Twenty-five women with breast implants participated in semistructured interviews designed to reveal their "mental models" of the processes potentially causing local (ie, nonsystemic) problems. The authors analyzed their responses in terms of an "expert model," circumscribing scientifically relevant information. Most of the women interviewed had something to say about most elements in the expert model. Nonetheless, gaps in their mental models undermined decision making about their implants. One woman misunderstood the terms used by the medical community to describe implant failure (eg, rupture, leak, and bleed). Another exaggerated the implants' vulnerability to direct impacts, such as car accidents. Participants also overestimated their ability to detect localized problems and to select medical remedies. Although they were generally satisfied with their own implants, many participants were dissatisfied with the decision-making processes that lead to their choice. Their interviews are interpreted by the form and content of communications that women with implants need to help them manage their health decisions better.
Index Terms: breast implants, decision making, informed consent, mental models, risk communication
The silicone breast implant issue has been one of the most hotly debated and closely watched topics in the history of women's health. The controversy surrounding the suggested association between silicone breast implants and connective tissue disease was punctuated by the US Food and Drug Administration (FDA) decision to restrict the general use of silicone breast implants in 1992. The debate continues to be played out in both scientific and legal arenas as researchers publish epidemiologic studies and courts hear implant-related cases.
As the debate over systemic connective tissue disease continues, local complications of implants, such as rupture and capsular contracture, have received much less public notice. Indeed, a recent, highly anticipated report from the Institute of Medicine concluded that the primary safety concern associated with breast implants is not systemic disease but local complications, resulting in pain, disfigurement, and other medical problems. For women with implants, these are potentially significant health problems, capable of disrupting the lives of some women and causing concern for others. Although the physiology of these problems has been studied extensively, little is known about how women with implants think about them, or what kind of risk-related information would be most helpful to them.
As a first step toward providing women with relevant information, we conducted the study reported in this article. We used mental model methodology to characterize women's beliefs about local complications of breast implants, including their views on the risks and benefits of implants and the management of possible risks. The mental model method uses semistructured, open-ended interviews that focus respondents' attention on scientifically relevant issues and still allow them to express their beliefs freely in their own words. Understanding women's current perspectives in depth is essential to creating communications that address their concerns in terms that make sense to them. It is also a precondition for devising structured (eg, multiple-choice, short-answer) tests of knowledge that could be administered more efficiently than in-depth interviews. Without initially casting so open a net, researchers cannot know the full range of lay persons' misconceptions and concerns or even the terms in which accurate beliefs are intuitively formulated.
The mental model approach attempts to be faithful simultaneously to scientific and individual realities, both of which are essential for effective communication. It has been used to design communications for a variety of topics in which laypeople must understand technical information to make effective decisions, whether as patients, consumers, workers, or citizens. These include infectious disease, radon in homes, electromagnetic fields, paint-stripper exposure, mammography, Cryptosporidium, climate change, and nuclear energy sources on spacecraft.[2-7] The process is conceptually simple, although each application poses its own challenges.
The first step in the method is to summarize the relevant scientific literature to determine what is worth knowing. This summary is called an "expert model." Our standard representation for this knowledge is in the form of an influence diagram, a form of directed graph suitable for considering information from multiple sources.[8-10] In it, each node represents a variable. An arrow connects two nodes if estimates of the variable at one node depend on estimates of the variable at the other. Sometimes, a cluster of variables on a common topic is combined in a node to simplify the representation and facilitate expert review of its contents. To sharpen our thinking, we attempt to make these models computationally tractable, even if we lack credible quantitative estimates of many of the variables and relationships. (See Figure 1, which shows an influence diagram for local complications of breast implants).
The second step is to develop a mental model interview protocol, based on the expert model. It begins with very open-ended questions (eg, "Tell me whatever you know or have heard or have thought about breast implants"). The intent is to encourage respondents to reveal, in their own natural mode of expression, whatever is on their minds, as well as to establish a respectful, cooperative tone. We want to know what respondents feel and think, with the eventual aim of helping people similar to themselves understand the issues better. Respondents are asked to expand on everything that they say while the interviewer keeps track of the topics that have been addressed by using the influence diagram.
When this questioning strategy stops producing additional information, respondents are asked whether they have anything to say about each general topic in the expert model, although we recognize that there is an increased risk of their devising an answer simply to have something to say. The interview concludes with focused questions on specific topics of particular interest to the investigators. These questions resemble the format of conventional structured surveys or tests. However, responses are still open-ended, providing an opportunity to see how the questions are interpreted and how respondents explain their answers.
Interviews are transcribed, then coded in terms of the nodes of the influence diagram. Additional nodes are created when the interviewer encounters nonexpert concepts. These can be either erroneous or correct beliefs that are not present in the expert model. Initial data analyses examine which topics are most and which are least elaborated in respondents' thinking. Subsequent analyses ask respondents what they have to say about each topic that they address.
The contrast between lay and expert beliefs provides the basis for deriving focused communications. This process allows for the possibility that lay people know things that the experts do not know or may define the problem differently from the way the community of experts do. This option is particularly important for emerging health problems and problems with far-ranging implications--situations in which the community of experts may have too little knowledge or a perspective that is too narrow. Thus, the terms expert and lay reflect the source of the beliefs, without prejudging their validity.
