A Diet Profiling Algorithm (DPA) to Rank Diet Quality Suitable to Implement in Digital Tools-A Test Study in a Cohort of Lactating Women

Abstract

Although nutrient profiling systems can empower consumers towards healthier food choices, there is still a need to assess diet quality to obtain an overall perspective. The purpose of this study was to develop a diet profiling algorithm (DPA) to evaluate nutritional diet quality, which gives a final score from 1 to 3 with an associated color (green-yellow-orange). It ranks the total carbohydrate/total fiber ratio, and energy from saturated fats and sodium as potentially negative inputs, while fiber and protein are assumed as positive items. Then, the total fat/total carbohydrate ratio is calculated to evaluate the macronutrient distribution, as well as a food group analysis. To test the DPA performance, diets of a lactating women cohort were analyzed, and a correlation analysis between DPA and breast milk leptin levels was performed. Diets classified as low quality showed a higher intake of negative inputs, along with higher energy and fat intakes. This was reflected in body mass index (BMI) and food groups, indicating that women with the worst scores tended to choose tastier and less satiating foods. In conclusion, the DPA was developed and tested in a sample population. This tool can be easily implemented in digital nutrition platforms, contributing to real-time dietary follow-up of patients and progress monitoring, leading to further dietary adjustment.

Keywords: breast milk leptin; diet algorithm; diet profiling; diet quality; dietary advice; eHealth.

Figure 1

Components and scores of the DPA, a diet profiling algorithm to assess diet quality. The figure shows the sequential checkpoints considered by the DPA, displaying the points attributed and the range of application and the proposed final outputs. Abbreviations: DPA: diet profiling algorithm; CH: carbohydrate; Fb: fiber; E: energy; P/Bw: protein/body weight ratio; BMDP: balanced macronutrient distribution parameter.

Figure 2

Flow chart of the DPA, a diet profiling algorithm to assess diet quality. Nutritional information is processed by the algorithm, taking into consideration the constraints fixed. Dietary records are used as input data to ultimately result in nutritional guidance at different levels of personalization. Abbreviations: DPA (diet profiling algorithm); BMDP: balanced macronutrient distribution parameter.

Figure 3

A sample of a mobile interface screens showing the DPA implementation as colored plates. (A) DPA score and the progress of the user; (B) tailored recipe suggestions; (C) the actual food groups; (D) some recommended improvements in order to provide personalized advice.

Figure 4

General dietary characterization of the cohort: (A) macronutrient distribution. Percentage of total energy intake coming from protein, fat, CH, SFA, and total sugars; (B) food groups. Percentage of total energy coming from twelve different food groups. Data are mean ± SEM (n = 59). Abbreviations: SFA: saturated fatty acids; CH: carbohydrates.

Figure 5

Dietary characterization of the cohort through DPA. (A) DPA score distribution; (B) BMI of women classified by their DPA score; (C) points obtained in each DPA element and classified by DPA score; (D) total energy intake (kcal) associated with the DPA score; (E) percentage of total energy coming from macronutrients in each DPA score; (F) total fat/total carbohydrate (F/CH) ratio defining the BMDP in each DPA score. Data are mean ± SEM (n = 23, 18, and 18, respectively). Statistics: groups with different letters are significantly different (LSD post hoc one-way ANOVA, p-value ≤ 0.05). Abbreviations: DPA: diet profiling algorithm; BMI: body mass index; CH: carbohydrate; Fb: fiber; SFA: saturated fatty acids; P/Bw: protein/body weight; BMDP: balanced macronutrient distribution parameter.

Figure 6

Food group analysis by DPA score. (A) Percentage of total energy coming from food groups classified by DPA score; (B) comparison of the reference food groups based on the Mediterranean diet with the food groups shown in the cohort and classified by DPA score. Data are mean ± SEM DPA scores 1, 2, and 3 (n = 23, 18, and 18, respectively). Statistics: (A) groups with different letters are significantly different (LSD post hoc one-way ANOVA, p-value ≤ 0.05); if not significant, single comparisons between DPA = 1 and DPA = 2 or 3 were assessed by Student’s t-test (*, p-value ≤ 0.05); (B) S, differences between DPA scores and Mediterranean diet (one-way ANOVA, p-value ≤ 0.05). Abbreviations: DPA: diet profiling algorithm.

Figure 7

Correlation analysis between BMI and breast milk leptin concentration in the cohort classified by DPA score (n = 23, 18, and 18, respectively). (A) Spearman’s correlation data between BMI and leptin at the first, second, and third month of lactation by DPA score; (B) scatter plot and linear regression trend lines with data from the second month of lactation. Statistics: Spearman’s correlation test, * p-value ≤ 0.05 (** = p-value < 0.01). Abbreviations: BMI: body mass index; DPA: diet profiling algorithm; r: Spearman’s rank correlation coefficient; p: p-value.