JMIR Publications

ABSTRACT

Background:

Accurate dietary assessment is needed in studies that include analysis of nutritional intake. Image-based dietary assessment applications have gained popularity in assessing diet, which may ease both researcher and participant burden compared to traditional pen-to-paper methods. However, few studies report the validity of these applications for use in research. Keenoa® is an imaged-based dietary assessment application intended for dietitians that recognizes food items using artificial intelligence, and permits for editing of food journals in-real time.

Objective:

To assess the relative validity of an image-based dietary assessment application — Keenoa®—against a three-day food diary (3DFD) and to test its acceptability of use in a sample of healthy Canadian adults.

Methods:

One hundred and two participants were recruited to complete two three-day food records. For two weeks, on 2 non-consecutive days and 1 weekend day, in random order, participants completed a traditional pen-to-paper 3DFD and the Keenoa® application. At the end of the study, participants completed the System Usability Scale (SUS). The nutrient analysis of the 3DFD as well as the Keenoa® data before (Keenoa-participant) and after it was reviewed by dietitians (Keenoa-dietitian) were analysed by ANOVA. Multiple tests, including the Pearson Coefficient, Cross-classification, Kappa score, % difference, paired T-test, and Bland-Altman test, were performed to analyse the validity of Keenoa®.

Results:

Seventy-two subjects completed the study. Most variables were significantly different between the Keenoa-participant and Keenoa-dietitian (P < 0.05) except for energy, protein, carbohydrates, fibre, vitamin B1, vitamin B12, vitamin C, vitamin D, and potassium. Significant differences in total energy, protein, carbohydrates, % fat, saturated fatty acids, iron, and potassium were found between the 3DFD and Keenoa-dietitian data (P < 0.05). The Pearson correlation coefficient between the Keenoa-dietitian and 3DFD ranged from 0.04 to 0.51. Differences between the mean intakes assessed by the 3DFD and the Keenoa-dietitian were within the 10% range except for vitamin D (misclassification rate = 33.8%). The majority of nutrients were within an acceptable range of agreement in the Bland-Altman analysis, except no agreement were seen for total energy, protein, carbohydrates, fat (%), saturated fatty acids, iron, potassium, and sodium (p < 0.05). Results from the SUS found that 34.2% of the participants preferred using Keenoa® in the future, versus 9.6% who preferred the 3DFD.

Conclusions:

The Keenoa® application provides an accurate assessment of dietary intake compared to the 3DFD. However, it was less reliable when analysing the average intake of energy, protein, carbohydrates, % fat, saturated fatty acids, and iron. Overall, Keenoa® showed better validity at the group level compared with the individual level; thereby suggesting it can be used when focusing on the dietary intake of the general population. Further research is recommended with larger sample sizes and objective dietary assessment approaches.