- Open Access
- Open Peer Review
Weekday snacking prevalence, frequency, and energy contribution have increased while foods consumed during snacking have shifted among Australian children and adolescents: 1995, 2007 and 2011–12 National Nutrition Surveys
Nutrition Journalvolume 16, Article number: 65 (2017)
There are limited data on the evolution of eating habits, including snacking, in Australia. This study aimed to understand snacking trends among Australian children over three previous National Nutrition Surveys.
Data were analysed from a single weekday 24-h recall in the National Nutrition Surveys 1995, 2007, 2011–12 among children 2-16y (n = 8258). A snacking occasion was defined as an eating occasion that occurred between meals based on time of day.
The percentage of children snacking increased over time (92.5 ± 0.5(SE) % in 1995, 98.1 ± 0.3% in 2007, and 95.8 ± 0.4% in 2011–12) (P < 0.001), particularly among those having four or more snacking occasions (7.1 ± 0.5% in 1995, 17.9 ± 0.6% in 2007, and 18.5 ± 0.8% in 2011–2) (P < 0.001). The mean number of snacking occasions increased from 2.0 ± 0.0 in 1995, to 2.5 ± 0.0 in 2007 and 2011–12 (P < 0.001). The energy contribution from snacking increased from 24.1 ± 0.3% in 1995 to 27.7 ± 0.3% in 2007 and 30.5 ± 0.4% in 2011–12 (P < 0.001), while the energy from discretionary food during snacking decreased from 56.5 ± 0.7% in 1995 to 47.3 ± 0.5% in 2007 and 47.9 ± 0.7% in 2011–12 (P < 0.001). There were differences in the top foods consumed during snacking: non-alcoholic beverages were prominent contributors in 1995 but not in 2007 or 2011, and pome fruit was the second top energy contributor during snacking in 2007 and 2011 but only fourth in 1995.
Snacking is a prominent dietary pattern that has increased over time in frequency and energy contribution. Foods and beverages consumed during snacking occasions include a mix of core foods and discretionary foods, and while the contribution of discretionary foods has decreased, there is still an opportunity to encourage consumption of more nutrient dense foods during snacking.
The prevalence of overweight and obesity among children and adolescents has risen across the globe [1, 2]. In Australia, in 2014–15 27.4% of children aged 5–17 years were overweight or obese , up from 25.7% in 2011–12, 24.7% in 2007, and 20.9% in 1995 . Rates of overweight and obesity have concurrently increased with major shifts in meal patterns, food choices and location of food consumption [5,6,7,8]. In the United States (US), snacking has become more prevalent and all snacking combined, contribute as much to total energy intake as a meal [9,10,11,12]. In Northern Ireland and Great Britain, both energy and portion size of snacks increased, but not snacking frequency . Snacking is a prevalent eating behaviour in both Mexico [14, 15] and Brazil  and has also increased among countries where it is less traditional including China [17, 18] and Spain .
There is a long-held concern that increased snacking contributes to obesity , but a number of recent reviews [21,22,23] and a meta-analysis of cross sectional studies and a case-control study,  found that eating frequency and snacking were not associated with obesity, whereas a recent cross-sectional study found that only after adjustment for misreporting, snacking frequency was associated with BMI percentile and waist circumference in children, but not in adolescents . However, the concern is that snacking contributes a significant proportion of total energy to children’s diets [10, 11, 14, 16, 18] and increased eating frequency is associated with greater total energy intake [9, 15, 26,27,28], which may contribute to the childhood obesity epidemic. Further, the type and hence quality of the ‘snack’ choice may influence total energy intake and the association between snacking and obesity .
Limited studies report on trends in meal and snacking patterns among children. It is necessary to capture changes in eating behaviour for successful, culturally relevant recommendations and intervention programs, as unhealthy eating patterns during childhood and adolescence can have implications in the development of chronic disease [29, 30] into adulthood. In Australia, dietary guidelines are food-based rather than meal-based. Understanding meal patterns can provide insight for the food-based approach in recommendations with regards to specific timing of food consumption and the distribution of food groups across the day. Given the shifts in dietary patterns worldwide and the rise in childhood obesity, there is a need to evaluate changes in eating patterns among Australian children. The objective of this study is to investigate the changes in snacking patterns in terms of prevalence, energy and nutrient contribution to total daily intakes and food groups consumed among children and adolescents across three nationally representative nutrition surveys in Australia, spanning 16 years.
Comparison of three surveys
Data from three nationally representative nutrition surveys among Australian children and adolescents were used: the 1995 National Nutrition Survey (NNS) , the 2007 National Children’s Nutrition and Physical Activity Survey (NCNPAS) , and the 2011–12 National Nutrition and Physical Activity Survey (NNPAS) . The 1995 NNS and 2011–12 NNPAS were conducted by the Australian Bureau of Statistics (ABS). The 2007 NCNPAS was carried out by the Commonwealth Scientific and Industrial Research Organisation and the University of South Australia. The 1995 NNS and 2011–12 NNPAS surveyed participants aged 2 years and over; while in the 2007 NCNPAS participants were aged 2–16 years.
For all three surveys, households were selected from all Australian states and territories, across urban and rural dwellings. The 1995 NNS and 2011–12 NNPAS used a stratified multi-stage area sampling plan to identify private households, whereas the 2007 NCNPAS recruited households using Random Digit Dialling. For the 1995 NNS and 2011–12 NNPAS, one adult and one child were randomly selected from each household, while one child only was selected for the 2007 NCNPAS. In the 1995 NNS and 2011–12 NNPAS interviews were conducted throughout the year to account for seasonal variation in intakes, while in the 2007 NCNPAS interviews were conducted between February and August 2007. For all three survey interviews were conducted on all seven days of the week. Data were collected by proxy for younger children, and with the assistance of an adult member of the household for older children. Trained interviewers used the 24-h dietary recall method to collect information on foods and beverages consumed on the day prior to the interview. The 1995 NNS interviewers and the 2007 NCNPAS interviewers used a three-pass method, whereas the 2011–12 NNPAS data were collected with an enhanced five pass method. The 1995 NNS used a ‘3 phase multiple pass’ method developed by the United States Department of Agriculture . The 2007 NCNPAS used a standardized, computer-based, three-pass method which employed the following steps: 1st pass – a ‘quick list’ of all foods and beverages consumed, 2nd pass – time and place of consumption for each item, 3rd pass – a ‘recall review’ for corrections and additions . The 2011–12 NNPAS used the Automated Multiple-Pass Method developed by the Agricultural Research Service of the United States Department of Agriculture . This is a five-pass method which involved the following steps: phase one – ‘quick list’, phase two – ‘forgotten foods’, phase three – ‘time and occasion’, phase four – ‘detail cycle’, phase five – ‘final probe’. Two days of recall were obtained in all three surveys, however, in 1995, only 10% of participants had a second day of recall compared to 57% in 2011–12 and 100% in 2007. More information on the three surveys is summarised in Additional file 1: Table S1.