Creating the Expert Model
Before the development of breast implants, women could receive silicone injections directly into their breast tissue to increase breast size.[11,12] The first breast implants were introduced in the 1960s. Current estimates suggest that about 1 million women in the United States have had breast implants. The implants are used primarily for cosmetic augmentation; however, breast implants are also used for reconstruction following breast cancer surgeries, traumatic chest injuries, or congenital abnormalities. Approximately one third of women with silicone implants received them for reconstructive purposes.[1,13,14]
Before the FDA placed a moratorium on the use of silicone breast implants in 1992, the majority of implants were made of silicone gel inside a silicone elastomer envelope. Silicone implants are now available for reconstruction purposes only to participants in clinical trials. Although saline-filled implants were available concurrently with silicone-filled implants, they were not as popular as silicone because of their propensity to deflate, as well as their less-natural feel and look. Although saline-filled implants have been the sole choice for women seeking cosmetic augmentation since 1992, the number of implant surgeries has reportedly increased threefold since then. Saline implant manufacturers are now beginning an aggressive marketing push with a new advertising campaign. This increase in saline implant use is likely to produce a parallel increase in concern over local complications because the risks of developing complications are similar to those with silicone implants.
Most of the elastomer envelopes covering the silicone gel had a smooth surface texture until the 1980s. However, design changes intended to decrease tissue hardening led to textured elastomer and polyurethane foam covers. Although highly successful in decreasing hardening around the implants, the polyurethane-covered implants were taken off the market when studies showed that a byproduct created by the degradation of the shell was carcinogenic in animals.[16,17] Breast implants are placed either below the pectoral muscle or above the muscle but below the mammary gland. Placement depends on the amount of the patient's available breast tissue and on the surgeon's preference.
The three main types of local complications are capsular contracture, ruptured implants (and the associated local consequences), and complications of mammography screening. We summarize each briefly and then present an "influence diagram" summarizing their precursors and consequences.
The human body naturally forms a fibrous capsule around the implant, as it would for any other object that is too large to be phagocytized and too inert to produce a toxic reaction. This capsule keeps the implant in position. However, in some women, it tightens and hardens, resulting in what is commonly called capsular contracture. Gabriel et al conducted one of the largest population-based studies of local complications in a cohort of Minnesota women with implants. Their data showed that 17.5% of women who received implants between 1964 and 1991 developed capsular contracture severe enough to require a surgical procedure (at a mean time of 7.8 years after the initial implantation). Many possible etiologies of capsular contracture have been studied, including filler material,[19-26] placement of the implant,[27-34] surface texture,[35-40] and bacterial infection.[28,41-45] Women who have undergone more complicated procedures, such as reconstruction, tend to have higher rates of contracture.
Implant failures are characterized by their degree of severity: silicone bleed, leak, and frank rupture. A rupture can be intracapsular, when the fibrous tissue surrounding the implant holds the escaped silicone in place; or extracapsular, when silicone migrates from within the fibrous capsule. An intracapsular rupture is "silent" when there is no indication, either symptomatically or aesthetically, that the implant is no longer intact. Even an extracapsular rupture may be silent. The clinical symptoms most often reported in association with implant rupture include palpable masses (granulomas), pain, and change in the shape or size of the breast. To the best of our knowledge, no studies address long-term outcomes for women who have asymptomatic ruptures, nor is it known with any accuracy what percentage of women with rupture will ever have various symptoms. Furthermore, no studies address the likelihood of an intra-capsular rupture evolving into an extracapsular rupture, which is more likely to be clinically significant.
The case reports and epidemiologic literature describe several possible causes of implant failure. One is closed capsulotomy, a painful procedure that is now mostly abandoned, in which the surgeon manually breaks the fibrous capsule surrounding an implant. Other possible causes are trauma or injury to the breast, compression from mammography, accidents during implantation or explantation, manufacturing defects, and normal wear. Two other factors are the age and design of the implant (or its "generation"). Implants produced between the early 1970s and the early-to-mid-1980s had thin outer shells. This design responded to complaints that the original thick-shelled implants did not look or feel very natural. Researchers have found a higher rate of ruptured implants from this second generation.[46,47] Findings from larger studies, mainly with second-generation implants, predicted that they would lose their integrity between 8 and 14 years after implantation.[48,49]
Breast Cancer Screening and Breast Implants
Mammography requires even compression of breast tissue between the x-ray plate and a paddle. However, the amount of breast tissue that can be compressed is reduced in an augmented breast. As a result, government guidelines recommend additional screening views for women with implants, including a technique that pushes the implant out of the field of view to increase the amount of tissue available for imaging. Nonetheless, current epidemiologic evidence does not suggest that those women with implants who develop cancer present with more advanced disease.[54-59] The effectiveness of mammography as a screening tool for breast cancer in all women depends on such factors as the quality, type, and processing of film; the radiographer's skill; and the use of additional modalities, such as ultrasound. Mammography's effectiveness with augmented breasts depends on the type of implant and its positioning, the number of views taken, as well as the implant's type and positioning, among other things (eg, degree of contracture).
