Constructing a regional adolescent health and wellness index for British Columbia, Canada

(1)

Constructing a Regional Adolescent Health and Wellness Index for British Columbia, Canada

Gina Chrissy Martin BSc, University of Victoria, 2008

A Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of

MASTER OF SCIENCE in the Department of Geography

(2)

Supervisory Committee

Constructing a Regional Adolescent Health and Wellness Index for British Columbia, Canada

by

Gina Chrissy Martin BSc, University of Victoria, 2008

Supervisory Committee

Dr. Peter Keller, (Dean of Social Science, Department of Geography)

Co-Supervisor

Dr. Les Foster, (Department of Geography, and School of Child and Youth Care)

(3)

Abstract

Supervisory Committee

Dr. Peter Keller, (Dean of Social Science, Department of Geography)

Co-Supervisor

Dr. Les Foster, (Department of Geography, and School of Child and Youth Care) Co-Supervisor

The purpose of this thesis is to construct an index of adolescent health and wellness for British Columbia, Canada, using the most recent available data. A three- round Delphi study is used in order to decide on what indicators to include in the index and each indicator’s relative weight. Spatial multi-criteria analysis (MCA) is utilized to combine the indicators into a single measure. The spatial MCA, technique for order preference by similarity to an ideal solution (TOPSIS) method was applied to the adolescent population as a whole and to examine male and female variation. This revealed that adolescent health and wellness is not experienced equally across the province. The Health Service Delivery Areas (HSDAs) Fraser South and Fraser North proved to have the greatest levels of adolescent health and wellness while the Northwest has the least. A rural/ urban gradient in adolescent health and wellness was revealed at the HSDA level. Male and female adolescents also experience health and wellness differently, with females achieving higher health and wellness across all HSDAs in the Province when directly comparing the two genders. The findings of this research are useful in informing

discussions of resource allocation for reducing inequalities and inequities and in order to target future research.

(4)

List of Tables

Table 1: Adolescent population (12-19), 2008 ... 6

Table 2: Summary of identified indicators, consolidated responses and number of participants who identified the indicator... 39

Table 3: Breakdown of panel participation... 43

Table 4: Results of round three Delphi survey ... 47

Table 5: Results of Krippendorff's α analysis... 50

Table 6: Results of ICC... 52

Table 7: Summary of adolescent health and wellness indicators and data sources ... 55

Table 8: CV of indicators under consideration for the index, n=14 ... 66

Table 9: Criteria weighting methods (n=number of criteria/ indicators)... 87

Table 10: Weighting of indicators ... 90

Table 11: Input statistical data for index... 92

Table 12: Combination rules for spatial multi-criteria analysis... 95

Table 13: Pearson’s correlation matrix of indicators in BCAHWI, N= 14 ... 98

Table 14: Data definitions and function in the index... 101

Table 15: Design of sensitivity analysis applied to both rank sum and rank reciprocal weights ... 110

Table 16: Comparison of BC adolescent health and wellness index scores by three different weights ... 113

Table 17: Pearson’s correlation coefficient of the three BCAHWI score using various weighting schemes ... 114

Table 18: Results of sensitivity analysis for rank sum (light grey shading indicates a change in ranking from the original) ... 117

Table 19: Results of sensitivity analysis for rank reciprocal (light grey shading indicates a change in ranking from the original) ... 117

Table 20: Patterns of clusters using Ward's method of hierarchical clustering (rank sum weighting) ... 119

(8)

(9)

List of Figures

Figure 1: Health Service Delivery Area boundaries ... 5

Figure 2 Adolescent percent of total population 12-19 years of age, 2008 ... 6

Figure 3: Map showing School District boundaries and Health Service Delivery Areas. 11 Figure 4: Delphi study administration process (adapted from Okoli & Pawlowski, 2004) ... 31

Figure 5: Participant panel by self identified position... 36

Figure 6: Steps taken in data analysis Questionnaire 1, Section 2 (round one)... 37

Figure 7: Screenshot of round three surveys... 46

Figure 8: Screenshot of TOPSIS selection method... 105

Figure 9: Results of BCAHWI using rank sum ... 115

Figure 10: Results of BCAHWI using rank reciprocal... 115

Figure 11: Results of BCAHWI using equal weights... 115

Figure 12: Results of BCAHWI using three different weighting schemes... 116

Figure 13: Map of 4 clusters produced by hierarchical cluster analysis... 120

Figure 14: Results of female BCAHWI, calculated using rank sum weights... 121

Figure 15: Results of male BCAHWI, calculated using rank sum weights... 122

Figure 16: Index score for females and males when the maximum and minimum values of both genders are applied ... 123

Figure 17: Relationship between BCAHWI (using rank sum weights) and the percent of the total population ... 126

Figure 18: Relationship between BCAHWI (using rank sum weights) and the population density rate ... 126

Figure 19: Relationship between BCAHWI (using rank sum weights) and the average family income, 2006 ... 127

Figure 20: Relationship between BCAHWI (using rank sum weights) and the incidence of low income in economic families, 2006... 127

Figure 21: Relationship between BCAHWI (using rank sum weights) and percent of students classified as rural/small town... 129

(10)

(11)

List of Abbreviations

AHP Analytical Hierarchy Process

AHS Adolescent Health Survey BC British Columbia

BCAHWI British Columbia Adolescent Health and Wellness Index BMI Body Mass Index

CA Census Agglomeration

CCHS Canadian Community Health Survey CMA Census Metropolitan Area

CV Coefficient of Variation CWI Child Wellbeing Index E Expert

GIS Geographic Information Systems HSDA Health Service Delivery Area I 1 Family Connectedness

I 2 Freedom from Abuse

I 3 Physical Activity

I 4 Healthy Diet

I 5 Freedom from Chronic Conditions (including mental health conditions) I 6 School Connectedness

I 7 Good Mental Health I 8 Positive Peer Influences I 9 Positive Adult Mentors

I 10 Adolescents Feeling They Are Good at Something I 11 Tobacco/ Alcohol Use of Teen Mothers

I 12 Healthy Weight

I 13 Literacy

I 14 Suicide

I 15 Illicit Drug Use

I 16 Adolescent Pregnancies

I 17 Community/Cultural Connectedness I 18 Residing Outside the Parental Home I 19 Educational Achievement

I 20 Adolescent Crime I 21 Self Rated Health

I 22 Tobacco Use

I 23 Housing and Neighbourhood I 24 Child Welfare Contacts

ICC Intraclass Correlation Coefficient MAUP Modifiable Areal Unit Problem MCA Multi-criteria Analysis

MCS McCreary Centre Society OWA Order Weighted Average

(12)

RAND Research and Development Corporation RST Rural Small Town

SAW Simple Additive Weighting

SD School District

SE Standard Error

SHIS Salutogenic Health Indicator Scale

TOPSIS Technique for Order Preference by Similarity to an Ideal Solution WHO World Health Organization

(13)

Acknowledgments

First and foremost, I would like to thank my supervisors, Dr. Peter Keller and Dr. Les Foster, for providing advice, support and encouragement throughout the research process. Their experience and guidance have been invaluable to me during these past two years.

