This is the old version of the H2O platform and is now read-only. This means you can view content but cannot create content. You can access the new platform at https://opencasebook.org. Thank you.
99 Wn.2d 609 (1983)
664 P.2d 474
The Supreme Court of Washington, En Banc.
May 26, 1983.
 Leonard W. Schroeter and Janet Lane Eaton (of Schroeter, Goldmark & Bender), for appellant.5
Williams, Lanza, Kastner & Gibbs, Joel D. Cunningham, and Mary H. Spillane, for respondent.6
Bryan P. Harnetiaux and Robert H. Whaley on behalf of Washington Trial Lawyers Association, amici curiae for appellant.7
This appeal raises the issue of whether an estate can maintain an action for professional negligence as a result of failure to timely diagnose lung cancer, where the estate can show probable reduction in statistical chance for survival but cannot show and/or prove that with timely diagnosis and treatment, decedent probably would have lived to normal life expectancy.9
Both counsel advised that for the purpose of this appeal we are to assume that the respondent Group Health Cooperative of Puget Sound and its personnel negligently failed to diagnose Herskovits' cancer on his first visit to the hospital and proximately caused a 14 percent reduction in his chances of survival. It is undisputed that Herskovits had  less than a 50 percent chance of survival at all times herein.10
The main issue we will address in this opinion is whether a patient, with less than a 50 percent chance of survival, has a cause of action against the hospital and its employees if they are negligent in diagnosing a lung cancer which reduces his chances of survival by 14 percent.11
The personal representative of Leslie Herskovits' estate initiated this survivorship action against Group Health Cooperative of Puget Sound (Group Health), alleging failure to make an early diagnosis of her husband's lung cancer. Group Health moved for summary judgment for dismissal on the basis that Herskovits probably would have died from lung cancer even if the diagnosis had been made earlier, which the trial court granted.12
The complaint alleged that Herskovits came to Group Health Hospital in 1974 with complaints of pain and coughing. In early 1974, chest X-rays revealed infiltrate in the left lung. Rales and coughing were present. In mid-1974, there were chest pains and coughing, which became persistent and chronic by fall of 1974. A December 5, 1974, entry in the medical records confirms the cough problem. Plaintiff contends that Herskovits was treated thereafter only with cough medicine. No further effort or inquiry was made by Group Health concerning his symptoms, other than an occasional chest X-ray. In the early spring of 1975, Mr. and Mrs. Herskovits went south in the hope that the warm weather would help. Upon his return to the Seattle area with no improvement in his health, Herskovits visited Dr. Jonathan Ostrow on a private basis for another medical opinion. Within 3 weeks, Dr. Ostrow's evaluation and direction to Group Health led to the diagnosis of cancer. In July of 1975, Herskovits' lung was removed, but no radiation or chemotherapy treatments were instituted. Herskovits died 20 months later, on March 22, 1977, at the age of 60.14
At hearing on the motion for summary judgment, plaintiff  was unable to produce expert testimony that the delay in diagnosis "probably" or "more likely than not" caused her husband's death. The affidavit and deposition of plaintiff's expert witness, Dr. Jonathan Ostrow, construed in the most favorable light possible to plaintiff, indicated that had the diagnosis of lung cancer been made in December 1974, the patient's possibility of 5-year survival was 39 percent. At the time of initial diagnosis of cancer 6 months later, the possibility of a 5-year survival was reduced to 25 percent. Dr. Ostrow testified he felt a diagnosis perhaps could have been made as early as December 1974, or January 1975, about 6 months before the surgery to remove Mr. Herskovits' lung in June 1975.15
Dr. Ostrow testified that if the tumor was a "stage 1" tumor in December 1974, Herskovits' chance of a 5-year survival would have been 39 percent. In June 1975, his chances of survival were 25 percent assuming the tumor had progressed to "stage 2". Thus, the delay in diagnosis may have reduced the chance of a 5-year survival by 14 percent.16
Dr. William Spence, one of the physicians from Group Health Hospital who cared for the deceased Herskovits, testified that in his opinion, based upon a reasonable medical probability, earlier diagnosis of the lung cancer that afflicted Herskovits would not have prevented his death, nor would it have lengthened his life. He testified that nothing the doctors at Group Health could have done would have prevented Herskovits' death, as death within several years is a virtual certainty with this type of lung cancer regardless of how early the diagnosis is made.17
Plaintiff contends that medical testimony of a reduction of chance of survival from 39 percent to 25 percent is sufficient evidence to allow the proximate cause issue to go to the jury. Defendant Group Health argues conversely that Washington law does not permit such testimony on the issue of medical causation and requires that medical testimony must be at least sufficiently definite to establish that the act complained of "probably" or "more likely than not"  caused the subsequent disability. It is Group Health's contention that plaintiff must prove that Herskovits "probably" would have survived had the defendant not been allegedly negligent; that is, the plaintiff must prove there was at least a 51 percent chance of survival.18
Pursuant to CR 56(c), summary judgment is appropriate only where there is no genuine issue as to any material fact and the moving party is entitled to judgment as a matter of law. All reasonable inferences from the evidence must be resolved against the moving party, and in favor of the nonmoving party. Lamon v. McDonnell Douglas Corp., 91 Wn.2d 345, 588 P.2d 1346 (1979).19
This court has held that a person who negligently renders aid and consequently increases the risk of harm to those he is trying to assist is liable for any physical damages he causes. Brown v. MacPherson's, Inc., 86 Wn.2d 293, 299, 545 P.2d 13 (1975). In Brown, the court cited Restatement (Second) of Torts § 323 (1965), which reads:21
One who undertakes ... to render services to another which he should recognize as necessary for the protection of the other's person or things, is subject to liability to the other for physical harm resulting from his failure to exercise reasonable care to perform his undertaking, if
(a) his failure to exercise such care increases the risk of such harm, ...
