An Interview with Ryan Leslie, VP of Analytics and Health Economics
By Jane Clabby
On March 20, I attended IBM’s Smarter Analytics Leadership Summit in New York City. I was pleased to discover that the Client Panel, “Leading with Analytics”, included Ryan Leslie, VP of Analytics and Health Economics at Seton Healthcare. I wrote a blog about Seton Healthcare and Watson about a month ago, and this was a great opportunity to hear more about that project.
Seton Healthcare is using Watson technology to improve patient care and reduce patient readmission rates. According to Leslie, 80% of Seton’s patient data is unstructured, including physician and nurse notes, discharge summaries and case notes. A standard chart review doesn’t reveal any of this information, and in a complex condition such as congestive heart failure which has to be diagnosed early for effective treatment, these details are critical. By using this unstructured data, Seton Healthcare has been able to institute a program that tells them which patients have the most potential for positive outcomes. By revealing patterns in unstructured patient data, they can identify those patients that will benefit most from a particular treatment plan, and avoid costly readmissions.
The most important information, according to Leslie, is garnered by reading between the lines — looking at patient information such as does the patient have access to transportation? Do they have a primary care physician? Do they have a social support system? Answers to these questions can identify risk factors for readmission. Leslie said that through the study, Seton Healthcare had identified another risk factor that frequently led to readmission. During physical examinations, patients were asked to tilt their head so physicians could examine the jugular vein. This simple test revealed vein patterns that were frequently seen in readmitted patients. By using this test rather than expensive array of lab tests, Seton Healthcare is better able to identify at-risk patients at a much lower cost.