Leveraging Health IT to Improve Quality and Safety

Twitter icon
Facebook icon
LinkedIn icon
e-mail icon
Google icon

"You can't improve what you can't measure." Boy, if I had a nickel for every time I've heard that phrase in the past four days . . . well, I wouldn't be a rich woman, yet, but I would have a pocketful of change. But what does that phrase mean for health IT? In two short words: data mining. The importance of that capability, as well as several others, was the focus of this morning's session on next generation IT requirements for improving quality and safety.

Ramin Khorasani, MD, and David Avrin, MD, presented on IT tools that can be used to improve patient safety and imaging quality by streamlining ordering, facilitating clinical decision support and hastening the delivery of critical results. "Safety is becoming the core mission in many hospitals," Khorasani said. "It's not just spoken. It's actually practiced." And with health care reform in the national focus, he said, smart radiologists will begin improving their quality and safety initiatives using health IT now in anticipation of more stringent requirements to come. “Allmedication, labs and radiology must be ordered electronically by 2011, and by 2014 we have quality measures coming into place," he said. "Imaging has become quite a focus in terms of quality improvement and waste reduction.”

Khorasani outlined how cutting-edge IT tools can be used to improve both the appropriateness of imaging and the access to imaging services and reports across the enterprise. "Why change?" he said. "What's wrong with pencil, paper and phone?" Then he proceeded to outline the answer. Short version: everything. Long version: the endless series of opportunities for human error throughout the ordering and scheduling process, wait times up to ten minutes on the phone and the limitations of most RIS systems, which only allow for 256 characters of text in spite of however much information the referring physician included in his or her requisition.

Compare that with the system Khorasani has in place at Brigham and Women's in Boston: "If anybody in our system tries to order a radiology exam, we launch an application that allows them to order electronically, which integrates it into our EMR and RIS/PACS environment and is connected to our payor database so we also get the acquisition number.” Decision support is also included: "Now, physicians can cancel it, but if they do we capture what they did, so we can see how they performed. So if a physician tries to do a lumbar MRI, he’ll see whether someone’s done it before, and if he tries to order the exam the system will ask for more information first." If a physician doesn't like the system's verdict, he or she can e-mail Khorasani automatically with feedback. "How many e-mails do I get?" he said. "Zero. We have beaten them into submission.” He laughed.

An analytics tool then generates real-time statistics reflecting data from both the RIS and PACS, showing how many orders were entered by each type of physician in the hospital and what percentage of those were signed by the physicians themselves. "This is the definition of meaningful use," Khorasani said. The system also tracks how regularly the evidence-based guidelines are followed. Prior to implementation, physicians followed them forty percent of the time; now the guidelines are followed 92% of the time.

David Avrin, MD, also spoke in the session, touching on recent improvements in reporting that can be leveraged for better quality. Avrin, a major advocate of voice recognition, urged those present not to hold off on VR because of its speed -- which has improved greatly -- or because of pride over being forced to do themselves what transcriptionists used to take care of. He described the system the Mayo Clinic used up until a few years ago to improve report delivery speed, which involved a transcriptionist with an IBM Selectric typewriter sitting next to every single radiologist, taking dictation onto reports in triplicate form. "Unless you want to revert to that system, voice recognition is the only option," Avrin said.