Martin v. American International Assurance Life Co., 2003 SCC 16

Martin v. American International Assurance Life Co., [2003] 1 S.C.R. 158, 2003 SCC 16

American International Assurance Life Company Ltd.

and American Life Insurance Company Appellants

v.

Dorothy Martin Respondent

Indexed as: Martin v. American International Assurance Life Co.

Neutral citation: 2003 SCC 16.

File No.: 28540.

2002: October 28; 2003: March 21.

Present: McLachlin C.J. and Gonthier, Iacobucci, Major, Bastarache, Binnie, Arbour, LeBel and Deschamps JJ.

on appeal from the court of appeal for british columbia

Insurance — Life insurance — Interpretation of accidental death benefit provision in policy — Insured dying from overdose caused by intravenous injection of Demerol — Accidental death benefit provision stipulating coverage would be provided only for deaths effected through “accidental means” — Whether category of deaths caused by “accidental means” narrower than that of “accidental deaths” — Whether insured’s death effected through “accidental means”.

The insured, Dr. E, was a physician who developed an addiction to opiate medications. He completed a residential treatment program. However, after a painful orthopaedic injury, he became physiologically dependent on both morphine and Demerol and was placed on a program of gradual withdrawal from these drugs. Dr. E was found dead in his office and the coroner concluded he died from an overdose caused by an intravenous injection of Demerol. The level of Demerol found in his blood was at the low end of the range for lethal doses. Dr. E’s life insurance policy stipulated that coverage would be provided only for deaths effected through “accidental means”. At trial, the claim for coverage under the provision was dismissed. The Court of Appeal allowed the appeal and determined that the respondent beneficiary could recover under the policy.

Held: The appeal should be dismissed.

The phrase “accidental means” in the insurance policy does not refer to a narrow subclass of the broader category of “accidental deaths”. Both phrases connote a death that was in some sense unexpected. To determine whether death occurred by accidental means, it is necessary to look to the chain of events as a whole, and consider whether the insured expected death to be a consequence of his actions and circumstances. The central question is whether the insured expected to die. The circumstances of the death may point to the answer. However, if the answer is unclear when the matter is viewed solely from the perspective of the insured, the court may consider whether a reasonable person in the position of the insured would have expected to die. In this case, the circumstances surrounding Dr. E’s death support the inference that his death was effected through “accidental means”. Dr. E did not expect to die but simply made a miscalculation concerning how much Demerol his body could tolerate. In concluding otherwise, the trial judge erred in his appreciation of the law and the facts. The appeal is dismissed. The respondent is entitled to payment of the accidental death benefit.

Cases Cited

Considered: Columbia Cellulose Co. v. Continental Casualty Co. (1963), 1963 CanLII 476 (BC CA), 43 W.W.R. 355, aff’d (1964), 1964 CanLII 630 (SCC), 42 D.L.R. (2d) 401n; Sloboda v. Continental Casualty Co., 1938 CanLII 204 (AB SCAD), [1938] 2 W.W.R. 237; Smith v. British Pacific Life Insurance Co., 1965 CanLII 91 (SCC), [1965] S.C.R. 434; Aguilar v. London Life Insurance Co. (1990), 1990 CanLII 7931 (MB CA), 70 D.L.R. (4th) 510, leave to appeal granted, [1991] 1 S.C.R. v, notice of discontinuance filed, [1991] 3 S.C.R. v; Leontowicz v. Seaboard Life Insurance Co. (1984), 1984 ABCA 346, 8 C.C.L.I. 290, leave to appeal refused, [1985] 1 S.C.R. ix; Candler v. London & Lancashire Guarantee & Accident Co. of Canada (1963), 1963 CanLII 155 (ON HCJ), 40 D.L.R. (2d) 408; referred to: Reid Crowther & Partners Ltd. v. Simcoe & Erie General Insurance Co., 1993 CanLII 150 (SCC), [1993] 1 S.C.R. 252; Landress v. Phoenix Mutual Life Insurance Co., 291 U.S. 491 (1934); Fenton v. J. Thorley & Co., [1903] A.C. 443; Glenlight Shipping Ltd. v. Excess Insurance Co., 1983 S.L.T. 241; Mutual of Omaha Insurance Co. v. Stats, 1978 CanLII 38 (SCC), [1978] 2 S.C.R. 1153; Canadian Indemnity Co. v. Walkem Machinery & Equipment Ltd., 1975 CanLII 141 (SCC), [1976] 1 S.C.R. 309; Johnson v. Mutual of Omaha Insurance Co. (1984), 1984 CanLII 1818 (ON CA), 45 O.R. (2d) 676, aff’g (1982), 1982 CanLII 2087 (ON HCJ), 39 O.R. (2d) 559; Cornish v. Accident Insurance Co. (1889), 23 Q.B.D. 453; Brown v. Continental Casualty Co., 108 So. 464 (1926); Bertalan Estate v. American Home Assurance Co. (1999), 1999 CanLII 6170 (BC SC), 68 B.C.L.R. (3d) 118; Thompson v. Prudential Insurance Co. of America, 66 S.E.2d 119 (1951); Allred v. Prudential Insurance Co. of America, 100 S.E.2d 226 (1957); Wagner v. International Ry. Co., 133 N.E. 437 (1921); Horsley v. MacLaren, 1971 CanLII 24 (SCC), [1972] S.C.R. 441; Corothers v. Slobodian, 1974 CanLII 187 (SCC), [1975] 2 S.C.R. 633; Housen v. Nikolaisen, [2002] 2 S.C.R. 235, 2002 SCC 33.

