Written by Thomas Davis
Emergency departments should reconsider using the HEART score and its variants due to significant concerns regarding the rule’s creation and performance.
Why does this matter?
Chest pain is a high risk chief complaint in the emergency department. In the United States, the HEART score has been largely adopted as a simple way to risk stratify patients. It is easy to memorize and creates a perception of medicolegal protection. But how does the HEART score (and its variants) measure up against the Annals of Emergency Medicine’s methodological standards for decision rules?
A Shot Through the HEART
After performing a structured analysis using the Annals’ framework, the authors concluded that the HEART score has several weaknesses that clinicians must consider:
The content of the HEART is more important than it looks: The HEART score was never formally derived or validated. Instead, it was created to be a cute and simple mnemonic that assigns an arbitrary scoring system. Not all risk factors are the same, and some important ones are completely ignored.
Better than gestalt? It’s debatable. To beat gestalt, we must achieve a greater sensitivity or lower resource utilization (without lowering sensitivity).
Poor interrater reliability: The kappa scores of history are more frequently closer to random than perfect agreement. What would stop an expert witness from saying that your patient deserved to be a HEART score of 4 when you assigned a 3?
Lackluster performance: Pooled sensitivities from 3 meta-analyses have an underwhelming 96-97% sensitivity. The lower bounds of the confidence interval should make you even less confident (93%).
A Methodological Appraisal of the HEART Score and Its Variants. Ann Emerg Med. 2021 Apr 28;S0196-0644(21)00118-9. doi: 10.1016/j.annemergmed.2021.02.007. Online ahead of print.