Optimizing Decision Support Along the Imaging Chain

From image order entry to reporting results to the referring physician, several spots on the medical imaging chain provide opportunities for using decision support.

“Clinical decision support for imaging can aid the ordering clinician, the radiologist and technologist by providing patient specific, exam specific information and integrating it to recommend the most appropriate course of action going forward,” said Katherine Andriole, PhD, of the radiology department at Brigham and Women’s Hospital .

Speaking at a session at SIIM 2012 on Friday, Andriole mapped the typical patient encounter, pointing out where decision support can provide information and suggestions for optimal care.

“I argue we can have decision support injected at each point of this chain,” she said.

For example, when the exam is ordered, a clinical decision support tool can provide reminders and guidance about the appropriateness of the exam. At the exam protocol stage, software can optimize that acquisition process. Software can also increase the conspicuity during processing to help the radiologist identify lesions, for example, and provide reference guides or related cases for comparisons. Finally, when the results are being reported, the radiologist can rely on reminders about certain information to be included and templates for the report.

At Brigham and Women’s, officials developed a portal to improve the protocol step of the process, Andriole explained. Protocolling the exam details the set of instructions for the test, and officials wanted to improve this before the patient arrived on site or received a potentially suboptimal exam.

For the old protocol process, a radiologist might want to view the RIS, the hospital information system for labs ordered, the medications system for ordering contrast, and yet another system to view prior studies. It can be a tedious process, so radiologists may rush through it, she said.

“So we created a protcolling portal that had all the things we anticipated they might want to review in one screen,” she said. The portal pulls from the various systems, including computerized order entry that requires the clinician explain the reason for the exam.

For clinical decision support to be truly effective, however, it must be based on strong evidence that is unambiguous and actionable, said Ramin Khorasani, MD, vice chair of radiology at Brigham and Women’s, who also spoke at SIIM 2012.

An order entry system that stops the clinicians and suggests the test may not be optimal or that others can be effective won’t cut it, he said. “Opinions that are ambiguous, take them off. Those are hallway conversations,” he said.

His organization developed the Radiology Management Program, which includes knowledge delivery at every point of care. In it, evidence on which those notices are based is updated routinely, and the tools are embedded in to the work flow - all critical elements to ensuring a decision support tool isn’t ignored.

Another way to ensure physicians follow the guidance? Have consequences, Khorasani said. There should be some kind of impact or feedback, or the tool will not change behavior.

For example, a primary care physician at Brigham and Women’s who orders a lumbar spine MR must answer several questions in the system about severity of pain and prior treatment. Those answers can be mapped to the ACR guidelines, and the ordering physician can easily click a link to see the evidence, Khorasani explained. If the system then alerts the physician that the MR is not the appropriate test, and the clinician hits “ignore,” he or she is then directed to have a peer-to-peer consults before moving forward. An imaging professional is paged for the consult and can then provide a number to authorize the order in the system.

They incentivize the process by providing ordering physicians the ability to schedule the exam online, he said. “To me this is one of the big drivers of adoption,” he said, adding his organization has seen the number of those exams that meet the appropriateness criteria rise.

“If you really think you can put something in front of people saying, ‘Oh, think about this,’ an ambiguous suggestion,” he added, “you will be waiting a long time to optimize your practice.”