Causality and the Bradford Hill criteria

In epidemiology we often look for a causal relation - that is if something is the cuase of the effect. Is it for example the fish itself that is good for the heart? (it could be that people who eat fish also do more exercise - and then the fish may not be the cause - but simply just associatied with the increased health)

To assess causality one can use the Bradfor Hill criteria - here is my interpretation of the 9 criteria - with one additional criterium proposed by other clever people (clever like Bradford Hill :-)