What to Know Before Hiring a Data Scientist

Before Mercy Health System revamped its analytics, its nursing scheduling system wasn’t working well. Its shortcomings forced the 44-hospital organization, based in St. Louis, Missouri, to lean on expensive agency work. Nurses couldn’t predict what their weeks might look like, and they weren’t happy.

Eventually, with the help of advanced technologies from SAP Health, Mercy turned it around. The health system found $4.3 million in savings and reduced its turnover rate, the fruits of a wider high-tech overhaul. But there was another key component to this initiative: A few years ago, Mercy began hiring data scientists.

“They’ve more than proven their value on projects we’ve appointed them to,” Curtis Dudley, Mercy’s vice president of integrated performance solutions, told Healthcare Analytics News™ today at HIMSS 2018. “They’ve helped me build better products on the dashboard side, knowing how to correctly calculate a metric, which methods of aggregation I should use. They’ve just been wonderful.”

So, since data science is driving much of the change across healthcare, a natural question pops up for Dudley, who now has some experience in this area: What do health systems need to know before—and after—hiring data scientists?

It’s all about storytelling. Dudley said he searches for data experts who know how to turn numbers and insights into a coherent, captivating tale. It’s not enough to get the information. Data scientists must be capable of selling them to a room of executives or clinicians.

“A lot of them are good at the mechanics, but being able to tell a story with the data and be OK communicating that story and the concepts in front of people—that’s the hard part,” Dudley said.

During the interview process, he asks candidates to tell a story about a project they once worked on. Then he sits back and listens, picking up on the data scientist’s logic, language, and storytelling structure. If Dudley can understand the project and what it produced, then that particular storyteller might be a good fit for Mercy.

Another piece of advice: “When you can find them, you’ve got to hold on to them,” Dudley noted. There is great demand for data scientists in healthcare and nearly every other field, which means they have legs.

So far, in healthcare, executives and others don’t yet properly value data scientists, Dudley said. Data scientists often aren’t paid well enough, and their titles don’t match the scope of their impact. They are, after all, the people who are reaping outcome-driven insights, hand in hand with clinicians, and targeting inefficiencies.

“They’re just not recognized the way that they should be yet,” Dudley added, “because I think that the work that they’re doing is life-changing.” And with all the stock that healthcare has placed in data and analytics, it might be time for the industry to reckon with that.

Related

The Healthcare Technology Winners of 2017

Data Science Bowl 2018: A Deep Learning Drive

The Underrated Challenges of Building a Learning Model