The Panic Virus
Page 16
While no specific legislation came out of Burton’s hearings, they served as yet another potent illustration of the connective power of the online world. In her prepared remarks, a Louisiana housewife named Shelley Reynolds told of how she’d used the Internet to make contact with “thousands and thousands of parents from across the country,” all of whom had identical stories to her own:
Child is normal, child gets a vaccine, child disappears within days or weeks into the abyss of autism. If you doubt me, I invite you to attend the Hear Their Silence rally on the [National] Mall, where our Open Your Eyes project will be displayed. View the thousands of pictures we have collected and realize that forty-seven percent of those who participated believe that vaccines contributed to the development of their child’s autism. Parents like me are relying on you. . . . I know my children. I know what happened to my son. As far as I am concerned, the needle that silently slipped into my baby’s leg that day became the shot heard round the world.
* * *
It was 1993 when the son of Albert Enayati, an Iranian Jew whose family had immigrated to the United States in the 1970s, was diagnosed with autism. At first, Albert was so despondent he spent his nights weeping. By the end of the decade, he’d become a dedicated activist who led the New Jersey chapter of CAN. It was through that network that he first met Sallie Bernard, and the two bonded over their anger about the presence of thimerosal in vaccines. In 1999, shortly after he was laid off from the multinational drug manufacturer Pfizer, Enayati dedicated himself to collecting research that confirmed his suspicions. In 2000, just as Burton began his hearings, Bernard, Birt, Enayati, and Redwood founded the Coalition for Sensible Action for Ending Mercury-Induced Neurological Disorders, or SafeMinds. One of the group’s first goals was to publicize the results of the research Enayati and the others had been conducting. Bernard, the Harvard-educated marketing consultant, was adamant that the best way for the group to disseminate its message was to get its work published in a medical journal. In order to be taken seriously, she explained, they had to speak “their language—scientists aren’t going to read anything unless it’s written in scientific jargon.” That summer, the leaders of SafeMinds completed “Autism: A Novel Form of Mercury Poisoning.” Its summary read:
A review of medical literature and US government data suggests that: (i) many cases of idiopathic autism are induced by early mercury exposure from thimerosal; (ii) this type of autism represents an unrecognized mercurial syndrome; and (iii) genetic and non-genetic factors establish a predisposition whereby thimerosal’s adverse effects occur only in some children.
While the result successfully aped the abstruse style unique to academia—one passage read, “These findings may be linked with HgP because: (a) Hg preferentially binds to sulfhydryl molecules (-SH) such as cysteine and GSH, thereby impairing various cellular functions; and (b) mercury can irreversibly block the sulfate transporter NaSi cotransporter NaSi-1, present in kidneys and intestines, thus reducing sulfate absorption”—the team of parents was unable to convince a respected academic journal that its work deserved an audience. That did not mean the paper would not appear in print: In December 2000, it was accepted by Medical Hypotheses. The New England Journal of Medicine this was not: Medical Hypotheses proudly eschewed peer review, a process it said disapprovingly “can oblige authors to distort their true views to satisfy referees.” In the “Aims and Scopes” section of its guidelines to writers, the journal emphasized that it had no desire to “predict whether ideas and facts are ‘true’ ”—in fact, it was eager to print “even probably untrue papers” so long as they spurred discussion.
“Autism: A Novel Form of Mercury Poisoning” was published in the spring of 2001. The crux of the paper’s argument could be found in a table comparing ninety-four traits that it claimed were characteristic of both autism and mercury poisoning. The dozens of citations—“ASD [autism spectrum disorder] references in bold; HgP [mercury] references in italics”—may have looked impressive, but the list was little more than a meaningless grab bag of attributes. A number of the “symptoms” the authors had ascribed to autism—temper tantrums, unprovoked crying, grimacing, sleep difficulties, rashes, itching, diarrhea, vomiting—were, as any parent could attest, also “symptoms” of infancy and early childhood. In places, the traits listed to illustrate similar symptomologies were actually contradictory: The “physical disturbances” heading included both diarrhea and constipation, “unusual behaviors” listed both agitation and staring spells, and “psychiatric disturbances” included both mood swings and flat affect. When it came to specifics, some symptoms were compared seemingly for no reason other than that they both affected the same set of behaviors: repetitive behavior and impaired speech for autism, loss of muscle control and slurred speech for mercury poisoning. Other times the reactions the paper listed were directly opposite to each other: increased sensitivity to sound was cited as a common symptom of autism, mild to profound hearing loss for mercury poisoning.
