4.7 Silicone nightmare

The Dow Corning corporation was considered one of the most ethical and progressive corporations in the world with a code of conduct that was cited as a model in case studies by business schools. In the 1977 version of this code, the company agreed that, "Dow Corning accepts as our responsibility a recognition, evaluation and sensitivity to social needs. We will meet this responsibility by utilizing our technological and management skills to develop products and services that will further the development of society." The Dow Corning code focused primarily on conduct in the global marketplace, covering issues like refusing any kind of payment or bribe, avoiding political contributions and respecting the rights of employees.

Here was a company that, on paper at least, was committed to making the world a better place. But one of its products, a silicone breast implant, became the focus of an ethical controversy that would force the company into bankruptcy and make it a pariah.

The Dow Corning Corporation was created in 1943 by The Dow Chemical Corporation and Corning, Incorporated as a jointly-owned business centered non-silicone technology. Dow Corning's first triumph was a sealant used to protect the ignitions in Allied fighter planes from failing at high altitudes during the Second World War. When the war was over, Dow Corning explored non-defense applications of silicone, eventually creating more than 5,000 silicone products. Dow Corning continued to emphasize research and expansion throughout the 1950's and 1960's.

During this time, the company ordered a substantial amount of testing on silicones' effects, both on organisms and the environment. Typically, silicone was found to be chemically inert, failing to cause harmful reactions in rats, monkeys, or even human embryonic cells. With such characteristics, silicone seemed the perfect candidate for use in medical applications, for instance, in synthetic coverings for burn patients and in a coating on needles to ease insertion.

To encourage research in this area, Dow Corning opened the Center for Aid to Medical Research (CAMR) as a source of silicone for in-house and independent medical researchers. Thus, in the early 1960,'s Dow Corning supplied Texas plastic surgeons Frank Gerow and Thomas Cronin with silicone for their medical implant device research. Gerow and Cronin, using Dow Corning silicone, invented the first silicone breast implant as a device to aid women who had undergone mastectomies. At that time, the Food and Drug Administration had no regulations governing implantable devices. Companies like Dow Corning had to determine the safety of devices on their own. The surgeons conducted clinical trials of the implant prior to product introduction in 1964. In the next several years, the implant grew popular for cosmetic surgery as well as reconstructive, and Dow Corning cornered both markets.

Through their Center for Aid to Medical Research, CAMR, Dow Corning also invented many other important medical devices. For example, first introduced in the late 1950’s, Dow Corning created the hydrocephalic shunt, a silicone valve implanted in a child’s head to relieve the effects of hydrocephalus, "an excess of cerebrospinal fluid in the cranial cavity causing enlargement of the skull and mental retardation". In 1952, Doctors F. E. Nulsen and E. B. Spitz originally developed the technique of treating hydrocephalus by inserting a valve into the skull, diverting the excess water from the ventricle to the jugular vein. Later, they used a shunt valve developed by a Dow Corning employee, with encouraging results.

4.7.1 Development of the First Breast Implant

The original Cronin implant was developed for use by mastectomy patients to replace ones made of sponge, which tended to harden and appear less natural over time. This new mammary prosthesis was a breast-shaped silicone sac filled with silicone gel. The sac was in elastomer form, and the gel was high-density. The elastomer had a much more tightly woven atomic pattern, which kept the gel inside the sac. The gel was firm, yet pliable, so that it successfully simulated natural breast tissue. With over 10 years of research already completed on it, silicone was a natural candidate for use in breast reconstruction. In addition, it had already been utilized in other medical applications, such as the hydrocephalic shunt and life-saving pace-maker. These cases could provide information on the long-term effects of silicone on the host.

1962 marked the first implanting of a mammary prototype. For the next two years, selected surgeons used the implants in clinical trials to obtain information on their performance, both long and short term, before Dow Corning took the Cronin implant fully to market in 1964. Additional support for the implants was found in the hydrocephalic shunt’s performance , since it used the same elastomer. By 1962, about four-thousand shunts had been placed in children’s brains without any apparent ill effects.

Also, in 1962, the National Institute of Health funded the Battelle Memorial Institute to conduct research on the stability of silicone implants in animals, among them, Silastic∆, which was used in breast implants. In particular, the study concentrated on the polymers' tensile strength and elongation, along with the reaction of the implant site to the polymers. (Tensile strength is a measure of the polymer’s pliancy.) The studies used mongrel dogs as test sites, implanting samples of all five materials in each. The plastics were recovered after six, eleven, and seventeen month intervals. Their tensile strength was recorded before implantation and after each removal, to track any loss, along with any elongation due to implantation. After 17 months, Silastic∆ showed little decrease in tensile strength and slight elongation. At the same time as the introduction of the Cronin implant in 1964, Dow Corning contracted Food and Drug Research Laboratories, an independent research company, to complete more long-term testing on the implants, which had already been followed for two years in the clinical studies and seventeen months in the Battelle study. Results encouraged the continued use of silicone in medical implants, especially in mammary prostheses.

Although implants were first targeted at mastectomy patients, even Cronin and Gerow would have been able to foresee a market for breast enhancement as well. Thus, other manufacturers developed similar implants, in response to a market which grew as women opted for cosmetic breast procedure. However, Dow Corning, where the implant originated, remained the industry leader.

4.7.2 A "New and Improved" Implant?

In 1968, Dow Corning started updating the breast implant, a process that would take approximately seven years. First, they developed a seamless envelope, which provided a smoother finish and a more natural appearance. Now came the tough part. Doctors were requesting a softer, more natural gel formulation, so the breast implant product team went to work. Research scientist Jack Roberts worked on the new gel formulation.

By the first half of 1971, he had found one, and sent it for preliminary toxicity testing. Dow Corning employed a standard, widely used chemical test involving human embryonic cells. The experimental gel was introduced to some embryonic cells, to investigate if any of them changed. Although the cells rarely reacted adversely, they did so when introduced to this gel. However, such a reaction did not necessarily indicate that the gel was unsuitable for medical use. Dow Corning had two options: complete a large, expensive battery of tests to determine if this gel was suitable for implantation, or substitute a similar polymer into the original test.

