Chapter 1: The Health of Canada's Young People: a mental health focus – Introduction

Introduction

by William Pickett, Matthew King, and John Freeman

The HBSC survey

The Health Behaviour in School-aged Children (HBSC) Study is a cross-national research study conducted in collaboration with the World Health Organization (WHO) Regional Office for Europe. The study aims to increase understanding of mental and physical health and their determinants in populations of young people. It involves health surveys conducted with students in classroom settings, with a focus on the early adolescent years (ages 11-15). HBSC is administered every four years following a common research protocol.

The HBSC survey was first developed in 1982 by researchers from three European countries. The project has since expanded to include 43 participating countries and regions. HBSC is now coordinated by a multi-disciplinary network of researchers from Europe, North America and Israel. Each country's research network consists of a Principal Investigator(s) and an affiliated research team. The most recent survey, the eighth in the cross-national series, was conducted in 2010. This represents the sixth cycle of the HBSC survey in Canada. A total of 26,078 young Canadians from 436 schools participated in 2010.

The international HBSC network represents a strong and organized research collaboration. All country investigators contribute to the development of the study and provide expertise on a variety of focused health topics. Research that involves the sharing of theory and skills across disciplines is encouraged within the network. A number of HBSC country teams are also involved in such efforts as the Schools for Health in Europe. Internationally, findings from HBSC are now routinely used by organizations such as the European Union and European Commission (2009), the United Nations International Children's Education Fund (UNICEF, 2010) and the Organization for Economic Cooperation and Development (OECD, 2009) to provide evidence in support of cross-national health promotion initiatives.

In terms of its central coordination, the Child and Adolescent Health Research Unit, University of Edinburgh in Scotland, is the International Coordinating Centre (ICC) of HBSC. It is responsible for the coordination of all international activities within the HBSC research network including the production of HBSC survey protocols and international reports; planning and organizing the network's semi-annual meetings of researchers; facilitating network communications; and acting as a central resource centre. International data collection is managed at the International Data Centre (IDC) at the University of Bergen, Norway. The IDC is responsible for coordinating all data management activities pertaining to the international data files, for both current and past cycles of the HBSC survey. HBSC's collaboration with the study's primary partner, the WHO Regional Office for Europe, creates opportunities for the transfer of knowledge generated by HBSC to participants in the WHO system.

Purpose and objectives of HBSC

The main purpose of the HBSC is to inform and influence health promotion and health education policy and programs at national and international levels, as well as to increase understanding of young people's health and well‑being.

Core objectives of the Canadian HBSC initiative are as follows:

1. To conduct national and international research on health behaviour, health and well‑being, and the social contexts of school-aged children.
2. To contribute new theoretical, conceptual, and methodological knowledge pertaining to these areas of research.
3. To compare health experiences among young people in Canada with those in other HBSC member countries.
4. To disseminate findings to the relevant audiences including researchers, health and education policy makers, health promotion practitioners, teachers, parents and young people.
5. To develop partnerships with other agencies who deal with adolescent health, in order to support health promotion efforts with populations of school-aged children.
6. To inform policy making and program development.
7. To contribute national expertise on health behaviour and on the social determinants of health in school-aged children.

Theory underlying HBSC

HBSC researchers believe that young people's health should be considered in its broadest sense, encompassing physical, social and emotional well‑being. As per standard and widely accepted definitions, health is viewed as a resource for everyday living, and not just the absence of disease.

HBSC researchers are intentional about considering positive aspects of health, as well as risk factors for future ill health and disease. This theoretical approach is informed by modern theories such as "assets-based approaches" to adolescent health research (Scales, 1999), in which factors that positively and negatively influence the health of youth populations are systematically examined.

In addition, HBSC research is almost always based upon population health thinking that focuses upon determinants of health, defined quite broadly (Health Canada, 1996). Among youth, these determinants include characteristics of home, school, peer group and neighbourhood settings, as well as socio-economic and behavioural influences. This theoretical approach suggests that health is influenced by the interaction of individual behaviours and environmental factors within these contexts.