For comparative purposes, data among children aged 2–16 years in each survey were examined. Due to significant differences in meal patterns between weekdays and weekends [35, 36], weekend day recalls were excluded. Only day 1 data were used in 1995 and day 2 data were used in 2007 and 2011–12 when day 1 was a weekend day. Data were weighted to represent the Australian population according to weightings provided by the Australian Bureau of statistic for each survey. A total of 2340 children and adolescents were included in 1995, 3637 in 2007, and 2281 in 2011–12. The interview components of the 1995 and 2011–12 were conducted under the Census and Statistics Act 1905. Ethics approval was not necessary for those surveys. In 2007, ethics approval was obtained from the National Health and Medical Research Council registered Ethics Committees of the Commonwealth Scientific and Industrial Research Organisation and the University of South Australia.
Meal and snacking definitions
Recommendations made by Johnson et al.  were taken into consideration in the development of an objective definition of snacking. There are several approaches used in the literature to define snacking: snacking based on time of day [13, 37], food-based [38,39,40] participant-defined snacking [41, 42], meal patterns , or methods that incorporate social, physiological or situational cues (e.g. after waking or based on hunger or hormones present) . Although participant-defined eating occasions were available for the 1995 and 2011–12 surveys, they were not captured in the 2007 survey and hence it was not possible to use this measure. Further, we felt it was necessary to develop an objective measure of snacking, where the participant’s perception of what constitutes a snack or meal was removed entirely from the analysis, as ‘snacking’ and ‘snack’ may be interpreted in different ways by different people or within different contexts, snacking is often associated with less healthful food choices and ‘snack foods’ can be classified by participants as both a snack and as a meal (e.g. cheese on toast) .
Each food and beverage consumed at the same time of day is reported as a single eating occasion, fully capturing eating patterns as they are consumed. A time of day approach includes all foods and beverages irrespective of energy content or time between eating episodes. This enables all foods and beverages such as diet soft drinks or water, or very low energy foods (i.e. celery sticks) to be further investigated in relation to other components of daily food intake (such as the effect of that eating occasion on subsequent food consumption). It can misclassify specific meals as snacking and snacking as meals, but only when relating to a participant’s perception of the type of eating occasion. Therefore, the time of day approach was utilised to define when meals and snacking occurs. This was determined using previously published methodology using population energy distribution across the day, by plotting daily percent of energy by time of day in 30-min increments .
A meal time period began at the first increase in the percent of daily energy and ended in the trough. Main meals were defined as the three largest peaks of energy, and snacking as the three lowest peaks of energy across the day (online Additional file 2: Figure S1). Main meal and snacking times were similar for all three surveys: breakfast occurred between 05.30–09.30 h (09.00 in 2007); morning snack 09.30–11.30 (09.00 in 2007), the midday meal 11.30–14.30, the afternoon snack 14.30–17.00 (17.30 in 2011–12), the evening meal was 17.00–21.30 (17.30 in 2011–12, 21.00 in 2007) and the late night snack 21.30–05.30 (21.00 in 2007). A sensitivity analysis was performed to investigate the effect of different main meal and snacking times for the three surveys. The times given above for the 2011–12 survey were replicated for the 1995 and 2007 surveys, and absolute energy and energy contribution from snacking, prevalence of consumption at each snacking period, frequency of snacking, and number of snacking occasions were calculated. The results were compared with the results using the survey-specific times, and were found to be sufficiently similar to justify the original methodology (Additional file 3: Tables S2 and S3).
An eating occasion was defined as one or more food or beverage items consumed at the same time of day. A participant who reported a glass of orange juice at 8.15 am, then reported toast and jam at 9.00 am had two eating occasions during ‘breakfast’. Eating occasions that occurred during breakfast, midday and evening meals were meals and all eating occasions that occurred between these meals were classified as snacking. A snacking occasion was defined as one or more food or beverage items consumed at the same time of day within a snacking time period. The prevalence of main meals and snacking consumption was estimated. The mean number of snacking occasions as well as the frequency of snacking were assessed by calculating the percentage of children having zero, one, two, three, or four or more snacking occasions per day.
Dietary intake data
Dietary intake data were analysed for each survey using the survey specific Australian Food Composition Database (AUSNUT) developed by Food Standards Australia New Zealand (FSANZ) [46,47,48]. The AUSNUT database is survey specific, resulting in a different database for each survey, and slightly different food groups. While daily energy and macronutrient intakes have been previously published , we present mean daily energy intake, mean daily energy during snacking, the percent contribution of energy during snacking to total daily intake, and the contribution of carbohydrate, protein, total fat and total sugars to total energy intake during snacking were estimated for each survey.
Discretionary foods are defined by the Australian Dietary Guidelines as foods and beverages not necessary for a healthy diet and are generally high in saturated fat and/or added sugars, added salt or alcohol and low in fibre . The ABS provided coding for discretionary foods in the 2011–12 NNPAS . Using the matching files created by FSANZ, foods from the 1999 AUSNUT database (used for the 1995 NNS) were cross-referenced with 2007 and 2011–13 (used for the 2011–12 NNPAS) to code for discretionary foods [52, 53]. Since there was no matching file for 2007 to 2011–13, a two-step matching approach was used for 2007. When a food group did not have any matching food, the nutrient profile was compared with 2011–13 foods. Foods with similar nutrient content were then grouped together and classified accordingly as discretionary. When a food matched more than one food group in the 1999 or 2007 database, the food was categorised as discretionary if at least one of the 2011–13 foods was discretionary.
