Science on Trial

Related articles

Jury selection began last week in California in a case where employees will argue that IBM knowingly exposed them to chemicals used in the manufacturing of chips and disc drives that caused a variety of cancers, birth defects and other ailments. If found liable, IBM faces hundreds of millions of dollars in damages.

IBM (and its suppliers) are the defendants here, but science is also on trial in this landmark case. Will the jurors be swayed by the emotional appeal of ailing workers? Or will they evaluate plaintiff claims in the context of sound science and epidemiology (the study of the cause of human disease)? The outcome of this trial has implications that go far beyond IBM. If science does not prevail, the semiconductor-electronics industry, which has fueled our nation's economy in recent decades, will come under legal assault and consumers, as usual, will end up assuming the costs in the form of higher prices and fewer innovations.

The case presents a real challenge for the IBM defense team. Juries often tend to sympathize with workers who are convinced that a big corporation made them sick. To complicate matters, the overwhelming majority of American consumers are seriously misinformed about the relationship of chemicals, the environment and human disease. Years of activist and media propaganda about alleged "carcinogens" in the air, water, food and workplace have convinced people that trace exposures to industrial chemicals represent a threat to human health when, indeed, they do not.

To succeed, defense attorneys must give jurors a crash course in epidemiology, emphasizing at least 7 points which will undermine the claims of plaintiffs lawyers that these chemicals caused the diseases claimed by defendants:

#1. The word "carcinogen" needs to be strictly defined. In the vast majority of cases, when the media refer to a "carcinogen," they mean a chemical that has been shown to cause cancer but only at very high doses and only in laboratory animals. For example, in reports about the IBM trial, readers learn that workers may have been exposed to the "carcinogen" trichlorethylene. They were not informed of the essential fact that it is a chemical that has been shown to cause cancer in laboratory animals but not in humans.

#2. Extrapolation from animal data in predicting human cancer risk is fraught with uncertainties. Mainstream public health scientists do not accept data from one or two high-dose tests on one animal species as evidence that a given chemical poses a human cancer risk.

#3. Even in cases where a worker is exposed to a chemical that is a known human carcinogen, this does not necessarily mean that a given plaintiff was at risk. It is the dose that makes the poison. For example, during the trial it may be argued that IBM workers were exposed to benzene, which at high doses over long periods of exposure has been shown to increase the risk of leukemia. But if the level of exposure was negligible, then so is the risk.

#4. The word "cancer" is a term used to describe a number of different diseases, most of which have different causes. In studies of occupationally-linked cancers, the risk is generally site-specific. For example, workers may claim that that their lung, colon, cervical and breast cancers are caused by benzene, but even setting aside the question of exposure level, there is no evidence of benzene's effects on these sites in the body.

#5. In recent years, Americans have come to believe there is a known "cause" of every case of cancer or birth defect. "There must be some reason I developed breast cancer and my best guess is it had to be from chemicals on the job," a plaintiff might argue. But the reality is that while we have learned a great deal about risk factors for human cancer, a substantial number of causes are simply unknown or related to genetics.

#6. Exposure to some specific occupational chemicals has been scientifically shown to increase cancer risk. For example, in addition to the high-dose benzene association with leukemia, vinyl chloride has been linked to liver cancer and 2-naphthylamine to bladder cancer. The overwhelming majority of such links were established decades ago before today's worker protection programs were put in place.

#7. Epidemiologists identify occupationally-caused disease by determining if there were significantly higher rates of specific illnesses among workers than in a comparable group of persons. More refined epidemiological analysis would identify higher rates of the specific disease among workers more highly exposed, an observation that would further advance the argument for causality. Absence of elevated cancer (or other disease) rates in workers suggests the lack of a causal relationship.

Will the attorneys and their designated expert witnesses be able to give jurors a crash course in epidemiology so that they have the tools to sort out facts from speculation? It is well worth the concerted effort to do so to protect all of us from verdicts based on popular lore instead of scientific reality.