THE RELATIVE VALIDITY OF THE REPEATED 24H RECALL FOR ESTIMATING ENERGY AND SELECTED NUTRIENT INTAKES OF RURAL GHANAIAN CHILDREN

Objectives: To investigate the relative validity of the 24 h recall fo r rural southern Ghanaian children. Design: Cross-sectional study, in which food consumption was estimated for the same two days, using the recall and weighed record dietary techniques. Setting: Two villages in the Greater Accra Region of Ghana. Subjects: 72 rural children (39 fe males and 33 males; 56.2 +/- 9.4 months). Results: Overall agreement i n the number of foods reported was noted for 42% of records; 53% of re calls under-reported and 5% over-reported the number of foods consumed . Over 65% of snacks compared to less than 6% of main meal foods were missed on recalls. Average recalled portions were similar to weighed w ith the exception of cereal staples; for Slepor 250 g vs 295 g, P = 0. 007 and for Gidantuba 272 g vs 260 g, P = 0.02. Average daily recalled intakes of energy and most nutrients were lower (P less-than-or-equal -to 0.05) than weighed intakes in Slepor, but similar in Gidantuba. In both villages, the recalled and weighed nutrient intakes per MJ, and percentage energy from food groups, were similar. Intraclass correlati ons between recalled and weighed intakes ranged from 0.06 (energy) to 0.78 (vitamin A) per day, and per MJ from 0.36 to 0.76 for Fe in Gidan tuba and Slepor, respectively. Less than 35% of recalled energy and nu trient intakes were within +/-10% of weighed intakes. Conclusions: The recall could be substituted for the weighed record to estimate averag e intakes of energy and most nutrients, dietary quality and food consu mption patterns, provided the average recalled serving of cereal stapl es was accurately estimated. For assessment of individual intakes or s nack consumption, however, the recall technique was invalid.