By Richard Hayles

For many years Canadian courts have maintained a distinction between two types of personal accident insurance policies: policies which insure against the consequences of "accidental injuries," and those which cover injuries caused by "accidental means." The former provides coverage so long as the result of the insured's act is an unintended injury, while under the latter there is no coverage unless the incident causing the loss, as distinct from the loss itself, is accidental.

Although the distinction between accidental injuries coverage and accidental means insurance is often criticized and strikes many as somewhat artificial, it is the conclusion of this article that the distinction is still good law in several Canadian jurisdictions.

The distinction is derived from a series of 19th century English cases which were applied in two key decisions of the Supreme Court of Canada in the early 1960's. Many recent decisions in trial courts and provincial courts of appeal have either criticized the distinction or rejected it outright; other courts have insisted that the distinction is binding law because of the earlier Supreme Court decisions. As there are no recent Supreme Court cases on the issue, the law has become confused.

The situation is aptly described by Fleury, J., in a recent Ontario case:¹ 

The case law is far from clear. To use the words of Cardozo J. in Landress v. Phoenix Insurance Co., 291 U.S. 491 (1933) "this branch of the law [has been plunged] into a Serbonian bog". The dictionary defines a Serbonian bog as follows: "difficult position with no way of escape; in Nile, delta in which whole armies were said to have disappeared". When one reviews the history of the jurisprudence surrounding the definition of what constitutes "injuries caused by accidental means", "accidental injuries", "accident", one quickly understands how armies of lawyers could have disappeared into this bog.

The comment of Cardozo, J. is quoted from a dissenting opinion he delivered in a case in which the United States Supreme Court accepted the distinction between accidental injuries and accidental means. More recent American authority indicates that the distinction has now been abolished in most United States jurisdictions.² 

The time has come to reconsider the distinction between accidental injuries and accidental means in view of the criticism it has received in Canada and of its rejection by the American courts.

Accidental Injuries vs. Accidental Means

The word "accident" is an everyday word which is to be interpreted in the "ordinary language of the people"; it is not a term of art and has no technical legal definition.³ 
In its popular sense the word refers to "an unlooked-for mishap or an untoward event which is not expected or designed.⁴  " It has also been described as "something fortuitous and unexpected", and as the opposite of bodily infirmity resulting from disease or natural causes.⁵ 

Spence, J. once suggested that one could determine whether an event was an accident or not by asking what word a witness to the occurrence would use to describe it.⁶  For some reason this common sense approach to the interpretation of the word "accident" is abandoned by the courts when they are called upon to construe policies which cover losses "caused by accidental means" as opposed to "accidental" losses.

The difference between injuries which are themselves accidental and injuries caused by an accident is subtle. To understand it one must carefully examine the type of policy language which courts have held to create accidental injuries coverage and contrast it with the language which has been held to create accidental means coverage.

The coverage clause considered in Mutual of Omaha Insurance Co. v. Stats⁷  is typical of the type of language used to describe accidental injuries coverage:

"injuries" means accidental bodily injuries received while the insured is insured under the policy which result in covered loss independent of sickness and all other causes . . . .

Under this wording the loss is covered if it is unexpected. The fact that the loss may be the unforeseen result of a deliberate action has no effect on the insurer's liability.

Under accidental means coverage the event which causes the loss must itself be accidental. Injuries which are the fortuitous result of deliberate actions undertaken by the insured fall outside the coverage provided by the policy.

The coverage clause in Tracy-Gould v. Maritime Life Assurance Co.⁸  is an example of the language used to create the accidental means type of coverage:

If an employee suffers . . . (loss of life) . . . as a result of an injury suffered from accidental, external and violent means, the Insurer will pay . . . (the insurance benefit).

The reference to "external and violent" means is not necessary, and the narrower type of coverage is effected so long as the language of the coverage provision contemplates a cause of the injury separate and distinct from the injury itself. In Columbia Cellulose Company v. Continental Casualty Company,⁹  for instance, the following wording was held to create the accidental means type of coverage:bodily injury caused by an accident . . . and resulting directly and independently of all other causes . . . .

