The Doctor Who Fooled the World

by Brian Deer

  “Well, I can’t really comment,” he says.

  That was enough for me. There was something going on here. If the paper’s senior author couldn’t offer a better answer, I suspected the disparity was real. And if that child’s case was wrong, what else might be wrong in that five-page, four-thousand-word paper?

  I was kind of tempted to give it a spin. But how could I investigate a clinical case series? Here was medical information of the highest security: anonymized patients, child patients, developmentally challenged patients. The chances of discovering who their parents were, and when their kids showed the first signs of autism, were probably about as good as winning a lottery for which I’d not bought a ticket.

  But then, even before I’d reported back to Nuki, my attention to this matter was sharpened. We got a complaint from Ms. Two that was so over-the-top that, in my opinion, its unstated, but transparent, objective was to get Mr. Lawrence taken off his story.

  “I remain deeply shocked that such a journalist who, in my opinion is neither well informed nor particularly intelligent, should be let loose as a representative of a newspaper with the reputation of the Sunday Times,” she wrote, in a three-page email headed “Serious Concerns re Sunday Times journalist,” to the newspaper’s editor, John Witherow. “The questioning began with a launch into the exact nature of what happened on the day my younger son had received his MMR vaccine down to questions about where I worked, what the surgery was like, what time of day it would have been.”

  And in case that didn’t do it, her charges went further. I only vouch for the bit about my bladder.

  Surprised and shocked by the tone . . . stopped little short of interrogation . . . repeatedly displayed arrogance . . . did not appear to know . . . consistently revealed a dangerous bigotry and clear ignorance . . . exceptionally insulting . . . complete waste of my time . . . methods seemed more akin to the gutter press . . . his whole appearance was shoddy and shifty . . . kept turning the same tape over every time it ran out . . . he paid many trips to the toilet saying that only just drinking tea again was affecting his bladder yet prior to that he had said he was a regular tea drinker.

  The next day, Nuki got a phone call from Wakefield’s publicist, a man named Abel Hadden. And later, I got a warning from a lawyer called Clifford Miller (who would one day resurface representing Wakefield) hilariously trying to gag me. In a dense two-pager, littered with comedy legalese, he claimed that what he called the “gratuitous license” granted me to record the interview was “void ab initio,” demanded that I “deliver up” my recordings within twenty-eight days, and told me that the use of “the words spoken” by Ms. Two would infringe his client’s “copyright literary work.”

  Call me suspicious. But, after her letter, I figured she had something to hide.

  NINETEEN

  Cracking the Coombe

  With Richard Barr’s lawsuit crashed and burned, I assumed I was alone in the wreckage. Other journalists who’d spent any time on Wakefield had mostly either offered themselves as his mouthpiece or platformed a crossfire of “experts.” Nobody, it appeared, was out looking for a story: old-style newspaper reporting.

  But, just five weeks after my visit with Ms. Two, came an investigation about which I knew nothing. At the Coombe Women’s Hospital, on the south side of Dublin, a lawyer and two scientists, retained to advise the drug companies, arrived at the front desk for a showdown with John O’Leary and his miracle measles-finding machine.

  “The Coombe,” as locals called it, was in a rough part of town. With Holles Street to the east, and the Rotunda north of the Liffey, it was one of the Irish capital’s three maternity units. Not renowned for molecular biology. Ringed by thin-walled terraced cottages and gang-friendly public housing, it wasn’t really the kind of place you’d want to walk past at night, much less take your scientific riddles.

  “This isn’t the Great Ormond Street of Dublin, trust me,” says a friend whose brother was born at the Coombe, and who knew the neighborhood well. “They’ve spruced it up a bit lately, so it’s not as much of a shithole as it used to be. An awful place, it is.”

  It was, nevertheless, a place where fear of vaccines was reborn for the twenty-first century. Second only to the twelve-child Lancet paper, O’Leary’s cherished machine (“a thousand times more sensitive”) was as integral to the delivery of Barr’s Venus De Milo as any midwife or obstetrician on the hospital’s wards to the latest baby Patrick or Mary.