Dow Corning chose the second option, since Roberts already had a gel similar to the first. Although Roberts' second gel overcame the embryonic cell hurdle, it also had to undergo a penetration test to measure its stiffness. For example, the gelatin we eat is a fairly stiff gel. Roberts and associates allowed a weighted probe to descend for a span of time into the gel. The farther into the gel the probe made it, the softer the gel. After tracking this characteristic in the second attempt gel, Roberts found that this gel was growing softer over time. This tendency was undesirable for an implant gel, since softer consistency could cause more diffusion. Also, if the implant ruptured, more gel would migrate from the area of implantation. Thus, this gel, too, was rejected.

By June of 1974. Roberts had joined another product team, and Kim Anderson had joined the mammary prosthesis team in his place. She was a chemist by training and had been with Dow Corning since February of 1970. Taking over where Roberts left off, Anderson sought to understand the objectives of her mission clearly. As the team members explained to her, they had been trying to develop a more responsive implant gel, one that more closely simulated the behavior of actual breast tissue. Originally, breast implants had been designed for women with mastectomies. Now they were being used increasingly for cosmetic reasons. Dow Corning's competitors were marketing new gels, softer and more responsive, partly in response to this increased need.

An important restriction given to Anderson was to include only ingredients in her gel formulations that had previously been safety-tested, and/or had successful medical implant histories. This meant the new gel would be a conservative design, which should ensure a safer implant for the customer. By January of 1975, Anderson and the Implant Development team (PMG) had created a product with a more responsive gel that passed the embryonic cell reaction and weighted probe tests.

Dow Corning geared up to engineer a major effect on the breast implant market with this new, improved product, creating a special Mammary Task Force to complete the final development of this new product for marketing by June of 1975. However, Dow Corning knew that two more questions had to be answered before the new gel could go to market:

  1. Can we manufacture the product?
  2. Can it suitable for long-term human use?

While Anderson had been working on the new gel formulation, she had simultaneously addressed a manufacturing issue. Presently, the gel was divided into Parts A and B for production, with the ratio of ingredients needed A to B at 100:1. This incongruent ratio would not allow for true simultaneous production in Dow Corning’s batch processing. (Since a batch of A required 100 times the number of ingredients as B, it took much more time.) In order to improve efficiency, DCC asked Kim Anderson to divide the batch ingredients for production more equitably. Her work ended in the masterbatching of the gel, with the ratio of ingredients needed A:B a much more equitable 3:1, allowing for improved efficiency in production. The new gel formula could essentially drop into this new process, and the question of whether Dow Corning could manufacture the new formulation was easily answered in the affirmative.

Anderson and the development team moved on to the second question, concerning the safety of the product. In addition to Anderson, other scientists on the task force included a biocompatability expert along with Anderson's laboratory manager. Chemists and biologists disagreed on the second question. In the chemists’ opinion, no additional testing of the new gel was needed, since it utilized only components which had been previously tested and used in medical implant devices. In contrast, the biologists argued that, since the softer, more responsive gel was made from a novel combination of familiar ingredients, it needed further testing. The biologists recommended to management a two to four week study of the worst case scenario, the insertion of silicone gel without any elastomer envelope. This kind of study simulated a situation where there was total leakage, with silicone flowing freely throughout the body; no human being would ever experience this. The studies would be performed in monkeys, rats and rabbits. In the interest of safety, Dow Corning delayed production until the biologists’ tests had been performed.

In March/April 1975, the results of a Dow Corning two-week study on the effect of silicone gels injected subcutaneously into rats and monkeys were delivered. The current Cronin gel acted as the control gel, and the scientists at Dow Corning tried out three new gels, including the New Production gel, the one they tentatively planned on producing; an experimental High-fluid gel; and a Low-cross linker gel. Specifically, Dow Corning wanted to investigate any tissue reaction, tendency to systemic migration, or differences in general response to the gels among the rats and the monkeys.

One iteration of the study produced an increase in the silicon in the axillary lymph node of the rats, but this result could not be replicated. No "grossly observable" tissue reaction in the monkeys was seen. However, in at least one monkey, gel moved from its original implant site. Another monkey was injected with the same formulation at multiple sites, and the gel migration result was replicated. Although this was an experimental gel and not the one tentatively scheduled for production, Dow Corning was concerned, and narrowed the acceptability criteria to rule out this gel and others with similar formulations.

In addition, Dow Corning also contracted out the requested rabbit study to Biometric, an independent research laboratory in New Jersey. The experiment involved the four gels previously discussed, as well as 28 rabbits, certain of whom were to be sacrificed after 7, 14, 21, and 90 days to reveal what, if any, effects the injected silicone gels had on them. At the end of the first twenty-one days, the only negative effect was a moderately acute inflammatory reacton at the implantation sites. This reaction was less in the twenty-one day rabbits than those sacrificed after seven and fourteen, so the inflammation appeared to decline over time. Overall, the test results had proved positive. Eager to get its implants out, Dow Corning asked Biometric if the test could be shortened by 10 days without incidence, and Biometric answered in the affirmative. After 80 days,

The majority of implant sites were entirely free of any reaction at all. These histopathologic changes observed during the 80 day course of this study were, in our opinion, due to the trauma of implantation and not due to the test gels.

Thus, the extra studies were complete and could be added to the collection of independent and in-house research on silicone already available. Anderson knew that the product team had been working for upwards of four years on this project, and it seemed like the new gel’s time had come.

4.7.3 Enter the ethicist

John Swanson was the one permanent member of Dow Corning’s Business Conduct Committee and therefore played an important role in shaping and maintaining the company’s award-winning ethical policy. His wife Colleen decided to get the new Silastic breast implants in 1974. Almost immediately afterwards, she started to have a variety of symptoms--migraines, lower back pains, rashes and fatigue that plagued her for years. In the late 1990s, her daughter told Colleen about a television program in which another woman with similar symptoms blamed them on Dow Corning implants.

Now Swanson was caught in an ethical dilemma. At work, he listened to his colleagues complain about lawsuits like the Stern case, in which the plaintiff was awarded 1.5 million dollars in damage on the grounds that Silastic implants had caused her auto-immune disorders. They also expressed concern about the Food and Drug Administration’s new role in regulating medical devices. The FDA was asking all breast implant manufacturers to prove their devices safe and effective, or face a moratorium on further sales. At home, Swanson heard Colleen’s accusations that the company he worked for her had hidden the true risk of her implants. She wanted them out--right away.