The Columbia Cellulose case is one of the two key Supreme Court of Canada decisions referred to at the beginning of this article. The other is Smith v. British Pacific Life Insurance Co.¹⁰  A review of these decisions will help to illustrate the distinction between accidental injuries coverage and insurance of the accidental means variety.

The Supreme Court Decisions

In Columbia Cellulose the insured was a business executive who had a fatal heart attack while engaged in a whirlwind inspection tour of his employer's manufacturing plants. The medical evidence as to the cause of the attack was in conflict. The physician called by the plaintiffs testified that the insured suffered from arteriosclerosis. Tissues lining the coronary artery began to bleed as a result of over-exertion, leading to the formation of a blood clot or thrombosis. This brought on the heart attack.

The doctor who testified for the defendants agreed that the insured suffered from arteriosclerosis and that the insured died of a heart attack. In his opinion, however, the narrowing of the arteries itself could have brought on the attack, and death from disease was as likely as death having resulted from over-exertion.

The trial judge accepted the evidence of the defendants' medical witness and dismissed the claim on the grounds that the plaintiffs had failed to show that the death "was a risk that the defendant assumed by the policy of accident insurance upon which the claim is based.¹¹  "

The British Columbia Court of Appeal pointed out that the trial judge seemed to have made a factual finding to the effect that death was caused by disease rather than by accident. This should have been enough to dispose of the appeal, but the court went on to consider the case on the assumption that the plaintiff's theory was correct and the heart attack was brought on by a blood clot caused by over-exertion.

The Court of Appeal reviewed the older English authorities which deal with accidental injury policies as opposed to accidental means policies. Some of these cases uphold the distinction while others do not. Essentially, the Court of Appeal in Columbia Cellulose rejected the English cases which were inconsistent with the distinction because the wording of the policies in those cases was different from the policy language before the court in Columbia Cellulose, but applied the cases which supported the distinction (although the language in those policies also differed from the policy in Columbia Cellulose).¹² 

Sheppard, J.A. concluded that the event insured against by an accidental means policy consists of three components: (1) a bodily injury; (2) an accident; and (3) that the accident cause the bodily injury. Since the cause of the bleeding of the artery was deliberate exertion by the insured, there was no accident causing the injury and the death was not covered.

This decision was upheld by the Supreme Court of Canada without written reasons.¹³ 

In Smith, the insured was a welder with a history of heart disease and arteriosclerosis. He suffered a heart attack a few months before his death, and his doctors advised him to stop working.

The insured's car became stuck in a snow drift while the insured was on a hunting trip. A friend got out of the car to dig out the tires and push while the insured stayed at the wheel, trying to assist by rocking the car and shifting from forward to reverse. The insured did no heavy work. After several hours of intermittent effort, the insured had another heart attack, which was fatal.

The coverage provisions of the policy stated that the insurance covered "loss resulting solely from Bodily Injury . . . and which is sustained during the life of this policy through Accidental Bodily Injury . . . ," and a rider to the policy provided an additional benefit for loss of life from ". . . bodily injury caused by an accident."

The beneficiary was successful at trial, but this decision was reversed by the Saskatchewan Court of Appeal. A further appeal to the Supreme Court of Canada was dismissed.

For the purposes of argument the court assumed that the stress of being stuck in the snow, the trouble involved in getting free, and the insured's rocking action caused his blood pressure to rise which triggered the attack. The question to be determined, then, was whether or not this sequence of events constituted an accident as this term was used in the policy.

The court declared that it did not, and approved the decision of the British Columbia Court of Appeal in Columbia Cellulose. Since the exertion of driving, handling the steering wheel and rocking the car was deliberate, the loss was not caused by an accident.¹⁴ 

It was open to the court in Smith to hold that the cause of death was, ultimately, the insured's car accidentally running into a snowbank and becoming stuck. Using this approach, the court could have found an uninterrupted chain of causation leading to the death of the insured. By focussing on the sequence of events which took place after this accident, the court makes it clear that only the immediate cause of the loss will be taken into account.

