Closing the dietary fibre gap - developing a novel dietary fibre screening tool (SCREEN-IT) for the UK population: validity, reproducibility and usability insights · Norton V, Lovegrove JA, Weech M et al.

Results

Four participants in the validity assessment reported daily energy intake from eNutri-FFQ which was considered implausible in accordance with non-individualised cut-offs advised by Willett (female: 500–3500 kcal/d and male: 800–4000 kcal/d) [45]. Three participants did not complete both dietary assessments within the agreed timeframe; hence, such data was excluded from subsequent analyses (Figure S3). Seventy participants (55.1 ± 18.7 years) were included in the validity analyses as summarised in Table 1. Most participants were within the healthy body mass index (BMI) range, living with a partner, had a university-based education and were either employed or retired. Overall, 155 participants (51.7 ± 18.2 years) were included in the reproducibility and usability assessment (six and four participants did not complete both assessments or were outside agreed timeframe, respectively) consisted of balanced numbers across the three age groups with similar demographics to the validity stage (Table 1).

*BMI (body mass index) calculated from self-reported body weight and height.

SCREEN-IT validity

Mean estimated dietary fibre intake was significantly lower (3.9 g/d) using SCREEN-IT compared with eNutri-FFQ (Fig. 3; p < 0.0001). The agreement was classified as “acceptable to good” despite differences between methods for ICC, percentage difference, quartile agreement and weighted kappa (Table 2). The exact and extreme quartile agreement were identified as 52.9% and 10.0% of the data, respectively. The Bland-Altman plot (Fig. 4) highlighted that 5.7% of the data points were outside the limits of agreement (-12.71 to 4.88 g/d) and mean bias (-3.91 g/d) demonstrated higher dietary fibre estimates for eNutri-FFQ than SCREEN-IT. SCREEN-IT was significantly quicker to complete (by 26-min); however, both tools had acceptable usability with higher scores (6.6 points) cited for SCREEN-IT compared with eNutri-FFQ (Fig. 3; p < 0.05).

SCREEN-IT reproducibility and usability

Overall, there were no significant differences in mean dietary fibre intake (between SCREEN-IT 1 and SCREEN-IT 2 [17.2 ± 0.5 vs. 17.2 ± 0.6 g/d]); 3.2% and 6.5% meeting the 30 g/d recommendation, respectively. It was demonstrated that SCREEN-IT was reproducible with either “acceptable or good agreement” for ICC, percentage difference, quartile agreement and weighted kappa (Table 2). Fig 4 illustrated that 7.1% of the data points were outside the limits of agreement (-8.39 to 8.31 g/d); however, the mean bias was close to zero (-0.04 g/d). There was a tendency for SCREEN-IT to be completed marginally faster the second time (3.4 vs. 3.8 min; p = 0.06).

The youngest cohort had higher mean dietary fibre intakes compared with older cohorts post SCREEN-IT 2 (p = 0.04); however, no age-related differences were reported for SCREEN-IT 1 (Fig. 5). Age-derived reproducibility was either “acceptable or good in agreement” for ICC, percentage difference, quartile agreement and weighted kappa (apart from 18 to 40 years quartile agreement was perceived as poor) (Table S2). Between 5.8 and 7.7% data points were outside the limits of agreement for each age-group but mean bias was close to zero (Fig. 5). Older adults spent more time completing SCREEN-IT 1 compared with younger cohorts (3.4 vs. 4.7 min; p = 0.002). This trend continued at SCREEN-IT 2; however, there was no significant difference (p = 0.10). The SUS values indicated acceptable usability for all age-groups (84.9/100); however, the middle-aged cohort (41–64 years) perceived the tool more usable compared with the other age groups (Fig. 5; p < 0.0001).

SCREEN-IT was completed using different sized devices: smartphones (small; 31.6%), tablets (medium; 7.7%) and laptops/desktop computers (large: 60.6%) with no significant differences in perceived tool usability between devices (p = 0.90).

The feedback on SCREEN-IT was very positive (81.4% of comments) with key themes emerging such as easy-to-use “all well laid out, very easy to understand, examples of foods are very nice, means I know how to answer”, functional “I think it is the best tool that I have ever used. I have learnt a lot from it so far and I have liked the idea of changing my meal diet”, thought-provoking “it’s a very useful tool for making you actually count how much fruit, veg etc. you are actually eating vs. how much you think you are eating”, enjoyable “I liked the tool. It’s easy to use, provides good explanations and was fun (never thought I would say a questionnaire could be fun!!!)” and concise “quick and easy. Two-week time period easier to recall than one month”. The negative comments (less than 20% of comments) were centred on two areas: portion size estimation “the portion samples aren’t easy to translate to what goes on the plate each day (it’s not weighed out before each serving! )” and more agile approach “I sometimes find the portion comparisons a bit limiting for me”.