Daniel J. Glanz, Alexandria, Va., (Richard J. Mudd, Washington, D.C., on brief) for appellant.

David B. Smith, Trial Atty., U.S. Dept. of Justice, Washington, D.C. (Elsie L. Munsell, U.S. Atty., Nash W. Schott, Asst. U.S. Atty., Alexandria, Va., on brief), for appellee.

Before HALL and ERVIN, Circuit Judges, and SWYGERT, Senior Circuit Judge of the United States Court of Appeals for the Seventh Circuit, sitting by designation.

SWYGERT, Senior Circuit Judge:

This case arises out of the admitted negligence of the National Institutes of Health, Department of Health and Human Services. The United States District Court for the Eastern District of Virginia dismissed the plaintiff's claim that she had been deprived of her chance of survival. For the reasons set out below, we hold that the district court's factual determination that plaintiff failed to prove that the delay in treatment substantially reduced her chance of survival is not clearly erroneous. Therefore, we affirm the judgment of the district court.

* Plaintiff Virginia Waffen was a thirty-eight year old mother of two children at the time she filed this suit. She was referred in November 1974 to the Arthritis and Rheumatism Branch of the National Institutes of Health ("NIH") in Bethesda, Maryland and was an in-patient for a month starting December 1974; she was an out-patient until 1981 receiving treatments for lupus (systemic lupus erythematosus) and nephritis. She had been smoking cigarettes since the age of eighteen and smoked two packs a day in later years. During this out-patient period, NIH took two sets of chest x-rays, one on March 16, 1977 and one on February 1, 1979, both of which were read as "negative."

On March 26, 1981 Waffen was admitted for the second time to the Clinical Center at NIH for symptoms consistent with lupus. At the time of her admission a standard chest x-ray was ordered. The radiologist's report was printed out on April 3, 1981 and concluded:

Impression: 3 X 5-cm. area of soft tissue density seen within the posterior segment of right upper lobe. Possibilities include area of consolidation vs. mass lesion, and clinical correlation and follow-up examination recommended. (emphasis added)

The radiologist's report was "misplaced" and not included in Waffen's file. The doctors who prepared and signed the discharge summary, attending physicians Cheryl Rubin and John Decker, had not seen the report, but nevertheless stated that Waffen's chest x-ray was "within normal limits." NIH has stipulated that the doctors were clearly negligent.

On October 9, 1981 Waffen's local internist, Dr. John Antus, admitted her to Prince William Hospital in Virginia, because she was suffering with fever, shaking chills, sweats, and cough. Dr. Antus' admission notes stated that "last chest x-ray done at NIH in April this year was normal." On the next day, a chest x-ray was taken, disclosing a mass measuring 4-cm. in diameter.

On October 28, 1981 a pathologist at the Prince William Hospital diagnosed a biopsy from Waffen's right lung as a malignant "infiltrating carcinoma." Waffen returned to NIH for yet another chest x-ray on November 5, which was compared by the Radiology Department with the one taken earlier. The comparison stated that "the mass lesion in the posterior segment of the right upper lobe is again seen and has grown somewhat in size. It now measures approximately 5 X 5-cm." It was at this point that NIH realized that an "error" had been committed.

When Dr. Santoro of NIH became aware of their error between November 12 and November 16, 1981, he immediately reported it to his supervisor, Dr. Decker (who had signed the erroneous discharge summary), and to the director of the Clinical Center. Dr. Santoro did notify the plaintiff's referring internist, Dr. Antus, of the error, but neither Dr. Decker nor Dr. Santoro advised Waffen's Georgetown thoracic surgeon, Dr. Hufnagel, even though he had scheduled her for surgery on November 20. Dr. Hufnagel testified that he was not aware of the lost x-ray until after the suit was commenced in 1983. Dr. Santoro also did not advise Waffen of the error, basing that decision on his belief that the stress of the news would add to the risk of the major surgery she faced and that it might exacerbate her lupus condition.

On November 20, 1981, Dr. Hufnagel surgically removed the upper right and middle lobes of the plaintiff's lung. From this was taken a tumor mass measuring approximately five to six centimeters. The pathologist diagnosed Waffen as having adenocarcinoma of the right lobe in her lung, but found no evidence of metastatic disease in the lymph nodes. On December 30, 1981 Dr. Hufnagel noted that Waffen had been suffering from "definite adenocarcinoma," but believed that patients "of her type seldom had recurrences."

