In Johnson v. Am. United Life Ins. Co., 2013 U.S. App. LEXIS 10528 (4th Cir. May 24, 2013), the Fourth Circuit reversed the District Court’s holding, a decision we reported in March of last year.   Johnson v. AUL, 2012 U.S. Dist. Lexis 32718 (M.D.N.C. 2012).   At the District Court, Magistrate Judge Patrick Auld concluded that, under the circumstances of the case, a death resulting from driving while intoxicated was not an “accident” for purposes of an Accidental Death & Disability (AD&D) benefit under an ERISA-qualified employee benefit plan.  The reversal at the Fourth Circuit illustrates again the struggle to define the word “accident,” in a situation involving a driver who intentionally becomes highly intoxicated and intentionally drives, knowing the inherent dangers, yet probably does not intend to crash, sustain injury and/or perish.

In Johnson, a participant of an employee benefit plan insured by AUL died after his truck left the road at high speed, hit a sign, and overturned several times. The post-mortem toxicology report showed a blood-alcohol concentration (BAC) of .289, more than three times the legal limit.

As did Judge Auld in the District Court below, the Fourth Circuit Court of Appeals explored a spectrum of interpretations of the word “accident.”  AUL argued for the definition adopted in Eckleberry v. ReliaStar Life Ins. Co. 469 F. 3d 340 (4th Circ. 2006), in which the Court interpreted the policy’s definition of “accident” to exclude losses from death or injury that were “reasonably foreseeable.”  Under the Eckleberry test, AUL argued, Mr. Johnson’s death was not the result of an accident because injury or death from driving while intoxicated was reasonably foreseeable.  The Fourth Circuit rejected AUL’s argument, distinguishing Eckleberry in two pivotal ways.  First, unlike the policy in the Eckleberry case, AUL’s policy did not empower it with discretionary authority sufficient to trigger the “abuse of discretion” standard of review.  Secondly, the plan in Eckleberry defined “accident” to suggest a “reasonable foreseeability” test, while, by contrast, the term "accident" was undefined in AUL's policy.  (This is not uncommon.  As the Court recognized in the seminal case, Wickman v. Northwestern Nat’l Ins. Co., 908 F. 2d 1077, 1087 (1st Cir. 1990), the word “accident” eludes articulation.)  

Reviewing AUL’s denial de novo, the Johnson Court characterized the term “accident” as ambiguous because it was undefined, and applied the contra proferentum doctrine. (Cf.  Carden v. Aetna Life Ins. Co., 559 F.3d 256, 260 (4th Cir. S.C. 2009) in which the Court held that the doctrine did not to apply in a review for abuse of discretion.). 

Moving on from Eckleberry, the Court considered two other interpretations that were  skewed “against the drafter” more than the “reasonable foreseeability" test.  The first was the test espoused by the Court in Wickman, supra:  When there was no evidence of the deceased’s actual (subjective) intentions and expectations (as is often the case), the Court asks whether a reasonable person would have viewed the injury as highly likely to occur as a result of the deceased’s intentional conduct.  If so, then the loss was not the result of an accident.  

Secondly, the Court considered the definition of “accident” under N.C.G.S. § 58-3-30(b), the test adopted by Judge Auld in the District Court.  Under this statute, which uses an “accidental result” test, a loss resulting from an intentional, voluntary act is still accidental if the injury (or result) is unanticipated and unexpected, unless the result was substantially certain to occur from the actions taken.    

When Judge Auld applied this test, he found that “a crash by a speeding driver in Mr. Johnson’s [intoxicated] condition [is] as much an anticipated and expected result as a bullet hitting the head of someone who chooses to play Russian Roulette,” (giving a nod to the Wickman Court’s illustration of an unreasonable expectation of survival, even if death were not actually intended.)  However, the Court of Appeals came to the opposite conclusion: While Eckleberry’s “reasonable foreseeable” test would most likely exclude coverage here, evidence of driving while intoxicated, even at a BAC level of .289, by itself, did not establish that the insured’s death was “substantially certain,” under the statute’s definition, or even “highly likely,” under the Wickman test.  The Court’s conclusion was based upon statistics published by the CDC that an intoxicated driver’s chances of a fatal crash are 1 in 9,128.  (The other 9,127 apparently survive.) Query:  How many drunk drivers with a BAC of .289 make it home safely? 

The proverbial “nail in the coffin” was the fact that another provision in the policy, providing a “Seat Belt Benefit,” specifically excluded recovery when the insured was driving while legally intoxicated.  The Johnson Court interpreted this exclusion to suggest that the policy’s drafters considered a death when driving while intoxicated to constitute an accident, choosing to exclude benefits based on legal intoxication only with respect to the Seat Belt Benefit, but not in general.