Toxic Knowledge: On Precision Medicine and a Doctor’s Search

I MET FRANK1 six years ago, when he visited my office at Columbia University Medical Center with swelling in his legs and more than 8000 mg of protein in his urine each day (roughly 8000 mg more than the trace amounts of protein a healthy kidney should leak). A kidney biopsy revealed a rare disease called membranous nephropathy, and over the next two years his disease went into remission on a combination of immune-suppressing and blood pressure medications. Since then, he’s seen me every six months for a routine check-up and surveillance labs. Fortunately, these visits have been utterly forgettable.

A few months ago, however, Frank told me that he felt “miserable,” but his misery, he said, had nothing to do with his kidneys. His two-year-old daughter had been diagnosed with severe developmental delay, and he and his wife had been shuttling her from one specialist to another without arriving at any semblance of a diagnosis or treatment plan. “We were in denial for a while, hoping she was just going to be a quiet kid or that, as the second child, she had no need to talk because her older sister was so loud,” he said, trying to crack a smile at his joke. “But now it’s clear there’s something wrong, although no one can tell us what exactly is wrong, or why it’s wrong.” Frank, an accountant, squinted and used his left thumb and index finger to rub his eyes. He might have been pushing back tears, but the action reminded me of the way he used to react to his lab results, before his disease was in remission, when he tried to make sense of the abnormal numbers I was showing him. He makes his living by making numbers fit; he couldn’t make his own numbers fit back then, and now he couldn’t make sense of his daughter’s case, either.

My hospital — in fact, my entire university — has a new research mission: precision medicine. The other term for this is personalized medicine. Regardless of terminology, the goal is clear. Using advances in DNA sequencing, we can run a patient’s entire genome in a few days at less than $2,000. By comparison, this same effort a decade ago took years and cost about 3 billion dollars. The promise of such rapid sequencing is that we can tell a patient, particularly one without a diagnosis, what disease she has, what in her genome is responsible for this disease, and which available therapeutics might be able to help. If the therapies are not available, then, with the aid of large sums of pharmaceutical industry money, we should be able to create them.

This mission was announced last spring with big fanfare by David Goldstein, our new Director of the Institute for Genomic Medicine, recruited away from a rival university. In his introductory lecture, Dr. Goldstein relayed the story of two children who were facing certain death until whole genome sequencing uncovered the roots of their diseases. He fought back tears when he told us how one of the patients, with just a simple therapy of vitamin B, went from being completely immobile to slapping high fives with her researchers.

I sat in the audience thinking of my mentor from fellowship, a man named Phil Klemmer, who often referred to the results of genetic testing as “toxic knowledge.” You and your patient have this information now, he’d warn, but what in the world are you going to do with it? These conversations occurred almost a decade ago, and a year in genetics research is about as weighted as a year in a dog’s life. Dr. Klemmer used his “toxic knowledge” admonition primarily in discussions about testing family members of patients with polycystic kidney disease, one of the few inherited kidney diseases we could test back then. Today, I can send a patient into a commercial lab and test him or her for dozens of genetic kidney disorders with a single blood draw.

Most genetic diseases are currently untreatable or “treated” only by relatively extreme measures, like prophylactic bilateral mastectomies in response to a breast cancer gene. The high-fiving girl on vitamin B therapy is the exception and not the norm, Dr. Klemmer would argue. On the other hand, his younger colleagues called him a dinosaur for that thinking, and genome sequencing technology is currently the best chance I have at answering the question that nearly every patient poses to me at his or her first appointment: “Why did this happen to me?” The information that comes out of trying to answer this question with “precision medicine” might, in turn, provide the answer to this same patient’s next question: “Can you make me better?” As Dr. Goldstein finished his talk, my thoughts drifted away from Dr. Klemmer and toward Frank, who six years ago would have asked those questions about himself but now was asking them about his daughter.

In the first essay of his new collection, Changing the Subject: Art and Attention in the Internet Age, Sven Birkerts compares his fears in the 1980s of nuclear fallout to his current concerns about “the stimulus barrage of modern life.” Today’s era of unlimited access to data has created a pandemic of TMI (too much information) that has replaced Birkerts’s fear of deadly contamination from a wholly different TMI (the acronym 1980s journalists used for Three Mile Island). Birkerts is interested in more than just clever wordplay (the TMI vs. TMI comparison is his, not mine). “Many of us,” he writes, “feel a nagging sensation of weightlessness and transparency, a persistent anxiety that the larger, deeper, satisfaction of living is missing.” This despair is the true subject of his essays. By subscribing to an internet-connected global community, and by subjecting our health to the results of genomic sequencing, are we giving up our own opaque individuality for access to what will, in the end, just be toxic knowledge?

A medical school professor once refused to answer a question my classmate posed because she began with the word, “Why.” “Why questions are meant for philosophers and religious scholars,” this professor said. “Doctors should only try to answer how questions.” His response seemed cheeky then (circa 2000) and today would sound antiquated. Nearly every major university hospital in this country is launching its own version of a precision medicine program to match (and receive funding from) the $215 million Precision Medicine Initiative announced by President Obama in his 2015 State of the Union address. These programs and this Initiative are built upon the idea that why questions in medicine should indeed be asked and, when answered, will provide much more than toxic knowledge. A White House press release touts precision medicine as “a new model of patient-powered research that promises to accelerate biomedical discoveries and provide clinicians with new tools, knowledge, and therapies to select which treatments will work best for which patients.”

