Introduction and factual setting
To some extent this paper is a case comment on Hemmings v Peng.1 More particularly, it focuses on the issue of remoteness of damage (now generally referred to as "legal causation") in the context of a medical malpractice claim.2
The plaintiff in Hemmings was a 29-year-old woman who suffered a cardiac arrest on the operating table during a Caesarian section, as a result of which she sustained severe brain damage. The plaintiff was morbidly obese, which made her a higher risk patient. The C-section procedure began, in accordance with normal practice, under a regional anesthetic, but the anesthetist decided to convert to a general anesthetic after the plaintiff started to panic and flail about and could not be calmed. The infant was delivered some 8 to 11 minutes into general anesthesia, but shortly after delivery the plaintiff experienced cardiac arrest and brain damage ensued.
An action was commenced against the treating obstetrician, two obstetricians who provided care at the hospital (one of whom performed the C-section), and the anesthesiologist. An additional defendant was the hospital, which employed a nurse who had conducted an allegedly insufficient telephone consult with the plaintiff several weeks prior to the delivery. Damages were agreed by the parties at $12 million. At trial, judgment was rendered against the treating obstetrician, the anesthetist, and the hospital. The action was dismissed as against the remaining defendants. The appeals made by the treating obstetrician and the hospital were allowed; that of the anesthetist was dismissed. The end result was that only the anesthetist was held liable.
The finding at trial that the treating obstetrician had owed a duty of care to the plaintiff and had breached the applicable standard of care – i.e. the finding that he had been negligent – and that his negligence had been a "but for" factual cause of the plaintiff's injuries, was affirmed on appeal. However, his appeal succeeded on the issue of legal causation. It was held that the plaintiff's injuries were too remote to be legally recoverable from him. The finding of negligent conduct on the part of the nurse was not challenged on appeal; instead, the nurse's argument was that her breach of the standard of care did not cause the plaintiff's injuries. Here again the appeal failed on the issue of factual causation but succeeded on the issue of remoteness. Only in the case of the anesthetist did the appeal fail on all issues, although legal causation apparently was not raised in his appeal and in any event was not considered.3
The lengthy appeal decision included consideration of issues that do not call for comment in this paper.
Two separate types of causation
Causation is a necessary element in a tort claim. That single term, however, encompasses two separate issues: factual causation, which is generally determined via the "but for" test, and "legal causation", which is a more recent, but now generally adopted, label for what traditionally has been identified as remoteness of damage.4
(a) Factual causation
As the name implies, factual causation requires a finding that the defendant's wrongdoing was a necessary factual cause of the loss or harm. As mentioned, this is normally determined via the "but for" test, which is articulated as follows: "[T]he plaintiff must show on a balance of probabilities that 'but for' the defendant's negligent act, the injury would not have occurred."5
(b) Legal causation
Legal causation, on the other hand, asks whether the harm "is too remote to warrant recovery"6 regardless of whether the defendant's wrongdoing was a factual cause of the plaintiff's harm or loss. Stated differently, the remoteness inquiry asks whether "the harm [is] too unrelated to the wrongful conduct to hold the defendant fairly liable".7
Legal causation does not, however, apply equally to all claims. "[T]he rules developed in legal causation, such as foreseeability, do not readily apply in [equitable compensation claims]... [While] the common law imports ideas of foreseeability (or reasonable contemplation) and remoteness into its assessment of causality...[those] considerations do not readily enter into Equity's assessment of fiduciary accountability".8
The general test
The general test for legal causation (remoteness of damage) is outlined in the leading Mustapha decision as follows:
Since The Wagon Mound (No. 1), the principle has been that "it is the foresight of the reasonable man which alone can determine responsibility"...Much has been written on how probable or likely a harm needs to be in order to be considered reasonably foreseeable. The parties raise the question of whether a reasonably foreseeable harm is one whose occurrence is probable or merely possible. In my view, these terms are misleading. Any harm which has actually occurred is "possible"; it is therefore clear that possibility alone does not provide a meaningful standard for the application of reasonable foreseeability. The degree of probability that would satisfy the reasonable foreseeability requirement was described in The Wagon Mound (No. 2) as a "real risk", i.e. one which would occur to the mind of a reasonable man in the position of the defendant...and which he would not brush aside as far-fetched...9
As discussed below, that passage contains problematic language.
