Abstract
After 5600 families of children diagnosed with autism filed claims with the National Vaccine Injury Compensation Program in the United States, the court selected ‘test’ cases consolidated into the Omnibus Autism Proceedings, held from 2007 to 2008, to examine claims that vaccines caused the development of autism. The court found all of the causation theories presented to be untenable and did not award damages to any parents. We analyze the Omnibus Autism Proceedings as a struggle within the scientific field between the scientific orthodoxy of the respondents and the heterodox position taken by the plaintiffs, suggesting that the ruling in these cases helped to shore up hegemony on autism causation. Drawing on the literature on non-knowledge, we suggest that only the respondents had enough scientific capital to strategically direct non-knowledge toward genetic research, thereby foreclosing the possibility of environmental causation of autism. The plaintiffs, who promote a non-standard ontology of autism, suggest that the science on autism remains undone and should not be circumscribed. In analyzing the Omnibus Autism Proceedings with field theory, we highlight the way in which scientific consensus-building and the setting of research agendas are the result of struggle, and we show that the strategic deployment of non-knowledge becomes a major stake in battles for scientific legitimacy and the settling of scientific controversies.
Keywords
According to the Centers for Disease Control (CDC), the prevalence of autism in the United States has increased from 1 in 2500 children in 1989 (Eyal et al., 2010) to 1 in 68 children in 2014 (CDC, 2014). 1 Most scientists suggest that this exponential increase in prevalence is due to better reporting and diagnostic substitution, which occurred after autism was added to the Individuals with Disabilities Education Act of 1991 (Baker, 2008; Eyal et al., 2010; Fombonne, 2001, 2003). But there are many who suggest that this growing ‘autism epidemic’ can best be explained by environmental factors. The most controversial theory promoted suggests that the increase in the number of childhood vaccinations added to the routine schedule in the 1990s and 2000s accounts for the increase in the prevalence of autism among children (Kaufman, 2010).
There are different historical accounts of the development of both the link between vaccines and autism, and also the consensus formed against the linkage in the scientific mainstream. Jeffrey Baker (2008) suggests that the controversy around the vaccine–autism link arises from the convergence of three histories: thimerosal (an ethylmercury preservative introduced after World War I to prevent bacterial contamination of vaccines), methylmercury and its neurotoxic properties, and the shift toward biological explanations of autism, which offered hope that it could be ‘cured’. Baker suggests that when the rates of autism began to rise in the early 1990s, parents of newly diagnosed children were confronted with a medical community that knew little about it, which caused many parents to search out ‘alternative’ autism research.
In 1998, British gastroenterologist Andrew Wakefield proposed a hypothesis linking the MMR (measles, mumps, and rubella) vaccine, inflammatory bowel disease, and autism. The publication of the study in The Lancet launched a controversy in Britain despite the ‘failure of large epidemiological studies to confirm its results’ (Baker, 2008: 249). This controversy soon spread across the Atlantic, uniting parents of children with autism and parent groups opposed to compulsory immunization (Baker, 2008). In 2000, Congressman Dan Burton instigated a series of congressional hearings on autism and vaccine safety, after his grandchild was diagnosed with autism following 12-month vaccines (Baker, 2008). This history converged with a Food and Drug Administration (FDA)-initiated risk assessment of thimerosal in vaccines in 1998–1999. ‘Although noting that there was no evidence that the use of thimerosal as a vaccine preservative had caused any true harm’, the American Academy of Pediatrics (AAP) and the CDC agreed that ‘“thimerosal-containing vaccines should be removed as soon as possible” given the concerns raised by the Environmental Protection Agency’s guidelines’ (Baker, 2008: 250). As David Kirby (2006) notes, it was parents of children with autism who first took up the theory that thimerosal might explain the ‘autism epidemic’, which led to the publication of a study in Medical Hypotheses linking autism symptoms to mercury harm (Bernard et al., 2001).
Through the early 2000s, a series of events challenged the emerging theory of a possible vaccine–thimerosal–autism link. Between 2001 and 2007, there were eight Institute of Medicine (IOM) hearings that confirmed the safety of childhood vaccinations (Baker, 2008). Mark Largent (2012) suggests that by 2002, several epidemiological studies from around the world were produced to refute the link between vaccines and autism and the scientific debate surrounding autism causation ended around this time (p. 122). For example, Largent (2012) reports that Wakefield’s co-authors on The Lancet article broke with him when he began publicly stating that the MMR vaccine and autism were definitively linked, although the paper only suggested a possible causal pathway (pp. 108–109). A series of papers published after 1998 challenged Wakefield’s findings (Largent, 2012: 119–122). In 2004, The Lancet retracted the 1998 paper (although Wakefield did not agree to the retraction) (Largent, 2012: 131–132). Wakefield’s General Medical Council hearings began in 2007, and his medical license was revoked in 2010 (Largent, 2012: 135).
Despite the settling of scientific consensus on the lack of connection between vaccines and autism, parental concerns remained (Largent, 2012: 122). After more than 5600 families filed cases related to autism with the National Vaccine Injury Compensation Program (NVICP), the court selected several ‘test’ cases consolidated into the Omnibus Autism Proceedings (OAP) held from 2007 to 2008, to ‘examine the evidence for claims about injuries caused by the MMR vaccine, thimerosal-containing vaccines or a combination of the two’ (Silverman, 2012: 221; see also Kirkland, 2012: 238). In the end, the court found all of the causation theories presented to be untenable and failed to award damages to any of the parents who had filed claims. Anna Kirkland (2012) suggests that the OAP effectively shut down any lingering debate about whether vaccines were a potential cause of autism in children. We suggest that the rulings on these cases (in 2009 and 2010) played a role in clinching and shoring up scientific hegemony on broader issues in autism causation and in steering future research investment. Both the respondents’ experts and the Special Masters of the OAP argue that future research should focus on genetics and that research into environmental causes should only be considered as part of gene–environmental interactions.
The OAP hearings have been analyzed by two previous scholars (Kirkland, 2012; Silverman, 2012), both of whom argue that it was a case of credibility battles over the ‘soundness’ of science and the expertise of those called to testify for the families and the government. We agree that the credibility of both the science and expertise presented was being adjudicated in these proceedings, but we analyze the OAP as a struggle within the scientific field between the scientific orthodoxy of the respondents and the heterodox position taken by the plaintiffs (Bourdieu, 1991). In addition, we argue that non-knowledge was a primary tool wielded in this battle.
The respondents (representing the US government) relied upon the existing scientific record, which they suggested shows sufficiently that vaccines are not a causal factor in the development of autism. They argued that there was no longer scientific uncertainty in this arena of autism research and that post-natal environmental insults are not viable risk factors for the development of autism. Instead, they claimed that the non-knowledge that continues to exist in autism research is within the realm of genetics. Therefore, the respondents argued that genetics should be the focus of research. In this way, they circumscribe the uncertainty surrounding autism in order to refocus attention on the potentialities of the genome. In order to strategically direct non-knowledge in such a fashion, they relied upon the scientific capital of the experts called by the government and the growing consensus among the scientific mainstream about autism causation. As Linsey McGoey (2009) has argued, non-knowledge can be strategically deployed by those with enough social capital to express ‘legitimate doubt’ (p. 155).
