JOHNSON, Circuit Judge:
Appellant Talmadge H. Perry brought this action pursuant to the Swine Flu Act, 42 U.S.C.A. § 247b et seq., and the Federal Tort Claims Act, 28 U.S.C.A. § 1346(b), seeking damages for injuries allegedly sustained in connection with a swine flu inoculation. Trial was held in the United States District Court for the Northern District of Georgia, on the issues of diagnosis and causation. The district court found that appellant had failed to prove by a preponderance of the evidence that the vaccine
I. STATEMENT OF THE CASE
A. The Facts
Appellant was inoculated with swine flue vaccine on November 5, 1976. That evening he suffered a fever and chill. Over the next few days he experienced dizziness and nausea. During the last part of November 1976, Perry experienced some numbness on the left side of his body, and began to have difficulty with balance and coordination. Because of his dizziness and coordination problems, Perry took a leave from his job as a Georgia State Trooper. On December 6, Perry was admitted to DeKalb General Hospital, where he was placed under the care of Dr. Russell Wallace, a neurologist, who continued to see him as a patient up to the time of the trial.
Appellant told Dr. Wallace that in 1968 he had had a series of symptoms similar to those he was currently experiencing. In particular, he had had a three-month episode of dizziness that had made it impossible for him to drive his car. Although appellant did not place much significance on this episode during the trial, he stated on a questionnaire filled out for the Georgia Highway Patrol that his dizziness had forced him to resign from the Highway Patrol in September 1968.
During appellant's hospitalization in December 1976, Dr. Wallace performed an exhaustive array of tests in an effort to diagnose and treat appellant's condition. Physical examinations revealed slightly slurred speech, mild incoordination and numbness on the left side of the body and nystagmus, a dancing or shimmering of the eyes. The results of numerous neurological tests and a lumbar puncture were all normal.
Appellant has continued under Dr. Wallace's care for the past six years. He has been hospitalized on several occasions and subjected to more neurological tests, all of which produced negative results. He has continued to have varying degrees of incoordination of his left leg and arm, and has experienced problems with sleeping, concentration and decreased endurance. He has suffered depression as a result of his debilitation.
B. The Evidence
During the trial the court heard testimony from four physicians. The first, testifying for appellant, was Dr. Wallace, the neurologist who had been his treating physician. Dr. Wallace described the tests that had been performed in December 1976, and stated that his diagnosis at the time of appellant's discharge in December 1976 was possible inflammation of the brain stem of unknown cause. In January 1977, Dr. Wallace explained, he had been questioned by appellant's counsel as to whether he believed that the vaccine could have been a cause of appellant's affliction. Dr. Wallace stated that appellant's condition was difficult to define and its cause could not be identified; he also stated that he believed the vaccine "triggered" appellant's problems, though he could neither prove nor dispute that it was the cause of them. When appellant was readmitted to the hospital in February 1977, Dr. Wallace's discharge diagnosis was "brain stem lesion, questionably related to swine flu immunization." Dr. Wallace ultimately diagnosed appellant's condition as encephalitis. He stated at trial that he believed "to a reasonable medical certainty" that appellant's encephalitis was caused by the swine flu vaccine. He testified that he could not explain precisely how the vaccine had caused the illness, but he believed that the connection was supported by the time of the onset of appellant's symptoms, the fact that the swine flue vaccine had been connected with other neurological disorders such as Guillain-Barre Syndrome (GBS), and the fact that encephalitis in general tended to be triggered by stressful occurrences such as vaccinations. Dr. Wallace admitted, however, that prior to the trial he had heard of no documented connection between encephalitis and the swine flue vaccine, and that he could not rule out the possibility that the temporal proximity of appellant's vaccination
Dr. Kibler, a neurologist under whom Dr. Wallace trained, testified for the government. After examining appellant's medical records and conducting a physical examination of appellant himself, Dr. Kibler diagnosed appellant's condition as labyrinthitis (or inflammation of the inner ear). He stated that affliction was perhaps similar to the episode that appellant had suffered in 1968. He based his diagnosis on the fact that appellant's neurological tests had been normal, and that one-sided incoordination or numbness was not typical of organic brain disease. He did state, however, that as appellant's treating physician, Dr. Wallace was in a better position to make a diagnosis than he was. When asked whether he believed the swine flu vaccine could have been the cause of appellant's condition, Dr. Kibler stated that it's "something that one would think about," but "somewhere along the line you have to show some kind of statistical relation to make that connection valid".
