Optum's big data push gets bigger

In January 2013, the Optum business at Minnetonka-based UnitedHealth Group teamed with the Mayo Clinic to launch a lab in Boston for research to improve the quality and lower the cost of health care. Called OptumLabs, the center was designed to work on "big data" research projects by collaborating with outside groups. In November, OptumLabs announced the U.S. Department of Health and Human Services was joining the initiative — an important step forward, since the federal agency is the nation's largest single health care payer by virtue of the Medicare and Medicaid health insurance programs. Dr. Paul Bleicher is chief executive for the group, which directly employs about 40 people and now partners with more than 20 outside groups. He says some of the research published in the past year shows the promise for OptumLabs going forward.

Q: Can OptumLabs point to any successes where research projects improved the understanding of cost or quality in health care?

A: We've had a number of interesting studies to date. Two of them, actually, came from the Mayo Clinic to address those two topics.

One in the spring of this year was a study that looked at the new anticoagulants … in comparison to the old standard of care warfarin in patients with atrial fibrillation. That study was examining the safety of the medications, especially around bleeding risk.

They were able to do something with the data because there was so much of it and there was so much depth to the data. They were able to identify that bleeding risks changed with these newer anticoagulants in older patients — I believe it was about age 75 where it switched. That's a very important and subtle finding that will hopefully with other evidence make its way into medical practice to help improve the quality of care and the outcomes of care.

The other study, which was very recently published, addressed … the overutilization of hemoglobin A1C testing in individuals who had stable Type 2 diabetes, and found that there was significant over-testing. There's an opportunity for something that is a very common test to be managed into a more appropriate testing interval for reduced costs.

Q: How would you compare the health care industry with other industries in the use of big data?

Health care is very different. First of all, there are very serious privacy issues, which have to be carefully dealt with in terms of the management of the de-identified data that we use. And when necessary, there are the appropriate regulatory approvals, etc. You can't simply then quickly try things out in a bunch of different ways.

If you think about what they say in carpentry — measure twice before you make your cut — in health care you've got to measure a whole bunch of times and be sure before putting something in place. That's because it's expensive to put things into place, and most things … have the potential to improve things but also have risks associated with them. The appropriate thoughtfulness and balance and research is really essential.

Q: How do you measure whether OptumLabs is doing what it should do?

A: We've been speaking of OptumLabs solely in terms of research productivity, and OptumLabs actually has much more than that. We are involved as well in translation, which means taking findings that you identify in research and putting them into practice. And we're involved in innovation.

One of the ways we can measure is simply on the quality of the publications that come from OptumLabs. … But we also want to look at the ways in which they're being implemented by our partners, and the changes that they're bringing to the delivery of care. Finally, we have begun — it's early days — but we have begun to develop innovations at OptumLabs that are hopefully going to be incorporated into products that will help to improve health care in general.

A: One of the things that OptumLabs is really interested in at this point is the voice of the patient. We're interested in being able to identify data that can be — again, with privacy in mind, in a de-identified fashion — combined with our data to be able to get data that track more directly to the patient's experience.

We're living at time when there is simply an explosion of information about patients, whether it's from devices that they might wear or information from the Internet of things — which is beginning to provide data from your car from your home you name it — as well as social networks and more formal surveys. When we can begin to put that together, this represents the Holy Grail.

It's an opportunity to truly understand patients not just from their interactions on an irregular basis with the health care system, but on a day-to-day basis. And I should actually say consumers, rather than patients, because we're also interested in wellness and making sure that consumers don't become patients.

Christopher Snowbeck • 612-673-4744