Monday, March 24, 2008

Practicing Patients

By THOMAS GOETZ
Todd Small was stuck in quicksand again. It happened, as always, on the floor of the Seattle machine shop where he worked. His shift complete, Small was making the 150-yard walk from his workstation to his car, when he realized that his left leg was sinking deep in the stuff. Though this had happened before — it happened nearly every day now — he stopped and glanced down at his feet. His Nikes looked normal, still firmly planted on the shop’s concrete floor. But he was stuck, just the same. His brain was sending an electrical pulse saying “walk,” but as the signal streaked from his cerebellum and down his spinal cord, it snagged on scar tissue where the myelin layer insulating his nerve fibers had broken down. The message wasn’t getting to his hip flexors or his hamstrings or his left foot. That connection had been severed by his multiple sclerosis. And once again, Small was left with the feeling that, as he described it, “I’m up to my waist in quicksand.”
For the 400,000 Americans with multiple sclerosis, Todd Small’s description will most likely ring true. Muscle stiffness is a hallmark of the disease, and “foot drop” — the term for Small’s quicksand feeling — is a frequent complaint. The condition is usually treated, as it was in Small’s case, with baclofen, a muscle relaxant that works directly on the spinal cord. Every day for 14 years, he took a single 10-milligram pill. “My neurologist always told me if you take too much it will weaken your muscles. So I never wanted to go over 10 milligrams.” It didn’t seem to have much effect, but he carried on as best he could.
Small would have continued just as he was had he not logged on last June to a Web site called PatientsLikeMe. He expected the sort of online community he’d tried and abandoned several times before — one abundant in sympathy and stories but thin on practical information. But he found something altogether different: data.
After choosing a user name and filling out a profile, Small was asked to list his symptoms and treatments. He entered the 200 milligrams of Provigil he takes daily to fight fatigue along with the Tysabri injection he takes to slow the progress of his disease. And then he clicked on baclofen, and the Web site informed him that nearly 200 patients registered at PatientsLikeMe were taking the drug. He clicked again, and up popped a bold bar graph, sectoring those 200 across a spectrum of dosages. And there it was. Contrary to what his neurologist told him years ago, 10 milligrams wasn’t the maximum dose. In fact, it was at the low end of the scale. “They’re taking 30, 60, sometimes 80 milligrams — and they’re just fine,” Small recalls. “So it hits me: I’m not taking nearly enough of this drug.”
A few days later, Small asked his neurologist to up his dosage. Now Small takes 40 milligrams of baclofen a day. His foot drop isn’t cured — there are no miracles in M.S. — but he has found that after 14 years, he can walk to his car without sinking into quicksand. “Oh, man, I really dreaded that walk,” Small recalled when I spoke with him recently. “All shift, it’d be in the back of my mind. Am I going to have trouble? Is it going to get me? Now I almost got it figured it out. I don’t struggle like I used to.”
There are a little more than 7,000 Todd Smalls at PatientsLikeMe, congregating around diseases like Parkinson’s, multiple sclerosis (M.S.) and AIDS, all of them contributing their experiences and tweaking their treatments. At first glance, the Web site looks like just any other online community, a kind of MySpace for the afflicted. Members have user names, post pictures of themselves and post updates and encouragements. As such, it’s related to the chat rooms and online communities that have inhabited the Internet for more than a decade.
But PatientsLikeMe seeks to go a mile deeper than health-information sites like WebMD or online support groups like Daily Strength. The members of PatientsLikeMe don’t just share their experiences anecdotally; they quantify them, breaking down their symptoms and treatments into hard data. They note what hurts, where and for how long. They list their drugs and dosages and score how well they alleviate their symptoms. All this gets compiled over time, aggregated and crunched into tidy bar graphs and progress curves by the software behind the site. And it’s all open for comparison and analysis. By telling so much, the members of PatientsLikeMe are creating a rich database of disease treatment and patient experience.
And that’s no small thing. As modern medicine has reduced infectious disease in the U.S., the country’s population — like those in other developed nations — has shifted to longer-term, chronic ailments like heart disease and diabetes, diseases that patients must cope with for years and even decades. What’s more, conditions like high cholesterol and obesity demand years of vigilance to minimize their chances of leading to more serious health problems.