I would also like to express my gratitude to the anonymous members who served on the panel of this study. Their time and participation formed the backbone of this research and their insights were crucial in the success of this project.

Special thanks to the data providers of this research- BC Stats, BC Perinatal Database, those involved in the wellness mapping project at UVIC and the McCreary Centre Society. In particular, I would like to express my extreme gratitude to Dr. Weihong Chen, Dr. Colleen Poon and Dr. Elizabeth Saewyc from the McCreary Centre Society, for their support and advice during this process and for presenting opportunities for students to share their research with each other.

Last but certainly not least I would like to thank my friends and family. I am particularly thankful to my best friends, Amy Lanthier, Coura Niang, and Stephanie Welters, for being the brightest part of my life and my partner Jed Long for his ongoing support. Importantly, thank you to my parents, Joseph Martin and Joanne MacDougall, for their unconditional love no matter what endeavour I am undertaking. I would also like to acknowledge my grandmother, Janet Currie, who served as a role model of kindness, caring and independence throughout my life.

(14)

1.1 Introduction

Up until the last decade, adolescent health inequalities have received less attention than those of adults and young children (Currie et al., 2008). In Canada there has been relatively limited research into the health of the adolescent population (Geddes et al., 2005). During the transitional time from childhood to adult status the adolescent is

expected to move from requiring adult monitoring to exercising self control and behaving in a socially responsible way (Dahl, 2008; Gaudet, 2007; Tonkin, 2005). Due to the distinctive health behaviours and concerns that affect adolescents, it is evident that this faction of the population deserves specific attention. This research seeks to create a composite index of adolescent health and wellness in British Columbia (BC) and to examine its geographical variation.

The ecological/population approach to health asserts that the environment in which one lives is linked with people’s health and wellness (Hills & Carroll, 2009). It recognizes that individuals are embedded within social, political and economic systems that shape behaviours and access to resources and has been the subject of renewed attention in health studies (Etches et al., 2006). The focus of this approach is on population level change, measured with numeric indicators of health and wellness, rather than on individual change and interventions (Roussos & Fawcett, 2000). Health and wellness indicators quantify behaviours or situations of a population over time, across groups, or across geographical units (Corbett, 2008). These indicators are used with the goal in mind

(15)

of reducing health and wellness inequalities and improving the health and wellness of the population under study (Etches et al., 2006). A goal of this research is to consider the unique issues that contribute to adolescent health and wellness when examining population health.

Furthermore, focus in population health studies has increasingly taken a wellness

perspective rather than an illness perspective. It has been established that health research should strive not only to examine negative but also positive indicators of health (Bringsen et al., 2009). This follows from the World Health Organization (WHO) definition of health as “a state of complete physical, mental and social well-being and not merely the absence of disease or infirmity” (WHO, 1948), beginning a societal shift that moved to looking at health from a holistic viewpoint rather than a purely physiological one. Wellness definitions have varied within the literature, but commonly measure states of positive health on a continuum and from a holistic viewpoint. Taking a wellness

perspective can be done by drawing on definitions of wellness to promote the generation and maintenance of population health (Miller & Foster, 2010).

This study seeks to draw on the WHO’s holistic definition of health and use a wellness perspective in order to examine adolescent health and wellness. This perspective is supported by growing research that has found that, rather than an emphasis on youth problems, focus should be given to elements of youth that help adolescents grow up healthy (He et al., 2004). The media often stereotype youth as at risk. This trend of examining negative aspects of adolescence leads policy and decision makers to focus on

(16)

problems that affect adolescents rather than supporting factors that make adolescents healthy and well (Cicognani et al., 2008). Positive elements of health and development are protective against risk taking behaviour and facilitate a healthy transition to adulthood (Smith & Barker, 2009). Looking at key factors of those who choose healthy lifestyles has been found to be important in understanding and fostering healthy populations (He et al., 2004; Moore et al., 2004). Although it is difficult to pinpoint what makes an

adolescent population healthy, it is evident that looking at what makes the population well in addition to what makes it unwell is important in getting a complete and accurate picture into the lives of adolescents (Moore et al., 2004).

It is the goal of this work to utilize geographical analysis, including visualization, to explore geographic variation of adolescent health and wellness using key indicators that are tailored to adolescents in the Province of British Columbia (BC), Canada, in order to create an index of adolescent health and wellness. Maps of health and disease have a long history of presenting data for visualization of complex geographical information for purposes including: hypotheses generation, surveillance, to highlight areas that appear at risk and to aid policy formation (Elliot et al., 2000). The power of using a population approach is that it lends to visualization of the similarities and differences between areas. Patterns can be seen and further interpreted. It has been noted that few studies examining adolescents have looked at multiple positive factors (Vesely et al., 2004). Indices can be used as a tool for combining multiple indicators of health and wellness and for ranking geographic areas. These composite indices have more explanatory power than single indicators (Boyle & Torrance, 1984; Bringsen et al., 2009; Frohlich & Mustard, 1996).

(17)

The results of the analysis presented in this thesis are valuable to inform policy makers in regards to adolescent health and wellness in BC for use in decision making (Etches et al., 2006; HCC, 2007). By learning from the areas with high levels of adolescent health and wellness, it is hoped that geographical health inequalities can be reduced and the

Province itself can become healthier (HCC, 2007).

1.2 Study Area and Demography

The Province of BC is the area under study. The data are examined at the Health Service Delivery Area (HSDA) for which data are readily available. The disadvantage underlying HSDAs is that they are relatively large geographic units. Further, two of the HSDAs (North Vancouver Island and North Shore/ Coast Garibaldi) correspond to non-

contiguous polygons. There are 16 HSDAs in BC. The HSDAs are administrative areas defined and legislated by the provincial ministries of health. They represent geographic areas of regional health authorities and are subject to change (Statistics Canada, 2009a). Figure 1 displays the area boundaries.

