by Jo Marchant
And to my friend in the sandpit I would say this: we no longer need to abandon evidence and rational thinking in order to benefit from the curative properties of the mind. The science is there. Let’s take a look at what it says.
Until a few months after his second birthday, Parker Beck from Bedford, New Hampshire, seemed to be a happy, healthy little boy. Then he began to withdraw from the world. Parker stopped smiling, speaking or responding to his parents. He woke frequently during the night, made odd, high-pitched screams and developed repetitive habits such as spinning around and banging his head with his hands. After seeking medical advice, his parents, Victoria and Gary, heard the words they dreaded: their son was showing classic signs of autism. Despite their efforts to gain the best treatment for their son, Parker continued to deteriorate. Until April 1996, that is, when Parker was three. Then something amazing happened.
As is common in children with autism, Parker also had gastrointestinal problems, including chronic diarrhea. So Victoria took him to see Karoly Horvath, a gut specialist at the University of Maryland. At Horvath’s suggestion, Parker underwent a routine diagnostic test called an endoscopy, in which a camera on the end of a flexible tube is inserted into the intestinal tract. The test itself didn’t reveal anything useful. But almost overnight, Parker began to make a dramatic recovery. His gut function improved, and he started sleeping soundly. And he began to communicate again—smiling, making eye contact, and from being almost completely mute, he was suddenly naming flashcards and saying “Mommy” and “Daddy” for the first time in over a year.
The label of autism covers a wide spectrum of disorders characterized by problems with language and social interaction, and it affects around half a million children in the U.S. Although some children show impaired development from birth, others like Parker appear normal but then regress. Some of the individual symptoms can be treated with drugs. Educational and behavioral therapies (for children and parents) can make a huge difference. But there is no effective treatment or cure. For Victoria, Parker’s sudden transformation seemed like a miracle.
She persuaded the hospital to tell her every detail of the endoscopy procedure that Parker had received, right down to the dose of anesthetic they used. After a process of elimination, she became convinced that the change in her son’s symptoms was due to a dose of a gut hormone called secretin. This hormone stimulates the pancreas into producing digestive juices, and was given to Parker as part of a test to make sure that his pancreas was working properly. Victoria believed that there was a connection between her son’s gut problems and his symptoms of autism, and concluded that the hormone must have triggered his dramatic improvement.
Desperate to get another dose of secretin for Parker, Victoria called and wrote to the physicians at the University of Maryland to tell them about her theory, but they showed no interest. She also contacted autism researchers and doctors around the country, sending home videos that documented Parker’s progress. Finally, in November 1996, her story reached an assistant professor of psychopharmacology at the University of California, Irvine, Kenneth Sokolski, whose son Aaron had autism. Sokolski persuaded a local gastroenterologist to give Aaron the same diagnostic test. He too started making eye contact and repeating words.
This was enough to persuade Horvath at the University of Maryland to infuse a third boy with secretin—and he showed the same response. Horvath also gave a second dose to Parker, and Victoria noted another surge in her son’s progress. In 1998, Horvath published a report in a medical journal of the three boys’ secretin treatment, claiming a “dramatic improvement in their behavior, manifested by improved eye contact, alertness and expansion of expressive language.”1
Horvath refused to give Parker any more doses after that, citing concern that secretin was not licensed for use as a treatment. Victoria eventually found another doctor who was willing to treat Parker, however, and on October 7, 1998, his story was broadcast to an audience of millions on the NBC show Dateline. The program showed the videos of Parker becoming a playful, connected little boy, and featured testimony from other parents who had tried the hormone after hearing of Parker’s progress. “After that secretin, no more diarrhea, potty trained, looking in the eyes, talking, saying, ‘Look how pretty outside!’ ” enthused one mother. “He was staring right in my face, looking at my eyes, looking like, ‘Mom, I haven’t seen you in a year,’ ” said another.2 The Dateline program claimed that of 200 children with autism who had been given the hormone, more than half showed a positive response.
