Food Behaviours and Health Indicators in Manitoba Adolescents and Relation to the Healthy Eating Index

Adolescence is a vital period of growth and development, both of which are dependent on adequate nutrition; however, concerns persist about poor nutrition and inappropriate food behaviours. In addition to nutrition assessment, the context of food and health behaviour is necessary to understand how dietary choices are shaped and related to diet quality. This study describes food-related behaviours and health indicators associated with dietary quality among adolescents in Manitoba, Canada. A stratified two-stage sampling method was used to collect data on the diet, food behaviours and health indicators of 1587 grade nine students. Diet quality was analysed using the Healthy Eating Index-Canada. Several food behaviours and health indicators varied by gender and school region (urban, rural, northern). The Independent Samples t-test and one-way ANOVA (analysis of variance) assessed differences between groups on the Healthy Eating Index-Canada. Higher Healthy Eating Index-Canada scores were found for those eating family dinners more frequently; consuming breakfast and lunch more frequently; consuming breakfast at home; eating lunch and morning snacks at school; purchasing fewer meals and snacks from cafeterias and vending machines; believing that food and nutrition education is important; not attempting to lose weight; being classified as ‘healthy weight’; and getting more sleep. Many Manitoba youth are exhibiting food and health behaviours that increase their risk of having a poor diet.


Introduction
Adolescence is a critically important period of growth and development [1]; however, concerns about obesity and poor nutritional status in adolescence continue to grow among public health authorities [2,3]. In Canada, 12% of children aged 5-17 are obese and 19% are overweight [4]. Poor diet is associated with excess weight and the excessive consumption of highly processed, high fat and sugar foods [5][6][7]. These foods are ubiquitous in Canada, with the highest proportion being consumed by children and adolescents [8]. They are readily available, high in calories and tend to have low nutrient density [9]. Diets high in ultra-processed foods (e.g., convenience foods, sweet and salty snacks, fast food) are also associated with increased risk of obesity and all-cause mortality and numerous illnesses including cardiovascular diseases, cancers and respiratory diseases [10]. In adolescents,

Materials and Methods
The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Joint Faculty Research Ethics Board at the University of Manitoba (protocol code HS21666 J2018:040).

Participants
The Food and Nutrition Security for Manitoba Youth (FANS) study included a selfadministered web-based survey that assessed dietary intakes, food behaviours and health-related indicators of grade nine students attending public school in the province of Manitoba, Canada. Eighteen of 37 school divisions were approached, with 14 agreeing. Within the agreeing divisions, classes with 10 or more grade nine students were invited to participate. Parents of participating students signed written consent forms and students provided assent at the start of the survey. A detailed description of the study design and rationale is reported elsewhere including demographic information on participants [28].
School divisions and schools were classified into northern, rural and urban regions of the province: school divisions within the Winnipeg Health Region are urban; divisions in the Northern Health Region are classified as northern; and remaining divisions are considered rural [29].

Measures
Dietary intake was assessed using a single 24 h dietary recall as part of the WEB-Q (Waterloo Eating Behaviour Questionnaire), a validated online tool for measuring the food and nutrient intake of adolescents [30]. All surveys were completed in the classroom during a weekday.
Diet quality was assessed using a Healthy Eating Index, a measure for assessing how closely an individual's diet, or set of foods consumed, aligns with dietary guidelines [31], such as Canada's Food Guide (2007) [32]. This study used a modified Canadian version of the Healthy Eating Index (HEI-C) [33]: 'low fat vitamin-D fortified milk and dairy-free beverage alternatives' and 'seafood and plant-based proteins' were criteria added to align with newer Canadian dietary recommendations [34]. Mean nutrient intakes and food group intakes were used to compute the HEI-C score (0-100), the main outcome variable.
Using the online WEB-Q, food behaviour was assessed via questions about frequency of meal consumption, meals consumed with family members and food purchasing habits. Health indicators included questions about eating-related weight control and sleep behaviours. Self-reported health and life satisfaction measures were included [35,36]. Body mass index (BMI) was calculated using self-reported height and weight, and classified using World Health Organization z-scores [37].
Students self-identifying as 'other' gender were not included in analyses because of small numbers, and because the HEI-C calculation uses sex-specific (male and female) dietary guidelines (32 and 33).