The principles which emerge from an analysis of the Columbia Cellulose and Smith decisions are:

(1) there is a distinction in law between "accidental injuries" coverage and "accidental means" coverage;

(2) the distinction is founded on the language of the coverage provisions of the policy ;¹⁵ 

(3) under accidental means coverage, the insured must show not only that the loss was accidental in that it was unintended, but that the cause of the loss was itself unintentional;

(4) a loss which is the unintentional result of a deliberate action is not covered by accidental means insurance;

(5) under accidental means coverage, the immediate cause of the loss must be shown to be accidental.¹⁶ 

Mutual Of Omaha V Stats

There are a number of decisions in provincial superior courts and courts of appeal, beginning in 1979, in which the principles established in Columbia Cellulose and Smith have been criticized and, in some instances, disregarded. In other provinces, courts have either expressed support for the two Supreme Court of Canada decisions or held themselves to be bound by the decisions. As a consequence, it is now in doubt as to whether or not Columbia Cellulose and Smith are still good authority.

The divergence of judicial opinion first appeared shortly after the 1970 decision of the Supreme Court of Canada in Mutual of Omaha Insurance Co. v. Stats.¹⁷  Courts critical of the earlier Supreme Court decisions have used Stats to distinguish Columbia Cellulose and Smith, and have suggested that in Stats, the Supreme Court may have overruled its earlier decisions.¹⁸ 

It is necessary to examine the Stats decision to determine whether or not the court intended to retreat from the principles established in Columbia Cellulose and Smith.

The first point that comes to light when one examines Stats is that neither Columbia Cellulose nor Smith are cited in the case report. Although it has been suggested that the court must have been aware of its earlier decisions at the time it delivered its reasons in Stats,¹⁹  it is only reasonable to suggest that the court would have referred to the earlier cases if it felt they were relevant, certainly if it intended to overrule them.

The second point derived from Stats is that the policy language under consideration was the type of language which creates "accidental injuries" coverage rather than "accidental means" coverage.²⁰  It is submitted that the court in Stats failed to refer to Columbia Cellulose and Smith because those decisions were irrelevant to the type of coverage under consideration in Stats.

The insured in Stats was killed instantly when she drove her car at high speed straight into the side of a building. Prior to the collision the insured had spent the morning visiting friends. Although the witnesses all stated that her behaviour that day was unremarkable, tests performed after her death revealed that she was grossly intoxicated.

Given that the policy covered the accidental results of deliberate or negligent acts on the part of the insured, it was not necessary for the court to consider whether or not the actions of the insured leading up to the collision could be considered an accidental cause distinct from the collision itself. Instead, the discussion focussed on cases in which the insured had been held to have deliberately "courted the risk",²¹  and the majority concluded that an event may be an accident for insurance purposes although it was brought about by the negligence of the insured.²² 

Cases Subsequent To Stats

The present confusion in the authorities arises from a series of cases in which the reasoning in Stats was applied to the "accidental means" type of coverage. Most of these cases involve alcohol related deaths, and are therefore factually similar to Stats.

The most important of these cases is CNA Assurance Co. v. MacIsaac,²³  decided one year after Stats. The insured had two policies, one of which provided accidental injuries coverage while the other provided accidental means insurance. She was found dead in her room after a night of drinking. Based on the medical evidence, the trial judge found that the insured had passed out and fallen in a awkward position which impeded breathing. He concluded that her death resulted from asphyxiation, which he ascribed to the combination of two factors: first, paralysis of the breathing centre in the brain due to intoxication, and second, collapsing in a position which made breathing difficult. He was not convinced that alcohol alone would have cause death without the awkward fall.

