Is the hand scale an appropriate tool for guiding and estimating food portions? An evaluation among free-living adults

The hand scale, which uses the number of hands (e.g., fists and palms) to estimate portion size, is widely used in practice. However, evidence regarding its appropriateness in free-living settings is limited. This study aimed to examine the relationship between actual food consumption (g/meal) and the self-reported number of hands used for estimating portion size and to develop and evaluate models for estimating food consumption per meal based on the self-reported number of hands. In this cross-sectional study, 1081 free-living Japanese adults aged 20–69 years completed a 4-day weighed dietary record, documenting food consumption in grams alongside the number of hands used to estimate portion size (fists for grain foods and fruits, palms for protein foods, and handfuls for vegetables) at breakfast, lunch, and dinner (12,148 meals in total). Correlations between the actual food consumption and the self-reported number of hands were examined. Models for estimating food consumption, incorporating the self-reported number of hands and participant characteristics, were developed and evaluated using a randomized 1:1 split-sample approach. Spearman correlation coefficients between the actual food consumption (g/meal) and self-reported number of hands were 0.59 for grain foods, 0.85 for fruits, 0.72 for protein foods, and 0.76 for vegetables. In the development group, models using the self-reported number of hands explained 34 %–65 % of the variance in food consumption per meal. In the evaluation group, despite wide limits of agreement, actual and model-estimated food consumptions were comparable (mean difference: −2.5 to −0.3 g/meal), with moderate to strong correlations (Spearman correlation coefficient: 0.61–0.75). In conclusion, the hand scale can serve as an appropriate portion size estimation tool. Further research is warranted to evaluate its applicability across different foods and populations.