How Gwinnett Medical increased effectiveness of its analytics effort
Analytics applications can help organizations improve operations—but first, the individuals using those systems must have faith in the data and be able to quickly find the results they need to take action.
Those have been important keys for getting value out of data, and making it actionable, according to Beth Grimes, director of enterprise data analytics at Gwinnett Medical Center (GMC) in Georgia.
The organization has been able to increase the effectiveness of its analytics efforts by pursuing an organized approach to using analytics, which has helped with acceptance of results and faster action on what the data is showing.
The program has been most helpful in GMC’s intensive care units, where the organization started an intensivist program earlier this year. In this program, it wanted to start an initiative to measure how the hospital and physicians performed on a variety of measures, and then use that as a springboard to determine best practices in the ICU.
More broadly, the organization wanted to build a culture of use of analytics by front-line clinicians, with trust both in the data on which findings are made and in the results that are found. It’s essential to build buy-in within a large healthcare organization, staffed by a relatively small analytics staff.
The medical center is part of Gwinnett Health System, which also operates Gwinnett Medical Group and Sequent Health Physician Partners. The system employs 5,000 associates and has 800 affiliated physicians serving more than 400,000 patients annually. Gwinnett Medical Center is a nationally recognized, not-for-profit healthcare network with acute-care hospitals in Lawrenceville and Duluth.
Grimes says GMC had data solutions that were functional but not well-used, so it was looking for a new solution to build a business intelligence solution that would provide dashboards to users and support the hospital’s new intensivist program.
Because data was coming from a variety of systems, clinicians sometimes questioned the results they were seeing from reports, and clinicians couldn’t manipulate data themselves, resulting in time lags and delays if results were questioned or if refinements were sought, Grimes says.
The hospital looked at some ICU-specific analytics solutions as well as some enterprise solutions. Ultimately, the hospital selected Dimensional Insight’s Diver Platform, an enterprise-wide business intelligence platform.
But putting in the system was only a first step, Grimes says. Getting the physicians to embrace it was a separate challenge, one with which many hospitals struggle. Grimes says the attributes of Diver and the way that GMC approached physician performance management made the buy-in process go much more smoothly.
She outlined three keys to achieving support for the system and the growing use of analytics by clinicians:
Building trust in the data. The organization had a plan to overcome physicians' data distrust by setting defined measures with concrete definitions for each measure, resulting in greater provider trust.
Not jumping the gun on measurement. GMC obtained an initial baseline on certain measurements prior to performance rankings to create a positive change environment, allowing for open discussions and decreasing team conflict.
Eliminating inconsistency across systems. The organizations ensured that the same metrics are used across all care systems to accurately identify physician-led opportunities for performance improvement.
The biggest concern on the physician level, Grimes says, was whether the data was accurate. “Once we were able to prove that physicians can trust the data and that they can access the data down to the patient level, it was very quick and very easy,” says Grimes.
GMC received a set of defined measures along with definitions for each of those measures. Grimes says the measure set was critical, as any user can easily look up the definition for a measure to know exactly what it comprises, resulting in even greater trust in the data.
A second concern for physicians was that they would be measured on their performance right at the start of the intensivist program. However, GMC decided at the outset that it would not assess performance in this manner. Rather, the hospital decided to focus on obtaining a baseline on certain measures. From that baseline and insights into the data, GMC would then begin to measure best practice changes for the intensivist program. The organization used Diver to create unique time period comparisons to compare (i.e. the first six months of the new intensivist program to the last six months) the progress they’ve been able to make.
Once the intensivist program as a whole has been able to make progress towards achieving the best practices they’ve defined, then GMC will start to use the data to identify physician-led opportunities for performance improvement. Grimes says this approach to physician performance management is critical, as it creates a positive change environment, allows for more open discussions, and decreases team conflict.
For clinical nurse specialists, who have quickly adapted to using the Dimensional Insights program for readmission and mortality data, it was important to demonstrate to them that Diver uses the same Centers for Medicare and Medicaid Services metrics that they’ve been used to seeing from other systems.
“We want to use the data to improve performance across the board,” Grimes adds. “As we align our best practices with our data, down the road we can start to identify the smaller opportunities where those best practices continue to need to be tweaked, but for now, that’s not the focus.”
Grimes is working with leadership at GMC to get them in the routine of using the analytics system on a daily basis. Then, it will help leaders use the analytics system to find the answers they need—enabling them to dive down into the details and to make decisions in their meetings.
“That’s the most exciting thing,” she concludes. “People are reaching out and asking ‘How do I find this?’ They know the data is there—it’s at their fingertips, and they don’t have to wait for me to get back to them. We’re a big hospital system with facilities across the town—and for an analytics staff, we have just me and three other people. Since we have 5,500 people in our employment, it’s important for them to answer their own questions. I don’t have to have a meeting about it every time.”