HBSC Canada

Within Canada, the research team affiliated with HBSC consists of six independent scientists, the Canadian project manager, project staff, and graduate students who work with the investigators. Two Principal Investigators (J. Freeman, W. Pickett) oversee the work of the Canadian team. The Canadian team includes specific expertise from the fields of education, clinical psychology, kinesiology, and epidemiology. Additional members of the team from the Public Health Agency of Canada (the main national funder of the survey) provide specific expertise related to youth health policy. HBSC Canada and its members are active contributors to the development of the HBSC Study Protocol, national and international reports, with individual scientists having their own research programs that focus on the HBSC data.

HBSC Canada participates in public education activities and supports research initiatives at the international, national and provincial/territorial levels. In 2009, a new partnership was established with the Joint Consortium of School Health (JCSH), to help increase the relevance and impact of HBSC research within Canada. An additional partnership was formed with Health Canada to support expanded samples for the survey in six provinces (BC, AB, SK, ON, QC and NL) and a census of eligible young people in the three Territories (YK, NWT, NU).

Canadian HBSC scientists hold additional operating grants from the Canadian Institutes of Health Research, the Heart and Stroke Foundation and other agencies to support their multi-disciplinary research activities.

Canadian reports

Following completion of each survey cycle in Canada, the Canadian HBSC team has produced a major report that provides an overview of the national survey findings. Recent national reports have focused on: (1) healthy settings for the health of young people (2005-06) (Boyce et al., 2008); (2) a general profile of the health of young Canadians and factors that influence health (2001-02) (Boyce, 2004); and (3) trends in the health of Canadian youth (1997-98) (King et al., 1999). The report from the current 2010 cycle focuses on mental health in young Canadians, consistent with national priorities in the Public Health Agency of Canada and in our country as a whole.

Methods

The student questionnaire

The student questionnaire represents the core source of information in the HBSC survey. These questionnaires were administered to school classes, typically by teachers, and were filled out by individual students during one 45-70 minute classroom session. Survey items covered a wide range of topics pertaining to health and its determinants in populations of young people. Almost all of the questions could be answered by checking off a response option to the question, as opposed to items with more open-ended response categories.

Internationally, the student questionnaire was developed in a collaborative fashion by HBSC researchers. Questionnaire content and the protocol surrounding its administration were ratified at semi-annual research meetings of the research network. In Canada, there are three sets of items that comprise our national questionnaire; 1) "Mandatory Items" that all countries use; 2) "Optional Package Items" that focus on particular aspects of adolescent health, each used by only some of the participating countries; and 3) additional items developed specifically for the Canadian survey. The optional packages and Canada-only items used in the 2010 questionnaire were selected via a collaborative process. Factors considered during this process were the research interests of the HBSC Canada research team, input from core staff at the Public Health Agency of Canada and Health Canada, and priorities indicated by the Joint Consortium for School Health (JCSH). As a principle, efforts were also made to retain a core set of items on each survey to facilitate the monitoring of trends over time. There were two versions of the Canadian questionnaire, one for Grades 6, 7, and 8, and the other for Grades 9 and 10. The Canadian questionnaire was made available in English, French, and Inuktitut. The research ethics of the study were granted clearance by Research Ethics Boards from both Queen's University and PHAC/Health Canada.

Three levels of consent were required before a student could participate in the HBSC survey. School jurisdictions of the sampled schools were first approached for permission to invite their schools and students to participate. At the second level, school principals were approached to participate. At the third level, both active parent consent (students were required to return a signed consent form to participate) and passive parent consent (students were permitted to participate if they did not return the parent consent form indicating the parents refused permission to participate) were used. As per the ethics agreements at Queen's University and with Health Canada, the participating school jurisdictions and schools selected the consent type that was consistent with local norms.

