Josephine Murray, in No. 96-5685 v. United of Omaha Life Insurance Company Mutual of Omaha Insurance Company American Home Assurance Company Hartford Accident and Indemnity Company John Hancock Mutual Life Insurance Company Legionnaire Insurance Trust Program Insurance Company of North America. Josephine Murray v. United of Omaha Life Insurance Company Mutual of Omaha Insurance Company American Home Assurance Company Hartford Accident and Indemnity Company John Hancock Mutual Life Insurance Company Legionnaire Insurance Trust Program Insurance Company of North America. Appeal of Hartford Accident and Indemnity Company, in No. 96-5747. Appeal of American Home Assurance Company, in No. 96-5748. Appeal of Mutual of Omaha Insurance Company and United of Omaha Life Insurance Company, in No. 96-5749

145 F.3d 143 Josephine MURRAY, Appellant in No. 96-5685, v. UNITED OF OMAHA LIFE INSURANCE COMPANY; Mutual of Omaha Insurance Company; American Home Assurance Company; Hartford Accident And Indemnity Company; John Hancock Mutual Life Insurance Company; Legionnaire Insurance Trust Program; Insurance Company Of North America. Josephine MURRAY v. UNITED OF OMAHA LIFE INSURANCE COMPANY; Mutual of Omaha Insurance Company; American Home Assurance Company; Hartford Accident And Indemnity Company; John Hancock Mutual Life Insurance Company; Legionnaire Insurance Trust Program; Insurance Company of North America. Appeal of HARTFORD ACCIDENT AND INDEMNITY COMPANY, Appellant in No. 96-5747. Appeal of AMERICAN HOME ASSURANCE COMPANY, Appellant in No. 96-5748. Appeal of MUTUAL OF OMAHA INSURANCE COMPANY and United of Omaha Life Insurance Company, Appellants in No. 96-5749. Nos. 96-5685 and 96-5747 to 96-5749. United States Court of Appeals, Third Circuit. Argued Jan. 21, 1998. Decided May 19, 1998. Ralph E. Polcari (Argued), Thomas J. DiChiara, Drazin & Warshaw Red Bank, NJ, for Josephine Murray. Michael J. Zaretsky (Argued), Chorpenning, Good, Carlet & Garrison, Clifton, NJ, for Hartford Accident & Indemnity Co. B. John Pendleton, Jr. (Argued), McCarter & English, Newark, NJ, for United of Omaha Life Ins. Co., Mutual of Omaha Ins. Co. Chris E. Piasecki, Psak, Graziano, Piasecki & Whitelaw, Middlesex, NJ, for American Home Assurance Co. Joseph J. Garvey, Kelaher, Garbey, Ballou, & VanDyke, Toms River, NJ, for John Hancock Mutual Life Ins. Co. Before: BECKER, STAPLETON, Circuit Judges and FEIKENS, District Judge.* OPINION OF THE COURT BECKER, Chief Circuit Judge. — 1 — This appeal arises from a suit on five accidental death benefit (ADB) policies (with the five defendant insurance companies), each of which provides benefits for a death that is accidental and independent of all other causes. The plaintiff, Josephine Murray, is the widow of Arthur Murray who, having been admitted to a hospital for treatment of a serious ailment, died after his condition worsened following an accidental fall which resulted in a fractured hip. A jury found that Mr. Murray's death was within the scope and conditions of the policies, and returned a verdict in favor of the plaintiff. The district court, however, determined that, because Mr. Murray suffered from a number of diseases and medical conditions prior to the accident, his death was not independent of all other causes, and entered judgment as a matter of law in favor of the defendants. — 2 — At bottom, this appeal forces us to decide whether, under New Jersey law, a jury could reasonably find that Mr. Murray's death was accidental and independent of all other causes. Unfortunately, New Jersey law is opaque on what is the critical question here: whether a plaintiff can prevail if she can prove (1) that the insured's pre-existing condition or disease, though active and symptomatic, was under medical control, and that the insured was expected to live a productive life for the foreseeable future (measured in terms of years); and (2) that the accident was the direct, efficient, and predominant cause of his death. Lacking a procedure that would enable us to certify this question to the New Jersey Supreme Court, we are forced to predict how that court would decide the issue. We predict that the court would answer in the affirmative, and hence we conclude that the district court incorrectly interpreted New Jersey law. Because the evidence was sufficient to sustain a verdict under New Jersey law as we interpret it, the order of the district court rendering judgment as a matter of law in favor of the defendants must be vacated. Although the district court's charge to the jury was erroneous under the law as we now predict it, the findings necessarily implied by the jury's verdict under the incorrect instructions make clear that the jury would have reached the same conclusion under correct instructions, and thus we direct the district court to reinstate the original verdict in plaintiff's favor. I. Facts and Procedural History — 3 — On February 18, 1992, Mr. Murray, age 71, was admitted to the hospital due to swelling and pain in his foot. While hospitalized, he developed gangrene, and on March 2 his foot was amputated. For several weeks following the surgery, his temperature was high and he remained hospitalized. This fever broke on April 6, and he was scheduled to be discharged on April 11. On April 10, however, his temperature began to rise again, and when it continued to rise on the morning of April 11, the discharge plans were canceled. At approximately 11:45 p.m. on the evening of the 11th, Mr. Murray fell while walking from the bathroom to his hospital bed, resulting in a fractured hip. — 4 — Mr. Murray's temperature continued to rise on April 12 and he became gravely ill. Although he was a "high risk" for surgery due to pre-existing diabetic, heart, and liver conditions (described infra ), Mr. Murray's attending physician determined that he "[would] be totally incapacitated and probably [would] not heal without surgical intervention" on the broken hip. Mr. Murray's condition apparently stabilized over the subsequent week, and the hip surgery was performed on April 22. Following that surgery, his condition steadily worsened. He suffered acute respiratory distress, progressive sepsis, acute renal failure, and a cardiac arrest. In addition, he also experienced a left pleural effusion. This condition was caused primarily by heart failure and low protein, and only appeared after the hip fracture. Finally, Mr. Murray required an emergency hemodialysis because of hyperkalemia. He died on May 10, 1992. The death certificate listed the causes of death as renal failure, sepsis, and renal transplant. A. Mr. Murray's Prior Medical Condition — 5 — Prior to his fall, Mr. Murray had been diagnosed with a significant number of medical problems; those from which he suffered at the time of his death are catalogued in the margin.1 He had coronary artery disease, a condition relatively common among older Americans, and suffered intermittently from atrial fibrillation. While Mr. Murray's heart was not fibrillating at the time he came into the hospital, he subsequently developed partial atrial fibrillation