Research questions:
1. What are the effects of the programme on children's dietary habits, physical activity, sedentary behaviour and body weight compared to standard care?
2. How is the programme implemented with regard to dose, fidelity, feasibility and acceptability?
3. What are the mediators of the effects of the programme?
Recruitment of schools, participants and setting:
Key persons in municipalities in the Stockholm region received a one-page invitation letter explaining the aim and content of the intervention and were invited to contact the project leader for more information. Primary schools were eligible to participate if the proportion of parents with university education was less than 50 percent compared to 57 percent at national level. In total 18 schools with 18 school nurses agreed to participate but one school dropped out before recruitment of parents had started. All parents with a child in pre-school class (5-7 years old) were eligible to participate. Informed, written consent was obtained from all participating parents (350 children and families).
Study design:
The study will be conducted as a cluster-randomised parallel trial with randomisation at school level, with a wait-list control group. Baseline data was collected in September-October 2017 (T0). Data will again be collected 6 months after the baseline at the end of the intervention in April-May 2018 (T1), and at follow-up in April-May 2019 (18 months post baseline, T2). Mediators will also be measured on one occasion during the intervention, in February 2018 (TM). After the T2 measurement, the control schools will be offered the intervention, including MI-training of the school nurses.
The intervention:
The programme builds on Social Cognitive Theory and consists of four components:
1. Health information: A 12-page brochure based on Swedish guidelines for diet and physical activity contains facts and advice concerning parenting in relation to healthy food and family meal times, sweets, snacks, ice-cream, soft drinks, fruit and vegetables, physical activity, sedentary behaviour, screen time, sleep, and a theme regarding cooperation between parents. The text is simple and short with many illustrations and available in Swedish, English, and Arabic. Parents will receive a printed version of the brochure from the school before the first MI session and it will also be available on the project website throughout the intervention.
2. The MI session will be scheduled by the school nurse as part of the ordinary health visit, where the child's height and weight is measured, and which also includes a discussion about the child´s overall health. The aim of the MI session is to increase positive parenting and interplay regarding the child's dietary and physical activity behaviour in the home environment. Parents in the intervention group will be offered one 20 to 30 minute face-to-face MI session by the trained school nurse and another session, depending on need, as determined by the nurse. During the session (conducted without the presence of the child), the parents choose a specific behaviour that they wish to change or maintain. Parents are supported in identifying a target behaviour using an agenda-setting tool and in the exploration of personal values and the advantages of change or maintenance, and assisted in goal-setting if applicable. School nurses will undergo MI training. MI competence will be measured prior to start, post training and at the end of the intervention.
3. Nine manual-based classroom sessions of approximately 30 minutes duration will be delivered by the teacher to the children. Various pedagogic materials are provided for the sessions including a workbook for children to be completed with the parents. Teachers will receive video-recorded instructions of how to use the teacher's manual ahead of the start of the intervention.
4. Parents complete a web-based test of their own risk of developing type 2 diabetes in order to make them become conscious of a healthy lifestyle for the whole family. The test contains 8 questions and yields a score between 0 and 26 points. Based on the result, the parent is classified as being at low, somewhat elevated, medium, high or very high risk. In case of medium or higher risk (≥15 points) the parent will be advised by automatic feed-back to consult the local primary health care centre. General information on health behaviours that minimise the risk for diabetes is provided to all parents who take the test, regardless of score.
Control group:
Children in the control schools will receive treatment as usual according to prevailing guidelines for school health services and in addition parents receive the 12-page health information brochure (component 1 of the intervention). Usual treatment means a visit of the child accompanied by parents to the school nurse soon after the child begins school for a discussion regarding the school situation, the child's overall well-being, relations with friends, possible health and learning problems, and diet and physical activity. The curriculum in pre-school class prescribes that pupils should be provided with opportunities to experience different forms of physical activity, and to develop an understanding of how health is affected and can be promoted in various ways.
Outcome evaluation:
All questionnaires are filled in via the project website (www.enfriskskolstartplus.se) with personal accounts for each family. Paper versions will also be available. In order to increase the response rate to the questionnaires, weekly reminders are sent out via e-mail and text messages. All outcomes are assessed as the difference between the intervention and the control group directly after the end of intervention at 6-months post baseline, and at follow-up 18 months post baseline, adjusted for baseline values.
Primary outcomes:
The primary outcome is the intake at home of a composite score of indicator foods of importance for energy-balance and health. Intake will be assessed using food frequency questionnaire, the Children's Eating Habits Questionnaire which measures intake retrospectively in the home during the preceding month. A new smartphone-based method will also be used to estimate the intake of indicator foods eaten at home (volume) prospectively through standardised photos taken by parents over two weekdays and one weekend day. A relative validation study will be conducted in parallel during baseline, where both methods will be compared to multiple 24-hour recalls. The intention is to use the photo-based method for the primary outcome in the intervention study, if shown to be valid.
