6. POWER-INDUSTRY SCIENCE AND POWERLINE EMF HEALTH HAZARDS
Neither scientists nor the public can rely on power-industry research or analysis to help decide whether powerline electromagnetic fields affect human health because power-industry research and analysis are radically misleading.
To decide whether powerline EMFs affect human health, it is necessary to produce scientific data by means of appropriate experiments, and it is necessary to analyze data to infer its meaning and overall significance. Production and analysis of data are distinct activities, and both are expensive. Over-simplistic as it may sound, whoever pays for EMF bioeffects research and analysis determines what data is produced and the way it is interpreted.
Soon after the possibility that powerline EMFs were health risks was raised in a legal dispute involving the New York Public Service Commission, power companies and their trade associations, particularly the Electric Power Research Institute (EPRI), became massively involved in EMF bioeffects research. Subsequently, the power industry dominated funding of the effects of powerline EMFs, both in terms of absolute dollars and compared with dollars from non-industry sources.
More than twenty-five years have elapsed since the power industry began its EMF activities, and it is now possible to evaluate the industry's role. I will show here that the power companies and their trade associations were deeply deceitful regarding the information they provided to scientists and to the public regarding the potential health hazards of powerline EMFs.
POWERLINE EMF RESEARCH AT BATTELLE
Battelle is a private company that performs contract research of many different types for many different organizations. Battelle began powerline EMF activities on behalf of the power industry in March, 1976, and this relationship has continued to the present, without interruption. The dimension of Battelle's involvement with EMFs is hard to discern exactly, but it far exceeds in scope and impact that of any other group or organization that has performed EMF research. Battelle has probably received more than $100,000,000 in funding for EMF research, and its employees have made more than 1000 presentations and reports dealing with EMF bioeffects issues.
Battelle's EMF research mostly involved the effects of powerline EMFs on rats, mice, and pigs. The experiments consisted of exposure of the animals to EMFs, followed by many different kinds of physiological measurements. Various investigators at Battelle designed and conducted the experiments, disseminated the results, and defended them in scientific forums. Most of the Battelle experiments, presentations, and reports were negative, by which I mean that the studies, either on their face or as interpreted by the Battelle investigators, failed to suggest that powerline EMFs were health risks.
The Battelle investigators urged that the negative studies were presumptive evidence of powerline safety, and disinterested scientists who reviewed Battelle's negative studies frequently agreed that the negative results suggested that powerlines were safe. But the Battelle investigators designed their studies and handled their data intentionally to produce negative results, and to produce the perception that the results were negative even when they were positive. Under these conditions, the negative studies did not justify an inference of powerline safety because the negativity was made, not found.
Negative Results by Design
Battelle investigators designed and performed many EMF studies in which the measured parameter had no plausible sensitivity to EMFs. In these cases the results were foreseeably negative because one would not expect an effect due to the EMFs. For example, in a study of the effects of powerline EMF exposure on heart rate in rats, the heart rate of the animals was measured only after the animals were removed from the EMF and then confined in narrow tubes so that they could not turn, rear, or make other normal movements. It would be expected that the stress of confinement in the tubes would alter heart rate, thereby obscuring any effect due to powerline EMFs; not surprisingly, the study was negative.
In another study, Battelle investigators measured the effect of powerline EMFs on visual evoked potentials in the brains of rats. Such potentials are sometimes used to diagnose pathological changes in the visual systems of patients, but there was no evidence whatever to suggest that evoked potentials would be a worthwhile parameter to measure in connection with EMF exposure. This was particularly the case in view of the method used by the Battelle investigators to measure the potentials. Normally, electrodes are attached to the head of the subject using an electrically conducting adhesive. This method of attachment minimizes the stress caused by the measurement process itself, thereby protecting the integrity of the results. The Battelle investigators, in contrast, drilled holes through the skulls of the rats and placed the electrodes directly on the brain, thereby making the measurements hopelessly insensitive to the effects of EMFs. The results were negative, but not finding an EMF-induced change that one had no reasonable expectation would occur was not evidence that powerline EMFs were safe. Nevertheless, that was Battelle's rationale for the study and the way the results were interpreted.
The question whether powerline EMFs are stressors is important because stress can worsen the consequences of any human disease, and Battelle investigators tried to show that powerline EMFs were not stressors. In these experiments, however, they built special cages that confined the test animals in abnormally small areas. For example, mice were confined to cages that were only 2 inches high, and rats in cages that were only 4 inches. Federal guidelines for caging mice and rats stipulated cages having minimum heights of 5 and 7 inches, respectively, precisely because that was the veterinary consensus regarding what was appropriate for stress-free housing conditions for each species. The published results of Battelle's studies using abnormally small cages indeed failed to find evidence that powerline EMFs were stressors, but that conclusion was foreordained by the way the animals were housed. Both the EMF-exposed and the control animals were already stressed as a result of their crowded living conditions.