The trial judge found in favour of the beneficiary and the insurers appealed. The Court of Appeal, citing Stats, held that the insured had not "courted the risk" in that no reasonable person would have foreseen that death would result from the amount of alcohol she had consumed. The death was therefore an "accident" covered by both policies.²⁴ 

Hart, J.A., speaking for the court, referred to the distinction between the accidental cause of a loss and the accidental result of a purposeful act as "artificial", and said "I would make no distinction between the cause of the injury and the result and find that if the cause would not be expected by the ordinary reasonable person to produce the result which was itself unexpected, that the death would be accidental and come within the meaning of the policy".²⁵  Neither Columbia Cellulose nor Smith were referred to by Hart, J.A.

MacIsaac has been followed in subsequent decisions in the Nova Scotia courts.²⁶  In Tracy-Gould v. Maritime Life Assurance Co., the Court of Appeal took note of the fact that neither MacIsaac nor Stats made any reference to the earlier cases on "accidental means" insurance, and Matthews, J.A., said: "With deference, I cannot conclude that the attention of the court in MacIsaac was not drawn to those two decisions any more than I would conclude that in Stats the attention of the Supreme Court of Canada was not drawn to them. In my opinion, Hart, J.A. preferred for good reason, the opinion in Stats, and Stats ignored them."²⁷ 

Courts in Ontario have adopted the views of the Nova Scotia Court of Appeal.²⁸  The courts of British Columbia and Alberta, on the other hand, have held that Columbia Cellulose and Smith are still binding authority,²⁹  and the comments of the Manitoba Court of Appeal in Aguilar v. London Life Insurance Co., although obiter, indicate support for the Alberta position.³⁰ 

The decision of the Alberta Court of Appeal in Leontowicz v. Seaboard Life Insurance Company³¹  is the leading authority among the more recent series of cases that support the distinction between "accidental means" and "accidental injuries".

The insured in Leontowicz fell asleep in a car shortly after she and other guests rapidly consumed a large amount of liquor which was left over at the end of a party. She died about an hour later, and a blood test showed an alcohol level sufficient to depress the respiratory centre and stop breathing. The insured was an inexperienced drinker, and the trial judge found that her death was completely unexpected.

The act of drinking liquor, however, was deliberate, and the Court of Appeal allowed the appeal and dismissed the beneficiary's action. The policy in question covered "bodily injury caused by an accident", and the court held that it was bound by Columbia Cellulose and Smith to conclude that this language did not cover death resulting from the deliberate consumption of alcohol.

The judgment of the court was delivered by Stevenson, J.A.(now of the Supreme Court of Canada). He reviewed the relevant authorities, and noted that the issue before him did not arise in Stats because the type of coverage there was different.³²  He also inferred that the earlier Supreme Court of Canada decisions had not been drawn to the attention of the Nova Scotia Court of Appeal when it decided MacIssac.³³ 

Although he found in favour of the insurer, Stevenson, J.A. was less than enthusiastic in his acceptance of the distinction between the two types of coverage:³⁴ 

I am attracted by the robust attitude adopted by the English Courts . . . and share Cardozo J.'s anxiety . . . . I am also mindful of the general proposition that in construing insurance policies one must resolve doubt and ambiguities in favour of the insured. I accept the proposition that accident and negligence are not mutually exclusive; a negligent cause may nonetheless be an accidental one. I conclude, however, we are bound by the decisions of the Supreme Court of Canada to respect the distinction between cause and result and on those authorities, the beneficiary must be able to point to an accident. I cannot distinguish those cases, cannot ignore them, and am unable to sustain the beneficiary's suggestion that their authority is weakened by the later cases.

A motion for leave to appeal from this decision was dismissed by the Supreme Court of Canada.³⁵  As no written reasons were given, it is impossible to conclude that the Supreme Court intended to affirm its decisions in Columbia Cellulose and Smith.

Current Position

As things now stand, the appellate courts in Nova Scotia and Ontario have declared themselves free to ignore the distinction on the basis of Stats. The distinction has been endorsed, however, by the British Columbia Supreme Court and the Manitoba Court of Appeal, and it has received the reluctant approval of the Alberta Court of Appeal. There is no way to reconcile these authorities, and the law will remain uncertain until the Supreme Court of Canada provides a definitive statement.