The plaintiffs or petitioners (representing the parents), on the other hand, painted the field of autism science as wholly uncertain – the probabilities of risk are unknown and could come from many different sources. The parents and scientific experts called by the plaintiffs were closely linked to an embodied health movement (Brown et al., 2004) referred to at the time as DAN! (Defeat Autism Now!), 2 which promotes an ontology of autism that challenges the scientific mainstream. One of the reasons that these parents established a movement at odds with the scientific mainstream was that their parental experiences (and knowledge of their children) challenge a genetic account of autism causation. As we explain in further detail below, these parents saw their children developing normally and then suddenly regressing during the ages of the bulk of their childhood vaccinations (between 12 and 24 months). This regressive form of autism is not recognized as legitimate by the scientific mainstream, and scientific experts directly challenged the parents – suggesting that they simply did not see the early symptoms of autism. This challenging of the parents’ ‘situated knowledges’ (Haraway, 1988) caused many parents to join an alternative scientific movement, which posits a unique ontology of autism wherein autism is perceived to stem from a number of environmental assaults, vaccines among them. For these parents, then, the science on autism is undone (Frickel et al., 2009); they see non-knowledge everywhere and therefore suggest that scientific research should not be limited to the field of genetics.
In this way, the respondents and petitioners both argue that non-knowledge is a major feature of autism science, but they perceive it quite distinctly. The respondents argue that research into vaccine causation would produce knowledge that is not worth pursuing, or ‘negative knowledge’ (Knorr Cetina, 1999), while genetic research lies in the realm of ‘specified ignorance’ (Merton, 1987) or ‘extended knowledge’ (Gross, 2007), an area that is promising for research investment. Thus, experts for the respondents suggest that the future of autism research should be directed toward genetic research and that environmental causes (of all kinds and not simply in relation to vaccines) should only be considered as they relate to gene–environment interactions. 3 The parents, on the other hand, trust in their own parental, situated knowledge (and the experts they put their trust into have listened to their stories and have legitimated them), but they also see autism research as uncertain and undone (Frickel et al., 2009). Given an entire realm of potential harms not yet understood, the parents urge the government to invest in a wide array of research initiatives to effectively determine autism causation.
The NVICP itself is the perfect place for such a battle over non-knowledge, as the standards of evidence required are not only ambivalent but also variable. The program was set up to provide more lenient standards for proving causation than tort litigation, in order to make compensation more efficient and realizable. However, the standards of proof used in the vaccine court have shifted over time. Whereas more lenient standards were used in early proceedings (as in the case of Hannah Poling, which we discuss briefly below), the parents in the OAP hearings were asked to establish more rigorous standards of causal evidence. In the OAP, the standards of proof look more similar to tort litigation than in previous NVICP hearings.
Scientific capital and the field of autism science
Pierre Bourdieu (1991) argues that the scientific field is characterized by struggle or competition between actors and institutions, vying to valorize their own capital, control the distribution and exchange rate of capital in the field, and thus control what counts as legitimate scientific inquiry and research. Official science is not a unanimously recognized system of norms and values agreed upon by some ‘scientific community’, but is instead the result of struggle. In fact, Bourdieu (1975) suggests that the consensus of the ‘scientific community’ is an official fiction ‘which is not in the least fictitious because the symbolic efficacy from which it derives its legitimacy enables it to perform a symbolic function similar to that performed for liberal ideology by the notion of “public opinion”’ (p. 24).
We argue that autism science is a scientific field that is neither well-bounded nor settled. First, at the time of the OAP, the field was neither coherent nor secure because of the widespread questioning by parents and the proliferation of ‘authorities’ and ‘experts’ in mainstream and alternative scientific communities. In addition, there are a multitude of disciplines that make up the field from epidemiology, pediatrics, psychology, psychiatry, neuroscience, speech and language pathology, genetics, pharmacology, immunology, psychopathology, internal medicine, and gastroenterology – making it more like the archipelago of specializations Panofsky (2014) suggests also describes the field of behavior genetics. Finally, since its inception the field has been consistently bound up in political battles, thereby making it heteronomous and not well-bounded. Therefore, during the OAP hearings, the boundaries of the field, as well as the eligibility for membership in it, are part of what is at stake in the struggle:
[I]t is indeed because the definition of the stake of the struggle is a stake in struggle … that one endlessly runs into the antinomy of legitimacy: in the scientific field, as elsewhere, there exists no judiciary for legitimizing claims to legitimacy, and claims to legitimacy carry a weight proportional to the symbolic power of groups whose specific interests they express. (Bourdieu, 1991: 14)
And yet, we suggest that the OAP is exactly this ‘judiciary for legitimizing claims of legitimacy’ within the field of autism.
Every battle between those who profess orthodoxy and are currently dominant in the field (and therefore control the definition and valuation of capital) and those who express a heterodox position stands out against a backdrop of doxa – that which goes without saying because it is made up of the ‘set of presuppositions that antagonists take for granted’ because ‘they constitute the implicit condition for discussion and contention’ (Bourdieu, 1991: 9). In other words, battles between orthodoxy and heterodoxy must take place between agents who are already consecrated by the field and are endowed with the proper habitus to ‘play the game’. In the OAP, each side relies upon scientific expertise, and the importance of double-blind clinical trials is never in question. And yet, the boundaries of the field are nonetheless at stake in this battle, as are the boundaries between authentic and ‘false’ science (p. 8). Some agents in any field hold enough capital to be able to wield ‘power over capital held by other agents’, who are then ‘capable of consecrating certain objects by devoting their investments to them’ (p. 13), and this is precisely what we end up seeing in the OAP – the consecration of genetics as the future of autism research and the delegitimation of other research into causality. The dominant members of the scientific field ‘try to impose the definition of science that best conforms to their specific interest, that is, the one best suited to preserving or increasing their specific capital’ (p. 13).
In the following table, we show the differential capital of the experts called by the respondents versus those called by the plaintiffs (Table 1). This chart represents a selection of the experts whose testimonies were most heavily relied upon.
Respondent and plaintiff expertise.
JAMA: Journal of the American Medical Association.
As the table clearly illustrates, the respondents’ experts are positioned at major teaching hospitals in the United States, Canada, and the United Kingdom, some of them directing departments and laboratories; their papers are published in leading medical journals and many of their top papers are cited hundreds and even thousands of times. When they are asked on the stand or in their expert reports to give their credentials of expertise, they include the following: editing major journals in their fields; serving as consultants for the CDC, IOM, AAP, and National Institutes of Health (NIH); receiving large grants from the NIH and World Health Organization (WHO); sitting on scientific advisory councils for major research centers; and serving as presidents of national organizations representing their disciplines. Although it is clear from the table that the plaintiffs’ experts have less scientific capital than the respondents’ experts, some plaintiffs’ experts are also affiliated with academic institutions and have major publications in leading journals. They are, therefore, still positioned within the scientific field, although often in dominated positions.