On February 4, 1982, when Waffen returned to the clinic, NIH doctors first reported to her their failure to get the report on her chest x-ray in March 1981. She was advised of the possibility of a tort claim and the option of treatment elsewhere, but she elected to remain at NIH and indicated that she did not plan to sue the doctors or the Clinical Center. From February 4, 1982 until May 18, 1983, additional x-rays taken at NIH showed no change. She reported symptoms of tingling and numbness in her upper right arm on May 19, 1983, but an x-ray was negative. However, the symptoms of tingling and numbness returned in September 1983 accompanied by headaches. A bone scan and chest x-rays suggested a recurrence. On October 26 a chest x-ray showed a "3 X 5-cm. spherical mass" and metastatic spreading to a rib. Finally, a biopsy completed on a node in Waffen's neck on November 10, 1983 showed recurrent adenocarcinoma.

From December 12, 1983 to January 23, 1985 the plaintiff underwent radiation at Georgetown, but it had to be discontinued due to a lupus related skin reaction. She is currently being treated at Georgetown Hospital. It is uncontested that Waffen's cancer is terminal, and she has no hope of long-term survival.

On December 15, 1983 plaintiff made an administrative claim against the United States under the Federal Tort Claims Act, 28 U.S.C. Secs . 1346(b), 2671 et seq., alleging medical malpractice by NIH. The claim was never formally accepted or denied. On July 30, 1984 the plaintiff filed her complaint in the United States District Court for the Eastern District of Virginia, Alexandria Division. On March 8, 1985, Waffen filed an amended complaint charging nine counts of medical malpractice, including failure to communicate, failure to supervise medical care, and abandonment. She charged that the acts of the doctors were grossly negligent, wilful, and deliberate.

A bench trial was held on April 9, 1985. On May 1 the district court ordered plaintiff's claims dismissed; on May 13 the plaintiff filed a motion to reconsider or, in the alternative, to grant a new trial. The district court denied the motion, and plaintiff then filed this appeal.

II

This case raises a thorny question of what constitutes legal injury and proximate causation in the medical malpractice context. The answer must be determined by State of Maryland law, which governs. There is, however, an absence of Maryland caselaw on facts like those presented here, leading to a dispute between the parties concerning the basic legal analysis which should be employed.

The first part of our analysis is straightforward. The general principles which ordinarily govern in negligence cases also apply to medical malpractice claims under Maryland law. A prima facie case of medical malpractice must consist of evidence which (1) establishes the applicable standard of care, (2) demonstrates that this standard has been violated, and (3) develops a causal relationship between the violation and the harm complained of. Fitzgerald v. Manning, 679 F.2d 341, 346 (4th Cir.1982). As in any other case founded upon negligent conduct, the burden of proof in a medical malpractice claim rests upon the plaintiff. Shilkret v. Annapolis Emergency Hospital Ass'n, 276 Md. 187, 349 A.2d 245, 247 (1975); Paige v. Manuzak, 57 Md.App. 621, 471 A.2d 758, 766-67 (1984).

Plaintiff has met her burden of proof with respect to the first two elements of the cause action: duty and breach. Defendant NIH has stipulated that it violated the applicable standard of care by its negligence. Specifically, NIH has admitted that its failure to communicate to Waffen the results, in a timely manner, of the x-ray taken on March 26, 1981 and to suggest to her that further evaluation was mandatory and urgent constituted a breach of the standard of care due and owing her.

The elements of the cause of action which are in dispute, and on which we must focus, are those including "harm" and "causation." Plaintiff has framed the issue on her appeal as being whether the negligence of defendant NIH removed or diminished any substantial possibility of survival remaining to her. Defendant NIH contends that the type of injury alleged by Waffen is not compensable under Maryland law, except perhaps in cases involving abandonment of a patient by a physician. The district court held that Maryland does recognize a harm in the loss or diminution of the chance to survive, but that the plaintiff failed to prove causation, and failed to present the evidence necessary to prove that prompt revelation of the March 26, 1981 x-ray would have resulted in any substantial possibility of survival.

The district court looked for its standard on these elements to Hicks v. United States, 368 F.2d 626 (4th Cir.1966), and to Thomas v. Corso, 265 Md. 84, 288 A.2d 379, 390 (1972), in which the highest court in Maryland cites Hicks with approval. Hicks is not only the leading case in this circuit, but is one of the seminal cases in this emerging area of law. It is imperative, therefore, that we understand the precedent that has been established by this case.

