To our knowledge, this obesity prevention intervention is one of the first cluster-RCTs to have targeted first-time parents of infants within an existing health service infrastructure. The current study focused on lifestyle behaviors in first-time mothers post intervention and showed positive impact on two of the targeted dietary patterns, characterized mainly by high-fat or processed foods. The intervention did not influence other maternal health behaviors, having no impact on the consumption of healthier foods (as reflected in the “Fruits and vegetables” pattern), physical activity or TV viewing time.
First-time mothers represent an important target group for lifestyle interventions for several reasons. Firstly, young Australian women aged 18–30 y have gained weight faster than their older counterparts over the last decade . Ball and colleagues report that this trend can be accounted for in part by a decrease in physical activity, and an increase in fast-food consumption . Secondly, first-time mothers face additional barriers likely to influence the adoption and maintenance of healthy lifestyle behaviors, due to their new commitments towards child care, housework and shopping [33, 34]. This is a time of transition and greater vulnerability, with significant changes in their biological life, time availability, and both domestic and social situations . Given this context, first-time mothers are likely to be receptive and sensitive to advice and recommendations . In fact, high levels of recruitment and retention were observed in this study (86% of all women agreed to participate, and 9% only lost to follow-up), with results further suggesting that young mothers seem responsive to changing their own behaviors. Finally, the current study provides unique evidence regarding the potential to utilize young mother’s existing social networks as a vehicle through which to deliver health promotion interventions.
The current findings give new insights into the effectiveness of multidimensional interventions, which seek to promote parenting skills for positive lifestyle behaviors; increase knowledge; provide social support; reduce perceived barriers; and encourage feelings of self-efficacy . However, not all behaviors promoted through the InFANT were improved post-intervention. The absence of beneficial effects on the healthier dietary pattern (“Fruits and vegetables”) and on both physical activity and TV viewing outcomes may reflect the low “dose” of intervention regarding parenting health behaviors. Our results also suggest that young mothers may find modifying some aspects of food consumption easier to achieve than modifying their physical activity and TV behaviors. This latter hypothesis is supported by work undertaken by Ball and colleagues where young women (18–32 years) reported adoption of a range of healthy eating patterns was more achievable than the adoption of a set of physical activity behaviors . In a previous RCT specifically aimed at promoting physical activity in women with young children, Miller and colleagues have demonstrated that enhancing self-efficacy and partner support through community engagement could be a means to increase physical activity in mothers .
Promoting active play and limiting sedentary time for children were important foci of InFANT, however more time was devoted to discussions regarding children’s food, nutrition and feeding styles. It is possible that the emphasis on child feeding strategies and promoting a healthy diet for children may have influenced mother’s own eating patterns. Indeed, an underlying premise of the design of InFANT was that parents are likely to be receptive to information which will support them to achieve their aims of providing ‘the best’ for their child [18, 35]. A further explanation of our null findings for improved physical activity behaviors may relate to our measure of physical activity. In this study physical activity was assessed using a self-reported questionnaire and the precision of the resulting measurement is relatively low. It is possible that we lacked power to be able to show any significant effect of the intervention on this specific behavior. Objective measurements of physical activity, using accelerometers or pedometers for example, are recognized to provide relatively higher accuracy and precision.
With regards to the differential effect of the intervention on dietary patterns, the following hypotheses may be posited. Firstly, due to taste issues, it may be more demanding to increase consumption of healthy foods - such as those which characterize the “Fruits and vegetables” dietary patterns (i.e. raw vegetables, legumes, cooked vegetables, non-fried fish, nuts other than peanuts, and fruits) - than to reduce high-fat and processed foods. Secondly, lack of time for purchasing, storage and cooking, along with perceptions of relatively high costs of these foods, may be additional barriers for improving adherence to healthy dietary patterns during this particular stage of women’s lives [32, 37]. Finally, from a methodological point of view, the FFQ did not account comprehensively for mixed recipes. As a result, we may have missed part of the vegetable consumption for example, as vegetables are often ingredients of complex dishes. We cannot exclude that mothers receiving the intervention increased their use of vegetables in mixed dishes.
Strengths and limitations
The strengths of the current study include the cluster-RCT design and the high response and retention rates. Considering overall diet through the dietary pattern approach is an additional advantage, as compared to the traditional single-food group approach, which does not account for colinearity among all dietary components . Despite subjective choices inherent to factor analysis, Newby and colleagues reported in their review that reproducibility has been observed between most studies which have identified patterns in adults . This consistency over national and international studies was confirmed for the InFANT study, as previously described . Some limitations of the study need to be acknowledged. First, we acknowledge that the differences between the mean change scores of each treatment arm observed for the "High-energy snack and processed foods" and the "High-fat foods" patterns are small, although statistically significant. This is likely to come from the low dose prevention intervention, as well as the challenge it is to influence maternal behaviors. Nonetheless, small changes in multiple health-related behaviors, if sustained (and thus accumulated) over the life course, are likely to favorably impact on maternal energy balance. Second, although all socio-economic positions were represented in this study, university educated women are over represented, with 54.2% reporting high education level (university degree or higher). In addition, participants in this study were older than the average age of first-time mothers in Victoria (32.3 years compared to 29.1 years) . These characteristics may have implications for generalisibility. However, they may be partially explained by the inclusion of only urban Melbourne residents in the trial, who are likely to differ from the broader Victorian population. Another limitation of the study is the number of mothers excluded from the analyses at post-intervention (34%), mainly due to missing data regarding the main outcomes (123 mothers out of the total 185 excluded from the analyses). Mothers excluded from the analyses did not differ significantly from the included mothers in terms of socio-demographic characteristics. However, less statistical power was available to assess the effects of the intervention.