Serra Majem L, Morales D, Domingo C, Caubet E, Ribas L, Nogués R M
Unidad de Medicina Preventiva y Salud Pública, Facultad de Medicina, Universidad de Barcelona.
Med Clin (Barc). 1994 Nov 19;103(17):652-6.
Methods of dietary collection at an individual level include various procedures with different estimates of food, energy and nutrients intakes. The aim of this study was to compare two dietary methods widely used in nutritional epidemiology: a 24-hour recall and a semiquantitative food frequency questionnaire.
One hundred and fifty-five people selected at random from the population of the Valle de Arán (Lérida, Spain) were surveyed. Nutritional intake information was obtained from both methods.
The frequency method overestimated all the nutrient intakes only before adjustment by energy intake, except for cholesterol. After adjustment by energy intake estimations from both questionnaires were very similar; the percentages of energy supplied by proteins, fatty acids and carbohydrates were very similar, too; the polyunsaturated/monounsaturated fatty acid ratios, were almost the same. Results show a level of correlation between the questionnaires that ranges from 0.81 to 0.23. Nutrients that showed the highest crude correlation between both questionnaires were alcohol, energy, carbohydrates, sodium and iodine (R > or = 0.7) while vitamins C, B1, B6, B12, and A, niacine, folic acid, iron, phosphorus and potassium showed the lowest (R < or = 0.5). Energy adjusted correlations were quite different to unadjusted ones particularly for lipids, monounsaturated fatty acids, carbohydrates, fiber, cholesterol, zinc and sodium, for which a lower coefficient was calculated after adjustment, and vitamins B2, B6, C, D and potassium, for which a higher coefficient was estimated after energy adjustment.
Both methods make similar assessments when adjusting for caloric intake, but food frequency questionnaire tends to overestimate food consumption.