During the court proceedings, the respondents’ experts regularly called the evidence used by the plaintiffs’ experts into question, either denying legitimacy to the scientific forums in which the evidence was published or denying that the relevant researchers had enough publications to be granted expert status. In the following two excerpts from respondents’ expert reports,
4
Drs Ward and Brent (experts for the respondents) observe that the studies upon which the plaintiffs’ experts rely for proof appear in either journals that are non-indexed or that do not appear in the National Library of Medicine, thereby calling their validity into question:
Brian Ward (expert for the respondents): Kinsbourne [expert for the plaintiffs] uncritically refers to the Bradstreet manuscript (Bradstreet, 2004) in support of his contention … This report of 3 cases was published in a non-indexed journal. Non-indexed journals are either brand new (ie: no track record of scientific rigor) or are thought to be insufficiently rigorous to be included in the medical-scientific literature search engines (eg: PubMed) … Non-indexed journals are generally ignored by scientists as sources of reliable scientific information. (Exhibit BB, Expert Report of Brian J Ward, p. 6) Jeffrey Brent (expert for the respondents): Dr. Aposhian [expert for the plaintiffs] cites a 1975 study by Koller and a 1981 study by Koller and Vos as evidence of immune dysregulation related to mercury exposure. A careful search of the National Library of Medicine database failed to identify any paper by Koller and Vos. (Exhibit L, Expert Report of Jeffrey Brent, p. 18)
Dr Brent similarly delegitimizes claims made by plaintiff experts because their findings were reviewed and rejected by prestigious panels:
Jeffrey Brent (expert for the respondents): I am aware of the studies by Geier and Geier (2003, 2004, and 2006) purporting to find a 2–6 fold statistically significant increase in the risk of developing autism associated with the administration of thimerosal-containing vaccines. These studies are methodologically flawed and nearly impossible to interpret and their results and reported correlation coefficients are highly implausible. In fact, such prestigious bodies as the 2004 IOM (2004 IOM Report, pp. 55–62, 65, Table 9), the American Academy of Pediatrics, the CDC, and the World Health Organization have reviewed the Geier and Geier studies and found them unconvincing and uninformative on the question of the hypothesized association between thimerosal-containing vaccines and ASDs. (Exhibit L, Expert Report of Jeffrey Brent, p. 19)
Dr Bustin (for the respondents) calls the broader expertise of plaintiff experts into question because of their lack of publication record in a particular field:
Stephen Bustin (expert for the respondents): A more exhaustive literature search for the authors does not reveal any high level of expertise, either in the field of autism or in RT-PCR [reverse transcription polymerase chain reaction]: A. Krigsman: 1 publication (2002) on laryngeal dysfunction; K(D) Hepner: 2 publications (2000, 2002) on scaffold proteins. Only one paper (2002) mentions PCR, and then as a method for amplifying DNA, not RT-PCR for quantitating mRNA levels. J Segal and SJ Walker have published one paper in this general area 11, but this did not use RTPCR. SJ Walker has published one short review in this general area. (Exhibit UU: Critique of Dr. Karin D. Hepner’s Letter, dated 21 May 2007, concerning the plausibility of an MMR-ASD causality hypothesis, by Stephen A Bustin, PhD, p. 11)
In these excerpts, the scientific authority capital (as measured by peer review and published successes) of the experts for the plaintiffs is called into question. It is also obvious that the respondents’ experts hold much more social authority (as measured by academic credentials) than those called for the plaintiffs. In the OAP, then, the respondents’ experts wield power over the capital of the plaintiffs’ experts and thus are able to delegitimize their scientific claims.
Bourdieu (1991) explains that the objective possibilities that are offered to agents involved in the field are determined by the relation between the universe of possibilities (which is determined by the state of scientific problems, theories, and beliefs) and the resources these agents can mobilize ‘which define … the universe of things “to be done”’ (p. 10). Bourdieu himself suggests that this process is not always conscious and is determined by actualizing the potentialities that lay dormant in the field, but, in the OAP, we see the strategic deployment of non-knowledge as a means of actualizing possibilities for future research. Partially at stake in this particular field battle is the issue of which knowledge gains investments and which does not.
The strategic production of non-knowledge
Matthias Gross (2007), Frank Knight (2006), and others make a distinction between risk and uncertainty, where risk assumes that the probabilities are known whereas uncertainty assumes they are not (and non-knowledge is the ‘realm that escapes recognition’ – Gross, 2007: 748). We suggest that these two frames are performatively deployed by the respondents and the plaintiffs. The respondents represent autism science as a field that contends with well-understood risks. Based on the existing evidence, there seems to be little to no risk posed by environmental factors for the development of autism; therefore, research should be directed where risk is greatest – at the genetic level. The plaintiffs, on the other hand, suggest that autism has multiple etiologies and results in divergent forms, which are little understood and not well researched. The distinction between risk and uncertainty indicates different perspectives on the ‘doneness’ of the science of autism, taken by those on either side of the courtroom.
We draw on McGoey’s (2009) research, which suggests that uncertainty is both generative and performative. It not only creates demand for resolution (which strengthens the authority of those who claim it), but it can also be strategically deployed for political or economic gain. 5 Following McGoey, we believe that the ability to express and manipulate uncertainty requires a specific amount of social capital. However, our account differs from McGoey’s in a two ways. First, in our analysis of the relationship between uncertainty and social capital, at stake is who has enough scientific capital to determine which non-knowledge is worthwhile and which is not – the ability to police the boundary of the known and unknown for the purposes of speculation about the value of future knowledge. McGoey, on the other hand, focuses on taming or not taming uncertainty. For example, reporting on how the drug company Merck was able to hide knowledge of the possible side-effects of Vioxx, McGoey (2009) suggests that Merck’s power comes from ‘emphasiz(ing) the impossibility of taming uncertainty’ (p. 154).
Second, we are cautious about McGoey’s tendency to reify distinctions between lay and expert knowledge. McGoey criticizes accounts of uncertainty by Adriana Petryna (2002), Veena Das (1995), and Brian Wynne (1996), which suggest that the laity is capable of strategically deploying uncertainty to contest expert advice and challenge political authority. McGoey (2009) argues that, in these cases, experts were called on behalf of the laity to strategically deploy uncertainty (p. 155). However, as we will illustrate in the case of the OAP, there is no clear distinction on the side of the plaintiffs between laity and experts. The parents involved in DAN! have a high degree of scientific knowledge and have come to depend on the doctors and scientists they call as experts precisely because those experts take their parental knowledge seriously. The plaintiffs deploy uncertainty strategically, even if they are delegitimated in the process.
Our approach to understanding the different usages and approaches to non-knowledge, by both sides of the OAP, is similar to that of Frickel et al. (2009) in their account of ‘undone science’. Experts for the respondents attempt to set a future scientific research agenda that privileges genetic knowledge, deemed to be in the area of ‘extended knowledge’, or that which deserves further attention and investment. The preferred agenda classifies research into environmental causes as ‘negative knowledge’ or knowledge that will remain ‘undone’ because it is considered unimportant or illegitimate (Gross, 2007; Knorr Cetina, 1999). The science exploring vaccine causality and other research into environmental causes of autism is not so much undone as underdone because the experts on whom the parents rely are trying to do this research on the edges of the scientific mainstream.
Methods
Public records (including rulings, briefs, transcripts, and decisions) from the OAP were accessed through the website of the Court of Federal Claims. 6 Six cases were held between June 2007 and July 2008, and the relevant transcripts and rulings totaled over 10,000 pages. Each day of testimony produced 200–300 pages of transcripts. There were three Special Masters who commented during proceedings, six attorneys for the respondents, three for the plaintiffs, 21 expert witnesses for the respondents, ten expert witnesses of the plaintiffs, and seven family members for the plaintiffs.
The transcripts are only available as protected PDFs, so file users’ activities are limited to opening, saving identical copies, and doing keyword searches. This prevented us from using qualitative data analysis software. Instead, we created a spreadsheet to summarize sections of testimony and assigned codes to those sections. This system allowed us not only to perform rudimentary open coding but also to summarize key pieces of evidence, tensions, and interpretation. We then created memos, summarizing each day’s testimony and explored how the codes operated. While primary coding and summarizing was done primarily by Underman, both authors read all the transcripts and met at least biweekly during coding to discuss codes and findings.
The OAP and the NVICP
The nature of the court system makes the NVICP an interesting setting for studying struggles over scientific capital and non-knowledge. As we will show, the structure and functioning of the NVICP had to adapt in order to accommodate the OAP hearings, and this had important results for the (non-)knowledge produced and the settling of the field of autism research.
