Researchers from technology companies and universities are trying to figure out how to apply analytic tools to a wide range of information to improve healthcare and find new cures for diseases. Most of these "big data" efforts involve the use of clinical records. But an article in the Atlantic says researchers could achieve more by analyzing data from patient surveys and other sources than by scouring electronic medical records.
Author Robert Litan, vice president for research and policy at the Ewing Marion Kauffman Foundation, which recently issued a report on big data, admits that lab and test results might be useful in this kind of research. But other data that could be important--on patients' health behavior, life and job histories, and genomic makeup, for example--is either not included in clinical records or is inaccessible because it is locked up in physician notes, he writes.
"What researchers could most profit from are databases that link genetic information, medical histories, areas where patients have lived and worked, and how patients behave," says Litan, who is also a senior fellow at the Brookings Institution.
Litan cites an ongoing study by Susan Love. M.D., that has so far collected self-reported data from 365,000 women who are "concerned" about breast cancer. Only some of the women have had breast cancer, he points out, so the study should help researchers compare the histories of the two groups to identify factors that might cause the disease. He urges organizations fighting cancer and other diseases to follow the example of Love's campaign.
Other researchers who are trying to harness big data in healthcare take a somewhat broader view of suitable data sources. For example, scientists at the State University of New York at Buffalo recently announced they were partnering with IBM and Revolution Analytics to fight multiple sclerosis by analyzing huge databases, according to InformationWeek Healthcare.
These researchers are seeking correlations among genetic, economic and environmental factors that might play a role in causing MS. They will utilize medical records, lab results, MRI scans, and patient surveys, plus genomic data obtained from the National Institutes of Health and other sources.
In the Kauffman Foundation report, 31 experts suggested that nonprofit organizations that study disease should put together a national health database. In the future, they said, job-related health benefit packages should instruct employees on how to contribute personal health information to this database. Of course, they added, the data should be de-identified to protect privacy.
The Obama Administration, meanwhile, has allotted $200 million to a big data initiative that will focus on several industries, including healthcare and health IT.