I think it’s okay for doctors to ask — and try to answer — why questions, because these are the questions our patients are asking. I’m cautious, though, about guaranteeing that these answers will, with rare exceptions, rise above the kind of toxic knowledge described by Dr. Klemmer and Sven Birkerts. Still, I’d argue that, from a doctor’s perspective, why questions represent the ultimate search, and “search” is a deliberate word choice. My first exposure to Birkerts’s writing was a 2004 essay in The Believer (“The Possibility of the Search,” later reprinted in the collection Reading Life: Books for the Ages) about Walker Percy’s The Moviegoer, in which Birkerts analyzed the protagonist’s existential curiosity, what Binx Bolling repeatedly referred to as “the search.” The process is more important than “the find” for Binx, by which I mean the search itself is what keeps him alive and wards off malaise. Binx and, if we trust his biographers, Percy too, ended their search in the tenets of Catholicism. Birkerts, in that 2004 essay and in virtually every essay of this most recent collection, ends his search in books, in reading, and, specifically, in the act of contemplating what he’s read. In “The Hive Life,” Birkerts writes that “an intent and properly contextualized contemplation of almost anything can create around it a sense of gravitas. [It’s] all very Walker Percy, this hypothetical play with the metaphysics of mattering.”

All this is a somewhat longwinded way of answering a question a patient posed to me the other day in clinic. In the waiting room, he’d been reading a hospital-issued pamphlet about precision medicine. “This is some high tech stuff you guys are doing,” he said with a smile, stuffing the pamphlet into his backpack. “You’ll do anything to help your patients, right?” He clearly wanted an affirmative answer, and for the sake of time I nodded and said yes so we could move on to the reason for his visit. But the true answer to his question, at least for me, is more complicated than just helping patients. Doctors will do anything to keep open the possibility of a search. We’re curious about genes for the same reasons Birkerts dives into Ralph Waldo Emerson, Virginia Woolf, Teju Cole, and Joseph O’Neill. We’re looking for answers. If we stop looking for answers, if we relinquish “this hypothetical play with the metaphysics of mattering,” then we stop living in a way that accrues meaning. Better to keep looking for answers and scoff at the notion that there could ever be too much information, even if we run the risk of trespassing onto the turf of religious scholars and philosophers.

Genome sequencing might reveal a mutation for a patient with a mystery illness, but in most instances we won’t know the function of the mutated gene and its encoded protein. The next step is to force the mutation in an animal (a “knockout” model in which the gene is literally removed from a mouse’s DNA) and see if the mystery disease develops. This knockout model also provides insight into all the potential functions of the mutated gene. In the best-case scenario, the mutant gene only encodes one function relevant to the mystery disease (e.g. an enzyme that allows us to absorb a specific vitamin) and that function, once exposed, reveals a targeted therapy. In the more realistic scenario, the gene’s encoding function is complex, and any attempt at targeting that function (assuming a targeted therapy already exists — in most cases it doesn’t) runs the risk of affecting myriad body systems, necessitating an often lengthy trial in an animal model before moving on to human subjects.

At my suggestion, Frank made an appointment for his daughter with my hospital’s Discover Program for Undiagnosed Disorders. Whole genome sequencing identified a mutation in his daughter associated with Costello syndrome. I told Frank that I’d never heard of the disease. He wasn’t surprised. “There’s less than 300 cases in the world,” he said, “and there’s only one doctor in the country who’s seen more than one.” He and his wife had already spoken on the phone with this doctor and were planning a trip to Philadelphia to see him. “But he already warned us over the phone that there wasn’t much he could do, not yet,” Frank continued. “Right now, the best care is intense physical and occupational therapy, which we’re already doing.” I asked Frank if it helped having a diagnosis, a name, and an explanation for his daughter’s symptoms. I’m sure the geneticists who’d tested Frank’s daughter were thrilled about finding a mutation and “diagnosing the undiagnosed,” a phrase typical of precision medicine publicity materials, but did their efforts truly help the patient and her parents? “Yes and no,” Frank answered. “It will help when there’s a real treatment, but at least we know what we’re trying to treat.”

The precision medicine initiative boasts that sequencing can be done within days with minimal costs. But the fact is that sequencing is just the first step of what may be a multiyear process for the patient and physician. To be sure, Frank and his wife can now tailor their internet reading toward a specific disease, rather than just Googling a combination of their daughter’s symptoms, but having a Google alert for “Costello syndrome” will not impact them or their daughter today. They still must wait for the doctors to make sense of this mutation in their daughter and the other hundred or so children with the same mutation. Birkerts’s Cassandra-like warnings about internet search engines and GPS devices therefore fit precision medicine better than my mentor’s admonitions on toxic knowledge: “We are not seeing the triumph over the unknown. We are seeing, rather, the differential between what we can achieve and what our superengineered machines can achieve. Where their reach concludes, the unknown resumes, and it is no less infinite than it was before.” The search goes on, to the delight of doctors and hopefully to the eventual benefit of our patients, and the search will always go on, I expect, regardless of how precise our medicine becomes.


1 To protect patient confidentiality, Frank’s name and identifying information have been changed.


Andrew Bomback is a physician and writer in New York. His essays have recently appeared in The Millions, Vol. 1 Brooklyn, Ohio Edit, Hobart, Entropy, Human Parts, and Essay Daily.