Disease management has become a national reality, a common experience that demands uncommon attention. And PatientsLikeMe is a tool that allows patients to manage their disease with a sophistication and precision that would have been unimaginable just a decade ago. The 7,000 members of PatientsLikeMe, in other words, are beta testers — they may be the vanguard of how we all will care and treat our résumé of chronic diseases. They’re not typical patients, in the sense of waiting for advice from a doctor. They are, rather, co-practitioners treating their conditions and guiding their care, with possibly profound implications. “People who use it will live longer; people who don’t won’t,” boasts Jamie Heywood, the provocative co-founder of PatientsLikeMe. “That’s evolution.”
Of course, turning patients’ experiences into usable data raises a host of questions for medicine. When patients take the reins of their own treatment, what role do doctors play? What’s to keep patients from misinterpreting the streams of data and finding false hope — and what’s stopping them from embarking on unproven and even risky treatments or dosages? And what happens if the real-world information at PatientsLikeMe contradicts the clinically proved protocols of medical science?
For many in the medical community, these are unexpected and unsettling questions. And they reflect the broader concern that many doctors have with the Internet’s emergence as a powerful source of medical information. According to a 2006 study by the Pew Research Center, 80 percent of Internet users, or about 113 million adults in the U.S., get health information from the Internet. Some physicians disdain this trend, bemoaning their hyperinformed patients as “Googlers.”
Yet even doctors who consider themselves advocates of online medical information raise concerns about patients self-medicating or self-treating. Erik Ensrud, a neurologist at Brigham and Women’s Hospital in Boston, specializes in many of the diseases that PatientsLikeMe focuses on. “The nervous system is the most complicated system in the body,” he says. “It’s so often much more difficult to quantify and study than we think it will be.”
For the members of PatientsLikeMe, though, the Web site can seem like a revelation — it’s a community forged not only around shared circumstance but also around a shared purpose. They say they are part of a project that may change the way their disease is treated and could change the way medicine is practiced. It’s an experiment in the future of medicine, running in real time.
PatientsLikeMe started with a single case of amyotrophic lateral sclerosis. In 1998, Stephen Heywood, a 29-year-old carpenter, learned that he had A.L.S., a neurodegenerative disorder commonly known as Lou Gehrig’s disease. Jamie, his older brother, quit his job to find a cure. An M.I.T.-trained mechanical engineer with a knack for neuroscience, Jamie founded the A.L.S. Therapy Development Institute in Cambridge, Mass., the following year. So began a radical quest to save Stephen’s life. They tried experimental drug therapies, they tried a stem-cell transplant and they tried a neural implant, each effort building on the previous one. After six years, Stephen was among the most documented A.L.S. patients in the world.
The sheer volume of Stephen’s data gave Jamie an idea — a notion hatched while browsing Match.com, the online dating site. The aspiring singles there had posted a trove of information about themselves — their likes and dislikes, their dating histories, their height and weight — all to find a perfect match. Jamie realized that a similar tool might be useful in the realm of disease, for treatment rather than romance. If patients shared their information and could find someone with a similar symptomatology and disease history, then they might better plot their own course of treatment and care. And it didn’t just have to be A.L.S. — it could work for any disease.
Jamie tapped his brother Ben, who also went to M.I.T. before earning his M.B.A. at U.C.L.A., and Jeff Cole, a college classmate of Ben’s who’d spent several years building dot-coms. They started with Stephen’s own case history, breaking it down into drug dosages, symptom severities and so on. In March 2006, PatientsLikeMe opened for business. Within a few months, the company added communities for M.S. and Parkinson’s disease. Last fall the company opened a community for people with H.I.V. Earlier this month, the company placed its biggest bet on the model of collective experience, creating a community for the millions of Americans with a mental-health condition.
The Web site gathers patient information on two levels: first there’s a quantitative breakdown of symptoms and dosages, data that the software instantly turns into charts and graphs. Second are the forums, where members share advice and provide more nuanced feedback on a certain drug or treatment issue. The site is designed so that relevant data bolster the conversations in the forums and vice versa.
“Our job is to allow a conversation with the computer that will match a conversation between two patients,” Jamie explains. “Then we capture that dialogue and turn it into useful, clean data.”