(18)

Figure 1: Health Service Delivery Area boundaries

In 2008, there were an estimated 442,663 people who fell into the age category of 12 – 19 living in BC. This accounts for approximately 1 in every 10 British Columbians (10% of the population of BC). Table 1 shows the proportion of adolescents (age 12 – 19) in each HSDA in order of the proportion of the population that adolescents make up (highest to lowest). It is possible to see that the regions in Northern BC have the highest proportion of adolescents while Vancouver has the lowest. Figure 2 illustrates this same information visually.

(19)

Table 1: Adolescent population (12-19), 2008

HSDA # HSDA Name Population 12 - 19 years of age % of Total Population

51 Northwest 9715 13.0

53 Northeast 8073 12.1

52 Northern Interior 16785 11.8

21 Fraser East 31673 11.5

23 Fraser South 75954 11.2

43 North Vancouver Island 13024 11.0

14 Thompson Cariboo 23384 10.6

33 North Shore/ Coast Garibaldi 28795 10.5

22 Fraser North 60885 10.4

12 Kootenay Boundary 8122 10.4

11 East Kootenay 8126 10.3

42 Central Vancouver Island 26241 10.2

13 Okanagan 34479 10.0

31 Richmond 18606 9.8

41 South Vancouver Island 32578 8.9

32 Vancouver 46223 7.3

British Columbia 442663 10.1

(Data source: BC Stats)

(20)

1.3 Data Availability and Limitations

In an ideal world it would be desirable to have individual data for each adolescent in a study area obtained through interview, self-reporting and administrative data. Access to such primary data (or releasing such confidential data), clearly is not feasible or realistic. In reality researchers usually work with secondary data consisting of existing sources of data collected by agencies and other researchers as part of other and usually larger studies. Collections of primary data are beyond the scope of this study.

Data sources that track and link adolescent health are limited. It has been identified that the availability of data sources is a challenge when examining child and adolescent health (Currie et al., 2008) and so without collecting primary data it is only possible to deduct information from a range of aggregated data sources. Boundary selection is forced by limitations in the data (Staines & Jarup, 2000). The following reports on secondary data available and accessible to incorporate into an index of adolescent health and wellness for BC.

1.3.1 The McCreary Centre Society Adolescent Health Survey, 2008

The McCreary Centre Society (MCS) Adolescent Health Survey (AHS) was undertaken in 1993, 1998, 2003 and 2008. These surveys of approximately 30,000 students in grades 7 to 12 in BC schools are collected to produce statistically reliable estimates at the HSDA level for each grade surveyed (Saewyc & Green, 2009). The AHSs are a rich dataset that cover a range of topics.

(21)

There are a number of considerations when using the AHS data. First, data are collected in schools and in the 2008 survey it was left up to each school district (SD) to decide on the consent procedure used prior to the survey (active or passive consent by the parents). Active (parental signed) consent vs. passive (parental notification) consent is argued by the MCS to have a significant effect on response rates to some questions in the survey which may impact statistical significance of subsequent analyses (Saewyc & Green, 2009). Moreover, the MCS uses a cluster -stratified sampling design in that the data are collected in randomly selected classes. Consequently individual responses are clustered at the classroom level. Using SPSS Complex Samples software (www.spss.com), accounts for the cluster stratification and differential population sizes and scales estimates to the enrollment of each school district.

Not all SDs agreed to participate in the 2008 AHS. Of the school districts 50 out of 59 or approximately 85% participated. This implies incomplete data across BC (Provincial Health Officer, 2008). The 50 school districts that did participate contained 92% of all students in grades 7 to 12 enrolled in BC public schools at the time of the survey

administration (Saewyc & Green, 2009). Non-participation of school districts occurred in two HSDAs, leaving two geographic gaps in the dataset. No SDs in the Northeast

participated. Fraser East only had two small SDs that took part, leaving this area under represented. In order to account for this the MCS reports combines estimates of Fraser South with Fraser East when reporting the estimates at the HSDA level. This is not useful when combining various data sources into an index, consequently Fraser East is not used.

(22)

Fraser South on its own has a large enough sample size to report. Data from the AHS are available by gender.

1.3.2 The Canadian Community Health Survey 2007/2008

The Canadian Community Health Survey (CCHS) is a nationwide cross sectional survey undertaken by Statistics Canada that is designed to examine health outcomes at the HSDA level (Bell et al., 2007a; Provincial Health Officer, 2008). The CCHS targets persons aged 12 or older who live in a private dwelling (Statistics Canada, 2009b). Persons not surveyed are: those living on Indian Reserves or Crown lands, those in institutions, fulltime members of the Canadian Forces and residents of very small remote regions. Statistic Canada states that the CCHS covers approximately 98% of the

Canadian population age 12 and over (Statistics Canada, 2009b).

In BC, the sample size for the 2007/2008 CCHS was 14,651 for those who agreed to share their responses with the provinces (McKee et al., 2009). The sampling was

designed to ensure an over-representation of those aged 19; if the interviewee was 12-15 then verbal permission had to be granted by the parents/guardians. In order to ensure privacy during the personal interviews or on the phone, if privacy was suspected to be breached by the interviewer (i.e. suspecting another person listening in) then the survey was coded as refusal. All items, for this age group, regarding income and food security were answered by parents/ guardians at the end of the survey. Approximately half of the surveys were conducted in person and the other half were conducted over the phone (Statistics Canada, 2009b). Data from the CCHS are available by gender.

(23)

1.3.3 BC Ministry of Education School Satisfaction Survey 2007/2008

Beginning in the 2000/2001 school year, the BC Ministry of Education has been undertaking a School Satisfaction Survey of students at the grade levels of 3/4 (grade 3 students are only surveyed if there are no grade 4 students within the school),7, 10 and 12. These surveys are carried out annually and are provided to all students, but

participation is optional. Still there is consistency in the results over time, suggesting that there is little bias that results from participation rates. For the 2006/2007 survey there was a 98.9 percent participation rate of eligible public schools (Provincial Health Officer, 2008).

The purpose of this survey is to measure various issues relating to the school environment, including health and safety. These data are deemed to be a complete population dataset (Provincial Health Officer, 2008). Data are available at the SD level; there are 59 districts in BC. The districts, in some places, nest logically inside HSDAs while in other areas there is considerable discrepancy in boundary overlap. Figure 3 illustrates this.

(24)

Figure 3: Map showing School District boundaries and Health Service Delivery Areas

1.3.4 BC Stats

BC Stats has census data and BC ministry data, from various years, much of which has been aggregated to the HSDA level. Crime and educational information is included in this. These data are not available by gender. All area boundary shape files used for mapping in this thesis were provided by BC Stats.

(25)

1.3.5 Other Data Sources

The BC Ministries of: Healthy Living and Sport, Children and Family Development, and Education as well as the BC Perinatal Database have relevant data available on

adolescents throughout the Province.