The influence diagram summarizing these processes is shown in Figure 1. Although complicated (as befits the problem), it can be read in a straightforward fashion. Going from left to right, one proceeds from the factors increasing the risk of problems, to the problems themselves, then to their potential treatment and consequences. Most relationships in the diagram are causal (eg, trauma compromises the integrity of the implant; doctors' recommendations affect perceived risk). However, in some cases, the variable (eg, age) is just a predictor of the associated variable. Nodes are connected when the scientific literature indicates some chance of a relationship, even if the current best guess is that there is probably no relationship at all. Subsequent research could erase some of these links and potentially create new ones in which no relationship is currently suspected.
Development of the protocol followed the general philosophy described above. We created open-ended introductory questions, follow-up prompts, and questions directing participants toward specific topics in the expert model. In this process, we were able to draw on an ongoing study of women's mental models of mammography and breast cancer that addresses some related issues.[3,60,61] We took particular care to establish a respectful tone and to reassure participants that the goal of our research was to serve the needs of women like themselves on topics outside of the ongoing litigation. We tested the protocol on 2 personal acquaintances with implants before using it with 25 strangers. The final interview protocol is available from the authors. Its administration by telephone took 1 hour on average (range: 45-90 minutes). We anticipated that participants would have questions about current research on implants; so at the end of the interview, we told the participant how to obtain FDA's comprehensive guide to implants. Promising this resource during the interview made it easier to avoid responding to participants' (more and less direct) requests for information. The interviewers made it clear that we were not physicians and were not familiar with the particulars of each case. It is our strong impression that participants were comfortable with this relationship, as shown by the length and frankness of the interviews. Each was conducted by one of two female interviewers. One was a clinical psychologist; the other was a human relations specialist.
We recruited participants through advertisements in local newspapers with some national circulation (The Washington Post and The Toronto Globe and Mail). The ads said:
Women of all ages who currently have or formerly had silicone breast
implants are asked to participate in a phone interview to gather
information and personal experiences and opinions about their implants.
Both cosmetic augmentation and reconstruction patients are needed.
Confidential interview results will assist researchers in developing a
large-scale study of the outcomes of silicone breast implant placement.
Twenty-six women responded but 1 declined to participate after she learned more about the study, which gave us a total sample of 25 women--8 from Canada and 17 from the United States. At the end of the interview sessions, we offered participants $20 for themselves or for the charity of their choice in gratitude for their time. We have no clear impression of how this selection process may have biased our sample and the content of our interviews. By comparison, the mammography study, which involved telephone interviews of comparable length, recruited participants through cold calls to women drawn at random from a national sampling frame. We sought 2 to 5 women for each cell of a design, with age, education, and race/ethnicity as factors. In some cases, we required up to 20 calls to fulfill the quota. Although the eventual interviews proved to be as collaborative and as intense as those we report in this article, it is not clear that they involved more representative samples.
In our final sample, the mean time since implantation was 17 years (range: 9-31 years). Eighteen participants had implants for cosmetic purposes and 7 for reconstruction. Their mean age was 52 years (range from 32 to 72 years). Seven had had implants removed and not replaced; 16 had their original implants. The reason for the implant was related to age at time of implant (r = 0.43, p [is less than] .05); cosmetic implants were more likely among younger participants. The reason was not related to interviewer, year of implant, country, or whether the implant had been removed. All but 1 respondents reported being White (she was Asian). Median reported income was US$50,000-$75,000.
Although the population of women with implants is not well characterized, demographic descriptions show them to be predominantly White, living in the South and West, aged 35 to 44 years at the time of implant, and with a relatively high family income.[13,63,64] Over the years, the number of reconstruction implantations has increased. By the end of 1990, about 60% to 70% of implant surgeries were performed for cosmetic reasons.[1,13,63] Thus, our sample was generally similar to the reported population of women with implants (late 1980s to early 1990s) in terms of age, race, income, and reason for implantation.
With participants' consent, we recorded and transcribed the interviews. The transcriptions were checked against the tapes, and the final versions were divided into statements. We then coded the transcribed interviews into the nodes of the expert model. Initially, several interviews were coded independently by two individuals. We resolved differences by discussion. Once the coding procedure seemed stable, however, a single individual coded subsequent interviews. Responses to structured questions, generally at the end of the interview, were simply tallied. Participants' names were separated from the tapes and data files before we conducted any data analysis.
The mean number of statements (per participant) coded into each node are summarized in Figure 2. Deeper shading represents more statements. These means range from 0 (nodes i and j, irrigation of the surgical pocket and implant) to 25.6 (node y, the physical consequences for women arising from problems with the interventions or interventions designed to alleviate them). The median for most nodes is somewhat smaller than the mean as a result of the distribution of mentions for individual respondents having been skewed by a few respondents who provided many statements.
The responses can also be described in terms of the proportion of participants who mentioned each link at least once (data not shown). In this study, almost all statements could be coded to one of the expert model nodes--indicating that the model covered the topics on these women's minds, even if the content of those beliefs varied from the corresponding expert opinions.