Stephen died the day after Thanksgiving in 2006, but his profile, under the user name ALSKing101, still contributes to that dialogue. Like many people with advanced A.L.S., in his last three years, Stephen was living in a wheelchair, completely paralyzed and breathing through a mechanical vent. One night his breathing tube snagged and disconnected, and he suffocated while he was sleeping. Such a death isn’t unknown for A.L.S. patients, and many PatientsLikeMe members — about 75 are on ventilators — have expressed fears about dying that way. Their main concern is that suffocation is a painful, unpleasant way to go.
As it turns out, the PatientsLikeMe community tries to allay that fear. At least two other members of the site have, in fact, had their breathing tubes accidentally disconnected and suffocated until passing out — only to be found and resuscitated, just in time. They have, in other words, almost died in just the way Stephen did, in the way so many have feared, but lived to tell about it: “Long story short, my vent hose came off in ’97, and I was certain I was going to die. I stared at the clock for exactly 10 minutes until I got tunnel vision then passed out. Nothing painful, stressful or anxious about it. It was like I just got sleepy and went to sleep. Suffocation was nothing like what I expected. Lol.” Laugh out loud, indeed.
The company is located in the brick-walled basement of a former twine factory, less than a mile from the M.I.T. campus in Cambridge. With just 15 employees, it is very much in start-up mode. Cole, who is 36, manages the site’s design and software team; Ben Heywood, also 36, handles overall operations and management; Jamie, 41, is the frenzied visionary, ever eager to sketch out how powerful a database of 7,000 patient profiles can be. The company, which is financed by private investors, eschews advertising; the business model instead seeks to exploit the value of the databank itself. By and large, this means working with pharmaceutical companies. The company has had discussions with several drug firms to sell anonymized patient data on various symptoms and treatments. They have also recruited PatientsLikeMe members to be participants in clinical drug trials. Ultimately, the company expects that the data will generate insights with considerable scientific — as well as economic — value.
One afternoon in late November when I visited the office, Jamie turned to a nearby whiteboard and traced out an x-y axis, slashing a descending line from left to right. “We have the ability to run a probability engine,” he said. “We can mathematically model each patient. We can tell them what’s going to happen in their life. We can tell you when you’ll need a wheelchair.” He made a mark along the line. “And we can even tell you the day you’ll die, with remarkable certainty.”
Of course, Jamie’s spiel is part bravado. The company hasn’t started to predict specific patients’ prognoses — though they have filed for a patent covering such prediction tools. And not all quests for scientific precision are likely to succeed or prove worthwhile. But for the patient coping with the cascade of day-to-day decisions that come with managing a disease, the site’s aggregations offer a remarkable tool.
Every day, like 453 other members with M.S., Leslie Fournier gives herself an injection of Copaxone, a drug that impedes the disease’s progress. Together, the group has generated what amounts to a checklist of best practices. Fournier has learned that by doing the injection herself, “John Wayne-style,” rather than using an auto injector, she can avoid some of the bruising that’s a particularly aggravating side effect. And she has learned that the optimal needle stick goes through the skin and into the fat layer — but no further. “You don’t want to get muscle,” she relates. “Last time I did that, I couldn’t use that leg very well for a day and a half.”
Fournier, a self-described information junkie, relates all these details with a blasé precision. Her willingness to put so much information about herself online may seem immodest, even imprudent. But not to her. “I don’t worry too much about that,” she says. “Pretty much everybody I know over 45 has some kind of medical condition. Some people have had cataracts surgery, or they have high blood pressure, or high cholesterol or diabetes. Everyone has something. And if everyone has something, that really levels the playing field.”
“Any disease that is treated as a mystery and acutely enough feared will be felt to be morally, if not literally, contagious,” Susan Sontag wrote in 1978, in “Illness as Metaphor.” “Contact with someone afflicted with a disease regarded as a mysterious malevolency inevitably feels like a trespass; worse, like the violation of a taboo.” The antidote to such stigma, she suggested, is “to rectify the conception of disease, to demythicize it.”
But the myths surrounding disease are hard to dispel, our impulse to hide in metaphor difficult to shake. We still talk of wars on cancer and plagues of lawyers. “Leper” endures as a metaphor for stigma, long after leprosy itself has been eliminated from most of the world.