1.4 Research Goals and Questions

There are three main goals that guide this project:

1. to use geographical analysis and visualization capitalizing on geographic information systems (GIS) capacity and other statistical analysis to examine adolescent health and wellness across the Province;

2. to establish an adolescent health and wellness index; and

3. to show geographic variations of adolescent health and wellness to inform policy and decision making.

There are four major research questions that guide this project:

1. What are the key indicators of adolescent health and wellness in BC, Canada? 2. Can a spatial multi-criteria analysis (MCA) contribute to the understanding of the

geographic variation of adolescent health and wellness in BC, Canada? 3. What is the pattern of adolescent health and wellness in the Province, as

established through a BC Adolescent Health and Wellness Index (BCAHWI)? 4. Can this approach be applied to examine male and female patterns in adolescent

(26)

1.5 Literature Review:

A review of relevant literature is undertaken to provide a foundation for this study. Definitions of adolescence are briefly explored in Section 1.5.1 to support that

adolescence is a unique timeframe in the human lifecycle and that it requires indicators that reflect its unique nature. An overview of contemporary adolescent health and wellness issues will be presented in Section 1.5.2. Section 1.5.3 will examine the use of indices in health research. The literature review provides a foundation to guide and develop the methodological and theoretical techniques used in this study.

1.5.1 Defining Adolescence

Clear age boundaries are helpful to provide starting points and cut offs for policy and research. Defining adolescence becomes particularly challenging because it is marked by dynamic development and has many biological and social influences (Beaujot & Kerr, 2007; Dahl, 2008; Gaudet, 2007). There is a strong case in the literature that adolescence begins with the onset of puberty (CPS, 2008; Dahl, 2008; Tonkin, 2005). Typically puberty occurs at 12 years of age in females and 13 in males. Puberty is a period of significant transformation marked by rapid physical growth, onset of sexual maturity and sexual interests and may commence mid-childhood, sometimes as early as age 9 in girls. During puberty strong emotional influences affect the capacity for self regulation and impact decision making (Dahl, 2008; Tonkin, 2005).

The WHO defines the years of adolescence as ages 10-19 (WHO, 2008). In BC, as set out by the Child, Family and Community Service Act (1996), once an individual turns the age of 19 they are no longer considered a child; this corresponds with BC’s legal drinking

(27)

age and the age at which one can buy cigarettes (Tonkin, 2005). Of course it is recognized that many adolescence experiment with drinking and smoking at a much earlier age. The Canadian Paediatric Society (CPS) uses a definition that is less age focused. It defines adolescence as beginning at the onset of physiologically normal puberty, and ends when adult identity and behaviour are established. According to the CPS this period of development corresponds roughly to the ages of 10 to 19 years, which is consistent with the WHO’s definition. In addition, for administrative and research purposes, it is useful to define adolescence by middle and high school years as this group faces many similar challenges and issues within society (CPS, 2008). At 19 most people have graduated from high school.

Due to data constraints a lower limit of age 12 or grade 7 and an upper limit of 19 or grade 12 were used in this study. In certain cases data did not meet this range in its entirety and so only a subset of this group is represented. For example the juvenile crime rate is only available from BC Stats for ages 12-17.

1.5.2 Adolescent Health and Wellness

Adolescence is recognized as a period of increased desire for independence,

experimentation, and an aspiration to discover the world and is characterized by change, growth and risk (Stangler & Zweig, 2008). Many adolescents will experiment with risky behaviours, including unsafe substance use and sexual experimentation exposing

themselves to health risks, while some continue on such a path of high risk behaviour well into adulthood thereby incrementally increasing their exposure to health risks (CCSD, 2008).

(28)

Adolescents are often considered vulnerable or at risk of poor health and wellness outcomes due to increased independence from parents and social protectors along with increased peer influences (Chassin et al., 1988). Contemporary focus of much North American research has pointed out the negative elements of adolescence, such as unhappiness, anxiety, depression and harmful behaviours. This creates a view of adolescents as placing a strain on society (Moore et al., 2004; Scales, 2001). It is

important that policy and decision makers know what types of risky health behaviours are prevalent, but research should also examine positive elements of adolescent health and wellness (Moore et al., 2004). Positive health outcomes are associated with a beneficial transition into adulthood as well as an enhancement of the present health and wellness of adolescents (Stagner & Zweig, 2008).

1.5.3 Health Indices: An Introduction

Indices are created by combining indicators. In order to understand how adolescent health and wellness varies by region it is possible to construct an index using techniques that draw on past studies of deprivation and health index construction. Over the last 20 years there have been a large number of studies that focus on indices of material deprivation and how they vary by area. Socio-economic deprivation indices have become a common tool in developed nations and are frequently used in policy making (Gatrell, 2002). These indices rank geographic areas using composite indices made of multiple indicators of socio-economic characteristics (i.e. Bell et al., 2007a; Frohlich & Mustard, 1996). Such indices have been shown to be a useful tool in studies of population health (Gatrell, 2002). Also, there have been a number of health, well-being and wellness indices

(29)

produced in order to examine the health of populations (i.e. Bobbit et al., 2005; Bradshaw and Richardson, 2009; Bringsen et al., 2009; Bradshaw et al., 2009; Foster & Keller, 2007).

Indicators reflect the domains (dimensions or categories) of health defined by the study. Each domain contains component indicators (Bringsen et al., 2009; Bradshaw et al., 2009). In choosing appropriate indicators, domains are used to provide a framework for health and vary by the population group under study. The principles that should be aspired to for each indicator is that the indicator should: relate to the domain, measure a major feature in health and wellness, be up to date, be capable of being updated on a regular basis, be statistically robust and be available at the level under study (Bradshaw et al., 2009; Pencheon, 2008). Keeping with these principles, it is clear that in order to gain a comprehensive representation of adolescent health and wellness appropriate indicators must be selected (Moore et al., 2004).

Combing the indicators can be done in many ways from a simple additive way to using more complex multivariate statistical methods. A score for an area is then calculated (Bell et al., 2007a; Bringsen et al., 2008; Gatrell, 2002). A spatial method of multi-criteria analysis embedded in GIS has been implemented to construct a socio-economic deprivation index (Bell et al., 2007a). No matter the method used in index construction, there are two fundamental questions that must be addressed. The first is selecting which of all possible indicators should be included in the index. The second is deciding on what weights should be assigned to each indicator (Frohlich & Mustard, 1996).

(30)

The approach I am using is novel in that it: 1) incorporates a positive health and wellness perspective when examining adolescent health and wellness, and 2) utilizes a spatial MCA in the creation of the index and to observe and analyze the findings.

