Terence Hines - Pseudoscience and the Paranormal

Let us repeat the marble demonstration—but this time the floor has patterns of lines drawn on it that represent power lines. Again, 1,600 marbles are thrown into the room and scatter at random. Will it be possible to find squares in which there is a concentration of both “power lines” and marbles? Certainly. And, again, the coincidence will be due purely to chance factors, not to any ability of the drawn lines to attract marbles to them. To determine if the relationship between power lines (or any other variable) and disease is real, it is necessary to show that the association occurs more frequently than would be predicted by chance alone.

By 1997, after eighteen years of research on the issue, it was clear that there was no association between power lines and cancer. Studies claiming to show such an association suffered from various serious flaws in methodology and statistical procedures (see National Research Council 1997; Linet et al. 1997; Lacy-Hulbert, Metcalf, and Hesketh 1998, for highly detailed reviews). One study (Feychting and Ahlbom 1993), done in Sweden, at first, looked like truly strong evidence that power line exposure was casual in childhood leukemia. The results looked so powerful, in fact, that the Swedish government was planning to switch children from schools near power lines to schools farther from such lines. In the published study, sixty-six risk ratios are reported. A risk ratio (RR) is a number giving the relative risk of some outcome (i.e., disease, accident, etc.) in a group exposed to some condition relative to a control group that is not exposed. If the RR is 1, there is no difference between the risk of the outcome in the exposed and nonexposed groups. RRs of less than 1 indicate that the exposed group is less likely to suffer the outcome. (Lest this seem implausible, note that the RR for dying in a traffic accident is less than 1 for a group “exposed” to wearing seat belts). RRs of greater than 1 indicate that the risk of the outcome is greater in the exposed group. Note, by the way, that RRs can’t be negative, so that there is much more “room” to find RRs of greater than 1 than those of less than 1. Generally, RRs of 2 and above, if found consistently, are considered good evidence that exposure to whatever is being studied does have a real effect on the probability of suffering the outcome. For example, the RR for contracting lung cancer in smokers is 20 compared to nonsmokers; that is, smokers are about twenty times more likely to get lung cancer than nonsmokers.

With that background in mind, the Swedish study (Feychting and Ahlbom 1993) seemed rather impressive. One of the RRs it reported was 3.8 for leukemia in children. In other words, children exposed to power line radiation were almost four times more likely than nonexposed children to develop leukemia. But even in the context of the published paper, that single result loses some of its terror-inducing properties when one notes that the paper reported sixty-five other RRs for different types of cancer and different measures of exposure to EMFs. Of the total of sixty-six RRs, twenty-three were 1 or less. Still, it was the 3.8 RR that was the focus of attention. It turned out, however, that the published paper didn’t tell the whole story. When the results of all the comparisons that the investigators made were released in the mid 1990s, that single RR of 3.8 turned out to be much less significant. The investigators had in fact calculated a total of nearly eight hundred different RRs (Frontline 1995). Of this huge number, the RR of 3.8 for leukemia was simply the one that happened to be the largest. If one calculated eight hundred RRs for such things as picking your nose, using pens with red ink, or listening to National Public Radio, it would be extremely surprising if, just by chance, you didn’t turn up one or two RRs of 3.5 to 4! When the real nature of the results of this study became clear, the Swedish government abandoned its plans to bus children to new schools.

It has also been claimed that studies of animals and of individual cells in culture have shown harmful effects of exposure to EMFs. As is so often the case, a few initial and preliminary studies seemed to show positive results. Later, more carefully controlled studies showed nothing. Lacy-Hulbert, Metcalf, and Hesketh (1998) have provided a comprehensive review of this aspect of the EMF research.

In the United States in 1997, the National Research Council published a report that concluded that there was no evidence of any harmful health effects of EMFs. The same year Linet et al. (1997) published a major study in the New England Journal of Medicine of 638 children who had ALL and found that these children had had no greater exposure to EMFs than control children without ALL. In an accompanying editorial Campion (1997) noted that “[i]t is sad that hundreds of millions of dollars have gone into studies that never had much promise of finding a way to prevent the tragedy of cancer in children” (p. 46).

The most recent instantiation of the “EMFs cause cancer” claim is that the radiation from cellular phones causes brain cancer, especially in the brain tissue in the part of the brain closest to the phone itself. In 1992 a tragic case in which a woman died of brain cancer was highlighted in the media when her husband had become convinced that her use of a cellular phone had caused her cancer. He went on a crusade against cell phones that included suing the major cell phone manufacturers. (Somewhat surprisingly, Paul Brodeur does not appear in this part of the story). Keller (1993) has provided a good history of the early stages of this controversy. One result was that the stock of cell phone manufacturers plummeted in value. It also led to widespread worry among cell phone users.

At the time, no epidemiological studies had been done of cell phones. The publicity over this case, however, led to the initiation of several studies of the risks of cell phones. Muscat et al. (2000) studied 469 patients with primary cancerous brain tumors. These patients had not used cell phones at a greater frequency or for a greater time than a control group that did not have any brain cancer. Inskip et al. (2001) reported a similar study of 782 patients and also found that cell phone use was not related to brain cancer, nor was there any association between the side of the brain in which the tumor was located and the side of the head to which the phone had been held, if the patient had been a cell phone user. Finally, Johansen et al. (2001) examined cancer rates in 420,095 cell phone users in Denmark between the years 1982 and 1995. Within this huge sample, there was no greater incidence of brain tumors than would be expected. In addition to examining brain tumors, this study examined the incidence of cancer in all organs of the body. Overall, cell phone users had a significantly lower rate of cancer than expected. This was due to reduced rates of “lung cancer and other smoking related cancers” (p. 203). This effect, in turn, was probably due to the fact that cell phone users are somewhat more affluent than nonusers and thus tend to be more aware of real health risks such as smoking.

Polychlorinated biphenyls (PCBs) are a class of chemicals that were used in numerous industrial processes until the 1970s. Through various routes, they entered the environment in small concentrations. The cause of most of the worry over PCBs is the fact that they are found in freshwater fish that are often eaten by humans. Specifically, it is claimed that exposure to PCBs in the workplace causes cancer and that the children of women who ate fish contaminated with PCBs show various physical and mental abnormalities. It is this latter claim that has caused the furor over PCBs because, while they are no longer found in the workplace, they are still present in the environment and, as noted, in fish. In the Hudson River Valley in New York State, worry about the effects of lingering PCBs is so great that the Environmental Protection Agency has ordered General Electric to spend hundreds of millions of dollars dredging the river bottom to remove contaminated soil. The important question thus becomes whether exposure to PCBs by eating PCB-contaminated fish really poses a hazard to human health.