This court heretofore has not faced the issue of whether, under section 323(a), proof that the defendant's conduct increased the risk of death by decreasing the chances of survival is sufficient to take the issue of proximate cause to the jury. Some courts in other jurisdictions have allowed the proximate cause issue to go to the jury on this type of proof. See McBride v. United States, 462 F.2d 72 (9th Cir.1972); Hamil v. Bashline, 481 Pa. 256, 392 A.2d 1280 (1978); Kallenberg v. Beth Israel Hosp., 45 A.D.2d 177, 357 N.Y.S.2d 508 (1974); Jeanes v. Milner, 428 F.2d 598 (8th Cir.1970); Hicks v. United States, 368 F.2d 626 (4th Cir.1966). These courts emphasized the fact that defendants'  conduct deprived the decedents of a "significant" chance to survive or recover, rather than requiring proof that with absolute certainty the defendants' conduct caused the physical injury. The underlying reason is that it is not for the wrongdoer, who put the possibility of recovery beyond realization, to say afterward that the result was inevitable. See also Wolfstone & Wolfstone, Recovery of Damages for the Loss of a Chance, Pers. Inj. Ann. 744 (1978).23
Other jurisdictions have rejected this approach, generally holding that unless the plaintiff is able to show that it was more likely than not that the harm was caused by the defendant's negligence, proof of a decreased chance of survival is not enough to take the proximate cause question to the jury. Cooper v. Sisters of Charity, Inc., 27 Ohio St.2d 242, 272 N.E.2d 97 (1971); Hiser v. Randolph, 126 Ariz. 608, 617 P.2d 774 (Ct. App. 1980); Hanselmann v. McCardle, 275 S.C. 46, 267 S.E.2d 531 (1980); Cornfeldt v. Tongen, 295 N.W.2d 638 (Minn. 1980). These courts have concluded that the defendant should not be liable where the decedent more than likely would have died anyway.24
The ultimate question raised here is whether the relationship between the increased risk of harm and Herskovits' death is sufficient to hold Group Health responsible. Is a 36 percent (from 39 percent to 25 percent) reduction in the decedent's chance for survival sufficient evidence of causation to allow the jury to consider the possibility that the physician's failure to timely diagnose the illness was the proximate cause of his death? We answer in the affirmative. To decide otherwise would be a blanket release from liability for doctors and hospitals any time there was less than a 50 percent chance of survival, regardless of how flagrant the negligence.25
We are persuaded by the reasoning of the Pennsylvania Supreme Court in Hamil v. Bashline, supra. While Hamil involved an original survival chance of greater than 50 percent, we find the rationale used by the Hamil court to  apply equally to cases such as the present one, where the original survival chance is less than 50 percent. The plaintiff's decedent was suffering from severe chest pains. His wife transported him to the hospital where he was negligently treated in the emergency unit. The wife, because of the lack of help, took her husband to a private physician's office, where he died. In an action brought under the wrongful death and survivorship statutes, the main medical witness testified that if the hospital had employed proper treatment, the decedent would have had a substantial chance of surviving the attack. The medical expert expressed his opinion in terms of a 75 percent chance of survival. It was also the doctor's opinion that the substantial loss of a chance of recovery was the result of the defendant hospital's failure to provide prompt treatment. The defendant's expert witness testified that the patient would have died regardless of any treatment provided by the defendant hospital.27
The Hamil court reiterated the oft-repeated principle of tort law that the mere occurrence of an injury does not prove negligence, but the defendant's conduct must be a proximate cause of the plaintiff's injury. The court also referred to the traditional "but for" test, with the qualification that multiple causes may culminate in injury. Hamil, at 266.28
The court then cited Restatement (Second) of Torts § 323 (1965) as authority to relax the degree of certitude normally required of plaintiff's evidence in order to make a case for the jury. The court held that once a plaintiff has introduced evidence that a defendant's negligent act or omission increased the risk of harm to a person in plaintiff's position, and that the harm was in fact sustained, "it becomes a question for the jury as to whether or not that increased risk was a substantial factor in producing the harm". Hamil, at 269. See also C. McCormick, Damages § 31 (1935); Wolfstone & Wolfstone, supra at 744.29
The Hamil court distinguished the facts of that case from the general tort case in which a plaintiff alleges that a  defendant's act or omission set in motion a force which resulted in harm. In the typical tort case, the "but for" test, requiring proof that damages or death probably would not have occurred "but for" the negligent conduct of the defendant, is appropriate. In Hamil and the instant case, however, the defendant's act or omission failed in a duty to protect against harm from another source. Thus, as the Hamil court noted, the fact finder is put in the position of having to consider not only what did occur, but also what might have occurred. Hamil states at page 271:30
Such cases by their very nature elude the degree of certainty one would prefer and upon which the law normally insists before a person may be held liable. Nevertheless, in order that an actor is not completely insulated because of uncertainties as to the consequences of his negligent conduct, Section 323(a) tacitly acknowledges this difficulty and permits the issue to go to the jury upon a less than normal threshold of proof.
(Footnote omitted.) The Hamil court held that once a plaintiff has demonstrated that the defendant's acts or omissions have increased the risk of harm to another, such evidence furnishes a basis for the jury to make a determination as to whether such increased risk was in turn a substantial factor in bringing about the resultant harm.32
In Hicks v. United States, supra, the Court of Appeals set forth the rationale for deviation from the normal requirements of proof in a case such as the one presently before us. The following quotation from Hicks, at 632, is frequently cited in cases adopting loss of a chance because it succinctly defines the doctrine:33
Rarely is it possible to demonstrate to an absolute certainty what would have happened in circumstances that the wrongdoer did not allow to come to pass. The law does not in the existing circumstances require the plaintiff to show to a certainty that the patient would have lived had she been hospitalized and operated on promptly.
Under the Hamil decision, once a plaintiff has demonstrated that defendant's acts or omissions in a situation to  which section 323(a) applies have increased the risk of harm to another, such evidence furnishes a basis for the fact finder to go further and find that such increased risk was in turn a substantial factor in bringing about the resultant harm. The necessary proximate cause will be established if the jury finds such cause. It is not necessary for a plaintiff to introduce evidence to establish that the negligence resulted in the injury or death, but simply that the negligence increased the risk of injury or death. The step from the increased risk to causation is one for the jury to make. Hamil, at 272.35
In Jeanes v. Milner, supra, the plaintiff mother brought a malpractice action for the death of her child from throat cancer, claiming delayed diagnosis of 1 month caused a shortened life span and pain and suffering. The United States Court of Appeals for the Eighth Circuit, reversing a dismissal for insufficient evidence on the element of proximate cause, held at pages 604-05:36
We cannot agree with the District Court's holding that "there is no evidence from which the jury could find that the delay of approximately one month in the transmission of [the] slides could have been the proximate cause of [Tommy's] failure to recover from his cancer, or to increase his pain and suffering or to shorten his life." Nor can we agree that the jury could "only find a verdict for the plaintiff based on speculation and conjecture."
The Supreme Court of the United States has spoken to a contention similar to that argued here by the doctors and the Infirmary. In Lavender v. Kurn, 327 U.S. 645, 653, 66 S.Ct. 740, 90 L.Ed. 916 (1946), the Court stated:
"It is no answer to say that the jury's verdict involved speculation and conjecture. Whenever facts are in dispute or the evidence is such that fair-minded men may draw different inferences, a measure of speculation and conjecture is required on the part of those whose duty it is to settle the dispute by choosing what seems to them to be the most reasonable inference."
The recent case of James v. United States, 483 F. Supp. 581 (N.D. Cal. 1980) concerned the failure to diagnose and  promptly treat a lung tumor. The court concluded that the plaintiff sustained its burden of proof even without statistical evidence, stating at page 587:38
As a proximate result of defendant's negligence, James was deprived of the opportunity to receive early treatment and the chance of realizing any resulting gain in his life expectancy and physical and mental comfort. No matter how small that chance may have been — and its magnitude cannot be ascertained — no one can say that the chance of prolonging one's life or decreasing suffering is valueless.