  The visitors were led by Gillian Aderonke Dada. She was both a lawyer and a medical doctor. Aged forty, and praised as “confident” and “inspiring,” she worked for a heavyweight firm of commercial solicitors—which had earlier been retained in the DTP trials—that day representing the three vaccine manufacturers: SmithKline Beecham, Aventis Pasteur, and Merck.

  The class action was dead. But a string of hopeless appeals kept its sweating corpse technically breathing. So the companies exploited this ventilated state, with their eyes on the United States. Spurred by Dan Burton’s congressional hearings, Lenny Schafer’s republishing of British media reports, and Wakefield’s appearance on CBS’s 60 Minutes, thousands of parents were being recruited, coast-to-coast, for a yet bigger lawsuit on the way.

  This was where the story gets briefly complex, as I am obliged to pursue Wakefield, Barr, Kirsten Limb, and Ms. Two to the molecular level of evidence. When I learned of it later, it would prove the biggest challenge in mastering all those interminable -ologies behind which sheltered the real people and specific facts.

  So while I plodded forward with my inquiries in London, waiting with Dada in the Coombe’s reception area were two top biomedical detectives. One was Malcolm Guiver, head of molecular diagnostics at the British government’s public health laboratories in Manchester. The other: Stephen Bustin, reader (later professor) in molecular science at London’s Queen Mary school of medicine. Both were specialists in the polymerase chain reaction and had worked for years with the same equipment as O’Leary’s, the ABI Prism 7700.

  Compared with later devices for amplifying DNA, the 7700 was a heavy metal beast. Although only ninety-four centimeters (twenty-seven inches) wide, it weighed 130 kilograms (286 pounds) with an attached Apple desktop computer. Finished in the color of aircraft interiors, its front sloped to a bevel above a row of ventilator slats that ran nearly the length of the machine. Wrapping its right front corner was a plastic, lift-up window, behind which a boxlike, heatproof cover shielded its exquisitely delicate reactions.

  The manufacturer’s manuals were a walk in the park. I’ve owned washing machines with more baffling instructions. O’Leary’s operatives lifted the window and raised the cover, exposing a “plate” of ninety-six tiny wells, which rested on a heating “block.” Into these they popped “runs” of sealed plastic tubes containing bowel tissue, blood, or cerebrospinal fluid, each in a solution of chemical reagents. Around these in the runs were placed various control tubes for comparison: some negative, containing, for example, distilled water; some positive, spiked with measles.

  Technicians set the computer. And off it all went: automated molecular amplification.

  If any samples in the tubes contained what O’Leary hoped to find (single strands of measles virus RNA), the machine allowed these to be converted into DNA: first a single strand, complementary to the RNA, then the full and famous double helix. Rapidly heating the tubes made this come apart, like the ladder unzipping, after which an enzyme, called the “Taq polymerase,” went to work with a miracle of life.

  Taq rebuilt the ladder’s rungs by creating fresh nucleotides of adenine, thymine, cytosine, or guanine—A, T, C, or G—to create partners for the separated strands. Then the machine cooled the tubes, and the ladders reformed: only now there were twice as many. Thus each single RNA strand became one, then two complementary DNA strands, which exponentially doubled, and doubled with each cycle, until billions of copies were present.

&nb
sp; Each tube was laser-monitored, second by second, and with the 7700 counting the cycles, it graphed how much (if any) of the target was present at each cycle, so as to calculate how much was there in the first place.

  Quick. Easy. Lab staff liked to think so. But therein lay traps for the unwary. Although Wakefield’s Carmel business planned to advertise to investors that its 7700 would diagnose both Crohn’s disease and “autistic enterocolitis,” its manufacturer said such use was improper. The machine’s operating manuals, technical guides, and brochures expressly warned, in bold, standalone statements:

  For Research Use Only. Not for use in diagnostic procedures.

  One reason for this was that, when working with such sensitive molecular technologies, if something could go wrong, it would. Despite a digital stream spewing data from the machine, which the Apple mashed together into real-time line graphs, you never stepped in the same river twice. Behind all the gadgetry lay biological phenomena: calling for art and integrity from researchers and technicians, working more like violinists than mechanics.