after the accident. He also had hypertension, which was elevated on the morning of April 11, prior to the accident. Still, Dr. Scotti, plaintiff's expert, opined that there was no reason to believe that Mr. Murray was in any imminent danger of an acute heart attack. Moreover, the infarctions noted on his discharge summary, see n. 1, supra, occurred after the accident, and, according to evidence adduced at trial, were caused by the fall. — 6 — In addition, Mr. Murray suffered from renal disease, which had necessitated two earlier kidney transplants (the first in the 1970s and the second in the 1980s) and required continuous medication to suppress his immune system and to prevent his body from rejecting the transplanted organ. As a result, he had "some immune deficiency." Dr. Scotti testified, however, that Mr. Murray's kidney function was "fine" prior to the accident, and that he did not appear to be in any danger of imminent kidney failure at that time. — 7 — Sometime after his second kidney transplant, Mr. Murray developed Diabetes Mellitus. According to Dr. Nasberg, an endocrinologist, this condition was also under control prior to the accident. Additionally, he suffered from peripheral vascular disease (i.e. the obstruction or narrowing of the blood vessels to the legs and feet), which led to the transmetatarsal amputation of his foot. Dr. Scotti nonetheless testified that "there was no evidence that his foot was a problem" after the amputation. Although Mr. Murray still suffered from peripheral vascular disease after the amputation--which could have lead to complications if he was to suffer another similar injury--he did not have any active gangrene or any specific symptoms prior to his fall. — 8 — As noted supra, Mr. Murray also suffered from a fever of unknown origin while in the hospital for treatment of his foot. Although the defendant offered testimony at trial that this fever was indicative of pneumonia, and that it was that pneumonia that lead to Mr. Murray's death, Dr. Scotti testified to the contrary. In his opinion, Mr. Murray's clear chest and lack of cough prior to the accident would make the possibility of pneumonia highly unlikely. Furthermore, he had suffered from hepatitis for the fifteen-year period prior to 1992, and Dr. Scotti hypothesized that the most likely cause of the pre-accident fever was chronic active liver disease. B. Dr. Scotti's Theory — 9 — It was not disputed by Dr. Scotti at trial that had Mr. Murray been a perfectly healthy seventy-one year old man at the time of the accident, he would probably not have died as a result thereof. Yet, it was Dr. Scotti's testimony that the accident caused Mr. Murray's death. In sum, Dr. Scotti opined that the hip fracture triggered a "cascade of events" that directly lead to Mr. Murray's death--and that had it not been for the fracture, Mr. Murray would not have died on April 11. — 10 — Briefly put, the "cascade" theory works as follows. According to Dr. Scotti, all of Mr. Murray's pre-existing conditions (heart problems, kidney disease, etc.) were under control at the time he went into the hospital for treatment of his foot. The fall and subsequent hip fracture either aggravated the pre-existing conditions out of their controlled state, or caused new conditions (such as the Non-Q-wave infarction) to arise, which ultimately lead to death. For example, Dr. Scotti testified that the Fat Emboli Syndrome noted on Mr. Murray's discharge summary was specifically caused by the fractured hip.2 It was this condition which most likely caused his temperature to rise significantly after the accident. Moreover, it was Dr. Scotti's opinion that the fall ultimately spurred on Mr. Murray's liver disease (i.e. hepatitis B). — 11 — As for Mr. Murray's renal failure, it too appears that the jury could have reasonably inferred from the testimony that this problem also was triggered by the accident. Dr. Scotti testified that "the end stage renal failure was certainly a result of the cascade effect" caused by the fall. Indeed, Dr. Genovese-Stone, defendants' expert at trial, conceded that Mr. Murray was not in kidney failure the day before he fell, that there was nothing in the medical records to indicate that his kidney was about to fail, and that, with his medication, Mr. Murray's kidney condition was under control prior to the fall. — 12 — C. Mr. Murray's Insurance and the Litigation Thereon — 13 — At the time of Mr. Murray's death, he owned accidental death benefit policies worth various amounts from the five defendant insurance companies.3 Plaintiff Josephine Murray was the named beneficiary under each of these policies, and after Mr. Murray died, she filed claims for benefits pursuant to them. Each of the insurance companies independently reviewed Mrs. Murray's claim, determined that Mr. Murray's death was not accidental, and refused to pay benefits. — 14 — On May 10, 1994, Mrs. Murray filed a complaint in Superior Court, Monmouth County, New Jersey, against the insurance companies, who removed the case to the District Court for the District of New Jersey. On September 13, 1996, a jury returned a verdict in Mrs. Murray's favor. After the jury was excused, the district court granted the insurance companies' renewed motion for judgment as a matter of law, and dismissed the complaint with prejudice. Mrs. Murray appeals.4 The insurance companies crossappeal, contending that the district court erred in charging the jury. II. The District Court's Opinion — 15 — The parties have stipulated that Mrs. Murray is entitled to recover benefits under the present ADB policies only if her husband's death "was the result of bodily injury directly from an accident and independent of all other causes." — 16 — They also agree that this language must be interpreted according to the law of New Jersey. Sitting in diversity, we must predict how the highest court of New Jersey would rule. See Robertson v. Allied Signal, Inc., 914 F.2d 360, 378 (3d Cir.1990). We begin by reviewing the district court's interpretation and application of New Jersey's law in this matter. — 17 — In its opinion granting judgment as a matter of law to the defendants, the district court read New Jersey law to require a case by case evaluation which takes into account both the mechanism by which death occurs and the "nature and severity of the underlying conditions or diseases." (See Amended Transcript of Trial, Sept. 13, 1996 at 12-13 [hereinafter "Transcript "].) The "telling rule" derived from the New Jersey cases, according to the district court, is that: — 18 — where the medical proofs establish affirmatively that the active disease, with which the insured was afflicted and for which he was being treated, not only was competent to contribute to his death, but in the opinion of his own physician, did in fact operate in conjunction with an accidental injury to produce the demise, the Court has no alternative but to take the case from the jury. — 19 — From this, the district court found New Jersey to prescribe a two-pronged test. First, in order to recover, the insured must show that his