In most classes, students were given instructions for completing the questionnaires at the beginning of the session. They completed the surveys individually at their own pace. Teachers were given the option to read through the survey question by question and have the students follow along at the same pace. This methodology was to be employed only in classes where literacy level was a concern. Students were guaranteed anonymity by sealing their unsigned survey in an envelope.

One exception to the standard survey administration process was in Yukon. There, the Yukon Bureau of Statistics administered the survey. Two experienced researchers from this Bureau travelled to schools in each community to administer the surveys using a standard methodology. In Grades 6 and 7, the survey was read aloud to accommodate the varied comprehension levels of students. For students in the higher grades the surveys were filled in at the students' own pace under the supervision of the researchers. These methods were complementary to those used in other provinces and territories, to ensure that the overall study methodology was not compromised.

The student sample

The international HBSC Network requires samples of young people aged 11, 13, and 15 years, to achieve target sample sizes for three developmental periods. In Canada, the HBSC research team sampled students in Grades 6 through 10. The Canadian sample was enlarged to approximately 26,000 students in 2010. Past cycles had averaged 7,000 to 10,000 students. This expansion was undertaken to achieve representative samples for 9 of Canada's 13 provinces and territories.

Provincial Samples. In each of the provinces, a systematic, two-stage cluster sample approach was used to select whole classes of students to participate in the study. A list of schools within eligible and consenting school jurisdictions was created initially, and then schools in the sample were selected for study from this list. The number of classes in specific schools was estimated based on the grades in the school, the numbers of teachers, the total enrollment, and enrollment by grade, while accounting for known variations in class structure by province. Classes were given an approximately equal chance of being selected. They were ordered on the sample lists according to school jurisdiction, language of instruction, public/Roman Catholic designation, community size, and community location within a province. Classes were proportionally distributed according to these characteristics. In most provinces one or two classes per grade were targeted in each school selected for study. Directions were provided for school administrators to randomly select the participating classes in their schools. The 2010 sample does not include students from Prince Edward Island or New Brunswick. Although the inability of these provinces to participate was unfortunate, their exclusion had little impact on national estimates. In addition, private and special schools including on-reserve schools were also not included to maintain consistency with past survey cycles.

Territorial Samples. For each of the three territories the sample consisted of all students in Grades 6 through 10. All eligible students were invited to participate. This sampling frame represented an attempt to obtain a census for the full student population in this range of grades, excluding private and special schools.

Response. At the student level, approximately 77% of the estimated students in the sample participated in the study. Fewer than 10% of students declined to participate or spoiled their questionnaires intentionally. For the remaining non-participants, the most common reasons for student non-participation were attributable to: failure to return consent forms; failure to receive parental consent; or absence on the day of survey administration.

The 2010 HBSC survey was administered in 436 Canadian schools. Table 1.1 provides details surrounding the national sample of 26,078 students. As mixed grade levels might exist in some classrooms selected for study, some Grade 5 and Grade 11 students were included. Table 1.2 further describes the number of participating schools, then students across the provinces and territories.

The administrator questionnaire

A second source of data for the 2010 HBSC was the Administrator questionnaire, an eight-page survey instrument to be completed by principals or their designate in each of the 436 schools that participated in the survey. The survey covered a basic description of the school and its student population, and individual modules on specific topics that were specified as priorities by HBSC Canada, the Public Health Agency of Canada, Health Canada, and the Joint Consortium for School Health. The latter included modules on: physical activity, school facilities, healthy eating, the school (social) climate, and neighbourhood factors. Administrator questionnaires were completed by representatives from 407 of the 436 (93%) participating schools.

Other data sources

Two additional sources of data were available in the 2010 cycle of the HBSC. Members of the HBSC Canada research team had secured operating grants from the Canadian Institute of Health Research to develop in-depth descriptions of the neighbourhoods surrounding participating Canadian schools. Each school was located on a map using address and postal code information, and characteristics of the area surrounding the school (using a 1 km to 5 km circular buffer) were abstracted using geographic information systems and data from the Canada Census of Population (Statistics Canada, 2006) and commercially available databases. More details about these data appear in Chapter 6 (Neighbourhoods).