The NVICP was established by the 1986 National Childhood Vaccine Injury Act, which compensates individuals injured by vaccines on a ‘no-fault’ basis. Funded by a 75-cent tax on vaccines (Keelan and Wilson, 2011; Kirkland, 2012), the program was established on the logic that, if the pharmaceutical industry were to be financially responsible for vaccine injury, the national supply of vaccines would be threatened (Keelan and Wilson, 2011).
While at least part of its intended function is to protect pharmaceutical companies from litigation and financial damages, the NVICP was established with contested goals. As Kirkland (2013) argues, the vaccine act was a ‘political compromise between doctors, parents and industry … Parents received compensation, doctors could keep vaccinating and manufacturers could keep producing without fear of bankruptcy’ (p. 12, italics added). Moreover, since its inception, the NVICP has been a site of struggle over what constitutes adequate causal proof of vaccine injury. As Betsey Grey (2011) argues, because the intended objective of the vaccine program has always been ambivalent, the definition of causal proof has remained vague. On the one hand, if the objective is to ‘encourage widespread vaccination of the population by ensuring that vaccines are not incorrectly blamed for causing injury’, then the standard of proof should be more stringent (Grey, 2011: 348). On the other hand, if the objective is to maximize industry liability, the standard should err on the side of leniency. In relation to the very stringent standards of proof required by tort litigation, the NVICP was intended to require more lenient standards. However, over time, the standards of proof used in vaccine courts have shifted toward standards of tort litigation, and parent-activists have come to suspect that the ‘political compromise’ has broken down (Kirkland, 2013: 12). It is important to note that, while plaintiffs’ attorneys come from private practices, respondents’ attorneys are from the Department of Health and Human Services (Kirkland, 2012) and the Vaccine Litigation Group of the Office of Constitutional and Specialized Torts (Davenport, 2000).
The NVICP has a Vaccine Injury Table to ‘expedite claims for which there is a known relationship between a vaccine and an adverse event’ (Kirkland, 2012: 238).
7
An off-Table claim, such as autism, requires the petitioner to bring the case before a ‘Special Master’ of the Court of Federal Claims. Initially, most successful claims were compensated because injuries fell under the Vaccine Injury Table. Now, off-Table claims account for 90 percent of all claims (Grey, 2011: 345–346). Kirkland (2013) argues that more and more claims were considered off-Table as of 1995, when new childhood vaccinations were added to the routine schedule and the Table was re-vamped with a narrower understanding of vaccine injury (pp. 18–19). The issue of causal proof has plagued off-Table claims because the Vaccine Act does not supply a standard of proof for cause of injury (Grey, 2011). In the vaccine court, petitioners have to ‘prove causation-in-fact by the preponderance of evidence standard’ (Kirkland, 2012: 238). This rests on a three-part evidentiary test:
1. a medical theory causally connecting the vaccination and injury; 2. a logical sequence of cause and effect showing that the vaccination was the reason for the injury; and 3. a showing of a proximate temporal relationship between vaccination and injury. (Althen v. Secretary of Health and Human Services (Althen III), cited in Grey, 2011: 379–80)
The amount and kind of necessary evidence is not clear, nor is the question of who bears the burden of proving that factors unrelated to the vaccine may have caused the injury (Grey, 2011: 356). This means that what counts as sufficient causation, and the preponderance standard, are left to the discretion of the Special Master in each case (Grey, 2011). In the OAP, the respondents spent a great deal of time making the case for genetic causation, precisely in order to argue that ‘factors unrelated to the vaccine’ caused the ‘injury’ of autism. In other words, they built their case by providing an alternative narrative of causation.
The Special Masters who serve in the NVICP are attorneys, appointed by the judges of the Court of Federal Claims (Kirkland, 2012). They have no scientific training or expertise, outside of hearing cases. While the eight original Special Masters were trained for 2 days in NVICP-relevant medical issues, none of the subsequently appointed Special Masters have received this training (Davenport, 2000). Thus, judgments are mostly based on expert testimony and the Special Masters’ particular readings and interpretations of scientific evidence presented in court.
In the course of the 2000s, when the OAP hearings took place, the vaccine court became increasingly stringent in terms of what evidence would be required for compensation (Kirkland, 2013). Prior to the OAP hearings, the court compensated a number of claims related to autism (Holland et al., 2011), as well as other injuries not recognized by available mainstream scientific evidence (Offit, 2008). In fact, Paul Offit (2008) claimed that the court had ‘turned its back on science’ (p. 2090). The most high-profile case was that of Hannah Poling, a then 9-year-old girl, who was awarded compensation in 2008 for autism exacerbated by vaccine injury (Donovan, 2010; Holland, et al., 2011; Keelan and Wilson, 2011; Offit, 2008). The case of Poling was originally part of the OAP, but was broken off because it was unique (Donovan, 2010). The NVICP ruled that being treated with an unusually high number of vaccines had exacerbated Poling’s underlying mitochondrial enzyme deficit and caused her encephalopathy (Offit, 2008). The case was judged on the basis of the plausibility of the causal mechanism that the plaintiffs put forward, although the plaintiffs lacked peer-reviewed studies and other hallmarks of ‘good science’. Poling’s parents called a press conference following the decision and declared that the court had ruled that vaccines do cause autism (Donovan, 2010). This declaration received some national media coverage, including an interview with Larry King (Donovan, 2010). The situation prompted a number of high-profile critics, such as Paul Offit, to scrutinize the scientific reasoning of the NVICP: ‘Going forward, the NVICP should more rigorously define the criteria by which it determines that a vaccine has caused harm’ (Offit, 2008: 2091; see also Keelan and Wilson, 2011).
During the 2007–2008 OAP hearings, three different biological causation theories were tested:
1. The combination of the MMR vaccine and the ethylmercury preservative, thimerosal, damages the child’s immune system so that her body cannot eradicate the measles virus, resulting in inflammation and the development of autism. 2. Thimerosal alone causes the development of autism through neuroinflammation in the brain. 3. The MMR vaccine alone causes the development of autism. (Keelan and Wilson, 2011; Kirkland, 2012)
Cases were selected to test each theory, although the final theory was dropped because ‘most of the evidentiary material addressing it would be covered in the first set of test cases’ (Keelan and Wilson, 2011: 2017). In the end, the Petitioner’s Steering Committee selected three cases to test Theory 1: Cedillo v. Secretary of HHS, Snyder v. Secretary of HHS, and Hazlehurst v. Secretary of HHS. Theory 2 was also tested by three cases: Dwyer v. Secretary of HHS, King v. Secretary of HHS, and Mead v. Secretary of HHS. In February 2009, the vaccine court ruled against the first three test cases and in March 2010 against the second three. Because of the factual record established by the ruling in these six cases, it is unlikely that any future cases of autism will be awarded compensation through the NVICP unless other petitioners can produce new evidence illustrating causation (Kirkland, 2012: 256n2).
This article mainly focuses on the most extensive and productive of the six cases: the 12-day case of Cedillo v. Secretary of HHS (12 days) and the 15-day joint hearings of King v. Secretary of HHS and Mead v. Secretary of HHS. As we will illustrate, the vaccine court functioned less in its designated role as a non-adversarial adjudicator of individual claims and more in the role of a ‘modernized court’ of mass tort litigation. As Sheila Jasanoff (2002) describes, when class-action lawsuits began to challenge the institutional capacity of the court system, the courts had to start making decisions about culpability for injury for groups, as opposed to individuals. As more and more cases were decided on the basis of statistical and other scientific demonstrations of risk, tensions around the admissibility of scientific evidence heightened. A series of Supreme Court decisions led to judges having to ‘think like scientists’ about the admissibility of evidence.
