Assessment of the Diagnostic Accuracy and Reliability of the HEART Score Calculated by Ambulance Nurses Versus Emergency Physicians

Chest pain is a common complaint for consultation of emergency medical services worldwide. Currently, ambulance nurses (AN) base their decision to transport a patient to the hospital on their own professional experience. The HEART score could improve prehospital risk stratification and patient treatment. The aim of this study is to investigate the interrater reliability and predictive accuracy of the HEART score between AN and emergency physicians (EP). A retrospective analysis on data of 569 patients 18 years and older included in two prehospital HEART score studies. The endpoints are interrater reliability (intraclass correlation [ICC]) and predictive accuracy for major adverse cardiac events within 30 days of the HEART score calculated by AN versus EP. Predictive accuracy is sensitivity, specificity, positive predicted value (PPV) and negative predicted value (NPV). Interrater reliability was good for total HEART score (ICC 0.78; 95% CI 0.75-0.81). However, focusing on the decision to transport a patient, the ICC dropped to 0.62 (95% CI 0.62-0.70). History and Risk factors caused the most variability. Predictive accuracy of HEART differed between AN and EP. The HEART score calculated by AN was sensitivity 91%, specificity 38%, PPV 26%, and NPV 95%. The HEART score calculated by EP was sensitivity 98%, specificity 32%, PPV -26%, and NPV 99%. With a cut-off value of 0-2 for a low HEART score, predictive accuracy significantly improved for the HEART score calculated by AN: sensitivity 98%, specificity 18%, PPV 22%, and NPV 98%. Our study shows a moderate interrater reliability and lower predictive accuracy of a HEART score calculated by AN versus EP. AN underestimate the risk of patients with acute chest pain, with the largest discrepancies in the elements History and Risk factors. Reconsidering the cut-off values of the low-risk HEART category, as well as a carefully developed training program, will possibly lead to a higher interrater reliability of the HEART score and higher predictive accuracy used by AN.

The interobserver agreement of the HEART-score, a multicentre prospective study

BACKGROUND AND IMPORTANCE

Chest pain is one of the most common presentations to the emergency department (ED). The HEART-score is used to assess the 30-day risk of developing a major adverse cardiac event (MACE). The HEART-score enables clinicians to classify patients in low, intermediate, or high-risk groups though little is known as to whether this can be done reliably and reproducibly in a prehospital setting.

OBJECTIVE

The aim of this study was to compare the interobserver agreement of the HEART-score between ambulance nurses and ED physicians.

DESIGN, SETTINGS, AND PARTICIPANTS

Patients ≥18 years, with chest pain of suspected cardiac origin presented by ambulance to the EDs of four regional hospitals, were prospectively enrolled between October 2018 and April 2019.

OUTCOMES MEASURE AND ANALYSIS

The primary endpoint was interobserver agreement of the HEART-scores calculated by ambulance nurses compared to those calculated by ED physicians. Agreement was measured using Cohen's Kappa (K) both for overall HEART-score and dichotomized HEART categories. A secondary endpoint was the occurrence of a MACE at 30 days after inclusion.

MAIN RESULTS

A total of 307 patients were enrolled of which 166 patients were male (54%). The mean age was 64.8 years. In 23% (95% confidence interval, 18-27), patients were scored in the low-risk category by both ambulance nurses and ED physicians. The K for the overall HEART-score compared between ambulance nurses and ED physicians was 0.514. The K for the low-risk category versus intermediate and high-risk category was 0.591. Both are defined as 'moderate'. MACE within 30 days occurred in 64 patients (21%). In the low-risk group as defined by the ambulance nurses, there was a 7% risk of MACE compared to an average 5% MACE risk in the ED physician group.

CONCLUSIONS

The moderate interobserver agreement of the HEART-score does not currently support the use of the HEART-score by ambulance nurses in a prehospital setting. Training for prehospital nurses is vital to ensure that they are able to calculate the HEART-score accurately.