APPEAL from a judgment of the British Columbia Court of Appeal (2001), 196 D.L.R. (4th) 427, 4 W.W.R. 404, 149 B.C.A.C. 249, 86 B.C.L.R. (3d) 4, 25 C.C.L.I. (3d) 1, [2001] I.L.R. I‑3957, [2001] B.C.J. No. 325 (QL), 2001 BCCA 130, setting aside a decision of the Supreme Court of British Columbia (1999), 1999 CanLII 6128 (BC SC), 16 C.C.L.I. (3d) 180, [1999] I.L.R. I‑3721, [1999] B.C.J. No. 1523 (QL). Appeal dismissed.

Peter H. Griffin, David Norwood and Nina Bombier, for the appellants.

David A. Critchley and Robert B. Kearl, for the respondent.

The judgment of the Court was delivered by

The Chief Justice —

I. Introduction

1This appeal concerns the proper interpretation of an accidental death benefit provision in a life insurance policy, which stipulated that coverage would be provided only for deaths effected through “accidental means”. I conclude that the circumstances surrounding Dr. Easingwood’s death support the inference that his death was effected through “accidental means”. It was caused by a miscalculation that fell within the provision covering death by accidental means. I would therefore uphold the Court of Appeal’s conclusion that the respondent is entitled to payment of the accidental death benefit, and would dismiss the appeal.

II. Facts and Background

A. Dr. Easingwood’s Death

2The insured, Dr. Edward Joseph Easingwood, was a 46-year-old family practice physician. In the course of treating a peptic ulcer, he developed an addiction to opiate medications. He completed a residential treatment program in 1994 and returned to practice in 1995. However, after a painful orthopaedic injury in the spring of 1996, he became physiologically dependent on both morphine and Demerol and had to stop work once more. His physician placed him on a program of gradual withdrawal from these drugs. He was able to return to work in mid-October of 1996. He had spoken with his friends in the days preceding his death and sounded enthusiastic and was planning for the future.

3On October 29, 1996, Dr. Easingwood told his spouse that he was going for a drive in an attempt to relieve the pain in his leg. He drove to his office. The following morning, he was found dead in the office. The Coroner described the scene of death as follows:

Dr. Easingwood’s body was found lying prone in his office with his broken glasses on the floor beside him. In his right hand was a bloody tissue. He was dressed in street clothes and his jeans were partially pulled down.

4The Coroner found that Dr. Easingwood died from an overdose caused by an intravenous injection of Demerol. The Coroner noted that the level of Demerol found in his blood was 2.4 mg, which was at the low end of the range for lethal doses. Toxicology reports indicated that phenobarbital was also found in his blood. Phenobarbital has an additive effect upon Demerol. There was no evidence to explain how the phenobarbital entered his system.

