Panel Calls for Google Maps of Human Disease

An expert panel today called for creating a massive data network that would combine cutting-edge genomic and molecular data on patients' diseases with their routine medical records. Such a database would be a boon for research and help move medical care into the era of "precision medicine," the panel says.

The National Research Council panel was formed last year at the request of the National Institutes of Health (NIH) to consider how the more than 100-year-old system of disease classification should be changed to reflect insights from molecular biology. But the committee decided that "our challenge is bigger," said panel co-chair Susan Desmond-Hellmann, chancellor of the University of California, San Francisco, at a press briefing today. Instead of a new classification, the nation needs a live network of data on individuals' molecular tests and health records. This system would be used to develop a new disease taxonomy and personalize medical care, according to the 108-page report, titled Toward Precision Medicine: Building a Knowledge Network for Biomedical Research and a New Taxonomy of Disease.

Precision medicine is already emerging in cancer diagnosis and treatment, the report says: some patients now receive drugs matched to a specific molecular marker in their tumor, and relatives can be tested for certain cancer risks. By contrast, a middle-aged man diagnosed with type II diabetes typically receives a 50-year-old drug that may or may not help him. And no type II diabetes risk tests are available for family members.

What's needed, the report says, is for patient's health records to be combined with layers of genomic and other molecular measurements, such as blood proteins and the microbes in a patient's gut. Like the GPS data used to make Google maps, these data could be plumbed in detail by researchers and used more superficially by others, such as doctors to treat patients, the report says. Separate databases would be combined to form a single network.

Some modest efforts like this already exist. For instance, the health care organization Kaiser Permanente is building a genetic database on 500,000 patients in the San Francisco area that will be used for disease studies. "We want to do it on a bigger scale," said panel co-chair Charles Sawyers of Memorial Sloan-Kettering Cancer Center in New York City at the briefing. (Another example, he says, is a plan by the Faroe Islands to sequence all 50,000 of its citizens' genomes and use the data for research and health care.)

As a pilot project, the report recommends sequencing the whole genomes of 1 million Americans and combining the data with medical histories to look for genetic links to disease. That may sound expensive—even if sequencing costs drop to $1000 per genome, it would cost $1 billion—but $1000 is in the range of what a routine MRI scan costs, Desmond-Hellmann pointed out. Another pilot project would use metabolomic profiles of patients' blood to help predict which patients with insulin resistance will go on to develop type II diabetes.

Creating the network over the next decade or two shouldn't require new funding, the report says. "This is not the Human Genome Project," said Sawyers. "It's taking advantage of things happening anyway and bringing them together and doing it at the point of care." NIH needs to redirect resources and push for more long-term studies that combine research with health care, the report says. Building the network might also require a revision of patient privacy rules and an "evolution" in the public's attitudes about allowing researchers to use their medical data.