Take healthcare. The advent of the electronic medical record (EMR), for example, has suddenly created a data bonanza for healthcare providers and academic researchers studying healthcare challenges. With so much data, solving big health issues should be easier than ever.
Now, hang on a minute. Despite what lots of people out there think, having lots of data doesn't mean being able to answer the questions. People sometimes forget about the at times arduous task of turning that data into insight. As Brian Denton, a healthcare researcher, told me in a recent phone interview, "It's applying the analytics model that makes it possible to turn data into useful information that doctors can use at the point of care."
INFORMS Healthcare 2013 conference that took place in Chicago in late June. In its second year, the conference is a reflection of the growing momentum in healthcare analytics right now. "There's an enormous growth in the number of problems people are working on and the number of people working on them," said Denton, who, among other credentials, is an associate professor in the Department of Industrial and Operations Engineering at University of Michigan and INFORMS secretary.
The excitement in healthcare analytics comes on two fronts: The first is healthcare delivery and management, and the second is the medical side of healthcare. Denton's interest and work falls in the second area, where his experience exemplifies how increased amounts of data, when poured into sophisticated analytics models, can change patient care.
For the past five years or so, Denton's team has been collaborating with investigators at the Mayo Clinic in Type 2 diabetes research funded by the National Science Foundation. The goals are to help define better treatment guidelines for high blood pressure and high cholesterol -- two major factors for diabetes -- at a national level, and to provide decision support to physicians at the point of care. In a nutshell, Denton said, "We've been using large amounts of historical data to develop models that can be used to simulate what happens when you apply guidelines for treatment to patients with Type 2 diabetes."
Initially, the team looked at how to control one risk factor. "We asked a simple question, that is, 'If and when should patients start using statins to control or lower their cholesterol?' "
Then, over time, it began looking at multiple drug treatments. "There are more drugs than just statins for controlling cholesterol, so we looked at using quantitative models to define a plan for how to use these kinds of medications together to reduce cholesterol."
From there, the team branched out to looking at other risk factors -- and along the way developed a model for blood pressure independent of what it had done for cholesterol. Then came bringing the two together into a model that "coordinates drug treatments for both cholesterol and blood pressure," Denton said.
That brings us to the present, and the team's exploration of blood sugar.
- The breadth of our model is growing in terms of how comprehensive it is regarding medical treatment decisions for patients with Type 2 diabetes. And the data we've been using has been growing as well. Our ability to harness large datasets to calibrate the models we've been working with has improved over time.
In its initial work, the team used data collected as part of a small study at Mayo Clinic. "So that's a very specific population of people in one region." More recently, it's been with a much bigger and more comprehensive dataset compiled from a far-flung population.
The dataset comes from more than 100,000 people across the US who have Type 2 diabetes, and includes data on various medical factors, such as blood pressure, cholesterol, and blood-sugar levels, and how they've changed over a 10-year period, Denton said. "Twenty years ago, we just didn't have access to datasets like this. It's the availability of the data that's making these kinds of studies possible."
And it's groundbreaking analytical work that's making the difference between having data and gaining meaningful intelligence from it. Denton's group is the first to develop an analytics model for optimizing treatment decisions.
As yet, however, the group is still doing the groundwork for the decision support system physicians would use at the point of care. Denton said he hopes to begin developing the point-of-care system within the next year. "We're getting close to having most of the risk factors covered in our prototype software, and at that point it becomes a question of implementing in a decision support system for the point of care."
What other innovations in healthcare can we attribute to more data and better analytics models? Share your ideas below.