Covering Disruptive Technology Powering Business in The Digital Age

Hospital analytics teams are on the verge of a formidable challenge: evolving their skillsets as fast as data is growing and relevant technologies are coming into play.

“The skills your team is going to need in three years are not the skills they have today,” said John Showalter, MD, chief healthcare information officer at the University of Mississippi Medical Center. “They’re going to need to be trained.”

And that training is going to be complex because there are only a few people who can really master analytics well, according to Sriram Vishwanath, a professor of engineering and computer science at the University of Texas, Austin, even if many claim they can.

“If anyone tells you they can do healthcare big data, I’d take that with a truck load of salt,” Vishwanath said. “Healthcare big data is hard.”

That’s why leading health systems including Advocate Health Care, Geisinger Health Care System, the University of California San Francisco and UMMC are assembling specialized teams.

Critical skills
Geisinger chief data officer Nicholas Marko, MD, rattled off an array of skills that enterprise analytics teams will need: statistical analytics, dashboarding, interactive visualization, predictive analytics, modeling and simulations, algorithmic research and development, as well as data-centric collaboration and real-time analytics.

Danyal Ibrahim, MD, chief data and analytics officer at Saint Francis Hospital and Medical Center agreed that future analytics skills will be different than what exists today and sees some skills already shifting to new departments.

“SQL skills are more data science, even data warehousing, architecture and design – those can live partly in the analytic team rather than pure IT,” Ibrahim said.

While that may raise some eyebrows, Ibrahim suggested working through suspicions by showcasing early success stories and Marko recommended identifying key people and making them feel like part something bigger in working with the analytics team.

Showalter added that even something as simple as teaching analytics teams about existing features in software programs already in place can help. At UMMC, they showed certain employees how to use pivot tables in Excel, for instance, which got many of them thinking about visualizations as a gateway to more sophisticated data presentation techniques.

Analytics teams core competencies
Advocate Health Care, for its part, focuses on three core competencies: technical, analytical, and deployment, according to Tina Esposito, vice president of Advocate’s Center for Health Information Services.

On the technical front, that means having employees who work to pull data together, use tools including Hadoop and generate appropriate data models. For the analytical piece of Advocate’s team, Esposito said the aim is to ensure they can leverage the technical aspects by having data scientists and statisticians. And the deployment function encompasses unsung heroes — experts with a clinical background who are focused on ensuring that what the technical and analytical teams develop actually make it to the front lines and get used routinely.

Danyal Ibrahim, MD, chief data and analytics officer at Saint Francis Hospital and Medical Center built his team on the following pillars: break down corporate and people silos, redesign the team for where you’re going, develop a strong data platform, design and adopt meaningful metrics and, ultimately, harness data presentation capabilities to make it appealing and actionable.

“The largest stores of data at Saint Francis live in finance, IT and quality,” Ibrahim said. “We need to bring them all together, create a forum for working through problems. Analytics is a team sport.”

Avoiding a tempting mistake
Showalter recommended steering clear of one mistake that can be all-too-easy to make.

“Don’t purchase tools your people cannot use, there are some you need a PhD in data science for,” Showalter said. “Don’t buy that. Outsource instead.”

Geisinger’s Marko echoed that, saying healthcare organizations need to understand exactly what they want before making any purchasing decisions — and that is not nearly as easy as it might sound.

“If you don’t know what you’re trying to answer you can certainly find someone to sell you a solution based on a bunch of numbers,” Marko said.

The common problem, however, is that a year later that product may or may not actually solve your problems.

‘Analytics is a people space’
Coordinating, managing, extracting value for the analytics team and ultimately thinking about the strategy really demands that one person be appointed to make ultimate decisions.

Whether that individual bears the title Chief Data Officer, Chief Medical Information Officer, even CIO or others matters less than empowering a leader who can speak the language of both executives and data scientists.

Penn Medicine Chief Data Scientist Michael Draugelis explained that the person in charge must also be able to interface with clinicians. “Usually they’re the compass heading in the right direction,” Draugelis said.

Marko added that the analytics team chief also has to able to manage a bright group of data scientists, which is particularly important because talent is extremely competitive right now, so a leader needs to know how to attract and retain employees while average turnaround time in this niche is about two years.

Marko also joked that CDO really stands for Chief ‘Diplomacy’ Officer because they spend so much of their time navigating the various departments and personalities of the broader organization.

“Analytics is a people space,” Geisinger’s Marko said. “It’s not a hardware space or software space. Those tools are important but absolutely your analytics enterprise will succeed or fail based on the people you have.”

This article was originally published on www.healthcareitnews.com and can be viewed in full

(0)(0)