We found few and variable operational definitions of PHD, despite American College of Cardiology/American Heart Association recommendations to report specific intervals. Worryingly, definitions of symptom onset, the most elusive component of PHD to establish, are uncommon. We recommend that researchers (a) report two PHD delay intervals (onset to decision to seek care, and decision to seek care to hospital arrival), and (b) develop, validate and use a definition of symptom onset. This will increase clarity and confidence in the conclusions from, and comparisons within and between studies.