Hicks involved an action brought under the Federal Tort Claims Act ("FTCA"), 28 U.S.C. Secs . 1346(b), 2671 et seq., to recover damages as a result of the death of a woman treated at a Navy dispensary in Virginia. The decedent was a diabetic brought to the dispensary in the middle of the night with the sudden onset of intense abdominal pain and vomiting. After a cursory ten-minute examination, a Navy doctor on duty diagnosed the patient as having gastroenteritis and told her that she had a "bug." She was released with medication for pain relief and instructions to return in eight hours. The woman took the medication and lay down. Later, after she drank a glass of water, she vomited and fell unconscious. She was rushed back to the dispensary and pronounced dead. An autopsy revealed that she had suffered a hiatal hernia, and death was due to strangulation of the intestine.

The uncontradicted testimony at the trial was that, with prompt surgery, the decedent would have survived. The court held that the dispensary physician's negligence in failing to make a thorough examination and in omitting standard diagnostic tests led to an erroneous diagnosis. As a result, the dispensary physician's negligence nullified whatever chance of recovery she might have had; the negligence was therefore the proximate cause of her death. The court concluded, in an oft-cited passage by Judge Sobeloff: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.

368 F.2d at 632 (citation omitted) (emphasis in original). The district court in the case before us interpreted this language in Hicks to adopt a standard for a cognizable harm in Maryland based on the loss of a "substantial possibility of survival." Transferring that standard to the facts of the instant case, the judge framed the issue as "whether NIH's negligence removed or diminished any substantial possibility of survival remaining to the plaintiff had she been operated on shortly after March 26, 1981." Jt. App. at 83. The Government argues that Hicks is merely "dicta." However, the same language has been repeated and given credence in two Maryland cases, one appearing on the heels of Hicks and in one decided only a short time ago.

The first such case is Thomas v. Corso, 265 Md. 84, 288 A.2d 379, 390 (1972), in which the plaintiff was hit by an automobile and brought to a hospital emergency room. He had been unconscious earlier, had been bleeding from his nose and mouth, and was complaining of pain in his right hip. The hospital employed no residents or interns so the registered nurse on duty telephoned the private practicing physician "on call" that night and notified him of the plaintiff's low blood pressure reading. The doctor did not come to the hospital to examine him, did not order that x-rays be immediately taken, and did not make a diagnosis. Instead, the doctor only instructed the nurse to admit Corso for observation, and to order x-rays for following morning. Corso was placed in the hall outside the nurses' station in a unit waiting for a bed. Complaining of pain, calling for water, and breathing poorly, he went into shock. The doctor arrived to pronounce him dead three and one-half hours after he entered the emergency room; the cause of death was listed as traumatic shock and hip and pelvic fractures.

The defendant in Corso argued before the Maryland court that the plaintiffs had failed to establish the causal connection between the doctor's negligence and decedent's death. The court disagreed and cited to the passage in Hicks referred to above. The state court ruled that the physician's admission that he believed he might have helped Corso and that the lack of treatment increased Corso's danger of losing his life, combined with a second doctor's testimony that Corso's chance for survival was linked to treatment, were sufficient to justify a jury finding of a "substantial possibility of survival which was destroyed by the failure of [the physician] to examine, diagnose and treat." Id. at 390. The court held that the physician's negligence was one of the direct and proximate causes of Corso's death and affirmed the trial court verdict for the plaintiff.

Here the defendant NIH argues that just as the language of Hicks recognizing a cause of action for the loss of a chance to survive was "dictum," so the reference to it in Corso was simply repeated "dictum." There is a recent Maryland case which recognizes the continuing viability of this theory of liability. Hetrick v. Weimer, 67 Md.App. 522, 508 A.2d 522 (1986), held that the circuit court's jury instructions had been given in error because they did not adequately present the appellants' theory, which rested on Hicks and Corso. The case was, on this ground, reversed in part and remanded for a new trial.

Hetrick involved the postnatal care of an infant. The baby was delivered nine weeks premature, by Caesarean, born in extremely poor condition. The defendant pediatrician and neonatologist for the community hospital first introduced himself to the baby's mother in the operating room. He spent the following ten hours trying various methods to improve the baby's cardiopulmonary responses and blood sugar level. The baby was then transferred to a Baltimore neonatal intensive care unit, where he died of perinatal asphyxia. 508 A.2d at 524-25.

