In implementing BI, hospital officials face KM challenges
He also stresses the “stepping stone” nature of the process. “It is decision support evolution, in which you start small and grow from there,” he says. “We are still in the start-small phase.”
Second BI effort at Hennepin
Hennepin County Medical Center, a 469-bed hospital in Minneapolis, relaunched its clinical analytics effort in January 2010, It turned to a performance analysis tool called Insight Enterprise from Mediware, which merges data from a half-dozen clinical applications, including Epic Systems’ (epic.com) EHR, says Sam VanNorman, manager of performance reporting and analysis.
A previous BI effort had fizzled in part because the tool chosen wasn’t sophisticated enough, and in part because there wasn’t enough executive support at the time, VanNorman recalls.
“Executives would get a dozen indicators at the department level, but they couldn’t drill down,” he says. Now, the chief nursing officer is able to study measures of patient satisfaction, and then drill down in a report to look more specifically at where issues might be arising.
Besides the more sophisticated dashboards, Hennepin is building an enterprise data warehouse. “There is definitely a push from executives for a way to combine clinical and financial data,” he says. “For instance, they may want to look at the correlation between staffing and patient satisfaction.”
VanNorman says one valuable lesson learned from the effort is that it should be interdisciplinary rather than seen as an IT project. Hennepin has a knowledge management team, and VanNorman works closely with its leader as well as with information security, finance, a clinical group and supply chain experts. “It is important for them to be involved in developing the first 100 indicators on the dashboards,” he says, “since they will be responsible for answering when things appear to go awry.”
Although the feedback has been positive, VanNorman says he also has had to learn how clinicians want to receive reports. “Most of them are not quantitative data people,” he notes. “They like working with patients, not poring over data. They want to look at a report for five minutes, see something that needs correcting and move on. But they do like getting this feedback and working on performance improvement.”
Divurgent’s Sirois says that eventually hospitals will begin pulling medical device data into data warehouses to aid with clinical decision support. “That is further down the road,” she says. “Automating that process is difficult because there are a lot of moving parts.”