1.6 Research Framework and Structure of the Thesis

This study asks stakeholders, employed in the public sector, with expertise in adolescence and/or health and wellness to consider what they feel are the most important indicators of adolescent health and wellness and to identify each indicator’s relative importance. A spatial MCA then is carried out in order to create a BC Adolescent Health and Wellness Index (BCAHWI).

This thesis has been structured into two distinct phases. The goal of the first phase is to identify the key indicators of adolescent health and wellness in BC and their relative weights of importance. This is done by conducting a three-round Delphi study of a panel of people with expertise in the field of adolescence and/or health and wellness. This phase is addressed in Chapter 2 of this thesis.

The second phase of this research uses the indicators selected from the former study and their associated weights in order to construct a BCAHWI using the spatial MCA

technique for order preference to the ideal solution (TOPSIS) method. This phase is addressed in Chapter 3 of this thesis.

(31)

A summary of the research, conclusions and recommendations for future studies is presented in Chapter 4.

(32)

2. TOWARDS A REGIONAL ADOLESCENT HEALTH AND

WELLNESS INDEX USING AVAILABLE DATA IN BRITISH

COLUMBIA, CANADA

2.1 Abstract

This research seeks to identify key indicators of adolescent health and wellness in British Columbia (BC) available from secondary sources. This information will then be used as a basis for construction of an index of adolescent health and wellness for the Province. A three-round Delphi study was utilized to examine what a panel of expertise feel are the most influential indicators of adolescent health and wellness. A review of available data was undertaken to identify what data could represent the indicators identified. The Delphi study identified 27 indicators as most influential in measuring adolescent health and wellness in BC. After reviewing the available data from secondary sources, 24 indicators were considered appropriate for inclusion in an adolescent health and wellness index.

2.2 Introduction

In order to measure adolescent health and wellness in BC it is important to identify indicators that are relevant to adolescents living in the Province, since what defines this period of human growth can vary regionally, culturally and individually. To date it has not been established what indicators are of most importance to researchers, service providers and decision and policy makers within BC. One objective of this study is to identify a set of potential indicators of adolescent health and wellness and determine each indicator’s relative importance by collaborating with a mixture of researchers, service providers, and decision and policy makers, in order to create a composite index of adolescent health and wellness for BC. The indicators and index values can then be

(33)

mapped using GIS as a visualization tool to examine the pattern of adolescent health and wellness in BC (Jankowski & Nyerges, 2001).

By using collaboration with stakeholders with expertise in the index construction, local knowledge can be utilized to identify issues that are specific to the Province (Bell et al., 2007a; Boyle & Torrance, 1984); thus, enhancing the validity of research and joining together people with “diverse skills, knowledge, expertise and sensitivities” (Israel et al., 1998). Together with indicator value weights, derived from a panel of expertise’s

opinions, the selected indicators are used to compute a ranking of the HSDAs in BC.

2.3 Background

In order to establish a composite index of adolescent health and wellness it is important to pick the most influential indicators to populate the index and to choose appropriate weights for them. In order to examine these issues the appropriate literature on building composite indices was reviewed. The findings of this are summarized in Section 2.3.1. Some limitations of current indices are addressed in Section 2.3.2. Finally an introduction to the Delphi technique is presented in Section 2.3.3 as a method of selecting and

weighting the relative importance of each indicator.

2.3.1 Building Composite Indices

The health and wellness of a population is made up of many components. Measuring it is a multi-attribute and evaluative issue. Multiple indicators are needed to address questions of spatial patterns of adolescent health and wellness. Creating an index becomes part of a decision process that includes the following questions: 1) which indicators should be

(34)

selected for evaluation criteria, 2) who should select the indicators (Jankowski &

Nyerges, 2001), and, 3) what is the relative importance that each indicator should have in the index (Frohlich & Mustard, 1996; Malczewski, 1999)? Selecting indicators and then combining them into a single scale are two important steps; it has been established that when building health indices, the indicators used must fall within different levels of human health and the procedure must then produce an index of cardinal values (Boyle & Torrance, 1984). It is the position of this study that deriving weights from a panel of expertise has information that can aid in an exploratory analysis of adolescent health and wellness in BC.

There have been many composite indices developed to study health and wellness of the human population at various regions and scales. A selection of these is examined below.

The Foundation for Child Development created a Child and Youth Well-being Index, for the United States. It uses equal weights for seven domains (each comprised of an unequal number of indicators); this influences the meaning of the index because each of the items in a domain with a lower number of indicators has more impact on the resulting index score than a domain with a higher number of indicators (Moore et al., 2008). More data are included for adolescents than other age groups. As a result the index more accurately reflects adolescence than infancy, preschool or childhood (Mitic & Leadbeater, 2009). This index is used to track changes in child and youth well-being over time.

(35)

In a Swedish study, Bringsen et al. (2008) has developed a Salutogenic (positive factors that affect human health and wellness) Health Indicator Scale (SHIS) that is based on measuring health from a holistic view point. This scale was developed using primary data. The indicators fall into ten dimensions [domains] of human health (perceived stress, illness, energy, physical function, state of morale, psychosomatic function, expression of feelings, cognitive ability, social capacity and self-realization). The questions asked of participants had both positive and negative wording. An example of this is the indicator of sleep, which falls under the dimension of psychosomatic function, which is associated with positive wording, “slept well,” and negative wording, “had sleeping problems.” Both positive and negative wording reflect health as a continuum. A written

questionnaire was delivered in which participants were asked to rate how much they agreed with either of the opposing statements. The index was validated by examining the correlation of the indicators with self-rated health and self-rated sick leave. The level of measurement was the individual (Bringsen et al., 2008).

An index of wellness in BC was derived utilizing data from the CCHS 2005 in the BC Atlas of Wellness. The method used in developing the index was establishing areas that were statistically significantly higher or lower than the provincial average for 26 CCHS wellness indicators. The amounts of positive and negative scores were then summed to derive a net wellness score (Foster & Keller, 2007). The level of measurement is at the HSDA level.

(36)

A Child Wellbeing Index (CWI) (2009) was created exclusively for children in England (Bradshaw et al., 2009). This index was created for the small area level (the 32,482 Lower Super Output Areas). Of note is that it was restricted because data on children are largely collected by surveys which lack the robustness to be broken down to the small area level. It was undertaken due to the fact that during public consultation the British Government was called to produce separate indices for different groups of the population. This index includes the following seven domains: material wellbeing, health, education, crime, housing, environment, and children in need. Indicators were limited to data available at the small area units. The indicators reflect both a positive and negative impact on child well-being; examples are burglary rate and percentage of green space (Bradshaw et al., 2009).

