Health IT executives talk big data

BOSTON—How are larger organizations approaching opportunities in big data and health analytics? Executives from Boston Children’s Hospital, UMass Memorial Health and Partners HealthCare shared their thoughts at the Big Data Healthcare Analytics Forum on Nov. 21.

David Fairchild, senior vice president for clinical integration at UMass Memorial Health—which is pursuing an accountable care organization shared savings arrangement next year—spoke about its efforts using analytics in its ICU telemedicine monitoring program.

An algorithm helps the hospital predict “what will happen before it happens,” and in the process reduces mortality, he said. Specifically, the team is able to detect subtle changes in patients’ vital signs before they become apparent and work with the patient’s care team to quickly remedy problems.

This program allows for monitoring of patients in smaller medical settings, so patients can avoid trips to the larger hospital. “It has a very nice element to it.”

Other areas where UMass Memorial has sought out opportunities for big data include aggregating claims and hospital ambulatory data from a population health point of view, Fairchild said.

When asked about opportunities in big data, Jon Bickel, senior director of business intelligence and clinical research IT at Boston Children’s Hospital, said the 300-bed institution is exploring the use of predictive analytics to reduce length of stay for high-risk patients. The goal: free up space during the busiest times of the year, including October and February. Buy-in still is an issue, he said, and leadership must be willing to recognize any excessive length-of-stay in units.

James Noga, vice president and CIO of Partners HealthCare, said population health management is critical for its at-risk contracts. As such, “we look to find out what’s directing population and volume,” with an analysis on how patients utilize services, like home care and specialists, and whether there are unnecessary expenses.

Insights are sought on high-cost patients identified through risk stratification, and Partners looks at total medical expenses and “drills into cohorts,” he said. It has developed a dashboard with prescriptive and predictive analytics capabilities to target care services, he said.

Buy-in to achieve data analytics initiatives is never easy. “It’s a difficult time for clinicians” as hospitals ask them to implement “one more thing,” said Fairchild. “Here’s another button, here’s another dashboard,” and solutions that function outside of the EMR are a “non-starter,” he said.

To improve matters, “I focus on things to make physicians' lives easier,” he said.

“Docs are very busy, changing practices has to be done in a delicate way,” added Bickel, noting that it’s tough to maintain their interest, as crises change on a daily basis.

To ensure physicians stay engaged, Noga said a good component of their financial compensation requires that they focus on population health.

However, he said that physicians are increasingly demanding evidence that health IT is working as intended. He said they currently are combing through 2,500 clinical decision support tools and eliminating those that lack evidence.  

In lessons learned, Noga said they’ve learned to do things in a controlled route before they propagate out. Also, “we’ve come to the realization that we need to invest in true data scientists,” he said. “They really drive the appropriate use of data.”

He also advised to engage universities to access students studying data science. “Most universities are mission-driven,” he said, thus students are willing to work in community hospitals.