Statistical Analysis
Study data were analysed using SAS Version 9.4 (SAS Institute Inc., Cary, NC, USA) (variable derivation) and IBM SPSS Statistics for Windows Version 27.0 (IBM Corp., Armonk, USA) (tables and statistical outputs) statistical software packages. Chi-square test was used to determine significant associations and follow-up pairwise comparisons were carried out using Bonferroni correction for food behaviour and health indicator variables, by sex and region. Fisher-Freeman-Halton Exact test was performed with BMI classification categories by region due to low expected cell counts for underweight participants. For continuous descriptive data, a Shapiro-Wilk test was used to confirm normality. Independent samples t-test or one-way ANOVA with Tukey's honestly significant difference (HSD) post hoc test was conducted with 2 and 3 comparison groups, respectively, to determine differences in mean HEI-C scores. Statistical significance was accepted at p ≤ 0.05. All analyses used unweighted sample data.

Recruitment and Participants
The participating students represented 37 schools from 14 school divisions across Manitoba. Of the 1587 students who participated, 8%, 14% and 78% were from northern, rural and urban regions of the province, respectively. Half (50.5%) self-identified as female and 44.7% self-identified as male; 4.8% identified as "other" or did not disclose their sex.

Food Behaviours and Health Indicators
Just over three-quarters of students had dinner with their family 5-7 days per week (Table 1); however, almost one in ten dined with family only 'once per week' or 'never'. Slightly more males than females dined with family most days. There was no difference in family dinner frequency by region. Almost two-thirds of students ate breakfast 5-7 days per week while over 80% had lunch this frequently. More males than females ate breakfast and lunch regularly. Significantly fewer northern students ate breakfast regularly (half) compared to other regions. Lunch frequency did not vary by region.
The majority of students consumed breakfast and dinner at home and lunch at school. Slightly more males than females breakfasted at home. Daytime snacks were mainly consumed at school while most had evening snacks at home. Most (86%) rural students ate lunch at school while 44% of northern students had lunch at home.
Fewer than one in five students purchased food from a cafeteria more than twice a week while only 6% made purchases from vending machines this frequently. There were no differences between males and females with regard to cafeteria and vending purchases. Significantly more northern students (26%) than students in other regions ate cafeteria food more than twice a week while rural students made the most vending purchases.
Significantly more females (68%) than males (55%) had taken a food and nutrition course in school. Rural students had the highest participation in food and nutrition classes: 72% compared to 50% in northern students. Over 80% of students agreed that learning about food and nutrition in school is important.
Self-reported health varied by sex: 68% of males reported excellent/very good health compared to 55% of females.
Almost one third of females reported attempting to lose weight compared to 19% of males; however, almost one quarter of males were trying to gain weight compared to 9% of females. More urban students were trying to gain weight than in other regions.
The majority of students were classified as having a healthy weight, while significantly more males were classified as overweight and obese [28]. Weight classification varied by region, with significantly more students classified as overweight in northern and rural regions compared with urban.
Almost half (46%) of students reported that they did not get enough sleep regularly, with half getting seven hours or fewer on the previous night. Significantly more females than males did not get enough sleep, with eight percent of females getting less than six hours per night regularly. Significantly more northern students had less than six hours of sleep the previous night, and twice as many northern students (11%) regularly slept less than six hours per night compared with rural students, who slept the most. The main reasons for insufficient sleep were homework, screen activities and 'other reasons' (Figure 1). hours per night regularly. Significantly more northern students had less than six hours of sleep the previous night, and twice as many northern students (11%) regularly slept less than six hours per night compared with rural students, who slept the most. The main reasons for insufficient sleep were homework, screen activities and 'other reasons' (Figure  1).

