New Tool Mines Patient Record for CIN Risk Factors

Although contrast-media induced nephropathy (CIN) has always been a serious patient safety issue, until now only qualitative tools have been available to aid physicians in predicting a patient's risk. Interested in a better alternative to current methods, researcher Garry Choy, M.D., M.S., set out to develop a quantitative tool for stratifying risk—specifically a prediction score—that could aid in the triage of patients needing a contrast-enhanced CT.

In 2008, during his radiology residency at Massachusetts General Hospital (MGH) in Boston, Dr. Choy parlayed a $30,000 RSNA Research & Education Grant (see sidebar) into a search-engine based informatics platform that has since been integrated into the hospital's IT/RIS system.

"This clinical decision support tool acts as a safety net and assists physicians by automatically mining the data in the medical record in real-time, seeking to identify risk factors and preventing the performance of a contrast-enhanced scan when there is the potential for harm to a patient," Dr. Choy said.

Tool Addresses Leading Cause of Hospital-acquired Renal Failure

CIN is the leading cause of hospital-acquired renal failure—accounting for up to 12 percent of all cases—and leads to longer hospital stays, higher mortality rates, and increased medical costs. Most troubling of all: there is no available treatment to reverse or even mediate the condition, Dr. Choy said. Identifying at-risk patients is critical for physicians to determine the best course of action, whether to administer contrast at all, reduce the dose and/or implement prophylactic strategies, he said.

Data points currently used in the qualitative screening for patients at risk of CIN include serum creatinine, estimated glomerular filtration rate and conditions such as pre-existing renal insufficiency, diabetes, congestive heart failure, dehydration, advanced age, multiple myeloma and malignancy affecting the kidneys, ureters or bladder.

Drawing on his background in operations management and engineering, Dr. Choy created a program capable of searching the MGH electronic medical record (EMR) for keywords and targeted concepts relevant to identifying factors and data within the EMR to determine if a particular patient is at risk for CIN.

Dr. Choy, along with Michael Zalis, M.D., and Mitch Harris, Ph.D., developed the project with scientific advisor G. Scott Gazelle, M.D., Ph.D., M.P.H., a professor of radiology at MGH and Harvard Medical School and a professor in the Department of Health Policy and Management at the Harvard School of Public Health.

The team also retrospectively analyzed the records of more than 13,000 patients who had undergone CT scanning at MGH between 2005 and 2007 in an attempt to correlate the development of CIN with certain risk factors. "We performed statistical analysis to quantify the effect of various risk factors on the development of contrast nephropathy," Dr. Choy said. "The analysis was used to develop a risk prediction score—a number—that indicated the potential CIN risk."

The first version of their Queried Patient Inference Dossier (QPID) module was tested in a group of 100 patients—22 of whom developed contrast nephropathy—during a pulmonary embolism CT scan. The QPID module performed at a sensitivity of 81 percent—18 of the 22 patients were correctly identified—and a specificity of 43 percent.

At-risk Patients Could be Flagged During Radiology Order Entry

Currently, Dr. Choy and colleagues are analyzing a larger cohort of patients for further validation of their risk prediction model. They are also modifying their radiology order entry system so that orders for contrast-enhanced CT scans will be flagged when patients are at risk for contrast nephropathy.

"Dr. Choy's research is highly innovative and of great potential significance," Dr. Gazelle said. "Contrast-induced nephropathy continues to be an important problem in radiology and a better means of identifying and stratifying patients at risk for this condition is urgently needed."

Dr. Choy and colleagues plan to publish their RSNA-funded research and hope to secure further funding to continue to work on decision-support tools to improve quality and safety in other areas of radiology, as well as other areas in clinical medicine.

"The RSNA research grant has given me the opportunity to pursue my passion for informatics in the development of decision-support tools that will aid the clinician," Dr. Choy said. "I hope to continue a career focused on innovation in radiology and medical informatics."