A multinational, multidisciplinary team apparently has won $1 million for improving the algorithms that Netflix uses to recommend movies to its users. The winning team — which includes statisticians, computer engineers and machine learning experts from the US, Austria, Israel and Canada — developed a program that improves the accuracy of movie recommendations by more than 10%.

Ix innovation design is a new focus for the Center for Information Therapy, in part through our collaboration with IDEO and the California HealthCare Foundation. We are eager to identify more robust methods for driving breakthrough innovations in patient-centered HIT applications.

For years, I’ve wondered why we see so few examples in the health care world of the Netflix or Amazon approach to guiding consumers to the information that they want. A huge success factor in information therapy (Ix) is the degree to which information prescribing can develop algorithms that effectively translate individual pieces of data into information triggers. The more data we can collect and efficiently transfer through electronic infrastructure and the more information derived from the consumers themselves, the greater the chance that we can prescribe or deliver the right information to the right person at the right time.

Netflix knows what movies I request and often how I rate them. They also ask me (in a very quick, convenient way) to rate other movies that I’ve seen. They put these data points together and somehow electronically try to make sense of someone who loved “The Killing Fields” and “This Is Spinal Tap,” but who was not enamored with “The Matrix.” They seem to do a reasonably good job, and perhaps this new program that was developed by the interdisciplinary team will deliver better movie recommendations to me.

That’s all fun, but it’s not my health — or my family’s health. There, the stakes are higher, and the information needs are several orders of magnitude more complex.

In trying to meet that need and challenge, the X-Prize Foundation has developed a health care competition, putting $10 million on the line. Unlike Netflix, however, this isn’t all about 1’s and 0’s. Changing the way health care is delivered involves a lot more than modeling scenarios on a computer.

That said, we can probably borrow a lot from the computer engineers, statisticians and machine learning experts and transfer those learnings to Ix innovation. To do so, however, we need those experts to join forces with clinicians, patients and ethnographers/anthropologists in order to make sure that what we learn has meaning for consumers and can be embedded into the clinical workflow.

This entry was posted on Monday, June 29th, 2009 at 9:23 am and is filed under Engaging Consumers, Innovation. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.