An upcoming Centers for Medicare and Medicaid Services (CMS) pilot project to boost EHR use in 12 communities has garnered a great deal of media attention, but a much more modest CMS pilot may also have considerable impact.
This summer, CMS is developing a repository that will test its capability to receive quality-reporting data sent directly from both EHRs and clinical registries rather than Medicare claims forms. If successful, the pilot could make it easier for physician groups to participate in CMS quality efforts, thus boosting the number of physicians participating. It could also provide momentum for physicians advocating a shift from using quality measures based on claims codes to those based on clinical data.
Along with 12 registries, including the Society of Thoracic Surgeons Adult Cardiac Surgery Database, six ambulatory EHR vendors are participating in the PQRI (Physician Quality Reporting Initiative) pilot: Cerner Corp., eClinicalWorks, NextGen Healthcare Information Systems Inc., DocSite, Allscripts, and Anceta.
The voluntary PQRI program pays a 1.5 percent bonus to physicians who successfully collect and report their practice data in relation to a set of 119 performance measures. In 2007, about 16 percent of eligible physicians did so, according to CMS.
“We're excited about adding EHRs and registries,” says Daniel Green, M.D., medical officer in CMS' Office of Clinical Standards and Quality. “We think it will open up PQRI to many more participants.” Green says that many physicians are already reporting data to registries as part of their normal workflow, so this will reduce the burden of reporting twice.
The pilot involves five measures, three focused on diabetes, one on coronary artery disease, and one on heart failure. Some vendors are submitting actual data while others are submitting fictitious information. “It is simply a test to see if they can extract information from patient records, that it can be transmitted and received,” Green says. “We would hate to have people submit data pulled from health records and be denied the bonus payment because the data was not formatted properly.”
Bay Area Family Physicians, a seven-doctor family medical practice in New Baltimore, Mich., is working with NextGen on the pilot. The practice has participated in PQRI using claims data, says Doug Render, Bay Area's IT director, who adds that the CMS effort nicely parallels efforts by Michigan insurers to use EHR data to help physicians with quality improvement. Render believes sending discrete data from the NextGen EHR will give CMS more information to work with.
CMS' Green agrees. For instance, claims data might describe a Hemoglobin A1c reading above 9.0., whereas, “It might be more meaningful to know that the patient's reading went from 14 down to 9.1,” he says. “Using current measures, we would never know that, but with data from EHRs and registries we could see it.”
The idea of making reporting quality information easier is music to the ears of Hal Williamson, M.D., professor and chair of family and community medicine at the University of Missouri. None of the system's 350 physicians has participated in PQRI so far. After an internal pilot project, they determined it was just too much effort to go back through files and submit claims codes. “It's a huge administrative burden,” Williamson says, “but we want to participate.”
That's why 70 faculty and resident physicians will participate with Cerner in the CMS pilot project. According to Williamson, for the past two years the university has been working with Cerner to develop an analytics package that helps measure quality internally. “For us, submitting information from that effort to bodies like CMS is the next logical step,” he says. “Our IT staff is working with Cerner on whether we are measuring the right thing and measuring it accurately.”
Paul Tang, M.D., chief medical information officer at the Palo Alto Medical Foundation, a California multispecialty healthcare provider with approximately 150 physicians, says CMS' effort to get quality data from EHRs is significant.
“Making it less of a burden to report quality measures is important,” he says. “That is not a trivial step.” But Tang, who chaired a 2007 National Quality Forum panel on deriving quality measures from EHRs, wants to see a parallel development of clinical quality measures that can be pulled from EHRs, “so that when we get to 2014, we're not still relying on administrative codes.”
Claims data deals with information captured in bills, and every encounter has to have an associated diagnosis. Those claims codes were not designed to be descriptive of physical conditions. Sometimes they are not entered by a physician, but by a billing person. “The accuracy of diagnoses on billing codes is often not very good,” Tang says. But, he says, research has shown that when you work with clinical data, the accuracy is much higher.
A coded problem list in the EHR could replace billing codes to identify patient conditions for quality measurement, Tang adds.