Most of the women interviewed had something to say about most elements in the expert model. Nonetheless, we found gaps in their mental models capable of undermining the individuals' decision making about their implants. One was a misunderstanding of the terms used to describe implant failure (ie, rupture, leak, and bleed). Another was exaggerating the vulnerability of implants to direct impacts, such as car accidents. Participants also tended to overestimate their ability to detect localized implant problems and to select medical remedies for them. Although they were generally satisfied with their own implants, the women interviewed were dissatisfied with the decision-making processes that led to their choices.
With such a rich data set, one cannot report everything in a single article. Here, we focus on participants' understanding of the etiology, symptomatology, diagnosis, and treatment of local complications--important issues for any health program serving women with implants. Because those programs must address women as people, not just as the potential bearers of health problems, we also describe respondents' attitudes toward their implants and the associated social and medical processes. We begin with the latter, hoping that it will create an understanding of the respondents before we get to their medical beliefs.
Attitudes Toward Implants
The benefits of implants cited by the respondents included cosmetic changes, improved self image, reconstruction, psychological impacts, and better clothing selection or fit. Two thirds of respondents expressed a positive overall feeling toward their implants, one quarter a negative overall feeling, and some were neither positive nor negative overall. Not surprisingly, negative feelings were much more common for those women whose implants had been removed (r = 0.82, p [is less than] .001). These feelings were unrelated to years since implant, age at implant, current age, or interviewer.
Positive feelings were more common with cosmetic than with reconstructive implants (72% vs 57%), although the difference was far from statistically significant. Participants from the United States were somewhat more positive (81% vs 25%, p [is less than] .05) and much less likely to mention fear of silicone spontaneously (24% vs 88%) than the Canadian women. Nevertheless, 18 of these 25 women said that, knowing what they know now, they would have made different choices. Of these women, 11 would not get implants again; 4 would choose saline or soy oil implants instead of silicone; and 7 would do more research, get more counseling, or just consider other, unspecified, options.
When we asked respondents directly whether they would recommend implants to a friend who sought their advice, only 1 said that she would do so. One third would definitely recommend not getting an implant, whereas one fifth would recommend that the friend try to accept herself as she is. When asked about specific information that women needed to consider when contemplating an implant, 17 respondents mentioned health risks, 5 mentioned the longevity of the implants, and 8 said (in effect) "everything." Specific suggestions included considering recent scientific research findings, the impact of an implant on breast-feeding and pregnancy, details of the surgery, psychological impacts, and having a realistic view of the outcome. Many participants expressed displeasure with their own decision-making processes and with the roles of their physicians in those decisions. For example, 15 said that their physicians had not informed them about the risks, compared with 8 who said that they had been told; however, only 2 reported not being informed about alternatives, compared with 10 who said that they had been.
When asked about the possibility of local problems, more than half of the women mentioned leakage and hardening (or some other term connoting contracture). Other problems, in order of decreasing frequency, were ruptures, pain or discomfort, arthritis or lupus (and other immune system disorders), scar tissue, implant migration, rejection, degeneration, loss of breast sensation, and heart palpitations. In some cases, respondents believed that they had experienced such problems; in others, these were viewed as just possibilities. The nature of the interview precluded taking fuller medical histories.
From respondents' descriptions of these possibilities, it was clear that they did not share the expert community's terminology. Many admitted not having heard all three of the terms (bleed, leak, and rupture). When asked what these words might mean, respondents generally believed that "bleed" referred to fluid seeping slowly into the body, "leak" referred to fluid escaping through a small hole or seam, and "rupture" referred to catastrophic loss of all contents--as with an explosion. Some saw the three problems as a common sequence of events, with bleeding leading to leakage, then to rupture.
Rupture was often cited as potentially arising from an impact, such as a blow to the chest or even from mammography provided by inexperienced technicians. Eleven respondents specifically mentioned concern about having their implants rupture in an auto accident, either through direct impact or airbag release. A few believed that removal of the implant could cause rupture. Many viewed rupture as dangerous enough to require immediate medical attention. All but 3 believed that a woman could tell if she was having problems with her implant. The symptoms that respondents cited were (in decreasing frequency), changes in the shape or consistency of the breast, soreness, perceptible hardening or lumps, a general decline in health, movement of the implant, or a change in the feeling of its edges.
When asked what treatment was possible with problematic implants, 21 of the 25 women cited removal. Some thought that the recurring symptoms could be treated with drugs or, in 2 cases, with alternative medicine and diet to "detoxify" the body. Many worded about the psychological impacts of removal and the lack of social support for women who had it done. Typically, they saw little that women could do to reduce these risks. A small number (6) suggested that a skilled surgeon could reduce the risks.
When asked to quantify the risks of such problems, no respondent denied the possibility, but 8 could not give a time or probability of occurrence. Varying numbers of the remainder reported that all silicone implants started leaking immediately (2); that leaking started early or not all (3); that some time period was all that could be expected, typically 10 to 15 years out (6); and that lifetime probability ranged from small to 80% (10).
