Skip to content Skip to navigation

Using Data Analytics to Improve Clinical Performance and Its Reimbursement Outcomes: One Hospital’s Experience

June 20, 2016
by Mark Hagland
| Reprints
A. Thomas McGill, M.D. is leading an ongoing initiative at Butler Health System to leverage data to improve both clinical and financial outcomes

At Butler Health System, a 311-bed community hospital in Butler Pa., hospital leaders have come together to use data analytics to improve a range of patient care delivery processes and outcomes. Working with an analytics solution from Information Builders, a New York City-based business intelligence and integration company, Butler Health System clinical, IT, and financial leaders have been moving forward to focus in particular on examining and improving specific diagnostic and care delivery processes whose outcomes have financial impacts.

In all this, the leaders at Butler Health System are fortunate to have A. Thomas McGill, M.D., leading the charge. Dr. McGill, a practicing infectious diseases specialist, has been vice president of quality and safety at Butler Health for 10 years, and for the past four years, he has also been the organization’s CIO. Thus, his title and responsibilities encompass both quality improvement and IT activities and efforts at the health system. McGill spoke recently with HCI Editor-in-Chief Mark Hagland regarding the work that he is helping to lead at Butler Health. Below are excerpts from that interview.

You have a unique perspective on all this, being both the vice president of quality and patient safety for ten years at your organization, and also, for the past four years, the hospital’s CIO. Tell me about your and your colleagues’ pursuit of clinical performance improvement through data analytics.

Certainly. Especially because of my dual titles, in our analytics work, we are focusing on a combination of quality and safety improvement, as well as on financial analytics. And our particular focus has become all the metrics for which we are held accountable by external organizations—payers and regulators. We had long been working on analyzing some metrics, but the evolving mandates coming from the Medicare program and the commercial payers have particularly spurred activity here. Medicare has all its adjustment programs, and the commercial payers have their incentive programs. For example, the healthcare-acquired conditions program under Medicare penalizes a wide range of conditions acquired while patients are being treated—some of them infections and other conditions non-infectious.

A. Thomas McGill, M.D.

For example, we started working on venous thromboembolism [VTE] prophylaxis early on. We looked at the actual costs of patients in the same DRG [diagnosis-related group]—looking at whether they had a clot or not. And we found that per case, patients with a clot were costing us $9,000 per case. I was really surprised by that level of expense. We immediately saw something like 50 episodes or events involving venous thromboembolism, when we started doing our analysis back in 2012. So our baseline measurement was that we had found that we had been experiencing basically 50 or 60 such cases a year, and we were able to get that number down to 9 or 10. The drug Lovenox was still a brand-name drug. So when we first started, we said, OK, we’re going to spend $25 a day on this prevention—we knew it was the right thing to do for the patients, but didn’t know what it would be like financially, and even financially, it was a home run to prevent blood clots. In fact, we ended up avoiding a quarter of a million dollars in costs just through that improvement.

And you just start doing these one after another, and they start accumulating. And outside of infectious disease, that’s one of our best examples of where clinical care quality improvement has saved money.

Could you share any other examples?

Basically, at the turn of the century, we started doing active MRSA surveillance to find out whether patients were carriers of MRSA [the methicillin-resistant staphylococcus aureus bacterium] when they came into the hospital for care. And therefore, we wanted to know if they were carriers of MRSA when they came in, were we inadvertently spreading that to other patients and amplifying the amount of MRSA in our community? So Dr. Jernigan from the CDC [Daniel B. Jernigan, M.D., M.P.H., Deputy Director of the Influenza Division at the National Center for Immunization and Respiratory Diseases, for the federal Centers for Disease Control and Prevention] had this idea, and we bought into it and started looking at screening everybody. So we did a baseline study again: we were screening incoming patients, and made those results available to doctors considering potential antibiotic treatment. We found that our amplification rate was 25 percent, meaning that for every 1,000 people who were coming into our hospital as carriers of MRSA, 125 were going out, confirming that we were adding to the burden of MRSA in the community. So we started screening and treating for MRSA upon admission. And we didn’t see a change in the burden until we had done active surveillance. We didn’t affect our overall amplification rate until 80 percent of our units were under this program. And we understand that many healthcare workers move from unit to unit. What our data showed was that the hospital-based spread of MRSA can be prevented.

In other words, you and your colleagues found that clinicians were spreading the disease inadvertently?