It took just two weeks for Ferring Pharmaceuticals, the only U.S. company licensed to produce secretin, to sell out. Doses of secretin exchanged hands for thousands of dollars on the internet. There were stories of families mortgaging houses to afford it, or buying black-market batches from Mexico and Japan. In the following months, more than 2,500 children were given secretin, and success stories continued to flood in.
“There was tremendous excitement,” recalls pediatrician Adrian Sandler at the Olsen Huff Center for Child Development in Asheville, North Carolina. “Our phones were ringing off the hook, because parents of kids with autism who we were following wanted to have them treated with secretin.”3 But medical professionals were concerned about a potential public health crisis. With no hard data on whether secretin was safe to use in repeated doses, let alone whether it worked, more than a dozen clinical trials were urgently commissioned at medical centers across the country. Sandler led the first controlled trial to be published, of 60 autistic children.
As is the gold standard in such trials, Sandler’s participants were randomly divided into two groups. One group received the hormone, the other a fake treatment or placebo (in this case, an injection of saline). To be judged an effective drug, secretin would have to do better than placebo. The children’s symptoms were assessed before and after the injection by clinicians, parents and teachers who had no idea which treatment each child had received.
Sandler’s report appeared in the prestigious New England Journal of Medicine in December 1999, and the results were as surprising as they were damning.4 There was no significant difference between the two groups. The results from the other trials were the same: secretin showed absolutely no benefit when compared to the fake treatment. As a drug for autism, it was useless. The entire promise of secretin was apparently an illusion, invented by parents so desperate to see an improvement in their kids that they had literally imagined it. The secretin story was over.
Or was it? The conclusion in Sandler’s paper is one line long: “A single dose of synthetic human secretin is not an effective treatment for autism.” But what he didn’t write in that paper was how struck he was by the fact that both groups significantly improved. “The interesting thing for me was that kids in both groups got better,” he tells me. “There was a significant treatment response in the group that received secretin and in the group that received saline.”
Was it a lucky coincidence? As with many chronic conditions, symptoms in autism can fluctuate over time. One reason why it is so important to test new treatments against placebo is that any apparent change in symptoms after taking a medicine might be a result of chance. But Sandler was surprised by how big the improvement was.
The children in his trial were assessed on an official scale called the Autism Behavior Checklist, which covers a wide range of symptoms from whether they respond to a painful cut or bruise to whether they return a hug. The scale runs from 0 to 158, with higher numbers denoting more severe symptoms. The kids in Sandler’s placebo group started the trial with an average score of 63. A month after receiving an injection of the fake hormone (saline solution), they averaged just 45.5 That’s an almost 30% improvement within a few weeks—something that to many parents of kids with autism would seem like a miracle. What’s more, the effect was not evenly distributed. Although some children showed no response, others responded dramatically.
This pattern suggested to Sandler that the Becks and other parents convin
ced of the treatment’s benefits had not imagined the changes in their children. Their kids’ symptoms really did improve. But it had nothing to do with secretin.
—
BONNIE ANDERSON didn’t notice the water on her kitchen floor until it was too late. One summer evening in 2005, the 75-year-old had fallen asleep on her Davenport sofa while watching TV.6 She doesn’t remember what program was on, a decorating show, maybe, or an old movie (she doesn’t like the bad-language ones, or the bloody kind). When she woke it was dark, and she walked barefoot into the kitchen for a glass of water, without bothering to switch on the lights. But the water purifier had been leaking and she slipped on the wet tiles, landing flat on her back.
Unable to move, Bonnie felt an excruciating pain in her spine. “It was scary,” she says. “I thought, ‘My God, I broke my back.’ ” Her partner, Don, dragged her down the hall and put a blanket over her, and a couple of hours later she was able to get up onto the sofa. Thankfully she wasn’t paralyzed, but she had fractured her spine—an injury common in elderly people whose bones have been weakened by osteoporosis.
