Empowered Care: The MHS Doubles Down on Clinical Intelligence

Twitter icon
Facebook icon
LinkedIn icon
e-mail icon
Google icon
Albert BonnemaHealth care informatics solutions are increasingly robust, but it is often observed that most organizations do not utilize the data they aggregate to its fullest potential—yet. That is the challenge the military health system (MHS) found itself facing two years ago, and in its attempt to bend the health care cost curve with limited resources, its informaticists doubled down on the clinical business intelligence available in its electronic health record (EHR). As Albert Bonnema, MD, chief medical information officer for the Air Force Medical Service and its chief health care informaticist, noted, an integrated informatics strategy is critical to supporting population health, patient safety, quality improvement, and enhanced resource management. In a session at the 2012 conference of the Healthcare Information and Management Systems Society (HIMSS) in Las Vegas, Nevada, Bonnema described the “pain and agony” of electronic medical record modernization. “We had a lot of data, but we had scarce information,” he said in a talk entitled “Maximizing Clinical Business Intelligence in the MHS.” “Suddenly, we got to a choke point where we couldn’t spend anymore.” Data Warehousing The MHS began digitally aggregating health care information almost 20 years ago, Bonnema said. In 2000, its existing medical data repository became ARS Bridge, which in turn became the Clinical Data Mart in 2005 and the COHORT system in 2010. COHORT, Bonnema explained, features near-real-time data aggregation without normalization; although the warehouse is advanced compared to the MHS’ original data repository, it still presents problems. “Sometimes, with data warehouses, you get questions where you know the answer is there, but the data’s not set up to give it to you,” he said. “We had COHORT modeled well to answer business questions, but when it came to clinical questions, it was a little trickier.” Meanwhile, provider requests for clinical intelligence capabilities were mounting into “an insatiable demand,” Bonnema said. “People are seeing what’s out there in the literature and wanting to do this or this with the data, and it became frustrating when we couldn’t meet the demand,” he recalled. Part of the problem, he said, is that there was no real organizational ownership of the data warehouse, and as such, no one to take responsibility for its capabilities or lack thereof. “You have to constrain yourself to what strategy is going to drive you,” Bonnema said. For the MHS, that strategy included cost containment, management of population health, and improvement of the patient care experience; in order to meet those three goals, a new architecture for the data would be required. “It became apparent to me that our strategy was impeded by our data architecture,” he said. “How do you build one that’s flexible enough, and offers a secondary use of data?” Next-generation Capabilities The MHS’ new architecture for its data warehouse includes the databases of more than 40 legacy applications from the Air Force alone, Bonnema said; the databases were either virtualized or brought into a common model, while the necessary capabilities of the applications were migrated over. The end result was a clinical informatics engine in which workflow integration feeds clinical decision support, clinical decision support feeds measurement, measurement feeds peer review, and peer review feeds analysis and intervention. The clinical intelligence now available to MHS providers is comprehensive, Bonnema said—and sometimes surprising. “We can predict mortality simply by looking at what type of provider you’re assigned to,” he said. “There’s the thought that you can change your hospitalization and mortality rates in your entire health care system by helping patients select a provider based on their risk profile.” In another assessment, the MHS discovered that it was spending $16 million annually on MRIs to assess the cause of lower back pain, some of which were clinically unnecessary. Bonnema and team added a template to the EHR with the practice guidelines for assessing lower back pain, and, he said, the feedback from providers has been “excellent.” “It’s not a pop-up, and it’s not obnoxious,” he noted. “It’s been well received.” An electronic peer review tool allows clinicians to assess their own performance on a monthly basis: “Now we can look at the defect rate,” Bonnema said. “Did you image someone for acute low back pain at two weeks with no signs? We can make it a teaching event.” Looking Forward The MHS’ current clinical intelligence project is creating a data mart for its patient-centered medical home initiative, aggregating five terabytes of data from every military treatment facility with dashboards that will allow users to drill down into the information. “They can see their schedules and appointments, can associate workload with scheduling, can see trends, and the business intelligence architecture has been great,” Bonnema said. In the future, Bonnema hopes to create an MHS “research cloud” that would leverage the MHS’ electronic data going back to 1991 in partnership with private sector and academic researchers. He also sees potential for what he calls a clinical looking glass tool that would offer residents advanced analytics tools for their required annual quality studies. “If I was doing a residency today or had a lot of residents, this would be a great way to handle all those projects,” he said. In the short term, the MHS also will be implementing point-of-service clinical decision support based on the data contained in the warehouse. “Measurement is fundamental to quality,” Bonnema concluded. “This is really an interesting new enabler of quality.” Cat Vasko is editor of HealthIT Executive Forum.