Input from young people in Canada

The final source of information used in this report was direct input from young people. In March 2011, with the support of the Public Health Agency of Canada, the Students Commission, lead of the Centre for Excellence in Youth Engagement was contracted to host a 2-day youth engagement event. Invitees included approximately 20 young people in the HBSC age range from across Canada, adult participants from the federal government and the Joint Consortium for School Health, and researchers from HBSC Canada. The young people who were chosen to participate included a cross-section of boys and girls in the appropriate age range. They "represented" a diversity of provinces, community sizes, ethnic backgrounds, and languages.

A major focus of this youth engagement event was to obtain thoughts and insight from the young people on the core findings of the national report, and in particular the findings pertaining to mental health. This youth consultation provided an opportunity for HBSC researchers to consider these interpretations from young people when writing up the findings for each chapter. The young people were also asked to provide their opinions about the contribution of various social environments (home, school, peer group, and neighbourhood) to mental health outcomes, and the relative importance of other major health topics that are presented in this report. These views have been included throughout the report.

Approach to Analysis

Descriptive analyses

The vast majority of survey estimates presented in this report are proportions in simple bar-chart format, broken down by age and gender. The data from one response category (or combination of response categories) are typically presented. Ideally, confidence intervals should be provided for each of the survey estimates, providing the likely range of values to be found in the population being considered. This approach is not practical for a report of this size. However, due to the large sample size available, confidence intervals for virtually all overall estimates (stratified by gender) would be within plus or minus 3 percentage points, and similar estimates would be plus or minus 4 percentage points within strata defined by grade and gender. These estimates account for the clustered sampling design. The primary purpose of the descriptive analysis is to highlight general patterns. The statistical significance of each association has not been presented, given the large sample size (26,000 students) when working with the entire data set. With such a large sample size, the vast majority of coefficients, no matter how small, would be expected to be statistically significant and the presentation of probability values would not add value to the analyses. The same principles apply to the description of trends in proportions observed over time, as is presented for a few items in each chapter.

Survey weights

Results presented from the student data set are weighted. Each province or territory data set is weighted within the national file such that student responses for a particular province or territory contribute to the national results in proportion to the actual student population within the national grade group population. Weights are calculated for each of Grades 6 through 10 independently. Effectively, provinces and territories that are over-represented in the student data file are given a weight of less than 1, while provinces that are under-represented in the data set are given a weight greater than 1.

Composite measures

Several composite measures have been created for the purposes of examining relationships in the report. The composition of these measures is detailed when used later in the report.

A number of new items have been introduced into the HBSC questionnaire for the 2010 cycle of data collection, with the express purpose of having meaningful and valid measures of mental health. For this report, four standard indicators of health have been created based on underlying theory and on factor and reliability analyses: two negative indicators (emotional problems, behavioural problems) and two positive indicators (emotional well‑being and prosocial behaviours).

Three composite measures are introduced in the School chapter: school climate, teacher support, and student support. There is also a parent relationship measure introduced in the Home chapter. In the creation of composite measures, semi-continuous measures are created through summing up the responses for several items or by taking the mean across several items. The resulting distributions have been collapsed into thirds (e.g., low, middle and high) for ease of presentation. It should be understood that these three categories are relative measures rather than absolutes. For example, students in the "low" category on the parent relationship measure are considered to have a poorer relationship with their parents than those scoring "middle" or "high," but they are not necessarily considered to have a poor relationship with their parents in the absolute sense.

Relationships with mental health indicators

The major theme of this report is the mental health of young Canadians. In each chapter, as well as profiling recent data pertaining to its main focus, authors explore relationships between different contextual factors, health behaviours, and health outcomes and the four mental health indicators.

