Quality-of-life scores and health outcomes (expressed as quality-adjusted life-years, QALYs) were similar for the two groups, as were the average direct healthcare costs per patient.

“That we did not find a decrease in costs is probably due to the hesitance of physicians to discharge low-risk patients from the ED without further testing,” Dr. Poldervaart explained. “Hopefully, in time (and more publications of the HEART score now appearing almost weekly from all over the world) this effect on use of health care resources will become more apparent.”

The probability that HEART care dominated usual care was 71 percent, and the probability that HEART care was cost-effective and a willingness-to-pay threshold of 20,000 euros (around $22,000) per QALY was 99 percent. Physicians’ nonadherence to HEART guidance occurred in 41 percent of low-risk patients (e.g., admission and additional testing) and 12 percent of high-risk patients (e.g., no further diagnostic testing).

Reasons cited for nonadherence in low-risk patients included intuition (24 percent), an alternative diagnosis being more probable (11 percent), and logistics (9 percent). No reasons were given for the rest.

“The HEART score is a decision support tool, not a strict protocol,” Dr. Poldervaart said. “The HEART score is an accurate tool, but is no gold standard (just as usual care has no 100% sensitivity).”

“Further research should focus on identifying low-risk patients, since the main barriers to follow the rule are in these patients, and at the same time this is the group where considerable reduction in the use of health care resources and harm of overdiagnosis to low-risk patients can be achieved,” she added.

Dr. Udo Hoffmann, chief of cardiovascular imaging at Massachusetts General Hospital in Boston, told Reuters Health by email, “The HEART Score worked in this population, given that no cardiovascular death or MI occurred in this population and the majority of MACE was unstable angina or PCI, which can be medically managed.”

“In my opinion the results strongly support the use of the HEART score as recommended,” he said. “Low-risk HEART score patients can be discharged.”

Dr. Hoffmann added, “It would have been interesting to see whether the physicians’ adherence to recommendations would have increased, if the trial would have incorporated a pilot period, showing participating physicians the results of HEART score care in their own patients. In addition, knowing the motivation for non-adherence in more detail would probably have been very informative.”