When asked whether some women faced particular risks, few respondents mentioned any of the features of the implant design, other than its filler material or its specific placement. Some talked in general terms about the skill of the surgeon (9) or the placement of the implant (4). Almost none mentioned procedures known to affect the rate of implant problems (nodes h-j). Many theories about individual differences were advanced. One recurrent theme referred (in a general way) to differences in body chemistry affecting reactivity to the implant.
Descriptions of the placement of their own implants were generally correct, but imprecise (eg, "in front of the chest muscle," "between fat and muscle," or "under the muscle"). Few used the technical terms (eg, subglandular, submuscular) needed to interpret medical results in terms of their own circumstances. A few expressed incorrect beliefs, such as "attached to the ribs" and "above the diaphragm."
All but 2 of the respondents believed that women with implants required special mammography procedures. Specific differences included taking more pictures, avoiding pressure on the implant, pushing the implant out of the way to get a better view of the tissue, and having the technician administer the mammogram more gently. Of the 22 participants who had had a mammogram since receiving their implants, several reported experiencing pain during the procedure. One said that her regular mammograms established that her implants were fine.
Respondents typically believed that mammography was less effective for women with implants. Two offered that MRI or ultrasound was better. In response to a direct question, 15 said that women with implants are not more likely to develop breast cancer, 3 thought that they were, and 7 did not know. Responses did not reveal whether the 3 women seeing increased risk attributed it to greater incidence of breast cancer or to reduced ability to detect it, once there.
The health professional's ability to serve women with breast implants (or other health problems) depends on his or her understanding of women's current beliefs and concerns, as well as on the terms the patients use to describe them. Only with such knowledge is it possible to provide information that women want and need in terms that they can comprehend. Caring enough to identify the most pertinent message demonstrates respect and should contribute to the message's credibility and understandability. The women who participated in these interviews had a great deal to say about their implants--covering most topics in our expert model. However, their beliefs regarding local complications were fragmented and often inaccurate or expressed in terms incompatible with those used by medical specialists. They needlessly feared some problems (eg, rupture attributable to airbag deployment), but they were unaware of others (eg, undetected local problems). Thus, although they had the basis for a fairly complete understanding, the necessary connections and corrections had not reached them.
A limitation of this study is the way we obtained the respondent sample and its small size. Respondents to a newspaper advertisement might have reactions to support that are stronger than those of "average" patients. Furthermore, attempting to estimate the content of 26 data nodes (a-z) from 25 subjects is certain to capitalize on chance. We believe, however, that the research model we demonstrated will lend itself effectively to larger samples of respondents.
The importance of these issues, and of a consultative relationship, can be seen in our respondents' willingness to spend a substantial period of time talking with our interviewers and openly sharing personal information, which sometimes was an emotional event for them. Their readiness seemed to reflect desires to talk about their experiences and to help other women get better care. Their feelings about the inadequacy of their personal decision-making processes and the roles of medical personnel in those decisions reflect this desire.
Most of these women had positive overall attitudes toward their implants. Yet, most would have made different decisions, knowing what they know now. Thus, having reached their choice through an unsatisfactory decision-making process still weighed on them. The process that they recommended for women contemplating implants involved the sort of informed-consent procedure currently in place, partly as a result of the litigation over systemic health effects. Indeed, the Institute of Medicine Committee on the Safety of Silicone Breast Implants urged that "a national model of informed consent be developed for women undergoing breast implantation, and that the effectiveness of this model be monitored. If the model is successful, it could be applied to other implantable devices in the future."
Comparing the experts' and patients' mental models suggests some specific steps that could be taken once a suitable communication process is in place. One is to clarify terminology that is accepted by the medical community, but not understood as intended by women. All three terms, bleed, leak, and rupture, created some confusion, with the severity of rupture being typically overstated. Speculatively, the term "rupture" itself contributes to the common belief that implants could be broken open by direct blows, such as those experienced in accidents.
This particular misunderstanding could lead to inappropriate alarm or complacency. On the one hand, women with implants may feel vulnerable in situations that pose little or no risk (eg, mammography, airbag deployment). On the other hand, they may also expect a clearer signal of implant failure than is typically provided. None of our interviewees raised the very real possibility that a woman could have a ruptured implant and not know it, despite having several prompts for talking about problems in general and failures in particular. The thought that they could be living with a compromised implant would, presumably, be discomfiting to a majority of these women, who believed that the associated effects could be severe and that removal or drug treatment would be required.
In one respect, creating more accurate mental models seems relatively straightforward (although these suggestions, like all others, must be subjected to empirical test, using procedures like those summarized by Fischhoff and colleagues)[3,5,65] Almost all of the women we interviewed realized that women with implants needed special mammography procedures. Although a woman could not be expected to master the precise procedures or monitor mammography technicians, she does need to know that these special procedures pose no threat to an implant's integrity (nor do routine procedures). Radiologists and mammography technicians could reasonably be taught (and expected) to convey this message in clear, credible terms. The moderately common fear that mammography can cause rupture was not accompanied by an increase in the perceived risk of breast cancer. Thus, the concern is over the local complications.