Stigma endures in more material ways, as well — most pointedly in concerns over medical privacy. For decades, our personal health information has been protected largely by disorganization and chaos, suggests Mark Rothstein, a bioethics and health-policy professor at the University of Louisville. But now, with major health-maintenance organizations and hospitals finally adopting electronic records, fallow storerooms of paper records are being converted into neat and portable digital form. “Technology is racing ahead of society,” says Rothstein, who advises the Department of Health and Human Services on electronic privacy. “If this is sensitive information — sexual history, mental illness, substance abuse — even if nothing bad happens, people are subject to stigma and embarrassment.”
The government has tried to protect patient privacy, most notably with the Health Insurance Portability and Accountability Act, or Hipaa. Enacted in 1996 to help streamline the adoption of electronic records, Hipaa was modified in 2002 to address concerns over how those records might be misused or distributed to unauthorized people. The law stipulates that if 18 categories of personally identifiable information are stripped from records then they may be transferred for research purposes without a patient’s consent.
Perhaps not surprisingly, Hipaa satisfies no one — not the privacy advocates, who argue against most any form of electronic record, and not the medical researchers, who now face onerous restrictions on their work. PatientsLikeMe upends this dialectic; in technology terms, it routes around the problem. Since the company is an opt-in service and not a health-care provider, Hipaa doesn’t apply. Good thing, really, since the site identifies members’ cities and their ages, two of Hipaa’s 18 prohibited categories of personal information.
But Hipaa is extraneous not just legally but philosophically as well. Many PatientsLikeMe members volunteer even more information from those 18 categories. They not only post their photos but also post photos of their children and spouses. They add brief autobiographies and describe their conditions in precise detail — including potentially embarrassing particulars on sexual function, bladder control or constipation. And though they all have user names, most go by their first names on the site’s forums.
PatientsLikeMe’s privacy policy clearly states that this sharing carries risks. It acknowledges that since anybody can register at the Web site, anybody can look at member profiles. It makes clear that there’s no guarantee that registered members are, in truth, who they say they are. And it nods to the fact that, yes, this is a business, not a public service — some personally identifiable information may be sold to “approved vendors.” But this is boilerplate. The most striking notice is the company’s Openness Philosophy, a manifesto posted prominently on the site.
“Currently, most health-care data is inaccessible due to privacy regulations or proprietary tactics,” it declares. “As a result, research is slowed, and the development of breakthrough treatments takes decades. . . . When you and thousands like you share your data, you open up the health-care system. . . . We believe that the Internet can democratize patient data and accelerate research like never before.”
In 1990, Alan Westin, a political scientist at Columbia University and an expert in privacy issues, offered a useful taxonomy of Americans’ attitudes toward privacy. On one end of the spectrum were what he called privacy fundamentalists — the 25 percent of Americans who feel that their privacy is paramount and that no one, not the government or corporations or their family, should have access to their personal information without explicit permission. At the other end of the spectrum were the privacy-unconcerned — about 15 percent of Americans — who paid no mind to privacy issues and didn’t figure they had anything to hide. In the middle were the vast majority, the 60 percent whom Westin called privacy pragmatists: those who felt that they could give a company they trusted some information — birth date, ZIP code, telephone number — for particular benefits.
The members of PatientsLikeMe aren’t all radicals culled from the privacy-unconcerned, though no doubt some are. Most more likely fit the profile of pragmatists. “I know it sounds like really personal information, but it’s not like I’m putting my phone number up,” says Jennifer Jodoin, a hotel manager in Palm Beach, Fla., who has changed her M.S. medications based on information gleaned at PatientsLikeMe. “I’m not posting my address and saying, ‘Come on by.’ It’s an exchange of information to get help and to give help.”
That pretty much defines pragmatism, and it exemplifies these strategic information-for-benefit exchanges people make like taking a blood test to get life insurance or consenting to a background check to secure a new job. Only now, at places like PatientsLikeMe, information has a currency that’s far more liquid than ever. Converted into data and bundled with information from those like us, private information can be invested for both immediate gains and long-term returns.