Dr. Goldfield, an epidemiologist who testified for appellant, presented evidence of a statistical relationship between encephalitis and the swine flue vaccine. In formulating his conclusion, Dr. Goldfield looked at three sets of data involving cases of encephalitis: 1) the data used by the Center for Disease Control to analyze the relationship between the swine flu vaccine and various neurological disorders: 2) case histories that Dr. Goldfield had collected during his several years of work with swine flu cases: 3) data collected by the Annual Encephalitis Survey Summary. Using figures from the first two sources, Dr. Goldfield projected the incidence of encephalitis following administration of the swine flu vaccine in the general population. He then compared this figure to the incidence of encephalitis not specifically associated with the vaccine in the general population, to find the "relative risk" of encephalitis associated with the vaccine. Dr. Goldfield determined by this method that the relative risk was 8.3 or that, if the illness occurred within six weeks of the vaccination, there was an 87% probability that the affliction was caused by the vaccine. Dr. Goldfield admitted, however, that although he was not a neurologist, he had "rediagnosed" as encephalitis a number of the cases in the first two collections of data that neurologists had diagnosed differently. He also stated that this rediagnosis had increased the number of cases in which encephalitis had followed administration of the swine flu vaccine. He stated, however, that, even if one were to use a figure representing post-vaccine cases of encephalitis which was only half as large as the figure he used, the relative risk would still be 4. Or, in other words, there would still be a greater than 75% likelihood that the vaccine was the cause of the encephalitis.
Dr. Coatsworth, a neurologist who testified for the government, was the final expert witness. He had examined appellant's records and read the deposition testimony of the other physicians, but had not conducted a physical examination of appellant. He claimed that there was insufficient clinical and laboratory evidence on which to base a diagnosis or to conclude that there was a connection between appellant's affliction and the swine flu vaccine. He criticized Dr. Goldfield's epidemiological study connecting encephalitis with the vaccine, and found particular fault with Dr. Goldfield's "rediagnosis" of a number of cases in his data base. Dr. Coatsworth testified that the best diagnosis of appellant's condition was probably multiple sclerosis, although test results had not supported this diagnosis.
The district court accepted Dr. Wallace's diagnosis of encephalitis, but concluded that Perry had failed to prove by a preponderance of the evidence that his affliction
II. PROOF OF CAUSATION
A. Dr. Goldfield's Epidemiological Findings
The district court's conclusion was based primarily on its rejection of the epidemiological findings presented by Dr. Goldfield. These findings were central to appellant's case, as they established the crucial statistical connection between encephalitis and the swine flu vaccine. Appellant argues that Dr. Goldfield's testimony was unrefuted and that the court abused its discretion in refusing to accept the results of his study as conclusive proof of causation. While there was no contrary evidence presented concerning the relative risk of encephalitis for those who had been vaccinated for swine flu, Dr. Goldfield's study was subjected to substantial criticism during the course of the trial. Moreover, the district court offered four reasons for rejecting Goldfield's conclusions, none of which has been shown by appellant to be clearly in error.
The court criticized Dr. Goldfield first for "rediagnosing" a number of cases in his data base. When cross-examined concerning this practice, Dr. Goldfield could make no satisfactory explanation beyond the statement that on the basis of the records he had disagreed with the original diagnosis. (This answer was obviously not satisfactory to the district court, which observed that Dr. Goldfield, unlike many of the doctors making the original diagnoses, was not a neurologist.) Appellant offers no better refutation of this criticism. He argues that it was Dr. Coatsworth who "rediagnosed" cases; but Dr. Coatsworth's criticism focused not on the original data but on the use made of it by Dr. Goldfield. This neurologist concluded from a review of the data that only 16 of the 39 cases diagnosed as encephalitis by Dr. Goldfield satisfied him as being definitive examples of the disease.
The district court also criticized the fact that Dr. Goldfield had reached a conclusion as to the connection between encephalitis and the vaccine before commencing his research. Appellant argues that Dr. Wallace, too, formed an opinion about the connection without having conducted epidemiological research; but this response misses the point. A scientist who has a formed opinion as to the answer he is going to find before he even begins his research may be less objective than he needs to be in order to produce reliable scientific results. The district court, moreover, is not the first to have applied this criticism to Dr. Goldfield's work. Several of the courts which have heard his testimony in other swine flu cases have criticized the effect of this predisposition on his research. See e.g., O'Gara v. United States, 560 F.Supp. 786, 790 (E.D.Pa.1983) ("... Dr. Goldfield's analysis seems to me to be marked by a tendency to select figures and assumptions ... which will lead to the conclusion that the relative risk in the immunized population remains significantly elevated"); Robinson v. United States, 533 F.Supp. 320, 328 (E.D.Mich.1982) ("[Dr. Goldfield] uses calculations that are constructed so as to prove his hypothesis").
The district court provides two additional grounds for its finding — the study's lack of peer review and the demeanor of Dr. Goldfield himself — which reveal no clear error. Despite appellant's protestations, the examination of a scientific study by a cadre of lawyers is not the same as its examination by others trained in the field of science or medicine. And a district court's observations concerning the demeanor of a witness cannot be second-guessed by a reviewing court which had no opportunity to observe him. In view of the rule that questions concerning a witness's credibility are to be resolved in favor of the finder of fact,
B. Proof by a Preponderance of the Evidence
Appellant also argues that the court's conclusion that he failed to demonstrate causation by a preponderance of the evidence