In this study, food groups were reported at the sub-major food group classification. There were a total of 107 sub-major food groups in the AUSNUT database in 1999, 111 in 2007, and 132 in 2011–13. The top ten sub-major food groups that reported the most to energy contribution during snacking were reported for each survey. In addition, the mean energy intake of each food group consumed during snacking, the percent of consumers and the mean portion size in grams among consumers of the food group were calculated.
Anthropometric measures were collected by trained interviewers who took weight, height, and waist circumference measurements for all consenting and able respondents. Weight was measured using digital scales, height was measured using a stadiometer and waist circumference was measured using a metal tape measure. The BMI z-score was calculated using a child’s age, sex, height, weight and the World Health Organization growth reference standards for 2–4 year olds and 5–19 year olds [54, 55]. The standard normal distribution was then calculated for the BMI z-scores and children were categorized as: normal weight (< 85%), at risk for overweight (≥ 85% to <95%), overweight (≥ 95%). In children a waist circumference to height ratio of <0.5 is associated with a low risk of metabolic complications from obesity, whereas a ratio of >0.5 is associated with a higher risk . Therefore, a waist circumference to height ratio of 0.5 was used as a cut-off for waist circumference and risk of metabolic complications.
Under-reporting in dietary surveys can selectively affect foods that are reported, but also specific meals and snacks, especially those that are perceived by individuals as less socially desirable (i.e. consumption of junk foods or that skipping breakfast is ‘unhealthy’) than others . Energy intake to basal metabolic rate ratio (EI:BMR) was used to estimate number of under-reporters, i.e. participants with implausibly low energy intakes. In the 2011–12 NNPAS the EI:BMR was calculated by the ABS and provided as part of the survey data; for the 1995 NNS and 2007 NCNPAS we used the same methodology as the ABS to calculate participants’ BMR using age, sex, and weight, with no adjustment for activity levels (Additional file 4: Table S4). The surveys did not provide the requisite data on physical activity level (PAL) so an EI:BMR of 1.55 was used for the calculation of under-reporting cutoffs in line with the ABS recommendations from the Australian Health Survey. Participants were classified as under-reporters based on the Goldberg  cut-off limit of 0.9 for EI:BMR, which is the lower 95% confidence limit for a single day of data for a single individual, allowing for day-to-day variation in energy intakes and errors in calculation of EI:BMR. Under-reporting was included as a factor in the analysis, rather than excluding participants with low energy intakes . A more recent analysis of methods to control for under-reporting also suggests that excluding under-reporters may lead to selection bias and inflated associations .
The statistical package IBM SPSS Statistics (Version 23) was used for all analyses. Descriptive summaries were calculated for all variables of interest and chi-squared tests were used to determine statistical significance of categorical variables between survey years. ANOVA tables were produced to calculate standard errors of the mean, and post hoc pairwise comparisons using the Bonferroni correction to show pairwise significance between survey years. A general linear model was used to investigate the effect of age (2–3, 4–8, 9–13, 14–16 years), sex, energy intake, survey year and under-reporting on number of snacking occasions per day. The main effects of the factors were included, together with the interaction of age and sex, and age and year. P-values <0.001 were taken to indicate statistical significance.
There was a higher prevalence of at risk for overweight and overweight children in 2007 (32.7%) and 2011–12 (32.2%) compared to 1995 (27.5%) (Table 1), and a significant increase in mean BMI z-scores from 1995 to 2007 (P < 0.001). In addition, there was a significant increase in mean waist-to-height ratio, between 2011 and 12 and both 2007 and 1995 (P < 0.001).
Snacking prevalence and frequency
There was an increase in the proportion of children snacking, from 92.5% in 1995 to 98.1% in 2007, as well as an increase in the proportion of consumers during each snacking period (morning, afternoon, and late night) (Table 2). From 2007 to 2011–12 there was a slight decline in the percentage of children snacking overall and during each snacking period, however the prevalence was substantially higher in 2011–12 than 1995.
The frequency of snacking rose across survey years with a higher percentage of children consuming three or four or more snacking occasions in 2007 (25.5% and 17.9% respectively) and 2011–12 (26.9% and 18.5% respectively) compared to 1995 (18.3% and 7.1% respectively) (Table 2). The percentage of children who had fewer than three snacking occasions progressively decreased from 74.6% in 1995 to 56.6% in 2007 and 54.6% in 2011–12.
The mean number of snacking occasions per day increased significantly from 1995 to 2007 and 2011–12 (P < 0.001). Figure 1 represents the model for the mean number of snacking occasions with age group, sex, energy intake, year, and under-reporting, all showing statistical significance (P < 0.001), however the R-squared value is only 8%. Number of snacking occasions was directly proportional to energy intake for all three years investigated, decreased with age and was higher among females than males, though the effect size was quite small. From the figure, the largest sex difference is about 0.3 snacking occasions in 14-16y olds in 1995, the largest age group difference is about 0.7 (between 2-3y and 14-16y boys in 2011), and the largest year difference is about 0.5 (in 14-16y boys between 1995 to 2007). The standardised effect sizes (for sex, age group and year) are between 0.15 and 0.35 and hence, they are ‘small’. The low R-squared value implies that the variance in snacking frequency cannot be accurately explained by age, sex and energy intake alone. While the number of snacking occasions increased from 1995 to 2007 and 2011–12, the factors that were significantly associated with frequency of snacking remained the same in each survey year.
Energy and macronutrients
While total daily energy intake decreased from 1995, energy intake from snacking increased significantly (Table 2). As a result, the contribution of snacking to total energy intake also increased significantly. The contribution from protein to total energy during snacking increased from 1995, but the contributions from fat, carbohydrate, and total sugars all decreased. In addition, the proportion of discretionary energy that came from snacking increased (33.6% vs. 37.9% vs. 40.0%, P < 0.001), whereas the proportion of total snacking energy that came from discretionary foods decreased significantly from 1995 to 2011–12 (56.5% vs. 47.9%, P < 0.001).
