There are causes within causes within causes and each is usually characterized by multicausal density. So an everage citizen's health may be substantially determinable by knowing, country of residence, genetic heritage, eating habits, exercise habits, cleanliness and health conscientiousness. Pick one of those factors, say, eating habits. You can then look at childhood habits on adulthood health and longevity, adult habits, grazing versus set meals, fresh versus processed, organic versus fresh, variety versus constancy, etc. You could then pick one of those such as freshness and pursue a sequence of causes - regular fresh versus organic versus locality, percentage mix of fresh versus processed, match of fresh vegetables to presumed historical patterns of consumption, etc.
There are layers upon layers, each involving some degree of uncertainty. More critically, there is usually an escalating diminution of impact. For example, eating large volume of iron-rich vegetation such as spinach might halve your incident rate of stomach cancer but the original incident rate is only 1/100,000. Is that a worthwhile return on changed behavior? And what if voluminous consumption of spinach doubles your chance of esophageal cancer (1/1,000,000 rate); does that change the calculation? And what if eating that much spinach makes it difficult to eat some other food that has different beneficial aspects? The complexity of benefit calculation is enormously complex about essentially vestigial effects. In TJRadcliffe's terms, a presumed deep and narrow combination of choices and risks with an uncertain optimum outcome.
There is still a deep, underlying issue with the example of breastfeeding: there is a presumption that there actually exists a deep, narrow optimum in behaviour that science will clearly and unequivocally identify for us, and that we fall out of that optimum at our (or our children's) peril. While the observation that some effects are too small to be worried about is good to see, the very presumption that any behavioural science is able to provide useful guidance for the betterment of human life in the modern world needs to be questioned.
While clearly some behaviours are better for us and others worse, the incremental value of changing any given behaviour is generally not that big. Humans are extremely flexible creatures endowed by our genetics with enormously long lives for fairly obvious evolutionary reasons (grandparents, not parents, transmit culture, and cultural transmission is a reproductive advantage, so children who had long-lived grand-kin back in the stone age were more likely to have more children than those without.) The huge gains in lifespan in the past century are the result of better nutrition, public sanitation and public health, all of which are the results of good science (as well as good politics and good engineering.)
Given those huge gains, we are now playing around at the margins of what our bodies are capable of, so most of the leading-edge stuff we read about is going to have at best marginal effects. This is why so often studies have ambiguous results, or contradict other, similar work: the effects being measured are small. The importance we grant to them in our own lives should also be small.
Unfortunately, irrational, power-mad people take these small perturbations on our overall understanding and treat them as if they mattered as much as the discovery of anti-biotics or sterile technique in surgery. Highly emotional issues like breast-feeding and infant male circumcision (which has the most hate-filled, psychotic gang of anti-scientific fear-mongers opposing it on the basis of completely bogus scientific claims behind which shelter quite worthy moral grounds that intactivists are unfortunately too gutless to own) lead power-hungry cowards to wrap their personal moral views in a false cloak of scientific objectivity and then attack and oppress anyone who disagrees, which is a sure sign they do not care at all about science--knowing--but only about power. If they cared about science, they would be willing to discuss and acknowledge how small and relatively unimportant the various effects at issue actually are, and understand the importance of parental choice in these questions.
[I will now stand back while hate-filled psychotic intactivists who believe they have magic knowledge about people they have never met make false assertions about the state of my physical, moral and psychological being, combined with trivially false "scientific" claims that serve no purpose but to shelter their purely moral stance from honest scrutiny and debate.]