Where percentage probabilities and decreased probabilities are submitted into evidence, there is simply no danger of speculation on the part of the jury. More speculation is involved in requiring the medical expert to testify as to what would have happened had the defendant not been negligent. McCormick, supra.41
Chester v. United States, 403 F. Supp. 458, 460 (W.D. Pa. 1975) was a medical malpractice suit for negligent failure to diagnose and treat cancer of the esophagus. The court found that in November 1972, the decedent was exhibiting symptoms and complaints that were consistent with cancer. No tests were performed to determine what was causing the illness, and the court determined this was below the accepted medical standard of care. Further, the judge decided that the cancer was indeed present in November 1972, and could have been treated or resected before metastasis. The judge reasoned that, if there was a possibility that the decedent had carcinoma of the esophagus, the hospital doctors were negligent in treating him for hypertension only.42
The Chester court then awarded damages of $45,988.10, including $36,741.10 for loss of earning power, $7,500 for the loss of guidance, etc., for his minor children, and $1,747 for funeral expenses. This damage award is consistent with our reasoning in Wooldridge v. Woolett, 96 Wn.2d 659, 638 P.2d 566 (1981) where we held that a decedent's shortened  life expectancy is not recoverable as a separate item of damages, but will be considered as it affects the loss of value of his future earning capacity.43
Both counsel have agreed for the purpose of arguing this summary judgment that the defendants were negligent in failing to make a diagnosis of cancer on Herskovits' initial visit in December 1974, and that such negligence was the proximate cause of reducing his chances of survival by 14 percent. It is undisputed that Herskovits had less than a 50 percent chance of survival at that time. Based on this agreement and Dr. Ostrow's deposition and affidavit, a prima facie case is shown. We reject Group Health's argument that plaintiffs must show that Herskovits "probably" would have had a 51 percent chance of survival if the hospital had not been negligent. We hold that medical testimony of a reduction of chance of survival from 39 percent to 25 percent is sufficient evidence to allow the proximate cause issue to go to the jury.45
Causing reduction of the opportunity to recover (loss of chance) by one's negligence, however, does not necessitate a total recovery against the negligent party for all damages caused by the victim's death. Damages should be awarded to the injured party or his family based only on damages caused directly by premature death, such as lost earnings and additional medical expenses, etc.46
We reverse the trial court and reinstate the cause of action.47
ROSELLINI, J., concurs.48
I agree with the majority that the trial court erred in granting defendant's motion for summary judgment. I cannot, however, agree with the majority's reasoning in reaching this decision. The majority's reliance on Hamil v. Bashline, 481 Pa. 256, 392 A.2d 1280 (1978) and Hicks v. United States, 368 F.2d 626 (4th Cir.1966) is inappropriate for the reasons identified in the  dissent of Justice Dolliver. Moreover, the issue before us is considerably more complex than the apparently straightforward policy choice suggested by the interaction of the majority opinion and Justice Dolliver's dissent. I therefore agree with Justice Brachtenbach that those opinions fail to focus on the key issue. I decline to join Justice Brachtenbach's dissent, however, because the result he advocates is harsh, as he recognizes at page 642. In an effort to achieve a fair result by means of sound analysis, I offer the following approach.50
This action began in July 1979 with a complaint alleging that defendant Group Health Cooperative had negligently treated the decedent Leslie Herskovits. Plaintiff, the widow of decedent Herskovits, alleged in this complaint that defendant's failure to diagnose the decedent's lung cancer "led to and caused his death". The complaint sought damages for the medical expenses, disability, and pain and suffering of the decedent, together with pecuniary loss suffered by plaintiff, including loss of support, affection, and consortium.51
As discovery progressed, some undisputed facts were established. Mr. Herskovits had been a patient of the Group Health Cooperative for more than 20 years. In December 1974, he consulted a physician at Group Health for the treatment of a persistent cough. The physician prescribed cough medicine. Obtaining no relief from his cough, Mr. Herskovits consulted a physician outside Group Health, Dr. Jonathan Ostrow. Dr. Ostrow suspected that Mr. Herskovits had lung cancer, and recommended a medical procedure be undertaken to confirm his suspicions. The procedure was performed in June 1975, and revealed that Mr. Herskovits had cancer in the bronchus of his left lung. The lung was removed on July 1, 1975. Mr. Herskovits died of cancer on March 22, 1977, at the age of 60 years.52
Defendant moved for summary judgment in May 1981, on the ground that plaintiff was unable to produce testimony that earlier diagnosis would probably have prevented Mr. Herskovits' death from cancer. The trial court granted  the motion and dismissed the action, holding that plaintiff had "failed to produce expert testimony which would establish that the decedent probably would not have died on or about March, 1977 but for the conduct of the defendant". This holding was based on the court's conclusion that "under Washington law the loss of a possibility of survival is not compensable".53
Plaintiff's case was based on the testimony of the expert witness, Dr. Ostrow. This court's enquiry, therefore, is whether Dr. Ostrow's testimony, together with reasonable inferences therefrom, creates a prima facie issue of causation. Lamon v. McDonnell Douglas Corp., 91 Wn.2d 345, 588 P.2d 1346 (1979). The critical testimony of Dr. Ostrow, from his affidavit and deposition, may fairly be summarized as follows:54
1. There is a reasonable medical probability that defendant failed to take necessary steps to diagnose Mr. Herskovits' condition, and defendant therefore failed to meet the appropriate standard of care.55
2. Had reasonable care been exercised, Mr. Herskovits' cancer could have been diagnosed in December 1974 instead of June 1975.56
3. Unless removed, a cancerous tumor can be expected to increase in size over time, and the patient's chances of survival decline accordingly.57
4. There is no way of knowing how far the tumor in Mr. Herskovits' lung had developed by December 1974.58
5. If the tumor had been a stage 1 tumor in December 1974, decedent's statistical chance of surviving 5 years was 39 percent.59
6. When the tumor was discovered in June 1975, it was a stage 2 tumor. The statistical chance of surviving 5 years when the tumor has reached stage 2 is 25 percent.60
7. Dr. Ostrow summed up his opinion as follows: "By failing to properly evaluate Mr. Herskovits' condition as late as December 1974, Group Health probably caused Mr. Herskovits' chance for long-term survival to be substantially reduced".61
 Dr. Ostrow's testimony does not, therefore, establish a prima facie case that defendant's alleged negligence probably (or more likely than not) caused Mr. Herskovits' death. Rather, the testimony establishes only that the alleged negligence caused a substantial reduction in Mr. Herskovits' long-term chance of survival. Dr. Ostrow testified that if the tumor was at stage 1 in December 1974, the chance of survival was reduced from 39 percent to 25 percent. He did not, however, indicate the likelihood of the tumor's being at stage 1 in December 1974, either in terms of certainty, probability, or statistical chance. Therefore, the only indications from the record of the extent of the reduction in Mr. Herskovits' chance of long-term survival are that it was "substantial" and that it was at most a 14 percent reduction (from 39 percent to 25 percent).62
I turn now to consider whether this testimony is sufficient to create a material issue whether defendant's alleged negligence was a proximate cause of harm to plaintiff.63
Recently, we stated that64
[p]roximate cause must be established by, first, a showing that the breach of duty was a cause in fact of the injury, and, second, a showing that as a matter of law liability should attach.... Cause in fact can be established by proving that but for the breach of duty, the injury would not have occurred.
Harbeson v. Parke-Davis, Inc., 98 Wn.2d 460, 476, 656 P.2d 483 (1983). In the present case, we must determine whether cause in fact has been established.66
In medical malpractice cases such as the one before us, cause in fact must usually be established by expert medical testimony, and must be established beyond the balance of probabilities.67
In a case such as this, medical testimony must be relied upon to establish the causal relationship between the liability-producing situation and the claimed physical disability resulting therefrom. The evidence will be deemed insufficient to support the jury's verdict, if it can be said that considering the whole of the medical testimony the jury must resort to speculation or conjecture in  determining such causal relationship. In many recent decisions of this court we have held that such determination is deemed based on speculation and conjecture if the medical testimony does not go beyond the expression of an opinion that the physical disability "might have" or "possibly did" result from the hypothesized cause. To remove the issue from the realm of speculation, the medical testimony must at least be sufficiently definite to establish that the act complained of "probably" or "more likely than not" caused the subsequent disability.