pre-existing conditions were "inactive and under control, and were not sufficient to cause death." In the district court's view, if any pre-existing condition was "active, known, symptomatic, or progressive, and of such a nature and severity in and of itself so as to be capable of contributing to death," this first element would not be met. Second, the district court held that the insured must show that the accidental injury was the "direct, efficient and predominant cause of death, in that it set in progress the chain of events leading directly to death by exciting or triggering the pre-existing conditions into activity, and thereby hastened death so as to cause it to occur at an earlier period than it would have occurred but for the accident." — 20 — The district court then reviewed the evidence. Most notably, the district court referenced Dr. Scotti's testimony that all of Mr. Murray's conditions were controlled at the time of the fall; that these conditions were not likely to cause death at the time death occurred; and that, in Dr. Scotti's opinion, the fall was the proximate cause of Mr. Murray's decline, which led to his death. Indeed, according to Dr. Scotti, Mr. Murray would have lived for several more years but for the accident.5 Based on these facts, the district court concluded that judgment as a matter of law was not warranted on prong two of the New Jersey test (i.e. causation). — 21 — Nonetheless, the district court found that Dr. Scotti's testimony did not provide sufficient evidentiary support for plaintiff to meet her burden on the first prong of the test. This was because several of the diseases and conditions from which Mr. Murray was suffering, including arteriosclerotic heart disease, peripheral vascular disease, diabetes, hepatitis B, and renal failure, were "active" and because they, singularly or in combination, together with the accidental injury, in some manner were contributing causes of Mr. Murray's death. Moreover, the district court relied upon Dr. Scotti's testimony that if Mr. Murray had been a "perfectly healthy 71 year old," then it was much less likely that he would have died from the hip fracture, that his death was made "much more likely" by his combination of existing medical problems at the time of the accident, and that in Dr. Scotti's opinion there was "no way that the fall was the sole cause of death." While Dr. Scotti testified that all of Mr. Murray's pre-existing conditions were under control at the time of the accident, the district court found that — 22 — the particulars, the specifics of Dr. Scotti's testimony make it clear that he acknowledged that several of Mr. Murray's conditions were, in fact, not under control at the time of the accident, and that these conditions were among the conditions which contributed to his ultimate demise. — 23 — On this basis, the court found for defendants. — 24 — The New Jersey Supreme Court has never addressed the ultimate question before us, and there are conflicting strains in New Jersey law. Nonetheless, we predict that under these circumstances, the New Jersey Supreme Court's approach to resolving this dispute would differ from the approach utilized by the district court. — 25 — III. The New Jersey Jurisprudence on the Interpretation of Accidental Death Benefit Clauses — 26 — The question before us is whether the district court correctly predicted how the New Jersey Supreme Court would rule on the present facts. The district court deduced its two-pronged test from three cases dealing with the construction of ADB clauses in New Jersey: Runyon v. Commonwealth Casualty Co., 109 N.J.L. 238, 160 A. 402 (1932) ("Runyon II "); Kievit v. Loyal Protective Life Ins. Co., 34 N.J. 475, 170 A.2d 22 (1961); and Tomaiuoli v. United States Fidelity and Guaranty Co., 75 N.J.Super. 192, 182 A.2d 582 (1962). We will look to these cases and others to determine how the New Jersey courts have treated the construction of ADB limited coverage clauses in the past. Since, as we discuss infra, the holdings of these cases are largely dependent upon their particular facts, we will also make reference to modern intrajurisdictional trends and scholarly opinions on how these clauses should be construed. — 27 — A. The Strict Construction Model--Runyon I & II — 28 — The first major New Jersey opinions treating issues similar to those before us were the related cases of Runyon v. Monarch Accident Ins. Co., 108 N.J.L. 489, 158 A. 530 (1932) ("Runyon I ") and Runyon II, supra. Both cases dealt with the same accident, in which the insured slipped on an icy pavement, broke his hip, and died about five weeks later. The evidence showed that the insured had suffered for eleven years prior to the accident from a condition then known as "paralysis agitans" (known more commonly now as Parkinson's Disease), and the death certificate listed the cause of death as "fractured left hip from slipping on ice; contributory paralysis agitans (secondary) duration 11 years." Runyon I, 158 A. at 532. The insured's physician testified at trial that his use of the word "contributory" meant that if the insured had been in normal health, he would have "stood a much better chance" to recover from the shock of the hip fracture. See id. — 29 — The ADB policy in Runyon I provided that the defendant was liable only if death resulted "exclusively from bodily injuries caused solely by external, violent and accidental means." Id. at 531. The court held that "under such a policy, if the insured, at the time of the accidental injury, was also suffering from a disease, and the disease aggravated the effects of the accident, and actively contributed to the death occasioned thereby, there can be no recovery upon the policy." Id. On this basis, the court upheld the verdict for the defendant. — 30 — Runyon II involved the same facts, though construed under a different insurance policy. The relevant policy language in that case provided for benefits so long as death or bodily injury was caused: — 31 — directly and independently of all other causes by external, violent, and accidental means, which bodily injuries, or their effects, shall not be caused wholly or in part, directly or indirectly, by any bodily or mental disease, defect or infirmity. — 32 — Runyon II, 160 A. at 403. The trial court denied the defendant's motion for a nonsuit, and the jury returned a verdict for the plaintiff. The New Jersey high court reversed.6 Since it was uncontroverted that the insured's pre-existing condition had, to some extent, caused the death of the insured, the court held that the case fell squarely within the exempting condition of the policy--i.e. that the death was not caused directly and independently of all causes other than the accident. See id. B. The Modern Model--Mahon and Kievit — 33 — About thirty years after Runyon II, the New Jersey Superior Court Appellate Division broke from Runyon 's apparently strict dictates in Mahon v. American Casualty Co. of Reading, 65 N.J.Super. 