Figure 2 shows that in 2007 and 2011–12, energy intake was more evenly distributed throughout the day than in 1995. This pattern is most evident in the middle four meal and snack periods: in 1995 the midday and evening meals contributed 57.5% of total daily energy intake, compared to 51.7% in 2011–12; in 1995 the morning and afternoon snacking periods contributed 22.8% of energy intake compared to 29.7% in 2011–12. In contrast, there was almost no change between 1995 and 2011–12 for breakfast and the late night snack period.
Food groups consumed during snacking
Top food groups that contributed the most to energy intake during snacking included a combination of those that were predominantly core foods and beverages (bread, dairy milk, pome fruit) and predominantly discretionary (sweet biscuits, cakes, potato snacks) in all three surveys (Table 3) . Beverages were popular in 1995, with fruit and vegetable juices and drinks the top-ranked food group, and soft drinks, flavoured mineral waters and electrolyte drinks the ninth-ranked food group. In contrast, aside from dairy milk, these non-alcoholic beverages did not appear in the top ten in 2007 or 2011–12.
Among discretionary foods, the sub-major food group cakes, buns, muffins, scones, cake-type desserts appeared in the top ten in all three surveys, but their contribution to snacking energy intake varied from 5.4% in 1995 to 6.1% in 2007 and 5.7% in 2011–12. The portion size for this food group was 79.4 g in 1995 compared to 106.8 g in 2011–12. Although the energy contribution from sweet biscuits and potato snacks remained constant between 1995 and 2011–12, the prevalence of consumption of these food groups during snacking differed (22.3% vs. 23.7%, 11.6% vs. 12.1% respectively) as well as the amount consumed (31.8 g vs. 33.4 g, 31.7 g vs. 34.1 g respectively). Energy contribution from chocolate and chocolate-based confectionery to snacking was 4.1% in 1995, 3.0% in 2007, and was not in the top ten in 2011–12.
Among core foods, the contribution from dairy milk and regular breads to energy intake during snacking varied between 1995 and 2011–12 (19% vs. 15.3%, 23.7% vs. 20% respectively) due to a reduction in both portion size and prevalence of consumption. Pome fruit, however, showed the opposite trend in proportion of energy intake from 5.6% in 1995, to 6.9% in 2007 and 6.6% in 2011–12. This was due to changes in prevalence of consumption (22.1% in 1995, 29.8% in 2007) and portion size (153.4 g in 1995, 176.8 g in 2011–12).
This study investigated trends in snacking patterns among Australian children and adolescents using three nationally representative surveys. It showed that dietary patterns have shifted to include a greater prevalence and frequency of snacking, with more than double the number of children in 2011–12 having four or more snacking occasions per day than in 1995. These findings are consistent with those in both developing and developed countries [8,9,10,11,12, 16, 18, 19, 21, 43, 61, 62]. Over three decades in the US, the prevalence of snacking increased from 74% in 1977–78 to 98% in 2003–06, while the prevalence of children with less frequent snacking (one or two snacking occasions per day) decreased from 26% to 20% . Similarly in China, where a snacking eating pattern is less prevalent, the proportion of children who snacked between 2004 and 2006 more than doubled .
In the present study, the energy contribution from snacking in Australian children increased from 24% in 1995 to 30.5% in 2011–12. Similarly, snacking contributes over a quarter of total energy intake in the US [10, 11], 21% in Scotland , 23.3% in Brazil , and 32.5% in the United Kingdom , but less than 20% in other countries such as Russia and the Philippines , Mexico [14, 15] and China . Thus, this highlights the importance of having culture-specific snacking recommendations.
With the greater proportion of daily energy from snacking, we showed a concurrent decrease in energy contribution from main meals. Over the three surveys, the peaks in energy distribution across the day ‘flattened’. This has major implications for school-aged children in meeting their increased nutrient needs , as more food is now being consumed outside of traditionally nutrient-rich main meals, and hence, these main meals are contributing less to total energy intake and possibly to nutrient intakes. We found that the greatest change in energy distribution occurred in meals and snacks in the middle of the day. Breakfast and the late night snack, which among children are traditionally consumed at home rather than away from home, remained constant. It was not possible to report on foods consumed at home or away from home in Australia because location of food consumption was only reported in the 2007 survey. In other countries, food consumed away from home is on the rise [5,6,7, 64,65,66,67]. Several factors may play a role in changing meal patterns, including increased incomes and the ability to purchase more foods and beverages outside the home [68, 69], fewer family meals at home, dual-working parents , urbanisation [71, 72] and increased food availability and affordability .
The clear trend of a decrease in snacking with age across all three surveys in Australia is also reported worldwide [5, 10, 11, 14] with the greatest increase in snacking prevalence observed among the younger age groups [10, 14]. In our model, total energy intake was the second most important determinant of number of snacking occasions in all three survey periods. Studies report a positive association between frequency of eating occasions and energy intake [15, 27, 74, 75], but others report that snacking frequency may also assist with energy regulation and that snacking is not necessarily associated with poor diet quality . Our study showed that the relationship between snacking and diet quality is modulated by food choice during snacking, its frequency and portion size. The contribution of snacking to total discretionary energy increased 2011–12 compared to 1995, which was primarily due to the increase in the contribution of snacking to total energy. Main meals contributed the most to discretionary energy intake across all years and the proportion of snacking energy that was discretionary actually decreased. Total discretionary energy intake among children in Australia has declined between 1995 and 2007 [76, 77], and our present findings showed that it continued to decline in 2011–12. Despite snacking contributing less to total discretionary intake, it still represents just under half of discretionary energy intake. Thus an overall reduction in discretionary intake is still warranted, from both meals and snacking. Our findings imply that practical dietary recommendations around snacking need to be culturally relevant and age-appropriate with a focus on the nutrient density or quality of the snack. Under-reporting has increased between 1995 and 2011, and adjustment does not fully account for the fact that certain ‘less-healthy’ foods or beverages may be more likely to be under-reported than others . Thus, true differences in foods consumed across the three surveys may be due to underreporting alone.
