EXECUTIVE SUMMARY

Hospitals in Kaiser Permanente’s Southern California region have implemented a new approach to help identify which patients who present with chest pain require further testing or intervention and which can be discharged safely. The core of this approach is use of the HEART score, a tool developed in the Netherlands to help physicians risk-stratify such patients quickly. Full implementation of Kaiser’s approach followed an investigation of how it performed at 14 participating sites.

  • Between May and December 2016, investigators found that the HEART score performed well as a risk-stratification tool. Out of 12,128 patients evaluated using the HEART algorithm, the tool missed just 22 patients at 30 days post-ED visit. Investigators noted this is well below published national miss rates, which are in the 2% range.
  • Investigators also monitored compliance with the recommended use of the HEART algorithm. They discovered 12% of patients identified as low risk by the HEART score received noninvasive testing, were observed, or were admitted to the hospital.
  • While there have been no formal flow analyses, investigators noted that the new approach has affected efficiency positively, as more patients are discharged rather than remain in the hospital for further testing or procedures.

Patients who present with chest pain or other symptoms suggestive of acute coronary syndrome (ACS) are a concern, but emergency physicians struggle to determine which of these patients are at risk for an adverse outcome. Studies have shown that most will not suffer a heart attack or die in the near term, but how might clinicians identify the small number of patients for whom interventions (or at least further testing) would be beneficial? Adam Sharp, MD, MS, a research scientist and emergency medicine physician at the Southern California Permanente Medical Group in Pasadena, CA, explains that the quest for a new approach for this patient group has been a high priority for him for several years.

“[Chest pain] is a very common complaint. Emergency physicians see hundreds of thousands of these patients over the course of their careers, and there is a lot of variability,” he says. “I think [it is an area] that has a lot of opportunity for improvement from a patient care perspective, from a quality perspective, and from an efficiency/affordability perspective.”

Sharp is sharing how he implemented a solution to the problem he thinks would help other emergency providers. At the core of his approach is a standardized method for risk-stratifying patients that Sharp has found is both effective and manageable in the emergency setting. There are several decision-support tools that have been designed to help physicians risk-stratify patients with low-risk chest pain. Sharp’s own initiative began with an internal review of all these tools to determine effectiveness and suitability for Kaiser EDs.

Investigators found that while there was no clear winner among the tools in terms of effectiveness, the HEART score, a tool developed in the Netherlands, offered some advantages. “It was created from an ED perspective with information that every patient should have,” Sharp says. “That made it easier to actually put into place, and it was as good as any of the other options from our perspective.”

The next step involved integrating the HEART score algorithm into the electronic medical record (EMR) systems at 14 participating EDs so that the tool could be used to inform clinical decision-making for patients, Sharp explains. (Editor’s Note: See box at the end of this article for more information.)

Some information, such as lab results, would populate the HEART algorithm automatically in the EMR, while clinicians could easily enter other data points, Sharp notes.

Further, while clinicians maintained autonomy in their medical decision-making, the guidance from Kaiser was that patients with a score of 0-3 would be considered low risk and safe for discharge with follow-up by their primary care physician. Patients with scores of 4-6 would be deemed moderate risk and could be admitted to observation or to outpatient stress testing if troponin tests at zero and three hours were negative. Patients with scores in the 7-10 range would be considered high risk and should be admitted for inpatient evaluation and consultation by a cardiologist.1

The logistics of integrating the HEART score into the systems and flow of participating EDs were not difficult. The bigger challenge was convincing clinicians to use the new tool and to rely on the results in their clinical decision-making, Sharp observes. “Nobody wants to miss anything — in this case, a potential heart attack — if we could have identified it and done something to change it,” he says.

To address such concerns, Sharp and colleagues reviewed what the evidence base shows. Sharp indicated that there is no evidence that hospitalizing a patient with low-risk chest pain or sending such a patient for a stress test will somehow avoid the risk of a heart attack or even death for up to a one-year window following the ED visit.

“There is a lot of evidence to suggest that medical management of [low-risk chest pain] patients — just doing a good job of trying to mitigate the risks with blood pressure management, aspirin for those who need it, or whatever medical management may be ... is at least as good as the other interventions and tests that we have historically used to try and help people,” Sharp says. “[Furthermore], we know that based on science, if you take a low-risk population and do any diagnostic test on that group, there is more likely to be a false-positive than a true-positive. That leads you down a path where [patients] may get invasive procedures like coronary catheterization, which can cause all sorts of potential side effects. No one wants to go through a procedure like that if there is not a reason it is going to benefit them.”

Consequently, a big part of Sharp’s initiative involved engaging in discussions with clinicians about evidence and patient-centered care. “That is the part that really takes a lot of effort and culture change,” he says. “If you look at the ED, there is always a risk that when somebody goes home, something bad could happen. The more important question is: Have you identified the group that you might be able to help? That is what this assists with.”

Additionally, initiative leaders held a summit with both emergency and hospitalist physicians to present the evidence as well as their expectations that the clinicians would integrate the HEART algorithm into their decision-making for patients with potential ACS. The information presented at the summit was compiled into an online continuing medical education module that the physicians could access. Investigators noted that a small financial incentive was available to emergency physicians who participated in the quality improvement effort.