148, 167 A.2d 191 (1961). The facts of Mahon can be simply stated. The deceased, a nine year old boy who appeared to be in good health, was playing in his schoolyard during recess when he accidentally bumped his head against the head of a schoolmate. Soon thereafter, the boy developed symptoms prompting a medical examination. That examination and subsequent diagnostic operations revealed that the boy had a condition known as the Arnold-Chiari malformation. According to expert testimony at trial, the malformation was caused by a tumor deep within the boy's brain; while the tumor and the malformation existed prior to the head injury, they were in a "quiescent" state and it took the trauma of the head injury to cause the boy's acute symptoms. See id. at 194-95. Moreover, the expert testified that but for the tumor, the acute symptoms would not have occurred; yet, but for the head trauma, the tumor would not have caused the malformation to become acute "for a further interval of time." See id. at 195. — 34 — The court recognized that cases like Runyon II reflected a "pronounced tendency" in the law "to hold for the insurer as a matter of law, if there is uncontroverted evidence of causal contribution by disease or abnormality to the loss." Id. at 198. At the same time, the court noted a "not inconsiderable number of decisions" which held that such evidence could still support a verdict for the insured if the jury was to find that the accident operated as the proximate or predominant cause of the loss. See id. The court also distinguished between two types of limitation clauses found in ADB policies. The first (found in Mahon and in the stipulation here), known as a "limited coverage clause," is typically of the form: "loss resulting directly and independently of all other causes from injury caused by accident." Id. at 196. The second (found in Runyon II ) is known as an "exclusionary clause," and is of the form: "which bodily injuries, or their effects, shall not be caused wholly or in part, directly or indirectly, by any bodily or mental disease, defect, or infirmity." Id. Examining cases from several jurisdictions, the court concluded that the exclusionary clause constitutes a significantly more restrictive contract than the limited coverage clause. See id. at 199. — 35 — In light of these considerations, Mahon held that the "mere conjunction of disease or abnormality and accident, each 'but for' causes of the resulting disability, and neither alone efficient to produce it" does not necessarily bar recovery as a matter of law in a limited coverage case. See id. at 200. The court opined: — 36 — In what seems to us a preponderance of American jurisdictions, the test is whether the accidental injury as contrasted with the contributing disease or bodily condition, is the proximate cause of the disability or loss.... Pervading such cases is the philosophy that if the accident is a more substantial contributing cause of the resultant disability or death than the disease, the latter merely being a condition thereof, recovery is allowed. — 37 — Id. at 201. The court held that this result was not foreclosed by Runyon I. Although Runyon I mentioned a "general rule" that recovery should be barred if a preexisting disease aggravated the accident and "actively contributed" to the death occasioned by the accident, Mahon 's review of the case law suggested that Runyon I 's statement could not reflect the "general rule" unless "active contribution" was understood to mean "predominant cause." Mahon, 167 A.2d at 205. The court also distinguished Runyon II on the basis that the policy in that case involved an exclusionary clause, whereas the policy in Mahon only involved a limited coverage clause. We do not rely on this distinction, however, because, as we note below, the modern trend (at least in New Jersey) is to treat both clauses the same. See infra at n. 7. — 38 — The New Jersey Supreme Court evinced an intent to follow the modern liberal trend outlined in Mahon in Kievit v. Loyal Protective Life Ins. Co., 34 N.J. 475, 170 A.2d 22 (1961), the court's most recent treatment of the issue at hand. The insured in Kievit was a forty-seven year old carpenter, who appeared to be in perfectly good health, without any diagnosed diseases or conditions, and with no symptoms. The accident occurred when the insured was struck over the left eye by a two-by-four, after which he developed "tremors" and became totally disabled. Kievit, 170 A.2d at 24. Testimony at trial indicated that the insured actually had (although asymptomatic) Parkinson's Disease prior to the accident, which the accident had aggravated. See id. at 25. — 39 — The trial court entered judgment for the defendant insurance company and the Supreme Court reversed. This case, the court held, was distinguishable from both Runyon opinions. In those cases: — 40 — a patent, active disease was found to have contributed with the accident to the resulting death of the insured. We are here concerned with a latent, inactive condition or disease which was not accompanied by any symptoms and which was precipitated or activated by the accident into a resulting disability. — 41 — Id. at 27. Although the policy held by the insured in Kievit contained an exclusionary clause, the court did not find that Mahon was distinguishable on this basis.7 Rather, the Kievit court held that the key was the reasonable expectations of the insured. See id. at 30 ("[T]he court's goal in construing an accident insurance policy is to effectuate the reasonable expectations of the average member of the public who buys it."). — 42 — In what we consider the most critical passage of the opinion, the court held: — 43 — When the Company issued its accident policy to Mr. Kievit it knew that, although he was then about 48 and in good health, he could and presumably would keep the policy in force until he was 65 and that in the course of time he would undoubtedly be subjected to bodily conditions and diseases incident to the aging process. If the terms of the policy were read literally, the policy would be of little value to him since disability or death resulting from accidental injury would in all probability be in some sense contributed to by the infirmities of age. ... Such literal reading was never contemplated and it may fairly and justly be concluded that it also was never contemplated that indemnity would be unavailable where a condition or disease which was wholly dormant was activated and became disabling as the result of an accidental injury. — 44 — Id. at 30 (emphasis added). The court then further clarified its position by quoting the following passage from United States Fidelity & Guaranty Co. v. Hood, 124 Miss. 548, 87 So. 115, 120 (1921): — 45 — It is not sufficient to defeat the policy that the accident may have made some latent disease active, which disease contributed in some degree to the death. If the disease was active and of such character and virulence as to endanger life apart from the accident, but might not have done so had the accident not happened, then that may be said to be a proximate contributing cause. — 46 — Kievit, 170 A.2d at 30. Based on this rationale, the court held that