The rest of the paper was marked by the conspicuous omission of anything that might contradict the authors’ thesis in any way. Rather than address the known differences between ethylmercury and methylmercury—differences that were one of the main points used by those who argued that thimerosal was safe—the paper acted as if no such difference existed. (The only time either “ethyl” or “methyl” appeared was in the final sentence of the second paragraph: “The mercury in vaccines derives from thimerosal (TMS), a preservative which is 49.6% ethyl-mercury (eHg).”) There was no acknowledgment of the total absence of reports of autism in every previous study of either ethyl-or methylmercury poisoning, nor was there an acknowledgment that the one published study that had looked for elevated mercury levels in children with autism had not found any such evidence. Finally, and most significantly, there was the paper’s failure to provide even a single example of “detectable levels” of mercury in an autistic child. The paper’s authors tried to explain this omission by referring back to the unproven conjecture that had launched the project in the first place: “parental reports of autistic children with elevated Hg.”
All in all, the Medical Hypotheses paper, like almost all efforts undertaken for the sole purpose of proving what someone already believes to be true, was a tutorial in bad science. Instead of amassing evidence to prove a conjecture, it used the conjecture itself as evidence. The absence of supporting data was cited as further proof of the authors’ detractors’ failings: If tradition-bound doctors and researchers had given proper credence to personal anecdotes and untested observations, they would already have identified the subset of children that was uniquely susceptible to this “novel” form of poisoning. In the final evaluation, the paper’s title may have been more accurate than its authors intended: Autism was a “novel” form of mercury poisoning only in that it was entirely fictional.
32 The database’s official name—the Internet’s Grateful Med—belied the degree of expertise needed to navigate its contents. Its instruction page told users to start by entering a search term, after which they were directed to “click on the ‘Find MeSH/Meta Terms’ button located at the top of the page. IGM then displays the Metathesaurus Browser Screen. The IGM algorithm displays the MeSH terms that are the nearest to the search terms (the Ovid ‘mapping’ function provides quite similar results). Click on the best MeSH term.” And so on.
33 The inherent problem here is the tendency, conscious or otherwise, to define your population in such a way as to achieve a predetermined result—as occurs when Democrats and Republicans carve out ever-more serpentine congressional districts in order to protect their incumbents.
34 In order to understand why this is true, it’s important to keep in mind the role randomness plays in all areas of life. For example: Both of the boys who lived on my block and were in my school year when I was growing up were named Seth. This does not mean that one hundred percent of the male population is named Seth, or that one hundred percent of the population in my grade in
school were named Seth, or even that boys who lived in my town and were in my grade were at an elevated risk for being named Seth. This demonstrates one of the major hurdles in proving causality: In order to demonstrate that the apparent “Seth cluster” on my block was not just the result of random noise, you’d need to explain how and why it occurred and not simply demonstrate that it had occurred.
CHAPTER 12
THE SIMPSONWOOD CONFERENCE AND THE SPEED OF LIGHT: A BRIEF HISTORY OF SCIENCE
The Salk vaccine trials in the 1950s and the subsequent eradication of polio in the United States proved that with enough money, resources, and willpower, government agencies and private businesses could join forces to produce massive change. In the decades since, no comparably proactive effort has been made to safeguard public health. Instead, it has taken a controversy or a catastrophe to precipitate most advances, from the monitoring of drug safety to the implementation of reporting and compensation systems for vaccine-related injuries.
