OBJECTIVES: In a previous study, the 2006 appropriateness criteria (AC) for cardiac CT were applied to 251 patients. It was found that 46% of patients could not be classified, and two observers showed only fair agreement (Îº = 0.31) on the assigned appropriateness rating (appropriate, inappropriate, uncertain, or not classifiable). The conclusion was that the 2006 AC were difficult to apply. The AC were revised in 2010. The aim of this study was to determine if the rate of patients not classifiable and interobserver variability had decreased to the point at which the AC could be reasonably applied.
METHODS: Medical records of the 251 patients who were classified using the 2006 AC were reviewed by two observers, who attempted to assign the patients’ indications using the 2010 AC. Patients for whom no indications could be found were deemed not classifiable. A third observer settled disagreements. The numbers of patients not classifiable using the 2006 and 2010 criteria and the number of patients on whom the two reviewers disagreed on indications were compared using McNemar’s test.
RESULTS: One hundred fifteen patients (46%) were not classifiable using the 2006 AC. With the 2010 AC, the number of patients not classifiable decreased to 39 (16%) (P < .001). With the 2006 criteria, the observers disagreed on specific indications for 152 patients (61%). With the 2010 criteria, the observers disagreed for 118 patients (47%) (P = .002).
CONCLUSIONS: Using the 2010 AC, the number of patients not classifiable and the number of disagreements decreased. Although the rate of patients not classifiable has decreased to an acceptable level, the interobserver variability remains concerning.