Recording nutritional intake is useful because it allows associations between particular food components and health effects to be found. Metabolic studies in laboratories allow accurate recording of nutritional intake from subjects, but they do not reflect people in their natural free living environment, and clinical effects can bias the results. Free living studies therefore offer advantages over studies conducted in metabolic wards because they are more representative of the everyday lives of people and how they interact with food. Free living nutritional studies may require the study participants to record or recall information pertaining to their diet. For this reason the use of methods to record nutritional intakes, such as 24 hour recall and the food frequency questionnaire, are required in such studies. Much effort has been expanded in attempts to find reliable methods of collecting nutritional information from free living populations. But how reliable are these methods at providing researchers with accurate information of a subject’s food intake?
The validity of the food frequency questionnaire has been investigated using student1. Researchers asked the study participants to record their food intake on self-reporting forms. The actual food intake was known to the researchers because the students were eating in halls of residence and the meal plans were available. The results showed that the information recorded on the food frequency questionnaire correlated fairly closely to the actual intakes recorded by the subjects. Calculations that revealed the degree of over- or underestimation showed that foods that were major components of meals were estimated to greater accuracy than those foods that could be considered minor components. Therefore the food frequency questionnaire may be a valid method of obtaining food data for nutritional studies. The questionnaire is advantageous because it does not require interview of the subjects and is therefore more convenient. However, the exclusion of researchers in the data collection causes the food frequency questionnaire to remain controversial.
While it may be argued that the metabolic ward studies are the gold standard because of the ability to control and accurately record variables, such studies are by their nature artificial2. They also suffer from the problem of only being able to include small numbers of individuals, which invites the problem of selection bias, and are of short duration. The larger numbers of subjects in population studies is advantageous because it is more likely representivity of the population as a whole. In free living populations associations between dietary fats and plasma lipids are rarely found, whereas the findings from metabolic ward studies do show such associations. This may relate to the fact that studies in metabolic wards often involved changes to food intake, which can create artificial and even stressful conditions. In particular isocaloric substitution of one food for another requires changes to more than one food. Increasing saturated fat content usually involved reducing fibre content, and as such the cause of metabolic changes can never be known for sure.
In addition, researchers are often to blame for methodological problems associated with the use of data from free living populations, as they try to manipulate and use the data in the same way as data from metabolic ward studies. However, the use of different methodologies with free living populations may be appropriate. For example, researchers often take ‘snapshot’ samples of food intake in studies of free living people, forgetting that food intake varies over time. A food diary that recalls the food eaten in the preceding 24-hours, is obviously not truly representative of the food eaten over a longer time period. In this respect, the limitations of such data is often not emphasised in such studies. Recording nutritional intakes over longer periods, perhaps with multiple self-reporting questionnaires may offer a solution to this problem. Therefore, despite investigations showing good correlations between self-reported nutritional intake and actual nutritional intake, the reliability of recording nutritional intake in free living populations is still controversial.
RdB