John Swanson was in a position to see the breast implant problem from both the standpoint of a woman who felt the implants were destroying her body and from the perspective of the company that had created them and saw them as a safe product, given the scientific evidence. Here was the classic hero’s call to an inward journey, in which he would have to exercise moral imagination to decide whether his wife was right, or the company he had worked for--or whether it was possible to reconcile these two opposing views in any way.

For Dow Corning, this was a great opportunity, too. The managers and employees believed they were virtuous. The breast implant was created in response to a surgeon’s need, and the company followed an expanding market. Of course they wanted to make a profit, but they also felt they were providing a service, particularly to those women who had mastectomies.

Therefore, the company felt blind-sided by the controversy surrounding its new breast implants, and were not prepared for the wave of public outrage. Dow Corning’s ethical practices had been the toast of business schools for over a decade. Dow Corning’s Code of Conduct included the following statement:

We are committed to providing products and services that meet the requirements of our customers. We will provide information and support necessary to effectively use our products.

We will continually strive to assure that our products and services are safe, efficacious and accurately represented for their intended uses. We will fully represent the use and characteristics of our raw materials, intermediates and products--including toxicity and other potential hazards--to our employees, suppliers, transporters and customers.

Dow Corning’s policy committed the company to total honesty and integrity, and this policy was reinforced by the Business Conduct Committee, on which John Swanson served. When I first met with Barie Carmichael, Vice-President and Executive Director for Corporate Communications at Dow Corning, to discuss doing a case, she seemed surprised that I saw ethical issues in the scientific testing the company had conducted. If the science said the product was safe, that was the end of the story. But she was coming to realize that, "We were naive. This is a company in the middle of a cornfield in Michigan. We were not publicly traded and didn’t have to answer to public stockholders. And we were naive about politics and did not fully appreciate how Washington, D.C. worked, or how politics could affect the company. Not only that, but most of management saw the implant issue as a scientific one, not one of communications".

But science, as John Burnham and others have shown, is not held in universal esteem by the public and in some quarters, is even seen as a tool used by the powerful to oppress groups that are underrepresented in the scientific community--like women. The case of Lise Meitner and dozens of others show how hard it has been for women to break into what for years has been a primarily male fraternity. Is it any wonder many women felt suspicious about Dow Corning’s scientific claims--especially when other ‘scientific experts’ appeared to disagree?

According to Carmichael, Dow Corning was not surprised that some women were skeptical of the science. What surprised the company was that these individual reports of suffering seemed to carry more weight than scientific studies. Dow Corning did not value emotional ways of knowing as much as scientific ones. They also did not pay much attention to anecdotal reports of problems with the implants, preferring scientific data. Dow Corning was learning the importance of case-based reasoning the hard way. Real cases of human pain are more salient to the public than mountains of scientific evidence.

In 1991, a jury awarded Marian Hopkins $7.3 million in damages against Dow Corning on the grounds that her implants contributed to connective tissue disorder--even though she had been diagnosed before having implants . In April of 1992, the "Food and Drug Administration announced that breast implants filled with silicone gel would be available only through controlled clinical studies and that women who need such implants for breast reconstruction would be assured of access to these studies" (Kessler, 1992, p. 1713). This decision was interpreted by many women as meaning the FDA thought breast implants were not safe, when in fact the agency really intended to keep the product off the market until more testing could be done. The Hopkins suit and the FDA moratorium led to a wave of lawsuits that forced the company into bankruptcy. Colleen Swanson’s was one of them.

Here we go back to the issues we discussed at the beginning of Chapter 2, where we talked about science and truth. A relativist might argue that scientific truths are always the product of social negotiations. A sociologist with a more realist bent would amend this statement to say not solely the product of social negotiations, but negotiations can and should play an important role in what counts as truth--especially when this truth has important policy implications.

The scientists at Dow Corning were realists; they felt that they had the data and were astonished that anyone else could have failed to see it as they did. They did not realize that in order to establish a scientific fact, it is not sufficient to let the data speak for itself; one has to build a network of allies, and as the network grows, the scientific knowledge is transformed . As we will see below, studies of rats and monkeys were gradually supplanted, in terms of importance, by human epidemiological studies.

Irwin & Winn argue that, in general, "science...misunderstands both the public and itself" . Even within science, the process of creating consensus can take years, as in the controversy over continental drift . Indeed, in some areas of research, a single experiment may involve a large network of collaborators and funding sources and take as much as twenty years to complete .

One of the characteristics of scientific development that most plagues historians is the enormous diversity of viewpoints that can continue to persist long after it appears that a consensus has been reached. The difficulty arises not only because consensus is never total, but also because of the fact that consensus always means the consensus of a particular community. Scientists make up many communities, and these communities vary by subject, by methodology, by place, and by degree of influence. Science itself it s polyphonic chorus. The voices in that chorus are never equal, but what one hears as a dominant motif depends very much on where one stands. At times, some motifs appear dominant from any standpoint. But there are always corners from which one can hear minor motifs continuing to sound (Keller, 1983, p. 174).

One of the ‘minor motifs’ was Barbara McClintock’s work on gene transposition in corn, which was not integrated into the dominant consensus, or paradigm, for almost thirty years (see 2.2). The cold fusion case is also instructive (see 3.1). Initially, other scientists claimed to get a cold fusion effect, though the effect itself was a moving target: was it mainly temperature? did it include neutrons and increased tritium levels? did it require heavy water? Eventually, most of the scientific community concluded that there was no validity to claims for the existence of cold fusion, but Pons and Fleischmann continue to work in this area and receive funding from industrial sources . This ‘minor motif’ may lead to a new discovery, but as of this writing, it looks like a consensus over cold fusion has been achieved in the space of a few years.

Similar time-frames and complex networks can be required for epidemiological studies on health issues (Angell, 1996). One example is the controversy over whether silicone breast implants cause cancer, which began in the early 1980s and peaked in 1988. By 1993, animal, epidemiological and clinical studies showed no carcinogenic link and a consensus was achieved. It took about a decade for a consensus to emerge in this area, helped by the fact that the controversy shifted to other conditions, like autoimmune disorders. In their obligation to prove safety, Dow Corning and other implant manufacturers were confronted with a moving target; one problem like cancer would get settled as another like autoimmune disorders arose.

