The recommendations of STROBE-nut (Strengthening the Reporting of Observational Studies in Epidemiology—Nutritional Epidemiology) were followed (see Additional File 1).
Study design
Food4toddlers is originally a randomized controlled trial evaluating the effect of a digital dietary intervention targeting parents of 1-year olds (22). The current paper reports on cross-sectional baseline data from this study.
Sample and procedure
In 2017 we recruited parents of toddlers. They were recruited through social media posts on Facebook. Parents voluntarily notified that they wanted to participate after reading written information about the study. In total 404 participants volunteered to participate initially, however of these, only 298 participants filled in the baseline questionnaire. Data from these 298 participants were included in the present study. The study was approved by the Norwegian Centre for Research Data, and by the Faculty ethics committee and has been conducted in line with the Helsinki Declaration of 1985, revised 2008.
Instruments
The baseline questionnaire included questions of parental and child characteristics like age and educational level, further there were questions regarding use of mobile devices, food frequency questions of child diet and parental diet, frequency of family meals, feeding practices and other intervention specific questions. In the current study the variables use of mobile device during mealtime (independent variables) and feeding practices and shared family meal frequency (dependent variables) are presented. In addition to parent and child characteristics.
The use of mobile phone during mealtime, was reported like this: To what extent do you agree with the following statement: I often check my phone during meals. The participants were given the following response alternatives: Disagree, slightly disagree, neither agree or disagree, slightly agree and agree. We dichotomized this variable with the intention to differentiate between those who do not use mobile phones during meals at all and those who do. The participants who filled in “disagree” were defined as the “No phone use during meal” group, while the participants who filled in “slightly disagree, neither agree or disagree, slightly agree and agree” were defined as the “phone use during meal” group.
To assess feeding practices, we used the well-known and validated, Comprehensive Feeding Practices Questionnaire (CFPQ) (23). This questionnaire covers 12 dimensions of parental feeding practices. Because the participants of this study were parents of infants, following nine items were considered relevant: Child control (The child is allowed to control his/her own eating behaviours), Emotion regulation (The parents use food to regulate the child’s emotional states), Balance and variety (Parents encourage well-balanced food intake), Environment (Availability of healthy food at home), Food as reward (Use of food to reward child behavior), Modeling (Parents demonstrate healthy eating behavior), Monitoring (Parents keep track of child’s intake of less healthy food), Pressure (Used by parents to make child eat more), Restriction for health and weight control (Parents control the child’s food intake with the purpose of limiting less healthy foods and sweets and to decrease or maintain the child’s weight). The reliability of the ten feeding practices dimensions was evaluated using Cronbach alpha. All scores except Child control and Food as a reward (α = 0.3) showed acceptable reliability (α = 0.5–0.8) (detailed data not shown). The CFPQ items were previously translated from English into Norwegian in another project and a random sample of 10 items were back-translated into English (24). The quality of the translation was considered very good and this translation was therefore used in this project.
To assess the frequency of shared family meals, following questions were posed: How often does your child have the following meals together with their family? Herein, meals with family included parents that were married, living in cohabitant families, single or separated / divorced. The participants could choose from the following response alternatives: Never, 1, 2, 3, 4, 5, 6 and 7 times per week for breakfast, lunch, and dinner, respectively. Later, the weekly food frequency was divided into having every specific family meal, every day or not.
Parental age was assessed in years, child age in months. Participants reported their own and the other parent’s highest level of education (primary school or less, primary schools plus one year of further education, high school, vocational school, upper secondary school or less, college/university (≤ 4 years), college/university (> 4 years), other, don’t know). These responses were dichotomized into having no university/college education or having university college education and used as a proxy for socio-economic status (SES). Ethnicity of the parents were assessed as native if they were born in Norway or non-native if they were born elsewhere.
Statistical analysis
The statistical analyses were conducted using IBM SPSS Statistics for Windows, Version 25.0 (IBM Corp., Armonk, NY). Demographic data were described using descriptive measures. Continuous variables were described with mean and standard deviation, and categorical variables with frequencies and percentages. For the dependent study variable feeding practice, 9 out of 10 sub-scales had skewness values of ± 1.3 and kurtosis values of ± 1.6, which indicated that these variables are approximately normally distributed. Thus, linear regression analyses were conducted between the dependent variables (feeding practices) on the independent dichotomy variable (phone use / no phone use), and then controlled for SES, ethnicity and age using multiple regression analysis. Binary logistic regression analyses were conducted to calculate odds ratios on the relationship between the dependent dichotomy variable shared family meals and the independent dichotomy variable (phone use / no phone use) controlling for SES, ethnicity and age. Due to the homogeneity of the sample in terms of gender (98% were mothers), adjusting for this variable was not considered relevant. P-values < 0.05 were considered statistically significant and all tests were two-sided.