Defendant's expert witness testified that, given the baby's condition at birth, the infant would have an 85-90% chance of survival if given proper medical care; he also testified that 80-90% of all such infants who do survive would develop and grow as normal children. He then offered his opinion that the reason the Hetrick baby did not survive, i.e., the cause of the death, was that the infant suffered from continuing asphyxia due to lack of proper ventilation by the defendant physician.

The circuit court instructed the jury that in order to recover damages the plaintiffs must prove that the defendant physician's failure to use the required degree of skill and care was the "most likely cause" of the infant's death. The court specified:

But if you should find that Dr. Weimer [the physician] was responsible for the lack of oxygen and that was 50% of the cause of death and if you feel that the prematurity was 50% of the cause of the death, then that's the standoff again. We got two causes of action. There are two possible causes of death that are both equal. If that's the case, the plaintiff hasn't done what the law requires and you must find in favor of the doctor. The plaintiff has to show that the act for which the doctor is responsible for is better than 50%, 51%. That's better.

The plaintiffs objected to those instructions and to the court's refusal to grant their instruction, based upon Corso, to the effect that "all the plaintiff need prove is that failure to properly resuscitate took away a substantial possibility that the child would have survived." Id. The Maryland Court of Special Appeals cited Corso as support for its holding that "[a]ppellant's theory of the case--that [the physician] is liable if his negligent failure to ventilate [the infant] in a proper manner deprived the infant of a substantial possibility of surviving--is unquestionably a correct exposition of the law." Id. at 532.

The appellate court emphasized that the theory of the plaintiffs' case was not that the doctor caused the infant's death; instead, it was that the doctor's failure to do what was reasonable, proper, necessary, and appropriate to resuscitate the infant deprived the child of a substantial possibility of survival. Id. at 531. The instruction that the plaintiffs had "an obligation to prove that the physician's neglect was the probable cause of the patient's death is by no means the same as an obligation to prove that the physician's neglect deprived the patient of a substantial possibility of survival." Id. (emphasis in original). Because of the error in instruction, the appellate court reversed the judgment of the circuit court.

The recognition of a cause of action for a loss of a substantial possibility of survival, first articulated in Hicks, relied on in Corso, is now again reaffirmed in Hetrick. We believe there is no further basis for argument that this consistent line of case is built merely upon "dicta." Therefore, we make explicit that under the law of this circuit, and in particular of the State of Maryland, the loss of a substantial chance of survival is a cognizable harm.

By recognizing such a harm, we do not contradict our long-held rules on causation. The law of this circuit, and of the State of Maryland, has traditionally required a strict showing of causation in order to impose damages in medical malpractice cases. Speculation and conjecture are not enough. Instead, the plaintiff must submit proof that the injury complained of was "more likely" or "more probably" due to defendant's action rather than to any other cause. Fitzgerald v. Manning, 679 F.2d 341, 356 (4th Cir.1982). Another way to phrase the test for causation is that it is one of reasonable probability or reasonable certainty. Clark v. United States, 402 F.2d 950, 953 (4th Cir.1968); Davidson v. Miller, 276 Md. 54, 344 A.2d 422, 427 (1975).

The mere possibility that a defendant's conduct may have caused injury does not provide sufficient causation as a basis for liability. The plaintiff has the burden of introducing evidence which affords a reasonable basis for the conclusion that it is more likely than not that the conduct of the defendant was a substantial factor in bringing about the result. When the matter remains one of pure speculation or conjecture, the court must direct a verdict for the defendant. Prosser, Law of Torts Sec. 41, at 241 (4th ed. 1971).

These requirements for causation have not been set aside in this circuit. "Hicks laid down no new rule of law with respect to either negligence or proximate cause." Clark v. United States, 402 F.2d at 953 n. 4. Nor did Corso change Maryland law on causation; Maryland cases following Corso, such as Hetrick, have approved the traditional cause instructions with no changes. See also Marlow v. Cerino, 19 Md.App. 619, 313 A.2d 505, 509 (1974).

What was significant about Hicks, Corso, and Hetrick was their affirmation that a certain kind of harm (the loss of a "substantial possibility of survival") could be actionable. A cause of action based on a loss of survival could now be brought and compensation awarded. Defendant argues that the district court has relied upon "dicta" in Hicks to adopt a lesser standard of proof of causation, based only on a "possibility." We believe defendant confuses causation and harm. After Hicks, the plaintiff's burden of proving causation by a preponderance of the evidence is not lessened, but the plaintiff may now recover damages for showing such causation of a new kind of harm: loss of a substantial possibility of survival.