the accident was the proximate cause of the insured's disability, and that his pre-existing disease, "activated by the accident into an incapacitating condition," was not a disqualifying contributing cause. Id. at 31. — 47 — Thus, Kievit and Mahon can be read most plausibly to hold that the construction of limited coverage clauses in ADB policies depends upon some form of proximate causation analysis in order to fairly represent the reasonable expectations of the insured. At the very least, if the pre-existing condition is found to have been "active" (as opposed to the "latent" condition at issue in Kievit ), these cases would appear to require a factual determination whether the condition was "of such character and virulence as to endanger life apart from the accident" in some relevant medical sense, or whether the accident precipitated the condition into a resulting disability. See Kievit, 170 A.2d at 27, 30. — 48 — This approach suggests what scholars have deemed the modern tendency of courts dealing with limited coverage clauses like these: — 49 — It has thus been the tendency to hold that where the disease was not a direct, proximate, or concurring cause of the loss recovery would be allowed, regardless of the existence of such condition.... Other courts have taken a still broader view, consistent with that which the authors urge herein as being the better approach, that recovery may still be had where the diseased condition appeared actually to contribute to cause the death, where the accident was the prime or moving cause. This has come to be the more modern rule, irrespective of the stringencies of policy language, where injury is a proximate cause of death or disability, even though the result for which claim is made would, perhaps, not have occurred except for the preexisting condition. — 50 — John Alan Appleman & Jean Appleman, Insurance Law and Practice § 393 at 81, 85-90 (1981); see also Robert E. Keeton & Alan I. Widiss, Insurance Law § 5.4(b)(2) at 502 (1988) (stating that in ADB cases, "courts tend to interpret the coverage provisions and limitations so as to favor the interests of the beneficiaries when the evidence indicates the death or injury was essentially fortuitous."). For the reasons discussed infra, we believe that the New Jersey Supreme Court would follow this modern approach in the present case. — 51 — C. The Appellate Division Returns to Runyon--Tomaiuoli — 52 — Shortly after Kievit was handed down, the Appellate Division revisited the Runyon approach in Tomaiuoli v. United States Fidelity & Guaranty Co., 75 N.J.Super. 192, 182 A.2d 582 (1962). In that case, the insured, a seventy-two year old man who had a history of arteriosclerotic heart disease and diabetes, was involved in a minor traffic accident, from which he suffered no bodily injuries. Shortly thereafter, while still at the accident scene, the insured collapsed on the sidewalk from a heart attack, and died before an ambulance could arrive. The plaintiff, relying upon expert medical testimony, claimed that the car accident led to a "chain of events" that caused the insured's death. See id. at 586. The court ultimately found that the accident was "the precipitating cause of exciting the decedent to a degree greater than he was able to withstand physically by reason of the underlying systemic maladies with which he was afflicted, with the consequence that he collapsed." Id. at 587. The insurance policy, as in Runyon II, provided for the payment of benefits for accidental injuries resulting in death only if the loss resulted "directly and independently of all other causes." Tomaiuoli, 182 A.2d at 588. — 53 — The appellate division found this case to be factually and legally indistinguishable from Runyon II, and affirmed the trial court's judgment notwithstanding a verdict in favor of the defendant. The court found that in both cases: — 54 — the insured persons were, and had been, suffering from active diseases, progressive in nature, capable of producing fatality, and presenting symptoms which brought home to the victims and their respective doctors knowledge of their existence. Moreover, in both cases the treating physicians confessed inability to separate the effects of the bodily injury from the effects of the pre-existing active disease, and felt obliged to conclude that the total of such effects in combination produced the death. — 55 — 182 A.2d at 588. In this light, the court held that: — 56 — Where, as here, his medical proofs establish affirmatively that the active disease with which he was afflicted, and for which he was being treated, not only was competent to contribute to his death, but in the opinion of his own physician, did in fact operate in conjunction with an accidental injury to produce his demise, the court has no alternative but to take the case from the jury. — 57 — Id. at 590. — 58 — Thus, in the view of the Tomaiuoli court, Runyon I and Runyon II do not permit recovery when the pre-existing condition is an "[a]ctive, patent progressive disease which in its very nature is competent to contribute to death." Id. — 59 — While the court recognized that Kievit and Mahon "stand strongly for liberality of construction," it found that those cases only compelled a departure from the Runyon approach when the pre-existing condition was "[l]atent or [d]ormant, [a] symptom-free condition, possibly aggravated or exacerbated by bodily injury." Id. D. Application to the Present Case — 60 — Not surprisingly, defendants contend that our disposition of this case should be controlled by Runyon II and Tomaiuoli. They submit that the New Jersey Supreme Court would follow Tomaiuoli 's reading of the Runyon cases, Kievit, and Mahon and hold that, if a known and active preexisting condition, capable in itself of producing fatality, actively contributes to the insured's death, recovery under the applicable ADB policies would not be permitted. Accordingly, they would have us affirm the district court's determination that the fact finder must determine as a threshold matter whether the pre-existing disease falls into one of two categories: either the disease is (1) an active, patent, progressive disease that by its very nature is competent to contribute to death; or it is (2) latent or dormant and symptom free. Under this test, if the condition falls within category (1), there can be no liability. According to the defendants, the facts of the present case, unlike Kievit, could not fall within category (2). To the contrary, they contend that this case is factually indistinguishable from Runyon II and Tomaiuoli, placing it squarely within category (1). — 61 — Although Tomaiuoli clearly suggests that we should construe Runyon broadly and find no recovery any time the pre-existing condition is active and symptomatic, see 182 A.2d at 588-89, we predict that the New Jersey Supreme Court would reject this view in light of Kievit 's emphasis on the reasonable expectations of the insureds. Tomaiuoli, to be sure, recognizes the reasonable expectations doctrine, and the "liberality of construction" for which Kievit and Mahon