Outcome measurements
Outcome variables are described in detail in Table
1 and along with covariates are to be measured by mother-completed questionnaires at baseline (Time 1, age 4-7 months) and six months after the completion of each 12-week intervention module at age 13-16 months (Time 2) and 22-25 months (Time 3). Questionnaires are to be completed at home and brought to assessment clinics where maternal and infant weight and height/length will be measured. Child intake will be measured at Times 2 and 3 using a telephone 24-hour recall conducted by a dietitian and a 2-day (one week and one weekend day) food record completed by the mother. Randomisation should optimise chances of no group difference in Time 1 infant intake which, along with maternal intake, will not be collected due to the resource and participant burden implications. Assessors will be trained and will not be involved in intervention delivery. Where a number of assessment tools are not suitable across all age groups, we have selected for age appropriateness within constructs, rather than consistency across time points.
Table 1
Outcome measures for the study
Food intake records H1 | 3 non-consecutive days (including weekend day), using 2 × 24 h food records + telephone 24 h-recall. Standard protocol (including estimation of breast milk intake and standardised visual aids for serve size estimation) will match FITS [ 48] and other [ 49] studies; well accepted in pilot study. | | ✓ | ✓ |
Food preference H2 | The Wardle tool [ 37] adapted to Australian target foods. Mothers rate on 5-pt scale from 'likes a lot' - 'dislikes a lot' with option for 'hasn't tried it' | ✓ | ✓ | ✓ |
Feeding behaviour H3 | Children's Eating Behaviour Questionnaire [ 50]. Validated 35-item parent report of satiety responsiveness, fussiness, food responsiveness, enjoyment, emotional over/under eating. | | ✓ | ✓ |
Weight & growth | Recumbent length and weight. Weight, length and weight-for-length z-scores calculated using CDC EpiInfo (version 3.3.2). | ✓ | ✓ | ✓ |
Maternal
| | | | |
Feeding style and practices H3 | The Infant Feeding Questionnaire [ 36]. 20-item - under/over-eating, hunger, infant cues, scheduling, use of food to calm. | ✓ | ✓ | |
| Child Feeding Questionnaire [ 43] 28-item - 2-11 yrs. - feeding attitudes, practices, perceptions/concerns regarding weight. | | | ✓ |
Parenting skills H4 | Four brief scales from LSAC measuring warmth, irritability, consistency and overprotection (24 items)[ 41] | ✓ | ✓ | ✓ |
BMI | Height and weight using standard procedures | ✓ | ✓ | ✓ |
Covariates
A comprehensive range of sociodemographic, maternal and infant covariates will be collected. At the first contact (ie at birth) data collected on the larger eligible sample will include maternal age, education, ethnicity, marital status, household composition, self-reported pre-pregnancy weight status, perceived level of support with parenting, lifestyle (smoking and alcohol intake) and health problems (diabetes, preeclampsia) during pregnancy, birth weight and breast feeding intentions. Baseline assessments (Time 1) of those consenting to full enrolment will include maternal mental health, current breast/bottle/solids feeding (also collected from non-consenters), family income, parental employment, child care use, child health issues, maternal diet, activity, smoking and alcohol intake. Data to be collected at subsequent assessments are maternal lifestyle behaviours (e.g. activity and fruit and vegetable intake) and any demographic data that are likely to change (including marital status and birth of subsequent children).
Maternal covariates will be assessed at Times 1 and 3. These will include maternal body mass index and baseline infant feeding practices, attitudes and beliefs assessed using the 20-item Infant Feeding Questionnaire [
36]. Maternal food preferences influence foods made available to the child and hence child food preferences, and will be described using the Wardle tool [
37] (Table
1) with an additional option 'like but don't usually eat' (based on our pilot study feedback that this addition was warranted). Maternal concerns regarding their own weight and eating-related issues influence child feeding practices [
38]. Maternal restrained eating will be determined using the Restraint Scale, a validated, widely used 21-item scale [
39]. The 5-item Weight Concern Scale [
40] will assess maternal perceptions regarding their own weight gain, body weight and shape.
Child covariates assessed will include detailed data on early infant feeding (breast, bottle, type of formula, exclusive breast feeding, use of other fluids, age of introduction of solids), early growth rate from birth to baseline, temperament and child care experience.
To enable comparisons with normative Australian data, NOURISH will use demographic, parenting, child temperament and child care measures that were developed and validated for the nationally-representative Longitudinal Study of Australian Children (LSAC) [
41].