Before Hicks, a plaintiff could only recover damages if he could prove that defendant's conduct was the cause of the medical condition creating the injury. After Hicks, plaintiff could recover damages if he could prove that defendant's conduct was the cause of a substantially reduced chance of survival from that medical condition. Specifically, Waffen does not allege that the conduct of NIH caused her to develop cancer which harmed her; rather, Waffen seeks to prove that the negligence of NIH caused her a harm in that it substantially reduced her chance of recovery from the cancer. To prove a medical malpractice case, after Hicks as before, the plaintiff must show all four elements of the cause of action: duty, breach of that duty, causation, and damages. The law of this circuit has not changed with respect to the first three elements, but in Hicks we recognized that a new species of legal injury exists, for which damages may be awarded.

Recognition of a cause of action based on a "substantial possibility" is not in conflict with traditional rules of causation. It has never been the rule of this circuit that a plaintiff is required to prove to an absolute certainty what would have happened in circumstances that the negligent wrongdoer did not allow to come to pass. Nor must the plaintiff "negative entirely the possibility that the defendant's conduct was not a cause." Prosser, Law of Torts, supra, Sec. 41 at 242. It would suffice if the plaintiff introduces evidence from which a reasonable person may conclude that it is more probable than not that the event was caused by the defendant. The fact of causation is incapable of mathematical proof, because it is impossible for a person to state with absolute certainty what could have happened if a defendant had acted in another manner. Restatement (Second) of Torts, Sec. 433B (1965).

A factfinder is entitled to make inferences, particularly in cases where the defendant has not only harmed the plaintiff substantively, but has also destroyed the plaintiff's opportunity of proving the damage. In a leading case in this circuit, a shipmaster's failure to turn back to save a drowning seaman destroyed the reasonable possibility of his rescue. Although the defendant's omission had "obliterated all possibility of evidence" as to whether the seaman was still alive at the time he was discovered missing, the factfinder nevertheless found proximate causation and imposed liability. Gardner v. National Bulk Carriers, Inc., 310 F.2d 284, 287 (4th Cir.1962), cert. denied, 372 U.S. 913 , 83 S.Ct. 728, 9 L.Ed.2d 721 (1963).

The confusion that has persisted is based upon confusing "causation" and "harm." Hetrick has clarified the issue, citing J.H. King, Jr., "Causation, Valuation, and Chance in Personal Injury Torts Involving Preexisting Conditions and Future Consequences," 90 Yale L.J. 1353-79 (1981). As the Hetrick court suggests, in analyzing problems of this nature, it is better to consider the loss of a substantial chance of survival as a different type of loss with a different measure of damages than the loss of life, instead of treating the former as a variation on the burden of proving causation in a claim for negligently causing the patient's death.1 The destroyed chance itself is the compensable loss. Hetrick, 508 A.2d at 531 n. 6.

In the line of cases following Hicks, including today's decision, we have employed an analysis that is not inconsistent with our long-held rules of causation. Instead, we have recognized the existence of a harm where the defendant's negligence destroys an individual's substantial chance of survival. It is this harm that the district court held could form the basis of a cause of action in the instant case. Because that holding is consistent with the precedent of this circuit, we uphold that portion of the court's decision.

III

Having upheld the district court's decision that a cause of action may be brought under Maryland law for loss of a substantial possibility of survival, we turn to the claim before us in order to determine whether Waffen has proved such a compensable harm as a result of defendant NIH's negligence. In reviewing the district court's decision, this court is bound by the basic premise that a finding of fact by the district court sitting without a jury may be set aside on appeal only if "clearly erroneous." In FTCA cases, which are tried without a jury, factual determinations, including damages, are governed by the clearly erroneous standard of review. United States v. Pendergrast, 241 F.2d 687, 689 (4th Cir.1957); Fed.R.Civ.P. 52(a). A reviewing court must not weigh the evidence de novo or disturb the finding simply because it might have reached a contrary result on the same evidence. Zenith Radio Corp. v. Hazeltine Research, Inc., 395 U.S. 100, 123, 89 S.Ct. 1562, 1576, 23 L.Ed.2d 129 (1969).

