Tools Leverage Data, Drive Quality Improvement in Radiology

CHICAGO - New informatics tools are emerging that can help radiologists leverage patient and other medical information to more accurately interpret images, avoid medical errors, and reduce unnecessary imaging, according to speakers at RSNA 2014.  

Like most physicians, radiologists are facing a growing pressure to increase efficiency and quality. They are being asked to help reduce unnecessary imaging and procedures, improve the accuracy of their image interpretations, and work more effectively with their physician colleagues. A new generation of health informatics tools presented at session called “Leveraging your Data: Informatics Approaches and Solutions to Improve Imaging Care Delivery” aim to help radiologists meet these growing demands by making relevant patient information, guidelines, and medical references available to them in real time.

“I expect these tools will evolve rapidly,” said Michael E. Zalis, MD, an associate professor of radiology at Massachusetts General Hospital.

Zalis explained that electronic medical records have made an unprecedented amount of patient information available to physicians. But these tools were largely designed to aid with billing, so they often don’t provide information in a way that is useful to clinicians. Relevant information may be lost in the sea of data, or context that is relevant to the patient’s care may be missing.

 “The patient’s story gets lost,” Zalis said. “We need tools to extract that.”

 To help physicians get the information they need out of electronic health records, Zalis and his colleagues at MGH developed a system called Queriable Patient Inference Dossier (QPID) that makes the electronic records searchable using ontologies. Physicians can use the system to search by clinical concepts or to automate certain queries. For example, Zalis said a radiologist could use the system to search a patient’s records for medical implants, and the system retrieves information about all relevant devices without having to search for each by name. 

Arun Krishnaraj, MD, MPH, the director of the division of body imaging at the University of Virginia Health System, tested the QPID system with fellows at his institution and found it reduced interpretation time by 20%. He explained that a normal electronic medical record may only provide the radiologist with a line describing the reason for the order. The QPID system can allow the radiologist to quickly access patient information necessary to rule out possible diagnoses and narrow down the options to the appropriate choice. For example, it allowed his team to diagnose a 39-year-old woman with a hepatic adenoma without needing a biopsy or additional imaging.

“We saved the patient a diagnosis that is wrong and anxiety inducing,” Krisnaraj said.

Informatics tools can also be used to allow radiologists to view images and annotation from medical atlases alongside their patient’s images, explained Cree Gaskin, MD, vice chair of informatics at the University of Virginia Heath System. For example, he explained viewing an annotated image of an age appropriate bone scan alongside a patient’s scan can allow radiologists to more quickly and accurately distinguish an injury from a normal developmental feature.

Some tasks can also be automated. For example, Cree noted integrated software for determining bone age was preferred by the system’s physicians compared to manual studies, and produced fewer errors, 0% compared with 23%.

Decision support tools can also be integrated into electronic health records to help physicians order fewer inappropriate or unnecessary images, explained Giles Boland, MD, vice chair the department of radiology at MGH.

“You are not creating any value if you order the wrong test,” Boland said.

But a decision support tool based on American College of Radiology guidelines for imaging can help physicians determine if imaging is appropriate given the patient’s records. Such a system can also help flag safety issues or notify a physician if an image has already been taken.

He noted that studies have demonstrated that these systems reduce imaging, which may reduce income for radiologists. However, he noted that emerging payment systems also reward value.

“It’s about what’s best for the patient,” he said.