Skip to content Skip to navigation

Mining for Data

August 3, 2010
by Jennifer Prestigiacomo
| Reprints
Web-Exclusive Interview: Carol Steltenkamp, M.D., CMIO of UK HealthCare

University of Kentucky Healthcare (UK HealthCare) in Lexington, Ky. recently transitioned to a new medical transcription service and data mining solution from Atlanta-based Webmedx. Already implemented in two of the three UK HealthCare campuses, CMIO Carol Steltenkamp, M.D., talks to Healthcare Informatics ’ associate editor Jennifer Prestigiacomo about the benefits they’re already seeing from the system.

Healthcare Informatics: Can you tell me a little about how the integration between your transcription service and your campuses is going?

Carol Steltenkamp, M.D.: The Chandler campus [the Lexington, Ky.-based UK Albert B. Chandler Hospital, UKHC’s main academic medical center] [is linked] to Webmedx, and that relationship is going well. We [just integrated] the other campus, a community hospital [UK Good Samaritan Hospital] that we acquired about three years ago. On June 27 we brought 49 systems onboard. With the nature of a community hospital being what it is, there had been some special circumstances and some arrangements made to handle different physicians and their personal requests in how they like to do things. I’ll give Webmedx credit; they helped us to support what was appropriate to support in helping us to ease the transition for some of those doctors to change the way they had always done it that really wasn’t the most effective, efficient way of doing dictation transcription.

HCI: With the three choices for physician documentation [physician self-completion with front-end speech, medical transcriptionist editing with back-end speech, or traditional transcription], do you know which option has been most popular thus far?

Steltenkamp: No, I don’t. And again, it’s easing in those physician preferences so that it best meets their workflow and gets the necessary information into the system. But I say overall when you take everything into consideration, it’s going to be mostly back-end. That’s been the in most hospitals the traditional way it’s been done, as was the case in this hospital[UK Good Samaritan Hospital].

HCI: What do you plan on using the data mining solution to do?

Steltenkamp: Because we opened a 52-bed emergency department [the Lexington, Ky.-based UK Chandler Emergency Department] we had to postpone our go-live with the analytics. I’ll tell you how it works, and how we plan on using it. They are able to find terms [through] natural language processing. The terms that we ask them to look for helped us to get information back, so that saves our quality people the time from having to comb through every dictated report to find information. And one of the [quality elements] we’re going to be going after is the ejection fraction. In the natural course of dictation, when one of our cardiologists is doing an echo [echocardiogram] or a cardiac cath [catheterization] report, we are looking forward to pulling out [that information] using the analytics tool.

HCI: Why did you decide to start by mining ejection fraction results for acute myocardial infarction and congestive heart failure patients?

Steltenkamp: Because that has been something that in many instances we need for our quality reporting as it relates to our patient safety goals and CMS [Centers for Medicare & Medicaid Services]core measures

HCI: Will this be one of the quality measures you will be reporting to CMS in 2011?

Steltenkamp: Yes.

HCI: What other clinical elements are you looking at or will be mining in the future?

Steltenkamp: Some of the others as we look toward [the future] will be noted compliance to medication, so how well the patient reports that they are doing. So, that is going to be another big one. Also, the kind of education they got regarding their medication and their self-report if they are taking it or not.

HCI: How has this solution helped your health system reach meaningful use requirements?