For example, it wasn’t until 1955, in the wake of the Cutter Incident, that the National Institutes of Health established the Division of Biologics Standards, and it wasn’t until 1962, a year after the sedative thalidomide was shown to cause birth defects when used by pregnant women, that Congress passed a law that required drug manufacturers to prove their products were safe before bringing them to market. In 1974, when the parents of Anita Reyes were awarded $200,000 in their lawsuit against Wyeth Pharmaceuticals, sixteen years had passed since the verdict in Gottsdanker v. Cutter had shown the need to clarify culpability questions relating to vaccines—and still there was no nationwide policy in place governing drug manufacturers’ ultimate liability.
The lack of comprehensive regulations meant that all the parties involved, from patients and pediatricians to drug manufacturers and the federal government, were unsure of their rights and obligations under existing immunization mandates. The American Academy of Pediatrics was among the first groups to recognize this as a disaster in the making, and in the 1970s its members began to lobby for a uniform vaccine compensation system. Without such a mechanism, they warned, there was an ever-present risk that manufacturers’ liability fears could result in a drop in supplies and a spike in prices. During the swine flu scare in 1976, Congress had enacted a stopgap measure only after it became clear there was no other way to guarantee the production of enough vaccine to protect the country. Since then, lawmakers had continually professed concern about the need for a solution to the liability issue, and just as continually had found reasons to delay taking action.
Throughout it all, vaccine-related tort claims continued to rise, from twenty-four in 1980 to 150 in 1985, a five-year span during which total damages awarded to plaintiffs topped $3.5 billion. By then, the AAP’s warnings had proven to be prescient, as the number of companies willing to distribute vaccines in the United States fell from several dozen in the mid-1960s to just four. The panic that followed the 1982 broadcast of “Vaccine Roulette” prompted new promises of reform, but every time a deal appeared to be imminent, discussions between the various parties broke down. It wasn’t until 1986, when the departure of one of the two remaining suppliers of the DPT vaccine prompted the CDC to warn that the country was on the verge of running out of an adequate stockpile, that a compromise was finally reached. Just hours before adjourning for the year, Congress passed the National Childhood Vaccine Injury Act (NCVIA), which aimed to create a simple, transparent process for dealing with injury claims.
The central feature of the NCVIA was a uniform compensation program that would be funded by a surtax on each dose of vaccine. The program would be implemented by a group of federal judges designated as Special Masters, who would preside over a “Vaccine Court” that existed outside the traditional legal system.35 The law also established the Vaccine Adverse Event Reporting System (VAERS), which would funnel reports of the thousands of vaccine injuries claimed to have occurred each year to a single, centralized database managed by the CDC and the FDA.
VAERS was, without question, a major step forward in monitoring the efficacy of vaccines already on the market. Still, VAERS’s voluntary, post-facto reporting system meant that it wasn’t until after concerns about vaccine safety had been raised that investigations got under way. By always playing catch-up, the federal government left an opening for reporters like Lea Thompson, self-styled renegades like Andrew Wakefield, and impassioned activists like the Mercury Moms to raise some legitimately troubling questions—and when they did, they felt it was their right to decide which answers were satisfactory. In each one of those situations, the underlying concerns were valid ones: Could the DPT vaccine cause permanent harm? Could the MMR vaccine cause gastrointestinal problems or developmental disorders? Could thimerosal cause neurological problems? Without satisfactory answers, it was impossible to stop the spread of the most outlandish strains of vaccine skepticism from the fringes to the mainstream.