Irwin & Winn argue that the boundary between science and the public is uncertain and often re-negotiated during controversies . Cases like the Swansons’ might have helped Dow Corning anticipate where the controversy was shifting and develop strategies for further testing. According to Keith McKennon, who became CEO of Dow Corning in February of 1992, Swanson raised the possibility of a meeting between McKennon and Colleen. McKennon thought it was a good idea; this was an opportunity to meet directly with someone who was having a problem with silicone breast implants.

On March 19th, McKennon took Dow Corning out of the breast implant business. He also provided $10 million in funding for more research, provided up to $1200 a piece in financial aid for women whose doctors recommended that their implants be removed and tried whenever possible to talk to women . But Colleen remained suspicious of McKennon’s motives and did not want to meet with him.

John Swanson and McKennon eventually had a second meeting, at McKennon‘s request. McKennon had finally heard about Colleen’s lawsuit, filed three months earlier, and he was angry. He would have liked to have talked with her. According to Swanson, McKennon asked what Colleen wanted-- ‘a zillion dollars?’--and offered to discuss a resolution . McKennon says he might have asked whether Colleen really wanted lots of money or whether there was another way to resolve the conflict. Swanson had moved into the legal, adversarial mode, in which you cannot afford to trust the other party, whereas McKennon was still searching for a more cooperative solution.

Colleen did eventually settle her lawsuit with the company, and John retired in 1993 with full benefits, to work as an ethics consultant and criticize his former employer. The conversation that never took place between Colleen Swanson and Keith McKennon might have represented a missed opportunity for Dow Corning to get a better understanding of the reasons why an intelligent woman who had every reason to be loyal to the company thought one of its products was causing her major health problems, and why she did not trust the company’s efforts to deal with the problem.

Richard Hazleton, who took over the job of CEO and Chairman of Dow Corning in June of 1993, had Swanson in mind when he asked,

How do you respond ethically to someone who is questioning your ethics? For instance, the major contributor to a book that has just been published about the breast implant controversy is an ex-employee who actually played an administrative role in the Dow Corning ethics program. He and his wife firmly believe that her current illness is due to her implants. That belief has led him to question the ethics of decisions and actions, some of which he was involved in, and some of which he wasn’t.

On the other hand, there are ethical questions around this individual’s actions. For instance, what were his responsibilities to act on his convictions while he was still employed at Dow Corning rather than two years after he retired? What were his obligations, while still an employee, to investigate his concerns regarding the company’s conduct and find out whether the accusations were true rather than waiting to raise them in a book for which he will be compensated? (.

Did John Swanson do enough to help the company he loved cope with its growing Frankenstein monster? Consider Gioia and the Pinto case, discussed in 4.3 above. Gioia played his role of Ford’s recall coordinator according to the schema and scripts he had learned; he was never able to step out of them and imagine how the problem would have looked to someone outside the organization. In Swanson’s case, the reverse seems to have happened. Understandably, he started looking at the problem from the perspective of his wife Colleen, adopting an outsider’s schema. He would have been well-positioned to try to see the problem from an inside view, as well--to try to understand why the company believed this was a safe and effective product, despite the experiences of women like Colleen. I do not mean to place too much burden on John Swanson, a decent person trying to do his best in a difficult situation. He knew he had a conflict-of-interest on the breast implant problem and tried to walk a tightrope that would allow him to do his job while offering his wife the support she deserved (see Byrne, 1993, for more details).

I think most of the people at Dow Corning were decent, also. Virtue is a necessary but not sufficient condition for moral action. Swanson might have been able to help his company understand why so many reasonable people saw their product as a monster. McKennon and Hazleton, in turn, might have helped Swanson see that, scientifically speaking, there was no monster. What some customers and outsiders saw as a monster other customers and company insiders saw as technological progress. Virtue needs to be complemented by moral imagination, by the ability to switch between perspectives and ultimately transcend them to arrive at another point of view. Could Keith McKennon and Colleen Swanson have worked jointly to create a compensation program that would have addressed the needs of dissatisfied customers without admitting that sicilone implants could cause disease? Dow Corning is moving towards such a policy now (see footnote below).

In the wake of the FDA’s moratorium and the wave of litigation, McKennon had to decide whether to take Dow Corning out of the implant market which was still only a small percentage of the company’s overall business. At first glance, this looks like an ethical no-brainer--until one considers patients with mastectomies who relied on Dow Corning to supply a product for which they felt a genuine need. As one women said after her mastectomy,

You just didn’t feel like you’re all there, you know, you just, it, my clothes didn’t... fit. It was just that, you know, I looked so sunken in, you know you could tell that I didn’t have anything...I came home from the hospital and my husband said, "Let’s see what you look like," and I said "No." I said, "I cannot believe what I look like, what a person looks like, what a woman looks like without breast tissue. And I just could not wait till the day came that I could have something done" .

After getting her silicone implants, another mastectomy patient described her feeling as follows:

So I just wanted to, sort of, to look like I did before I started this, and I’m very happy with the result...And I just think it’s, it’s sort of a self-esteem...thing for a woman, you know, it makes you feel like a whole person...? .

A 1991 study suggested that about 90% of the mastectomy patients were satisfied with their implants and about 95% of the augmentation patients were satisfied with theirs . Even so, in March of 1992, McKennon announced that Dow Corning would leave implant business and would fund a $10 million research program to determine the effects of silicone in the human body. McKennon also announced a fund for women who were unable to afford the removal of implants.

Dow Corning pulled the product, sold the medical devices part of their business and declared Chapter 11 in May of 1995, after an effort to settle the huge number of lawsuits (almost 20,000) failed. Was this just desserts for a company that had produced a faulty product, or a victory of superstition over science?

4.7.4 Junk Science?

John Burnham lamented that by the 1980s, "science probably did not exist any more at the popular level. Superstition did" . Similarly, Marcia Angell, editor of the prestigious New England Journal of Medicine, portrayed this controversy as a victory of superstition over science.

The breast implant controversy shows every sign of continuing on its irrational course for years. Only an unyielding commitment to scientific evidence can stop it, and that does not seem very likely, given the money and passions involved. If all parties had accepted the discipline of evidence at the outset, the controversy would never have reached such proportions. It would hardly have gotten off the ground. But without a commitment to objective data, people were free to believe whatever they liked. Instead of basing their conclusions on the evidence, they willed the evidence to their favored conclusions (Angell, 1996, p. 209).