Maintain Physician Autonomy

Between May and December 2016, investigators found that the HEART score performed well as a risk-stratification tool. Out of 12,128 patients evaluated using the HEART algorithm, the tool missed just 22 patients at 30 days post-ED visit. Investigators noted this is well below published national miss rates, which are in the 2% range.

Investigators also monitored compliance with the recommended use of the HEART algorithm across the 14 participating sites. They found that 12% of patients identified as low risk by the HEART score received noninvasive testing, were observed, or were admitted to the hospital. This suggests there is room for improvement in persuading more physicians to use and rely on the HEART score in their decision-making. Sharp notes that decisions about care involve both the physician and the patient.

“Institutionally, because this has been a priority and incentivized in some ways to get physicians to use [the HEART score], they all use it. That doesn’t necessarily mean that they all follow the recommendations we have made for low, moderate, and high risk,” Sharp observes. “That goes back to the autonomy of physicians and the shared decision-making that goes on between the doctors and the patients they are treating. There always should be some variability there. The overwhelming majority of physicians are now using HEART scores for patients they are seeing.”

Efficiency has improved, too, although investigators did not perform any formal flow analyses, Sharp reports.

“We certainly see that more patients — predominantly, those in the lower-risk categories — are able to go home and not stay in the hospital and be observed [or admitted]. Fewer patients are being referred to noninvasive stress-testing centers,” he says. “These are clearly efficiency markers.” Such gains not only affect efficiency, but also patient care, Sharp suggests.

“We know in the ED there are always patients that we can spend more time on. This helps to reassure both patients and physicians that there is a large, low-risk group that really doesn’t warrant the extra observation and testing,” he says. “Physicians can then spend that time on patients who would otherwise be waiting for the limited resources that we have.”

For chest pain patients in particular, the HEART score provides a way to talk with patients about the risks and benefits associated with different care options.

“This creates a standard, evidence-based way to have those discussions,” Sharp says. “No one tells the physicians that they have to do anything in particular. They maintain their autonomy, as they should, to make what they consider to be the best clinician decisions for the patient, but here is some pretty strong evidence to discuss with patients and to use in decision-making.”

Are patients receptive to the information? Sharp says his group has not studied that aspect. However, in his own experience as a practicing emergency physician, Sharp has found most patients to be appreciative.

“There is no one right decision that is a cookie-cutter answer for everyone. I would say anecdotal experience has been overwhelmingly positive,” he says. “In fact, patients seem far less worried about this miniscule risk than the physicians.”

Sharp acknowledges that there are some patients who want to undergo available medical interventions — even when their risks are low. He notes that this is a valid opinion worth further discussion; however, patients should know about possible complications related to those interventions. “Some patients are still willing to take those chances, and that is part of the clinical decision-making process,” Sharp adds.

While more data from the initiative are forthcoming, using the HEART score is a standard practice now in Kaiser’s Southern California region. Kaiser EDs in other regions are beginning to adopt similar strategies, either with the HEART algorithm or similar tools as part of a standardized approach, Sharp reports. However, such strategies don’t imply that patients found to be at low risk for ACS simply should be discharged. “I think it is important to emphasize to these patients that they need to manage their chronic diseases,” he stresses. This may involve treatment for diabetes, high blood pressure, weight management, dyslipidemia, or other conditions, Sharp says. While emergency physicians do not typically provide ongoing treatment for chronic problems, they can refer patients to a primary care physician for such follow-up.

“I personally do have these discussions with patients,” Sharp says. “When I say there is no evidence that bringing the patient into the hospital or doing a stress test is going to reduce his risk of having a heart attack or potentially dying, I note that what we do know is that managing his diabetes well or managing his blood pressure, weight, and exercise ... are things we know will decrease the patient’s risks.” These are factors that patients and physicians can work together to control, and there is no need for additional cardiac testing to do that, Sharp advises. “It helps to identify things patients do have control over and can do to mitigate the risks of down-the-road problems,” he says. “Those are discussions that can happen between physicians and patients regardless of the system you are in.”

The HEART Algorithm

Sharp and colleagues integrated the HEART score algorithm into the electronic medical record systems at 14 participating EDs so that the tool could be used to inform clinical decision-making. Represented by the acronym HEART, clinicians were prompted to note a patient’s:

  • History — low, moderate, or high risk (0-2 points);
  • ECG — whether the results show a new ischemic change, a nonspecific change, or normal readings (2, 1, or 0 points);
  • Age — older than 65 years, 45-64 years, or younger than 45 years of age (2,1, or 0 points);
  • Risk factors — known coronary artery disease, a prior stroke, or peripheral artery disease (2 points) or other risk factors, including smoking, diabetes, or high blood pressure. Three or more risk factors = 2 points, one to two risk factors = 1 point, and no risk = 0 points;
  • Troponin — greater than 0.12 ng/mL (2 points), 0.041 ng/mL-0.12 ng/mL (1 point), and 0-0.40 ng/mL (0 points).

 

REFERENCE

  1. Sharp AL, Broder B, Sun BC. Improving emergency department care for low-risk chest pain. NEJM Catal 2018. Available online at: https://bit.ly/2v65DCc. Accessed July 27, 2018.

SOURCE

  • Adam Sharp, MD, MS, Research Scientist, Emergency Physician, Southern California Permanente Medical Group, Pasadena, CA. Email: Adam.L.Sharp@kp.org.