We found that a mixture of traditional ‘snack foods’ (i.e. sweet biscuits, and fresh fruit) and ‘meal’ foods (i.e. bread and milk) were consumed during snacking in Australia. Non-alcoholic beverages (fruit and vegetable juices and soft drinks) were among the top contributors of snacking energy in 1995 but did not feature in the top ten in 2007 or 2011–12; pome fruit was the fourth-highest contributor in 1995 and the only fruit in the top ten in 1995, whereas it was the second-highest contributor in 2007 and 2011–12, and tropical fruits also appeared in the top ten. Cakes and desserts remained a top contributor to snacking energy in all surveys, in line with previous Australian reports on trends in food groups consumed [76, 77, 79]. Varying trends in foods and beverages consumed during snacking are reported among children and adolescents and seem to be country-specific. They include a mixture of both core and discretionary snack foods and beverages. For example, fruit, milk and sugar-sweetened beverages as snacks have increased in China . In contrast, the percentage of fruit and milk consumed as a snack decreased in the US between 1977 and 78 and 2003–06 , but the energy density of snacks remained the same. A more recent US study, using nationally representative nutrition data from 2003 to 2010, also reported a decrease in energy from discretionary foods among children of all weight categories aged 2–5 years and 12–19 years, but not among those overweight aged 6–11 years . In contrast, in the Philippines, milk and soft drinks consumed as snacks more than doubled between 1994 and 2002 while fast food intake did not change . The effect that snacking has on children’s diet quality is therefore dependent on food choice. In a sub-sample of American children, weekday self-reported snacking was positively associated with age-adjusted diet quality among 9–11 year olds and negatively associated with diet quality in those 12–18 years old , but foods consumed as snacks were not reported in that study. Among British adolescents, a positive relationship existed between diet quality and number of snacks per day but this was dependent on the nutritional quality of the snack food or beverage consumed . A recent study among British children and adolescents reported that snack frequency was associated with a lower diet quality score . Similarly, meal frequency has been associated with higher diet quality among children, but the association of snacking frequency with diet quality was dependent on the definition used: time of day or energy contribution .
As a consequence of the mixture of nutrient-dense and nutrient-poor food choices during snacking, emphasis on public health and nutrition intervention programs should encourage nutritious snack food choices based on the core food groups, especially since a snacking dietary pattern has become more prevalent. For example, in Australia, the primary school in-class Crunch&Sip® program was introduced in 2005 by the government to encourage students to consume fruit or vegetables and water during class and develop healthy eating behaviours. Nearly half of surveyed government schools have implemented this program into >80% of classrooms  and perhaps the increase in fruit intake that was observed on weekdays during snacking in the present study could be, in part, due to this program. It is especially important to focus on the provision of foods and beverages to children, as when given the choice between less nutrient dense and more nutrient dense snack options, children tend to choose the less nutrient dense one: they choose sugar-sweetened beverages over fresh fruit .
The strengths of our study include the use of an objective definition of snacking based on time-of-day. This removes any participant bias for what constitutes a ‘snack food’. Even among children’s care-givers, snacks can be defined by type of food, portion size, time of day, eating episode, location, or even purpose [85,86,87]. A study on adults, by Leech et al. revealed that methodologies based on participants’ self-report of the eating occasions and time-of-day lead to significant differences in snacking frequency but non-significant differences in eating frequency, energy intake and energy contribution from snacking . Further research is needed in order to reach to a consensus on an appropriate definition of meals and snacks [25, 88]. Furthermore, only weekdays were included for consistency in meal patterns, as evidence has shown that weekend patterns are quite different from weekdays [35, 36, 89]. Lastly, prevalence of under-reporting across the three years was accounted for in the general linear model for number of snacking occasions model rather than by excluding those participants from the analysis, which has recently been shown to increase results bias .
The present study displays several limitations. Although the trend analysis was based on three nationally representative surveys based on one 24-h dietary recall, the methodology of the recall collection was different across surveys and included both face-to-face (1995 and 2011–12) and telephone interviews (2007). However, results from the US show little difference between in-person and telephone interviews using the automated multiple pass method , so we would expect this to have a minimal impact on our findings. The eating pattern analysis was derived from one weekday 24-h recall which is not necessarily representative of usual intakes. In addition, the 2007 survey did not collect data in all months of the year, which may miss differences if snacking patterns in September to January are different from those in the survey months of February through August. It is possible that snacking patterns could differ according to whether the dietary recall day was a school day, but the surveys did not collect data specifying this. Furthermore, the coding of discretionary foods was limited by some differences in food groups across surveys. A systematic approach was used to code for discretionary foods across the surveys in order to minimise bias and accurately capture changes in food group intakes. A limitation to the time of day definition approach is that the time of eating may have changed over the time period studied in response to external influences (e.g. differences in school schedules), which may affect the way the consumption peaks are distributed in each survey and therefore the timing used to define snacking. This effect was evaluated in our sensitivity analysis and found to be minimal. Further, by classifying meals and snacking on the basis of aggregated energy intake data is unlikely to detect important between-subject differences or subpopulations who consume meals or who snack at unconventional times of the day.
Lastly, the present study is based on three cross-sectional surveys, which does not enable us to establish any causal link between snacking and obesity.
In conclusion, in this study investigating trends in weekday snacking across three nationally representative surveys 1995, 2007 and 2011–12, a large proportion of Australian children and adolescents were found to be snackers, with increases in snacking frequency and the contribution of snacking to total energy, but a decrease in the contribution of snacking to discretionary energy. Both discretionary and core foods were consumed during snacking, highlighting the need for targeted snacking recommendations specific to the local culture. Further research to profile timing and reasons for foods consumed during snacking, which include both nutrient dense and nutrient poor choices, will help understand shifts in dietary patterns and nutrient intakes over time and advise health promotion strategies on population-based food and beverage recommendations that are culturally relevant.
