Validation of a web-based dietary assessment program against 24-h recalls in adults with type 1 diabetes

Nutrition Data is a web-based program for nutrition analysis and registration of diet and exercise. It may aid dietary assessment and carbohydrate counting in people with type 1 diabetes (T1D) but requires validation.

To assess relative validity of Nutrition Data in measuring energy, carbohydrate and other macronutrient intake and evaluate the program's user acceptability, in adults with T1D.

In this validation study, we analyzed data from 42 participants (median age 46.5 years, 45% women) from the DANCE study, a randomized controlled trial comprising of individuals with T1D in Sweden. Mean intakes of energy, carbohydrates, fat, protein, alcohol, fiber, sugars and saturated fat from 2 days registered in Nutrition Data were compared against the respective intakes acquired by unannounced 24-h-recalls of the same days. Paired sample t-tests and Wilcoxon matched-pairs signed rank tests were used to compare mean intakes between the two methods, and Spearman's rank correlation and Bland Altman plots were used to assess agreement between the methods. Usability and user acceptability of Nutrition Data were assessed with a questionnaire.

There were no significant differences in mean dietary intakes between the two methods. Spearman's correlation coefficients ranged from r = 0.79 for energy intake to r = 0.94 for carbohydrate intake (% total energy intake) (p < 0.001 for all outcomes). The Bland-Altman plots showed no clear patterns of bias, though limits of agreement were relatively wide. Most participants found Nutrition Data easy to use (70%), helpful for carbohydrate counting (88%) and would recommend it to others (73%).

The web-based program Nutrition Data showed good validity in assessing intake of energy and macronutrients compared to 24-h recalls and high user acceptability in Swedish men and women with T1D, and could, therefore, be used to facilitate diet registration and carbohydrate counting.