It is only when the appellate court has made an independent review of the evidence, and it is left with the definite and firm conviction that a mistake has been committed, that the district court's judgment should be reversed. United States v. United States Gypsum Co., 333 U.S. 364, 395, 68 S.Ct. 525, 541, 92 L.Ed. 746 (1948). Upon a review of the record in the instant case, we conclude that no mistake has been committed. The factual determination made by the district court--that Waffen failed to prove that the delay in treatment substantially reduced her chance of survival--is not clearly erroneous. For that reason, we affirm the decision of the district court.

Let us review the evidence that plaintiff has presented. First, and most importantly, Waffen has demonstrated that the tumor in her lung did increase in size from 3.5X5-cm. on March 26, 1981, when the first x-ray was taken, to 4X5-cm. on October 9, 1981, when a chest x-ray was taken at Prince William Hospital. In general, accepted medical opinion tells us, the larger the tumor in a given organ, the worse the prognosis. Certain jurisdictions accept proof of any increase in the size of a tumor, standing alone, as sufficient to sustain a cause of injury for loss of survival. Courts that take this position do so on the basis that an increase in size, however small, represents the spread of cancer cells into once healthy tissue and constitutes an injury per se. See Evers v. Dollinger, 95 N.J. 399, 471 A.2d 405, 410 (1984); Cloys v. Turbin, 608 S.W.2d 697, 701 (Tex.Civ.App.1980).

We are not willing, however, to accept such a standard based on any one factor, where harm may be potentially "de minimus." Mere increase in the size of a tumor, however minute, is not per se an actionable injury. The standard of Hicks, we believe, requires proof that the defendant's negligence contributed in a specific way to the ultimate harm suffered. Such a requirement is also contained in Maryland law, which demands proof in any negligence action of an "actual loss or injury to the plaintiff." Scott v. Watson, 278 Md. 160, 165, 359 A.2d 548, 552 (1976).

Similarly, the fact that plaintiff has proved that the negligence of NIH caused a seven-month delay in her treatment is also not dispositive. There is no question that delay in treatment almost invariably results in a more serious prognosis. But, we agree with the Fifth Circuit that more is required in the way of proof of harm than testimony that generally "the sooner you find an [injury], the better the chance of obtaining a good result." Bryant v. Rankin, 468 F.2d 510, 514-15 (8th Cir.1972). The passage of time is not a dependable bellwether, particularly in malignancy cases. The point at which micrometastases grow into distant or secondary tumors--which may occur at any time and be invisible when it does--is critical to determining life expectancy.

More accurate gauges by which to determine whether a legal harm has occurred are: (1) the chance of survival if properly treated according to medical assessments generally recognized in the field, and (2) the extent to which the plaintiff's chance of survival was reduced by improper departure from the standard of care. See Daniels v. Hadley Memorial Hospital, 566 F.2d 749, 757 (D.C.Cir.1977). Both of these factors militate against recovery in the case at bar.

The defendant's expert, Dr. Aisner, presented extensive support for the view that tumor/nodal involvement/metastasis ("TNM") staging is the most widely used assessment in cancer diagnosis and treatment because of its predictive capacity. "Staging" means a method of describing the anatomic progression of the cancer at the time of a diagnosis. The American Joint Committee for Cancer Staging and End Results Reporting ("AJC") has adopted the TNM system of classification for use in lung cancer cases, as representing the closest medical science can get to definitive staging. The system is based on a comparison of three factors: (1) the size of the primary tumor, (2) the status of the regional lymph nodes, and (3) the presence or absence of distant metastases.2

Under the TNM assessment, Dr. Aisner testified, Waffen's tumor was at T2 NoMo or Stage 1, on both March 26 and November 5, 1981: greater than 3-cm. with no evidence of nodal involvement or distant metastasis. Dr. Aisner testified that patients with T2 NoMo stage lung cancer have a 40-60% probability of remaining disease-free for five years. According to Dr. Aisner, the size of the primary tumor is the least important among the various aspects of lung cancer because at the point that a tumor is large enough to be detected by chest x-ray, it "is approximately 1-cm. in size and already contains a billion cancer cells." While early detection is useful, by the time the lesions are demonstrable on chest x-rays mestastasis has already been formed in the vast majority of individuals. For those individuals who experience a disease recurrence, the overwhelming evidence suggests that the metastases were formed at the time of the earliest possible surgical intervention. Thus, the size of the lung cancer tumor is not as important a factor, Dr. Aisner testified, as is nodal metastasis.