The NVICP also increased scrutiny of evidentiary standards, in light of the OAP’s consolidation of 5600 individual cases into three test cases, and the controversial Poling decision. The program uses the ‘Daubert standard’, which requires that evidence in proposed testimony be supported by theory that is
1. tested and is falsifiable; 2. peer-reviewed; 3. generally accepted within the scientific community; and, 4. the error rate must be provided, if known. (Daubert vs. Merrell Dow Pharmaceuticals, Inc, cited in Jasanoff, 1995: 63)
According to Kirkland (2013), because vaccine and mercury critics published evidence that aimed to meet the Daubert standard criteria, the plaintiffs had access to evidence that only fell short of the criterion of ‘general acceptance within the scientific community’. The court’s own standards were sufficiently open that the Special Masters could choose to be more lenient, or to look closely at ‘contests between experts’ (Kirkland, 2013). The Special Masters thoroughly scrutinized the credibility of the expert witnesses and the veracity of the scientific evidence presented. That is, they consistently ‘assess(ed) offers of expert testimony using the same principles that scientists themselves would use under the circumstance’ (Jasanoff, 2002: 46). Thus, the OAP hearings strictly complied with the Daubert standard of evidence and that necessarily entailed battles over legitimacy (Kirkland, 2012).
The NVICP’s shifting standards of evidence made the OAP an ideal site of struggle in the field of autism science. Ambiguity of admissible scientific evidence enabled both sides to attempt to impose definitions of ‘good science’, with each side wielding scientific capital to uphold its own interests. However, no one in the OAP hearings disagreed on scientific doxa. As we will show, the debate was over the ontology of autism.
DAN! and situated knowledge
One of the key struggles in the OAP was over the nature of autism itself, since causation and intervention strategies reflect the ontology ascribed to it by the plaintiffs and the respondents. To understand how these contestations over ontology played out in the OAP, we consider the historical formation of the field of autism science from the plaintiffs’ position. As we will illustrate, the ontology to which the plaintiffs subscribed was established long before the hearings. We explain the formation of this alternative standpoint on autism as part of a response to mainstream science’s attempts to delegitimize parental knowledge about the timing of the onset of autism. This view of autism reflects the parents’ commitment to subjective and situated knowledge (Silverman, 2012), in general, and a belief about a regressive subtype of autism, in particular.
Bernard Rimland, a psychologist with an autistic son, was the first to promote the idea of parent-researchers playing a primary role in the development of autism science. He championed the idea of parents gathering data on different treatments they had tried on their children and sharing this information with scientists and other parents (Eyal and Hart, 2010; Silverman, 2012). Rimland formed DAN! in the mid-1990s, with physician Sydney Baker and parent-researcher Jon Pangborn (Eyal and Hart, 2010; Silverman, 2012). DAN!’s goal is to bring parents and researchers together to share knowledge and resources about autism; its members believe that autism is a treatable – if not wholly curable – syndrome (Silverman, 2012). DAN! is one of the only parent groups that actively attempts to pioneer autism treatments (Silverman, 2012). Eyal and Hart (2010) call DAN! an agonistic network that actively blurs the boundaries between mainstream and alternative medicine. Its treatments depend on a network of providers and laboratories, which offer testing of methods (Silverman, 2012).
Rimland became convinced in 1968 that vaccines were associated with autism (Eyal and Hart, 2010). At DAN! meetings, he reached out to other parents, asking whether they too believed vaccines had caused their children’s autism (Silverman, 2012). In partnership with scientists and physicians, the parents developed and deployed a framework of observation and intervention in their children’s lives. As Eyal et al. (2010) write,
In the vaccination controversy one discerns what is probably Rimland’s most significant legacy: an alternative network and model of research that is a direct challenger, competitor, enemy and detractor of experimental science’s ethic of ‘virtual witnessing’ – an alternative network that attempts to subordinate virtual witnessing to the ethical vocation of autism parenting. (p. 238)
This formation enabled DAN! to champion alternative research and theories of disease associated with the vaccine–autism link. DAN! physician Jacquelyn McCandless (2005) says,
Members of DAN! are bonded by their belief that autism and associated ASDs are primarily medical disorders with behavioral and cognitive impairments that are by-products of the physical illnesses these children suffer … these are physically ill children with real medical disorders. (p. 4)
Thus, from its inception DAN! had its own fully developed ontology of autism.
Within DAN!, political advocacy plays a ‘secondary role at best’ (Silverman, 2012: 180). Thus parents involved in DAN! have tended to affiliate with other, more directly political organizations (p. 207). For example, some of these parents have been either in contact with or were members of the Coalition for SafeMinds (Sensible Action For Ending Mercury-Induced Neurological Disorders), an organization that motivates parent-researchers to connect with politicians (Silverman, 2012). SafeMinds took an active role in the controversy early in the 2000s, especially in relation to the Congressional hearings initiated by Representative Dan Burton, who supported the vaccine–autism link (Silverman, 2012). SafeMinds also criticized the National Institute of Mental Health (NIMH) and the CDC for restricting parent-researchers’ access to vaccine data. SafeMinds wanted to use data from both the Vaccine Adverse Event Reporting System and the Vaccine Safety Database, data that NIMH and CDC argued were inappropriate or confidential. DAN! members Dr Mark Geier and his son David Geier 8 used Vaccine Adverse Event Reporting System data to publish one paper on the link (Silverman, 2012), on which the plaintiffs relied heavily in the OAP.
The OAP plaintiffs’ medical evidence, discussion of treatments, and theories of autism line up with those of DAN! and SafeMinds. The medical tests the plaintiffs offered in both test cases met DAN!’s standard of evidence-gathering, which draws from Rimland’s legacy as parent-researcher. The parents were connected to the lab in the Cedillo case, UniGenetics, through DAN! and other parent networks. Many of the experts for the plaintiffs testified that they had presented at DAN! conferences. The discussion of treatments in both cases relies on proof of efficacy gathered according to DAN!’s alternative standard. The type of parenting described in the trials reflects what Silverman (2012) calls ‘intensive motherhood’ or ‘Autism Mommies’. The parents report many hours of investigation, self-education, and advocacy for their autistic children, learning about the most recent treatments and therapies, and seeking out providers to aid them.
Parents associated with DAN! began to question and challenge mainstream autism science, in light of what they perceived as a unique ‘regressive’ subtype of autism, which they argue is a distinct phenotype from ‘classical’ autism (with potentially different pathology and biomarkers) and which causes the loss of language and social skills previously acquired by children. These parents claim that their children began life developing normally and then suddenly (usually in concert with a battery of vaccines) began to lose skills (e.g. the ability to speak, to respond to their own names, to make eye contact, etc). As Silverman (2012) notes, a significant factor in the OAP was that these claims privilege subjective, experiential knowledge acquired from raising their children on a daily basis. As governmental experts privileged population-based statistical knowledge (p. 198), DAN! parents remained uncompelled by genetic explanations of autism:
[P]arents who nonetheless believe that vaccines caused their child’s autism are not impaired, in denial, or unable to comprehend statistical correlations. Their individual experiences and observations contradict the claims of scientists and the evidence against an association between immunizations and autism. (Silverman, 2012: 200)
As we will show in the next section, parents’ participation in DAN! and other counter-hegemonic scientific movements produced an influential ontology of autism. Importantly, this ontology informed the plaintiffs’ emphasis on treatment as a cure and reflected the parents’ hope of recovering the children they knew prior to the autistic regressions.
Ontology
In this section, we explain each side’s presentation of the ontology of autism, in terms of etiology, diagnostic criteria, morphology, recommended treatments, and prevalence rates. We have reproduced these side-by-side in the following table for the sake of comparison (Table 2).
Comparison of divergent ontologies of autism.