• • •
If there was one thing citizens, scientists, and safety monitors agreed on in the wake of the 1999 announcements about thimerosal, it was the need to analyze the data that was already available. An obvious first step was a comprehensive review by the Institute of Medicine (IOM), an independent organization devoted to “providing objective and straightforward answers to difficult questions” related to health and science policy. (The scientists who serve on IOM committees all do so on a volunteer basis.) From the outset, the IOM knew it needed to address the concerns of those most suspicious of vaccines if it wanted its work to have any validity. As a result, this report, unlike the four vaccine safety reviews the institute had conducted between 1988 and 1994, was mandated to address “the significance of the issue in a broader societal context” in addition to answering questions about scientific plausibility. As the process evolved, the committee’s efforts led to invitations to Sallie Bernard, Barbara Loe Fisher, and Lyn Redwood to participate in planning meetings and conference calls that would typically have been off-limits to members of the public.
An awareness of the snowballing anxiety surrounding vaccines had the exact opposite effect on a group of scientists who were asked by the CDC to review a massive collection of HMO patient records called the Vaccine Safety Datalink (VSD). The aim of that effort was to get vaccine experts to reach a consensus on what, if any, further research should be done to determine the possible side effects of the vaccines that had been administered over the previous decade. From the outset, the group’s members sought to conduct their work outside the public eye, mindful of previous instances in which complex science had been poorly translated for a general audience: Inevitably, it seemed, data was misinterpreted, preliminary results were assumed to be definitive, tiny sample sizes were blown out of proportion, and the mere act of conducting an experiment was taken as an admission that a given hypothesis was correct.
The first round of studies of the VSD data was completed in early 2000, and that June, a total of fifty-one specialists assembled for a closed-door, two-day meeting. The gathering took place at Simpsonwood, a United Methodist Church conference center on 227 acres of woodlands located about an hour’s drive north of Atlanta. Included among the dozens of participants were pediatricians, infectious disease specialists, and immunotoxicologists. Robert Chen and Frank DeStefano, the two CDC scientists who’d critiqued Andrew Wakefield’s 1998 paper for The Lancet, were there, as was John Clements, the top vaccine specialist at the WHO. Scientists from Aventis Pasteur, Merck, Wyeth, and SmithKline Beecham, the four pharmaceutical companies that continued to distribute vaccines in the United States, were also in attendance.
On the morning of Wednesday, June 7, Walter Orenstein, the head of the CDC’s immunization program, called the meeting to order. “We who work with vaccines take vaccine safety very seriously,” he said. “Vaccines are generally given to healthy children and I think the public has, deservedly so, very high expectations for vaccine safety as well as the effectiveness of vaccination programs.” The seriousness of the concern
s about a “possible dose-response effect” linking thimerosal in vaccines with “certain neurologic diagnoses” necessitated “a careful scientific review of the data.” The rest of the day, Orenstein said, would be given over to presentations relating to the VSD data, while the following day’s session would be devoted to free-ranging debate of the various recommendations of how to proceed.
The pivotal presentation of the conference was given by Thomas Verstraeten, a CDC epidemiologist who’d led a team that examined the cumulative level of ethylmercury exposure in 110,000 children when they were one month, two months, three months, six months, and one year old. Verstraeten opened his talk with a quip about how the invariably ambiguous results produced by studies that merged toxicology with environmental health made it “very hard to prove anything.” Because of that, he said, his study was the one “that nobody thought we should do.”
Even a cursory look at the potentially confounding factors that Verstraeten had to take into consideration showed the depth of the challenge he’d faced. Since there was no definitive data about ethylmercury’s half-life (the time it takes for half the amount of a given substance to be eliminated from the bloodstream), Verstraeten used previously established half-life figures for methylmercury in order to estimate the aggregate amount of thimerosal in infants’ bodies at any given point.36 Because of the privacy concerns inherent with using HMO patient records, he’d been unable to question (or even identify) the people whose data he was using—which meant he had no way of accounting for possible selection bias on the part of parents (it was possible that “the same parents that bring their children for vaccination would be the same parents that bring their children for assessment of potential developmental disorders”), or doctors (there was “a potential that certain health care providers use more hepatitis B [vaccine] at birth and would also be more likely to diagnose some of the outcomes”), or both.