O'Donoghue v. Riggs, 73 Wn.2d 814, 824, 440 P.2d 823 (1968).69
The issue before the court, quite simply, is whether Dr. Ostrow's testimony satisfies the standard enunciated in O'Donoghue. We must decide whether Dr. Ostrow's testimony established that the act complained of (the alleged delay in diagnosis) "probably" or "more likely than not" caused Mr. Herskovits' subsequent disability. In order to make this determination, we must first define the "subsequent disability" suffered by Mr. Herskovits. Therein lies the crux of this case, for it is possible to define the injury or "disability" to Mr. Herskovits in at least two different ways. First, and most obviously, the injury to Mr. Herskovits might be viewed as his death. Alternatively, however, the injury or disability may be seen as the reduction of Mr. Herskovits' chance of surviving the cancer from which he suffered.70
Therefore, although the issue before us is primarily one of causation, resolution of that issue requires us to identify the nature of the injury to the decedent. Our conception of the injury will substantially affect our analysis. If the injury is determined to be the death of Mr. Herskovits, then under the established principles of proximate cause plaintiff has failed to make a prima facie case. Dr. Ostrow was unable to state that probably, or more likely than not, Mr. Herskovits' death was caused by defendant's negligence. On the contrary, it is clear from Dr. Ostrow's testimony that Mr. Herskovits would have probably died from cancer even with the exercise of reasonable care by defendant. Accordingly,  if we perceive the death of Mr. Herskovits as the injury in this case, we must affirm the trial court, unless we determine that it is proper to depart substantially from the traditional requirements of establishing proximate cause in this type of case.71
 If, on the other hand, we view the injury to be the reduction of Mr. Herskovits' chance of survival, our analysis might well be different. Dr. Ostrow testified that the failure to diagnose cancer in December 1974 probably caused a substantial reduction in Mr. Herskovits' chance of survival. The O'Donoghue v. Riggs standard of proof is therefore met.72
I note here that two other problems are created by the latter analysis. First, we have never before considered whether the loss or reduction of a chance of survival is a compensable injury. And second, this analysis raises the issue of whether an action for reduction of the chance of survival can be brought under the wrongful death statute, RCW 4.20.010.73
Confronted with these problems, and with the first impression choice between the two approaches to the issue before us, I turn to consider how other jurisdictions have dealt with similar cases.74
One approach, and that urged by defendant, is to deny recovery in wrongful death cases unless the plaintiff establishes that decedent would probably have survived but for defendant's negligence. This approach is typified by Cooper v. Sisters of Charity, Inc., 27 Ohio St.2d 242, 272 N.E.2d 97 (1971). The court in that case affirmed a directed verdict for defendant where the only evidence of causation was that decedent had a chance "maybe some place around 50%" of survival had defendant not been negligent. The court said in 27 Ohio St.2d at 253-54:75
In an action for wrongful death, where medical malpractice is alleged as the proximate cause of death, and plaintiff's evidence indicates that a failure to diagnose the injury prevented the patient from an opportunity to be operated on, which failure eliminated any chance of  the patient's survival, the issue of proximate cause can be submitted to a jury only if there is sufficient evidence showing that with proper diagnosis, treatment and surgery the patient probably would have survived.
This case was followed in Hiser v. Randolph, 126 Ariz. 608, 617 P.2d 774 (Ct. App. 1980). The Arizona court explicitly rejected both Hamil v. Bashline and Hicks v. United States (the leading cases cited by plaintiffs in the case before us) and agreed with Cooper that "the mere loss of an unspecified increment of the chance for survival is, of itself, insufficient to meet the standard of probability". 126 Ariz. at 613.77
On the other hand, plaintiff cites seven cases in support of her position. Hicks v. United States, 368 F.2d 626 (4th Cir.1966); Jeanes v. Milner, 428 F.2d 598 (8th Cir.1970); O'Brien v. Stover, 443 F.2d 1013 (8th Cir.1971); McBride v. United States, 462 F.2d 72 (9th Cir.1972); Kallenberg v. Beth Israel Hosp., 45 A.D.2d 177, 357 N.Y.S.2d 508, aff'd, 37 N.Y.2d 719, 337 N.E.2d 128, 374 N.Y.S.2d 615 (1974); Hamil v. Bashline, 481 Pa. 256, 392 A.2d 1280 (1978); James v. United States, 483 F. Supp. 581 (N.D. Cal. 1980). The complexity of the issue and elusive nature of the rationales of these opinions persuade me to discuss them at some length.78
The first case, Hicks, was an action under the Federal Tort Claims Act, to recover damages for the death of a 25-year-old woman from an undiagnosed bowel obstruction. The uncontradicted testimony of the plaintiff's experts was that "if operated on promptly, [the patient] would have survived". 368 F.2d at 632.79
The District Court, construing the laws of Virginia, concluded that there was insufficient evidence that the concededly erroneous diagnosis was a proximate cause of the woman's death. The defendant urged the Court of Appeals to affirm the District Court, arguing that even had surgery been performed immediately, it was mere speculation to say that it would have been successful. The Court of Appeals rejected the argument and in what has become an  oft-quoted passage said:80
When a defendant's negligent action or inaction has effectively terminated a person's chance of survival, it does not lie in the defendant's mouth to raise conjectures as to the measure of the chances that he has put beyond the possibility of realization. If there was any substantial possibility of survival and the defendant has destroyed it, he is answerable. Rarely is it possible to demonstrate to an absolute certainty what would have happened in circumstances that the wrongdoer did not allow to come to pass. The law does not in the existing circumstances require the plaintiff to show to a certainty that the patient would have lived had she been hospitalized and operated on promptly. Harvey v. Silber, 300 Mich. 510, 2 N.W.2d 483 (1942).
368 F.2d at 632. The court in Hicks also drew an analogy to another federal case which imposed liability on a ship's master for failing to attempt to rescue a seaman who had fallen overboard. Gardner v. National Bulk Carriers, Inc., 310 F.2d 284, 91 A.L.R.2d 1023 (4th Cir.1962). The court in Gardner rejected the argument that any breach of duty by the ship's master was not a cause of the seaman's death. It held that once the evidence sustained the reasonable possibility of rescue, total disregard of the duty imposed liability.82
The court in Hicks v. United States concluded its opinion with the observation that the defendant's "negligence nullified whatever chance of recovery [the patient] might have had and was the proximate cause of the death". 368 F.2d at 633.83
Hicks is susceptible to two interpretations. On one hand, the court's use of phrases like "substantial possibility of survival" (368 F.2d at 632) and "whatever chance of recovery" (368 F.2d at 633) suggests that a less than 50 percent chance of survival would create a prima facie case of proximate cause. The analogy to the duty of a ship's master reinforces the suggestion that only a "reasonable possibility" of survival need be established. Gardner v. National Bulk Carriers, Inc., 310 F.2d at 287.84
 On the other hand, however, the court had no need to decide such a broad question. There was uncontroverted testimony before the court that the decedent would have survived if operated upon promptly. At the very least, this testimony would establish a probability of survival. All the court had to decide, therefore, was that causation was established by a showing that an operation probably (rather than certainly) would have saved the patient's life. Two passages in the opinion seem to confirm that this is the rationale of the case. First, the court, in the passage quoted above, said that "[t]he law does not ... require the plaintiff to show to a certainty that the patient would have lived". 368 F.2d at 632. Second, the court, in explaining the similarity between Harvey v. Silber, 300 Mich. 510, 2 N.W.2d 483 (1942) and the facts before it, points out that in Harvey there was testimony to the effect that an operation would probably have saved the decedent's life, and this was sufficient to establish proximate cause. 368 F.2d at 632 n. 2.85
Hicks v. United States, therefore, appears to be authority for no more than the proposition that proximate cause may be established on a probability of survival. This, of course, is entirely consistent with the existing principles of this state under O'Donoghue v. Riggs, and provides little direct support for plaintiff in the present case.86
Subsequent cases, however, have relied on the expansive dicta in Hicks to allow recovery for the reduction of a less than probable chance of survival. The first such case was Jeanes v. Milner, 428 F.2d 598 (8th Cir.1970). This was a malpractice action brought by the mother of a 13-year-old boy who died of throat cancer. The plaintiff alleged that the defendant's negligence had delayed diagnosis of the cancer for over a month. The plaintiff presented expert testimony that during that month the decedent's cancer had progressed from stage 1 to stage 2. Testimony also indicated that patients whose cancer is diagnosed at stage 1 have a statistical survival rate of 35 percent; patients at stage 2 have a survival rate of 24 percent.87
 The Court of Appeals, reversing the District Court, held that this was sufficient evidence from which to infer that the patient's "life would have been saved or at least prolonged and his pain lessened had he received early treatment." 428 F.2d at 604. This conclusion was supported by reference to Hicks, which was described as "Persuasive". 428 F.2d at 605.88
Both Jeanes and Hicks were cited in another case from the Eighth Circuit in which the plaintiff alleged a negligent delay in the diagnosis of cancer. O'Brien v. Stover, 443 F.2d 1013 (8th Cir.1971). Causation was apparently established in O'Brien by testimony that "there is an overall 30 per cent survival rate from this type of cancer and that chance is considerably improved the earlier the cancer is discovered". 443 F.2d at 1018. Among the elements of damages which the court approved was reduction of the patient's "chances for survival, or at least living longer and more comfortably". 443 F.2d at 1019.89
The fourth case cited by plaintiff, McBride v. United States, is based upon the proposition that90
When a plaintiff's cause of action rests upon an allegedly negligent failure to give necessary treatment, he must show, with reasonable medical probability, that the treatment would have successfully prevented the patient's injury.