The analytical approach used in these analyses involves the development of models using a specific form of regression called logistic regression. Each model examines the potential relationship between specific variables as independent predictors and specific mental health indicators as dependent variables, while stratifying by gender, and controlling statistically for grade and a single measure of socio-economic status (perceived material wealth). Coefficients from each of the models are used to produce prevalence estimates for the mental health indicators in relation to each predictor of interest, controlling statistically for the other factors. For simplicity of presentation, the probability of being in the "top third" of the indicator is modeled and presented graphically. Similar to the descriptive analyses, formal tests of statistical significance are not presented, although because of the robust nature of the sample, virtually all noted associations achieve significance by conventional statistical standards.

An example of the findings from one such regression model is provided in Figure 1.1. It summarizes the relationship between body weight and emotional well‑being, separately for boys and girls, while mathematically accounting for any group differences in grade and socio-economic status. For boys, this model shows that the probability of being in the highest ("top third") group of young people classified according to their emotional well‑being is 43% for normal weight boys, 39% for overweight boys, and 34% for obese boys. There is a clear relationship between reporting a higher body weight and lower emotional well‑being. A similar pattern exists for girls, although the overall percentages of girls in all weight categories reporting high emotional well‑being is much lower than for boys. Each of these associations is statistically significant, although formal statistics and confidence intervals are not provided in order to simplify the presentation.

An obvious complication surrounding interpretation of these relationships is the issue of causality. It is unclear in many situations whether the health behaviour leads to the mental health outcome or the mental health outcome leads to the health behaviour. Most likely, there is reciprocal causation with regard to health behaviours and mental health. Reciprocal causation suggests that we need a multi-pronged approach to the issue, such that we ignore neither health behaviour nor mental health outcome under the likely false impression that ameliorating the one will directly have positive effects on the other.

1.1 Sample findings from a multivariable regression analysis: relationship between emotional well‑being and body weight in boys and girls (%)

[Click to enlarge Figure 1.1]

Organization of the report

This report is organized as follows. Chapter 2 provides a national profile of mental health indicators for young people, and provides focused results pertaining to the major theme of the present report. Chapters 3 through 6 provide descriptive overviews of the environments that impact the health of young people, including home, school, peer and neighbourhood. This set of chapters is followed by a series of chapters that are organized conceptually around topics that remain priorities for the public health system in Canada, including injury (Chapter 7), healthy living (Chapter 8), healthy weights (Chapter 9), risky behaviour (Chapter 10), and bullying (Chapter 11). Each of Chapters 3 to 11 contains a statistical review of health measures, a brief summary of how selected indicators relate to the four mental health outcomes, some perspective from young people, and a summary and implications section. The final chapter provides a summary of the core messages for public health that emerged from these national analyses.

References

• Boyce, W.F., King, M.A. & Roche, J. (2008). Healthy settings for young people in Canada. Ottawa: Public Health Agency of Canada.
• Boyce, W.F. (2004). Young people in Canada: Their health and well‑being. Ottawa: Health Canada.
• European Commission (2009). Questions and answers on the EU Youth Health Initiative. Brussels: Europa Rapid Press Releases. Memo 09/326. Available at: http://europa.eu/rapid/pressReleasesAction.do?reference=MEMO/09/326&format=HTML&aged=0&language=EN&guiLanguage=en Accessed May 11, 2011.
• Health Canada. (1996). Strategies for population health – investing in the health of Canadians. Probe, 30:106-108.
• King, A.J.C., Boyce, W.F. & King, M.A. (1999). Trends in the health of Canadian youth. Ottawa: Health Canada.
• Organization for Economic Cooperation and Development (OECD). (2009). OECD Family Database (2009).
• OECD: Directorate for Employment, Labour and Social Affairs. Available at: http://www.oecd.org/document/4/0,3746,en_2649_34819_37836996_1_1_1_1,00.html, Accessed April 3, 2011.
• Scales, P.C. (1999). Reducing risks and building developmental assets: Essential steps for promoting adolescent health. Journal of School Health, 69:113-119.
• UNICEF. (2010). The Children Left Behind: A league table of inequality in child well‑being in the world's rich countries. Innocenti Report Card 9, UNICEF Innocenti Research Centre, Florence.