Although women were aware of distinctly local problems, those beliefs sometimes merged with concerns over systemic problems, as in the belief in drug treatment following the removal of compromised implants. We speculate that women's mental models of local problems have suffered as a result of the great attention paid to systemic problems. Not only has the distinction been blurred (leading to confusion between the two classes of problems) but women have not received the clear, focused messages that they need about local problems. Such messages should benefit from the relative lack of controversy about local complications. Although the relevant science is imperfect, no one denies that these problems exist. Thus, it should be possible to avoid the confusion that arises when laywomen watch experts and advocates argue about the reality of systemic problems.
Our findings also suggest that communicating the facts about local complications should not be that difficult. Women already have many of the pieces. The few common misconceptions seem readily addressed by describing how implants can fail, what consequences can follow, and what symptoms should be apparent. This also seems to be a case in which the expert model provides an easily followed template for presenting information. That model begins with those particulars of individual implants most relevant to the risks of local problems (eg, placement, implant design). Women with implants should keep them in mind when considering the relevance of particular messages. It may require learning personal details about their implants that they were not told or do not remember. The model continues with events that could compromise the integrity of an implant, with associated consequences and with options for treating or preventing problems.
Some messages will apply to large classes of women. For example, those who have been counting on self-monitoring to detect problems need to know about its accuracy and the availability of other measures. Other messages will have to target subgroups of women with specific circumstances[4,65] and questions. Seeing the fuller picture might make a woman more or less concerned overall. How these concerns are resolved should depend on the balance that each woman seeks between the various possible physical and psychological consequences of her action options (eg, explantation, watchful waiting, and neglect). Doing so might require some of the same individualized counseling that our interviewees thought necessary for implantation decisions. The psychological processes set in motion by implantation, local problems, and their treatment might prove as complex as the physiological processes. Such counseling might provide some of the social support that our interviewees found was lacking in their current situations. If so, then it might reduce the (aversive) uncertainty about what is happening (and what to do), while shifting the distribution of consequences in a positive direction.
The Breast Implant Public Health Project provided financial support for this study through a grant from the Dow Corning Corporation. We wish to express our thanks to Liz Adams, Barbara Blasick, Julie Downs, and Susan Houston for their help with the conduct of this research, as well as to anonymous reviewers. The views expressed are those of the authors.
For further information, please address communications to Martha Embrey, MPH, Department of Environmental and Occupational Health, The George Washington University School of Public Health and Health Services, 2300 K St, NW, Washington, DC 20037, firstname.lastname@example.org.
[1.] Bondurant S, Ernster V, Herdman R, eds. Safety of Silicone Breast Implants. Washington, DC: Institute of Medicine (US), National Academy Press; 1999.
[2.] Bostrom A, Fischhoff B, Morgan MG. Characterizing mental models of hazardous processes: A methodology and an application to radon. Social Issues. 1991;48:85-100.
[3.] Fischhoff B. Why (cancer) risk communication can be hard. National Cancer Institute Monographs. 1999;25:7-13.
[4.] Fischhoff B, Downs J. Communicating foodborne disease risk. Emerg Infect Dis. 1997;3:489-495.
[5.] Fischhoff B, Bostrom A, Quadrel MJ. Risk perception and communication. In: Detels R, McEwen J, Omenn G, eds. Oxford Textbook of Public Health. London: Oxford University Press; 1997.
[6.] Fischhoff B, Downs J, Bruine de Bruin W. Adolescent vulnerability: A framework for behavioral interventions. Applied Preventive Psychology. 1998;7:77.
[7.] Morgan MG, Fischhoff B, Bostrom A, Lave L, Atman CJ. Communicating risk to the public. Environmental Science & Technology. 1992;26:2048-2056.
[8.] Burns WJ, Clemen RT. Covariance structure models and influence diagrams. Management Science. 1991;39:816-834.
[9.] Clemen RT. Making Hard Decisions: An Introduction to Decision Analysis. Boston: PWS-Kent; 1991.
[10.] Howard R. Knowledge maps. Management Science. 1989;35,903-922.
[11.] Ashley FL, Braley S, Rees TD, Goulian D, Ballantyne Jr DL. The present status of silicone fluid in soft tissue augmentation. Plast Reconstr Surg. 1967;39:411-420.
[12.] Smith LF, Smith TT, Yeary E, McGee JM, Malnar K. Squamous cell carcinoma of the breast following silicone injection of the breasts. J Okla State Med Assoc. 1999;92:126-130.
[13.] Cook RR, Delongchamp RR, Woodbury M, Perkins LL, Harrison, MC. The prevalence of women with breast implants in the United States--1989. J Clin Epidemiol. 1995;48:519525.
[14.] Gabriel SE, Woods JE, O'Fallon WM, Beard CM, Kurland LT, Melton III LJ. Complications leading to surgery after breast implantation. N Engl J Med. 1997;336:677-682.
[15.] Oldham, J. Companies debut 1st ads for saline breast implants. Los Angeles Times. 20 May 1999; C1.