Australian Bureau of Statistics
Australian Health Survey
Australian Food Composition Database
Body Mass Index
Energy intake to basal metabolic rate ratio
Food Standards Australia New Zealand
2007 National Children’s Nutrition and Physical Activity Survey
2011–12 National Nutrition and Physical Activity Survey
National Nutrition Survey
Lobstein T, Baur L, Uauy R, TaskForce IIO. Obesity in children and young people: a crisis in public health. Obes Rev. 2004;5(Suppl 1):4–104.
World Health Organisation. Interim Report of the Commission on Ending Childhood Obesity. Geneva: WHO; 2015.
Australian Bureau of Statistics. Australian Health Survey: First results, 2014–15. ABS, Canberra, 2015.
Australian Bureau of Statistics. Profiles of Health, Australia, 2011–13: Children's risk factors. ABS, Canberra, 2013.
Adair LS, Popkin BM. Are child eating patterns being transformed globally? Obes Res. 2005;13:1281–99.
Kant AK, Graubard BI. 40-year trends in meal and snack eating behaviors of American adults. J Acad Nutr Diet. 2015;115:50–63.
Nielsen SJ, Siega-Riz AM, Popkin BM. Trends in food locations and sources among adolescents and young adults. Prev Med. 2002;35:107–13.
Nielsen SJ, Siega-Riz AM, Popkin BM. Trends in energy intake in U.S. between 1977 and 1996: similar shifts seen across age groups. Obes Res. 2002;10:370–8.
Duffey KJ, Popkin BM. Causes of increased energy intake among children in the U.S., 1977-2010. Am J Prev Med. 2013;44:e1–8.
Jahns L, Siega-Riz AM, Popkin BM. The increasing prevalence of snacking among US children from 1977 to 1996. J Pediatr. 2001;138:493–8.
Piernas C, Popkin BM. Trends in snacking among U.S. children. Health Aff (Millwood). 2010;29:398–404.
Zizza C, Siega-Riz AM, Popkin BM. Significant increase in young adults' snacking between 1977-1978 and 1994-1996 represents a cause for concern! Prev Med. 2001;32:303–10.
Kerr MA, Rennie KL, McCaffrey TA, Wallace JM, Hannon-Fletcher MP, Livingstone MB. Snacking patterns among adolescents: a comparison of type, frequency and portion size between Britain in 1997 and Northern Ireland in 2005. Br J Nutr. 2009;101:122–31.
Duffey KJ, Rivera JA, Popkin BM. Snacking is prevalent in Mexico. J Nutr. 2014;144:1843–9.
Taillie LS, Afeiche MC, Eldridge AL, Popkin BM. Increased Snacking and Eating Occasions Are Associated with Higher Energy Intake among Mexican Children Aged 2-13 Years. J Nutr. 2015;145:2570–7.
Duffey KJ, Pereira RA, Popkin BM. Prevalence and energy intake from snacking in Brazil: analysis of the first nationwide individual survey. Eur J Clin Nutr. 2013;67:868–74.
Liu Y, Zhai F, Popkin BM. Trends in eating behaviours among Chinese children (1991 -1997). Asia Pac J Clin Nutr. 2006;15:72–80.
Wang Z, Zhai F, Zhang B, Popkin BM. Trends in Chinese snacking behaviors and patterns and the social-demographic role between 1991 and 2009. Asia Pac J Clin Nutr. 2012;21:253–62.
Ribas-Barba L, Serra-Majem L, Salvador G, Castell C, Cabezas C, Salleras L, Plasencia A. Trends in dietary habits and food consumption in Catalonia, Spain (1992-2003). Public Health Nutr. 2007;10:1340–53.
Booth DA. Mechanisms from models--actual effects from real life: the zero-calorie drink-break option. Appetite. 1988;11(Suppl 1):94–102.
Bellisle F. Meals and snacking, diet quality and energy balance. Physiol Behav. 2014;134:38–43.
Johnson GH, Anderson GH. Snacking definitions: impact on interpretation of the literature and dietary recommendations. Crit Rev Food Sci Nutr. 2010;50:848–71.
Larson N, Story M. A review of snacking patterns among children and adolescents: what are the implications of snacking for weight status? Child Obes. 2013;9:104–15.
Kaisari P, Yannakoulia M, Panagiotakos DB. Eating frequency and overweight and obesity in children and adolescents: a meta-analysis. Pediatrics. 2013;131:958–67.
Murakami K, Livingstone MB. Associations between meal and snack frequency and overweight and abdominal obesity in US children and adolescents from National Health and Nutrition Examination Survey (NHANES) 2003-2012. Br J Nutr. 2016;115:1819–29.
Chapelot D. The role of snacking in energy balance: a biobehavioral approach. J Nutr. 2011;141:158–62.
Duffey KJ, Popkin BM. Energy density, portion size, and eating occasions: contributions to increased energy intake in the United States, 1977-2006. PLoS Med. 2011;8:e1001050.
Evans EW, Jacques PF, Dallal GE, Sacheck J, Must A. The role of eating frequency on total energy intake and diet quality in a low-income, racially diverse sample of schoolchildren. Public Health Nutr. 2015;18:474–81.
Goran MI, Ball GD, Cruz ML. Obesity and risk of type 2 diabetes and cardiovascular disease in children and adolescents. J Clin Endocrinol Metab. 2003;88:1417–27.
Mikkila V, Rasanen L, Raitakari OT, Pietinen P, Viikari J. Consistent dietary patterns identified from childhood to adulthood: the cardiovascular risk in Young Finns Study. Br J Nutr. 2005;93:923–31.
McLennan W, Podger A. National Nutrition Survey: User's Guide - 1995; ABS cat no. 4804.0. ABS and DHFS, Canberra, 1998.
Commonwealth Scientific and Industrial Research Organisation. The 2007 National Children's Nutrition and Physical Activity Survey. Australian Social Science Data Archive. Australian Commonwealth Department of Health and Ageing, and University of South Australia, Canberra, 2009.