B. The Insurance Policy

5The policy, at several points, describes itself as an “Accidental Death Benefit Provision”. However, the clause granting coverage refers to deaths effected through “accidental means”:

BENEFIT

Subject to this provision’s terms, the Company will pay the amount of the Accidental Death Benefit . . . upon receipt of due proof that the Life Insured’s death resulted directly, and independently of all other causes, from bodily injury effected solely through external, violent and accidental means . . . .

6The appellant insurers maintain that Dr. Easingwood’s death was not effected through “accidental means”. They argue that his self-injection of that particular dosage of Demerol was a deliberate act, and that his death was a consequence that he must have foreseen as possible, given the amount of the dosage. The respondent, on the other hand, asserts that Dr. Easingwood’s death was “accidental”. She rejects the argument that “accidental means” is narrower than “accidental death”; and she asserts that, in any event, it is reasonable to infer that Dr. Easingwood died by “accidental means”, mistakenly believing that he was administering a non-lethal dose.

III. Judicial History

A. British Columbia Supreme Court (1999), 1999 CanLII 6128 (BC SC), 16 C.C.L.I. (3d) 180

7At trial, Josephson J. dismissed the claim for coverage under the provision. He expressed doubt as to whether there was a real distinction between insurance policies covering “accidental death” and those covering only deaths caused by “accidental means”. But he took the relevant test for “accidental means” to be whether the insured’s injuries were caused by an “unlooked-for mishap” or “an untoward event which was not expected or designed” (para. 5). Josephson J. then inferred from Dr. Easingwood’s experience as a drug-user and knowledge as a medical practitioner that he would not have been unaware of the risks posed by injecting this amount of Demerol, particularly in combination with phenobarbital. On this basis, Josephson J. concluded that Dr. Easingwood’s death was not effected through accidental means.

B. British Columbia Court of Appeal (2001), 86 B.C.L.R. (3d) 4, 2001 BCCA 130

8The Court of Appeal allowed the appeal. Writing for a unanimous bench, Huddart J.A. also questioned the usefulness of the distinction between “accidental deaths” and deaths by “accidental means”. However, she held that it was not necessary to decide this question. In her view, it was “enough, for the purposes of this appeal, to look at the action that caused the injury and all the circumstances surrounding it in a holistic way and to ask whether in ordinary and popular language the event as it happened would be described as an accident” (para. 26). Huddart J.A. then inferred from the circumstances of Dr. Easingwood’s death that it was more likely than not that he had not intended to give himself a potentially lethal dose. Because an unintentional overdose would be regarded as an accident by the ordinary person, the court held that Dr. Easingwood’s death occurred accidentally, and that the respondent could therefore recover under the policy.

IV. Analysis

A. The Distinction Between “Accidental Means” and “Accidental Death”

9The first question to be considered is whether deaths caused by accidental means form a subclass of accidental deaths. To put the question another way, is the category of deaths caused by accidental means narrower than that of accidental deaths?

10The insurers argue that the category of deaths caused by accidental means is narrower, in that it excludes accidental deaths that are the natural effects of deliberate actions. In their view, a death is only caused by accidental means when both the death and the actions that are among its immediate causes are accidental.

11This view has been adopted by Canadian courts in Columbia Cellulose Co. v. Continental Casualty Co. (1963), 1963 CanLII 476 (BC CA), 43 W.W.R. 355 (B.C.C.A.), aff’d (1964), 1964 CanLII 630 (SCC), 42 D.L.R. (2d) 401n (S.C.C.); Sloboda v. Continental Casualty Co., 1938 CanLII 204 (AB SCAD), [1938] 2 W.W.R. 237 (Alta. C.A.); Smith v. British Pacific Life Insurance Co., 1965 CanLII 91 (SCC), [1965] S.C.R. 434; Aguilar v. London Life Insurance Co. (1990), 1990 CanLII 7931 (MB CA), 70 D.L.R. (4th) 510 (Man. C.A.) (leave to appeal granted, [1991] 1 S.C.R. v; notice of discontinuance filed, [1991] 3 S.C.R. v); Leontowicz v. Seaboard Life Insurance Co. (1984), 1984 ABCA 346, 8 C.C.L.I. 290 (Alta. C.A.) (leave to appeal refused, [1985] 1 S.C.R. ix).