Bonnie lives with Don in a small, white bungalow in Austin, Minnesota. She worked for 40 years as a telephone operator for the town’s main employer, Hormel Foods (makers of Spam), and has stayed active into her retirement. She has orange makeup, big white hair and a busy social life, and loves nothing more than an 18-hole round of golf, a sport she has played all her life. But the accident left her devastated. She was in constant pain and couldn’t even stand up to do the dishes. “I couldn’t sleep at night,” she says. “I couldn’t play the golf I wanted to play. I’d go and sit in the den with a heating pad.”
A few months later, Bonnie took part in a trial of a promising surgical procedure called vertebroplasty, which injects medical cement into the fractured bone to strengthen it. Don drove Bonnie to the hospital—the Mayo Clinic in Rochester, Minnesota—just before dawn on a cold October morning. She walked out of the hospital after the procedure, and felt better immediately. “It was wonderful,” she says. “It really took care of the pain. I was able to go back to my golfing, and everything I wanted to do.”
Almost a decade on, Bonnie is still delighted with the outcome. “It was a miracle how well it turned out,” she says. Although breathing problems are now starting to slow her down, she isn’t limited by her back. “I have a birthday coming up, I’ll be 84,” she chuckles. “But I still plan on playing a little golf this summer.”
The vertebroplasty apparently healed the effects of Bonnie’s fractured spine. Except there’s something Bonnie didn’t know when she took part in that trial: she wasn’t in the vertebroplasty group. The surgery she received was fake.
In 2005, when Bonnie slipped on her wet floor, the technique of vertebroplasty was rapidly gaining popularity. “Orthopedic surgeons were doing it. Physiatrists [rehabilitation physicians] were doing it, anesthesiologists were doing it,” says Jerry Jarvik, a radiologist from the University of Washington in Seattle. “Anecdotally there were lots and lots of reports as to how effective this procedure was. You’d get them on the procedure table, inject the cement, and they’d effectively jump off cured.”7
Bonnie’s surgeon at the Mayo Clinic, David Kallmes, says he too had seen “positive” results from the procedure, with around 80% of his patients getting substantial benefit from it.8 But nonetheless he was starting to have doubts. The amount of cement that surgeons injected didn’t seem to matter much. And Kallmes knew of several cases in which cement was accidentally injected into the wrong part of the spine, and yet the patients still improved. “There were clues that maybe there was a lot more going on than just the cement,” he says.
To find out what, Kallmes teamed up with Jarvik to do something groundbreaking—at least in the field of surgery. They planned to test the effectiveness of vertebroplasty against a group of patients who would unknowingly receive a pretend operation. Although such placebo-controlled trials are routinely used to test new drugs like secretin, they are not generally required for new surgical procedures, partly because it often isn’t seen as ethical to give patients fake surgery. Kallmes points out, however, that with surgery just as with drugs, untested therapies risk harming millions of patients. “There’s nothing unethical about a sham trial or a placebo trial,” he says. “What is unethical is not doing the trial.”
Kallmes and Jarvik enrolled 131 patients with spinal fractures, including Bonnie, at 11 different medical centers worldwide. Half of them received vertebroplasty and half received a fake procedure. The patients knew that they only had a 50% chance of receiving the cement, but Kallmes went to great lengths to make sure that the sham surgery was as realistic as possible, so that the trial participants wouldn’t guess which group they were in. Each patient was taken into the operating room, and a short-acting local anesthetic was injected into his or her spine. Only then did the surgeon open an envelope to discover whether the patient would receive the real vertebroplasty or not. Either way, the operating team acted out the same predetermined script, saying the same words, opening a tube of the cement so that its characteristic smell of nail polish remover filled the room, and pressing on the patient’s back to simulate the placement of the vertebroplasty needles. The only difference was whether or not the surgeon actually injected the cement.