Australian Bureau of Statistics. Australian Health Survey: First results, 2011–12; ABS cat no. 4364.0.55.001. ABS, Canberra, 2014.
Moshfegh AJ, Rhodes DG, Baer DJ, Murayi T, Clemens JC, Rumpler WV, Paul DR, Sebastian RS, Kuczynski KJ, Ingwersen LA, et al. The US Department of Agriculture Automated Multiple-Pass Method reduces bias in the collection of energy intakes. Am J Clin Nutr. 2008;88:324–32.
Hart CN, Raynor HA, Osterholt KM, Jelalian E, Wing RR. Eating and activity habits of overweight children on weekdays and weekends. Int J Pediatr Obes. 2011;6:467–72.
Yates L, Warde A. The evolving content of meals in Great Britain. Results of a survey in 2012 in comparison with the 1950s. Appetite. 2015;84:299–308.
Ziegler P, Briefel R, Ponza M, Novak T, Hendricks K. Nutrient intakes and food patterns of toddlers' lunches and snacks: influence of location. J Am Diet Assoc. 2006;106:S124–34.
Berteus Forslund H, Lindroos AK, Sjostrom L, Lissner L. Meal patterns and obesity in Swedish women-a simple instrument describing usual meal types, frequency and temporal distribution. Eur J Clin Nutr. 2002;56:740–7.
Jimenez-Cruz A, Bacardi-Gascon M, Jones EG. Consumption of fruits, vegetables, soft drinks, and high-fat-containing snacks among Mexican children on the Mexico-U.S. border. Arch Med Res. 2002;33:74–80.
Nicklas TA, Morales M, Linares A, Yang SJ, Baranowski T, De Moor C, Berenson G. Children's meal patterns have changed over a 21-year period: the Bogalusa Heart Study. J Am Diet Assoc. 2004;104:753–61.
Hartmann C, Siegrist M, van der Horst K. Snack frequency: associations with healthy and unhealthy food choices. Public Health Nutr. 2012;16(8):1487–1496.
Maffeis C, Grezzani A, Perrone L, Del Giudice EM, Saggese G, Tato L. Could the savory taste of snacks be a further risk factor for overweight in children? J Pediatr Gastroenterol Nutr. 2008;46:429–37.
Macdiarmid J, Loe J, Craig LC, Masson LF, Holmes B, McNeill G. Meal and snacking patterns of school-aged children in Scotland. Eur J Clin Nutr. 2009;63:1297–304.
Wadhera D, Capaldi ED. Categorization of foods as "snack" and "meal" by college students. Appetite. 2012;58:882–8.
Fayet F, Mortensen A, Baghurst K. Energy distribution patterns in Australia and its relationship to age, gender and body mass index among children and adults. Nutr Diet. 2012;69:102–10.
Zealand FSAN. AUSNUT 2011–13- Australian Food, Supplement and Nutrient Database for Estimation of Population Nutrient Intakes. Canberra: FSANZ; 2014.
Food Standards Australia New Zealand. AUSNUT1999. FSANZ, Canberra, 1999.
Food Standards Australia New Zealand. AUSNUT2007. FSANZ, Canberra, 2007.
Australian Bureau of Statistics. Australian Health Survey: Nutrition First Results - Foods and Nutrients, 2011-2012; ABS cat no. 4364.0.55.007. ABS, Canberra, 2014.
National Health and Medical Research Council. Australian Dietary Guidelines. NHMRC, Canberra, 2013.
Australian Bureau of Statistics. Australian Health Survey 2011–13: User's Guide. ABS, Canberra, 2014.
Food Standards Australia New Zealand. 2011–13 AHS and NNS food classification concordance file: access to Excel file. [http://www.foodstandards.gov.au/science/monitoringnutrients/ausnut/ausnutdatafiles/Pages/concordance.aspx], Accessed on January 2016.
Food Standards Australia New Zealand. AUSNUT 2007 - AUSNUT 1999 Matching File [Excel] in AUSNUT 2007 Supporting Documentation. [http://www.foodstandards.gov.au/science/monitoringnutrients/ausnut/pages/ausnut2007supporting4063.aspx], Accessed on January 2016.
World Health Organisation. Growth reference 2–4 years. Geneva: WHO; 2014.
World Health Organisation. Growth reference 5–19 years. Geneva: WHO; 2014.
Weili Y, He B, Yao H, Dai J, Cui J, Ge D, Zheng Y, Li L, Guo Y, Xiao K, et al. Waist-to-height ratio is an accurate and easier index for evaluating obesity in children and adolescents. Obesity (Silver Spring). 2007;15:748–52.
Becker W, Welten D. Under-reporting in dietary surveys--implications for development of food-based dietary guidelines. Public Health Nutr. 2001;4:683–7.
Black AE. Critical evaluation of energy intake using the Goldberg cut-off for energy intake:basal metabolic rate. A practical guide to its calculation, use and limitations. Int J Obes Relat Metab Disord. 2000;24:1119–30.
Jessri M, Lou WY, L'Abbe MR. Evaluation of different methods to handle misreporting in obesity research: evidence from the Canadian national nutrition survey. Br J Nutr. 2016;115:147–59.
Vainik U, Konstabel K, Latt E, Maestu J, Purge P, Jurimae J. Diet misreporting can be corrected: confirmation of the association between energy intake and fat-free mass in adolescents. Br J Nutr. 2016;116:1425–36.
Hess J, Slavin J. Snacking for a cause: nutritional insufficiencies and excesses of U.S. children, a critical review of food consumption patterns and macronutrient and micronutrient intake of U.S. children. Nutrients. 2014;6:4750–9.
Popkin BM, Duffey KJ. Does hunger and satiety drive eating anymore? Increasing eating occasions and decreasing time between eating occasions in the United States. Am J Clin Nutr. 2010;91:1342–7.
Buttriss J. Nutrition, health and schoolchildren. Nutr Bull. 2002;27:275–316.
Briefel RR, Johnson CL. Secular trends in dietary intake in the United States. Annu Rev Nutr. 2004;24:401–31.