GIS based Order Weighted Average (OWA) Multi-criteria Analysis (MCA) was used by Bell et al. (2007a) to construct a socio-economic deprivation index for the Vancouver Census Metropolitan Area. Both OWA and MCA are well known in spatial analysis but have been used little in social epidemiology (Bell et al., 2007a). Municipal Health

Officers were used in indicator selection and weighting. It was concluded from this study that local knowledge can play an important role within quantitative analysis in public health research (Bell et al., 2007a).

Bradshaw and Richardson (2009) created an Index of Child Well-being for comparison of the 27 countries of the European Union, Norway and Iceland. It uses 43 indicators

(37)

then used to classify the indicators (health, subjective well-being, personal relationships, material resources, education, behaviour and risks, and housing and the environment). In this index the child rather than the parent, family or household was used as the focus of analysis. Indicators of present well-being are given priority on the grounds that the present life stage is valuable in its own terms. The inclusion of indicators that represent what the children think and feel reflects accordance with the United Nations Convention on the Rights of the Child (1990) determination, that “the primary consideration in all actions concerning children must be their best interest and that their views must be taken into account.” Average z scores were used to compute the index score. Bradshaw and Richardson (2009) used equal weights citing that there is no theoretical justification for unequal weights but also stated that there is no theoretical justification for equal weights.

Bobbit et al. (2005) developed a County Level Index of Well-being for Larimer County, Colorado. This index attempts to include both “strength and problem” indicators. In order to weight the indicators they used stanine (standard nine) scores (data are

standardized and then the z-scores are used to create stanine scores which range from 1 – 9). They found that data availability, reliability and validation were the greatest

challenges to constructing an index of this type (Bobbit et al., 2005). It was stated that the reliability of the index was high as they used credible sources to populate it.

(38)

2.3.2 Limitations of Current Indices

There are three limitations that are present in the majority of current health indices:

1) Most widely used indices created to date measure negative outcomes and behaviours. The widespread use of socio-economic deprivation indices illustrates that current health measurement tends to measure aspects of ill or poor health rather than health in general (Bringsen et al., 2008). These measurements are needed and useful but only provide part of the picture. In order to complement the measurement tools that examine negative impacts on health, health and wellness indices can be created in order to meet both a health and wellness perspective thereby using indicators that measure health status and considers subjective experience (Bringsen et al., 2008).

2) Different groups within the population face varying health issues which are not

reflected in many indices that measure the population as a whole. Adolescents are part of a unique cohort that has many distinct health behaviours. Creating a composite index that combines adolescents with the population as a whole, or younger populations, therefore may not reflect the true health and wellness of this group.

3) Expert opinion is often overlooked in index construction. This can leave out local knowledge that maximizes the opinions and expertise of key stakeholders and public health agents (Bell et al., 2007a). Although not common, there are examples of

stakeholder based index construction created by Bell et al. (2007a) and Jarman (1984); these are both used as deprivation indices.

(39)

2.3.3 The Delphi Technique

In order to address the questions of: 1) what indicators should be included in the index and, 2) what are the relative weights of each indicator, the Delphi technique was used. The Delphi technique is considered a useful tool when evidence is limited, where

subjective evidence can be of benefit (Vernon, 2009) and when little factual data exist or there is uncertainty surrounding the topic being investigated (Malczewski, 1999; Syed et al., 2009; Vernon, 2009). This technique is fitting to the questions at hand as it has been established that measuring health and wellness of humans is an “inexact and changing science” (Millar & Hull, 1997). It has been deemed a strong methodology for answering questions based on expertise from a panel of selected participants (Okoli & Pawlowski, 2004; Malczewski, 1999).

The Delphi technique was originally developed by the RAND (Research and

Development) Corporation in order to help structure communication and decision making around complex issues (Beech 1999, Okoli & Pawlowski, 2004; RAND, 2009; Syed et al., 2009; Vernon, 2009). It uses a series of questionnaire rounds to illicit and distribute information to and from a panel of members with expertise on the topic under study (Beech, 1999; Hanafin & Brooks, 2005). The technique has the advantage of being a relatively quick and inexpensive way of gathering expert opinions thereby bringing together a wide range of experience, with the goal to reach consensus (Beech, 1999; Hanafin & Brooks, 2005; Normand et al., 1998; Okoli & Pawlowski, 2004; Syed et al., 2009; Vernon, 2009).

(40)

Although there is flexibility and variation in how the Delphi technique is implemented there are four characteristics that typify a Delphi study: 1) a panel of expertise (defined by the context of the study), 2) anonymity, 3) iterations with controlled feedback (continuation of surveys past one round that are aggregated and analyzed with the feedback controlled), and 4) statistical group response (statistically summarizing each item under consideration), which allows panel members to compare their responses to that of the group collective in anonymity (Normand et al., 1998; Ospina et al., 2007; Vernon, 2009). Several variations of the Delphi technique exist; one type is a ranking Delphi which elicits a weight based on the comparison of variables (Normand et al., 1998; Okoli & Pawloski, 2004). This lends itself to answering the question of what relative importance each indicator should have in the index (Frohlich & Mustard, 1996; Malczewski, 1999).

There are many benefits to using the Delphi technique: 1) it is able to bring together the opinions of a collection of expertise, 2) it is flexible in its design, 3) the anonymity of this technique can minimize the dominance of one or more participants or group pressure allowing freedom of expression and equality of opinions (Dalkey, 1969; Normand et al., 1998; Vernon, 2009), and 4) the outcomes of the study are derived from information obtained from participants in a collaborative process (Jankowski & Nyerges, 2001).

(41)

2.4 Methods

A three round Delphi study was conducted from July, 2009 to December, 2009. The study was reviewed and approved by the Human Research Ethics Board at the University of Victoria (Certificate # 09-194) (Appendix A). An outline of the participant sample of the Delphi study is presented in Section 2.4.1. The methods used in this study are

presented in Section 2.4.2. Section 2.4.3 summarizes the means to which reliability of the results were investigated. Section 2.4.4 discusses acquiring secondary data to populate the index.