In recent decisions, the Supreme Court has placed a new emphasis on the value of certainty in the interpretation of contracts, while showing a reluctance to rewrite contract language in order to obtain a just result.³⁶  Although this approach has been applied to the interpretation of insurance contracts,³⁷  it remains to be seen whether or not it will lead the Court to maintain the distinction between accidental injuries and accidental means. Since many lawyers and judges admit to difficulty in understanding the distinction, it is far from clear that a decision by the Court to uphold the distinction would promote certainty.

All of the courts which have dealt with the issue have done so by purporting to follow precedent. There has been little discussion as to whether or not there is any basis for the distinction in logic, justice or public policy, and these are the questions which really need to be addressed by the courts.

Analysis

The problem of causation in insurance contracts is essentially one of defining the risk, since an insurance policy responds to a loss only if it is caused by an event which the insurer has agreed to cover. For this reason causation is fundamental to the adjustment of any insurance claim, and problems of causation can never be avoided altogether.

The task of defining the risk is of equal importance to insured and insurer; the insured is interested in obtaining coverage in order to provide against a particular eventuality, but the insurer cannot agree to take on an unlimited risk. The insurer always wants to define the risk narrowly, while the insured prefers an expansive definition. Both parties, however, require a clear understanding as to what the risk is so that they may know how to govern their affairs. The current state of the law fails to meet the needs of either party to the insurance contract.

The most useful approach to the problem of causation is to ask ourselves what risk the insurer is offering to cover when it sells personal accident insurance, and what coverage the insured expects to obtain. The interests of both sides will be protected if these two objectives can be articulated as a defined risk which the insured is willing to pay for and the insurer is willing to cover.

If asked, insurer and insured would agree that the risk involved in personal accident insurance is that of an unexpected injury resulting in the disability or death of the insured. This particular risk is covered so long as the word "accident" appears in the coverage clause of the policy. What, then, is accomplished by maintaining the distinction between accidental injuries coverage and accidental means coverage? Since the distinction tends to narrow the scope of coverage, it can operate only to the detriment of the insured. It should therefore be done away with unless insurers can point to some legitimate interest which it protects.

The distinction protects insurers in two respects. First, it tends to exclude coverage in situations in which an underlying illness contributes to the loss. The facts in issue in Columbia Cellulose and Smith are examples, and it is submitted that these two cases would have been decided the same way if the coverage in question had been of the accidental injuries variety, since the plaintiffs were unable to point to any accidental injury which brought about the insureds' heart attacks.

It is now well established that a loss which is caused by a combination of injury and illness is covered by accident insurance so long as the underlying illness was dormant and would have remained so but for the supervening injury.³⁸  There is no coverage, however, where the illness itself causes loss without the occurrence of an accidental injury. This arises from the interpretation of the word "accident" itself, and the insurer's position is not advanced by the distinction between accidental injuries and accidental means.

The second type of situation in which the distinction tends to favour the insurer arises where the insured has engaged in some sort of reckless or negligent conduct which has increased the chance that an accident will occur. An instance of this would be the alcohol related death of the insured in Leontowicz.

Although it is perfectly legitimate for the insurer to try to limit its exposure to losses which arise from hazardous activities, it would be more fair if this was accomplished by means of a specific exclusion clause. A clear exclusion where drinking is associated with the accident would promote fairness in that the insured would know precisely what was and was not covered. Insurers should not be able to bring about the same result indirectly by means of the archaic distinction between accidental injuries and injuries caused by accident.

Since there is no need to continue the distinction, it should be abolished.

1. Lupia v. The Imperial Life Assurance Co. of Canada [1993] I.L.R. 1-2943 (O.C.G.D.)

2. Knight v. Metropolitan Life Insurance Co. (1968), 7 Life Cases (2d) 1143

3. Mutual of Omaha Insurance Co. v. Stats (1978), 87 D.L.R. (3d) 169 (S.C.C.), pp. 181 and 192

4. Ibid., p. 182

5. Columbia Cellulose Company v. Continental Casualty Company (1963), 43 W.W.R. 355 (B.C. C.A.), p. 359; aff'd 42 D.L.R. (2d) 401n (S.C.C.)