According to defendant's expert, Waffen's chance of survival never diminished because her tumor grew very slowly, there was little change in the primary mass, and the T2 NoMo stage never changed during the seven-month delay period. Dr. Aisner also cited Waffen's continued use of cigarettes and her long-term use of the drug Prednisone for treatment of lupus (which has the effect of depressing the immune system) as negative "host factors" which kept the original chance of survival in March at the low end of the range.

This same medical conclusion was reached by Dr. Henderson, originally called by Waffen to submit a written opinion on her behalf. Dr. Henderson concluded that Waffen's chance of survival remained in the forty percent range during the delay period and that "[m]ore likely than not earlier diagnosis could not have changed the outcome..... My feeling is that it would not have made a difference in terms of percentage chance of cure if Mrs. Waffen's cancer were diagnosed earlier." Jt.App. at 455. After Dr. Henderson's opinion was offered as evidence, plaintiff moved to substitute another expert witness.

It is only Dr. Melvin Shiffman, the expert who was finally called by Mrs. Waffen at trial, who arrived at a different conclusion. In order to explain his opinion that the metastasis responsible for Waffen's recurrent disease occurred within the seven-month delay period, Dr. Shiffman relied on an assessment method called "doubling time." This assessment is based on the amount of time necessary for a tumor to double in numbers, or in the number of cells which are preparing to divide.

Dr. Shiffman acknowledged that "doubling time" was not recognized by a consensus of the medical community as a method for determining prognosis or modality of treatment, as was the TNM system. He characterized Waffen's tumor as fast-growing, in contradiction to every other oncological specialist called upon in the instant case, all of whom characterized Waffen's tumor as slow-growing. He further testified that based on his theory that cancer cells grow by a division of cells, progressing geometrically, the earliest time at which the metastasis could have spread to the nodes was in July 1981. Dr. Shiffman therefore concluded that because the spread of cancer had occurred during the seven-month delay caused by defendant's negligence, her injury was the result of NIH's negligence.

After consideration of both experts' testimony, the court found that Dr. Aisner, defendant's witness, offered better reasoning for his opinion. We cannot find the court's holding to be clearly erroneous. In a medical malpractice case, the factfinder is entitled to weigh the testimony of the experts and decide the credibility of each opinion; similarly, whether the witness is qualified to testify as an expert at all is within the sound discretion of the judge. Greenstein v. Meister, 279 Md. 275, 368 A.2d 451, 457 (1977).

The district judge in this case was presented with two contrasting expert opinions. Dr. Aisner was a Board-certified oncologist, professor, and world-renowed author. He had written or contributed to a number of books; one of his works, Lung Cancer: Basic Principles, was the basis of much of his testimony. Dr. Aisner cast doubt on the tumor-growth method relied on by plaintiff's expert, saying that it was widely unaccepted for prognosis because it was speculative and had essentially no clinical data to validate its application to any one individual. The district judge found that Dr. Aisner persuasively attacked the "doubling time" concept by pointing out that tumors do not grow predictably, but erratically, and that the cell composition of a tumor mass is not itself homogeneous, either as to active/dormant status, or with regard to size.

On the other hand, plaintiff's expert, Dr. Shiffman, was a surgeon and lawyer who was not a Board-certified oncologist or certified in related areas and had never published scholarly articles on lung cancer. Dr. Shiffman could present no authority for his contention that the growth of a tumor within a TNM staging category has an effect on survival rates and, therefore, could not substantiate his opinion that Waffen had suffered a 15-20% decrease in survival rate.

We see no error in the district court's decision to accept the opinion of defendant's expert. Plaintiff argued that "the real issue in this case" was whether Waffen "might have been saved by earlier treatment." We disagree. The word "might" is not enough. The issue is whether a plaintiff can prove by a preponderance of the evidence that a defendant's negligence actually deprived her of a substantial possibility of survival.

What the extent of the increased risk must be to constitute loss of a "substantial possibility" is up to the factfinder, within certain limitations. First, the lost chance cannot be so insubstantial as to amount to sheer speculation. It is true that other jurisdictions have held that "[n]o 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." James v. United States, 483 F.Supp. 581, 587 (N.D.Cal.1980). We cannot accept a de minimus standard. If we adopted a rule that allowed the awarding of damages without any showing of a real injury, it would run afoul of 28 U.S.C. Sec . 2674, which prohibits the award of punitive damages against the United States. Flannery v. United States, 718 F.2d 108, 110-11 (4th Cir.1983), cert. denied, 467 U.S. 1226 , 104 S.Ct. 2679, 81 L.Ed.2d 874 (1984). Further, it disregards the bedrock "substantiality" requirement of Hicks and Corso.