2.4.1 Delphi Study Participant Sample

The panel was selected with the goal of utilizing local knowledge to obtain age

appropriate indicators that are place-specific to issues pertaining to adolescents in BC. The Delphi technique does not require a statistical sample but rather that the panel must be qualified in that they have a deep understanding of the issue under investigation (Okoli & Pawlowski, 2004). The foremost criterion is that participants are qualified and

knowledgeable of issues pertaining to adolescence and/or health and wellness. It is therefore deemed that the individual is able to identify factors that are indicative of adolescent health and wellness based on this experience and expertise. A broad definition of a person with expertise was utilized in order to obtain a range of experience and background within the panel. The definition of a person with expertise for the purpose of this study is: An individual who is experienced and qualified in the field of youth and/or health and employed in the public sector within the Province of BC and has been

(42)

expertise may be drawn from practice as a decision maker, researcher or service provider (Hanafin & Brooks, 2005). This study uses the rationale that professional knowledge, which may also include privileged information, places one in an expert position (Vernon, 2009). This position is utilized to identify indicators of adolescent health and wellness specific to BC.

In order to get a large breadth of knowledge it was decided to obtain participants with various experiences. A heterogeneous panel (employed in several different areas of the public sector) was considered advantageous in order to gain a wider understanding of adolescent health and wellness than could be addressed with a homogeneous panel (Vernon, 2009). In group decision making, heterogeneous groups have been found to be more creative than homogeneous ones (Okoli & Pawlowski, 2004). There is no

prescribed guideline to the number of members in a panel when using the Delphi technique (Vernon, 2009). From the experience of past studies 10 – 18 participants on a panel is the recommended number (Okoli & Pawlowski, 2004) and this range was decided as the goal for this study.

Twenty-one people were asked to participate in the study. Those asked to participate were from the Provincial Government of BC ministries of: education, children and family development, health services, healthy living and sport, aboriginal relations and

reconciliation, labour and citizen’s services and the Office of the Representative for Children and Youth. All deal with health and/ or adolescent issues. There were also participants working in the non-government public sector in the field of adolescent

(43)

community services or health promotion. Public sector employees are exposed to both professional and privileged knowledge and are employed to work towards the public good, and so were reasoned to be the most appropriate panel for this study.

Each participant was contacted by email and invited to participate in the study (Appendix B). Email was chosen as the medium because it is quick, cost effective and responses are sure to be legible (Hanafin & Brooks, 2005; Okoli & Pawlowski, 2004; Syed ey al., 2009). Access to email may be seen as a biasing factor (Okoli & Pawlowski, 2004), but due to the fact that all participants are employed at agencies requiring access to email, it is not a source of bias. At this time the invited participants were also encouraged to clarify any questions or issues they had pertaining to the study. Of the 21 invited participants 19 accepted the invitation to participate and signed a Letter of Consent (Appendix C) (90% response rate). This exceeded the criterion of obtaining between 10 to18 participants but it was decided better to overshoot the participant population to account for possible attrition through subsequent rounds of surveys. Reminders were also sent out to those who had not responded by a certain date at various stages in the study. Reminders have been shown to aid in achieving a high response rate in a Delphi study (Syed et al., 2009).

(44)

2.4.2 Delphi Study Design

This study followed the steps of a ranking type Delphi (Okoli & Pawlowski, 2004). Generally these steps are: 1) brainstorming the important indicators, 2) narrowing down the list of indicators, and 3) providing a rank for the indicators (Okoli & Pawlowski, 2004). Figure 4 summarizes the steps taken in this research.

Figure 4: Delphi study administration process (adapted from Okoli & Pawlowski, 2004)

Fitting with the traditional Delphi technique it was deemed that an open-ended question was best to begin the brainstorming process of the Delphi study in round one (Vernon, 2009). Although some studies begin with a list of indicators upon which each panel member is asked to remark (Vernon, 2009), this was ruled out as the possibilities of

(45)

indicators are large and could be overwhelming to the study participants. For instance, in a 2008 report entitled “Population and Public Health Indicators for British Columbia” (2008), prepared by The Population and Public Health Evidence and Data Expert Group on behalf of The Provincial Health Services Authority (PHSA), there were 246 public health indicators identified for the general population of BC (PHSA, 2008).

When developing an index to be used in general applications it is up to the researcher to formulate the definition of health and wellness so that the indicators are selected in accordance with this definition (Boyle & Torrance, 1984). The following short

definitions of adolescence and health and wellness were sent out via email (Appendix D) to ensure that all participants considered the same definitions when selecting their

indicators and that a health and wellness perspective rather than an illness or ill health perspective would be considered in the indicator selection:

Definition of Adolescence: The Canadian Paediatric Society (CPS) defines adolescence as beginning at the onset of physiologically normal puberty, and ending when adult identity and behaviour are established. According to the CPS this period of development

corresponds roughly to the period between the ages of 10 and 19 years, which is consistent with the World Health Organization’s definition (WHO, 2008). For

administrative and research purposes, it can be useful to define adolescence by middle and high school years as this group faces many similar challenges and issues within society (CPS, 2008).

(46)

Definition of Health and Wellness: Increasingly, focus in population health studies has taken a wellness perspective rather than an illness perspective (Bringsen et al., 2009; Foster & Keller, 2007; PHAC, 2008). In 1948, the WHO defined health as “a state of complete physical, mental and social well-being and not merely the absence of disease or infirmity” (WHO, 1948). This began a societal shift to look at health from a holistic view point rather than a purely physiological one. Wellness definitions have varied within the literature but they tend to measure states of positive health on a continuum and from a holistic view point (Miller & Foster, 2010). It is the goal of this study to draw on the WHO’s holistic definition of health and use a wellness perspective.

In the second round participants were asked to validate the results of the first round (Appendix E). Then in the third (final) round each participant was asked to identify, using a Likert scale, which indicators they feel are “very influential” (5) to “not very

influential” (1). Participants could also answer “unsure” if they were uncertain of the importance of an indicator. (Appendix H). Indicators that had a group mean of 4 or more were deemed an influential indicator and retained for use in the index. This value was decided as the cut off a priori.The aggregate response scores are used in the next phase of this study (Chapter 3) to weight the indicators.

There has not been an established method of what value indicates consensus or the number of rounds that should be used in a Delphi study (Ospina et al., 2007; Vernon, 2009). The study ended after three rounds in order to keep the length of the study manageable and to reduce possibility of attrition.

(47)

2.4.3 Inter-Rater Reliability

Statistical techniques can help increase confidence in stakeholder responses by addressing inter-rater agreement (Bell et al., 2007b). In a past study Cohen’s kappa statistic was used to increase the confidence in the level of agreement of responses (Bell et al., 2007b). However, Cohen’s kappa is best for a sample of only 2 raters. Fleiss’ kappa can be used when the sample size is greater than 2 but Fleiss’ kappa is limited to nominal data. Cronbach’s alpha is another commonly used statistic except it is not appropriate for measures of rater agreement but rather it is in measuring the reliability of aggregate scales (Hayes & Krippendorff, 2007). In order to examine the agreement of the panel of expertise with ordinal data with over 2 raters, Krippendorff’s alpha and

Intraclass correlation (ICC) were employed (Field, 2005; Hayes & Krippendorff, 2007).

