Validity of photographs for food portion estimation in a rural West African setting.

<bold>Objective: </bold>To validate food photographs for food portion size estimation of frequently consumed dishes, to be used in a 24-hour recall food consumption study of pregnant women in a rural environment in Burkina Faso. This food intake study is part of an intervention evaluating the efficacy of prenatal micronutrient supplementation on birth outcomes.<bold>Subjects: </bold>Women of childbearing age (15-45 years).<bold>Design: </bold>A food photograph album containing four photographs of food portions per food item was compiled for eight selected food items. Subjects were presented two food items each in the morning and two in the afternoon. These foods were weighed to the exact weight of a food depicted in one of the photographs and were in the same receptacles. The next day another fieldworker presented the food photographs to the subjects to test their ability to choose the correct photograph.<bold>Results: </bold>The correct photograph out of the four proposed was chosen in 55% of 1028 estimations. For each food, proportions of underestimating and overestimating participants were balanced, except for rice and couscous. On a group level, mean differences between served and estimated portion sizes were between -8.4% and 6.3%. Subjects who attended school were almost twice as likely to choose the correct photograph. The portion size served (small vs. largest sizes) had a significant influence on the portion estimation ability.<bold>Conclusions: </bold>The results from this study indicate that in a West African rural setting, food photographs can be a valuable tool for the quantification of food portion size on group level. [ABSTRACT FROM AUTHOR]