What does the ‘discipline of evidence’ show? The only way to determine whether silicone implants pose a health risk in humans is to conduct epidemiological studies. In such a study, one compares a sample of women with implants with a sample who have none, to see if a significantly higher proportion of the former have a disease or complication than the latter. Several major studies of this sort were conducted recently, focusing on what is called ‘connective tissue disorder’--a broad category of auto-immune diseases like rheumatoid arthritis and lupus that can cause the kinds of symptoms experienced by Colleen Swanson.

In 1994 the Mayo clinic reported that 749 women with implants did not show a greater rate of connective tissue disorder than a similar group without implants. Similarly, in 1995, a Harvard Medical School questionnaire study of 1183 nurses with implants, including 876 with silicone implants, again found no association between implants and connective tissue disorder. The most recent and comprehensive study, conducted again by the Harvard Medical School, involved sending questionnaires to 395,543 women involved in the health professions, of whom 10,630 reported having implants. In this case, there was a small but significantly greater chance that women with implants would report connective tissue disease, at a rate of about one extra case of disease per year for every 3000 women. But even this very slightly higher rate could be explained by a tendency for women with implants to overreport symptoms because of the controversy surrounding implants--or, conversely, for women with implants to refuse to participate in the study because their attorney’s advised them not to. The only way to be certain is to check medical records. Whenever this has been done, no relationship has been found between silicone implants and diseases (Angell, 1996).

Earlier in this book, I referred to 'willing the evidence to fit one's favorite conclusions' as confirmation bias. Angell saw most of the signs of confirmation bias in the plaintiffs and their attorneys, fueled by the prospect of multi-million dollar awards.

But there was a kind of confirmation bias on the part of Dow Corning. Not a bias in terms of scientific evidence, but in terms of moral imagination. The company took the scientists’ view: if the data showed there were no problems, then everyone would agree. For the women like Colleen Swanson suing the corporation, the ‘data’ was their pain. Many of the women discounted the science altogether; others, looking at the fact that Dow Corning had funded many of the studies, thought the science had been bought from a Kuhnian standpoint, this might be evidence that the two sides are holding incommensurable views, meaning that they literally cannot understand one another. On the one hand, you have scientists like Marcia Angell and Kim Anderson, who think of themselves as seekers of truth. On the other hand, you have women like Sybil Goldrich who believe that the Kim Andersons and Marcia Angells have been bought.

Dow Corning was also surprised by the FDA’s finding that they had not proved the product safe. From a Popperian standpoint, the FDA presented Dow Corning with an impossible dilemma. In the same way you can never prove a theory true, you can never absolutely prove anything will be safe--tomorrow, you might discover an unanticipated interaction with other drugs, or chemicals, or a new medical condition. (Popper here borrows from Hume’s classic critique of induction). A company following Popper’s dictum would only be able to say that, given everything we know at this time, there is no evidence that this product is unsafe, but that we will continue to test for any possible future problems.

David Kessler, the head of the FDA at this time, defended this decision by noting that the Food, Drug and Cosmetic Act of 1976 required "a positive demonstration of safety--and the burden of proof rests squarely with the manufacturer" (Kessler, 1992, p. 1713). ‘Positive demonstration’ is vaguer than proof, and may allow for more ‘wiggle room’. Kessler goes on to deal with the classic Popperian/Humean criticism: "It is never possible to predict with certainty how a device will function 10, 20 or 30 years after its implantation; however, even basic characteristics that have some value in predicting future performance, such as tensile strength and fatigue resistance tested through cyclic loading, are missing in this case .

Popper brings us back to the crux of Dow Corning’s dilemma. They felt that they had done state-of-the-art testing and found no serious problems. Furthermore, given the long track record of silicone, they had reason to believe no major problems would emerge in the future. Kessler and the FDA felt the company had not done adequate testing to support this view.

One of the problems with research into implants is that new potential hazards kept being hypothesized. First there was a concern with cancer. After that was dismissed, scleroderma became a concern. Then it was connective tissue disorder. Epidemiological studies take several years, although one can sometimes ‘piggyback’ onto an existing study, searching for evidence of known diseases in the medical records and correlating them with implants. But such records do not contain information on new disorders like the special silicone syndrome hypothesized by Mark Lappe (1993) and Nir Kossovsky (1993), who developed their own theory--that silicone, an inert substance, can get coated with proteins in the body and then denature these proteins, changing their shape. The body’s immune system would then attack these proteins, creating an autoimmune disorder.

Kossovsky tested his theory by running a standard, but very difficult, test for antibodies (Taubes1995). In March of 1992, the Autoimmune Disease Center at Scripps Research Institute sent him blood he could test, from:

  1. 40 women who had auto-immune disease, but no implants;
  2. 10 healthy women with no implants;
  3. 10 women who had auto-immune disease and implants.

The blood samples were sent blind, meaning Kossovsky didn’t know which were which--he had to send his results back to Scripps, then they would tell him whether his test distinguished between women who had implants and autoimmune disease and woman who had no implants and autoimmune disease. If his theory were right, the test should show more or different antibodies among the 10 women who had implants, presumably because the immune system was attacking the proteins that were denatured by the silicone.

There were no differences between the no-implant and implant blood samples. Kossovsky’s test did distinguish between women who had autoimmune disorders and those who did not, but this had nothing to do with his theory. The initial result, therefore, was a disconfirmation of Kossovsky’s hypothesis.

He rejected the result. Instead, he took the Scripps data on 40 women with his own data on 249 women with implants and 47 healthy women who did not. He reported that 9 of the 249 women with implants had scored higher on his test than any other women in the sample. These 9 women showed no real distinguishing symptoms, but Kossovsky concluded that they were suffering from some kind of silicone-related disease, supporting his theory. But even if he did the very difficult test properly, a result of this sort could be due to the difference in sampling size between the groups--he was comparing a group of 249 to groups of 47 and 40, and so the 9 women in the larger group were equivalent to about 2 in the smaller ones. If he had taken a larger sample of women without implants, he might have found a few had scores on his test as high as those in the top 9 of the implant group (see 2.3.8.1). Combine this ambiguous, positive result with the negative Scripps result and the best one can conclude is that more research is needed, preferably by an independent lab.