And there’s a fortunate side effect to converting your health history into data — it strips it of metaphor, of secrecy, of stigma. It becomes concrete, manageable and valuable. It becomes, just as Sontag suggested 30 years ago, demythicized.
As diseases go, A.L.S. or M.S. or Parkinson’s or even H.I.V. are relatively rare afflictions, at least in the United States. Mental illness, on the other hand, afflicts a vast swath of the country. Nearly 60 million Americans have a diagnosable mental disorder, according to the National Institute of Mental Health, a population that includes everything from depression to bipolar syndrome to anxiety. For PatientsLikeMe, that population represents a huge market, not to mention the potentially lucrative bounty of data related to antidepressants and other mood-disorder drugs. But it also presents a challenge.
Creating a PatientsLikeMe mental-health community — or as they call it, a “mood community” — requires a new strategy for measuring mental health. The challenge is in part semantic. Where the argot around A.L.S. or M.S. is largely clinical, the vernacular around mental health is more subjective. The official diagnostic criteria for major depression, for example, include “feelings of worthlessness” and “indecisiveness.” So PatientsLikeMe faces an input problem: how to convert the ambiguities of mental illness into metrics?
Whatever its ultimate worth, the site’s answer is elegantly straightforward. Members can update their mood status every hour on a scale of 1 to 4, from very bad to very good. How they feel may be subjective, but the resulting data can be mapped across time. The site treats sessions of therapy as if they were a dose of Prozac; the type of therapy (say, group or individual) stands as the treatment, and the length of a session (say, 50 minutes a week) as the dosage.
Such efforts at precise measurement and comparison are not the norm in evaluating mental-health treatment. Americans spend about $12 billion a year on antidepressants, but we still have little understanding of how or whether they work. In 2006, the National Institute of Mental Health released the results of the largest and longest depression study ever undertaken, the Sequenced Treatment Alternatives to Relieve Depression Study, or Star-D. Star-D rejected typical clinical study design and aimed to create a real-world representation of how patients actually experience and treat depression: through trial and error, taking one drug after another, searching for one that helps.
The results were mixed. Star-D found that if you’re depressed and spent three months on a potent psychotropic drug, you had a one-third chance of achieving remission. After almost six months on drugs, your chances bump up to about fifty-fifty — a coin flip — and you still risk side effects like sexual dysfunction, insomnia and weight gain. Those may be good odds for the industry, but it seems a poor bet for a patient.
At PatientsLikeMe, Ben Heywood sees pure opportunity in such statistics. “Those odds just aren’t good enough,” he says. “So you try Wellbutrin, and after six weeks it doesn’t work. Then Prozac. Doesn’t work. Now what? Where do you go next?” Plugged in to a community of patients sharing their depression histories and treatments, Ben argues, patients could readily find someone with symptomatology close to their own, compare drug regimens and go straight to the drug that may be more likely to work for them.
This, for patients, is the promise of a mental-health community: better tailoring of treatments. It is also the promise of the site as a business. The pharmaceutical industry should be eager to improve the accuracy and efficacy of its treatments. After all, sometimes side effects can turn into blockbusters, most famously when Pfizer scientists learned that their hypertension treatment was causing erections in men, leading the way to Viagra. Since PatientsLikeMe forgoes advertising, selling its data to pharmaceutical companies is its best apparent way to make money. But so far, it seems, the drug industry has balked at the prospect of knowing so precisely what happens to their products after they reach the market.
If Jamie Heywood talks about his company as an incendiary force in the health-care industry — “This is the way it worked before ethics destroyed medicine,” he told me at one point — the American Medical Association would pretty much agree. From the perspective of many doctors,
PatientsLikeMe raises many more questions and quandaries than it answers.
Joe Heyman, a practicing physician in Amesbury, Mass., and the chair-elect of the A.M.A.’s board of trustees, praises the potential benefits of a tool like PatientsLikeMe for patients. He’s just uncomfortable with who’s minding the toolbox. “The idea of them is a wonderful thing,” he says. “But you know, as a physician I have a code of ethics. We have licensing boards and accreditation and all these people keeping an eye on us.” Heyman suggests that if physicians themselves aren’t sole stewards of the patient data — his first choice — then there should be some national standards or a law that covers a company like PatientsLikeMe that traffics in such data.
