“In my bioethics class, I leave students with one idea: Question everything,” says Craig Klugman, a professor and chair of the Department of Health Sciences. “That advice applies to the role of big data in health care specifically and in our lives generally.”

As a bioethicist and medical anthropologist, Klugman thinks about the evolving impacts of big data—both positive and negative—on intangibles, such as privacy and accountability.

“I’m 100 percent in favor of using technology to improve quality of life,” he says, “but every week a story breaks about some company that knows more about us than we think it should. Many companies use big data just to make money—not a problem in and of itself, but a merely commercial goal makes us a little wary. Big data is amazing, and the insights we can gain are often laudable, but are we paying a price? Maybe. It’s important that we keep asking that question.”

Here, Klugman talks about state-of-the-art bioinformatics (the use of computer science, statistics and engineering to analyze biological data) and ethics in the health care industry.

Q: Is big data changing health care delivery?

The answer is “yes and no”—or “not yet, but probably soon”—depending on the sector.

The Affordable Care Act requires physicians to keep electronic medical records [EMR]. Does that make data more available to diagnosticians? Not necessarily. Because different EMR systems don’t talk to each other, different institutions can’t share information. Nor would they necessarily want to, since they’re competitors. So, here the use of big data is still nascent: A patient who needs to go from one hospital or network to another basically starts over.  Within a network, big data can be analyzed to improve processes; but outside a network, no.

Insurance companies have been tracking massive amounts of data for decades, of course, and now they can do that more efficiently. For example, they can buy an AMA database of every prescription written by every physician; they can know what medicine you’re taking, whether you’ve filled a prescription, and whether you’re compliant with dosages. Also, they’re bringing together data from multiple sources. A case in point: In the 1970s, death rates from surgery were at 1.1 percent. By correlating big datasets, hospitals and medical schools were able to see how to change practices and achieve fewer errors and better patient outcomes. Now, the death rate is 0.1 percent.
In public health, we can now do a better job of tracing diseases, causes and outcomes based on all kinds of variables, from geography to patient ethnicity, age and lifestyle demographics. Here, big data is really useful in changing health behavior. In the past, researchers would have to trace thousands of people over decades to discover optimal health activities. But, with today’s computing technologies, they can quickly look retrospectively at 20 million people. That’s a great productivity improvement.

In pharmaceuticals, we’re seeing real innovation because of the emerging science of genomics, which uses big data to understand and draw conclusions about genetic information. In the future, we could be seeing targeted therapies in which specific drugs will work only on people with a specific gene. Theoretically, personalized medicine should deliver better outcomes with fewer side effects. That’s one way big data could have a huge impact.

Q: So, what’s the “question everything” angle?

Problems can develop when people are treated as data points. Some of the most egregious medical cases, such as the Nazi medical experiments and the Tuskegee syphilis study, came about when human beings were viewed as objects.

When we ask “What can we do with technology?” we should also ask “But should we do it?” For example, when it comes to using big data in business, we have to take a look at a loss of privacy. For what we give up, do we get enough in return?

Also, there’s always a risk in interpreting data. When you have big data sets, you’ll find correlations. But how does the analyst separate coincidences from meaningful, cause-and-effect relationships?  Even more important, perhaps: How does the data gatherer deal with incremental, unintentional findings?  Let’s say company X is studying disease Y and finds out subject Z has a different, unrelated problem. Should that person be told? Other questions: Does a person’s genome predetermine his future? Could people be discriminated against based on their DNA? Or because of “bad” behavior? Sounds like dystopian science fiction, but that kind of oversight into our personal, private lives is not that hard to imagine.

We need to be careful not to separate big data from its social context.

Q: How can educators make a difference?

When DePaul was exploring the ways we could prepare our students for a big data world, a committee was formed that included multiple colleges, and that was the right thing to do. Together, we got the big picture. As a result, we now have two new masters programs and several, industry-specific certifications within these. Our approach is exactly right, I think. At the same time, DePaul’s values tend to put ethics front-and-center in everything we do. In the end, that will make our students—high skilled, ready to go—really stand out in this growing market.