The HEART score (Six, Neth Heart J 2008) has likely become one of the most popular risk stratification tools for patients with potential ischemic chest pain. Two issues have always struck me as problematic: (1) the derivation is methodologically (Green, Ann Emerg Med 2014) grotesque, and (2) the 42-day major adverse cardiac event rate is 1.7 % even for the lowest risk patients, which may not be acceptable to clinicians.

Drs Steven Green and David Schriger have now been far more eloquent in a recent Annals of Emergency Medicine publication (2021). In brief, here are the "good" aspects:

1. The study population is relevant

2. Predictor variables recorded independently of outcomes

3. The rule is widely used, clinically sensible, easy to apply, and can be integrated into practice.

Here are the "not good" aspects:

1. Initial sample size (122 patients with 29 MACE) not justified and insufficient.

2. Many predictor variables (for example, patient sex, type of pain) were missed, and included variables had poor inter-observer reliability, especially "history".

3. No justification for coding (ie 0, 1, or 2 points) For example, patients > 65 years with 3 or more risk factors have an automatic score of 4 (moderate risk) prior to any evaluation. NOT GOOD. While HEART score was supposed to be "Apgar for chest pain" the points are too heavily weighted to EKG, risk, and age, and insufficiently for history and troponin.

4. Importantly, despite many, many studies on the score, no one has compared the rule to unstructured physician judgement.

5. The rule sensitivity is not acceptable--most physicians will accept a 1% miss rate, and the sensitivity of the HEART score and its variants is greater than this threshold.

6. The HEART score does not inform clinical decisions with respect to follow-up.

Overall, the score provides a simple, comforting algorithm--in addition to a catchy nickname--and this likely accounts for its widespread popularity. However, clinicians are encouraged to think rather carefully about widespread application of the HEART score or any of its variants.

Happy to engage in further discussion!