Estimation accuracy of bean bags as portion size estimation aids for amorphous foods

Abstract

Objective: A study was undertaken to explore the portion size estimation accuracy of bean bags as low-cost volumetric portion size estimation aids for amorphous foods.
Designs: Three observational, cross-sectional and three experimental/quasi-experimental developmental evaluation sub-studies were carried out.
Settings: Observational, cross-sectional: one retirement village and four schools. Experimental/quasi-experimental: one South African university.
Participants: Observational, cross-sectional: elderly persons, adult women (school educators), school children.
Experimental/quasi-experimental: university students. In total N = 541; > 3 800 observations.
Variables measured: Using a standardised set of bean bags (test object; volume range: 60–625 ml), volumes of different amorphous foods (reference objects: actual foods or representations) in varying portion sizes had to be estimated. Accuracy (outcome measure) was perfect if volumes of test and reference object were identical. Acceptable estimation accuracy allowed for misestimation by one bean bag size. Test–retest reproducibility was also assessed.
Analysis: Descriptive statistics (proportions perfect and acceptable accuracy).
Results: Across the sub-studies, perfect accuracy ranged from 22–65% depending on participants, reference food and portion size. Irrespective thereof, acceptable accuracy was noted in > 70% of observations. Reproducibility varied (range: 28–67% agreement).
Conclusions and implications: Perfect portion size estimation of amorphous foods remains challenging. When misestimation by 60–125 ml still serves the purpose of a dietary assessment, bean bags show promise for cost-effective food volume quantification, especially on group level in resource-limited settings.
Keywords: dietary assessment, portion size estimation, portion size estimation aids, accuracy, amorphous foods, bean