I don't disagree. It's actually my belief that most nutritional and health science is likely bunk. Partially, it is for the same reasons there is a lot of bad science, which you touch on. But it also has it worse than other fields - look how bad a lot of CS papers are, where you're dealing with a domain that is, relatively speaking, one in which you can generate immediate, obvious, and empirically demonstrable results. Now, imagine a field dealing with the most intricate and complex system known to man, where actual clear, concrete results may take decades to manifest, there are zillions of confounding factors, the code is insane spaghetti, truly controlled experiments are practically impossible, and every individual is slightly different and may operate differently. It's really a field where can have very little practical faith that it will ever be much good, but we can hope.
That's why I prefer for myself to hold to heuristics. We have good life expectancy and diet data going back a long time. We can see, for example, various periods and places where, excluding childhood mortality, people lived as long or longer than they do now - remarkable and slightly horrifying, considering the advances in medical science in the past 50-70 years. This included peoples with diets high in saturated fats. The best practical advise you can give anyone, I believe, is simply to eat traditional foods your ancestors have eaten for hundreds of years, and to therefore avoid anything prepared, sticking to fruits, vegetables, grains, dairy, and meats. And if at all possible, try to get the food you buy grown in a closer-to-traditional matter; ruminants should eat grass, not soy and corn, for example; oats, if possible, should not have been treated with Roundup. And naturally, whatever your diet, do not let yourself get overweight and try to maintain a high level of productive physical activity (of course, unproductive is better than none at all!)