The plaintiffs believe the causes of autism are complex and multiple. As their expert witness Dr Marcel Kinsbourne explains,
There are none and there are too many [causes of autism]. It is as complex a situation as anyone might find in medical science … I think most people working in the field of autism recognize that autism has many, many causes. (Cedillo trial, Day 5)
In general, the etiology presented by the plaintiffs suggests that, although a child may have a genetic predisposition to autism, it is generally ‘triggered’ by either pre- or post-natal environmental factors. Consider the following excerpts from the OAP transcripts:
Ms Chin-Chaplin (attorney for the plaintiffs): Doctor, when it’s thought that there’s a genetic susceptibility to developing autism, is that an indication that every child will develop autism? Dr Marcel Kinsbourne: No, on the contrary. There’s little doubt that some children happen not to encounter the provocative or triggering situation, whatever it is, and there’s a wide range of possibilities … [of not] becoming autistic. (Cedillo trial, Day 5) Mr Williams (attorney for the plaintiffs): First of all, there’s no such thing as a genetic epidemic. If autism was all genetic, you wouldn’t see a change in rates. You can only see an increase if something is triggering it, so it’s an interaction between the environment and the genetic susceptibilities of these children. (Opening statements, King/Mead trial, Day 1)
Such emphasis on genetic vulnerability to post-natal environmental insults is the foundation for the plaintiffs’ claim that vaccines cause autism. 9
The plaintiffs presented two different theories of causation in the different trials. The first (in Cedillo, Snyder, and Hazlehurst), built on Wakefield’s work on gut inflammation (Wakefield et al., 1998), suggests that thimerosal suppresses the immune response to the MMR vaccine, thereby reducing resilience to the measles infection, which builds in the gut and causes inflammation and developmental regression. The second theory (presented in the Dwyer and King/Mead trials) argues that thimerosal causes neuro-inflammation, through a build-up of mercury in the brain, which causes over-excitation. Chelation therapy is recommended to clear the mercury and consequently the neuro-inflammation. Despite the differences between these theories, the plaintiffs clearly view autism as at least partially an acquired illness.
As Annemarie Jutel (2009) has argued, there is often disjuncture between experiences and diagnoses, which can prevent therapeutic alliance across explanatory frameworks (p. 289): while experiences are personal and cultural, diagnoses are interventions by medical authorities. Indeed, the parents involved in DAN! had experiences with their children’s illnesses that were publicly delegitimized by medical authorities with competing ontological commitments. While the plaintiffs’ experts tended to agree with the respondents’ experts about the basic constellation of autism symptoms, the former emphasized that every autistic child is unique. According to this approach, diagnosis should be tailored to the individual child: ‘The fact is we look at individuality of the patient as our clinical approach, and one of our concerns is that children with autism, they are a very heterogeneous population’ (Dr Mumper, King/Mead trial, Day 4). This view echoes a central tenet of DAN! that ‘each child is biochemically unique’, and so, ‘there is no one-size-fits-all autism treatment’ (Eyal and Hart, 2010: 24).
For the plaintiffs, autism is an illness, as opposed to a disability. They believe autism is acquired, and therefore potentially curable. On this view, autistic children suffer a myriad of symptoms of their illness, which include gut dysfunction, yeast infections, allergies, sleep disorders, aggression, ‘self-stimming’ seizures, and hyperactivity (Eyal and Hart, 2010; Silverman, 2012). A large portion of trial testimony from Dr Arthur Krigsman deals with the existence of autistic entercolitis, a condition that plaintiffs’ experts argue explains the high prevalence of bowel problems in autistic children. Treatments address the underlying causes of autism, such as chelation and experimental drug therapies related to neuro-inflammation, as well as dietary and hormonal suppressants. Plaintiffs believe that an autistic child’s body is out of homeostasis, which means that a child with autism can be ‘recovered’ if a parent can hit on the precise treatment that will cure his or her own unique brand of autism (Eyal and Hart, 2010). Recovery requires a
delicate fine-tuning of chemical balance which can only be achieved in individual cases and only through trial and error … The result is a fuzzy zone where established and alternative medicines are indistinguishable, and the expertise of doctors is not superior, but additive to the expertise of the parents. (Eyal and Hart, 2010: 28)
The plaintiffs’ lawyers represent the work of parents and their medical allies as helping children ‘fight their way back’ from the ‘empty fortress’ of autism 10 (Mr Powers, Attorney for the plaintiffs, King/Mead trial, Day 1; see also Eyal and Hart, 2010: 28).
The plaintiffs’ expert witnesses explicitly argued that the recent sharp rise in autism is linked to greater variety of post-natal environmental insults, like childhood vaccines. They rejected the view that diagnostic substitution alone can account for the rise in prevalence of autism, which they construed as an epidemic:
Dr Mumper (expert for the plaintiffs): [W]e are trying to take care of a generation of children and being overwhelmed by their medical problems … we feel some urgency that we can’t wait for 10 or 20 years. These children seem to have a window of opportunity where if you treat their medical problems they get better. And with the timeline of applying for grants, getting the studies completed and analyzing the results … we are proceeding in good faith, using our best clinical judgment, realizing that we don’t have good case control studies for all that we do. (King/Mead trial, Day 4)
Such calls for development of a cure for autism conveyed a sense of urgency, which was framed in tension with the bureaucratic approaches of mainstream science.
On the other side of the courtroom, the respondents’ ontology of autism reflected a mainstream scientific episteme. Their experts repeatedly testified that autism is linked overwhelmingly to genetic risk. Drs Fombonne, Edwin Cook, Robert Rust, and Michael Rutter all claim that ‘[a] genetic contribution is postulated to be involved in perhaps 90 percent’ (Dr Rust, King/Mead trial, Day 8) or ‘the estimates are that about 90 percent of the liability to autism is genetically influenced’ (Dr Rutter, King/Mead trial, Day 12). When the respondents concede that environmental factors can be implicated in the development of autism, these are almost all pre-natal insults:
Dr Rutter (expert for the respondents): Height also has a heritability of about 90 percent … Between 1900 and about 1950 the average height rose by approximately 12 centimeters. That’s a big rise. We don’t know for sure what it’s due to, but it’s almost certainly due to improved nutrition and partly also to a reduction of the impairments caused by infections. So here is an example of something which is highly heritable, but nevertheless a major environmental factor could and did make a difference. Ms Ricciardella (attorney for the respondents): If there were an environmental influence, speaking to the heritability of a liability to autism, when in the course of development would that influence occur? Dr Rutter: It’s likely to be in the prenatal period. It could be I suppose in the very early postnatal, but the evidence suggests prenatal is more likely. (King/Mead trial, Day 12)
In the above exchange, the respondents’ witnesses wielded non-knowledge, suggesting that the relevant non-knowledge is around the topic of which genetic markers are connected to autism, as opposed to the topic of how much genetic makeup influences the acquisition of autism. We will return to this point in the next section. Moreover, estimates of 90 percent genetic causation are based on a twin study from 1977 (by Folstein and Rutter, 1977), which had small samples (n = 21) obtained through referral (Eyal et al., 2010: 14). In this case, the respondents’ witnesses rely heavily on a questionable study published 30 years earlier, a reliance similar to ones that led to plaintiffs’ witnesses losing credibility. Indeed, the 90 percent heritability rate is almost certainly inflated: a more recent study drawing on a larger, randomly sampled population estimates heritability at 19 percent for males and 63 percent for females (Liu et al., 2010). Eyal et al. (2010: 14) suggest that it may be impossible to know how much of autism is genetically heritable and that it might be a moving target.