Kant AK, Graubard BI. Eating out in America, 1987-2000: trends and nutritional correlates. Prev Med. 2004;38:243–9.
St-Onge MP, Keller KL, Heymsfield SB. Changes in childhood food consumption patterns: a cause for concern in light of increasing body weights. Am J Clin Nutr. 2003;78:1068–73.
O'Dwyer NA, Gibney MJ, Burke SJ, McCarthy SN. The influence of eating location on nutrient intakes in Irish adults: implications for developing food-based dietary guidelines. Public Health Nutr. 2005;8:258–65.
Jeffery RW, Utter J. The changing environment and population obesity in the United States. Obes Res. 2003;11(Suppl):12S–22S.
Du S, Mroz TA, Zhai F, Popkin BM. Rapid income growth adversely affects diet quality in China--particularly for the poor! Soc Sci Med. 2004;59:1505–15.
Anderson PM, Butcher KE. Childhood obesity: trends and potential causes. Futur Child. 2006;16:19–45.
Popkin BM. Urbanization, Lifestyle Changes and the Nutrition Transition. World Dev. 1999;27:1905–16.
Hawkes C. Uneven dietary development: linking the policies and processes of globalization with the nutrition transition, obesity and diet-related chronic diseases. Glob Health. 2006;2:4.
Thow AM, Hawkes C. The implications of trade liberalization for diet and health: a case study from Central America. Glob Health. 2009;5:5.
Kearney J. Food consumption trends and drivers. Philos Trans R Soc Lond Ser B Biol Sci. 2010;365:2793–807.
Llaurado E, Albar SA, Giralt M, Sola R, Evans CE. The effect of snacking and eating frequency on dietary quality in British adolescents. Eur J Nutr. 2015;55(4):1789–1797.
Rangan AM, Kwan J, Flood VM, Louie JC, Gill TP. Changes in 'extra' food intake among Australian children between 1995 and 2007. Obes Res Clin Pract. 2011;5:e1–e78.
Rangan AM, Kwan JS, Louie JC, Flood VM, Gill TP. Changes in core food intake among Australian children between 1995 and 2007. Eur J Clin Nutr. 2011;65:1201–10.
Macdiarmid J, Blundell J. Assessing dietary intake: Who, what and why of under-reporting. Nutr Res Rev. 1998;11:231–53.
Collins K, Watson JF, Collins CE. Food and beverage portion sizes in Australian children: a secondary analysis of 1995 and 2007 national data. BMC Public Health. 2014;14:517.
Bleich SN, Wolfson JA. Trends in SSBs and snack consumption among children by age, body weight, and race/ethnicity. Obesity (Silver Spring). 2015;23:1039–46.
Murakami K, Livingstone MB. Decreasing the number of small eating occasions (<15% of total energy intake) regardless of the time of day may be important to improve diet quality but not adiposity: a cross-sectional study in British children and adolescents. Br J Nutr. 2016;115:332–41.
Murakami K, Livingstone MB. Meal and snack frequency in relation to diet quality in US children and adolescents: the National Health and Nutrition Examination Survey 2003-2012. Public Health Nutr. 2016;19(9):1635–1644.
Nathan N, Wolfenden L, Butler M, Bell AC, Wyse R, Campbell E, Milat AJ, Wiggers J. Vegetable and fruit breaks in Australian primary schools: prevalence, attitudes, barriers and implementation strategies. Health Educ Res. 2011;26:722–31.
Beets MW, Tilley F, Kyryliuk R, Weaver RG, Moore JB, Turner-McGrievy G. Children select unhealthy choices when given a choice among snack offerings. J Acad Nutr Diet. 2014;114:1440–6.
Younginer NA, Blake CE, Davison KK, Blaine RE, Ganter C, Orloski A, Fisher JO. "What do you think of when I say the word 'snack'?" Towards a cohesive definition among low-income caregivers of preschool-age children. Appetite. 2016;98:35–40.
Chamontin A, Pretzer G, Booth DA. Ambiguity of 'snack' in British usage. Appetite. 2003;41:21–9.
Blake CE, Bisogni CA, Sobal J, Devine CM, Jastran M. Classifying foods in contexts: how adults categorize foods for different eating settings. Appetite. 2007;49:500–10.
Leech RM, Worsley A, Timperio A, McNaughton SA. Characterizing eating patterns: a comparison of eating occasion definitions. Am J Clin Nutr. 2015;102:1229–37.
An R. Weekend-weekday differences in diet among U.S. adults, 2003-2012. Ann Epidemiol. 2016;26:57–65.
The authors would like to thank Kate Tuck for her assistance with the literature interpretation and formatting of the manuscript.
Funding for this analysis was provided by the Nestlé Research Center, Lausanne, Switzerland (Nestlé Ltd.)
Availability of data and materials
More information on each of the surveys in this analysis can be obtained from the Australian government Department of Health (http://www.health.gov.au/internet/main/publishing.nsf/Content/health-pubhlth-strateg-food-monitoring.htm#07survey). The data generated that support the findings of this study are available from the corresponding author upon reasonable request.
Ethics approval and consent to participate
The interview components of the 1995 and 2011–12 were conducted under the Census and Statistics Act 1905. Ethics approval was not necessary for those surveys. In 2007, ethics approval was obtained from the National Health and Medical Research Council registered Ethics Committees of the Commonwealth Scientific and Industrial Research Organisation and the University of South Australia.
Consent for publication
FFM, AM, and PP declare that they have no competing interests. VP and ALE are employed by the funding body.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Characteristics of the National Nutrition Surveys 1995, 2007 and 2011–12. (DOCX 14 kb)
Percent of energy consumed by time of day and classification of meal and snack time periods: 1995, 2007 and 2011–12 National Nutrition Surveys. In 1995 n = 2340, 2007 n = 3637, 2011–12 n = 2281. (PDF 71 kb)
Sensitivity analysis: snacking patterns of children and adolescents 2-16y from the Australian National Nutrition Surveys 1995, 2007 and 2011–12. (DOCX 15 kb)
Basal Metabolic Rate (BMR) for each age and sex group. (DOCX 12 kb)