Study design and subjects have been presented [25, 26]. Briefly, overall aim of the BRA-study was to improve vegetable intake (primary outcome) among preschool children through changing the food environment and food-related practices in kindergarten and home (secondary outcomes). Child consumption of vegetables and vegetables served in the kindergarten were assessed with regards to frequency, variety and amount of vegetables served and consumed.
Children born in 2010-2011, attending public or private kindergartens in the counties of Vestfold and Buskerud, Norway were included. All regular kindergartens (n = 479) in the counties were invited by letter followed-up by a phone call to inform about the study and to motivate for participation. Seventy three kindergartens participated (response rate 15.2%) and parental consent was obtained for 633 children (response rate 38.8%). This study was conducted according to the guidelines laid down in the Declaration of Helsinki, and the Norwegian Center for Research Data approved all procedures involving human subjects.
The intervention and its components have been reported [27, 28]. Briefly, the intervention components to the kindergartens consisted of a one-day inspirational course (kitchen practice and theory) in addition to resources like aprons, a vegetable memory game, a hand blender, booklets, posters and brochures. The intervention components were related to the four determinants: availability, accessibility, encouragement and role modelling. The study focused on the potential for improving child vegetable consumption and vegetable serving in the kindergartens, by changing these determinants, as implemented in the intervention components.
Data collection
There were 313 (49%) children in the intervention group, representing 37 kindergartens and 70 departments, while corresponding numbers for the control group were 320 children (51%), representing 36 kindergartens and 65 departments.
The research team conducted direct observations to assess children’s vegetable intake at two meals (lunch and an afternoon snack meal) in one day in the kindergarten. A 5-day weighted vegetable diary was completed by the kindergarten staff to assess amount of vegetables served in the kindergarten. Data were collected at baseline (spring 2015) and at follow-up 1 (spring 2016) and a description of methods are provided elsewhere [25, 26]. By completing all data collections, kindergartens received a gift card of 2000 NOK.
In January 2016, process evaluation was conducted in the intervention group, both in kindergartens and at home. Due to a lack of significant effects upon children’s vegetable consumption in the home setting [28, 29], this paper reports on process evaluation from the kindergartens.
Process evaluation consisted of a paper-based questionnaire including several aspects assessing to what degree the intervention components were perceived useful or not for changing usual vegetable practices in the kindergarten department. The present study includes responses to the question “to what degree have the following components been important in getting started with the changes you have been working on?” (i.e. during the intervention period). Perceived usefulness of the components were assessed by four response alternatives: “not at all”, “to a small degree”, “to some degree” and “to a large degree”. Nine intervention components, described in Table 1, were combined into three groups of components: “posters”, “supplementary material” and “one-day inspirational course”.
Table 1
Kindergarten staff’s perceived usefulness of the given intervention materials in the BRA-study (n = 48)a
|
|
Not at all
|
To a small degree
|
To some
degree
|
To a large degree
|
“Posters”
|
A large poster (70 x 100 cm) with photos of vegetables
|
1 (2.1%)
|
7 (14.6%)
|
17 (35.4%)
|
23 (47.9%)
|
2 small posters (A4) on amounts of vegetables
|
2 (4.2%)
|
10 (20.8%)
|
20 (41.7%)
|
16 (33.3%)
|
4 small posters (A4) with ideas of “what to do” for each of the four determinantsb
|
4 (8.3%)
|
13 (27.1%)
|
19 (39.6%)
|
12 (25.0%)
|
“Supplementary material”
|
Four child-aprons
|
5 (10.4%)
|
11 (22.9%)
|
13 (27.1%)
|
19 (39.6%)
|
One fruit and vegetable memory game
|
2 (4.2%)
|
12 (25.0%)
|
18 (37.5%)
|
16 (33.3%)
|
One hand blender
|
5 (10.4%)
|
11 (22.9%)
|
16 (33.3%)
|
16 (33.3%)
|
“One-day inspirational course”
|
Practical training in the kitchen
|
2 (4.2%)
|
7 (14.6%)
|
24 (50.0%)
|
15 (31.3%)
|
A theoretical session
|
4 (8.3%)
|
8 (16.7%)
|
25 (52.1%)
|
11 (22.9%)
|
Make action plans
|
4 (8.3%)
|
8 (16.7%)
|
27 (56.3%)
|
9 (18.8%)
|
akindergarten departments
bavailability, accessibility, role model and encouragement
|
Data analysis
Children’s observed vegetable intake (grams/day) and served vegetables in the kindergarten (grams/day) were dependent variables in separate mixed effects models. For children’s observed vegetable intake, kindergarten, department and participant were used as random effects, while for vegetables served in the kindergarten, kindergarten and department were used as random effects. Time (baseline and follow-up 1) and perceived usefulness (“not at all/to a small degree”, “to some degree” and “to a large degree”) were used as fixed effects. The reference category for perceived usefulness was “not at all/to a small degree”. To avoid small subgroups, response alternatives “not at all” and “to a small degree” were merged into one category in analysis. We defined the effect of a component, as the difference in change of vegetable intake from baseline to follow-up 1 between the response alternative “to some degree”/“to a large degree” and the reference category “not at all/to a small degree”. The perceived usefulness factor thus measured the effect of being in one of the two other categories relative to being in “not at all/to a small degree”. A potential difference in effects of the perceived usefulness categories between the two time points may be attributed to the intervention. In particular it was of interest to assess if the intervention provided the same change in effect of “to some degree” and “to a large degree” levels. This was accommodated by an interaction effect between time and perceived usefulness. Test for significance was done by likelihood ratio tests. Corresponding models were considered for each of the two outcomes and each of the three groups of intervention components. Models including children’s vegetable intake were adjusted for covariates: maternal education, child gender and child year of birth. Participants were included in analyses if they had data on baseline and/or follow-up 1 and if they had data on all covariates. All analyses were conducted using Stata Statistical Software: Release 15. College Station, TX: StataCorp LLC. P values less than 0.05 were considered statistically significant.