OPINION
Anticipating an outbreak of swine flu in 1976, the federal government undertook a massive immunization effort through the National Influenza Immunization Program of 1976, 42 U.S.C. § 247b(j)(l) (hereinafter the Swine Flu Act). To encourage pharmaceutical companies to supply the vaccine, Congress relieved the companies of liability for injuries resulting from its use. It made the federal government liable in the companies’ stead under the Federal Tort Claims Act, 28 U.S.C. §§ 1346(b), 2671 et seq. Immunization began on October 1, 1976. Officials halted the program in mid-December of 1976, when they discovered a strong statistical association between the vaccine and an increased incidence of GuillainBarre Syndrome (GBS).
Mary Rose B. Gaul received a swine flu vaccination on November 5, 1976. On February 21, 1977, she entered the Wilmington Medical Center with severe muscular weakness, tingling and numbness in her limbs, and difficulty with speech and swallowing. Her doctors found that Mrs. Gaul had GBS.
In this action Mrs. Gaul and her husband allege that the swine flu vaccine caused her GBS. They seek damages from the government under the provisions cited above. The proceedings were bifurcated to allow separate trials of the liability and damages issues. This opinion constitutes findings of fact and conclusions of law on the issue of liability. The sole question for determination is whether Mrs. Gaul has shown by a preponderance of the evidence that her swine flu vaccination caused her to contract GBS.
I. BACKGROUND
A. Guillain-Barre Syndrome
Guillain-Barre Syndrome is a disease characterized by acute1 or subacute ascending paralysis of the peripheral nervous system. That GBS is called a syndrome denotes that medical science understands it only as an aggregation of signs and symptoms; its cause or causes are unknown.2
In GBS patients it is believed that the disease-fighting system which usually defends the body against harmful foreign substances attacks the myelin, the fatty substance that coats the peripheral nerves. The myelin insulates the nerves, like the plastic sheath that covers electrical wiring. This attack will often destroy whole segments of the myelin. This impairs the ability of the peripheral nerves to conduct elec*1124trical impulses from the brain which control muscular movement and reflexes.
An acute infectious illness, such as an upper respiratory or gastrointestinal infection, precedes approximately one-half of all cases of GBS.3 The interval between the illness and the onset of GBS varies. It may be as short as one week, and in most cases is between one and three weeks.4
All the expert witnesses in this case agreed on the criteria for diagnosis of GBS.5 Patients rapidly develop motor weakness, and mild sensory symptoms and signs such as tingling and numbness. Fifty percent reach the nadir of the disease by the end of two weeks, eighty percent do so by the end of three weeks, and ninety percent do so by the end of four weeks. Symptoms (i.e., weakness and tingling and numbness) are distributed relatively symmetrically; if one limb is affected, the opposite usually will be also. The symptoms ascend; they begin in the lower parts of the limbs and progress upward. Involvement of the muscles of the trunk and head sometimes follows. Recovery usually begins two to four weeks after progression of the disease stops.
B. The Facts
Mrs. Gaul received a swine flu vaccination on November 5, 1976. She experienced no immediate reaction to the vaccine. She testified that starting in December, she noticed occasional numbness in the fingers on her left hand, and in her right arm. Mrs. Gaul said she noticed these sensations only when inactive. They were, she said, “sort of vague things, that I would have ... sometimes and not at others.... ” Mr. Gaul recalled hearing about these symptoms in late December. Mrs. Gaul also testified that she felt uncharacteristically fatigued after her two-week December vacation. Notwithstanding Mrs. Gaul’s emphasis on these events in her testimony, she did not consult a physician about them at the time.
Mrs. Gaul visited her family physician, Dr. Benge, on February 8, 1977, for gastrointestinal difficulties. Mrs. Gaul had a long history of colitis, or irritable colon. Dr. Benge prescribed Enarax.6 The testimony conflicted as to whether Mrs. Gaul was suffering from an episode of colitis or a gastrointestinal infection. While Dr. Benge’s records7 indicate that he treated Mrs. Gaul on February 8 for an “acute gastroenteritis,”8 he would offer no opinion as to whether he treated her that day for a gastrointestinal infection. The evidence does not indicate that Mrs. Gaul mentioned any sensations of tingling or numbness to Dr. Benge on this visit.