12This view seems to me, however, to be problematic. Almost all accidents have some deliberate actions among their immediate causes. To insist that these actions, too, must be accidental would result in the insured rarely, if ever, obtaining coverage. Consequently, this cannot be the meaning of the phrase “ accidental means” in the policy. Insurance policies must be interpreted in a way that gives effect to the reasonable expectations of the parties: Reid Crowther & Partners Ltd. v. Simcoe & Erie General Insurance Co., 1993 CanLII 150 (SCC), [1993] 1 S.C.R. 252, at p. 269. A policy that seldom applied to what reasonable people would consider an accidental death would violate this principle.

13In my view, the phrase “accidental means” conveys the idea that the consequences of the actions and events that produced death were unexpected. Reference to a set of consequences is therefore implicit in the word “means”. “Means” refers to one or more actions or events, seen under the aspect of their causal relation to the events they bring about.

14It follows that to ascertain whether a given means of death is “accidental”, we must consider whether the consequences were expected. We cannot usefully separate off the “means” from the rest of the causal chain and ask whether they were deliberate. Cardozo J. emphasized in his dissenting judgment in Landress v. Phoenix Mutual Life Insurance Co., 291 U.S. 491 (1934), at p. 501, that “[i]f there was no accident in the means, there was none in the result”. The converse is equally true: if there was no accident in the result, there can be none in the means. As Cardozo J. went on to say, either “[t]here was an accident throughout, or there was no accident at all”. Hence, to determine whether death occurred by accidental means, we must look to the chain of events as a whole, and we must consider whether the insured expected death to be a consequence of his actions and circumstances.

15This interpretation of death by accidental means accords with the ordinary meaning of the phrase. Death by accidental means is death that has been brought about unexpectedly — or that constitutes, in Lord Macnaghten’s words, “an unlooked-for mishap or an untoward event which is not expected or designed”: Fenton v. J. Thorley & Co., [1903] A.C. 443 (H.L.), at p. 448. Usually we intend the consequences of our actions. However, sometimes our actions have unintended or unexpected results. When death is the unexpected result of an action, we say that the death was “accidental”, or that it was brought about by “accidental means” as opposed to “intentional means”. In ordinary language, then, “death by accidental means” and “accidental death” have the same meaning. As Lord Stott held in Glenlight Shipping Ltd. v. Excess Insurance Co., 1983 S.L.T. 241 (Sess. (2nd Div.)), at p. 245: “[a]n accidental occurrence is merely an occurrence which has been brought about by accidental means and the problem remains the same whether the question is asked: ‘Was [the insured’s] death accidental?’ or is put in the alternative form ‘Was it brought about by accidental means?’”

16This reading of the phrase “accidental means” also accords with the principle that insurance contracts must be interpreted in a manner that gives effect, as far as is possible, to the reasonable expectations of the parties (Reid Crowther, supra). The appellants suggest that the reasonable expectations of insurers will only be respected if death by “accidental means” is read as excluding any death that is the natural result of a deliberate action. However, it is not clear that most insurers do expect the phrase to be interpreted in this narrow way, and not clear that they could reasonably expect this. This is so for two reasons. Firstly, if death by “accidental means” were read in this way, it would imply that many actions which the ordinary person would unhesitatingly classify as “accidental” means of death were not. It would imply, for instance, that a person who drinks a lethal substance believing it to be water does not die by “accidental means”; or that where, as in Glenlight Shipping, supra, a person drives his car off the deck of a ferry in the erroneous belief that it has docked and is then drowned, death is not by “accidental means”. Secondly, the expectations of the insurers are not the only expectations at issue. The insured party also has expectations that must be taken into consideration. Any adequate interpretation of “accidental means” must attempt to strike a balance between these two sets of expectations, and the two sets of interests that underlie them. Insurers cannot reasonably expect the court to adopt an interpretation that gives more protection to their interests than to those of the insured.

17Finally, this interpretation of death by “accidental means” gives full weight to the word “means”. As discussed below, it is generally accepted that accidental death policies insure against miscalculations as to whether a certain action or event may bring about death. A deliberate act that results in death because of a miscalculation may be regarded as an “accidental means” of death.

18I conclude that the phrase “accidental means” in this insurance policy does not refer to a narrow subclass of the broader category of “accidental deaths”. “Accidental death” and “death by accidental means” connote a death that was in some sense unexpected. The two phrases have essentially the same meaning.

