Browsing by Subject "relative validity"

The aim of this study was to assess the relative validity of an interactive 24-hour recall used in a large randomized controlled trial among 15-month-old Malawian children. Relative validity studies should always be done when a dietary method is applied to a new population. The interactive 24-hour recall is a modified version of the traditional 24-hour recall. It was developed in 1995 in order to improve the validity of 24-hour recall in measuring nutrient intake in poor rural areas in Africa. The modifications of the interactive 24-hour recall are designed to make remembering food items and estimating portion sizes easier for the respondent. In the interactive 24-hour recall, the day before the recall interview the respondents are asked to use standard sized bowls and cups when eating, and to mark off each eaten food item on a picture chart containing pictures of local foods. The picture chart is intended to reduce memory lapses, and the use of cups and bowls is intended to facilitate estimating quantities of consumed foods. In this study, the relative validity of the interactive 24-hour recall was assessed among 44 children aged 15 months. The subjects were healthy children living in rural villages in Mangochi district, southern Malawi. To mimic the trial conditions, an intervention was included in this study. Half of the children were randomly assigned to receive a lipid-based nutrient supplement and the other half were assigned to follow their usual diet. The relative validity of the interactive 24-hour recall was assessed against a more accurate method of dietary assessment, the weighed food record (reference method). Intakes of energy, protein, fat, iron, zinc and vitamin A were measured by the two methods. Statistical analyses were done to assess the relative validity of the interactive 24-hour recall at both the population and the individual level. Food level data were analyzed to describe memory lapses and inaccuracies in portion size estimates as sources of measurement error. Furthermore, a description of differential measurement error was calculated to see whether such a bias might exist between the two study groups. Correction values developed in an earlier study in Malawi were tested to see whether they can be successfully used for adjusting values obtained by the interactive 24-hour recall. The relative validity of the interactive 24-hour recall was good at the group level, and moderate in estimating the intake of individuals. Memory lapses and portion size estimates introduced some errors to measurement. The interactive 24-hour recall performed well in assessing the intake of staple foods. Notable differential measurement error between the study groups did not exist. The correction values did not improve the results. Together, the results of the present investigation suggest that the interactive 24-hour recall is a relatively valid method for estimating the average energy and nutrient intakes of rural 15-month old Malawian children at the group level, and it is an acceptable method to be used to estimate selected individual nutrient intakes.