Overview
Following is an overview of the two basic principles governing the issue of legal causation.
(i) While reasonable foreseeability is the underlying general principle, it is not sufficient that some generic kind or type of harm (such as "property damage" or "bodily injury") was foreseeable to a reasonable person in the position of the defendant. It is necessary that the actual harm suffered by the plaintiff (although only to the first level of specificity – this will be explained below) was foreseeable; and
(ii) While foreseeability of a bare possibility of that actual harm is insufficient, there is no requirement that the harm be foreseeable as a probability (in the sense of a greater-than-50% chance of occurrence). Reasonable foreseeability of the actual harm as a "real risk" (which, by judicial definition, is generally a particular form of possibility) is sufficient.
Unfortunately, clarity in the case law has been in short supply.10
A - FORESEEABILITY OF THE ACTUAL HARM SUFFERED
Two separate foreseeability analyses
As mentioned above, the first of the two hurdles to be overcome when considering the issue of remoteness of damage is the obligation to show that the actual harm suffered by the plaintiff was, in all the circumstances, and at a particular minimum level of likelihood (i.e. as a "real risk"), a reasonably foreseeable consequence of the defendant's wrongdoing.. (The requirement for the minimum level of likelihood is the second hurdle and is discussed below.) As outlined below, a similar but not identical foreseeability analysis is made when determining whether the defendant owed the plaintiff a duty of care.
(a) Duty of care foreseeability analysis
The nature of the duty of care foreseeability analysis is described in the following comments:
The reasonable foreseeability inquiry requires the court to ask whether the type of injury to the plaintiff, or to a class of persons to which the plaintiff belongs, was reasonably foreseeable to someone in the defendant's position...The question [in this case] therefore is whether someone in [the defendant's] position would reasonably have foreseen economic loss to the [plaintiffs], or the class of plaintiffs to which they belong, as a result of their negligence.11
[I]t is well established that damages must be foreseeable as to kind, but not extent...12
[If] the requisite close and direct relationship is shown...the first stage of the Anns/Cooper framework will be complete, as long as the risk of injury was reasonably foreseeable.13
Under the Anns/Cooper framework, a prima facie duty of care is established by the conjunction of proximity of relationship and foreseeability of injury.14
The first two of those remarks expressly refer to the need for foreseeability of the generic type or kind of harm or loss that was sustained by the plaintiff, while the latter two appear to indicate, or at least suggest by way of omission, that foreseeability of injury of any kind will be sufficient when considering whether the defendant owed a duty of care to the plaintiff. The correct approach is set out in the initial two comments.15
(b) Remoteness foreseeability analysis
As in the duty of care analysis, the issue here involves the foreseeability of harm or loss to the plaintiff arising from the defendant's wrongdoing. There is, however, a subtle and perhaps (depending on the circumstances) important difference, one which is outlined in the following statements:
Remoteness, at its core, turns on the reasonable foreseeability of the actual injury suffered by the plaintiff.16
Remoteness is distinct from the reasonable foreseeability analysis within duty of care because it focuses on the actual injury suffered by the plaintiff, whereas the duty of care analysis focuses on the type of injury...17
Thus, in the remoteness analysis, the requirement is not (as it is in the duty of care analysis) that the generic type or kind of harm or loss that was sustained by the plaintiff was reasonably foreseeable to the defendant, but rather that the actual injury suffered by the plaintiff was reasonably foreseeable (at the "real risk" level of likelihood).
Foreseeable at what level of specificity?
The requirement for reasonable foreseeability of the "actual injury" sustained by the plaintiff is itself ambiguous. It leads to the question: at what level of specificity must the injury have been foreseeable? Should it, for example, have been foreseeable in Hemmings not just that brain damage might be suffered, but that (say) frontal lobe brain damage might be sustained? Or, taking it one level further, that a particular type of frontal lobe brain damage might occur? Those questions can continue to be asked down a spiral of levels of specificity. As indicated in the passage from Mustapha reproduced above,18 it appears that the required foreseeability is limited to the first level of classification. The passage shows that foreseeability only of some mental, as opposed to some physical, injury was necessary there. There was no need for foreseeability of the specific type or subtype of mental injury that the plaintiff sustained.
Therefore, the requirement in Hemmings was reasonable foreseeability of brain damage of some (any) kind or type, regardless of magnitude or severity.19 (The second issue – the level of likelihood at which brain damage of some kind was required to have been foreseeable – is considered below.)
