Craig Mundie thinks that if the U.S. really wants to solve its massive healthcare problem, it should subject it to big data practices.
Specifically, Microsoft’s (s MSFT) chief research and strategy officer said government and/or industry should bring the Internet model to bear on the problem of out-of-control healthcare costs, and that means sharing, not segregating, massive amounts of data: something that flies in the face of current HIPAA requirements. HIPAA is the Healthcare Insurance Portability and Accountability Act passed in 1996, which mandates strict privacy for patient records.
The big data concept, which calls for analyzing huge amounts of collected data — often in different formats — has become a rallying cry for vendors and customers that want to wring more value out of the information they already have but didn’t necessarily know how to capitalize on. Until now.
Speaking at the Techonomy 2011 conference Monday night, Mundie used the extensible business reporting language (XBRL) which underpins all financial reporting as an example of how the government can grease the skids for change. “XBRL was ginned up by the SEC …. and went from a standing start to full adoption by U.S. companies in less than four years,” said Mundie.
Between Medicare and Medicaid, the U.S. government pays for more than half (54 percent) of aggregate domestic healthcare costs. “If the government said, ‘I don’t care how the private sector does it; we’ll do it this way,’ the whole industry would flip,” he said.
Free data flow is necessary so costs can be compared between the providers that are so reluctant to share their information. The fundamental issue with healthcare is “perverse payment system”: Most people don’t care about what their care really costs because a third party foots the bill. “There is no braking component” to contain runaway costs, he said.
If, however, Americans paid based on outcomes, there needs to be a total view of data to allow comparisons between providers. But until the data is in a form that can be shared — with privacy concerns met — he insisted there’s no basis for comparison and no way to do quality control.
“We’re not serious yet about aggregating this data. If the data is not available, you can’t make wholesale change.”
He didn’t venture a guess as to how much could be saved overall, but said in a smaller test case, Microsoft (s msft) worked with a system of several hospitals that made their information available to a heath information exchange. “It slurped up all their Medicare data, and as soon as it had that data, they could query it to see just what was costing so much.”
Once the data is in one place, it’s not expensive to work with it, he noted, and it turned out that queries were able to identify a relatively small number of uninsured people who visited various emergency rooms around the system and ran up millions of dollars in charges. “Without this data, there was no way to find these people,” he said, adding that just eliminating redundancy and fraud alone would result in huge savings in a national context.
And, he insisted privacy issues can be addressed. Using metadata, “you can describe the provenance of the data to support sharing and also take privacy constraints into account,” he said. “Technology can be privacy-enhancing,” he noted.
But it’s highly unlikely insurance companies will drive this sort of change given the current privacy regulations and that their business is so profitable the way it is now, hence the need for a strong-spined government mandate.
Asked if these big data queries on patient care represent a “Google-like” (s GOOG) capability for healthcare, Mundie said it’s “really a Bing-like system for healthcare, but absolutely.”