On the other hand, the chance of survival need not have been fifty-one percent or more before it was reduced. It is true that several courts have held that the probability of survival had to have been more than fifty percent in order for liability to be present. Recovery is denied unless the plaintiff can establish that a decedent would probably have survived "but for" defendant's negligence. See Cooper v. Sisters of Charity of Cincinnati, Inc., 27 Ohio St.2d 242, 272 N.E.2d 97, 103-04 (1971); Hiser v. Randolph, 126 Ariz. 608, 617 P.2d 774, 778-79 (Ct.App.1980) (later overruled by the Arizona Supreme Court in Thompson v. Sun City Community Hosp., Inc., 141 Ariz. 597, 688 P.2d 605, 616 (1984) (en banc)); Hanselmann v. McCardle, 275 S.C. 46, 267 S.E.2d 531 (1980). The rationale is that the defendant should not be liable where the decedent more than likely would have died in any event. Cornfeldt v. Tongen, 295 N.W.2d 638, 640 (Minn.1980); University Hosp. Bldg., Inc. v. Gooding, 419 So.2d 1111, 1113 (Fla.App.1982), aff'd, Gooding v. University Hosp. Bldg., Inc., 445 So.2d 1015 (Fla.1984) (fifty percent chance to survive not enough). Courts that take this position do so on the basis that the principle of proximate cause requires the harshness of the result.

We believe that such an analysis confuses proximate causation and proof of harm. While interrelated, they cannot be merged into each other. As discussed above, the standard for proof of causation in this circuit after Hicks remains unchanged: it is one of proof by a preponderance of evidence. But that which is being proved to be proximately related to defendant's negligence is the legal harm of loss of a "substantial possibility." There is nothing in the language of Hicks which requires "substantial" to mean 51%, as opposed to 49%. "To decide otherwise would be a blanket release from liability for doctors and hospitals at any time there was less than a 50 per cent chance of survival, regardless of how flagrant the negligence." Herskovits v. Group Health Co-op, 99 Wash.2d 609, 664 P.2d 474, 477 (1983).

Any other rule would require pseudo-scientific precision and encourage a battle of experts between one who evaluates the lost chance at 49% and the other who estimates it closer to 51%. Thompson, 688 P.2d at 615. The better rule, we believe, is 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, it becomes a question for the factfinder to determine whether the increased risk amounted to a loss of a substantial possibility of survival. See also Hamil v. Bashline, 481 Pa. 256, 392 A.2d 1280, 1287-88 (1978).

The factfinder in the case at bar was simply not presented with evidence of such a loss of a substantial possibility. Instead, Waffen can only point to Dr. Aisner's admission that there was some "undefinable" chance that she "might have survived had she been treated on March 26, 1981." This is not enough to constitute proof of a legal harm under Maryland law.

Plaintiff argues that the court erred by requiring her to quantify her lost chance of survival in order to recover damages. To emphasize, we do not require a plaintiff to produce precise quantification of the lost survival chance. We recognize that medical science is not yet developed enough to detect micrometastasis, the clinically undetectable spread of cancer, and thereby give us a mathematical assessment of the rate of survival at every point. Clearly, the loss of life expectancy may be real but eludes quantification to some degree. On the other hand, we do require a plaintiff to show a cognizable injury, beyond sheer speculation. The Government need not rule out some remote possibility; rather, the plaintiff must produce evidence demonstrating, to a reasonable probability, an actual injury--a "substantial" loss.

The district court said that this is a "tragic" case, and added, "The decision the court makes is not one that affords any sense of satisfaction." We agree. Nevertheless, the court's factual determination that the plaintiff failed to prove that the delay in treatment caused by the negligence of NIH substantially reduced her chance of survival is not clearly erroneous. Therefore, we must affirm the judgment of the court.

1 Professor King considers just such a problem as that which confronts us here:

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.

Id. at 1363-64.

2 A malignant tumor sheds cancer cells into the blood stream and the lymphatic channels that drain these tissues. The cancer cells spread through the body and can establish secondary tumors in distant organs, which are known as "secondaries" or "metastases."