stand. See Tomaiuoli, 182 A.2d at 590. We believe, however, that Tomaiuoli unduly limits the reach of both of these principles. Accordingly, we predict that the New Jersey Supreme Court, if faced with the present facts, would not adopt Tomaiuoli. See generally Nationwide Mutual Fire Ins. Co. v. Pipher, 140 F.3d 222 (3d Cir.1998) (predicting that Pennsylvania Supreme Court would not extend holding of its earlier decision as suggested by intervening decisions of other courts). — 62 — We observe first that Kievit clearly rejected the proposition that there can be no liability if any known pre existing condition at all contributed to the insured's death. If that was the case, then an insured could only recover if he or she was in perfect health at the time of the accident. As another court, interpreting Kievit and other similar cases, has eloquently stated: — 63 — If the phrase is given literal effect, only the healthiest of individuals would be given the protection of their policies. Those suffering from even the slightest preexisting medical condition would be precluded from benefits--the purchased coverage would be illusory. The court will not construe the contract to defeat, rather than promote, the purpose of accident insurance. The court therefore concludes that literal application of the phrase "direct result, independent of all other causes" defeats the reasonable expectations of insureds. — 64 — Henry v. Home Ins. Co., 907 F.Supp. 1392, 1397 (C.D.Cal.1995). Such a result could not be consistent with the reasonable expectations of the insured. See Kievit, 170 A.2d at 26, 30; see also Sparks v. St. Paul Ins. Co., 100 N.J. 325, 495 A.2d 406, 414 (1985) ("The interpretation of insurance contracts to accord with the reasonable expectations of the insured, regardless of the existence of any ambiguity in the policy, constitutes judicial recognition of the unique nature of contracts of insurance."). — 65 — We recognize that this observation may be in tension with a literal reading of the Runyon decisions. However, those cases were based upon a strict construction approach to limited coverage and exclusionary clauses which we do not believe is tenable after Kievit specifically applied the reasonable expectations doctrine to this area of New Jersey insurance law. See, e.g., Werner Industries, Inc. v. First State Ins. Co., 112 N.J. 30, 548 A.2d 188, 191 (1988) ("At times, even an unambiguous contract has been interpreted contrary to its plain meaning so as to fulfill the reasonable expectations of the insured."); Kievit, 170 A.2d at 30 (rejecting literal interpretation of limited coverage clause).8 Accordingly, we predict that the Supreme Court of New Jersey would not apply a literal reading of the Runyon opinions to this case. — 66 — The question remains, however, whether that court would agree with Tomaiuoli that recovery can only be had if the pre-existing condition was latent, dormant, and symptom free. We predict not. Consider, for example, a situation in which an individual has been diagnosed with a slow-moving cancer that is likely to cause his death in ten or fifteen years and which is thus, by definition, not inactive. Let us further assume: (1) the disease carries with it certain symptoms, all of which are presently controlled by medication and treatment; and (2) the individual has been living a normal and productive life and is expected to do so until the cancer finally progresses to its terminal stage (i.e. in ten to fifteen years). Assume also that the individual suffers an entirely fortuitous accident and, due to serious injuries suffered in that accident, his previously controlled condition is aggravated, resulting in death. We doubt that, in light of the reasonable expectations doctrine, New Jersey would hold that his beneficiaries should be deprived of accidental death benefits just because his pre-existing disease, which acted in combination with the accidental injury to cause his untimely death, was not "inactive and symptom free." — 67 — What this hypothetical points out, we believe, is that the patent/latent categorization scheme suggested in Tomaiuoli is too simplistic, and yields results inconsistent with the reasonable expectations of insureds. As Appleman & Appleman, supra, have noted, while a move away from this scheme "is not as satisfactory as the ability to decree a clean line of demarcation, it might be more consistent with the realities of existence." Appleman & Appleman, supra, § 391 at 52-3. And, we add, it is more consistent with the ability of modern medicine to prolong life. While Tomaiuoli was correct in recognizing that Kievit distinguished its facts from the Runyon facts along the patent-latent axis, we believe--and predict that the Supreme Court of New Jersey would similarly believe--that Tomaiuoli erred (an error which the district court here followed) in raising this essentially factual distinction to the level of legal categorization. — 68 — It is important to note also that all of the New Jersey precedents discussed above are heavily fact-bound, relying upon the precise nature of the insured's pre-existing condition; the relationship between that condition, the type of accident, and the resulting disability; and the language of the applicable insurance policy. In none of these cases were the New Jersey courts faced with a factual scenario like the cancer hypothetical or the present case, in which the defendant suffered from arguably "active" diseases which the jury could find, based on medical expert testimony, were reasonably under control by virtue of the insured's then-existing medical treatments.9 While such a condition is not "dormant" in any technical sense of the word, it is functionally like a "dormant" condition in the important respect that it did not pose an immediate threat of death until triggered by the injury.10 Of course, such an analogy may or may not be appropriate on the facts of a particular case--as noted in Kievit, just because an active or virulent pre-existing condition might not have killed the insured had the accident not happened does not mean that the condition is not a proximate contributing cause. See Kievit, 170 A.2d at 30. But this is a question of fact, not a question of law. — 69 — Bearing in mind New Jersey's broad and liberal construction of insurance policies in favor of the insured, see, e.g., Dittmar v. Continental Cas. Co., 29 N.J. 532, 150 A.2d 666, 672 (1959), its adoption of the doctrine of reasonable expectations, and the harshness with which the rule advanced by these insurers would operate under some factual scenarios, we do not believe that the New Jersey Supreme Court would adhere to the insurers' mechanical reading of New Jersey law. To the contrary, we believe that the court would hold that in cases such as the present one, in which there is medical evidence to a reasonable degree of certainty that the insured's relevant pre-existing conditions, even if active and symptomatic and capable of ultimately causing death, were under control at the time of the accident, and that the insured was expected to live