The pervasive consequences of the crowding were shown by the Battelle results obtained between March, 1976 and November, 1977. During this period, Battelle investigators found only two positive effects that they considered to be potentially adverse, out of more than 380 parameters that they measured in their chronically crowded animals. These overwhelmingly negative results were reported in almost 50 contemporaneous presentations and papers.
Negative Results by Analysis
Some Battelle experiments yielded positive results. On their face, positive results would appear to a disinterested scientist to suggest that powerline EMFs were not safe, following the logic used with the negative studies that led to the opposite conclusion. But positive results were the opposite of what Battelle's clients wanted, and Battelle invoked various artifices to insure that positive results were not recognized as positive. One way this was accomplished involved the device of replication.
When the Battelle investigators found a positive effect, they routinely repeated the experiment. Superficially, this practice appeared to be an honest procedure, predicated on the possibility that the positive effect might have been a statistical fluke. Only the positive effects were usually replicated, however, even though the negative results might also have been statistical flukes. Thus, the routine procedure of replicating only positive effects created a pervasive bias in favor of the general conclusion that powerline-EMF studies were negative. Adding to this bias was the way the Battelle investigators interpreted the overall result when the replicate of a positive experiment was negative. In those cases, the Battelle investigators arbitrarily concluded that both experiments, taken together, were negative.
In some instances, both the first study and the second study of a particular type were positive; in that event the study was repeated a third time. If the results of the third study did not exactly match the results of the first study and the second study, then the set of three studies was considered to be a negative study. For example, they observed an inflammation of the prostate glands of rats that were exposed to EMFs for 30 days. The experiment was repeated, with the same result. The experiment was repeated for a third time, but the 67% increased rate of prostatitis in the EMF-exposed rats was not statistically significant. The investigators concluded that, overall, EMFs had no effect.
Battelle's strategic use of replication forced the inherent uncertainty in biological studies to favor the point-of-view of Battelle's clients. In theory, the results of biological studies must be certainly yes, certainly no, or somewhere in the middle. The Battelle investigators arbitrarily interpreted the two most likely outcomes in favor of the power industry.
Negative Results by Omitting Positive Results
If an investigator performs an experiment and then withholds some of the data, without explanation, it's easy to see that a disinterested scientist who reviews the published data might be misled. Relevant data was routinely withheld by the Battelle investigators.
For example, in one of their endocrinology studies, the Battelle investigators exposed a group of male rats to powerline EMFs for 30 days to assess whether or not the EMF was a stressor. The experiment consisted of recovering the blood of the exposed and control animals and analyzing for the presence of changes in corticosterone levels, which would indicate that the EMF was a stressor. I had previously performed the same experiment several times, and reported that corticosterone levels were altered as a consequence of the EMF exposure.
Using a fluorimetric technique, the Battelle investigators found that corticosterone in the blood of the EMF-exposed animals was 123±17(units of ng/ml), which was less than that in the control animals (175±50). Portions of the same samples were sent to the University of Rochester to be analyzed by competitive protein-binding radioassay, a different and perhaps more specific method of measurement. Using the radioassay method, the corticosterone levels in the exposed animals were found to be even more significantly different than the levels in the control animals (34.9±7.7 compared with 287.0±137.9).
The experiment was repeated using twice as many rats as previously. When the results were analyzed using the fluorimetric method, the exposed animals were again lower than the controls (150±16, compared with 193±32). The radioassay measurements also showed that the levels in the EMF-exposed rats were lower than in the controls (43.4±10.6, compared with 82.8±22.1).
The experiment was repeated a third time; in this case the blood samples were sent to the University of Kansas for analysis. Again, the levels were lower in the EMF-exposed animals (51.5±9.9, compared with 90.8±15.8). In a fourth experiment, rats were exposed for 120 days (4 times longer than the exposure in the first three experiments). Again, the levels again were lower in the EMF-exposed rats compared with the control rats (52±10 and 91±16, respectively). Battelle wrote to the study sponsor: "The data appears to be consistent with similar findings reported by Marino."
But then the 30-day experiment was repeated a fourth time, and there was no difference in the blood levels of corticosterone between the exposed and control rats (42.1±11.6 and 35.6±9.5, respectively). And the 120-day exposure experiment was repeated with the result that the corticosterone levels in the exposed animals was lower than in the controls, but not significantly so (64.4±6.2) compared with 76.5±8.0). When the Battelle investigators published their results, they included only the second of the four 30-day experiments, and the two 120-day experiments, and they concluded that EMF exposure had no effect on corticosterone levels.
