded correctly and made no error of principle.  Although he did not expressly cite it as a reason for his decision, the fact that the medical profession would not have recommended an abortion in these circumstances also supports the trial judge’s conclusion that Ms. Arndt would not have chosen an abortion had she been advised of the very small increased risk to the fetus posed by varicella.  As she admitted, she would not have opted for abortion “without significant convincing”.  On the evidence, that convincing would not have occurred.

71                               It is argued, however, that the trial judge erred in considering the increased risk of a second trimester abortion and the need for approval for an abortion on the ground that these matters were not in evidence.  In fact, there was evidence of increased risk of second trimester abortions (although this appears not to have been pointed out to the Court of Appeal).  The need to obtain committee approval for an abortion was a matter of law of which the trial judge could take judicial notice:  Laferrière, supra.  Consequently, I find no error on these counts.

72                               I find no legal basis for interference with the trial judge’s conclusion that the plaintiff failed to establish on a balance of probabilities that had she been given the required information concerning risk, she would have ended her pregnancy.

                                                                     VIII.

73                               Finding, as I do, that the trial judge did not err in concluding that the plaintiff had failed to establish Dr. Smith’s negligence caused her loss, it is unnecessary to consider whether the action would in any event have been statute-barred.  However, since the matter has been fully argued before us, it may be useful to offer guidance on the point for future cases.

74                               The Limitation Act, R.S.B.C. 1979, c. 236, s. 3(1)(a), provides a limitation period of two years in actions for “damages in respect of injury to person or property, including economic loss arising from the injury”.  The majority of the Court of Appeal concluded that this provision barred the plaintiff’s claim for emotional trauma, stress and loss of enjoyment of life, because these claims relate to “injuries” to Ms. Arndt, but did not bar her claims for loss and expense relating to the care of Miranda because these were not claims in relation to “damages in respect of injury to person or property”.

75                               In my view, Ms. Arndt’s claims for economic loss are not statute-barred because they do not relate to personal injury either to Ms. Arndt or to Miranda.  In Martin v. Insurance Corp. of British Columbia (1979), 13 B.C.L.R. 163 (S.C.), the court held that injury to person means physical, emotional or mental injury to the body or mind of the individual.  Unlike her claim for emotional trauma, stress and loss of enjoyment of life, Ms. Arndt’s claims for economic loss do not relate to physical, emotional or mental injury to her body.  Nor do the injuries represent physical, emotional or mental injury to Miranda because Dr. Smith did not cause the child's injuries.  The action for economic loss is for the expenses incurred in raising Miranda.  Dr. Smith’s negligent failure to fully inform the plaintiff could have caused loss only in the sense that it was a factor leading to Miranda’s birth.  It is therefore not caught by the wording of s. 3(1)(a) of the Limitation Act. 

                                                                      IX.

76                               I would allow the appeal and restore the trial judge’s order dismissing the action.  The defendant is entitled to costs on a party and party basis here and in the British Columbia courts.

Appeal allowed with costs, Sopinka and Iacobucci JJ. dissenting.

Solicitors for the appellant:  Harper Grey Easton, Vancouver.

Solicitors for the respondents:  MacLeod Smith, Vancouver.