Appropriate use criteria for diagnostic catheterization may not be reliable for guiding clinical decisions

Criteria developed to validate use of diagnostic cardiac catheterization may not be helpful for guiding clinical practice, a study found.

Researchers looked at the relationship between the criteria-judged appropriateness of cardiac catheterization in patients with suspected stable ischemic heart disease and the proportion of patients with obstructive coronary artery disease (CAD) and subsequent revascularization. They conducted a population-based, observational, multicenter cohort study of data from the Cardiac Care Network, a registry of all patients having elective angiography at 18 hospitals in Ontario, Canada, between October 2008 and September 2011.

Results were published by Annals of Internal Medicine on March 10.

Among 48,336 patients, 58.2% of angiographic studies were classified as appropriate, 10.8% were classified as inappropriate, and 31.0% were classified as uncertain. Overall, 45.5% of patients had obstructive CAD (range, 31.8% to 64.4% across the 18 hospitals), with 13.1% having significant left main or triple-vessel disease. Studies in patients who were found to have obstructive CAD were classified as 52.9% appropriate, 30.9% inappropriate, and 36.7% uncertain (P<0.001). Studies in those with significant left main or triple-vessel disease were 16.5% appropriate, 7.1% inappropriate, and 8.7% uncertain (P<0.001).

In the overall cohort, 33.4% of patients had revascularization: 40.0% of those in the appropriate group, 18.9% of those in the inappropriate group, and 25.9% of those in the uncertain category (P<0.001). Obstructive CAD was more common in patients with angiography done by interventionalists and at sites capable of percutaneous coronary interventions or coronary artery bypass grafting (P<0.001 for each). Obstructive CAD was found in more than half (54.4%) of angiographic studies that were based on noninvasive testing and rated as appropriate, compared with 47.3% of angiographic studies that were rated as appropriate and were done based on pretest probability without testing in symptomatic patients.

The authors expressed caution in applying the appropriate use criteria in isolation for ranking specific cases or hospitals, citing an “overarching assumption” that obstructive CAD is a meaningful quality metric for diagnostic angiography. “However, finding obstructive disease does not necessarily mean the test should have been done,” they wrote. “In fact, the test is more meaningful when it finds normal coronary arteries in patients who would otherwise have been subjected to primary prevention medications and whose actual cause of symptoms would have been misdiagnosed.”

The authors of an accompanying editorial said that the study highlights some of the challenges and opportunities of applying appropriate use criteria to large data sets. While “big data” could be used to reduce health care costs, an ideal system would consider evidence-based medicine, use uniform and comprehensive clinical data, provide point-of-care decision support, and aim to improve quality by reducing overuse and underuse.

“The AUC [appropriate use criteria] could be the backbone of such a system and, if trusted by all stakeholders, could provide a practice-level alternative to preauthorization requirements or indiscriminant reductions in reimbursement,” the editorial stated. “Physicians must embrace the opportunity for self-regulation that AUC offers to ensure that we remain advocates for our patients and stewards of our health system.”