2.4.4 Secondary Data Acquisition

After indicators have been selected by the participants, available data were then acquired from various data sources in BC in order to assemble the index. A review of various BC data sources was undertaken in order to ensure data were available and met certain criteria. This is addressed further below.

2.5 Analysis

Analysis of the three round Delphi study is presented in the following three Sections: 2.5.1, 2.5.2 and 2.5.4. The inter-rater reliability analysis conducted at the conclusion of round three is presented in Section 2.5.5. Furthermore, reviews of available data are presented in Section 2.5.6; a preliminary review is also presented in Section 2.5.3. Acquiring the data is discussed in Section 2.5.7 and analysis of the geographic variation

(48)

of each indicator is presented in Section 2.5.8. A brief summary of each indicator is included in Section 2.5.9.

2.5.1 Round One

In early July, 2009 (with the exception of one later acceptance) the participants were sent the first round of questionnaires (Appendix E). The first round questionnaire was divided into two sections. Section 1 presented a short series of questions pertaining to the

participant’s background and experience. Section 2 asked the panellists to identify up to 12 indicators that they felt are influential in measuring adolescent health and wellness in BC. A follow-up question asked for a brief reason why they felt that the indicator should be included in the list. This explanation was used in understanding and consolidating the various indicators and aided in classifying the indicators into domains

[categories/themes] (Okoli & Pawlowski, 2004).

Of the 19 participants who completed the Letter of Consent, 14 participants completed the round one questionnaire by September, 2009. The attrition rate was 26%. This was deemed acceptable as the 14 participants fell within the desired 10-18 Delphi participant sample size. The mean number of years each participant worked in their capacity in the field of adolescence and/or health was 15.6, with a maximum of 32 and a minimum of 6. There was a standard deviation of 8.5 years. Figure 5 shows the breakdown of the self identified positions of the participant panel.

(49)

Figure 5: Participant panel by self identified position

Schmidt’s (1997) guidelines for analysis of a Delphi study were utilized in analyzing the second section of round one. The first step is that the responses are consolidated into a single list. When several different terms were used for what appears to be the same issue, the terms were listed together and one consolidated description of the indicator was used (Hasson et al., 2000; Schmidt, 1997).

To reduce the complexity of the total indicator list, the indicators were divided into natural groupings by the researcher based on information derived from participants (Boyle & Torrance, 1984). Results were consolidated by the reseacher, duplicates were removed and the terminology was unified (Okoli & Pawlowski, 2004). The list of indicators was compacted by aggregating items that appear to have high similarity and then categorized based on knowledge from literature and thematic analysis (Aronson, 1994; Malczewski, 1999). These groupings are commonly referred to as domains in health literature. Domains vary by the population group under study.

50% 14% 29% 7% Decision Maker Researcher Service Provider Other

(50)

Thematic analysis refers to the iterative process of recovering identifiable themes within data through the process of reading and summarization. It is an inductive process in that themes emerge from the data (Aronson, 1994; Fereday & Muir-Cochrane, 2006; van Manen, 1990). Related literature is then referred to in order to validate the themes. The analysis was undertaken manually by the author with consultation from the supervisors of this research. No software was utilized as the sample of participants was small and

produced a manageable amount of text, this outweighed the time and cost that is

associated with using a software to do the analysis (Bryne, 2001). Figure 6 summarizes the process undertaken in the analysis of the second section of the round one

questionnaire. Of note is that one questionnaire did not contribute any indicators due to the fact that it did not list any specific indicators1; but, by acknowledging a response the participant was retained in the panel.

Figure 6: Steps taken in data analysis Questionnaire 1, Section 2 (round one)

1_{This participant offered many thoughts on indicators and their pros and cons as well as examples in the form} of past experience rather than a list of indicators

Step 1: Compile a list of all Questionnaire 1, Section 2 indicators with attached explanation

Step 2: Group each indicator into a broad theme that emerges from the data inductively

(iteration 1) and repeat (iteration 2 & 3)

Step 3: Create sub-themes, refine themes at each of the iterations, which establish each

indicator based on indicator title and associated explanation (iteration 4-9). After 3

subsequent iterations yield the same result then it is deemed that nothing new will be learned from subsequent rounds.

(51)

After nine iterations of thematic analysis 62 indicators were identified, with 19 being identified by 3 or more participants. To minimize researcher intervention all indicators were kept in the list even if they occurred only once (Hasson et al., 2000). At this point a results report was sent back to the participants for validation (Schmidt, 1997). Table 2 presents the results of round one.

(52)

Table 2: Summary of identified indicators, consolidated responses and number of participants who identified the indicator

Domain Indicator Consolidated Responses

Number of Participants who Identified the Indicator (out of 14)

General Health Physical Activity physical activity/exercise amount/ involvement in

recreation or sporting activities/ physical development 8 Healthy Weight healthy weight/obesity/ appropriate weight & height

for age/ overweight rate of grade 10/12 students/ BMI

or waist circumference 7

Healthy Diet healthy diet/ healthy eating/ fruit and vegetable

consumption/food choices/ nutrition/ food security 6

Freedom From Chronic

Conditions free of chronic disease/ chronic conditions 2

Self Rated Health self rated health 2

Screen time time spent on computers or alone/ in front of display

terminals; TV, computer or games 2

Time for leisure extra curricular activities/ leisure time 2 Freedom From Allostatic

Load freedom from allostatic load 1

Good Nutritional Knowledge good nutritional knowledge 1

Health Literacy health literacy 1

Oral/ Dental Care oral/ dental care 1

Physical and Mental Health

Admission Contacts physical and mental health admission contacts 1

Relationships Family Connectedness family connectedness/ loving supportive family/

family functioning/ relationship with parents, guardian & siblings/ social development and sense of belonging (family/extended)

9 Positive Peer Influence positive peer relations/ influence of peer pressure

/positive and supportive peer groups/ social

development and sense of belonging (peers) 4 Residing Outside Of Parental

Home number of children living out of parental home/ youth custody rates 2

Single Parent Families single parent families 1

Number of Children Living With Lone Female Parents

number of children living with lone female parent

1

Child Welfare Contacts child welfare contacts 1

Supportive Relationships supportive relationships 1

Positive Adults Mentors positive adult mentors/ adults to talk to 1 Collaboration Between Peers,

Family, Schools and Communities

strong links between peers, family, schools and