Secondary outcomes:
Secondary outcomes are physical activity and time spent sedentary, which will be measured objectively using accelerometers worn on a belt at the right hip (GT3X+, Actigraph, LCC, Pensacola, USA) over 7 consecutive days. Children with at least 500 min of activity registration per day for a minimum of three days, including at least one weekend day, will be included in the analyses. Body composition (height, weight and waist circumference) will be measured using standardised procedures using SECA instruments. Waist circumference will be measured over the umbilicus, with the child standing upright with arms alongside the body and after exhalation. BMI status will be defined according to the International Obesity Task Force. BMI standard deviation score (BMI sds) will be calculated according to the International Obesity Task Force references.
Covariates:
Parental educational level and occupation is self-reported and a composite score will be used to define socio-economic position consisting of the highest level of education attained, and the occupation (according to Statistics Sweden) by either of the parents. A dichotomised variable indicating low and high socio-economic position will be created. In addition, child sex will be used.
Process evaluation:
The process evaluation will include intervention dose, fidelity, feasibility, acceptability and a description of the context.
The dose will be estimated quantitatively in the following way. Parents will be asked in a non-judgmental way by the school nurse if they have read the brochure. The number of MI sessions per family will be noted. Compliance with the teaching sessions and workbook completion will be monitored by teachers in a logbook and whether they made any adaptations. The result of the diabetes test will be available from the website.
All MI sessions will be recorded and a random selection of 10-20 percent will be coded for MI competence/fidelity using a standardised protocol by reliable coders. The MI competence of all nurses in the intervention and control group will be assessed before the first MI training in May 2017 by recording them having a standardised conversation with a trained actor posing as a parent. This will be repeated after MI training in August/September 2017, and again after the end of intervention in May 2018.
After the end of the intervention, semi-structured interviews with parents, school nurses and head teachers will be performed to explore their views on the acceptability and feasibility of the intervention. In addition, focus groups will be conducted with the children to explore their perceptions of the activities in class and the home assignments.
The hypothetical mediators parental self-efficacy and observational learning/role modelling will be measured by questionnaire in both mothers and fathers. The instrument used assesses parental self-efficacy for influencing physical activity during the weekend, intake of vegetables, sweets and chocolates, and sugar-sweetened beverages, each in three challenging situations. Parental feeding practices, including modelling, will be measured using the Comprehensive Feeding Practices Questionnaire covering both positive and negative practices. The questionnaire will assess the parental feeding practices of involvement, environment, food restriction for weight control, restriction for health reasons, encouragement of balance, pressure to eat, monitoring, emotion regulation, food as reward, and modelling. For assessment of parenting practices regarding children's physical activity, 25 items will be used from an item bank currently being developed for this purpose. The items cover the parenting practices modelling, pressure, encouragement, guided choice, involvement/praise, co-participation, facilitation, monitoring, and restriction.
Data management:
All participants will receive a code and data will be depersonalised. All data collected through the project website and otherwise will be stored electronically in password protected folders on a secure data server at Karolinska Institutet. Measures to prevent loss of data will be taken through the use of systematic back-up routines throughout data collection and by data storage on servers with complete data back-up on a daily basis. Access to data will be restricted to the research personnel working directly with data entry or analyses.
Data analysis:
Differences in continuous demographic variables at baseline between intervention and control group will tested using independent t-tests for normally distributed data. Non-normally distributed variables will be analysed by non-parametric tests. For categorical variables, differences will be analysed by chi-square test. Effectiveness with regard to primary and secondary outcomes will be analysed by mixed-effect regression analyses with two levels (individual and school) with data at T1 or T2 as outcomes, according to the principle of intention to treat. First, a crude model will be tested for all outcomes at T1 or T2 with group as the predictor and adjusted for baseline values of the relevant outcome. Interaction between group and child sex or group and parental SEP will be tested and analyses stratified if significant interaction terms are found. All children with measurements at T0 will be included and missing data will be handled with multiple imputation. The effect that any missing data might have on results will be assessed by performing sensitivity analyses of augmented data sets. Analysis will also be performed per protocol including only those children whose parents have attended at least one MI session.
All statistical analysis will be performed blinded to group allocation. Interviews and focus groups will be transcribed verbatim and analysed using qualitative content analysis.
Sample size calculation:
Dietary intake of unhealthy foods, unhealthy drinks and healthy foods is the primary outcome. In the previous study, the intake of unhealthy foods was 1.7 ± 2.0 portions/day. In this study, the aim is to achieve a difference in intake of unhealthy foods of 0.8 portions/day between the intervention and the control group. With a power of 80% and significance level of 5% and an intra-class correlation coefficient of 2% (clustering at school level), at least 17 schools with 15 children in each school need to be recruited, 255 children in total.