Association with Healthy Eating Index-Canada
Several food behaviours and health indicators showed significant differences in mean Healthy Eating Index-Canada (HEI-C) scores (Table 2). Students who ate dinner with family members less than two days per week had significantly lower mean HEI-C scores than those sharing family dinners two or more days per week. Consuming breakfast and/or lunch 0-4 days per week also yielded significantly lower HEI-C scores. Students who consumed breakfast at home had higher HEI-C scores than those eating breakfast away from home, while those who ate lunch at school or other locations had higher scores than those eating lunch at home. Morning snacks at school were associated with higher HEI-C scores than those eaten at home; there were no differences in scores for the location of dinner or afternoon/evening snacks.

Association with Healthy Eating Index-Canada
Several food behaviours and health indicators showed significant differences in mean Healthy Eating Index-Canada (HEI-C) scores (Table 2). Students who ate dinner with family members less than two days per week had significantly lower mean HEI-C scores than those sharing family dinners two or more days per week. Consuming breakfast and/or lunch 0-4 days per week also yielded significantly lower HEI-C scores. Students who consumed breakfast at home had higher HEI-C scores than those eating breakfast away from home, while those who ate lunch at school or other locations had higher scores than those eating lunch at home. Morning snacks at school were associated with higher HEI-C scores than those eaten at home; there were no differences in scores for the location of dinner or afternoon/evening snacks.
Purchasing meals and snacks from cafeterias and vending machines more frequently was associated with lower HEI-C scores. Students who felt learning about food and nutrition in school was important had higher HEI-C scores, as did students reporting excellent or very good health.
Students attempting to lose weight had significantly lower HEI-C scores than those not trying to lose weight; trying to gain weight did not yield different scores than those not trying to gain weight. Students classified as overweight or obese had significantly lower HEI-C scores than those classified as healthy weight.
Students who got the most sleep per night (8-10 h) had significantly higher HEI-C scores than students getting 7 h or less. Those who usually slept less than six hours had the lowest HEI-C scores.

Discussion
This study explored the food behaviours and health indicators of grade nine adolescents, and their relationship to diet quality. It identified several food behaviours and health indicators associated with poor diet quality as measured by the Healthy Eating Index-Canada.
While the majority of students reported having regular dinners with families, almost ten percent rarely had family dinners (i.e., never or once per week). Further, those having the fewest family dinners had significantly lower diet quality. This is consistent with other studies showing that family meals are associated with higher diet quality (e.g., eating more fruits and vegetables) and healthful meal patterns during adolescence [38,39]. Family meals have other positive benefits including associations with better academic achievement, well-being and fewer behavioural issues [40].
Regular breakfast and lunch consumption was observed in most males, but fewer females. This may be part of weight regulation for some female participants, as 30% reported attempting to lose weight. Breakfast was less frequently consumed in the northern region, where half of students only ate breakfast 0-4 days per week. This could in part be due to household food insecurity, which is significantly greater in northern and remote communities in Canada [26]. Those consuming breakfast regularly had higher HEI-C scores, indicative of better diet quality. This is consistent with the literature showing that breakfast eaters have greater micronutrient intake than breakfast skippers [41]. Breakfast skipping is associated with being overweight or obese [42], and adolescents who skip main meals have lower fruit and vegetable intake [43].
Regional differences in meal/snack purchases at school through vending machines and cafeterias may be due to the availability of these amenities. Northern students purchase food at school least frequently, possibly because food is inaccessible or unaffordable. Northern Manitoba has higher rates of unemployment and lower household income than other regions of the province [44].
Most students, regardless of gender and geography, felt that taking a school-based food and nutrition course was important; however, almost one half of males and one third of females had not taken a course. This may be because home economics classes, which traditionally teach this subject, are not available or mandatory in all Manitoba school divisions [45]. This reflects a trend observed across Canada and elsewhere, where there has been a decrease in home economics programs in recent decades [46].
One quarter of students were attempting to lose weight, although 50% more females were doing so than males. Conversely, almost three times the number of males were trying to gain weight. These patterns reflect normative gendered expectations of body shape, including thinness for women [18] and muscularity for men [47]. Adolescents have been shown to internalize social media messages reinforcing these ideals, which may have negative mental health effects in some [48]. It would also be important to explore the relationship between these beliefs and taking food and nutrition courses.
Body mass index varied regionally, with significantly more students residing in northern and rural communities classified as overweight compared with urban students. This may be in part due to reduced access to healthy foods, and higher food prices [49]. Healthier foods such as milk, fruit and vegetables have higher transportation costs, which are reflected in the grocery aisle [50][51][52]. These foods may be replaced by 'other' foods (e.g., high fat/salt/sugar convenience foods, sugar-sweetened beverages) which are more shelf-stable and less expensive than fresh alternatives, yet contribute to poor diet and weight gain. This was reflected in the lower HEI-C scores for those classified as overweight and obese.
Although adolescents have a reputation for excessive sleeping, almost half the participants felt they did not get enough sleep regularly with fewer females than males regularly getting more than eight hours of sleep per night, and almost one in ten getting less than six hours. These results are consistent with other studies that have found higher sleep deprivation in adolescent girls [53]. Screen activities and homework were the most frequently reported reasons for poor sleep in this study, which is consistent with other reports [54]. Poor sleep habits in adolescents are associated with poor health outcomes and health behaviours including poor dietary choices [55].