a productive life for the foreseeable future (measured in terms of years), it is a question of fact for the jury to decide whether the pre-existing condition or the accident was the cause of the defendant's death or disability under an ADB policy. — 70 — The remaining question is the standard of causation. The traditional term is "proximate cause," but there appears to be some confusion in the case law and scholarly commentary as to how that term should be defined in this context. More specifically, there is some difference of opinion whether it should be defined in the same way as that term is used in standard negligence cases, or whether it should be defined as the "predominant cause." See, e.g., Appleman & Appleman, supra, § 362 at 484-85, 487 ("it is necessary only that the accident stand out as a predominant factor in producing the loss") (citing cases); Carroll v. CUNA Mutual Ins. Society, 894 P.2d 746, 755 (Colo.1995) (en banc) ("Courts when using the word 'proximate cause,' however, seem to intend no more that to distinguish between remote and predominant causes."). — 71 — The Mahon court appears to conflate the two terms: — 72 — [T]he test is whether the accidental injury, as contrasted with the contributing disease or bodily condition, is the proximate or predominant cause of the disability or loss, sometimes additionally qualified as the active, efficient, dominant, originating, or direct cause. — 73 — Mahon, 167 A.2d at 201. The New Jersey Supreme Court in Kievit uses the term "proximate cause" without defining it, although the court does discuss Mahon 's definition and seems to approve of it. See Kievit, 34 N.J. at 487-90, 170 A.2d at 29-31 (noting Mahon 's definition of proximate cause as the "direct, efficient, and predominant cause."). In consideration of the foregoing, we predict that the New Jersey Supreme Court would define "proximate cause" in this setting as the direct, efficient, and predominant cause of the insured's death. We note that this was the test the district court applied in instructing the jury and ruling on defendants' motion for judgment as a matter of law. — 74 — Adoption of the rule we have described would bring New Jersey in line with the modern approach taken by courts of other jurisdictions. See Appleman & Appleman, supra, at § 393; see also, e.g., Carroll v. CUNA Mutual Ins. Society, 894 P.2d 746, 755 (Colo.1995) (holding that benefits are recoverable "as long as one can show that the accident is the predominant cause of the [death]."); Life Ins. Co. of North America v. Evans, 195 Mont. 242, 637 P.2d 806, 808-09 (1981) ("Recovery may be had even though the disease appears to have actually contributed to the cause of death as long as the accident sets in motion the chain of events leading to death, or if it is the prime or moving cause."). We believe that New Jersey would follow this course. — 75 — Under this standard, the evidence adduced by the plaintiff would support a verdict in her favor. Although it is undisputed that some of Mr. Murray's conditions were active and symptomatic and capable of ultimately causing his death, Dr. Scotti testified that they were under control, that Mr. Murray would have lived with them for years, that the accident caused an exacerbation of the pre-existing conditions, and that the accident was the predominant cause of death.11 If so, the death could be found by the jury to be proximately caused by an "accident and independent of all other causes" within the stipulated limited coverage clause of the policy as we have construed it, and the judgment as a matter of law in favor of the defendants must be vacated. IV. The Jury Charge — 76 — The defendants contend that, even if the district court improperly granted judgment as a matter of law, the jury rendered its verdict pursuant to erroneous instructions, and a properly instructed jury would have had no choice but to find in their favor. They therefore assert that they are entitled to judgment as a matter of law on this alternative ground. — 77 — We agree that the jury charge as given by the district court did not conform to New Jersey law as we predict it.12 Nonetheless, we will not reverse a judgment where "it is highly probable that the error did not contribute to the judgment," McQueeney v. Wilmington Trust Co., 779 F.2d 916, 924 (3d Cir.1985), i.e., where the challenged error was harmless. We conclude that the findings necessarily implicit in the verdict of the jury compel the conclusion that the jury would have reached the same result had it been instructed according to the correct legal standard as we have explained it. Therefore, the error in the instruction was harmless.13 Cf. Commercial Union Ins. Co. v. International Flavors and Fragrances, Inc., 822 F.2d 267, 275 (2d Cir.1987) (if jury's findings would support verdict under proper instructions "no useful purpose could be served by submitting the same evidence to another jury"); H.C. Blackwell Co. v. Kenworth Truck Co., 620 F.2d 104, 107 (5th Cir.1980) (same); 9A Charles Alan Wright & Arthur R. Miller, Federal Practice and Procedure § 2540 (1995) ("If the trial court erroneously grants a renewed motion for judgment as a matter of law under Rule 50(b), the appellate court may reverse and order reinstatement of the verdict of the jury."). — 78 — In relevant part, the jury was instructed as follows: — 79 — You are instructed with respect to this first element that if you find that the disease condition of the insured at the time of the accident was either active, known, symptomatic, or progressive, and of such a nature and severity in and of itself so as to be capable of contributing to his death, then plaintiff has not met plaintiff's burden of proof with respect to this element, and you should end your deliberations and return a verdict for defendants. — 80 — If, on the other hand, you find that at the time of the accident, the disease condition of the insured was inactive and under control, and was not sufficient to cause death, then plaintiff has met plaintiff's burden of proof with respect to the first element, and you should proceed to consider the second element; which is, second, that the accidental injury was the direct, efficient and predominant cause of death, in that it set in progress the chain of events leading directly to death by exciting or triggering the pre-existing condition into activity, and thereby hasten death so as to cause it to occur at an earlier period than it would have occurred but for the accident. — 81 — According to this instruction, for the jury to have returned a verdict for the plaintiff, it necessarily must have found that, at the time of the accident, Mr. Murray's pre-existing conditions were: (1) inactive, (2) under control, and (3) not sufficient to cause death. — 82 — As we explained supra, whether the insured's conditions were "active" or "sufficient to cause death" is not dispositive. An insured suffering from conditions that are active, symptomatic, and ultimately capable of causing death may still recover under an ADB policy for an accidental death so long as there is