Imaging informatics & MU: The time is now

One of the key goals of healthcare reform is to boost quality while cutting cost. It’s a tall order, and success will hinge on thoughtful deployment of imaging informatics systems.

Yueyi I. Liu, MD, PhD, and Daniel L. Rubin, MD, both from the department of radiology at Stanford University in Stanford, Calif., outlined how various informatics systems intersect with meaningful use (MU) criteria in a review article published in the August issue of Academic Radiology.

“There is overlap in MU goals for the primary care physician and radiologist, and informatics methods to achieve MU in radiology need to be considered now in order to realize the incentives for participating as well as to be able to anticipate new directions in MU as radiology becomes a focus area of these criteria in the future,” wrote Liu and Rubin.

Decision support
MU requires hospitals to implement “one clinical decision support rule related to a high priority hospital condition along with the ability to track compliance with that rule.” The criterion dovetails with major cost and quality challenges in imaging. Escalating advanced imaging expenditures represent a financially unsustainable model. Part of the problem is the proportion of inappropriate exams, which may be as high as 26 percent of advanced imaging studies.

However, ordering physicians’ use of American College of Radiology (ACR) appropriateness criteria is low. Liu and Rubin attributed the lukewarm adoption to a lack of awareness and lack of integration into order entry systems. They noted that hospital systems that had integrated decision support into order entry realized substantial decreases in potentially inappropriate exams and improved adherence to evidence-based guidelines. The authors cautioned against one potential pitfall of such systems—alert fatigue—and recommended continuous monitoring of the system to ensure optimal utility.

The other form of decision support focuses on radiology interpretation. Systemic reviews have shown that computer aided detection can improve practitioner and clinical performance. According to Liu and Rubin, four features are independent predictors of improved performance:

1. Automatic provision of decision support as part of clinical workflow;
2. Provision of recommendations in addition to assessment;
3. Provision of decision support at the time and location of decision-making; and
4. Computer-based decision support.

NLP, data mining and dose tracking
MU stresses the need to “generate lists of patients by specific conditions to use for quality improvement, reduction of disparities, research or outreach.”

Existing processes rely on ICD-9 codes, which could be improved by use of natural language processing (NLP), according to the authors. They referred to a recent study which showed NLP outperformed administrative codes in detecting five of six common postoperative complications.

The ACR has established the National Radiology Data Registry, a collection of databases that allows radiology practices to benchmark performance on measures such as radiation dose, mean wait time and percent exam signed.

One of the challenges in benchmarking radiation exposure and other measures across sites is the lack of standard terminology across facilities. The Radiological Society of North America (RSNA) RadLex Playbook ID maps nonstandard names to a controlled vocabulary, which can normalize these names and facilitate comparisons. The controlled vocabulary also may improve the clarity of communication and reporting by unifying numerous similar terms with a preferred term.

Data sharing
MU also promotes data sharing among providers, an effort supported by numerous imaging initiatives. Integrating the Healthcare Enterprise aims to improve the ways computers share information and leverages the Cross Enterprise Document Sharing profile to facilitate sharing of documents. The RSNA Image Share project is building a secure, patient-centered mechanism to share images, which supports this goal.  

Remaining challenges
Although imaging informatics can help address cost and quality needs, implementation may present new challenges, including cost, customization and integration.

“Even though the financial incentives by Medicare and Medicaid will help alleviate the cost of implementing various systems, building a robust (instead of patchwork) system may incur a higher cost than is initially offset by incentives,” wrote Liu and Rubin.

The authors cautioned that “one size would not fit all.” Thus, local institutions need to dedicate personnel to align systems with their goals and needs on an ongoing basis.

Finally, the various systems are standalone technologies, and integration into the electronic medical record will be difficult, they predicted. However, Liu and Rubin emphasized that these informatics developments will meet the end goal—improved quality and efficiency.