Do time trends in food supply levels of macronutrients reflect survey estimates of macronutrient intake?

Abstract

Two types of data may be used to estimate trends in food and nutrient intake by the US population: per capita food supply estimates and survey estimates of individual intake. Because these data vary markedly in measurement goals and methods, we examined whether trends in food supply and survey intake estimates for fat, carbohydrate, and protein are reflective of one another. The data selected for comparison included all available survey estimates of mean intake by the US population (i.e., periodic estimates from 1965 to 1988) and all available per capita food supply estimates from a comparable time period (i.e., annual estimates from 1965 to 1985). The two types of data generally did not reflect the same trends. Furthermore, expressing macronutrient levels as percentage of calories rather than in grams affected the trend relationships. Our findings indicate that caution is needed in the selection and application of available data to estimate trends in macronutrient intake by the US population and in the interpretation of these data with regard to public health research, policies, and programs.