B. What Constitutes Death by Accidental Means?

19This brings us to the central question. What constitutes death by “accidental means”? As Spence J. pointed out in Mutual of Omaha Insurance Co. v. Stats, 1978 CanLII 38 (SCC), [1978] 2 S.C.R. 1153, at p. 1164, the word “accident” is “an ordinary word to be interpreted in the ordinary language of the people”. Hence, as the British Columbia Court of Appeal emphasized in the case at bar, we must focus on the ordinary person’s understanding of the phrase, and on “whether in ordinary and popular language the event as it happened would be described as an accident” (para. 26). Only in this way can the reasonable expectations of both the insured and insurer be protected. We must therefore inquire how the phrase “death by accidental means” is used in ordinary language.

20As a starting point, we note that the accidental nature of a particular means of death depends, in ordinary parlance, on the consequences that the insured had or did not have in mind. When we speak of an “accidental” means of death, we normally have in mind a situation in which someone’s action has had results that this person did not intend or expect. Unintentional or unexpected death is seen as accidental; intentional or expected death as non-accidental. In Canadian Indemnity Co. v. Walkem Machinery & Equipment Ltd., 1975 CanLII 141 (SCC), [1976] 1 S.C.R. 309, at pp. 315-16, Pigeon J. explained the term “accident” with reference to Halsbury (vol. 22, 3rd ed.) as “any unlooked for mishap or occurrence” (emphasis in original). Similarly, in Stats, supra, at p. 1164, this Court, per Spence J. quoted Lord Macnaghten’s comment in Fenton, supra, at p. 448, that “the expression ‘accident’ is used in the popular and ordinary sense of the word as denoting an unlooked-for mishap or an untoward event which is not expected or designed”. It follows that death is not non-accidental merely because the insured could have prevented death by taking greater care, or that a mishap was reasonably foreseeable in the sense used in tort law. Nor does a death that is unintended become “non-accidental” merely because that person was engaged in a dangerous or risky activity. As this Court emphasized in Canadian Indemnity, supra, at p. 316, the jurisprudence assigns a generous meaning to “accidental”, in the absence of language to the contrary in the insurance policy.

21The pivotal question is whether the insured expected to die. The circumstances of the death – what the insured said, or did, or did not do – may point to the answer. However, to the extent that the answer is unclear when the matter is viewed solely from the perspective of the insured, the court may consider whether a reasonable person in the position of the insured would have expected to die: Candler v. London & Lancashire Guarantee & Accident Co. of Canada (1963), 1963 CanLII 155 (ON HCJ), 40 D.L.R. (2d) 408 (Ont. H.C.), at p. 423; Johnson v. Mutual of Omaha Insurance Co. (1984), 1984 CanLII 1818 (ON CA), 45 O.R. (2d) 676 (C.A.), aff’g (1982), 1982 CanLII 2087 (ON HCJ), 39 O.R. (2d) 559 (H.C.); Stats, supra, at pp. 1164-65.

22The general rule that unexpected death is accidental has been repeatedly applied. In Glenlight Shipping, supra, the insured drove off the ferry into the sea, mistakenly thinking the ferry had reached the pier. Because he “did not appreciate through error what the consequences of his action would be” (p. 243), his death was held to be accidental. In Cornish v. Accident Insurance Co. (1889), 23 Q.B.D. 453 (C.A.), the insured failed to notice an oncoming train when crossing a railway track to get from one part of his farm to another, as he regularly did. The Court of Appeal found that he met his death “by what may be properly called an accident”; though, for other reasons, it held that the insurance policy in question did not cover it (p. 455). The insured in Brown v. Continental Casualty Co., 108 So. 464 (La. 1926), was held to have died accidentally when he mistakenly inhaled more chloroform than he expected to during a procedure that he used regularly to relieve headache and insomnia. Similarly, the death of the insured in Bertalan Estate v. American Home Assurance Co. (1999), 1999 CanLII 6170 (BC SC), 68 B.C.L.R. (3d) 118 (S.C.), was held to be accidental, because he failed to remove the mask connected to his bottle of nitrous oxide in a timely fashion, as he normally did. All these cases rest on a finding that the insured did not expect or intend to die. As the cases attest, the absence of the expectation of death can arise from a variety of factors. The death may be caused by events outside the insured’s control. The insured may have made a miscalculation, or may have had mistaken beliefs about his circumstances. The insured may have failed to perform a certain action in a timely manner, or have failed to undertake a necessary check. Or the insured may simply have miscalculated the effects of his action on his body.