Instead of continuing research, Kossovsky took a cold fusion turn and applied for a patent for his new blood test. He promptly marketed it as well, targeting trial lawyers who were looking for clients with breast implants or other silicone devices. Kossovsky also appeared in many trials as an expert witness. When lawyers for the defense asked to see his laboratory notebooks, he claimed they were lost in an earthquake. When results of epidemiological studies showed no link between silicone breast implants and autoimmune diseases, Kossovsky countered that the studies were not looking for the right kind of diseases--the breast implants might be producing some new kind of illness.

This example illustrates how confirmation bias can be maintained by a moving target: when studies disconfirm one relationship, simply propose another. Marcia Angell also points out that

Kossovsky’s observations...focus on one link in a long chain of postulated events. But before focusing on one link in a chain of possible causation, scientists usually first try to establish a connection between one end of the chain and another--that is, between the suspected cause and the disease. For example, first we found out that cigarette smoking is associated with lung cancer. Only then did scientists turn their attention to how cigarettes might cause the disease...In the breast implant controversy, there has been a tendency to do it backwards. Assuming there is a connection, some people have sought to explain how it works (Angell, 1996, p. 108).

Angell is defending an inductive model of science, here, in which interesting findings lead to hypotheses. In fact, the reverse quite often happens--interesting hypotheses lead to novel findings. There is nothing wrong with proposing a causal mechanism to explain a possible phenomenon--it could lead to focused research that would demonstrate the phenomenon. But Kossovsky’s results, like the early cold fusion claims, have not been replicated.

Kossovsky, Lappe and others were paid substantial sums to serve as expert witnesses in trials, testifying that silicone implants cause medical problems. The old discovery/justification distinction in philosophy of science would suggest that this does not matter--a scientist’s motives for introducing an idea have nothing to do with the truth value of the claims, which need to be independently tested. The important point is independently. The fact that Kossovsky confirmed his own theory could be dismissed as a bias motivated in part by financial incentives--not that he is lying, any more than Pons and Fleischmann were, just that he is human and more likely to see positive evidence in an ambiguous setting when there is a tangible reward.

Similarly, critics of much of the scientific research which shows no association between silicone implants and diseases have faulted the studies because many were funded by Dow Corning. The important question here is whether Dow Corning could buy the results it wanted, or whether the studies were genuinely independent. Marcia Angell defends the epidemiological studies on the grounds that they were published in refereed journals; the referees were not funded by Dow Corning. Furthermore, the studies are sufficiently detailed so the data can be inspected and criticized. Still, those scientists who cannot obtain funding may find it hard to participate in the debate--they may not be able to conduct studies of sufficient quality. Funding sources can create a kind of confirmation bias by leaving some voices out in the cold.-

Federal District Judge Robert E. Jones, overseeing breast implant cases in Oregon, appointed an independent panel of scientific experts to assess the evidence that silicone breast implants cause disease. Based on the panel’s findings, in December of 1996, Judge Jones ruled that lawyers cannot introduce evidence that implants cause disease since such evidence is not scientifically valid .

In effect, Jones was declaring that Kossovsky and other similar expert witnesses represented ‘junk science’. If this ruling is upheld, it will encourage other judges to appoint independent panels of scientific experts, rather than relying on the defense and the prosecution to find competing experts and see which side convinces the jury. Judge Samuel Pointer in Birmingham, Alabama, has appointed a national panel of scientific experts to answer two questions:

  1. Does existing scientific research indicate that breast implants filled with silicone can cause or exacerbate chronic conditions like autoimmune diseases?
  2. To what extent would disagreement with the conclusion to the first question "represent a legitimate and responsible debate within the field?"

Thousands of cases have been consolidated under Judge Pointer’s jurisdiction, so this panel could have a great influence, though the Judge will decide how much weight to give to their conclusions. His second question suggests that he is trying to determine if scientific consensus is emerging on this issue. As of this writing, the American Cancer Society, the American Medical Association, the American Society of Plastic and Reconstructive Surgeons, the American College of Rheumatology, the British Council on Medical Devices and the Food and Drug Administration "have concluded that there is no evidence that silicone breast implants" cause any kind of autoimmune disorder .

A recent article in Science argues that, "the experts need to have appropriate credentials, including knowledge, neutrality, and diligence. The National Institutes of Health (NIH) now maintain a roster of potential scientific reviewers who are checked for conflicts of interest. The NIH, universities, the American Association for the Advancement of Science (AAAS), the National Academy of Sciences (NAS), and other neutral organizations are existing resources to provide scientific guidance in the classroom. Scientific bodies should not wait for the court to seek advice: as scientists we should ensure that every court has at its disposal a listing of neutral experts with specified areas of expertise and acknowledgment of potential conflicts" .

The independent-panel-of-experts approach assumes that jurors and judges are not competent to decide what constitutes good and bad science. A complementary approach would be to make sure the public was better educated concerning the actual process of scientific investigation. This case shows how hard it is to understand the science, and how easy it is to manipulate to make it appear to confirm one’s perspective. For example, lawyers frequently would isolate single findings and results, whereas scientists know that one must look at a larger pattern of evidence--hence, their reliance on multiple epidemiological studies.

But this kind of research takes time. For example, a recent study in the Journal of the American Medical Association (JAMA) looked at the issue of what variables are confounded with silicone implants, and whether any of these variables were risk factors for connective tissue disease. Even if there were an association between implants and connective tissue disease, this would not necessarily imply a causal relationship. Another factor associated with both might account for the apparent relationship.

The JAMA study compared the characteristics of 80 women who had breast implants with 3520 women who had not . According to an American Medical Association press-release,

When compared to other women, the researchers found that women with breast implants were:

Of these factors, hair dyes have been associated with an increased risk of connective tissue disease. Therefore, any association between silicone implants and connective tissue disease might be explained by a third factor--the fact that women who undergo breast augmentation are also more likely to dye their hair.

More research will be needed to determine if there are other relationships of this sort. It is even possible that women with implants are more likely to engage in some behavior that reduces the probability of connective tissue disease, thereby masking any problems caused by the implants themselves. It is impossible to absolutely prove, beyond the shadow of a doubt, that no complex relationship between implants and some disease will ever be found. That is the nature of science.

