Should public health interventions aimed at reducing childhood overweight and obesity be gender-focused?

The Raising Healthy Eating and Active Living Kids in Alberta survey http://​www.​REALKidsAlberta.​ca aims to evaluate a comprehensive initiative by Alberta Health and Wellness to promote healthy body weights among children and youth through surveys in 2008 and 2010 among grade five students who are primarily 10 or 11 years of age. The present study reports on the 2008 observations. The study employed a one-stage stratified random sampling design. The sampling frame include all 1505 schools with grade five students, with the exception of private schools (4.7% of all Alberta children attend such schools), francophone schools (0.6%), on-reserve federal schools (2.0%), charter schools (1.7%), and colony schools (0.8%) [25]. Schools were stratified into three geographies: 1) urban: Calgary and Edmonton; 2) towns: other municipalities with more than 40,000 residents; and 3) rural: municipalities with less than 40,000 residents. Schools were randomly selected within each of these strata to achieve a balanced number of students in each stratum. Of the 184 invited schools, 148 (80.4%) participated in the study. These schools were attended by 5594 grade five students who received an envelope with a parent survey and consent form to take home. Of the 3758 (67.2%) students that returned completed consent form to school, 3645 (97.0%) had received parental consent to participate. Of these students, 3407 were present when evaluation assistants visited the school to conduct the survey, 6 students declined to participate, and 20 absent students completed and mailed their surveys, resulting in 3421 participating students (61.2% of total student population in those schools). These students completed the Harvard Food Frequency Questionnaire for Youth and Adolescents (FFQ) which is suitable for grade five students [26]. Students further completed a short survey on activities and behaviors, and had their heights and weights measured.

Standing height was measured by trained evaluation assistants to the nearest 0.1 centimeter without shoes using stadiometers and body weight to the nearest 0.1 kilogram on calibrated digital scales. Body mass index was calculated by dividing weight (in kilograms) by height (in meters) squared. Overweight (including obesity) and obesity were defined using the body mass index cut-off points for children and youth by the International Obesity Task Force [27].

Parents and students responded to activity questions on: a) travel to and from school; b) time spent to get to and from school; c) frequency of child's activities outside of school hours; d) activities at morning and lunch recess in the past seven days; and e) frequency of involvement in sports and physical activities in the past seven days. These questions, totaling 29 items, were by large adopted from the Physical Activity Questionnaire for Children (PAQ-C) which has previously been validated and demonstrated high reliability [28]. The 29 items were the basis of a composite score ranging from 1 to 6. Students with a score exceeding 3 were classified as 'physically active'. Students also provided answers on the frequency (< 3 times/week or ≥ 3 times/week) during which they were engaged in physical activities with a coach ("sports with coach"), without a coach ("sports without coach"), or with parents ("sports with parents") out of school hours. Parents answered questions on the number of hours (≤ 2 hours/day or > 2 hours/day) that their child spent playing videos games, watching television and using the computer out of school hours ("screen time"). These questions have been previously validated and were adopted from the National Longitudinal Survey of Children and Youth [29].

Based on students' responses to the FFQ and nutrient information from Canadian food tables [30], we calculated the nutrient intake for each of the participating students. Specifically, we calculated percentage of energy in their diets originating from dietary fat, and calculated whether students complied with Canadian guidelines of 6 servings of vegetables and fruits per day [31]. The questionnaires for students further included 6 questions on types of foods they would eat or purchase (see table 1).

Finally, the questionnaire completed by parents provided information on household income (≤ $50,000; $50,001-$75,000; $75,001-$100,000 and > $100,000) and parental education (secondary school or less, community college and graduate university).

We applied multilevel logistic regression methods to assess gender differences in overweight, physical activity and nutrition while acknowledging the nested structure of the data whereby student observations nest within those of schools. All analyses were weighted to account for the design effect such that estimates apply to the grade five student population of Alberta. We adjusted for the confounding potential of parental education attainment, household income and residency. Analyses pertaining to dietary outcomes ('having 30% or more of dietary energy from dietary fat' and 'eating 6 or more servings of vegetables and fruits daily') were further adjusted for calorie intake as is recommended for FFQ data [32]. Missing values for parental education attainment, household income were treated as separate covariate categories. Stata Version 10 (Stata Corp, TX, USA) was used to perform the statistical analyses. This study, including data collection and parental informed consent forms, was approved by the Health Research Ethics Board of the University of Alberta.