23The expectation test can be applied generally to all cases in which death appears to be accidental. In most of these cases, it is not difficult to determine whether death was expected or not. However, a small number of cases involving risk taking of various sorts merit further comment, if only because it is sometimes urged that they support a stricter test for whether death was not accidental, such as whether death was reasonably foreseeable, or whether the conduct posed a very high risk of death. The authorities clearly stipulate that the mere fact that someone has engaged in a dangerous or risky activity does not rule out the possibility that death was accidental, absenting special exclusion clauses in the insurance policy (Candler, supra, at p. 421; Canadian Indemnity, supra, at pp. 316-17). However, the decision to “court the risk” of death, as Spence J. phrased it in Stats, supra, at p. 1162, becomes at some point equivalent to an intention to die. Thus, when someone takes a risk that most people would expect to cause death, it is common to say of the death “That was no accident”. To say this is not to speak metaphorically, but to express a common view of where the category of accidents ends. The test does not change for cases involving conduct that brings with it a high risk of death; the test remains whether the death was designed or expected. The first question is always “What did the insured, in fact, expect?” If we cannot be sure, as is often the case, then we may ask what a reasonable person endowed with the factual beliefs of the insured and placed in the circumstances of the insured would have expected.

24In this small but difficult class of cases, trial courts must work out the results as best they can, having regard to the circumstances of death and to the wording of the policy. However, two particularly difficult types of case bear mentioning here.

25The first type of case involves people who engage in activities that carry an inordinate risk of death, whether for the psychological gratification in living on the edge, like the player of Russian Roulette (Thompson v. Prudential Insurance Co. of America, 66 S.E.2d 119 (Ga. Ct. App. 1951)), or in order to impress others with their bravado, like the young man who deliberately lay along the centre line of the highway with traffic approaching (Allred v. Prudential Insurance Co. of America, 100 S.E.2d 226 (N.C. 1957)). It is often difficult to ascertain from such circumstances what the insured’s own expectations were. However, as was stated in para. 21, where the insured’s expectations are unclear, a court may consider whether a reasonable person in the position of the insured would have expected to die. Although death is not certain to result from such dangerous activities, it is certainly within the realm of what a reasonable person would expect. In most cases of this type, then, as the Court of Appeals of Georgia said of the Russian Roulette player in Thompson, supra, at pp. 123-24, “[s]uch reckless abandon and exposure to a known, and obvious danger cannot be said to have been accidental, nor can it be said that [the] death was effected by accidental means”.

26A controversial case of this type was Candler, supra. The insured in that case had balanced himself on the coping of the patio of his 13th floor hotel suite, telling his friend “I’ll show you how much nerve I have” (p. 413). In spite of warnings by his friend, he continued to adopt various precarious positions on the coping, until he lost his balance and plummeted to his death. The policy defined the risk covered narrowly as “loss resulting directly and independently of all other causes from bodily injuries caused solely by accidental means” (p. 414). The trial judge found the death non-accidental, stating that although Candler “hoped and probably believed that he could accomplish the attempted feat without injury” (p. 422), and that to that extent, his death was unintended, nevertheless, the fact that he had performed the action in order to show off his nerve was conclusive evidence that “there was present in his mind the risk involved” (p. 422). Spence J. in Stats expressly declined to express a view on the correctness of this decision (p. 1165). One might speculate that the trial judge concluded that, despite a hope and belief that he would survive, the insured had knowingly adverted to the risk and must have, on some level, expected death. One might also surmise that the narrow terms of the policy may have led the judge to take a stricter view of “accident” than would otherwise have been the case. It suffices for us here to note that Candler does not rise to the level of changing the usual test that unintended and unexpected deaths are accidental.