How does this translate into advice for a company that wants to avoid what happened to Dow Corning? Based on this case, I would suggest the following:

1) Get potential stake-holders involved in the design process. Dow Corning felt it had done this--surgeons were intimately involved in the creation of the product, and also in the many modifications Dow Corning made to improve it. But there were other stake-holders, including the women who carried the implants and the FDA. For most drug manufacturers, interaction with the FDA is built-in. It wasn’t for Dow Corning, which was an insulated, privately-held company that had little interaction with government agencies or the broader public before this nightmare began.

2) Pay attention to cases. When women reported symptoms, even though the science showed no problems, the company should have been all over it, making a concerted public effort to find out what was going on, sponsoring more research, etc.--all without creating any impression that the company believed the product was flawed. Similarly, in the Pinto case, Gioia admitted that he should have paid more attention to isolated reports of fiery explosions involving the car On April 4, 1993, a six-year old girl in Ohio was mortally injured by an air bag which expanded when her mother hit another car. This incident was the first of a series of similar cases that signaled air bags could injure small children who were not wearing seat belts . But it was another 3 years and sixteen more deaths before warning labels were issued, and a debate over the cost and benefits of air bags still rages.

Anecdotal evidence can be an important warning sign. In the case of silicone breast implants, it is not clear that the warning signaled a real problem. A company needs to take anecdotal reports seriously without privileging them over scientific data. At the very least, consumers need to be informed of all possible risks. But how to do this without causing unnecessary panic, especially when the scientific evidence does not support the anecdotal concerns?

It would be refreshing if companies and regulators could be totally honest in a situation like this, saying that while the scientific evidence suggests no problems with silicone breast implants, some consumers do complain of symptoms like the following, and include a list and where to get more information. In our current adversarial system, this sounds like a ridiculously naive strategy, almost an invitation for lawyers to sue. But what if we adopted a different model, one in which stakeholders worked together to create a safer product? We will have more to say about this in the next section.

3) Document what you do as a company to anticipate and avoid future negative impacts. Barie Carmichael lamented the ‘killer memos’ that caused Dow Corning trouble in court. One example is a memo written by Chuck Leach to Bob Levier, head of biological testing at Dow Corning. Leach was concerned about research on the problem of capsular contracture. He noted that Dow Corning’s competitors were studying these problems, and noted that Dow’s customers among the plastic surgeons were asking whether Dow Corning was conducting similar research. Leach said, "I assured them, with crossed fingers, that Dow Corning too had an active contracture/gel migration study underway."

This memo was taken by the Associated Press and others as evidence that Dow Corning was lying to its customers, and it was used in court to discredit the company. Leach objected that he crossed his fingers as a sign of hope--he did not lie. He managed to get the AP to issue a correction, but the correction was published only in the Midland Daily News. .

Sometimes a ‘killer memo’ has to be written to draw attention to a problem--but the resolution of the problem contained in the killer memo should also be recorded. In this case, Leach pointed out that he knew Dow Corning had already done extensive safety research, but he was not sure whether the company had research under way to test a new hypothesis: that the well-known problem of gel-bleed might exacerbate the well-know problem of capsular contracture . In fact, Leach’s hopes were well-founded: research was under way on this topic. But the resolution to the incident was not remembered; only the ‘killer memo’.

4) Have a coherent framework for deciding whether new products will be harmful or beneficial, one that is easy to explain and justify to a broad spectrum of stake-holders. Dow Corning had a general code of conduct that included a commitment to "operate our facilities in a manner that meets or exceeds all applicable regulatory requirements. We will also plant and strive for continuous improvement in process efficiency, waste generation, and emissions to the environment.."

But these commendable goals were not accompanied by a philosophy like McDonough’s principles or the Natural Step that provided a mental model for how to accomplish them.

Consider the system of product classification developed by the environmental chemist Michael Braungart, who is the head of the Environmental Protection Encouragement Agency in Europe and has formed a partnership with William McDonough called McDonough Braungart Design Chemistry (http://www.mbdc.com/). Braungart classifies products into three categories:

  1. Consumables: These are items like food and detergent that need to be free of toxins and any chemicals that will not biodegrade.
  2. Products of Service: These include products like automobiles and televisions that cannot be biodegradable. According to Braungart, these sorts of materials would have to be leased from the manufacturer, who would take them back after use and recycle all the parts and materials into new products. This kind of recycling is not the same as what McDonough calls ‘downcycling", in which a complex technical product like a computer is melted down to make plastic parts for automobiles. In Braungart’s system, all the special metals and materials in the computer would have to be used again to make a new computer.
  3. Unmarketables: These include things that should never be produced, like toxins that cannot be degraded into something harmless when recycled. DDT and radioactive waste might be examples. Those unmarketables that have to be nmanufactured while alternatives are being sought should have molecular tags that identify the manufacturer (Hawken, 1997).

Consider where silicone breast implants might fit in Braungart’s scheme. They are not consumables. Firstly, there are groups that regard it as toxic. As this section has shown, these claims are dubious, but research continues. The point is, when has one done enough research to show that a product is not toxic? For any system of classification, one can only work from the best knowledge available at a particular time. Silicone one of the most bio-compatible matrials, though it certainly could act as an irritant when it leaked.

Secondly, silicone breast implants should not be biodegradable, unless it could be guaranteed that they would not degrade over several lifetimes. The fact that silicone is relatively inert is an advantage from the medical standpoint, but not from the standpoint of converting wastes into food.

Could these implants become products of service which could be recycled by the manufacturer? The silicone in the implants would last well beyond the life of their owner. There are all sorts of ethical issues about how you would get the implants back after use. One could ask people to voluntarily donate the silicone in their bodies, in a manner similar to organ donations. One might even be able to offer some kind of discount for a willingness to recycle, although this raises all kinds of ethical issues about the wealthy and the poor. Should rich people be the only ones whose bodies remain intact after death? Should Medicare cover the cost of recycling? Even if some kind of voluntary recycling program could be implemented, not everyone would agree to it.

Would silicone breast implants have to be classified as unmarketables, on recycling grounds? Silicone is used in a wide range of implants: it "is a component of artificial joints and heart valves, shunts and other tubings, disposable needles and syringes, and contraceptive implants (Norplant), as well as testicular and penile implants. Indeed, probably no American is without some silicone in his or her body, put there by some type of routine medical care--such as injections with silicone-lubricated needles and syringes" .

