Can you please link that study? Is it the one where hospital employees were observed from more than a year ago? I vaguely remember there was one from Thailand that was relevant at the time.

My methods are probably lacking severely from the perspective of a statistician.

However taking an honest look at the data, I don't see any of them coming to a different conclusion than I have. They could end up saying it's 500,000 or 1,500,000 damaged hearts and I might not even disagree.

The basic idea is that data is sufficiently random for symptoms to be represented equiproportionally. For diagnoses and other medical concepts it is different.