27The question of whether death was expected in cases of evident or high risk must, as in all other cases, be answered from the perspective of the insured. In Johnson, supra, the insured was a physician who habitually abused an anaesthetic by strapping a bathing cap containing this solution over his face, and removing it just before reaching the point of unconsciousness. He died from asphyxiation when he failed to remove the mask in time. The defendant insurer argued that this activity would be seen by the average person as carrying a substantial risk of death, and that the insured deliberately courted this risk. However, based on the deceased’s professional experience and his long history of using this method, the trial judge concluded that “it was reasonable for him to believe that he was not risking his life or courting death” (p. 571), even though this activity “would be considered dangerous by the average person” (p. 569). The Court of Appeal upheld the result.

28There is a second type of risk-taking case that merits comment. This is the case of the rescuer who puts himself in the way of death – for instance, the swimmer who dives into the ocean, knowing of the strong undertow, to help someone who has fallen overboard; or the person who lowers herself down a well to save someone overcome by gas fumes, and is herself overcome. When the rescue is viewed in the larger context of the events that trigger it, it becomes apparent that the death is unexpected. The rescue is but part of an unexpected chain of events, triggered by the danger of death to another human being. Death is not the result of the rescuer’s intentional decision to court death as a response to the danger of another. If the rescuer dies, we do not say that his death was designed, intended or expected. Rather it was part of a tragic, accidental sequence of events. As courts have noted in the context of negligence law: “Danger invites rescue. . . . The risk of rescue, if only it be not wanton, is born of the occasion” (Cardozo J. in Wagner v. International Ry. Co., 133 N.E. 437 (N.Y. 1921)), at pp. 437-38; see also Horsley v. MacLaren, 1971 CanLII 24 (SCC), [1972] S.C.R. 441, and Corothers v. Slobodian, 1974 CanLII 187 (SCC), [1975] 2 S.C.R. 633). Rescue is born of the occasion: it is a natural human response to peril. Moreover, because the rescuer’s conduct has high redeeming social value, we can rightly demand less caution in taking on the risk of death than we would demand of the Russian Roulette player. This policy consideration, too, supports recovery.

29Cases such as these offer guidance on how the expectation test applies in different circumstances. However, coverage under an accidental death benefit policy depends not only on the circumstances but on what the insurance contract stipulates. It remains open to the insurer, as the party that drafts the insurance contract, to narrow coverage by means of explicit exclusion clauses. If an insurer wishes not to offer coverage for deaths that occur in certain circumstances — or, for that matter, for any death that results from a deliberate or voluntary action — then an explicit exclusion clause to this effect can simply be added to the contract. Insurers remain free to limit accidental death coverage in any way they wish, provided they do so clearly, explicitly, and in a manner that does not unfairly leave the insured uncertain or unaware of the extent of the coverage.

30This approach does not place an unfair burden of proof on the shoulders of the insurer to show that death was not accidental. The onus is on the plaintiff to establish a prima facie case that the death was accidental, at the risk of non suit. The plaintiff must therefore adduce evidence that permits the trier of fact to infer, on a balance of probabilities, that the insured’s death was accidental, within the ordinary meaning of that word. The tactical burden then shifts to the insurer to displace these inferences. The burden of proof never shifts, but remains squarely with the plaintiff.

V. Application to the Facts

31It remains to apply the proposed approach to the circumstances of Dr. Easingwood’s death. The essential question is whether Dr. Easingwood expected to die. The facts surrounding the death may permit inferences as to his actual expectation or intention. Insofar as this is unclear, we may ask what a reasonable person with his expertise would have expected: see Candler, supra, at p. 423; Johnson, supra, at p. 676; Stats, supra, at pp. 1164-65.

32This is a question of factual inference, and absent palpable and overriding error, a Court of Appeal will not interfere with the trial judge’s conclusions: Housen v. Nikolaisen, [2002] 2 S.C.R. 235, 2002 SCC 33, at para. 23.

33The appellants here conceded and the trial judge found at para. 2 that “[t]he insured did not intend to cause his death”. However, the trial judge went on to hold, at para. 11, that the death “was not caused by ‘accidental means’”. He further stated, at para. 14: “this could not be described as an unlooked-for mishap or untoward event which was not expected or designed. Death would not reasonably be viewed as an unlikely result.” The trial judge based this conclusion on the fact that absent evidence to the contrary, Dr. Easingwood as a physician must have been aware of the “potentially lethal effect of consuming these quantities of drugs in combination” (para. 12).