Afterwards, all of the patients were followed for a month, and asked to rate their pain and disability using questionnaires. The study was published in 2009.9 And even though Kallmes had harbored some doubts about the procedure, he was shocked by the results. Despite all of the apparent benefits of vertebroplasty, there was no significant difference between it and the fake operation.
Both groups substantially improved, however. On average, their pain scores were reduced by almost half, from 7/10 to just 4/10. The disability score was based on a series of questions such as: can you walk a block, or climb stairs without holding a handrail? At the beginning of the trial, the patients answered no to an average of 17 out of 23 questions, a score that is categorized as “severe disability.” A month after the surgery, they scored on average just 11. Although some were still in pain after the procedure, others, like Bonnie, were practically cured. A second trial of vertebroplasty carried out in Australia was published around the same time, with very similar results.
The patients’ improvement was probably due to range of factors. Pain symptoms can fluctuate, and vertebral fractures do heal, slowly, over time. But both Kallmes and Jarvik believe that to produce such a dramatic improvement, there must have been something else going on—something in the patients’ minds. Just as with secretin, it appears that the mere belief they had received a potent treatment was enough to ease—and in some cases banish—their symptoms.
The phenomenon in which people seem to recover after they are given a fake treatment is called the placebo effect, and it is well-known in medicine. Clinical trials consistently show a strong placebo effect across a wide range of conditions, from asthma, high blood pressure and gut disorders to morning sickness and erectile dysfunction. In general, however, scientists and doctors view it as a mirage or trick: a statistical anomaly where people would have improved whether they received the treatment or not, combined with a morally dubious phenomenon in which desperate or gullible people are fooled into thinking they are better when they really aren’t.
Back in 1954, an article in the medical journal The Lancet stated that placebos comfort the ego of “unintelligent or inadequate patients.”10 Although doctors might not put it so bluntly today, attitudes haven’t changed much since then. The placebo-controlled trials introduced at around that time have been one of the most important developments in medicine, allowing us to determine scientifically which medicines work and which don’t, saving countless lives in the process. They form the bedrock of modern medical practice, and rightly so. But within this framework, the placebo effect is of no interest beyond being something to guard against in clinical trials. If a promising therapy is shown to be no better
than placebo, it is thrown out.
Trial results show that neither secretin nor vertebroplasty has any active effect. So according to the rules of evidence-based medicine, the improvements experienced by patients like Parker and Bonnie are worthless.
Yet when Sandler told the parents in his study of secretin that he had found no benefit for the hormone over placebo, a huge 69% of them still wanted it for their kids.11 Likewise radiologists have refused to give up on vertebroplasty. After Kallmes and Jarvik’s report was published, the pair were attacked in hostile editorials and personal letters, and even screamed at in a meeting. “People felt extraordinarily strongly that we were taking away something that was helping their patients,” says Jarvik. In the U.S., many insurers still cover the procedure, and even Kallmes still carries out vertebroplasties regardless of his trial results, arguing that for many of his patients there is no other option. “I see patients get better,” he says. “So I still do the procedure. You just do what you need to do.”
We see similar cases again and again. In 2012, a popular class of sleeping pills called Z-drugs was shown to be of little value after accounting for the placebo effect.12 The same year, the sedative ketamine was tested in a double-blind trial for cancer pain; previous studies had described its effects as “complete,” “dramatic” and “excellent,” yet it too proved to be no better than placebo.13 In 2014, experts analyzed 53 placebo-controlled trials of promising surgical procedures for conditions from angina to arthritic knees, and found that for half of them, sham surgery was just as good.14
Perhaps the doctors and patients in all of these cases really were fooled by a combination of random chance and wishful thinking. But by continuing to dismiss the experiences of so many people, I can’t help wondering if we are also throwing out something that could be of real help. So here’s my question. Might the placebo effect, instead of being an illusion that we should puncture, sometimes be of real clinical value—and if it is, can we harness it without exposing patients to potentially risky treatments?