The wrong question was asked in Hemmings
Insofar as the treating obstetrician was concerned, the trial judge framed the remoteness question as follows: "[W]as it reasonably foreseeable [that the plaintiff] would suffer a cardiac arrest from an anesthetic complication?"20 That was not, however, the correct question.21 "It is not necessary that one foresee the 'precise concatenation of events'; it is enough to fix liability if one can foresee in a general way the class or character of injury which occurred".22 Legal causation (remoteness) turns on the type of harm that was foreseeable, and more particularly whether the actual harm ultimately suffered was foreseeable (as a "real risk"), not the manner in which that harm occurred.
As stated in another recent appellate decision, "The question of legal causation is whether the actual injury suffered by [the plaintiff] is sufficiently related to the [defendant's] breach."23 The onus resting with the plaintiff in Hemmings was to show that the actual injury that was suffered –brain damage – was reasonably foreseeable (at a "real risk" level of likelihood). There was no need for foreseeability of the process or method by which the brain damage occurred – i.e. the anesthetic complication and the resulting cardiac arrest.
Although foreseeability of those matters (the anesthetic complication and the resulting cardiac arrest) was not within the scope of the remoteness test, the appeal court laid out a summary of the evidence "on the issue of whether it was foreseeable to a reasonable person in the position of one of the obstetrics defendants, at the time of their respective breaches of the standard of care, that [the plaintiff] would suffer a cardiac arrest arising from an anesthetic complication".24
The appeal court (as did the trial judge) misidentified the issue for consideration. The correct issue was whether, in all the circumstances, it was reasonably foreseeable (at the "real risk" level of likelihood) that the plaintiff might suffer some form, and at some level of severity, of brain damage in the event of negligent medical treatment, with account being taken not only of their specialized knowledge as physicians, but also their knowledge of the plaintiff's particular condition (her morbid obesity) and frailties (the "multitude of increased health risks...including cardiac arrest" that accompanied such obesity).25
The issue was not whether it was foreseeable that there might be an anesthetic complication leading to a cardiac arrest, but that was the basis upon which the appeal court concluded that the plaintiff had not satisfied her onus of proof on the issue of legal causation.26 For example:
If Dr. O'Brien could not anticipate, or reasonably foresee, that a C-section on [the plaintiff] would result in an anesthetic complication and cardiac arrest, how could Nurse San Juan, whose interaction with [the plaintiff] was limited to a telephone call twelve days before delivery?27
This is not to say that a different result would have obtained had the correct question been asked. Because the trial judge (as did the appeal court) asked the wrong question, we do not have the benefit of a finding by him on the correct issue. Not only was there a failure to ask the right question, but it may be that no evidence related to the issue in its correct form was led at trial.
As stated above, it is noteworthy that the anesthetist, the only defendant whose liability was affirmed, apparently did not raise the matter of legal causation on appeal, and no consideration was given there to the issue of remoteness in the claim made against him.
B - "REAL RISK" LEVEL OF LIKELIHOOD
Inappropriate terminology
The term "probability" is ambiguous. Strictly speaking, "probability" refers to something that is "probable". An event or outcome or happening of any sort is "probable" when the chance of that happening is greater than 50% (in lawyer-speak, when that happening is "more likely than not"). If the chance is less than 50%, the occurrence is "possible", not "probable", and therefore is a "possibility", not a "probability".
There is, however, a well-established second use/meaning of the term "probability", one which addresses likelihood generally, including those situations where the chance of occurrence is less than 50%. For example: "There is a 10% probability of rain tomorrow."28 When used in that sense, "probability" includes possibilities.
Ambiguity can sow confusion. How should the term "degree of probability" as used in Mustapha29 be interpreted: (a) to apply only where there is a greater-than-50% chance, or (b) to include chances that are below 50% (so long as a specified condition, as discussed below, is met)? As explained hereafter, it is virtually incontestable that the latter is the correct interpretation.
Were the strict meaning applied, there would be an immediate bar to a large number of otherwise worthy claims as to which foreseeability on a greater-than-50% level of likelihood could not successfully be asserted. On the other hand, the employment of a wider meaning – one that may include harm that is foreseeably less-than-50% likely to occur – would be a counterintuitive application.