Is there any alternative to silicone that would not have the same recycling problems? Implants made from triglyceride, a vegetable oil, are being offered as an alternative to silicone; this substance is used in intravenous injections; it can be metabolized by the body, but is also resistant to bacteria and fungi. Furthermore, Lipomatrix, the company that created it, is also including a code number on each implant on a miniiature microchip, which will facilitate an audit trail that can incorporate disposal. Therefore, implants made from triglyceride appear to follow the ‘waste into food’ analogy.

But it is not clear that soy, or any alternative, could cover the wide range of additional medical uses for silicone. One cannot have any medical device that would biodegrade in the body, nor can one recover such devices without patient consent.

Similarly, The Natural Step’s second principle holds substances produced by society should not systematically increase in nature. Again, silicone-based medical products--and all other silicone products--are likely to increase over time as human beings continue to use them.

It is not clear that either of these frameworks will work in the medical device area, in their present form. But remember that their emphasis is on continuous improvement--on adding environmental criteria to design and working towards them. Perhaps a way could be found to make medical devices that would take a long time to biodegrade--perhaps a thousand years--but would still eventually turn into organic materials that could be absorbed by the biosphere. Perhaps there should be some incentives for recycling medical devices, or at least a donation program similar to the ones for organs. Perhaps medical devices need their own classification system.

Having such a system or framework both forces a company to reflect and also provides a defense against future litigation and regulation. It shows that a company is not merely reacting to regulation, it is proactive--it has a long-range vision that promotes safety and sustainability. It is to be hoped that companies in the medical implant business will work with people like McDonough and Braungart on the development of such a system.

4.7.5 Can the legal system act as guardian against pollution?

Did the legal system protected the public from dangerous substances in the case of breast implants? Evidence suggests that Dow Corning underestimated the risk of rupture and silicone leakage and that there may be a very small additional risk of connective tissue disorder. Was that worth a multi-billion dollar lawsuit, bankruptcy and taking the product off the market? If this is the model for how medical devices ought to be handled, it is hard to imagine many companies being willing to pursue this market.

In the Dow Corning case, the legal system severely punished a company for polluting the human body in the absence of any real scientific evidence. The recent book A Civil Action describes the opposite: a case where a judge decided that groundwater from a site where toxic chemicals had been dumped could not have reached an aquifer . An appeals court upheld this conviction despite the fact that by the time of the appeal, the EPA and the USGS had established that the waste site had contaminated the wells drawing water from the aquifer. Here again, the court ruled against the science.

Carolyn Merchant, in her discussion of the partnership model of relations with nature, cites examples of companies, environmentalists and other stakeholders that form partnerships. The network formed to create Climatex Lifecycle is an example of such a partnership; it included the EPEA, an agency that functioned like a combination of a regulator and a consultant; Rohner Textil hired the EPEA to tell it what standards to meet and help it figure out how to meet them.

Could there have been a partnership between Dow Corning and its unhappy customers, a sense of working together towards a just solution that would keep a product many women wanted on the market, while helping those who wanted to get it out of their bodies? The legal model demands an adversarial relationship in court. Any effort to reach out to the other side can be interpreted as an admission of guilt. Of course, part of the goal of the adversarial court model is to promote settlement outside of court. But even these settlements can be motivated more by a need to eliminate the expense and grief of gambling on a trial than by a genuine desire to cooperate with another party in a just agreement.

One alternative is an independent panels of experts, like the one Federal District Judge Robert E. Jones used in December of 1996. These panels are supposed to be neutral, unlike the experts hired by attorneys to appear in court. But a sociologist of science who takes a radical perspective on Kuhn might argue that such panels could never be unbiased: the would probably be composed of normal scientists who would support the current paradigm and would not even understand a radically different view (Pinch, 1997). Such panels of experts are really a statement about our lack of confidence in the ability of the judicial system to understand science. Certainly the record on patent controversies has been poor, with frequent flip-flops in decisions about who invented what depending on which court was considering the appeal (Lewis, 1991:Hanson, 1982).

Under ideal circumstances, scientific experts presenting evidence would not be committed to either party in a litigation and juries would be so well-educated in matters scientific that they could reach an intelligent verdict. As Jasanoff argues, participants in the judicial system need a better understanding of science and scientists and engineers need a better understanding of the judicial system (Jasanoff, 1995). Public understanding of science is essential in a democracy, and reinforces the need for improved education in the processes by which invention and discovery really occur (see chapter 5).

Dow Corning is attempting to work out some kind of consensual agreement with all stakeholders that will enable it to emerge out of Chapter 11. As Gary Anderson, Dow Corning’s President, said, "We are committed to continuing negotiations and hopefully achieving a resolution acceptable to all parties. We are also committed to compensating legitimate claims in our case based on implant rupture or problems caused by our products. But we continue to believe that sound scientific evidence should be the basis for resolving those claims."

A spur to this sort of agreement is a ruling by Judge Arthur J. Spector, of the U.S. Bankruptcy Court, Eastern District of Michigan, calling for common issue causation trials to resolve whether scientific evidence supports the claim that silicone breast implants cause disease. If this question is put to experts selected by an independent panel of scientists, the answer would clearly be no.

What if one had a moral framework and protocols that would protect one from almost any kind of possible future litigation, and that rendered regulation unnecessary? McDonough actually got started on his path to such a framework by the threat of litigation. He agreed to design a building for the Environmental Defense Fund, then found out they were going to sue him for any health hazards associated with the air in the offices. He carefully researched all the possible hazardous chemicals and materials, and made him realize how many ordinary products released gases he considered toxic.

McDonough’s is not, of course, the only framework one could use . Both Braungart’s classification system and the Natural Step are complementary frameworks that could potentially be applied to a wide range of situations, though as we saw in the last section, they may need modification to work for medical devices. Any such framework should establish standards that guarantee, given everything that could have been known at a particular time, any product would exceed current standards for safety and environmental intelligence. It would put the company on record as having the goal of restoring the environment, demonstrating a real concern with long-term consequences. The commitment would have to be backed up with policies that ensured employees owned the companies values. Potentially, such a system could provide at least some protection against future legal action. To see how such frameworks could be used to produce and evaluate actual products, we need to consider a series of cases.

This page was last edited: Wednesday, July 14, 1999