462 F.2d at 75. This is the rationale also of Cooper v. Sisters of Charity, Inc., 27 Ohio St.2d 242, 272 N.E.2d 97 (1971), and clearly of no assistance to plaintiff in the case before us.92
The fifth case, Kallenberg v. Beth Israel Hosp., is a more extreme approach. The Appellate Division of the New York Supreme Court upheld a jury verdict of $55,000 in a wrongful death action. The jury had heard testimony from the plaintiff's witness that defendant's negligence was a "producing, contributing factor" in the death and further that "`if properly treated ... the patient still [would have had] a 20, say 30, maybe 40% chance of survival'". 45 A.D.2d at 179. The appellate court noted that proximate  cause is a jury question and concluded, without citing authority, that the testimony reviewed was sufficient for the jury to find that, but for the defendant's negligence, the decedent might have made a recovery.93
Plaintiff relies most heavily on the next case, Hamil v. Bashline. This was a wrongful death action based upon the alleged negligence of the defendant hospital's failure to treat the decedent, who subsequently died of a heart attack. An expert witness for the plaintiff testified that had the decedent received reasonable treatment he would have had a 75 percent chance of surviving the heart attack he was experiencing when admitted to the hospital. The defendant argued that the plaintiff had failed to establish a prima facie case because no testimony was introduced to show that the defendant's negligence did, to a reasonable degree of medical certainty, cause decedent's death.94
The Supreme Court of Pennsylvania acknowledged that the general principles of tort law in that state required that expert testimony "establish that the injury in question did, with a reasonable degree of medical certainty, stem from the negligent act alleged". 481 Pa. at 267. The court recognized an exception to this heavy burden of proof in situations governed by Restatement (Second) of Torts § 323(a) (1965). The court said in 481 Pa. at 269-70:95
Section 323(a) recognizes that a particular class of tort actions, of which the case at bar is an example, differs from those cases normally sounding in tort. Whereas typically a plaintiff alleges that a defendant's act or omission set in motion a force which resulted in harm, the theory of the present case is that the defendant's act or omission failed in a duty to protect against harm from another source. To resolve such a claim a fact-finder must consider not only what did occur, but also what might have occurred, i.e., whether the harm would have resulted from the independent source even if defendant had performed his service in a non-negligent manner. Such a determination as to what might have happened necessarily requires a weighing of probabilities.
(Footnote omitted.) The court then quoted the ubiquitous  passage from Hicks v. United States and observed:97
We agree with this statement of the law and hold that once a plaintiff has demonstrated that defendant's acts or omissions, in a situation to which Section 323(a) applies, have increased the risk of harm to another, such evidence furnishes a basis for the fact-finder to go further and find that such increased risk was in turn a substantial factor in bringing about the resultant harm; the necessary proximate cause will have been made out if the jury sees fit to find cause in fact.
(Footnote omitted.) 481 Pa. at 272. In a footnote the court elaborated upon the manner in which the plaintiff must demonstrate that the defendant's acts or omissions increased the risk of harm to the plaintiff. In this footnote, the court quotes from comment a to Restatement (Second) of Torts § 433B(1), to the effect that the plaintiff must establish by a preponderance of the evidence that the conduct of the defendant was a substantial factor in bringing about the harm to the plaintiff. The court continues:99
In the instant case, Dr. Wecht testified to a 75% chance of recovery had prompt treatment been administered to Mr. Hamil by Bashline; this was sufficient basis upon which the jury could have concluded that it was more likely than not that the defendant's omissions were a substantial factor in causing Mr. Hamil's death. Of course, as was done here, a defendant may present expert testimony to the opposite effect, i.e., the unlikelihood of survival even had the defendant exercised due care; as always, resolution of conflicting testimony is for the jury.
481 Pa. at 272 n. 9.101
It is clear that Hamil, like Hicks and McBride, stands for no more than a rejection of a reasonable certainty standard of proof, and an acceptance of a reasonable probability standard. Viewed thus, it advances plaintiff's case very little.102
The final case cited by plaintiff is James v. United States, 483 F. Supp. 581 (N.D. Cal. 1980). In that case, the District Court held that plaintiff's failure to establish a statistically measurable chance of survival did not rule out recovery. The court said, without citing authority:103
 As a proximate result of defendant's negligence, [the decedent] was deprived of the opportunity to receive early treatment and the chance of realizing any resulting gain in his life expectancy and physical and mental comfort. No matter how small that chance may have been — and its magnitude cannot be ascertained — no one can say that the chance of prolonging one's life or decreasing suffering is valueless.
483 F. Supp. at 587.105
This decision goes well beyond any of the other cases, allowing recovery for the loss of any chance, no matter how small.106
Having concluded this somewhat detailed survey of the cases cited by plaintiff, what conclusions can we draw? First, the critical element in each of the cases is that the defendant's negligence either deprived a decedent of a chance of surviving a potentially fatal condition or reduced that chance. To summarize, in Hicks v. United States the decedent was deprived of a probability of survival; in Jeanes v. Milner the decedent's chance of survival was reduced from 35 percent to 24 percent; in O'Brien v. Stover, the decedent's 30 percent chance of survival was reduced by an indeterminate amount; in McBride v. United States the decedent was deprived of the probability of survival; in Kallenberg v. Beth Israel Hosp. the decedent was deprived of a 20 percent to 40 percent chance of survival; in Hamil v. Bashline the decedent was deprived of a 75 percent chance of survival; and in James v. United States the decedent was deprived of an indeterminate chance of survival, no matter how small.107
The three cases where the chance of survival was greater than 50 percent (Hicks, McBride, and Hamil) are unexceptional in that they focus on the death of the decedent as the injury, and they require proximate cause to be shown beyond the balance of probabilities. Such a result is consistent with existing principles in this state, and with cases from other jurisdictions cited by defendant.108
The remaining four cases allowed recovery despite the plaintiffs' failure to prove a probability of survival. Three of  these cases (Jeanes, O'Brien, and James) differ significantly from the Hicks, McBride, and Hamil group in that they view the reduction in or loss of the chance of survival, rather than the death itself, as the injury. Under these cases, the defendant is liable, not for all damages arising from the death, but only for damages to the extent of the diminished or lost chance of survival. The fourth of these cases, Kallenberg, differs from the other three in that it focuses on the death as the compensable injury. This is clearly a distortion of traditional principles of proximate causation. In effect, Kallenberg held that a 40 percent possibility of causation (rather than the 51 percent required by a probability standard) was sufficient to establish liability for the death. Under this loosened standard of proof of causation, the defendant would be liable for all damages resulting from the death for which he was at most 40 percent responsible.109
My review of these cases persuades me that the preferable approach to the problem before us is that taken (at least implicitly) in Jeanes, O'Brien, and James. I acknowledge that the principal predicate for these cases is the passage of obiter dictum in Hicks, a case which more directly supports the defendant's position. I am nevertheless convinced that these cases reflect a trend to the most rational, least arbitrary, rule by which to regulate cases of this kind. I am persuaded to this conclusion not so much by the reasoning of these cases themselves, but by the thoughtful discussion of a recent commentator. King, Causation, Valuation, and Chance in Personal Injury Torts Involving Preexisting Conditions and Future Consequences, 90 Yale L.J. 1353 (1981).110
King's basic thesis is explained in the following passage, which is particularly pertinent to the case before us.111
Causation has for the most part been treated as an all-or-nothing proposition. Either a loss was caused by the defendant or it was not.... A plaintiff ordinarily should be required to prove by the applicable standard of proof that the defendant caused the loss in question.  What caused a loss, however, should be a separate question from what the nature and extent of the loss are. This distinction seems to have eluded the courts, with the result that lost chances in many respects are compensated either as certainties or not at all.
