Norway, as other Nordic countries, has a long-standing tradition of extensive and freely available health care for children under school age at the child health centre. Virtually all parents with small children use these child health centres, having regular encounters with public health nurses. This service aims at promoting health and preventing disease in the population. However, when comes to content and performance of conversation about food and feeding practices there are no guidelines regarding how to adapt the information to the individual child and to his or her family after infancy and up to two years of age.
The overall purpose of this study is to promote optimal nutrition for the child based on the public health nurse's family-focused conversations at the child health centre.
The design of the study is a cluster randomized controlled trial. Statistics Norway has on the basis of a systematic match drawn samples of five pairs of municipalities in Norway. It is random which municipality in each match is extracted respectively to intervention or control municipality.
Public health nurses on child health centres within intervention municipalities will use a communication tool about food and feeding practices in their encounters with participating children and parents when the child is 10-, 12- and 15-18 months old. In the control municipalities corresponding encounters will be as usual, without use of the actual communication tool. The parents will answer a food frequency questionnaire on behalf of their child when the child is 10 months old (baseline) and at the age of two years (final measurement).
Sample size consideration: The aim of the study is to reveal change in vegetable intake. The investigators keep level of statistical significance at the customary 5 % and statistical power of 80 % (beta equal to 20 %), the investigators would need 176 children in each group to reveal the change described above as statistically significant.
The investigators expect some drop out and given that their estimate might be slightly higher, the investigators want to enroll 300 children in each group to make sure our study is sufficiently powered.
Plan for missing data: Concerning missing data the investigators will use mixed models for repeated measure, imputation of missing data will therefore not be necessary. All data will be analysed according to Intention to treat (ITT) analysis.
Analytical principles: Unadjusted (or crude) differences between intervention and control group will be assessed with t-tests (for continuous variables) and Chi-square tests (when we test for association between pairs of categorical data).
In case of continuous outcome variables, the investigators will model the differences between groups (scores from the questionnaires) using uni- and multiple linear regression methods.
All regression models will be conditioned on district variable and not individual level because the randomization was performed on district variable level.
The investigators treat this study as an exploratory analysis, however, they will fit several statistical models and perform several tests therefore a correction for multiple testing will be performed. P-values \< 0,01 will be considered statistically significant.