The easy ability to hide data or to disclose only that portion that comported with the position of the study sponsor is one of the fundamental weaknesses in the use of trade-industry research results for making public-health determinations about the safety of powerline EMFs. In the endocrinology experiments, for example, if the Battelle investigators had disclosed all the data, the results would likely have been interpreted by disinterested scientists to show that powerline EMFs were stressors. But nothing is more clearly demonstrated by an analysis of the history of EMF bioeffects research than the fact that investigators or organizations that find results suggesting that powerline EMFs are health risks do not have their research contracts renewed. Thus, every instance of a positive effect found by the Battelle investigators created a conflict of interest for them, and in many cases this resulted in their failure to disclose pertinent data that should have been disclosed. In the endocrinology experiments, the Battelle investigators hid the data because it suggested exactly the inference that the power industry sought to avoid.
The Battelle studies involving rats and mice consisted of 12 different kinds of experiments, each of which was headed by a principal investigator who was answerable to the Task Leader, R.D. Phillips. Every instance in which it was possible for me to compare internal Battelle documents with the results of their published experiments I found serious instances of hiding of data, resulting in an altogether different public perception than if all the data were disclosed.
In the Battelle Cardiovascular Function studies, for example, male rats were exposed to powerline EMFs for 30 days and then removed from the field and placed in narrow tubes so that wires could be attached to facilitate measurement of heart rate. In the 1-hour period following removal of the rats from the field, the heart rate of the exposed animals did not differ from that of the controls. The investigators intended to repeat the experiment after 4 months' exposure, but found that the male rats grew too large to fit into the tubes. The experiment was therefore begun again with female rats, resulting in data for male and female rats after 1 month's EMF exposure, and for female rats after 4 months' exposure.
When the Battelle investigators reported their results on heart rate, they described only the results for male rats and for female rats exposed for 4 months, and concluded that there were no significant effects due to the EMF. But their report was misleading for several reasons. First, the unpublished data from the female rats exposed for 30 days was statistically significant, and showed an effect due to EMFs. This was remarkable because it suggested that the effect of the EMF could not be obscured even by the stress of confinement. Second, the reported data for female rats exposed was not the same as that in their monthly report, which seemed to show that the EMFs significantly affected the heart rate for about the first 20 minutes after the rats had been removed from the EMF. Thus, the conclusion of their publication that there were no EMF effects was not true if all the data was considered.
Battelle's Reproduction and Development study also resulted in data that was never publicly disclosed. The reproduction study began in January, 1978, and was intended to refute an earlier study published by me and my colleagues. The plan was to produce 3 successive generations of mice, and to code the data in such a way that some of the people who worked on the experiment could not determine what the results were during the experiment. In February, a second version of the same experiment began in a separate exposure facility 50 feet down the hall from the first exposure facility. Both experiments were scheduled for completion in December, 1978.
Some time prior to November 22, 1978, after only two generations had been born in each of the two experiments, the data codes were broken and the data was analyzed. The interim analysis showed that the EMFs affected the growth rate of the mice in both experiments, whereupon the experiment was changed to a 4-generation study. The fourth generation was born around March, 1979, but its existence was never disclosed.
The results from the first 3 generations showed that the EMFs consistently affected the growth rate of the mice. However, as described in Section 3, because the results were not exactly the same in the two experiments, the Battelle investigators concluded that there were no effects due to the EMF. Because the data from the fourth generation of mice was never disclosed, we can only speculate about how it might have affected the overall interpretation of the study. Perhaps Battelle's procedure of averaging the results of two positive experiments would not have yielded a negative result if the data from the fourth generation was also included. In that case, even scientists who accepted the averaging procedure would be constrained to agree that the overall results of the study were positive.
Negative Results by Argument
Battelle investigators frequently characterized their data as negative even when it was probably positive. By undercutting the obvious implications of their work, the Battelle investigators denied its use to those who might disagree with the power industry position. An outstanding example was the Battelle study of the effects of powerline EMFs on reproduction and development of pigs, which lasted more than 5 years and cost more than $7 million. During the study it began to appear that powerline EMFs produced many different biological effects. When the Battelle investigators published the study they identified a broad range of problems and claimed that these problems, not the EMF, was responsible for producing the biological effects in the pigs. Among the problems were infections, electrical fires, hysterical female pigs, and statistical fluctuations. In each instance where the data apparently disclosed a positive effect, the Battelle investigators chose a non-EMF cause and explained away the positive result.