To illustrate, consider the case in which a doctor negligently fails to diagnose a patient's cancerous condition until it has become inoperable. Assume further that even with a timely diagnosis the patient would have had only a 30% chance of recovering from the disease and surviving over the long term. There are two ways of handling such a case. Under the traditional approach, this loss of a not-better-than-even chance of recovering from the cancer would not be compensable because it did not appear more likely [than] not that the patient would have survived with proper care. Recoverable damages, if any, would depend on the extent to which it appeared that cancer killed the patient sooner than it would have with timely diagnosis and treatment, and on the extent to which the delay in diagnosis aggravated the patient's condition, such as by causing additional pain. A more rational approach, however, would allow recovery for the loss of the chance of cure even though the chance was not better than even. The probability of long-term survival would be reflected in the amount of damages awarded for the loss of the chance. While the plaintiff here could not prove by a preponderance of the evidence that he was denied a cure by the defendant's negligence, he could show by a preponderance that he was deprived of a 30% chance of a cure.
90 Yale L.J. at 1363-64.113
Under the all or nothing approach, typified by Cooper v. Sisters of Charity, Inc., 27 Ohio St.2d 242, 272 N.E.2d 97 (1971), a plaintiff who establishes that but for the defendant's negligence the decedent had a 51 percent chance of survival may maintain an action for that death. The defendant will be liable for all damages arising from the death, even though there was a 49 percent chance it would have occurred despite his negligence. On the other hand, a plaintiff who establishes that but for the defendant's negligence the decedent had a 49 percent chance of survival recovers nothing.114
 This all or nothing approach to recovery is criticized by King on several grounds, 90 Yale L.J. at 1376-78. First, the all or nothing approach is arbitrary. Second, it115
subverts the deterrence objectives of tort law by denying recovery for the effects of conduct that causes statistically demonstrable losses.... A failure to allocate the cost of these losses to their tortious sources ... strikes at the integrity of the torts system of loss allocation.
90 Yale L.J. at 1377. Third, the all or nothing approach creates pressure to manipulate and distort other rules affecting causation and damages in an attempt to mitigate perceived injustices. (Kallenberg v. Beth Israel Hosp. appears to be a good illustration of this tendency.) Fourth, the all or nothing approach gives certain defendants the benefit of an uncertainty which, were it not for their tortious conduct, would not exist. (This is reminiscent of the reasoning in the fertile dictum in Hicks v. United States.) Finally, King argues that the loss of a less than even chance is a loss worthy of redress.117
These reasons persuade me that the best resolution of the issue before us is to recognize the loss of a less than even chance as an actionable injury. Therefore, I would hold that plaintiff has established a prima facie issue of proximate cause by producing testimony that defendant probably caused a substantial reduction in Mr. Herskovits' chance of survival.118
The decedent's personal action for loss of this chance will survive to his personal representatives as provided by RCW 4.20.046. The family of the decedent should also be allowed to maintain an action for the lost chance of recovery by the decedent. I would interpret the wrongful death statute, RCW 4.20.010, to apply to cases of this type. Under this interpretation, a person will "cause" the death of another person (within the meaning of RCW 4.20.010) whenever he causes a substantial reduction in that person's chance of  survival.119
Finally, it is necessary to consider the amount of damages recoverable in the event that a loss of a chance of recovery is established. Once again, King's discussion provides a useful illustration of the principles which should be applied.120
To illustrate, consider a patient who suffers a heart attack and dies as a result. Assume that the defendant-physician negligently misdiagnosed the patient's condition, but that the patient would have had only a 40% chance of survival even with a timely diagnosis and proper care. Regardless of whether it could be said that the defendant caused the decedent's death, he caused the loss of a chance, and that chance-interest should be completely redressed in its own right. Under the proposed rule, the plaintiff's compensation for the loss of the victim's chance of surviving the heart attack would be 40% of the compensable value of the victim's life had he survived (including what his earning capacity would otherwise have been in the years following death). The value placed on the patient's life would reflect such factors as his age, health, and earning potential, including the fact that he had suffered the heart attack and the assumption that he had survived it. The 40% computation would be applied to that base figure.
(Footnote omitted.) 90 Yale L.J. at 1382.122
 I would remand to the trial court for proceedings consistent with this opinion.123
WILLIAMS, C.J., and STAFFORD and UTTER, JJ., concur with PEARSON,124
I dissent because I find plaintiff did not meet her burden of proving proximate cause. While the statistical evidence introduced by the expert was relevant and admissible, it was not alone sufficient to maintain a cause of action.126
Neither the majority nor Justice Dolliver's dissent focus on the key issue. Both opinions focus on the significance of the 14 percent differentiation in the patient's chance to survive for 5 years and question whether this statistical data is sufficient to sustain a malpractice action. The issue is not so limited. The question should be framed as whether all the evidence amounts to sufficient proof, rising above speculation, that the doctor's conduct was a proximate cause of the patient's death. While the relevancy and the significance of the statistical evidence is a subissue bearing on the sufficiency of the proof, such evidence alone neither proves nor disproves plaintiff's case.127
Proximate cause is neither easily defined nor readily understood. Dean Prosser begins his discussion of proximate cause with this disclaimer:129
There is perhaps nothing in the entire field of law which has called forth more disagreement, or upon which the opinions are in such a welter of confusion. Nor, despite the manifold attempts which have been made to clarify the subject, is there yet any general agreement as to the proper approach. Much of this confusion is due to the fact that no one problem is involved, but a number of different problems, which are not distinguished clearly, and that language appropriate to a discussion of one is carried over to cast a shadow upon the others.
(Footnotes omitted.) W. Prosser, Torts § 41, at 236 (4th ed. 1971).131
 Confusion is generated by the fact that proximate cause is a uniquely legal concept. It is not synonymous with the concept of cause in a philosophical sense because, hypothetically, an act may cause endless consequences. See Restatement (Second) of Torts § 431, comment a (1965). Proximate cause, however, represents the judicial limitations placed upon an actor's liability for the consequences of his or her conduct. See King v. Seattle, 84 Wn.2d 239, 249, 525 P.2d 228 (1974); cf. Hunsley v. Giard, 87 Wn.2d 424, 434, 553 P.2d 1096 (1976) (liability limited through judicial definition of duty).132
The boundaries of proximate cause are not self-determinative. The question is one of law, not fact alone, and it is one that necessarily involves a policy decision. See Probert, Causation in the Negligence Jargon: A Plea for Balanced "Realism", 18 U. Fla. L. Rev. 369 (1965); Green, The Causal Relation Issue in Negligence Law, 60 Mich. L. Rev. 543 (1962). Whether the case goes to the jury or the judge dismisses the claim for a failure to make a case for causation may depend on the actors and the circumstances involved.133
Seldom does a rule protect every victim against every risk that may befall him, merely because it is shown that the violation of the rule played a part in producing the injury. The task of defining the proper reach or thrust of a rule in its policy aspects is one that must be undertaken by the court in each case as it arises.