Mrs. Gaul visited Dr. Benge again on February 14, 1977. She testified that she had a fever, general aches and pains, and perhaps a sore throat. Dr. Benge’s medi*1125cal history indicates that Mrs. Gaul had muscular aches in the abdominal and shoulder area, some sore throat, a slight cough, and fever of up to 103 degrees. Dr. Benge prescribed Ampicillin, an antibiotic, to which Mrs. Gaul “responded well.” Mrs. Gaul missed work due to this illness from Tuesday, February 15 to Friday, February 18. Dr. Benge’s records indicate that towards the end of the week, Mrs. Gaul “felt quite well.” Again, there is no indication that she mentioned any sensations of tingling or numbness to Dr. Benge.
On Sunday, February 20th, 107 days after she received swine flu vaccine, Mrs. Gaul awoke with numbness in her arms. Mr. Gaul called Dr. Benge, who suggested that he examine Mrs. Gaul the next morning if the numbness had not gone away. When Mrs. Gaul awoke on February 21st with continued tingling and numbness and speech difficulties, Mr. Gaul took her to Dr. Benge’s office. Dr. Benge examined Mrs. Gaul and asked that the Gauls inform him of any change in her condition. Mrs. Gaul returned home; her condition deteriorated rapidly. She told Dr. Benge later that day that the numbness in her arms and speech difficulties had worsened, and that she also had significant numbness and tingling in her hands and feet. Dr. Benge’s reaction was that she had GBS. An ambulance brought Mrs. Gaul to the hospital at 5:00 P.M. on February 21st. She was examined there by Dr. Gersh, a resident in neurology and Dr. Chronister, a consulting neurologist. Both doctors concluded that she had GBS.
In this case, no one disputes that Mrs. Gaul had GBS. The parties dispute whether Mrs. Gaul’s inoculation with swine flu vaccine caused her GBS. The plaintiffs advance two theories in support of their contention that the vaccine did indeed cause Mrs. Gaul’s GBS. First, Mr. and Mrs. Gaul argue that Mrs. Gaul’s GBS began back in December of 1976, when she felt some tingling, numbness and fatigue. They contend that the disease did not become full blown immediately but smoldered for approximately two months before fully presenting itself. Mrs. Gaul’s GBS began in December, plaintiffs argue, but full onset of symptoms occurred later, in February. This is the so-called “late onset” GBS theory. Commencement of the illness in December would mean that GBS followed Mrs. Gaul’s inoculation by a period of three to seven weeks. This would fall well within the period of time in which a number of epidemiologists have concluded that inoculation with swine flu vaccine greatly increases the risk of contracting GBS. The government argues in response that Mrs. Gaul’s GBS began on February 20th, the day Mrs. Gaul awoke with numbness and tingling in her arms. The medical evidence, it contends, does not support the idea that GBS may begin without a full constellation of symptoms and then progress slowly to a later, full-blown onset.
The Gauls’ second theory is that an epidemiological study shows that inoculation with swine flu vaccine creates a relative risk9 of contracting GBS of more than two10 for up to sixteen weeks after inocu*1126lation. Since Mrs. Gaul contracted GBS in the sixteenth week after inoculation, the Gauls argue, it is therefore more likely than not that Mrs. Gaul’s GBS was caused by the inoculation. The government contends that the more persuasive studies show an enhanced risk of contracting GBS until only the tenth week after inoculation, and that the study the plaintiffs proffer must be rejected.
II. THE LATE ONSET THEORY
On the question of when Mrs. Gaul’s GBS began, plaintiffs presented the testimony of Dr. Milton Alter. Dr. Alter is chairman of the Department of Neurology at Temple University Hospital, and has had extensive experience in the diagnosis, treatment and epidemiologic study of GBS. Based on the occurrence of numbness, tingling and uncharacteristic weakness and fatigue in December, Dr. Alter concluded that Mrs. Gaul’s GBS began in late November or early December of 1976. He said that the disease did not become full blown until the following February 20th. Dr. Alter conceded that this interpretation of the events would make Mrs. Gaul’s case atypical. This did not, however, change his view.
