Medical imaging overuse has been pinpointed as a key driver behind the overall escalating health care costs in the United States. Across widely separated and unmanaged populations, up to 50% of tests are for cardiovascular imaging studies, of which approximately 25% have been previously estimated as inappropriate or, at best, noncontributory. To establish mechanisms that will ensure more efficient use of cardiac imaging techniques, the American College of Cardiology Foundation developed a task force to examine the use of appropriateness criteria (AC) for cardiac imaging in clinical practice. In 2008, the American College of Cardiology Foundation/American Society of Echocardiography Appropriateness Criteria for Stress Echocardiography were released. However, little was known about the direct applicability of the stress echocardiography AC in clinical practice.

An investigation by Mansour and colleagues featured in this issue of the Journal explored the clinical application of published AC for stress echocardiography at a single-center university hospital. By using a prospective methodology, two independent physicians reviewed the indications and test results of 289 patients undergoing stress echocardiography. The authors found that 88% of stress echo studies could be classified using the published AC; of these studies, 71% were deemed “appropriate” and 9% were rated “uncertain.” The study confirmed previous work done in this field, suggesting that approximately 20% of stress echocardiography tests performed in academic centers might be inappropriate. The authors also established for the first time the potential use of AC in improving the yield of stress echo, because a higher number of abnormal results were seen in patients whose tests were done for appropriate indications in comparison with those with uncertain or inappropriate indications.

The use of stress echocardiography for optimal patient care requires prudence for discerning the pre-test probability of coronary artery disease (CAD) because use of stress tests determines subsequent treatment and may lead to further diagnostic testing or other procedures. Therefore, if test results are not needed for proper clinical care, the inappropriate use of tests could lead to increasing cost of care without obvious clinical benefit. Not surprisingly, more than half of the inappropriate studies in the investigation by Mansour and colleagues were performed to screen for CAD in patients with low pre-test probability or in asymptomatic patients who had only low or moderate risk for CAD. The remaining inappropriate stress echo studies were performed for preoperative risk stratification either in the absence of cardiac risk predictors or before low-risk non-cardiac surgery. Efforts aimed at minimizing the overuse of stress imaging could focus on these clinical situations. One possible approach could be ordering a test through computerized data entry that mandates test referral using an electronic checklist for guiding physicians’ decision-making before referral for diagnostic testing. Mansour and colleagues also identified that a significant number of inappropriate stress tests were ordered from the preoperative clinic. This further substantiates the utility of AC evaluation studies in identifying a target audience for future educational initiatives.

Critics have pointed out a variety of potential problems with the development and use of AC, chief among them being the possibility of imprecision in the development of the criteria and the lack of evidence for their validity. The process for AC development is only partially evidence-based and is heavily weighted by expert consensus. Previous evidence suggests that the AC developed by an expert panel may show approximately the same degree of variation as the interpretation of some commonly used diagnostic tests. Recent studies have reported a wide range of the transthoracic echocardiography studies that were ordered for indications that were not addressed by the AC. Similar problems are apparent in patients referred for stress echocardiography, with recent reports, including the study by Mansour and colleagues, suggesting that up to 10% of stress echocardiography studies may have indications that are not addressed in the AC. Thus, studies that evaluate the application of the AC in clinical practice have an important role in providing useful feedback for future development and revision of the AC.

Another interesting aspect of the study by Mansour and colleagues was the revelation of a significantly higher number of inappropriate studies in women. The origin of this gender bias remains uncertain, but it is similar to gender differences reported in cardiac care and resource use. The authors suggest that this may be related, at least in part, to outdated methods for assigning pre-test probability of CAD in women. In any event, it should be emphasized that stress imaging studies are highly accurate for the detection of CAD in women. Because of the challenges in the initial diagnosis of CAD in women and the adverse consequences of the disease, at least the some relaxation in AC for ordering stress echo in women may be justified.

Rating a diagnostic test as “appropriate” or “inappropriate” may not reflect its clinical value accurately. It is a rather crude way to make “binary” what is a process requiring complex decision on why a test is ordered. Some studies categorized as “inappropriate” may in fact be very helpful clinically by reorienting the diagnostic evaluation or therapy (e.g., a patient with coronary risk factors and exertional dyspnea, in whom a negative stress echo study might support pursuing other causes, such as impaired left ventricular relaxation with sensitivity to heart rate increases, or lung disease), whereas some studies categorized as “appropriate” might not be needed to guide proper patient care (e.g., transthoracic echo in a patient with nonischemic cardiomyopathy, a left ventricular ejection fraction of 30% on prior studies, and increased dyspnea after stopping diuretic medications and retaining 12 lbs of edema fluid).

The study by Mansour and colleagues supports the need for adhering to AC while continuing the quest for refinement of rational approaches to the delivery of high-quality cardiac care. Further multicenter trials are needed to explore the variability in application of AC in clinical practice and to understand the impact of AC on health outcomes. Development of appropriate pre-testing decision-support tools may be useful for ensuring that tests are ordered for appropriate reasons and that inappropriate use of resources is minimized. In this manner, cardiac imaging can be used most effectively, benefitting patients, caregivers, and other stakeholders.