Malone, Ruminations on Cause-In-Fact, 9 Stan. L. Rev. 60, 73 (1956). For example, the standard of proof of causation involving an intentional wrongdoer or joint tortfeasors may often be relaxed, whereas stringent proof requirements have been applied if the plaintiff was partially at fault. Compare Madigan v. Teague, 55 Wn.2d 498, 348 P.2d 403 (1955) (joint tortfeasors found liable) with Scott v. Rainbow Ambulance Serv., Inc., 75 Wn.2d 494, 452 P.2d 220 (1969) (case dismissed because of plaintiff's failure to segregate his contribution to the damages).135
Malpractice suits represent a class of controversies where  extreme caution should be exercised in relaxing causation requirements. See Atkins v. Clein, 3 Wn.2d 168, 100 P.2d 1 (1940). The physician serves a vital function in our society, a function which requires the assumption of a duty to the patient. Yet, his profession affords him only an inexact and often experimental science by which to discharge his duty. Moreover, the tendency to place blame on a physician who fails to find a cure is great. Thus policy considerations do not, on balance, weigh in favor of abandoning the well established requirements of proximate cause.136
The majority proposes adoption of the substantial factor test for cases falling under Restatement (Second) of Torts § 323(a) (1965). This was the approach taken in Hamil v. Bashline, 481 Pa. 256, 392 A.2d 1280 (1978); however, application of the substantial factor test in these circumstances is truly novel. Usually the substantial factor test has been applied only in situations where there are two causes, either of which could have caused the event alone, and it cannot be determined which was the actual cause. See Madigan v. Teague, supra. For example, A and B both start separate fires which combine to burn C's house. Either fire alone would have caused the same result, but C cannot prove that "but-for" the negligence of either A or B the house would not have burned. Therefore, to prevent both A and B being relieved of liability, the "but-for" test is abandoned, and the question becomes whether the conduct of A or B was a substantial factor in causing the fire that injured C. Under this test, either A or B could be held liable for the damage.137
Except in situations where there are coequal causes, however, defendant's act cannot be a substantial factor when the event would have occurred without it. W. Prosser, at 244. Moreover, the substantial factor test has not been applied when one of the causes of the injury was the plaintiff himself. Hence in Scott v. Rainbow Ambulance Serv., Inc., supra, the case was dismissed because plaintiff failed to segregate her contribution to the damages from that of the tortfeasor; there was no discussion of the tortfeasor's  conduct in the context of the substantial factor test. Similarly, the substantial factor test should not be applied in the instant case.138
Furthermore, the instant case does not present evidence of proximate cause that rises above speculation and conjecture. The majority asserts that evidence of a statistical reduction of the chance to survive for 5 years is sufficient to create a jury question on whether the doctor's conduct was a proximate cause of the death. I disagree that this statistical data can be interpreted in such a manner.140
Use of statistical data in judicial proceedings is a hotly debated issue. See, e.g., Finkelstein & Fairley, A Bayesian Approach to Identification Evidence, 83 Harv. L. Rev. 489 (1970) (Finkelstein & Fairley I); Tribe, Trial by Mathematics: Precision and Ritual in the Legal Process, 84 Harv. L. Rev. 1329 (1971) (Tribe I); Finkelstein & Fairley, A Comment on "Trial by Mathematics", 84 Harv. L. Rev. 1801 (1971) (Finkelstein & Fairley II); Tribe, A Further Critique of Mathematical Proof, 84 Harv. L. Rev. 1810 (1971) (Tribe II); Dickson, Medical Causation by Statistics, 17 Forum 792 (1982); see also Brachtenbach, Future Damages in Personal Injury Actions — The Standard of Proof, 3 Gonz. L. Rev. 75 (1968). Many fear that members of the jury will place too much emphasis on statistical evidence and the statistics will be misused and manipulated by expert witnesses and attorneys. Tribe I, supra; Dickson, supra; People v. Collins, 68 Cal.2d 319, 438 P.2d 33, 66 Cal. Rptr. 497 (1968).141
Such fears do not support a blanket exclusion of statistical data, however. Our court system is premised on confidence in the jury to understand complex concepts and confidence in the right of cross examination as protection against the misuse of evidence. Attorneys ought to be able to explain the true significance of statistical data to keep it in its proper perspective.142
Statistical data should be admissible as evidence if they are relevant, that is, if they have143
 any tendency to make the existence of any fact that is of consequence to ... the action more probable or less probable than it would be without the evidence.
ER 401. The statistics here met that test; they have some tendency to show that those diagnosed at stage 1 of the disease may have a greater chance to survive 5 years than those diagnosed at stage 2.145
The problem is, however, that while this statistical fact is relevant, it is not sufficient to prove causation. There is an enormous difference between the "any tendency to prove" standard of ER 401 and the "more likely than not" standard for proximate cause.146
Reliance on statistics alone to prove proximate cause may lead to unjust results. A simple example will illustrate my point. Assume there are two cab companies in a town; one has three blue cabs and the other has one yellow cab. A pedestrian is hit by a cab, but doesn't know what color it was. In a suit for personal injury, plaintiff wants to admit the statistical fact that there is a 75 percent chance that she was hit by a blue cab. This fact has relevancy; it is admissible. But is it sufficient to prove the blue cab company more probably than not committed the act? No. If this were not the case, the blue cab company could be held liable for every unidentified cab accident that occurred.147
Thus statistics alone should not be sufficient to prove proximate cause. What is necessary, at the minimum, is some evidence connecting the statistics to the facts of the case. Referring back to the cab example, testimony that a blue cab was seen in the vicinity of the accident before or after it occurred or evidence of a recently acquired, unaccounted for, dent in a blue cab could combine with the statistical  evidence to lead a jury to believe it was more probable than not that this plaintiff was hit by a blue cab. See Finkelstein & Fairley I.148
Thus, I would not resolve the instant case simply by focusing on the 14 percent differentiation in the chance to survive 5 years for the different stages of cancer. Instead, I would accept this as an admissible fact, but not as proof of proximate cause. To meet the proximate cause burden, the record would need to reveal other facts about the patient that tended to show that he would have been a member of the 14 percent group whose chance of 5 years' survival could be increased by early diagnosis.149
Such evidence is not in the record. Instead, the record reveals that Mr. Herskovits' cancer was located such that corrective surgery "would be more formidable". This would tend to show that his chance of survival may have been less than the statistical average. Moreover, the statistics relied on did not take into consideration the location of the tumor; therefore, their relevance to Mr. Herskovits' case must be questioned. Clerk's Papers, at 41.150
In addition, as the tumor was relatively small in size when removed (2 to 3 centimeters), the likelihood that it would have been detected in 1974, even if the proper test were performed, was less than average. This uncertainty further reduces the probability that the doctor's failure to perform the tests was a proximate cause of a reduced chance of survival.151
Other statistics admitted into evidence also tend to show the inconclusiveness of the statistics relied on by the majority. One study showed the two-year survival rate for this type of cancer to be 46.6 percent for stage 1 and 39.8 percent for stage 2. Mr. Herskovits lived for 20 months after surgery, which was 26 months after defendant allegedly should have discovered the cancer. Therefore, regardless of the stage of the cancer at the time Mr. Herskovits was examined by defendant, it cannot be concluded that he survived significantly less than the average survival time. Hence, it is pure speculation to suppose that the doctor's  negligence "caused" Mr. Herskovits to die sooner than he would have otherwise. Such speculation does not rise to the level of a jury question on the issue of proximate cause. Therefore, the trial court correctly dismissed the case. See Restatement (Second) of Torts § 434 (1965); Scott v. Rainbow Ambulance Serv., Inc., 75 Wn.2d 494, 452 P.2d 220 (1969).152