The plaintiffs cited one article from the medical literature11 in support of Dr. Alter’s conclusion. The weight of medical opinion is contrary to the late onset theory, however.12 Moreover, in an article co-authored by Dr. Alter, called “Clinical Features of the Guillain-Barre Syndrome,”13 Dr. Alter outlines the diagnostic criteria used in the study.14 The first among these criteria is “acute or subacute onset of weakness with evolution of signs within two months.”15 If, as Dr. Alter testified, Mrs. Gaul’s GBS began in late November or early December and reached the full blown state on February 20th or 21st, the smoldering stage would have lasted from ten to twelve weeks. She would not, therefore, meet the criteria in Dr. Alter’s article.
Doctors Nelson, Gersh, and Chronister, all neurologists who treated Mrs. Gaul for GBS, testified that Mrs. Gaul’s GBS began on February 20, 1977. When asked to assume the truth of the symptoms Mrs. Gaul said that she experienced in December, none of them changed his opinion. Dr. David Pleasure, Professor of Neurology at the University of Pennsylvania and the author of numerous works on peripheral nerve diseases, agreed with the treating neurologists.
Whichever medical theory one subscribes to, the threshold issue is what symptoms Mrs. Gaul experienced during the period in question. Mrs. Gaul testified at trial that she had vague feelings of tingling and numbness and felt uncharacteristically *1127weak beginning in December of 1976. I am persuaded that, in retrospect, she believes she experienced these symptoms at that time. I do not think it more likely than not, however, that Mrs. Gaul had symptoms of GBS in December or January. First, she did not seek medical advice with respect to any unusual sensations during those months. Mrs. Gaul also did not mention tingling or numbness to Dr. Benge even when she was present in his office on February 8th and 14th. Further, she apparently did not say anything to Dr. Benge about earlier experiences of tingling and numbness when he examined her on the morning of February 21st after she had awakened with tingling, numbness and speech difficulties. Finally, and most importantly, when her history was taken by Dr. Gersh after admission to the hospital with severe neurological symptoms, Mrs. Gaul mentioned her flu shot, her cold and her gastroenteritis, but said nothing about prior experiences of tingling and numbness. Dr. Gersh’s history reads, in relevant part:
The patient is a 47 year old white female who had been entirely healthy in the past apart from “colitis.” She had a flu shot last October and her children [sic] had an upper respiratory infection just before admission. A week prior to admission she became ill with mild intermittent ache in the upper mid-abdomen, left shoulder pain, diffuse, bifrontal headache, malaise, sore throat, cough, and fever up to 103°. She was given antibiotics and antihistomines and the “cold” cleared up.
I consider it highly unlikely that someone in Mrs. Gaul’s position on February 21, 1977, would have failed to inform Dr. Gersh of prior episodes of tingling and numbness if they had in fact occurred.
In sum, I cannot accept plaintiffs’ contention that Mrs. Gaul began suffering from GBS in December. Even if one assumes arguendo the correctness of Dr. Alter’s opinion that GBS may begin mildly and smolder for a period of ten to twelve weeks before becoming severe, the record does not support Mrs. Gaul’s assertion that she experienced any notable problem in December. No part of the record created contemporaneously with those events bears her out. The evidence points the other way; rather than an atypical, smoldering case, her GBS seems to have fit rather neatly the accepted temporal and symptomatic profile16 of GBS. Accordingly, plaintiff’s theory that Mrs. Gaul’s GBS began in December, within four or so weeks of her swine flu inoculation and therefore well within the accepted period of demonstrated causal connection between swine flu vaccine and GBS, must be rejected.
III. THE ENHANCED RELATIVE RISK THEORY
Plaintiffs’ second theory focuses on epidemiologic studies of the effects of immunization on the rate of GBS cases. Plaintiffs contend that the epidemiological evidence shows that immunization results in a relative risk of contracting GBS of approximately 4 until at least the sixteenth week after vaccination. The government argues that the evidence only supports the finding of an increased relative risk of GBS from immunization until the tenth week after immunization. It argues that since Mrs. Gaul’s GBS occurred in the sixteenth week after inoculation, it is more likely than not that she would have contracted GBS even if she had not received a swine flu vaccination. Both sides presented expert testimony on the subject.
