2.6. Statistical Analysis

The whole statistical examination has been performed using the statistical software StatPlus (StatPlus Pro v7©, AlaystSoft Inc., Walnut, CA, USA). Categorical variables were summarized in crosstab and expressed as % of the own group and analyzed using a χ² test, and if the test gave a significant result, that was further analyzed with a z-test. Continuous variables were shown as median and interquartile range (IQR). The differences between medians were tested using the test U of Mann–Whitney. The Box-Plot graph was realized for the main results to visualize the values of distribution.

A two-sided p < 0.05 was considered statistically significant.

For the statistical analysis, we grouped together patients with an HFpEF (LVEF > 50%) and HFmrEF (LVEF between 41–49%), because we did not find any significant difference between the main variables (outcome; p-value 0.5, and hs-cTnI values; p-value 0.4). Furthermore, given the setting of ED, an ejection fraction of less than 40% was chosen as able to identify a compromised ventricular function, assuming that this value is less susceptible to variability between operators.