Pictogram-based method of visualizing dietary intake

Background: Pictograms have been shown by many studies to be an effective way of conveying information. An easy-to-understand pictorial description is essential for communication of dietary intake in the computer era.

Objectives: We proposed a novel approach that represents textual descriptions of dietary intake into a pictorial representation with the concept of pictograms. The computational implementation in terms of a webbased tool was investigated on how well the pictograms carry their intended message.

Methods: 1) We investigated how well the pictograms are comprehended in terms of subjects’ accuracy rate and response time. In the study (n = 90), pictorial variants with three types of food images (black-and-white sketch, colored sketch, and colored photograph) were tested. 2) We also investigated how well subjects were able to select the standard food size among various food portions with the use of the tool. A comparison was made against the current standard of an educational session taught by a registered dietitian. We recruited 86 university students who were asked to select a standard size out of five different size categories. Three types of shapes were used. The bowl is the container that is widely used in the participants’ country. The pork strip was to represent foods with elliptical cross-section. The apple was used to represent a baseball-like size and shape.

Results: Two pictograms with lack-andwhite food image were low of less than 50% in accuracy rate. The rest of the twenty-seven pictograms derived from portions of the nine foods were well understood with high accuracy rates (above 85%). Participants in using the tool without the dietitian’s session was better than participants in the dietitian education session in selecting a standard portion size of an apple (p < 0.0001; p = 0.0009 after adjustment for gender and age). The rate of correct bowl and pork strip size estimates were similar between the two conditions (p > 0.05).

Conclusion: The development of pictograms could be used as a computational visual aid for comprehending and identifying dietary intake. Broader investigation is required for considering the effectiveness of the pictograms on recall, measurement, or estimation as well as for further evaluation in the clinical practice.