Plaintiffs presented the testimony of Dr. Martin Goldfield. Dr. Goldfield is board certified in medical and public health and microbiology and has expertise in epidemiology and infectious diseases. He was in charge of part of the New Jersey Health Department during the outbreak of GBS associated with the swine flu immunization program.
Dr. Goldfield performed a statistical analysis using the data collected by the Center for Disease Control in Atlanta *1128(CDC) from the beginning of the Swine Flu Program on October 1, 1976, through the discontinuation of immunization on December 18,1976, until the end of data collection by CDC on January 29, 1977. His analysis leads him to conclusions different than those of others who have analyzed the same data.
Goldfield first calculates the rate of GBS in the unimmunized population in units of cases of GBS per million persons per week of observation. He calculates that in weeks one through twelve there was an average of roughly .22 cases of GBS per million persons per week. In weeks thirteen through sixteen, however, he says that because reporting of cases dropped after the twelfth week of the program (i.e., at about December 18, when immunization was stopped), the rate drops to an average of approximately .11 cases of GBS per million unimmunized persons per week. These numbers form the basis for comparison of the rate of GBS in the immunized population with the rate in the unimmunized population.
Dr. Goldfield next takes the number of immunized persons observed each week and the rate of GBS cases in the unimmunized population and calculates the number of cases one would expect to see if the rate of GBS was no different in the immunized population than in the unimmunized population. He compares the resulting expected number of cases to the actual number of cases observed among the immunized in each week. The ratio between the actual number of cases and the expected number of cases is the relative risk ratio. With these calculations, Dr. Goldfield generates Table 12 in Plaintiffs’ Exhibit 58.
Table 12
Ideal Method
Best Summary of Relative Risks of GBS17
Among Population Older Than Seventeen
By Weeks After Immunization
October 1,1976 to January 29,1977
From these data, Dr. Goldfield concludes that immunization increases the risk of contracting GBS through the sixteenth week after immunization by a factor of more than 4. If this is true, it would mean that there is a significantly greater than fifty-fifty chance that Mrs. Gaul’s GBS, which occurred in the sixteenth week following inoculation, was caused by the swine flu vaccine.
Dr. Goldfield’s analysis differs from those relied upon by the government in a number of ways. It is necessary to discuss only two of those differences. First, in his primary presentation, Dr. Goldfield used approximately forty million persons as the *1129total immunized population. The CDC figure, which I accept, was 43.3 million persons.18 This difference has a significant effect on Dr. Goldfield’s relative risk ratios because a correction of his analysis not only increases the immunized population by over three million persons but also decreases the unimmunized population by over three million persons. Anticipating this criticism, Dr. Goldfield prepared a supplementary set of tables, Plaintiffs’ Exhibit 59, using an immunized population of forty-three million. Table 12 shows the recalculated relative risks:
Table 12
Ideal Method
Best Summary of Relative Risks of GBS19
Among Population Older Than Seventeen
By Weeks After Immunization
October 1,1976 to January 31,1977
(Using CDC’s Estimate of 43.327 Mil.)
The second major difference between Dr. Goldfield’s analysis and that of the other epidemiologists who have studied the CDC data relates to his base rate, i.e., the rate of GBS among the unimmunized and thus the rate one would expect if there were no causal connection between GBS and the flu vaccine. As a result of a significant drop in reported cases of GBS (among the unimmunized as well as the immunized population) in eight states after December 18, 1976 (the week in which the inoculation program was terminated), Dr. Goldfield used an attack rate for the unimmunized population of .22 per million per week of observation for weeks one through twelve and an attack rate of .11 per million per week for weeks thirteen through sixteen.20 He considered this approach justified because the drop in reporting after December 18 was substantial and because the data led him to conclude that there was a drop of approximately the same amount in the reporting of GBS among both the immunized and unimmunized populations.
Dr. Schonberger, who made the initial evaluation of the CDC data, used a base rate throughout of .22 per million per week. He too noted the reporting drop off after December 18th. Based on a CDC follow-up investigation of that phenomenon, he excluded from his calculation of the rates of attack for both the immunized and unimmunized the data reported by the *1130eight states where there had been a significant drop in reporting. While Dr. Schonberger did not testify at trial and only limited portions of his deposition were offered,21 Dr. Nathanson, a highly qualified epidemiologist who did testify, adopted Dr. Schonberger’s approach in his analysis of the data.
