Personalized Medicine: How Physicians Fare

Doctors are already using genetic data to customize patients’ treatments, but they could soon be overwhelmed by logistical and legal issues arising from inexpensive genome screening, experts said at a colloquium cosponsored by AAAS.

Within the next two years, people will be able to get their entire genomes sequenced for less than $1000, the speakers said. But many doctors lack the tools to use this information, they cautioned, and could face malpractice suits if they ignore or misinterpret the data.

The $1000 screen “is a 10-mile-wide asteroid heading toward us,” said Hank Greely, a Stanford University law professor. “We don’t have a clue how we’re going to handle it.”

The conference, held 8 to 9 March and cosponsored by the Food and Drug Law Institute, the Sandra Day O’Connor College of Law at Arizona State University, and the Mayo Clinic, was the third personalized medicine colloquium organized by AAAS in 2009—2010. The latest event offered a unique perspective on changes that may be in store.

Personalized medicine is “a very slow train,” said Lee Hartwell (who received the 2001 Nobel Prize in Medicine), in part because it relies heavily on a small number of genetic markers that predict disease.

Hartwell, codirector of the Center for Sustainable Health at Arizona State, said the risk of disease associated with these markers often isn’t large enough to help doctors. “I don’t have a lot of hope of taking your DNA sequence and predicting your disease susceptibility,” he said. “Family history is still a better predictor than that for most common diseases.”

Protein biomarkers are much more useful than genes for the doctor in intensive care who needs to know whether a drug is causing a patient’s kidneys to shut down, said Prasad Devarajan, the director of nephrology and hypertension at Cincinnati Children’s Hospital. It’s a problem that’s “crying out for personalized medicine,” he said, since drug toxicity varies considerably from patient to patient.

For example, Devarajan has studied the protein NGAL, which is produced by injured kidneys. It can be detected in urine days before other signs of poisoning appear, he said, and that makes it a good example of how biomarkers can be used to tailor real-life treatments.

For Donald McAlpine, the trove of genetic data related to antidepressant medication is a “scouting report” that “makes prescribing more complex, but ultimately better.” McAlpine, an assistant professor of psychiatry at the Mayo Clinic, underscored the point by unfurling a scroll of genetic variations so long that it spilled over the edge of the podium. “All this information is hard for the busy clinician to manage,” he noted wryly.

But he and others suggested that the matter has serious ramifications. “Health care professionals are likely the most vulnerable to liability risks associated with personalized medicine,” said Gary Marchant, a professor of law and life sciences at Arizona State. Many physicians unfamiliar with genetics,

he noted, may soon encounter patients who want their doctors to use genetic data in treatment decisions.

Kenneth Offit, chief of clinical genetic service at Memorial Sloan-Kettering Cancer Center, led a 2003 study that found some women with mutations in the BRCA gene might reduce their risk of several cancers if their ovaries were removed. Eighteen months later, he said, “we saw the first malpractice suit” against a doctor who had not removed the ovaries of a patient with the mutations.

“This meeting revealed the still enormous gap between our ability to generate large volumes of data and to offer targeted treatments,” said Mark S. Frankel, director of the AAAS Scientific Freedom, Responsibility and Law program. “The gap is the greatest challenge we face in achieving a level of personalized medicine that makes a real difference in health care.”