A Better Model of Heart Attacks

In a previous post, I presented a causal model of heart attacks and commented that the model could be improved. Pictured above is an improvement that specifically includes coronary artery blockage. Compare it to the excellent discussion and animation available from the National Institute of Health. The way to read this causal diagram is to look at the input variables and trace their effects through the system (as you are tracing a single path through the diagram, the effects are positive until you encounter a negative sign):

(1) Genetics (family history) and life style (overweight, smoking, diabetes, etc.) can lead to coronary blockage.

(2) Intervention in the form of a diagnosis can lead to either prevention, drug treatment or angioplasty and stents, all of which can reduce blockage and reduce the chance of heart attack.

(3) Heart attacks, however, can also be caused directly by shocks (e.g., taking cocaine, emotional stress or pain, exposure to extreme cold, cigarette smoking, etc.) or heart spasms (tightening of a coronary artery that cuts off blood flow through the artery--the causes of coronary artery spasm are not well understood).

The causal diagram above seems to capture the current understanding of heart attacks but it doesn't explain why you or a sample from the general population will have a heart attack and it doesn't predict which interventions will reduce the chances of having a heart attack. This is because there are too many back-door paths that can create a heart attack.

You could (1) have a heart spasm, (2) continue smoking, (3) drop out of your cardiac rehab program, (4) gain weight, (5) become diabetic, (6) become a drug addict, (7) stop taking your prescription medications, etc. etc. Out of all this, the only thing you can't change are your genetics or family history and the stents you may or may not already have in one or more coronary arteries. If angioplasty and stents don't prevent future heart attacks, as some research suggests, it does not mean the angioplasty is an unnecessary or ineffective treatment. There are simply too many back-door paths to that future heart attack to draw that conclusion.