Mustapha has made it clear that it is the latter interpretation which governs. By adopting the "real risk" level of likelihood test, and then defining "real risk" to mean "[a risk] which would occur to the mind of a reasonable man in the position of the defendant...and which he would not brush aside as far-fetched", the court indisputably included within the scope of recoverable claims those where the foreseeable level of likelihood of occurrence of the harm or damage is less than 50% - perhaps considerably less than 50%. In fact, the "real risk" test virtually ensures that the grey area in which remoteness is a live issue will generally involve the reasonable foreseeability of the possible, not probable, occurrence of the actual harm ultimately suffered. That is because the foreseeability of a probability (in the more-than-50% sense of that term) of the occurrence of that actual harm will almost certainly satisfy the "real risk" test, and will therefore not be an issue requiring consideration.
Mustapha, in other words, decisively rejected the application of the narrower, more literal meaning of the term "probability", and permitted claims which fall within the purview of foreseeable "possibilities" to be advanced, but did so only on the condition that those claims satisfy the "real risk" test of level of foreseeable likelihood. There can be little doubt that that is the correct interpretation; otherwise, the adoption of the "real risk" test would be rendered meaningless. The rejection of the requirement for a greater-than-50% chance can, in fact, be seen in the following comment (by Dickson J.A. as he then was) made long before Mustapha was decided:
The test of foreseeability of damage becomes a question of what is possible rather than what is probable.30
In addition to being incompatible with the judicially-mandated "real risk" test, any requirement that the actual injury suffered by the plaintiff be foreseeable as a probability (in the more-than-50% sense of the term) would be unduly harsh and punitive.
It appears to have been recognized in Hemmings that the word "probability" is not to be applied in the narrow sense of requiring a greater-than-50% chance. Although the bald statement was made that "Mere possibility that the harm would occur is not sufficient",31 the court also expressly referred to the "real risk" test, thereby indicating that a possibility of the occurrence of the actual harm ultimately suffered will be sufficient if that foreseeable possibility amounts to a "real risk" as judicially defined, i.e. "one which would occur to the mind of a reasonable man in the position of the defendant...and which he would not brush aside as far-fetched".32 The statement "Mere possibility that the harm would occur is not sufficient" (emphasis added) accords with the statement made in Mustapha that "[P]ossibility alone does not provide a meaningful standard for the application of reasonable foreseeability".33 The test, in other words, can be satisfied by foreseeable possibilities, but only if those possibilities amount to "real risks". A "mere possibility", or "possibility alone", will not do the trick.
The following comments made in an earlier appellate decision spell out somewhat more clearly the potential for confusion engendered by the use of the ambiguous term "probability":
I observe that in the law of negligence concluding that something is "probable" does not imply that it is likely to happen. Rather, as indicated by Linden, Canadian Tort Law, at p. 117: If there is only a slight chance that an accident will occur, the court may hold that running such a risk is not unreasonable, because "people must guard against reasonable probabilities, but they are not bound to guard against fantastic possibilities"...In this context, the word "probability" does not mean that an accident must be more likely to happen than not; there need be only a real or substantial risk of harm. One chance of injury in 100 or even 1,000 may suffice.
The converse should also be true it seems to me; in other words, there may be "possibilities" that are not so remote or "fantastic" as to be unforeseeable because there is a real risk of harm that can and should be guarded against.
...[F]rom a reading of [the trial] judgment as a whole and the evidence...there was a real or substantial risk of harm or injury to the plaintiff. While undoubtedly it would have been better and less likely to cause misinterpretation if [the trial judge] had not used the word "possibility," the inapt use of this word does not constitute by itself, legal error.34
The final paragraph in that passage demonstrates the potential for confusion arising from the use of "probability"-related language. After first correctly saying that, in the context under consideration, the term "probability" is not to be given its strict meaning of a "more-than-50%" chance, and that a one-in-one-hundred, or even (depending on the circumstances) a one-in-one-thousand, chance of injury might qualify as a reasonably foreseeable "real or substantial risk of harm or injury", the court then said (in the final quoted paragraph) that the use of the term "possibility" was "inapt". That inconsistency – saying first that, depending on the circumstances, a level of likelihood less than 50%, perhaps even less than 1% (i.e. a possibility, perhaps even a very slim possibility), might satisfy the test, but then saying that a "possibility" is an unacceptable, or at least inappropriate, term for use in the "real risk" context – muddies the waters and results in needless confusion.
To read the full article click here
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