When this Keystone Kops of powerline EMF studies was published by EPRI, the written record extended to 7 volumes. Even if all of the data was present, Battelle's written and oral reports were so thoroughly hedged, it looked like the study was negative. The Battelle investigators pooh-poohed the inference that data which looked positive was actually positive. Obviously, independent investigators would be reluctant to assert that data was positive when the Battelle investigators themselves would not make that claim. The overall result, therefore, was that the Battelle pig study was generally accepted as negative.
Negative Significance of Concededly Positive Results
Battelle developed a novel strategy for insuring that inferences based on their data could not undercut the position of the power industry, even in those cases where the Battelle investigators admitted that the data was positive. This was accomplished by intentionally compromising the significance of the data using a confounder. The strategy was based on mathematical modeling that, on the surface, seemed designed to resolved a bona fide problem - the important issues of EMF dosimetry and scaling.
What EMF strength should be used in animal studies that will ultimately serve as a basis for answering the question of human risk? Should the animals be exposed to the same strength of EMFs as the people who live hear the powerlines? More? Less? The Battelle investigators performed many mathematical studies that seemed designed to deal with the dosimetry issues. On the basis of these calculations, they claimed that animal studies should be done at about 5 times the strength of the powerline EMFs to which people were exposed.
But the Battelle investigators arrived at the factor of 5 by making a series of assumptions in their calculations. By changing the assumptions, one could produce an infinite number of factors, each of which was as valid as the factor of 5 suggested by the Battelle investigators. Nevertheless, on the basis of their calculations, the Battelle studies were done using EMFs many times stronger than powerline EMFs.
Early in the course of the work, Battelle investigators discovered that the strong EMFs caused the hair on their mice, rats, and pigs, to vibrate. Since these animals, but not people, are completely covered with hair, one consequence of using high EMFs was to destroy the potential scientific significance of any positive effects that might occur. Any such effects could equally be attributed to chronic irritation of the animals due to causing the hair on their body to oscillate continuously, as well as to EMFs interacting with body tissues. The overall result was that the Battelle investigators reported some biological effects due to EMFs, thereby avoiding the absurdity of always failing to find anything, but they did not jeopardize the position of the power industry in doing so because the implications of the positive effects could be explained away. For example, Battelle investigators found that powerline EMFs retarded fracture repair in rats. As a potential explanation, they suggested that the hair vibration caused by the EMF may have increased muscular activity in their fractured legs, thereby inhibiting repair.
The artifact of hair stimulation was used like an ace in the hole. During an EMF blue-ribbon committee meeting, for example, a suggestion by a disinterested scientist that the positive results from a particular Battelle study suggested that powerline EMFs might be a health hazard typically resulted in a remark from the Battelle representative pointing out the potential role of the irrelevant mechanism of hair stimulation. Thus, Battelle's calculations rationalized the use of high EMFs which, in turn, virtually guaranteed that any positive data could not be used for evaluating human health risks.
Unreliability of Contract Research
There is a right way and a wrong way to do science. Scientific misconduct is the general name for the wrong way. I think that, in specific experiments, the powerline EMF research at Battelle was scientific misconduct. But the problem posed by the type of research performed at Battelle and other similar companies is far more serious for science and society than isolated cases of scientific misconduct. The process that produced the scientific data published by Battelle differed too greatly from the process normally employed to produce scientific data. Battelle's data, therefore, simply cannot be treated like data that was produced in the normal way. It does not matter what the data says or doesn't say, the process followed tainted every result.
To appreciate how radically research at Battelle differed from normal research, consider the following comparisons. The National Institutes of Health (NIH) and the National Science Foundation (NSF) are major sources of funding for scientific research in the United States. The primary goal of the research funded by both the NIH and the NSF is scientific truth, by which I mean scientific knowledge that is as reliable as can be obtained by human beings. The particular scientific knowledge sought is subject to priorities that are governed by the political process. The emphasis might be, for example, on breast cancer research, particle physics, developing the internet, HIV research, or space travel. As a result of competition among these and other priorities, the research favored by one group or another might not get done for lack of funds. But if it is done, the research is as reliable as human beings can make it.
Research funded by NIH and NSF is characteristically innovative. Both agencies explicitly seek new and novel solutions to particular problems, and the degree of innovation of a particular proposal is an important factor in the funding decision regarding it. Moreover, both agencies expect that proposed innovative research will be carried out by competent investigators having a demonstrated history of performing the types of studies proposed. It is difficult to imagine, for example, that the NIH would fund a grant application by a principal investigator who had never performed the type of studies proposed in his application or, even worse, had attempted to perform the experiments but failed because he lacked the necessary technical capability.