Although the respondents’ experts do admit that children present symptoms of autism somewhat differently, they consistently present it as a developmental disability characterized by language deficits as well as a social and repetitive behaviors, and insist that diagnostic, morphological, and treatment standards should be stable. Drs Fombonne, Cook, and Rust provide in-depth descriptions of the process of diagnosing autism, which do not recognize an etiological relationship to other conditions, such as autistic entercolitis. Disputing the plaintiffs’ claims of causal links between autism and other conditions, the respondents argued that, like all medical conditions, autism can co-occur with other conditions without their being etiologically related. Most of the experts also dismissed the notion of regressive autism as a distinct phenotype. They claimed that regression is fairly common, insisting that few regressive children start out completely ‘normal’ and arguing that post-regression outcomes are identical to non-regression outcomes. In the King/Mead trial, Dr Rutter provided a lengthy discussion of why regression is a common, but not a distinct, phenotype. 11
To some extent, the idea that autism is an epidemic was on trial, and the transcripts reflect a great deal of debate about the increasing prevalence of autism. The respondents’ experts argue, almost uniformly, that the rapid rise in autism in recent decades is due to diagnostic substitution:
Dr Fombonne (expert for the respondents): [There] could be evidence that there is an increased incidence of the disorder, but it could also reflect different factors … One of them has to do with diagnostic concepts and diagnostic criteria which have changed over the last 40 years … [I]t’s clear that the change in concepts and then how they are reflected in diagnostic systems have an impact on who meets criteria or not … (Cedillo trial, Day 11) Dr Rust (expert for the respondents): [V]irtually every week … a patient comes into my clinic … for cerebral palsy or comes for mental retardation or some other condition and it’s as plain as the nose on my face that these individuals have autism because I know what [it] looks like … So this is one of the most important explanations we have for what has appeared to us to be an increase in the prevalence of autism, but not an increase in the incidence of autism. (King/Mead trial, Day 8)
These experts vehemently reject the language of ‘epidemic’ in relation to autism. They frame autism as a disability, which can be manageable if discovered early and treatment is undertaken cautiously, with clinically tested therapies. Dr Rust describes children with autism as suffering from ‘isolation’ due to ‘communication difficulties’ that only educational and behavioral therapies can bridge:
Dr Rust: What we find is that what might appear to be anxiety or other things are so readily alleviated when a child is placed in an educational setting where there’s understanding on the part of the educators … But what a child with autism may have … we don’t know for sure because we can’t ask … (King/Mead trial, Day 8)
As such, treatment becomes a matter of translation, which can only be performed by those with proper training in educational and behavioral therapy.
Differences between the plaintiffs’ and respondents’ approaches to issues of etiology, diagnosis, morphology, treatment, and prevalence of autism reflect competing ontologies of autism. For the plaintiffs, autism is a curable illness associated with increasing exposure to environmental risks, such as vaccines or mercury. For the respondents, autism is a genetic disability that can be better managed through Applied Behavioral Analysis (ABA) therapy. On the respondents view, better awareness has led to more children being diagnosed earlier in life, which explains the recent rise in autism rates. For both groups, the extent to which autism is a knowable, manageable, and curable/treatable illness or disability is a function of ontology. These ontologies influence how each side frames non-knowledge in autism science, with the respondents foreclosing the possibility of environmental causation in favor of the unexplored science of genetics.
Genetic causation and hegemony
Dr Cook (expert for the respondents): It’s strongly genetic. It’s not purely genetic. Ms Chin-Caplan (attorney for the plaintiffs): Strongly genetic but not purely genetic. There’s environmental influences then. Dr Cook: Well, we don’t know what the other 8 percent is. (Dr Cook, Cedillo, Day 6)
Experts for the respondents attempted to establish genetic explanation as hegemonic in autism science. Their testimony circumscribed the unknown in autism science to questions about which specific genetic markers require further research. In this way, environmental causes were delimited to a space of negative knowledge – non-knowledge that is not worth further investigation. As the epigraph for this section demonstrates, while experts for the respondents admitted that current genetic research cannot explain 100 percent of autism causation, they were unwilling to consider environmental explanations:
Ms Renzi (attorney for the respondents): Dr. Rust, petitioners have put forward a hypothesis that thimerosal-containing vaccines cause autism. What is your number one reason that that hypothesis will be proven wrong? Dr Rust (expert for the respondents): That we’re coming to some considerable understanding of actually what happens in autism and why this terrible illness develops in children. This has been the progress of the decade of the brain and the very excellent work of a number of developmental neuroscientists. So we’re understanding this and other disorders with regard to problems in the working out of a genetic code for the development of the brain. (Hazlehurst, Day 3)
Here, Dr Rust relies on the established scientific record on autism to indicate that ‘problems in the working’ of the genetic code are responsible for the causation of autism, rather than environmental insults. He presents a progressive view of science, one that builds toward certainty, correcting erroneous conclusions as uncertainty is replaced by calculable risk. When he does concede to non-knowledge, he does so strategically to direct it toward genetics:
Dr Rust: As far as heritability is concerned, although again all the details aren’t worked out, this is strikingly one of the most heritable severe neurological disorders that we’re aware of. It may occur on the basis of quite a long list of genes … This is still something of a work in progress, but the concordance between identical twins is quite high and between siblings a reasonably high risk also for autism unfortunately … A great deal of jumping to conclusions regarding behavior of humans has been made in medicine and science for a long time … The point is that it’s clear that there’s a heritable aspect to this that this quite striking and supports the idea that … the thing that has to be there, is a genetic defect and other things are not necessary for it to express itself in that way. (Hazlehurst, Day 3)
Such claims about genetic causation are overstated, as we mentioned in the last section. Dr Rust does not admit uncertainty about how much genetic influence there is in autism causation, dismissing claims about environmental causes, specifically claims about vaccines and thimerosal, as ‘jumping to conclusions’, detailing the reasons why genetics is a more promising and certainly more scientific avenue of research. The quote above represents very clearly the struggle to define ‘good science’ in the OAP. By implying that the plaintiffs were illogical, that their evidence was unscientific, and that they were ‘jumping to conclusions’, Dr Rust portrays the respondents as representatives of good science. He relies on his scientific capital to wield genetic causation as the only relevant area of non-knowledge important in the field of autism science.
In the Cedillo case, Dr Fombonne similarly limits relevant non-knowledge around autism to the field of genetics. His testimony first takes issue with the notion of an autism epidemic and then moves into the plausibility of an environmental cause. On cross-examination, he again relies on existing scientific research, in this case twin studies, to shore up hegemony around genetics, although scientific consensus seems to be that how much autism is genetic is still somewhat uncertain:
Dr Fombonne: I think in terms of the etiology of autism what we know is that genetic factors play paramount contributions to the development of autism as estimated by heritability estimates from twin studies … which indicates that 92 percent of the population’s variance of this heavily defined disorder are accountable to gene effects, okay? Mr Powers (attorney for the plaintiffs): Right, yes, but not all of it. Not all of it. There’s no evidence to support that autism is entirely a genetic collection of symptoms? Dr Fombonne: Well, it’s actually a bit more complex than that. I mean, if you had 100 percent of heritability, yes, everything would be ascribed to gene effects of one sort or the other. The fact that the heritability estimates are slightly lower than 100 percent, you can see this discrepancy as two ways. It could be measurement error … and that we don’t capture the whole range of the phenotype … but you could as well say then maybe there is some kind of room for environmental factors. If that is the case the question is what it is. It’s not very likely in the sense that … fraternal twins have more of the same environment than siblings because they grow up in the same pregnancy, same womb, et cetera. So we would predict that if there were environmental factors, particularly those occurring during pregnancy, they are more shared by fraternal twins than by siblings. Therefore, you would expect that the risk in … fraternal twins, for autism would be slightly higher than there is for siblings from different pregnancies. In fact, it’s not the case. So it’s a way to look at environmental factors and their role in the genetics of autism, and the data are not very supportive of that. (Cedillo, Day 11)
Here, Dr Fombonne attempts to shut down debate over the extent to which autism is genetic and to foreclose almost all possibility that the causes of autism could involve environmental factors. He does not dismiss the fact of uncertainty in autism research, but instead frames the task of investigating the unknown entirely in terms of measuring genetic factors. Grappling with uncertainty about causes of autism, then, is a matter of measurement, and science’s task is to develop appropriate measurement tools.