The apparent harshness of this conclusion cannot be overlooked. The combination of the loss of a loved one to cancer and a doctor's negligence in diagnosis seems to compel a finding of liability. Nonetheless, justice must be dealt with an even hand. To hold a defendant liable without proof that his actions caused plaintiff harm would open up untold abuses of the litigation system.153
Cases alleging misdiagnosis of cancer are increasing in number, perhaps because of the increased awareness of the importance of early detection. These cases, however, illustrate no more than an inconsistency among courts in their treatment of the problems of proof. See Annot., Malpractice in Connection with Diagnosis of Cancer, 79 A.L.R.3d 915 (1977). Perhaps as medical science becomes more knowledgeable about this disease and more sophisticated in its detection and treatment of it, the balance may tip in favor of imposing liability on doctors who negligently fail to promptly diagnose the disease. But, until a formula is found that will protect doctors against liability imposed through speculation as well as afford truly aggrieved plaintiffs their just compensation, I cannot favor the wholesale abandonment of the principle of proximate cause. For these reasons, I dissent.154
DIMMICK, J., concurs with BRACHTENBACH, J.155
This is apparently a case of first impression. As is usually true in such instances, the court is called upon to make a policy decision. The issue before us is whether, when the chance of survival is less than a probability, i.e., less than 50 percent, proof that the  chance of survival — not the probability of survival — is reduced is sufficient to take the case to the jury.157
The majority answer in the affirmative cites several cases in support of this view and adopts the reasoning of Hamil v. Bashline, 481 Pa. 256, 392 A.2d 1280 (1978). I favor the opposing view and believe the reasoning in Cooper v. Sisters of Charity, Inc., 27 Ohio St.2d 242, 272 N.E.2d 97 (1971), also cited by the majority, is more persuasive. In discussing the rule to be adopted the Ohio Supreme Court stated:158
A rule, which would permit a plaintiff to establish a jury question on the issue of proximate cause upon a showing of a "substantial possibility" of survival, in our judgment, suffers the same infirmity as a rule which would permit proof of a "chance of recovery" to be sufficient. While the substantial possibility concept appears to connote a weightier burden than the chance of recovery idea, both derogate well-established and valuable proximate cause considerations. Traditional proximate cause standards require that the trier of the facts, at a minimum, must be provided with evidence that a result was more likely than not to have been caused by an act, in the absence of any intervening cause.
Lesser standards of proof are understandably attractive in malpractice cases where physical well being, and life itself, are the subject of litigation. The strong intuitive sense of humanity tends to emotionally direct us toward a conclusion that in an action for wrongful death an injured person should be compensated for the loss of any chance for survival, regardless of its remoteness. However, we have trepidations that such a rule would be so loose that it would produce more injustice than justice. Even though there exists authority for a rule allowing recovery based upon proof of causation by evidence not meeting the standard of probability, we are not persuaded by their logic....
We consider the better rule to be that in order to comport with the standard of proof of proximate cause, plaintiff in a malpractice case must prove that defendant's negligence, in probability, proximately caused the death.
(Citations omitted.) Cooper, at 251-52. This, it seems to  me, is a better rule.160
The majority states the variations from 39 percent to 25 percent in the decedent's chance for survival are sufficient evidence to "consider the possibility" that the failure of the physician to diagnose the illness in a timely manner was the "proximate cause of his death." This reasoning is flawed. Whether the chances were 25 percent or 39 percent decedent would have survived for 5 years, in both cases, it was more probable than not he would have died. Therefore, I cannot conclude that the missed diagnosis was the proximate cause of death when a timely diagnosis could not have made it more probable the decedent would have survived. "`It is legally and logically impossible for it to be probable that a fact exists, and at the same time probable that it does not exist.'" Cooper, at 253 (quoting Davis v. Guarnieri, 45 Ohio St. 470, 490, 15 N.E. 350 (1887)). It would be pure speculation, given these figures, for an expert, a jury, or anyone else to conclude the decedent would live more or less time within the 5-year period with or without the proper diagnosis.161
Two further comments: Factually, many of the cases cited by the majority, including Hamil v. Bashline, do not support its position. In McBride v. United States, 462 F.2d 72 (9th Cir.1972), the mortality rate was 15 percent for coronary patients admitted to the hospital and 30 to 35 percent for those outside the hospital. Thus, in any circumstance, the probability of survival was over 50 percent and was measurably greater in rather than out of the hospital. In Hamil v. Bashline, supra at 272 n. 9, the court stated the testimony observed "a 75% chance of recovery had prompt treatment been administered". In Hicks v. United States, 368 F.2d 626 (4th Cir.1966), the court observed, "Both of plaintiff's experts testified categorically that if operated on promptly, Mrs. Greitens would have survived, and this is nowhere contradicted by the government expert." Hicks, at 632.162
In Wooldridge v. Woolett, 96 Wn.2d 659, 638 P.2d 566 (1981), we held the value of a decedent's shortened life  expectancy was not recoverable as a separate item of damages in a survival action under RCW 4.20.046. Yet, in the present case the majority is willing to grant recovery, not where there is a direct cause for the shortened life expectancy, but on the more ephemeral basis of a statistical probability entitled "probability of survival".163
While these observations may not go to the heart of the majority position, they do make it a bit ragged around the edges.164
I believe the position articulated in Cooper v. Sisters of Charity, Inc., supra, is preferable and thus I dissent.165
Reconsideration denied July 21, 1983.166
 The wrongful death statute is probably the principal reason the parties focused on the death of Mr. Herskovits rather than his diminished chance of survival. As I have endeavored to demonstrate, this approach leads either to harsh and arbitrary results, or to distortions of existing tort principles and the potential for confusion. A liberal construction of the statute appears a more effective method of achieving the most desirable end. The word "cause" has a notoriously elusive meaning (as the writings on legal causation all agree) and it is certainly sufficiently flexible to bear the interpretation I give it in the context of RCW 4.20.010.167
 In effect, this approach conforms to the suggestion of Justice Brachtenbach in his dissent at page 640, footnote 3. The statistical data relating to the extent of the decedent's chance of survival are considered to show the amount of damages, rather than to establish proximate cause.168
 There is also a difference between the standard of proof for proximate cause to show liability and the standard of proof to show the amount of damages after liability is established. Courts are willing to relax proof requirements on the issue of damages, once liability is shown. See generally 22 Am.Jur.2d Damages § 23 (1965). Therefore, statistical data may be of greater value at the damage stage, especially with regard to future damages that are necessarily subject to some uncertainties and contingencies. See Brachtenbach, Future Damages in Personal Injury Actions — The Standard of Proof, 3 Gonz. L. Rev. 75 (1968).
September 29, 2015
"Herskovits v. Group Health Cooperative of Puget Sound"
Professor of Law
Stanford Law School
This is the old version of the H2O platform and is now read-only. This means you can view content but cannot create content. If you would like access to the new version of the H2O platform and have not already been contacted by a member of our team, please contact us at email@example.com. Thank you.