We reported a slightly higher overweight prevalence among boys than girls. This observation is consistent with other Canadian and US studies that reported slightly lower prevalence of overweight among girls [11, 34‐37]. When focusing on specific geographies, we found that boys residing in towns were substantially more likely to be overweight as compared to girls. This reinforces the need to considering geographic discrepancies in addition to gender-focused approaches in the development of preventive strategies.

Various studies have shown that boys engaged more in physical activity than girls [38‐40]. Physical activity counterbalances excess energy intake and the risk for overweight. This risk seems to be particularly present in rural schools and calls for programs to promote physical activity, particularly among girls during their transition from childhood to adolescence and adulthood, as longitudinal studies have shown a drop in activity levels during these years [5]. Specifically, a review of the literature suggested annual declines in physical activity of up to 2.7% among boys and 7.4% among girls between the ages of 10-17 years [5].

Our study population constitutes grade five students who are primarily 10 or 11 years old and primarily preadolescent. Adolescence is a period of physical, psychosocial, cognitive, and emotional change that may be linked to physical maturity and may affect physical activity behaviors [41]. Thomson et al. [42] demonstrated that gender differences in physical activity diminished when aligned on maturational age, suggesting that sexual maturation may be intricately involved in the adolescent decline in physical activity. As girls mature approximately 2 years before boys [43], relative more girls than boys are expected to have entered sexual maturation in our study population. This may have contributed to the higher physical activity levels among boys relative to girls observed in this study. A US-based study suggested that early-maturing girls are twice as likely to be obese compared with early-maturing boys [44]. If this were to apply to our study participants, we should anticipate changing gender differentials in overweight in the year to come.

We observed gender-difference in dietary intake and eating behaviors that parallel observations in the relatively few studies that focused on gender and nutrition among children and adolescents [4, 34, 36, 45, 46]. We observed higher calorie intakes among boys relative to girls, which is consistent with another Canadian study [4] and the fact that energy requirements among boys are slightly higher than among girls [37], though are inconsistent with studies reporting an absence in gender difference in children's energy intake [45, 47]. The limited number of regarding gender comparisons in the literature likely reflects the challenge of collecting and analyzing dietary information among children.

Differences in overweight and health behaviors between boys and girls may result from differences in biology (sex differences), from differences assumed to be due to society or culture (gender differences), or a combination of the two [6, 48]. With respect to health in general, there is support that integrating gender considerations into interventions will improve their effectiveness [24], though it is acknowledged that these gender considerations are often difficult to translate into health promotion and preventive programs [49]. In the Canadian public health care system, patient care absorbs the far majority of health sector resources, with less than 3% of health spending allocated for resources towards health promotion and primary prevention [50]. Therefore it is of utmost importance to invest these limited resources in effective preventive activities [51]. The lack of translation of knowledge about gender differentials in health and health behaviors into health promotion interventions may lead to a weakened potential for success and herewith inefficient use of resources [52]. Thus, studies into gender differences are important to increase our understanding of mechanisms of behavioral change to allow optimizing health promotion and interventions [53]. The present study contributes to this understanding for the prevention of childhood overweight and obesity. The results reinforce the needs for gender-focused health promotion and provide direction on how to deliver that.

Strengths of the present study include its large representative sample, measured heights and weights and its response rate that is considered high for school-based research. A further limitation relates to the cross-sectional design that necessitates caution with respect to interpretations related to directionality and causality. We had used the body mass index cut-off values by the International Obesity Task Force to define overweight and obesity [27]. These are increasingly established and some studies showed that they provided similar definitions for overweight though more conservative estimates for obesity [54‐57].

Self report is prone to error. As in other studies, in the present study this may have lead to an over-reporting of "structured" activities and sport activities [41]. The use of objective measures of physical activity (e.g. pedometers) would have permitted us to more accurately estimate physical activity levels of students, although it is acknowledged that the use of such devices may be logistically and financially challenging in large population studies. The limitations of self report also apply to the assessment of dietary intake, although the Harvard Food Frequency Questionnaire for Youth and Adolescents is a extensively validated and used instrument [58]. Studies have shown that FFQ tend to underestimate energy intake when compared with measured energy intake [59, 60].