Dr. Schonberger was able to calculate the extent to which the incidence of GBS in the vaccinated population exceeded the incidence of GBS in the unvaccinated population for any given week after immunization, within the temporal bounds of the study. Presenting this information on a graph, one obtained a standard epidemiological curve: the risk of GBS in the vaccinated population was seen to rise sharply above the background rate (the line representing the rate of GBS in the unvaccinated population) during the first two to three weeks following immunization. Thereafter, the line representing the incidence of GBS in the vaccinated population dropped sharply over the next three to four weeks and then tailed off gradually to a plateau level. As a result of the Schonberger study, the government stipulated to liability in cases of GBS occurring within ten weeks after receipt of a swine flu shot.
28. In the course of multi-district pretrial discovery, serious questions were raised concerning the validity of the Schonberger study and its conclusions. As a result, plaintiff's expert, Dr. Goldfield, obtained access to the CDC data base and performed his own study. In response, the government commissioned a panel of experts from a variety of disciplines, including Dr. Nathanson, to re-evaluate the CDC data base and produce a new epidemiological study.
29. For a variety of reasons, it is impossible to know what the true rate of GBS was in the vaccinated population following the administration of the swine flu shot, or what the true rate of GBS is in the unvaccinated population. The most important reason for this is that reporting of cases of GBS was almost certainly incomplete to some degree both when the CDC was collecting data in 1976-77, and during all other attempts to ascertain the "background" rate of GBS without regard to a particular triggering event. There may have been varying degrees of under-reporting between vaccinated and unvaccinated cases; between early-onset and late-onset cases; between mild and severe cases; and according to the geographical area in which data were accumulated. The degree of under-reporting may have been affected by the termination of the swine flu inoculation program in mid-December, 1966, and it was certainly drastically affected by the CDC's decision to stop soliciting reports at the end of January, 1977. Another reason why it is impossible to determine the true rates of GBS for the vaccinated population is that many reports of cases of GBS were incomplete or otherwise unreliable. The symptoms of GBS, especially in mild cases, are easily confused with those of numerous other neurological problems. The publicity which followed the discovery of a connection between the swine flu vaccine and GBS may have created a tendency for certain neurological disorders to be diagnosed as GBS to a greater extent at this time than previously, or to a greater extent among vaccinated patients than among unvaccinated patients.
30. Given the impossibility of determining the true rates of GBS in the vaccinated and unvaccinated populations, the role of the epidemiologist is to develop, from the available data, the best possible comparison of rates of GBS in the vaccinated and unvaccinated populations. The epidemiologist, with or without the help of experts from other disciplines, must make choices and assumptions concerning the reliability of data and the magnitude of error introduced by such factors as under-reporting. In making these choices and assumptions, the epidemiologist may consult, or test his tentative conclusions against, biological and anecdotal information.
31. A comparison of Dr. Goldfield's and Dr. Nathanson's testimony reveals there is substantial room for responsible epidemiologists to differ significantly on many of the key choices and assumptions to be made in analyzing the causal relationship between the swine flu vaccine and GBS. I feel somewhat disadvantaged by my own lack of epidemiological expertise in performing my task of moderating Dr. Goldfield's and Dr. Nathanson's professional dispute. However, I am guided by the following considerations.
First, I am impressed by the great range and depth of expertise represented on the Nathanson panel. Besides Dr. Nathanson, the panel included two other leading epidemiologists, a highly respected neurological expert (Dr. Maurice Victor), and a computer expert and statistician who had particular knowledge about the details of the data compilation and storage at CDC. Although Dr. Goldfield attempted to discredit the work of the Nathanson panel as reflecting a pro-government bias, I found Dr. Nathanson's testimony highly credible, and, from his description, the work of the panel would appear to have been scientific and impartial. Moreover, the panel's interdisciplinary approach to the problems created by the imperfect data base would appear to me to reflect sound scientific judgment. The fact that Dr. Nathanson relied in part on the findings of his colleagues
in formulating his own opinions neither makes those opinions inadmissible nor detracts from the weight to be accorded them.
Second, I am disposed to give greater weight to Dr. Nathanson's study because his conclusions appear to me to make more biological sense. All epidemiological studies of GBS and swine flu vaccine have revealed a sharp rise and fall of the relative risk in the vaccinated population within the first six to eight weeks following inoculation. After that point, all studies show the line representing the relative risk in the vaccinated population as leveling off, with some minor variations attributable to the increasingly small sample sizes as the number of weeks following immunization increases.
The crucial question for determination of the factual issue of causation in this case is whether the plateau level of the relative risk in the vaccinated population in the later weeks following immunization was approximately equal to the background rate of GBS in the unimmunized population or significantly elevated above the background rate. Dr. Nathanson testified that one would expect, as a biological matter, that the plateau level would represent the end of the epidemiological curve: a return to the background rate of GBS among the unimmunized, signifying the end of the period of elevated risk associated with the vaccine. Dr. Nathanson's analyses of the data bore this hypothesis out, and Dr. Goldfield's do not convince me to the contrary. Rather, Dr. Goldfield's analysis seems to me to be marked by a tendency to select figures and assumptions, when a range of reasonable alternatives is presented, which will lead to the conclusion that the relative risk in the immunized population remains significantly elevated for the maximum period which can be studied from the available data. This is particularly true of his assumptions concerning the degree of underreporting which occurred in the immunized, as opposed to the unimmunized, population following the termination of the immunization program in mid-December, 1976, and of his assumptions concerning the background rate of GBS. Thus, accepting Dr. Goldfield's testimony at face value, it only supports the conclusion that an epidemiologist could analyze the available data in such a way as to reach the conclusion that the relative risk in the immunized population remained significantly elevated above the background rate for an extended period of time. It does not establish that the various analytical assumptions and choices essential to reaching that conclusion are the only reasonable, or the most reasonable, assumptions and choices an epidemiologist could make.
For the above reasons, I am not persuaded that it is more probable than not that plaintiff's long-onset of GBS was causally related to the swine flu vaccine.
I note in this regard that, although my finding is based upon careful consideration of the evidence presented at this trial, I have also read and considered many of the reported and unreported decisions of other district judges in similar cases to which counsel have referred me. I found the opinion of Judge Schwarzer in Cook v. U.S., 545 F. Supp. 306 (N.D.Ca.1982), particularly helpful in my efforts to understand and evaluate the evidence in this case.
32. The government argued at trial that, even if one could conclude from epidemiological evidence that a case of GBS occurring 14 weeks after a swine flu inoculation was more probably than not caused by the shot, it is nevertheless more probable than not that plaintiff's GBS was caused by an antecedent respiratory infection rather than by the swine flu shot. I found the government's argument elusive and inadequately supported by the evidence. Consideration of the possibility that an antecedent illness caused plaintiff's GBS played no role in my finding that plaintiff failed to carry his burden of persuasion on the issue of causation.
CONCLUSIONS OF LAW
1. This court has jurisdiction over the subject matter of this controversy and over the parties.
2. The plaintiff having failed to carry his burden of persuasion on the issue of causation, judgment will be entered for the defendant.