  Hence Dr. Dada, and her two detectives. They’d come to collect the music. Unlike a violin, every performance was recorded, digitally archived, and disclosable to the lawsuit’s defendants. O’Leary had supplied the court with “experimental reports,” giving results on many samples from the Royal Free, well by well. But the Apple also logged a great deal more—some four thousand data points on each tube’s progress—which the drug companies wanted. Very much.

  Guiver, forty-six, with black hair and chiseled jaw, was a catch for more than his looks. With a PhD degree in molecular virology, his lab was among the first to get a 7700. And he’d filed a forty-nine-page analysis, based on reports from the lab, evaluating the Coombe’s performance. He alleged “wholly inappropriate” and “entirely unreliable” methods, “lack of judgment,” “inadequate controls,” and complained of “spurious false positives” from the machine.

  Big Pharma had paid for his professional opinions. But he wasn’t alone in his critique. There was a second analysis from a Dutch adviser, Bertus Rima, professor of molecular biology at the Queen’s University, Northern Ireland. He’d spent thirty years studying measles. And a third from a virology professor, Peter Simmonds, of Edinburgh, Scotland, who in the past decade had grabbed the cape as Britain’s most cited microbiologist.

  Rima and Simmonds took the same view.

  Meaningless . . . suspicious . . . no validity . . . dubious practice . . .

  Wholly unacceptable . . . results invalid . . . unreliable . . . contradictory . . . untenable

  It would take me weeks to get my head round judgments like these. But one simple issue, very easy to grasp, concerned the machine’s most basic operation. If you wanted, you could run it for as long as you liked. But if the lasers didn’t quickly detect a signal from the tubes—essentially after no more than about thirty-five cycles—then you needed to accept that your target wasn’t there, or tweak your preparations, and start over. Because by then the process would be running out of reagents—the juice—and the hiss from billions of random amplified molecules could, metaphorically, drown any tune.

  Every source concurred. Even on YouTube. But, to be safe, I ask the manufacturer. Later owned by a multinational, Thermo Fisher Scientific, the 7700 became a legacy product of its Life Sciences Group, whose vice-president for research and development, Vinod Mirchandani, briefs me from San Francisco. “We’ve learnt that, if you go beyond thirty-five,” he tells me, “and you’re starting to see some small signal, most likely they’re not the real thing.”

  He says that robust results are found sooner in the cycles: when the graphs on the Apple plot strongly rising lines, before these plateau as the process falters. In later cycles, he explains, on behalf of the machine’s makers, you can “get a lot of noise” and “false positives.”

  But the Irish pathologist and his team saw things differently in their enthusiasm for measles virus. Reports from the Apple that they filed in the lawsuit revealed that samples taken at Hampstead—including from Child Two—were often tested in runs of up to forty-five cycles. And, in one case, fifty heatings and coolings. Rima even said that Barr’s firm had told him that they planned to retest samples unblinded at Warwick with a, frankly ludicrous, seventy.

  Meanwhile, Bustin came at O’Leary by another route, after experiencing a personal epiphany. Aged forty-nine, with a cheeky manner and a PhD in molecular genetics, he was England’s Dr. PCR Geek. He wrote about it, taught it, probably dreamed in thermocycles. And within days of being hired to evaluate O’Leary’s methods, Bustin found himself wondering . . . and wondering.

  He’d been studying an Apple graphic of a 7700 plate generated from a disclosure in the lawsuit. Of the ninety-six wells, fifty-one were shown in use: labeled from A1 to E3. They’d been loaded in duplicate (two per child or type of control), which was common in this kind of work.

  But he’d learned at school that fifty-one was an odd number. It looked to him like E4 had gone missing.

  Right or wrong, he was keen to learn more. This anomaly might be a clue. What appeared to be absent was a “no template” control—meaning a tube of reagents, with Taq and other stuff, but no biological sample. No measles. If a tube in that well had tested positive for measles, it would suggest contamination: the scourge of PCR. And if that was the case, he speculated, but the well reported empty, then maybe someone had airbrushed the result.

  For six months Dada had been pressing for this visit. But after her experts’ analyses were disclosed to Barr’s people, dealings with the Coombe became fraught. First O’Leary was in Australia and couldn’t be contacted. Then none of fifty-nine dates from Dada were convenient. It would take £20,000 and up to five weeks to prepare for guests. The 7700 broke down.