Dr Cook’s testimony more clearly illustrates how the respondents used non-knowledge about genetics as a discursive tool to exclude the possibility of environmental explanations from autism science. His detailed testimony focuses on specific genes. For example,
Dr Cook (expert for the respondents): We are looking at the fifth area of the genome because as I mentioned before, the complexity of autism includes multiplicative inheritance with multiple genes conspiring, interacting to cause autism. That is what the twin data suggests when you go up from a 25-fold increase to an over 300-fold increase going from siblings to monozygotic twins. So we expect multiple variants contributing across the genome, which is very similar to what’s expected in Type I diabetes and Type II diabetes, asthma and other disorders. (Cedillo, Day 6)
Dr Cook’s testimony is full of detail about the areas of genetic code he and other scientists are currently investigating, looking for the markers for autism. His testimony clearly frames non-knowledge about autism in terms of genetics. He admits that there is still a great deal of uncertainty in the field of autism, but he speaks of it as uncertainty about genetic code. In doing so, he sets up a division between two types of non-knowledge: first, there is the extended knowledge – that which science should pursue – of genetics. Second, there is the negative knowledge – that which it should leave undone – of potential environmental factors in cases of autism.
The respondents’ arguments convinced the Special Masters both that genetic causation is the best available explanation of autism and that genetics is the most promising area for future autism research investment. In his decision in the Cedillo case, Special Master George Hastings rejected the plaintiffs’ arguments about environmental causation, stating instead that science has shown that autism is ‘strongly genetic’: ‘some non-genetic factors have been identified and accepted as factors in causing autism, but those consist of exposures during the early prenatal period’ (Cedillo decision, emphasis original). While Hastings admitted that science is ‘open to evidence of the possibility of postnatal factors playing a causal role’ in autism, he reaffirmed the respondents’ argument that ‘we know for certain that genetic factors and prenatal exposures can cause autism’ (Cedillo decision, emphasis original). In this way, the Special Masters affirmed the respondents’ deployment of non-knowledge, suggesting that genetics is where future research should focus. With these boundaries around non-knowledge, the respondents protected their scientific capital and defined the rules of science in their own interests.
Conclusion
Dr Mumper (plaintiff expert): But I think obviously the science is in its infancy here. (King/Mead trial, Day 5) Dr Rust (respondent expert): We are living in the golden age of science. (King/Mead trial, Day 8)
These two statements perhaps best capture the differences between the plaintiffs’ and the respondents’ views on scientific uncertainty and the ‘doneness’ of autism science. The respondents attempt to circumscribe their certainty about autism and their scientific expertise on the subject:
Ms Ricciardella (attorney for the respondents): Unlike Petitioners’ experts who broadly speculate about an unlimited universe of scientific possibilities, Respondent’s experts root their opinions in decades of meticulous, specialized research. (Opening statements, King/Mead trial, Day 1, emphasis added)
However, the plaintiffs constantly challenged the respondents to open up and reflect on both the possibilities of uncertainty and also on the non-knowledge that blurs the margins of the two groups’ orbits of certified knowledge. The plaintiffs charged the respondents with manufacturing certainty by putting on blinders to unanswered questions, and the respondents replied by shutting the black box. As the respondents’ attorney argued, the science is done:
Ms Ricciardella (attorney for the respondents): There is no scientific debate. The debate is over. There’s no scientific controversy. The only controversy is the media controversy … The credible scientific community has already spoken on this issue and has rejected it. (Opening statements, King/Mead trial, Day 1; also quoted in Silverman, 2012: 223)
As an expert witness for the respondents, Dr Rust stated that, although autism is a complex disease, scientists are getting better at clinical descriptions and progressively weeding out bad hypotheses:
Dr Rust (expert for the respondents): We do careful analyses of the increasingly abundant literature on these subjects and then we execute the well designed scientific investigations … we live in a golden age of science … Hypotheses are a dime a dozen. Anybody can make up an idea about what’s going on … We need to do the experiment and see whether we can either refine that experiment or abandon that hypothesis based on those conclusions. (King/Mead trial, Day 8)
On this account, while the respondents achieved their analyses of autism through gold standard methods and sound scientific testing, the plaintiffs are tilting at windmills – stuck, ideologically, on a theory of causation for which there is no evidentiary proof. Genetic questions are the only area of non-knowledge worth pursuing.
The plaintiffs, on the other hand, claimed that the science of autism is contentious:
It’s not scientific certainty because, frankly, the science is in dispute … that’s what you’re going to hear in all the test cases. The science is in dispute, and this issue is not scientific certainty. It’s more likely than not on a balance of the evidence. (Mr Thomas Powers, attorney for the plaintiffs, opening statements, Cedillo trial, Day 1)
As such, the plaintiffs claimed that the black box of autism should not be closed: non-knowledge in autism research is pervasive, and it lies in multiple areas:
Mr Tom Powers (attorney for the plaintiffs): The government’s own science, by the way, is still a work in progress. (Opening statements, Cedillo trial, Day 1) Mr Powers: The story that you will hear in this hearing, the presentation and the evidence, is a scientifically sound description about ongoing story, and we don’t know what the end is going to be, but we do know that where we are today is that we have a mechanism of injury that meets the standards of causation in this program, and there is medical evidence that these two children satisfy this burden and are entitled to compensation. (Opening statements, King/Mead trial, Day 1; our emphasis)
Following Gross (2007), we have argued that part of the dispute here is between frameworks of risk versus uncertainty. The respondents claim that the probabilities of risk are well researched and known. The only question of risk concerns which genes cause autism and which epigenetic factors influence genetic expression. The plaintiffs, however, operate under a rubric of uncertainty in which the probabilities are unknown and non-knowledge is unspecified.
The debate, then, is partially about the ‘doneness’ of autism science, hinging on what Bourdieu (1991) refers to as the ‘space of possibles’ within a field of objective potentialities. Some possibles will be actualized and others will be ignored, which is the result of a field struggle between agents differentially imbued with scientific capital.
The possibilities offered to agents in a field
depends on the relations of power between the agents and institutions that, having an absolutely vital interest in this or that of the possibilities put forth as instruments or stakes in the struggles for the ‘legitimate problematic’, strive with all the means and powers at their disposal to see that those possibilities are actualized that best suit their dispositions and their position, and thus, their specific interests … Struggle is established between agents who are unevenly endowed with specific capital. (Bourdieu, 1991: 12)
Therefore, what is ‘possible’ is not only the result of struggle but is determined by both the state of the field – including views on the ‘doneness’ of different research trajectories – and the resources that different players can mobilize toward their interests.
Footnotes
Acknowledgements
We would like to thank the following people for their very helpful feedback and input along the way: Gil Eyal, Anna Kirkland, Robert Johnston, Stefan Timmermans, Mariana Cranciun, Owen Whooley, Claudio Benzecry, and Matt Mahler. We also appreciated the anonymous reviews and editorial assistance we received from Social Studies of Science. A special thank you must also be extended to Paige Sweet for her research assistance.
Funding
This project was funded by the Office of Social Science Research and the Research Grant Support Program at the University of Illinois at Chicago.