Mediterranean-type diet patterns are consistently associated with reduced risk for cardiovascular disease, diabetes, and cancer in the general population. In contrast, several randomized controlled trials (RCTs) focusing on nutrient supplements have shown no or adverse long-term effects on long-term chronic disease. Food in its natural form is a nonrandom mixture of numerous molecules, orchestrated evolutionarily to maintain the life of the organism being eaten. Food synergy assumes that the biologically determined combination of nutrients and other bioactive substances found in food plays a concerted role in influencing health. Coupled with reduced risk in the Mediterranean-type diet pattern, food synergy implies that the concerted action of nutrients and other bioactive substances in fruit, vegetables, whole grain cereals, nuts, and legumes is beneficial for health. The assumption that single molecules work in isolation as they work in food violates the food synergy concept of concerted action and often leads to a partial picture. Public health nutrition strategies that focus on single nutrients have led to a flourishing diet supplement industry and advice to the public to eat low-fat diets. The latter is questionable in two respects: not all fats are equal, and industry efforts to comply spun-off products high in refined carbohydrates. It is time to rethink the research paradigm concerning diet and health. Reductionist research, though valuable, focuses on partial pathways, rather than the whole system integrating a lifetime of food intake with the long-term health of intact humans. Epidemiology provides this information, but is subject to residual confounding. RCTs are useful, but RCTs of food differ fundamentally from RCTs of drugs; for example, in terms of blinding, long-term adherence, and specificity of the reference treatment. All research inferences are most secure when based on convergent evidence from multiple research approaches. A balanced approach is therefore needed in nutrition research.