34In my respectful view, the trial judge committed several errors. First, he seems to have assumed that the mere fact that Dr. Easingwood had deliberately engaged in an activity that posed a high risk of death sufficed to render his death non-accidental. The trial judge found, at para. 11, that Dr. Easingwood was engaging in a “particularly hazardous” activity. Given this high risk, death “would not reasonably be viewed as an unlikely result” (para. 14). This set the legal threshold too low. As this Court affirmed in Stats, supra, death as a result of even highly dangerous activities may be accidental. The issue is not whether the activity was dangerous, or even whether death was likely, but whether the insured expected or intended to die. While the trial judge said the death could not be viewed as an unexpected event, he equated this with objective likelihood, not with whether the insured expected to die.

35Second, the trial judge gave no weight to the evidence indicating that Dr. Easingwood in fact did not expect to die. As discussed earlier, the test for accident is essentially subjective, although to the extent the insured’s actual intent is unknown, we must infer it from what a reasonable person in his or her position would have expected. The trial judge here seems to have approached the matter solely from the objective perspective of a reasonable person in Dr. Easingwood’s position and to have given no weight to his actual state of mind.

36The trial judge’s reasoning ignored the conceded fact that Dr. Easingwood did not intend to die. His lack of expectation to die is also supported by two sets of facts which the trial judge failed to consider.

37The first set of facts concerns the circumstances in which Dr. Easingwood’s body was found. The body was found in a dishevelled state inappropriate for someone who anticipates death as a potential result of his actions. He was lying prone in his office with his glasses broken on the floor beside him, with his jeans partially pulled down, revealing the site where he had injected the Demerol. These facts point strongly to the conclusion that Dr. Easingwood did not expect to die; indeed they suggest that he did not so much as turn his mind to the possibility that death would result from his actions. They suggest, instead, that he miscalculated how much his body could tolerate.

38The second set of facts not considered concerns Dr. Easingwood’s conduct in the days leading up to his death. The Coroner noted in her report that “[h]e had spoken to friends in the days preceding his death and sounded enthusiastic and was planning for the future”. She also noted that he admitted to his physician on September 17 that he was still taking Demerol for management of the pain caused by his orthopaedic injury. This conduct does not support the conclusion that Dr. Easingwood expected to die.

39These two sets of facts, combined with the fact the dosage of Demerol was at the low end of the scale, suggest that Dr. Easingwood was simply attempting to ease the pain in his leg and perhaps also to satisfy his addiction to painkillers, and that he would not have willingly taken on the risk of administering a potentially lethal dose.

40The trial judge’s third error was to introduce unwarranted assumptions into the reasonable person test upon which he based his conclusion. He inferred, largely on the basis of Dr. Easingwood’s knowledge as a physician and his experience with drugs, that Dr. Easingwood was aware of the potentially lethal effect of consuming Demerol and phenobarbital in combination and therefore was aware of the risk of death. However, in drawing this inference, the trial judge assumed that Dr. Easingwood deliberately consumed phenobarbital at the same time as administering the Demerol. The record provides no basis for this assumption. As discussed, the amount of Demerol he administered to himself was found by the Coroner to be a “low lethal dose”. Someone who injects such a dosage of Demerol would not necessarily believe it would cause death, even if he were an experienced physician and drug-user. In fact, the record is silent on when or how Dr. Easingwood came to have phenobarbital in his blood. Indeed, the Coroner herself drew no conclusions on this point. The trial judge was not entitled to assume that Dr. Easingwood consumed the phenobarbital and Demerol in combination, or that he was aware of the presence of phenobarbital when administering the Demerol. The mere fact that the dosage was on the spectrum of lethal doses does not show that it was probable that Dr. Easingwood would die from it. In fact, his experience may have led him to conclude that the chances of his dying were relatively small. A dosage that is lethal for one individual, with one history and one body-type, may not be lethal for another.

41I conclude that the trial judge erred in his appreciation of the law and the facts and that it is open to this Court to set aside his conclusion and to hold that Dr. Easingwood did not expect to die. The most reasonable inference from the known facts is that Dr. Easingwood simply made a miscalculation concerning how much Demerol his body could tolerate.

VI. Conclusion

42I would dismiss the appeal with costs.