In this paper, David Ludwig, Walter Willett, and Mary Putt argue that many short-term diet trials may produce misleading conclusions because they fail to account for “wash-in” and “washout” effects—the time it takes for the body to adapt to a new diet and for previous dietary effects to fade. When dietary interventions last only a few days or weeks, researchers may measure temporary metabolic responses rather than the diet’s true long-term effects. This problem is especially common in crossover trials, where participants switch between diets without enough time for the previous diet’s effects to disappear, potentially biasing results.

The authors explain that major dietary shifts—such as moving between high-carbohydrate and ketogenic diets—can take several weeks or even months for the body to fully adapt. During this period, hormones, metabolism, energy expenditure, gut microbes, and other systems are still adjusting. If a trial measures outcomes too early, the results may exaggerate, underestimate, or even reverse the true effect of the diet.

Because many feeding studies last only one or two weeks and sometimes omit washout periods entirely, the authors suggest that a large portion of the nutrition literature may be vulnerable to this type of bias. They recommend designing longer studies—often at least one to two months per diet phase—and carefully accounting for adaptation periods to produce more reliable conclusions about how diet affects chronic diseases.