medical evidence to a reasonable degree of certainty that the insured's conditions were under control at the time of the accident. Since the jury found that Mr. Murray's conditions were under such control, we have no reservations in concluding that the jury would also have found for the plaintiff under the standard as we have articulated it. — 83 — This is not the end of the analysis, however. As quoted above, the district court also instructed the jury that it must find that the accidental injury was the "direct, efficient, and predominant cause of death." Since we have predicted that the New Jersey Supreme Court would apply this same causation test, this prong is also satisfied. Indeed, the district court found that there was sufficient evidence to support the jury's conclusion on the causation prong under the instructions given. Since we have found congruence between the findings implicit in the jury's verdict and both the control and causation prongs of our standard, in addition to vacating the judgment of the district court, we will remand with instructions to reinstate the verdict in favor of plaintiff. — 84 — The parties shall bear their own costs. Footnotes [-s] * Honorable John Feikens, United States District Judge for the Eastern District of Michigan, sitting by designation [1] The discharge summary prepared by the hospital after Mr. Murray's death lists fifteen conditions under the heading "Final Diagnoses." That list includes: uremia; cellulitis of the left foot with gangrene; intertrochanteric and subtrochanteric fracture, right hip; peripheral vascular disease, severe; diabetes mellitus; end-stage renal disease, secondary to autosomal dominant polycystic kidney disease; status post kidney transplant; acute renal failure, secondary to sepsis; atherosclerotic heart disease; fever of undetermined origin; left pleural effusion; hepatocellular necrosis, idiopathic; fat emboli syndrome; pneumonia; acute, non-Q wave myocardial infarction [2] Dr. Scotti described Fat Emboli Syndrome as a condition in which the fatty materials from the patient's bone marrow escape into the bloodstream. These materials can then form small clots in the patient's lung and block the lung vessels. App. at 72A [3] Those policies (and their values) are the following: United of Omaha Life Insurance Company ($3,995.00); Mutual of Omaha Insurance Company ($80,200.00); American Home Assurance Company ($25,000.00); Hartford Life Insurance Company ($75,000.00); John Hancock Mutual Life Insurance Company ($26,250.00). Mr. Murray also owned accidental death policies from Legionnaire Insurance Trust Program and the Insurance Company of North America. Although these latter two companies were named in the complaint, they settled with Mrs. Murray and are not parties to this appeal [4] Removal of this action from state court to the district court was appropriate under 28 U.S.C. § 1441(a), as the district court had original jurisdiction pursuant to 28 U.S.C. § 1332. We have jurisdiction over this appeal under 28 U.S.C. § 1291. Our review is plenary, and we will sustain the district court's entry of judgment as a matter of law if "there is no legally sufficient evidentiary basis for a reasonable jury" to return a verdict in favor of Mrs. Murray. Rhone Poulenc Rorer Pharmaceuticals Inc. v. Newman Glass Works, 112 F.3d 695, 696 (3d Cir.1997). The evidence must be viewed in a light most favorable to the non-moving party. See Mosley v. Wilson, 102 F.3d 85, 89 (3d Cir.1996) [5] For example, Dr. Scotti testified that "... about five percent of people with hepatitis will go on to have chronic active liver disease but they will live for years and years. He was not in hepatic failure and was not--did not have cirrhosis, so that again before the fracture, his liver disease did not pose an immediate threat of death or death within a year or two or three or four." [6] The court was then known as the Court of Errors and Appeals [7] "In the instant matter we attach little significance to the presence of the exclusionary clause in view of the primary provision limiting coverage to loss from accidental bodily injuries, directly and independently of all other causes." Kievit, 170 A.2d at 30 [8] We have on several occasions recognized and applied New Jersey's adoption of the reasonable expectations doctrine. See, e.g., Oritani Savings & Loan Ass'n v. Fidelity & Deposit Co. of Md., 989 F.2d 635, 638 (3d Cir.1993); Vargas v. Hudson County Bd. of Elections, 949 F.2d 665, 671-72 (3d Cir.1991); Van Orman v. American Ins. Co., 680 F.2d 301, 308-310 (3d Cir.1982) [9] Tomaiuoli is the closest case to the present one on the facts. Mr. Tomaiuoli's heart disease and diabetes were diagnosed a few years prior to his accident and were monitored during regular checkups. His doctor noted "no signification [sic] changes" in his conditions during this time, and it appears that medication was prescribed at some point. See Tomaiuoli, 182 A.2d at 585-86. The court found additionally that "arteriosclerosis and its attendant symptoms are the invariable consequences of the aging process and are progressive in varying degrees depending upon the arterial fortitude of the individual." Id. at 587. The court does not make note of any testimony to the effect that Mr. Tomaiuoli's conditions were under control at the time of accident, however [10] Both Kievit and Mahon appear to attach significance to whether or not the insured was aware of his pre-existing condition prior to the time of the injury. In light of modern medicine's increasingly developed ability to detect diseases in their nascent stages, we fail to see how this distinction could continue to have much meaning. That is, we discern little difference between a disease or condition that is unknown to the patient and asymptomatic, and one that is known but that is being treated to the extent that it is considered "under control." The relevant question in both instances is the same--to what extent is the preexisting condition affecting the life (and life expectancy) of the patient? [11] As we noted in Part II, supra, the district court found that the "particulars, the specifics of Dr. Scotti's testimony" belie Dr. Scotti's testimony that all of Mr. Murray's pre-existing conditions were under control at the time of the accident. While the district court could be justified in reaching this conclusion on a de novo review of the facts, it is error in the context of a motion for judgment as a matter of law, since there is a legally sufficient evidentiary basis for a reasonable jury to accept Dr. Scotti's conclusion, and since the evidence must be taken in a light most favorable to the non-moving party. See supra note 4 [12] We of course do not mean to be critical of the district judge, who, given the opaqueness of New Jersey law in this area, could hardly have been expected to divine what our prediction would be [13] For the same reason we reject defendants' assertion that if the jury could reasonably have ruled in plaintiff's favor under proper instructions then a new trial should be ordered under the present circumstances