NIH and NSF research is peer-reviewed, by which I mean an independent and disinterested group of scientists evaluates individual proposals, compares them with other similar proposals and ranks them accordingly. The actual mechanism of evaluation differs between the two organizations, but in both cases it is usually true that a specific proposal is actually reviewed by peers of the individual making the proposal.
Individual peer reviews are regarded as privileged communications and not disclosed to the public; the reasons for this have never been clear to me. However, essentially all other aspects of the funding of scientific research by NIH and NSF are open and public, and can be examined by anyone who chooses to do so. Thus, the actual funded research proposal submitted by each investigator is listed on the agency internet servers, and can be obtained from the agency. Each of the annual reports submitted by investigators of funded projects are also public documents. The rules of both NIH and NSF also require that any data obtained using federal funds be freely available to all other investigators who request it. Thus, the traditional openness and free availability of scientific information that has been characteristic of the development of science throughout its history is part of the tradition and the specific rules of both the NIH and the NSF.
The conduct of powerline EMF research at Battelle differs markedly from each of the above mentioned characteristics of scientific research funded by NIH and NSF, and the Battelle research suffers badly in comparison. The ultimate goal of the Battelle EMF research was the economic advantage of the power industry, not scientific truth. Specifically, they sought to produce scientific information that supported the positions of the directors of the power companies. The willingness of the power companies to pay the hefty price for Battelle's EMF research reflected the power industry's judgment regarding priorities affecting its business, and had no necessary connection with scientific truth or public priorities. The industry's priorities translated into Battelle's goals which in turn determined Battelle's specific activities. If the industry-Battelle axis did result in scientific truth or if it fostered the welfare of the general public, those benefits would be accidents, not the result of design.
Whereas NIH and NSF research is innovative and competent, Battelle research was almost always reactionary. It is not possible to identify a single fruitful line of research that was initiated by Battelle investigators. On the other hand, it is almost always possible to identify a line of powerline EMF research that each Battelle report was designed to rebut, replace, or otherwise undercut. The hypotheses for the bulk of their studies was that a previously reported EMF-induced bioeffect was an artifact, and in most cases the Battelle investigators supported their hypothesis. NIH and NSF do not fund studies having such hypotheses. In contrast, the Battelle studies almost always had such a hypothesis. It is simply impossible for honest EMF investigators to establish scientific truth under these circumstances because anybody can perform a study and not find something that was found in a previous study by someone else. It simply takes no skill whatever to do this.
As a rule, the Battelle investigators had few publications prior to beginning powerline EMF research. If their negative powerline EMF publications were erased, they would still generally have few publications in the peer-reviewed scientific literature. This suggests that the Battelle investigators did not have the training and expertise necessary to perform the studies that they were hired to perform. The trade associations were indeed free to hire anybody they wanted to perform their research because it was trade-association dollars that were spent. Legally, therefore, no one can insist that their dollars should be spent only for innovative research done by competent investigators. On the other hand, it would be foolish to treat the Battelle work product as if it were done by competent scientists pursuant to innovative experimental designs.
Nothing about the research at Battelle was released to the scientific community or the public except for material that was approved by the power industry. In contrast to NIH and NSF research where documents describing the experimental design as well as the actual data are freely available to all interested persons, the experimental designs of the Battelle investigators and the data they obtained were not made available because the Battelle research was a private contractual affair between Battelle and a particular power company or trade association. No disinterested investigator was in privity with either party, and hence had no right to demand information regarding experimental design or the results of particular experiments. Thus, the legal nature of the relationship between Battelle and the power industry prevented full and open disclosure of pertinent scientific information. The essential nature of the industry-Battelle axis has been known since 1984.
In summary, different masters are served by the open grant system and the secret contract system that governed Battelle's activities. Consequently, the results produced in each system regarding EMF research must be evaluated differently. All claims, conclusions, reports, publications, and presentations by Battelle investigators should initially be doubted or disbelieved because they were primarily intended to serve the interests of Battelle's clients, not scientific truth. In some cases, the Battelle version of the facts may be correct, but the point is that the likelihood of biased information is too great for scientists or the public to believe what Battelle says to the same extent and in the same way they might believe information provided by disinterested scientists whose only goal was truth. Information from Battelle regarding the health risks of powerline EMFs can be rehabilitated and perhaps used in public-health decision-making only in circumstances that provide a mechanism to challenge the responsible Battelle investigators regarding the details of their work.
Marino Home Page | Research Interests
| 58415 |