[16.] Sepai O, Henschler D, Czech S, Eckert P, Sabbioni G. Exposure to toluenediamines from polyurethane-covered breast implants. Toxicol Lett. 1995;77:371-378.
[17.] Chan SC, Birdsell DC, Gradeen CY. Urinary excretion of free toluenediamines in a patient with polyurethane-covered breast implants. Clin Chem. 1991;37:2143-2145.
[18.] Nemecek JAR, Young VL. How safe are silicone breast implants? South Med J. 1993;86:932-944.
[19.] Asplund O. Capsular contracture in silicone gel and saline-filled breast implants after reconstruction. Plast Reconstr Surg. 1984;73:270-275.
[20.] Barker DE, Retsky M, Searles SL. New low-bleed implant--Silastic II. Aesthetic Plast Surg. 1985;9:39-41.
[21.] Cairns TS, de Villiers W. Capsular contracture after breast augmentation--A comparison between gel- and saline-filled prostheses. S Afr Med J. 1980;57:951-953.
[22.] Chang L, Caldwell E, Reading G, Wray Jr RC. A comparison of conventional and low-bleed implants in augmentation mammaplasty. Plast Reconstr Surg. 1992;89:79-82.
[23.] Gylbert L, Asplund O, Jurell G. Capsular contracture after breast reconstruction with silicone-gel and saline-filled implants: A 6-year follow-up. Plast Reconstr Surg. 1990;86:260-267.
[24.] McKinney P, Tresley G. Long-term comparison of patients with gel and saline mammary implants. Plast Reconstr Surg. 1983;72:27-29.
[25.] Reiffel RS, Rees TD, Guy CL, Aston SJ. A comparison of capsular formation following breast augmentation by saline-filled or gel-filled implants. Aesthetic Plast Surg. 1983;7:113-116.
[26.] Vistnes LM, Bentley JW, Fogarty DC. Experimental study of tissue response to ruptured gel-filled mammary prostheses. Plast Reconstr Surg. 1977;59:31-34.
[27.] Biggs TM, Yarish RS. Augmentation mammaplasty: Retropectoral versus retromammary implantation. Clin Plast Surg. 1988;15:549-555.
[28.] Burkhardt BR. Capsular contracture: Hard breasts, soft data. Clin Plast Surg. 1988;15:521-532.
[29.] Handel N, Jensen JA, Black Q, Waisman JR, Silverstein MJ. The fate of breast implants: Critical analysis of complications and outcomes. Plast Reconstr Surg. 1995;96:1521-1533.
[30.] Jarrett JR, Cutler RG, Teal DF. Subcutaneous mastectomy in small, large, or ptotic breasts with immediate submuscular placement of implants. Plast Reconstr Surg. 1978;62: 702-705.
[31.] Puckett CL, Croll GH, Reichel CA, Concannon MJ. A critical look at capsule contracture in subglandular versus sub-pectoral mammary augmentation. Aesthetic Plast Surg. 1987;11:23-28.
[32.] Vazquez B, Given KS, Houston GC. Breast augmentation: A review of subglandular and submuscular implantation. Aesthetic Plast Surg. 1987;11:101-105.
[33.] Woods JE, Irons GB, Arnold PG. The case for submuscular implantation of prostheses in reconstructive breast surgery. Ann Plast Surg. 1980;5:115-122.
[34.] Worseg A, Kuzbari R, Tairych G, Korak K, Holle J. Long term results of inflatable mammary implants. Br J Plast Surg. 1995;48:183-188.
[35.] Coleman DJ, Foo ITH, Sharpe DT. Textured or smooth implants for breast augmentation? A prospective controlled trial. Br J Plast Surg. 1991;44:444-448.
[36.] Ersek RA. Rate and incidence of capsular contracture: A comparison of smooth and textured silicone double-lumen breast prostheses. Plast Reconstr Surg. 1991;87:879-884.
[37.] Gasperoni C, Salgarello M, Gargani G. Polyurethane-covered mammary implants: A 12-year experience. Ann Plast Surg. 1992;29:303-308.
[38.] Hakelius L, Ohlsen L. A clinical comparison of the tendency to capsular contracture between smooth and textured gel-filled silicone mammary implants. Plast Reconstr Surg. 1992;90:247-254.
[39.] Pollock H. Breast capsular contracture: A retrospective study of textured versus smooth silicone implants. Plast Reconstr Surg. 1993;92:404-407.
[40.] Shapiro MA. Smooth vs. rough: An 8-year survey of mammary prostheses. Plast Reconstr Surg. 1989;84:449-457.
[41.] Ahn CY, Ko CY, Wagar EA, Wong RS, Shaw WW. Microbial evaluation: 139 implants removed from symptomatic patients. Plast Reconstr Surg. 1996;98:1225-1229.
[42.] Dobke MK, Svahn JK, Vastine VL, Landon BN, Stein PC, Parsons CL. Characterization of microbial presence at the surface of silicone mammary implants. Ann Plast Surg. 1995;34:563-569.
[43.] Gylbert L, Asplund O, Bergeeren A, Jurell G, Ransjo U, Ostrup L. Preoperative antibiotics and capsular contracture in augmentation mammaplasty. Plast Reconstr Surg. 1990;86: 260-269.