  But now here she was with her molecular detectives, being ushered toward the back of the building. O’Leary’s lab was bright and modern, with plenty of kit—and anomalies screwed to the walls. Above doors to side rooms, the visitors spotted a pair of signs that set the scientists murmuring with the lawyer. One read “Plasmid Room,” and the other “PCR Set-Up”: the first generally meaning where positive control materials were prepared, the second where amplifications were run.

  “The doors of these side rooms were single swing doors,” Dada explained later in a statement, filed at the Royal Courts of Justice, in London. “I did not see any facilities for changing coats and/or shoes.”

  Happily, Barr was her best witness to this. He’d also flown to Dublin for the visit. And he’d brought a camera, which he offered the visitors. But O’Leary wouldn’t allow any pictures. Nor, Dada said, would he release any data that hadn’t already been disclosed to her clients.

  It soon became clear that legal force would be needed. And the drug companies could afford to get it. Weeks later, after O’Leary refused to consent, a London judge issued a request to the Dublin courts to compel the pathologist to comply. This man who’d bragged in Washington of his “independent” achievements—on matters potentially critical to the safety of millions of children—had become uncommonly shy.

  The visit that Monday was productive, however. Although the visitors spent most of it in a conference room, waiting, Dada, Guiver, Barr, and an O’Leary associate witnessed Bustin pull a rabbit from a hat. With them huddled around his laptop, the PCR geek displayed a report on a sample of blood from a non-autistic child included in the Warwick unblinding.

  The report was negative for measles virus: good if you aimed to show that the vaccine caused autism. But Bustin also displayed the underlying raw data, from a small amount that O’Leary had disclosed. It was derived from the same child—same well; same tube—and was reported to be positive.

  Oops.

  “Richard Barr acknowledged that this was an important scientific issue,” Dada drily recounted in her statement.

  So here was another anomaly, spotted by the geek. And mo
re would be alleged in court. “He is beginning to find a mismatch,” counsel for SmithKline Beecham told the judge in London, weeks later. “These findings are raising very serious concerns.”

  Guiver, Bustin, Rima, and Simmonds all agreed that much was wrong. Along with excess cycling likely producing false positives—precisely the phenomenon the manufacturer warned me of—they thought the lab was awash with measles. “The overall impression is one of inadequate care, slipshod compliance with protocols, and a lack of basic understanding,” Bustin argued.

  Contamination, they believed, might have occurred at any stage. It might have arisen at Hampstead, even with Wakefield himself, who personally delivered samples to Dublin. It might be related to the plasmid room, and the drafty swing doors. Maybe the reagents, careless fingers, or pipettes. Rima recounted how, in his own laboratory, traces of mumps RNA (another paramyxovirus) had lingered on a bench for nine years.

  A bigger shock, however, was proposed in court: that the experimental reports, submitted in Barr’s lawsuit, didn’t give the true results from the machine.

  I can’t say. But, if the visitors guessed right, the upshot was surely serious. O’Leary’s claimed findings were pivotal to Barr’s lawsuit. They were the peg for the BBC’s Panorama program. They were the basis for the pathologist’s endorsement of Wakefield in front of the congressional committee. They were the occasion for Schafer’s “Wakefield Fest,” importing the claims into the United States. And they’d convinced countless parents of children with developmental issues that a cause, MMR, had been found.

  In reply, O’Leary was certain all was well: his lab hadn’t failed in any way. Its findings backed claims by parents like Ms. Two. The Coombe was clean of stray measles. “Appropriate environmental and laboratory set-up controls” were maintained, he insisted, in a report. Results of testing had shown “clearly and unequivocally” that no contamination had occurred.

  Better than that: he’d the means to prove these assertions—as he’d explained to the Washington lawmakers. What he’d called the “gold standard” was DNA sequencing: running out all the strings of As, Gs, Cs, and Ts after PCR amplification. This couldn’t merely confirm whether the virus was present, but “outrule” false positives, detect contamination, and definitively fingerprint the measles viral strains: as found wild in nature, brewed for vaccines, or reared in labs for experiments.