The difference between these approaches is significant for present purposes. If one adopts all of the assumptions of Dr. Goldfield’s “Best Summary” methodology, other than those concerning the size of the immunized population and the base rate, the relative risk ratio for weeks eleven through sixteen drops to 1.66, substantially below the more likely-than-not, relative risk ratio of 2. If one uses a population of 43.4 million and a base rate of .22 per million per week, the figures on the bottom of Table 12 change as follows:
Table 12 22
Weeks After Immunization Expected Number of GBS Cases Observed Number of GBS Cases Relative Risk
8-10 24.198 44 1.82
11-16 18.703 31 1.66
For several reasons, I conclude that it is more appropriate to use a base rate of at least .22 per million per week. I think it reasonable to infer that significant drops in reporting in some states immediately following the moratorium are reporting artifacts and that the data from those sources are not as reflective of actual attack rates as the remainder of the CDC data. If so, it seems reasonable to exclude the data from those states after December 18th in calculating relative risk ratios.24
Dr. Goldfield appears on the current record to be the only analyst to use a split base or to utilize a base rate of less than .22 per million per week. Even plaintiffs’ other expert, Dr. Alter, could see no justification for a split base and considered Dr. Schonberger’s .22 per million per week to be a conservative, understated figure for the natural rate of GBS occurrence. The view that .22 is a conservative figure was shared by others such as Dr. Nathanson and Dr. Dyck. Dr. Schonberger testified that he had received criticism that his .22 base rate was too conservative.
The conservative nature of a .22 per million per week base rate is also suggested by the GBS base rate found by two other studies in which case ascertainment methods are considered by epidemiologists to be excellent. While neither study was of a national population, they have some probative value as to the natural rate of GBS. The Michigan Study followed Michigan’s 6.3 million adults, 2.3 million of whom received swine flu vaccinations, from July 1, 1976 to April 30, 1977. It found a base *1131rate of .36 per million per week.25 The Olmstead County Study conducted by the Mayo Clinic utilized a base rate of .44 per million per week based on data collected between 1935 and 1977.
Finally, I am persuaded by Dr. Nathanson’s explanation of the difference in the CDC data between the attack rate in the immunized population and the attack rate in the unimmunized population after the tenth week. This explanation and his analysis also tends to confirm that Dr. Schonberger’s base rate was conservative. There was consistent testimony from both epidemiologists and clinicians (1) that not all GBS cases were reported, (2) that one would normally expect a higher percentage of vaccinated GBS cases than unvaccinated GBS cases to be reported because of the interest in a possible connection between swine flu vaccination and GBS, and (3) that one would normally expect better case ascertainment of severe cases in both populations than of milder cases. In accordance with recognized epidemiologic practice, Dr. Nathanson analyzed the CDC data by the severity of the case involved. In GBS cases with extensive motor involvement, he found that the attack rates in the vaccinated and unvaccinated populations were virtually identical after the eighth week. In GBS cases involving limited motor involvement, Dr. Nathanson found that beyond the eighth week the attack rate among the immunized exceeded that of the unimmunized by a small but relatively constant amount. The conclusion to be drawn from this analysis is that where case ascertainment was best there was no material difference in the attack rates after the eighth week and that the excess of the vaccinated rate over the unvaccinated rate beyond the eighth week shown in Dr. Goldfield’s analysis26 is attributable to the fact that there was greater unreporting of unvaccinated GBS cases in the limited involvement category than in vaccinated GBS case in that category. If these inferences are drawn, Dr. Goldfield’s analysis does not support the conclusion that Mrs. Gaul’s GBS was more likely than not caused by the swine flu vaccination.
IV. CONCLUSION
Mrs. Gaul received a swine flu vaccination on November 5, 1976. Thereafter, she contracted a severe case of GBS which commenced with a classic rapid onset on February 20, 1977. Neither the clinical evidence nor the epidemiologic evidence tendered in this case suggests that there was a causal connection between these events. Accordingly, judgment will be entered for the United States.