6. Above, note 3, p. 182

7. Above, note 3, p. 170

8. (1992), 98 D.L.R. (4th) 741 (N.S. C.A.), p. 743

9. Above, note 5, p. 357 (C.A.)

10. (1965), 51 W.W.R. 417

11. Above, note 5, p. 358 (C.A.)

12. Above, note 5, pp. 363 and 364 (C.A.)

13. Above, note 5 (S.C.C.)

14. Above, note 10, p. 423

15. A policy which uses both "accidental injuries" and "accidental means" language is ambiguous, and is to be interpreted against the interests of the insurer: Rubens v. Travelers Life Insurance Company of Canada (1992), 10 C.C.L.I.(2d) 113 (Alta. Q.B.)

16. In Ward v. Allstate Life Insurance Co. of Canada (1993), 13 C.C.L.I.(2d) 211 (B.C. S.C.) the court used the concept of proximate cause to support a finding of accidental death where the insured suffered an epileptic seizure, fell in an awkward position, and suffocated

17. Above, note 3

18. Tracy-Gould v. Maritime Life Assurance Co., above, note 8, pp. 749, 751, 752 and 755

19.Above, note 8, p. 752

20.Above, note 3, p. 176

21.See, for example, Candler v. London & Lancashire Guarantee & Accident Co. of Canada (1963), 40 D.L.R. (2d) 408 (Ont. H.C.J.)

22.Above, note 3, pp. 182 and 183

23.102 D.L.R. (3d) 160 (N.S. C.A.)

24.Above, note 23, pp. 166 and 167

25.Above, note 23, p. 168

26.Jones v. Allstate Insurance Company of Canada [1980] I.L.R. 1-1295 (N.S. S.C.); Tracy-Gould v. Maritime Life Assurance Co., above, note 8

27.Above, note 8, p. 755

28.Voison v. Royal Insurance Co. of Canada (1988), 66 O.R. (2d) 45 (C.A.); Golding v. Citadel General Insurance Co. (1990), 49 C.C.L.I. 296 (Ont. C.A.); Lupia v. The Imperial Life Assurance Co. of Canada, above, note 1

29.Tamelin v. Pioneer Life Assurance Co. (1989), 39 C.C.L.I. 306 (B.C. S.C.); Leontowicz v. Seaboard Life Insurance Company (1984), 8 C.C.L.I. 290 (Alta. C.A.), leave to appeal refused 61 N.R. 78 (S.C.C.)

30.44 C.C.L.I. 195, p. 199

31.Above, note 29

32.Above, note 29, pp. 294 and 295

33.Above, note 29, p. 297 

34.Above, note 29, p. 299

35.61 N.R. 78n 

36.B.G. Checo International Ltd. v. British Columbia Hydro and Power Authority (1993), 147 N.R. 81 (S.C.C.); Town of Fort Frances v. Boise Cascade Ltd. (1983), 143 D.L.R.(3d) 193 (S.C.C.); Consolidated-Bathurst Export Ltd. v. Mutual Boiler and Machinery Insurance Co. (1979), 32 N.R. 488 (S.C.C.). See also Mitsui Construction Company Limited v. Hong Kong (1986), 71 N.R. 285 (J.C. P.C.) 

37.Brissette v. Westbury Life Insurance Company (1992), 96 D.L.R.(4th) 609 (S.C.C.) 

38.Mitchell v. Fidelity and Casualty Co. of New York (1916), 9 O.W.N. 341 (H.C.J.); aff'd 10 O.W.N. 311 (Div. Crt.); aff'd 13 O.W.N. 181 (J.C. P.C.)

All transcript on this site is copyright © 2000  Gilbertson, Davis, Emerson LLP.
All rights reserved.
For comments or questions, please send mail to Linda Barnes