This cross-sectional study conducted with first-time mothers at 26-36 weeks of pregnancy in southwestern Sydney indicates that there are low proportions of mothers who report meeting the recommended intake of vegetables or fruit during pregnancy. More worryingly, there are significant proportions of mothers who reported consuming more than 2 cups (500 ml) of soft drink a day and have fast-food/takeaway more than 2 times a week. Significant inverse associations were found between the intake of water and soft drinks, and between fruit consumption and eating fast-food/takeaway, processed meat or chips. Soft drink consumption was also found to be positively correlated with fast-food/takeaway consumption. In addition, this study found that a range of poor dietary behaviours was associated with various demographic factors, although no single factor was associated with all the dietary behaviours.
To our knowledge, this is the first time that a range of dietary behaviours, rather than nutrient intake [18–21], of first-time mothers in mid to late pregnancy, have been investigated in Australia. Our findings support the conclusion of a previous study, from Queensland, that few pregnant women report meeting the guidelines for recommended fruit and vegetables intake , even though another study from Japan suggests that pregnant women are aware of the need to adopt healthy behaviours .
The positive correlation between soft drink consumption and eating fast/takeaway food in this study is consistent with other studies linking soft drink intake and poor dietary quality, such as a higher intake of saturated fat and a lower intake of key nutrients [26, 27]. However, in contrast to a previous study suggesting that soft drink consumption was negatively related to milk intake among mothers , we found no association between milk intake and soft drink consumption. This difference may be explained by the use of different dietary measures in the two studies, as the measurement of milk intake in our study covered a broad milk consumption, including milk used in tea, or coffee, cow milk, soy milk, milk on cereal and flavoured milks and was based on a set of short questions assessing dietary behaviours .
Without doubt soft drink consumption is one of the key modifiable dietary behaviours that need to be targeted in the prevention of poor nutrition and obesity . The inverse relationship between water and soft drink consumption, and between fruit and other takeaway/fast food suggests that encouraging water and fruit consumption could potentially lead to reduced consumption of soft drink and other takeaway/fast foods.
Our findings are consistent with other research in the general population linking dietary behaviours to SES and other demographic factors [11–14]. While most studies suggest SES and education are the stronger predictors of nutrient intake [11, 12, 14, 30], we found that some socio-demographic characteristics were associated with particular dietary behaviour. For example, household income was associated with vegetable consumption, mother's education level was associated with daily fruit consumption and mother's country of birth was associated with soft drink consumption. In addition, mother's age and marital status were associated with fast-food/take away consumption. However, there was no single demographic factor associated with all the dietary behaviours we examined. This is an important consideration when developing dietary interventions, as tailored interventions to meet the needs of specific demographic sub-groups of pregnant women may be more successful. Specific demographic characteristics of the target population should be considered in the development of a dietary intervention.
In addition, the study found that, even using self-reported weight status, which is likely to give an underestimate , there was a high prevalence of overweight and obesity in women prior to their pregnancy - 38% of mothers were either overweight (25%) or obese (13%). The overall rate of combined overweight and obesity is similar to the self-reported rate of women in the same age range in NSW, but the rate of obesity is slightly higher than that of NSW women . Although weight status and its relationships with dietary behaviours and socio-demographic factors are not a focus of this study, it is understood that maternal weight status and dietary behaviours play an important role in the future health of infants. These in turn are also likely to be influenced by mothers' socio-demographic factors. With an increased need for early dietary interventions that establish healthy feeding practices and patterns of healthy eating among pregnant women in order to prevent early onset of childhood overweight and obesity [5, 24], a good understanding of these relationships is warranted.
Limitations of this study
There are a number of limitations of the study. First, the selection of study participants may limit the generalisability of the study findings, given that the participating mothers were able to communicate in English. Second, the locality of the study area, southwestern Sydney, is one of the most socially and economically disadvantaged area of metropolitan Sydney, Australia , which may give rise to a range of potential confounders for which we were unable to control. Third, we may not have included other specific dietary behaviour changes that are important for nutrition in pregnancy. Fourth, due to the cross-sectional nature of this study, no casual relationships should be asserted between the variables we examined. In addition, there are differences in dietary assessment tools between different studies, as a result of which it is difficult to make any comparisons across studies. The inconsistencies found in different studies suggest that more research, particularly with standardised measurement tools for assessing dietary behavours, should be conducted to ascertain the levels of dietary intake and relationships between various dietary behavours, and between dietary behaviours and socio-demographic factors.
Caution also needs to be exercised when interpreting these self-reported data on dietary behaviours. In terms of recommended serves, the questions we used in this study do give reasonable rank-order validity but the quantities reported tend to be different in terms of absolute amounts as suggested by the validation studies [16, 17]. We cannot rule out the possibility that the mothers in this study over-estimated the perceived size of a serve of vegetable or fruit intake. As a result, the findings from this study that 7% of mothers reported meeting the recommended vegetable consumption and 13% reported meeting the recommended fruit consumption could be over-estimated
Traditionally, dietary assessment methods have been known to under-report energy intake  and therefore under-report the foods which are more energy dense. It has also been observed that food frequency questionnaires (FFQs) have a tendency to over-report fruit and vegetable consumption (less energy dense foods). This has been observed among both long FFQs [33, 34] and short FFQs [16, 17]. It is therefore more likely that the mothers may have over-estimated the perceived size and amount consumed of a vegetable or fruit, rather than under-estimated, but we do not have specific data on this.