The fundamental flaw in obesity research.

The basic problem with comparative diet trials is our inability to measure what people eat. All conventional instruments depend on subjects' reports. Most trials lack independent biochemical, physiological or genetic measures of intake. So, we do not know if subjects actually follow the diets being tested and compared. We can assess weight gain/loss, but we fail in a fundamental scientific requirement, accurately measuring the independent variable in a causal experiment. Worse, we know most subjects under-report their energy intake and its components, the obese especially. The problem is compounded by attempts to show diets' effects on other risk factors, like triglycerides. Researchers seek to correlate two variables, without having accurately measured one of them, producing misleading associations. The consequence is we do not know if the results of any current diet trials are valid or reliable. Developing rigorous measures of food intake is the highest priority in obesity research. That involves improvements in technology as well as science. We need: (1) biomarkers of intake for energy, macro- and micro-nutrients and other food components relevant to weight gain/loss; (2) field measuring instruments that are cheap, rapid, painless, non-intrusive and self-administerable; and (3) electronic data transmission systems that preclude subjects' ability to misreport.