Why is single sample rule out of non-ST elevation myocardial infarction using high-sensitivity cardiac troponin T safe when analytical imprecision is so high? A joint statistical and clinical demonstration