[44.] Netscher DT, Weizer G, Wigoda P, Walker LE, Thornby J, Bowen D. Clinical relevance of positive breast periprosthetic cultures without overt infection. Plast Reconstr Surg. 1995;96:1125-1129.
[45.] Virden CP, Dobke MK, Stein P, Parsons CL, Frank DH. Sub-clinical infection of the silicone breast implant surface as a possible cause of capsular contracture. Aesthetic Plast Surg. 1992;16:173-179.
[46.] Malata CM, Varma S, Scott M, Liston JC, Sharpe DT. Silicone breast implant rupture: Common/serious complaints? Med Prog Tech. 1994;20:251-260.
[47.] Peters W, Smith D, Lugowski S. Failure properties of 352 explanted silicone-gel breast implants. Can J Plast Surg. 1996;4:55-58.
[48.] Robinson OG, Bradley EL, Wilson DS. Analysis of explanted silicone implants: A report of 300 patients. Ann Plast Surg. 1995;34:1-7.
[49.] Rohrich RJ, Adams Jr WP, Beran SJ, Rathakrishnan R, Griffin J, Robinson Jr JB, Kenkel JM. An analysis of silicone gel-filled breast implants: Diagnosis and failure rates. Plast Reconstr Surg. 1998;102:2304-2308.
[50.] Bantick GL, Taggart I. Mammography and breast implants. Br J Plast Surg. 1985;48:49-52.
[51.] Sullivan DC, Linden SS. Imaging of augmentation mammaplasty patients. In: Georgiade NG, Georgiade GS, Riefkohl R, eds. Aesthetic Surgery of the Breast. Philadelphia: Saunders; 1990.
[52.] Agency for Health Care Policy and Research. Clinical Practice Guideline, Number 13, Quality Determinants of Mammography. Rockville, MD: US Dept of Health and Human Services; October 1994. Available from: US GPO, Washington, DC; AHCPR Publication No. 95-0632.
[53.] Eklund GW, Busby RC, Miller SH, Job JS. Improved imaging of the augmented breast. American Journal of Roentgenology. 1988;151:469-473.
[54.] Birdsell DC, Jenkins H, Berkel H. Breast cancer diagnosis and survival in women with and without breast implants. Plast Reconstr Surg. 1993;82:795-800.
[55.] Brinton LA, Malone KE, Coates RJ, Schoenberg JB, Swanson CA, Daling JR, et al. Breast enlargement and reduction: Results from a breast cancer case-control study. Plast Reconstr Surg. 1996;97:269-275.
[56.] Cahan AC, Ashikari R, Pressman P, Cody H, Hoffman S, Sherman JE. Breast cancer after breast augmentation with silicone implants. Ann Surg Oncol. 1995;2:121-125.
[57.] Carlson GW, Curley SA, Martin JE, Fornage BD, Ames FC. The detection of breast cancer after augmentation mammaplasty. Plast Reconstr Surg. 1993;91:837-840.
[58.] Clark III CP, Peters GN, O'Brien KM. Cancer in the augmented breast: Diagnosis and prognosis. Cancer. 1993;72:2170- 2174.
[59.] Deapen DM, Bernstein L, Brody GS. Are breast implants anticarcinogenic? A 14-year follow-up of the Los Angeles study. Plast Reconstr Surg. 1997;89:660-665.
[60.] Byram S. Breast cancer and mammogram screening: Mental models and quantitative assessments of belief. Dissertation. Department of Social and Decision Sciences, Carnegie Mellon University; 1998.
[61.] Silverman E, Woloshin S, Schwartz LM, Byram SJ, Welch HG, Fischhoff B. Women's interpretation of breast cancer risk and screening mammography: A qualitative interview study. Med Decision Making. 2001;21:231-240.
[62.] Food and Drug Administration. Breast Implants, An Information Update. March 1996;18:31.
[63.] Bright RA, Jeng LL, Moore RM. National survey of self-reported breast implants: 1988 estimates. J Long Term Eff Med Implants. 1993;3:81-89.
[64.] Handel N, Wellisch D, Silverstein MJ, Jensen JA, Waisman E. Knowledge, concern, and satisfaction among augmentation mammaplasty patients. Ann Plast Surg. 1993;30:13-20.
[65.] Fischhoff B. Giving advice: Decision theory perspectives on sexual assault. Am Psychol. 1992;47:577-588.
Dr Byram, who died of breast cancer in June 2001, was a post-doctoral fellow in the Department of Psychology at Carnegie-Mellon University in Pittsburgh, Pennsylvania, where Dr Fischhoff is a professor in the Department of Social and Decision Sciences. Ms Embrey is a research associate in the Department of Environmental and Occupational Health at The George Washington School of Public Health and Health Services in Washington, DC. Dr Bruine de Bruin is a postdoctoral fellow in the Department of Technology Management at the Technical University of Eindhoven, The Netherlands, and Ms Thorne is with Thorne Butte: Decision Partners, in Toronto, Ontario.
COPYRIGHT 2001 Heldref Publications