Limitations
There are several limitations with this study. The sample was not representative of all adolescents in Manitoba due to the complex nature of provincial school board structures, and the voluntary nature of participation at the school division, school, classroom and student levels. It was also limited by those who chose to come to school on the day of data collection. Self-reported dietary intakes are subject to under-reporting through the omission of food intake [56] and portion sizes may be inaccurate due to reliance on memory [57]; however, the use of a multi-pass technique in the WEB-Q tool including photographic images for portion sizes and prompts for complementary foods helped address these limitations [30]. The BMI calculations used self-reported height and weight, which have been found to be misreported in surveys; therefore, those in the overweight and obese categories could be underestimated [58]. Further, the food behaviours and health indicators used are not exhaustive. Other variables could provide more in-depth context into youth nutrition and health including physical activity and screen time. Finally, the study did not include schools in First Nation communities, where food costs and unemployment are high and incomes are low compared to other communities in Manitoba [59,60]. Consequently, the results presented for the north region may over-or underestimate some variables. There was, however, participation by schools from all regions of the province, and the results are consistent with those observed in other Canadian regions [61]. Despite these limitations, the study did include a large sample from three regions of the province. Dietary intakes estimated through the validated tools are appropriate for the types of analyses conducted. Further studies should examine additional behaviours impacting diet quality, and qualitative studies could provide further insights into youth experiences of food, nutrition and eating, and contribute to appropriate intervention development.

Conclusions
This study indicates that food and health behaviours were associated with diet quality in the study population; however, these were mediated by gender and the region of residence. Similar patterns were observed for many behaviours across participant subgroups (gender and region), and average HEI-C scores were poor, suggesting universal strategies to improve food behaviours and diet are required; however, the results also demonstrated that some targeted interventions may be necessary based on gender and region. In particular, strategies to address food insecurity and support school-based feeding programs in the north should be considered. Any intervention strategies need to be supported by multi-faceted policies and extend across sectors including education, food systems, health, social welfare and digital media [62].  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The datasets generated during and/or analysed during the current study are not publicly available due to the nature of the research and study participants not agreeing for their data to be shared with the public. Queries about data and materials should be directed to joyce.slater@umanitoba.ca.