this is my napkin math read as well, but aspirin is an inhibitor of both COX-1 and 2 with somewhat higher affinity for the former. this leaves you messing with something else that's important for regulating platelets, your gut, etc. even selective inhibitors of the latter aren't out of the woods yet. there's some literature that's starting to suggest COX-2 may be at least somewhat constitutive in some tissues (endothelium, hence coronary vasospasm) even if it's also mostly inducible.

not sure why no one has tried developing a prostaglandin e2 inhibitor instead. going further downstream might be safer.

This was my first thought.

It's a complicated topic, biochemically / physiologically, however. Guidelines and recommendations have changed many times in the past couple of decades. With the scale of epidemiological data collection and specific studies carried out in recent years, conclusions are still rather "mixed".

There are many reasons to think, mechanistically, that aspirin (acetylsalicylic acid) should come out "net positive" in its effects on human health. But, even some of the "prostaglandins", for example, can be important in endothelial function and health (in a positive way).

I will note that I haven't kept up with this topic much in at least 5 or so years, but, I can point to a couple of sources with some analysis / discussion that should be of use to most, I think:

https://www.cuimc.columbia.edu/news/aspirin-making-sense-cha...

https://www.uspreventiveservicestaskforce.org/uspstf/recomme...

https://jamanetwork.com/journals/jama/fullarticle/2791399

https://www.annualreviews.org/doi/10.1146/annurev-pharmtox-0...

I specifically remember concerns roughly 10+ years ago that some of the expected net benefits might, in fact, actually be net harms in the broader population - particularly for other NSAIDs (non-steroidal anti-inflammatory drugs), but, even, at some point not long thereafter, for "aspirin" as well. I may dig out that info if I get a chance to try to refresh, myself, and will comment again if / when I do.

I always think of Lewis Black's bit on this, though (while pointing out that this is the nature of science and its best to go with the best info available at any time, generally, IMO):

https://youtu.be/iRsTtzYhTxk?t=5m27s

Classic baby/bathwater disposal problem. There are plenty of medicines that are beneficial for some applications but harmful for others. Maybe all of them, to some extent? That's why the authors wrote dozens of other sentences in that article.