The ACTION Registry-GWTG in-hospital mortality risk model for heart attack patients has been updated to include cardiac arrest, and validated as a robust instrument for risk adjustment and benchmarking of mortality outcomes, according to a study published in the Journal of the American College of Cardiology.
The updated model is based on data from 2012 and 2013 and replaces an earlier version, which was based on 2007 and 2008 data. The revised risk model is more robust and generalisable to a specific population of heart attack patients, according to a press release. The model now includes cardiac arrest, an addition viewed by some as critical as cardiac arrest is often an important predictor of mortality following a heart attack. Hospitals in the registry are able to regularly access metrics comparing them to all other participating hospitals, including the in-hospital mortality risk. These metrics allow hospitals to make quality improvements in care.
Using data from the ACTION Registry -GWTG, researchers examined the records of 243,440 patients from 655 hospitals between January 2012 and December 2013 to create the risk model. The new model includes age; heart rate; systolic blood pressure; presentation after cardiac arrest, in cardiogenic shock, or in heart failure; type of heart attack; and the blood levels of creatinine; and troponin.
Researchers also developed a simple risk score predicting the likelihood of individual patients dying based on their unique profile. This score is designed to assist clinicians in determining their patients’ prognosis and optimal care plan.
The results showed that the overall in-hospital mortality rate for heart attack patients was 4.6%. The risk scores varied considerably, depending on the patients’ demographics and clinical features. For example, a younger heart attack patient without other risk factors and not experiencing cardiac arrest might have a less than 1% likelihood of dying, while an older heart attack patient with many other risk factors and presenting after cardiac arrest might have up to a 50% likelihood of dying.
Robert McNamara, the study’s lead author and associate professor of medicine at Yale School of Medicine in New Haven, USA, says, “The model performed well across a broad range of subgroups, including those with and without cardiac arrest and with and without other high-risk characteristics, illustrating the value of the model for benchmarking mortality outcomes even for hospitals caring for different types of patients.” He adds that “adjusting for cardiac arrest among heart patients is critically important and enables a fairer assessment for hospitals that care for these patients”. Furthermore, the model “should enhance research into best practices to further reduce mortality in heart attack patients.”
In an accompanying editorial, Peter WF Wilson, professor of medicine in the Division of Cardiology at Emory University School of Medicine in Atlanta, USA and Ralph B D’Agostino Sr, professor of mathematics and statistics at Boston University, Boston, USA, say the results from this study “provided very contemporary data demonstrating the progress made in the care of heart attack patients over the past decade.” They note that the results show “the dynamic nature of health risk appraisals,” in that there is now “extensive information related to risk factors, recent medications, past history of coronary disease, and severity of clinical presentation” available in the hospital setting.
Wilson and D’Agostino also pointed out that numerous other risk models have been developed. In addition to the 2000 model, TIMI (thrombolysis in myocardial infarction), investigators from GRACE (Global Registry of Acute Coronary Events) designed a risk score based on more than 24,000 patients with a diagnosis of acute coronary syndrome admitted to hospitals across numerous geographic regions. The HEART (history, ECG, age, risk factor, troponin) Score has been proposed to help determine whether patients being evaluated for coronary-related symptoms in the emergency department should be admitted to the hospital or sent home.
Given the growing number of risk models, Wilson and D’Agostino suggest that “a comprehensive cross-validation and comparison across at least some of the algorithms would help at this point.” For example, in their view, the HEART score will be more useful in situations where it is important to assess the significance of risk factors, while the ACTION-GWTG score will likely be the most useful for evaluating severely ill patients and determining interventions for them.
“It is likely that one score does not fit all,” they say. Each model provides a useful summary of risk to guide decision-making for patients with heart-related symptoms, depending on their severity and the duration of the follow-up period. “Consensus building would help to move this field forward for hospital-based management of patients evaluated for cardiac ischemia,” they conclude.