Heyman is also uncomfortable with how PatientsLikeMe encourages patients to take their medical information and then act on it. “Sometimes patients misunderstand what a doctor says to them,” he says, “so by the time it gets to the third or fourth party it’s like a game of telephone. It’s not as reliable as coming from the horse’s mouth.”
Heyman may be speaking from the perspective of the injured party — PatientsLikeMe could be seen as a direct challenge to physicians’ omniscience — but there’s still something to his concerns. PatientsLikeMe not only lets members track their disease; it also tacitly encourages them to take action. While PatientsLikeMe advises users to consult a doctor before changing their treatments or dosages — patients can print out an exhaustive status report, replete with graphs and charts, to take on doctors’ visits, and personal physicians can register as caregivers on the site — there will inevitably be patients who treat the Web site’s apparent rigor and depth as a substitute for a physician’s training and expertise.
Erik Ensrud, the neurologist at Brigham and Women’s Hospital, worries that patients could spot an animal study or early-stage trial investigating an off-label usage of a drug and then start taking that drug in the hope that the researchers are onto something. “Even if a treatment seems promising, it could result in a very serious reduction in lifespan,” he suggests. “We’re talking survival.”
In fact, some PatientsLikeMe members have already started doing pretty much what Ensrud warns against. Last November, the A.L.S. community was abuzz with word that researchers in Italy had found that taking lithium seemed to slow the progression of A.L.S. significantly. The Italian study hadn’t actually been published yet, but that didn’t stop 34 members with A.L.S. from soliciting lithium prescriptions from their doctors and coalescing into an ad-hoc clinical trial. There are now 109 members using lithium and tracking their progress with the data tools on the site. The company has rolled out new features to monitor the group with the hope that they will be able to lend a little credence — or cast a little doubt — on the Italian study in a matter of months.
Jamie insists that PatientsLikeMe isn’t encouraging A.L.S. members to start taking lithium. But he is unmistakably excited by the endeavor. As he sees it, the experiment perfectly illustrates how PatientsLikeMe might complement large-scale and long-term clinical research by conducting observational research “on the fly.” Drawing on the notion of personalized medicine, Jamie calls this “personalized research.” And it has a certain logic: for those who already have A.L.S., traditional science works at far too plodding a pace. “The system is broken for terminally ill patients,” says Hanns Riederer, a music producer in Los Angeles who has joined the group of A.L.S. members taking lithium. “It makes us wait five to seven years for results, when we don’t even have that time. Even if it’s half-true, it’s still groundbreaking. I don’t want to wait for something else. I don’t have time to wait.”
While the members of PatientsLikeMe may seem exceptional today, they will be fairly typical tomorrow. As we learn more about disease and are able to identify it earlier, more and more Americans will fall under the rubric of “patient” and will need to navigate an array of confusing health decisions. Doctors don’t like to admit it, but “most treatment decisions right now are still based on doctors’ judgments that don’t have real research behind them,” says Jodi Halpern, a physician and bioethicist at the U.C. Berkeley School of Public Health. “But it takes real data to make the right decisions, especially for patients. There’s a powerful improvement in health outcomes, people’s quality of life, when people are better-informed.”
Physicians, of course, have known this for decades; the idea of “evidence-based medicine” — that all decisions should be based on real data — was hatched in the 1980s. But the pace of traditional research is slow, and the number of outstanding questions far exceeds the body of evidence to answer them. Only now, with technologies like PatientsLikeMe, is the pace accelerating. What’s more, it’s becoming evident that data — the foundation of evidence and research — are present not just in laboratories or universities or the proverbial halls of science but in everyday life.
Really, when you start looking, information can be found everywhere. If we could gather in structured communities and create databanks to inform our approach to life decisions, not just health decisions but also gardening or parenting or car-buying decisions, we could do everything in a more informed manner. Were we all to avow a philosophy of openness and churn our experiences into hard numbers, we could presumably improve our odds in all sorts of decisions. Why not a PregnantLikeMe or a ParentsLikeMe or even, really, an all-encompassing PeopleLikeMe?
One evening, I suggested this bigger idea to Jamie and Ben. They shared a glance and a mock-conspiratorial smile, and then looked back at me. “We know,” Ben said. “We already own all those domain names.”

No comments: