Dean Clinic, a multi-specialty clinic serving southern Wisconsin, began creating a clinical analytics unit in 2007. The team works closely with the organization’s decision support, which is more focused on traditional business metrics, says Jennifer Close, vice president of operations. CMIO spoke with Close about Dean’s business intelligence efforts.
Why did Dean Clinic need a clinical analytics unit?
In my experience, analysts who think about business information and think about care delivery typically approach data differently. Business analysis tends to be more linear while clinical analysis is more curvilinear because care delivery is more curvilinear. No patient is the same as another and care delivery is a different experience for each patient.After Dean got a new CEO in 2006, the new CMO and I came over from Dean Health Plan. The plan had a well-established, robust clinical analytics division but the clinic did not. A gap analysis indicated that clinical analytics was the best way to help the organization achieve its goals by providing a richer and more robust story about the care and quality delivered to patients.The hardest part of setting up the unit was finding qualified candidates willing to make a leap from the health plan side. That’s where most of this talent lies. The unit now has five members with an open leader position. I would love for the team to be bigger but we’re investing in it slowly.
Was it difficult getting support for the clinical analytics unit from Dean’s leadership?
I was lucky that both the chief medical officer and chief operations officer already believed in the value of the skillset. For others it may be a hard sell because it’s really difficult to explain the difference between business analytics and clinical analytics and that you need both. We have found value in appreciating the synergy that really allows us to understand how the quality and service component connects to the value and cost components.
Is there an increasing focus on business intelligence in healthcare?
Business intelligence and healthcare analytics have always existed on the plan side. Health plans have much more experience leveraging their data. Now, providers are starting to understand that they need more robust information—not just for reporting but true analytics and interpretation of information to understand their business.
How did you identify opportunities for improvement through clinical analytics?
Dean has traditionally believed that service is our differentiator in the marketplace. We are a high-touch organization. We conduct customer satisfaction surveys, but we never had a good process for digging under the layers of that information to get better.Now, if we find that any department or provider that is not happy with the level of their service scores, we can identify specific areas most highly correlated to their satisfaction rating, so they can really focus on improvement. Since we implemented this process, we also have implemented a structured service excellence team which uses the data to coach providers. This is very different from our previous, one-size-fits-all approach.On the quality side, we have made improvements in blood pressure (BP) control. Dean is a founding member of the Wisconsin Collaborative for Healthcare Quality, a multi-stakeholder, voluntary consortium that publicly reports performance measurement data. Through the collaborative, we saw that our BP readings weren’t performing as strongly as we would like. We looked for missed opportunities to do better. If patients had a high reading, did the provider perform an appropriate clinical intervention, such as offer counseling or change medications? We created an interventional strategy, leveraged Lean management techniques and did a rapid improvement event to address the clinical inertia.We also identified that we had a lot of BP readings that were zero. That led us to retrain our staff on taking BP and accurately recording the readings. Another problem was with high readings. When a patient has a high reading, the national recommendation is to retake the reading in five minutes. We found that those second readings, often normal, were recorded in the progress note rather than a queryable field. Fortunately, we had analysts that understood that there was likely a second reading somewhere and wondered why we weren’t seeing that.
What role does the CMIO play in this effort?
CMIOs need to appreciate that the way they think about